# Paul Weismann

# Soft Law and its Importance in Ensuring Member States' Compliance with Union Law

https://doi.org/10.5771/9783748935865, am 02.04.2024, 08:01:05 Open Access – - https://www.nomos-elibrary.de/agb

Beiträge zum ausländischen öffentlichen Recht und Völkerrecht

Edited by

the Max Planck Society for the Advancement of Science represented by Prof. Dr. Armin von Bogdandy and Prof. Dr. Anne Peters

Volume 321

Paul Weismann

# Soft Law and its Importance in Ensuring Member States' Compliance with Union Law

Veröffentlicht mit Unterstützung des Austrian Science Fund (FWF): PUB 981-G und des Open Access-Publikationsfonds der Universität Salzburg.

**The Deutsche Nationalbibliothek** lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de

ISBN 978-3-7560-0008-1 (Print) 978-3-7489-3586-5 (ePDF)

#### **British Library Cataloguing-in-Publication Data**

A catalogue record for this book is available from the British Library.

ISBN 978-3-7560-0008-1 (Print) 978-3-7489-3586-5 (ePDF)

#### **Library of Congress Cataloging-in-Publication Data**

Weismann, Paul Soft Law and its Importance in Ensuring Member States' Compliance with Union Law Paul Weismann 714 pp. Includes bibliographic references.

ISBN 978-3-7560-0008-1 (Print) 978-3-7489-3586-5 (ePDF)

1st Edition 2024

© Paul Weismann

Published by Nomos Verlagsgesellschaft mbH & Co. KG Waldseestraße 3 – 5 | 76530 Baden-Baden www.nomos.de

Production of the printed version: Nomos Verlagsgesellschaft mbH & Co. KG Waldseestraße 3 – 5 | 76530 Baden-Baden


Onlineversion Nomos eLibrary

This work is licensed under the Creative Commons Attribution 4.0 International License.

# Preface

This work is based on my habilitation thesis which I completed at the Uni‐ versity of Salzburg. The continuous and thorough exchange of thoughts on all kinds of aspects relating to this text with my then academic supervisor, Professor *Stefan Griller*, was particularly helpful in all phases of my work on it. For this support I would like to express my sincere gratitude. The text was awarded the *Dr. Alois Mock-Wissenschaftspreis*, for which I am thankful to the eponymous foundation.

The original thesis was updated, in part shortened, in part extended, but the structure essentially remained the same. These efforts were facili‐ tated by the much-appreciated reviews submitted during the habilitation procedure and the review commissioned by the Austrian Science Fund FWF, which contained helpful instructions as to how to improve my arguments. My work on the thesis, and its revision respectively, was fur‐ thermore advanced by research visits at the Walter Hallstein-Institute for European Constitutional Law at the Humboldt University Berlin, at the Europa-Kolleg Hamburg and at the Amsterdam Centre for European Law and Governance at the University of Amsterdam. In this context, I would like to thank Professors *Matthias Ruffert*, *Markus Kotzur*, *Andreas Grimmel*, and *Christina Eckes*.

I am also grateful to the editors of the series 'Contributions on Compara‐ tive Public Law and International Law', Professors *Anne Peters* and *Armin von Bogdandy*, for having accepted my text for publication. Moreover, I would like to pay tribute to the team of Nomos for their support during the process of transforming the typescript into a book.

Also the generous financial support for the publication of this book by the following institutions is to be acknowledged: the Austrian Science Fund FWF and the Publication Fund of the library of the University of Salzburg. Without this aid, the publication of this book would not have been feasible.

> Paul Weismann Salzburg, April 2023

https://doi.org/10.5771/9783748935865, am 02.04.2024, 08:01:05 Open Access – - https://www.nomos-elibrary.de/agb

# Content Overview


*Content Overview*


# Table of Contents












https://doi.org/10.5771/9783748935865, am 02.04.2024, 08:01:05 Open Access – - https://www.nomos-elibrary.de/agb






vol volume WTO World Trade Organisation

# Table of Cases

#### *Court of Justice*


#### *Table of Cases*

Case 74/82 Commission of the European Communities v Ireland, ECLI:EU:C:1984:34


#### *Court of Justice: Opinions of the Advocates General*


### *Court of First Instance / General Court*


Case T-192/16 NF v European Council, ECLI:EU:T:2017:128


#### *Permanent Court of International Justice / International Court of Justice*


### *Other judicial bodies*

Arbitration Panel, Case Responsabilité de l'Allemagne à raison des dommages causés dans les colonies portugaises du sud de l'Afrique (sentence sur le principe de la responsabilité) (Portugal contre Allemagne) of 28 July 1928, Report of International Arbitral Awards II, 1028 (D.2.)

German Bundesverfassungsgericht, Case BverfGE 49, 89 (1978)

German Bundesverwaltungsgericht, Case 6 C 13/09 (2010)

German Bundesverfassungsgericht, Case 2 BvE 4/11 (2012)

Joint Board of Appeal, Case C v EBA, BoA-D-2022-01

https://doi.org/10.5771/9783748935865, am 02.04.2024, 08:01:05 Open Access – - https://www.nomos-elibrary.de/agb

# Table of Acts (Selection)

### *Council, European Parliament or European Parliament and Council (together)*


60


### *Commission*


#### *Other actors*


# I. INTRODUCTION

#### *1. Ambition and structure*

'Soft law' allows for the steering of human behaviour by legally non-binding rules. In spite of its – from a legal perspective – apocryphal character, it may dispose of a high authority resulting in high compliance-rates, that is to say in a large measure of effectiveness. Therefore, 'soft law' – unlike what its name suggests – is to be taken seriously, also from a legal point of view. It is created and applied within the framework of the various levels of public rule-making, that of national law, European Union (EU) law and public international law. In an attempt to 'abandon our nostalgia for a world of unequivocally binding law',<sup>1</sup> this book focuses on the creation of 'soft law' by the EU and its institutions and other bodies. It aims at improving the understanding of EU 'soft law' as a *quasi*-legal phenomenon in general and specifically of its importance in facilitating compliance with the law of the EU by its Member States (MS). Given the often considerable steering effect of 'soft law', the inherent discrepancy – legally non-binding acts shall lead to compliance with binding law – seems to be only a *prima facie* contradiction.

As a wide-spread tool in ensuring compliance with EU law, 'soft law' shall be tested against the EU primary law framework: To which extent do the Treaties allow for the adoption of 'soft law', what are its effects and its purposes, and how can the Court of Justice of the European Union (CJEU) review EU 'soft law'? On the basis of these findings, the use of 'soft law' in ensuring compliance with EU law by the MS shall be addressed. In various policy fields (eg banking supervision or energy policy), EU law provides for procedures which often prescribe the adoption of 'soft law' as a means of *persuading* MS to apply the relevant EU law. While some mechanisms only allow for 'soft law' acts to be adopted, others also envisage the (subsequent) adoption of law (that is to say: hard law as opposed to 'soft law') – in case persuasion does not work. Again other mechanisms only permit the adoption of hard law acts directed to the MS concerned.

<sup>1</sup> Trachtman, Direct Effect 656.

This work proffers a structural classification and enquires on the legal feasibility of these procedures, taking into account, *inter alia*, the most important principles enshrined in the Treaties, such as the principle of subsidiarity, the principle of proportionality, or the delicate institutional balance of the EU.

In conclusion, this thesis is about two main issues which are related to each other. It explores the competences of the EU to adopt 'soft law' and the effects, the purposes and the possibilities of judicial review of this body of rules more generally. On the basis of these findings, it then presents the way in which various EU actors – within the framework of different procedures and not rarely with the help of 'soft law' – try to ensure MS' compliance with EU law (compliance mechanisms). A profound understanding of the functioning and a thorough legal evaluation of these mechanisms – of which a selection shall be presented – is possible only on the basis of a more general legal account of EU 'soft law'. Therefore, the two main concerns of this work – EU 'soft law' and said compliance mechanisms – are dealt with consecutively. That way, the book ought to contribute to a better understanding of EU 'soft law' in general, but also to the disclosure of one important field of its application in practice (compliance mechanisms), thereby addressing some of the many legal concerns these issues raise.

Having a closer look, the book's structure is the following:

Part II explores the intriguing idea of 'soft law' beyond specific legal orders. Following a discussion of different concepts of this phenomenon which can be found in the literature, a definition of 'soft law' shall be provided which shall underlie the entire work: 'soft law' as a category of norms which are enacted by different bodies as an expression of public authority, and which are aimed at steering human behaviour in a legally non-binding way (1.).

This definition shall be complemented by a delimitation of 'soft law' from other categories of norms such as law, customs, morals or private rule-making, but also from non-normative output of public bodies. For the sake of stimulating the discussion, a related phenomenon shall be presented at this point: the effects of the law of the United Nations (UN) which is in conflict with EU fundamental rights as dealt with most famously in the *Kadi* cases on the one hand and, on the other hand, the effects of the law of the World Trade Organisation (WTO) in the EU legal order. It will turn out that the effects of these specific regimes in the given context are similar to those of 'soft law'. They are not considered binding – to the extent that they conflict with EU fundamental rights (in case of UN law) or to the extent that an individual/undertaking or a MS relies on them (in case of WTO law). This exemplifies the close proximity of law and 'soft law', and the resulting difficulties when it comes to distinguishing them from each other (2.).

After this overview, in Part III the emphasis shall be shifted to EU law or rather: EU 'soft law'. In an attempt to classify the various forms of EU 'soft law', and following an overview of the historical and current use of EU 'soft law' (1.), the actors concerned shall be depicted, that is both the originators and the addressees of EU 'soft law' (2.). When it comes to the competence of the EU to adopt 'soft law', it is unclear, not least from the case law of the CJEU, whether the principle of conferral applies in this context. In order to address this question, but also in order to get an idea of the structural meaning of 'soft law' in primary law, the Treaty provisions allowing for the adoption of recommendations and opinions as the standard expressions of EU 'soft law'<sup>2</sup> shall be looked at more closely. Here we can distinguish between general empowerments such as Article 292 of the Treaty on the Functioning of the European Union (TFEU) or authorisations only with regard to a specific task. But also the possibilities to adopt other forms of EU 'soft law' (such as guidelines or resolutions), as enshrined in the Treaties, shall be reflected upon. From these findings about the structure of 'soft law' competences in the Treaties, conclusions shall be drawn with respect to the (questionable) applicability of the principle of conferral (3.).

In a next step, the effects of EU 'soft law' shall be addressed, be they legal, factual or mixed (4.). A legal effect may be an obligation of the national authority addressed to *consider* 'soft law' or even to *provide the reasons for non-compliance*, which may again emanate from the legal order of the EU, eg the principle of sincere cooperation or the protection of individuals' legitimate expectations. Factual effects root in human nature, such as the propensity to follow an existing rule – irrespective of whether it is binding or not – rather than not to follow it. In this context, also the literary discus‐ sion on the phenomenon of 'nudging' shall be reflected upon. The effects of EU 'soft law', in the terminology applied here, are mixed if it is both legal and factual circumstances which facilitate compliance, for example the rule which allows the Council to deviate from a Commission proposal addressed to it (an act of EU 'soft law') only by a unanimous decision, instead of the qualified majority which usually applies in the underlying

<sup>2</sup> See Article 288 TFEU.

procedure.<sup>3</sup> Here the legal requirement, together with the factual difficulty to reach a unanimous decision, brings about an increased authority of 'soft law'. The subsequent chapter will shed light on the varying purposes of 'soft law', eg the preparation or even the substitution of law. These purposes shall be illustrated by various examples from legal practice (5.).

Eventually, the possibilities of judicial review of 'soft law' – as essentially determined by the CJEU – shall be fleshed out. True 'soft law' acts being excluded from an annulment procedure *qua* not being 'intended to produce legal effects vis-à-vis third parties',<sup>4</sup> a preliminary reference pursuant to Article 267 TFEU or a damages claim may be useful tools to achieve a review of EU 'soft law' by the Court (6.).

Following this account of legal questions regarding EU 'soft law' more generally, the focus shall then turn to the question how *individualised* EU 'soft law' – addressed to a certain MS or national authority – may be and is actually used to facilitate MS' compliance with EU law in the course of what was referred to above as compliance mechanisms. Part IV shall be dedicated to presenting such mechanisms. Following an introduction (1.), those compliance mechanisms laid down in primary law and a selection of those laid down in secondary law shall be depicted (2.). In terms of their scope we can distinguish between the general Treaty infringement procedure which allows the Commission (or a MS) to pursue all kinds of non-compliance with EU law on the part of a(nother) MS before the Court on the one hand, and special compliance mechanisms which are tailored to address non-compliance only in a specific field of EU law (eg transport law or deficit rules), on the other hand. On a different level, depending on whether the respective procedures provide for EU 'soft law' acts to be addressed to a certain MS, hard (no 'soft law' acts), mixed (both 'soft law' and legal acts) and soft mechanisms (only 'soft law' acts) can be discerned.

Finally, in Part V of this work, after a brief introduction (1.), these mechanisms shall be classified with a view to the EU actors involved (these are regularly administrative bodies like the Commission or European agen‐ cies), the policy fields concerned, their procedural structure, the different purposes of 'soft law' (in mixed and soft mechanisms) and the way in which the special compliance mechanisms deviate from the general one, that is the Treaty infringement procedure. Eventually, the question why each of the

<sup>3</sup> See Article 293 para 1 TFEU.

<sup>4</sup> Article 263 para 1 TFEU.

presented mechanisms is designed in its peculiar way shall be addressed (2.).

Thereafter, a legal analysis shall be proffered, tackling a variety of legal questions relating to our selection of compliance mechanisms (3.). At the beginning of this analysis, a material distinction shall be drawn between the implementation and the enforcement of EU law, as undertaken by EU ac‐ tors *vis-à-vis* the MS. The core rule of the former is Article 291 para 2 TFEU, the core rule of the latter are the provisions on the Treaty infringement pro‐ cedure. From these rules – but also with a view to the Treaties as a whole – the characteristics of implementation and enforcement under primary law shall be distilled, and subsequently each of the compliance mechanisms presented above shall be allocated to either category. As regards the com‐ pliance mechanisms laid down in secondary law, this allocation shall be coupled with a test of the adequacy of their respective Treaty base. Against the background of these findings, it shall then be scrutinised whether and, if so, in which way each individual compliance mechanism – or all of them in their entirety – affect(s) the EU's institutional balance. The basic assumption is that in the given context EU administrative bodies may only take decisions on the implementation of EU law, not on its enforcement. Decisions on the enforcement of EU law *vis-à-vis* the MS, taken by EU administrative bodies (eg European agencies), would negatively affect the (monopolistic) role of the CJEU (with an only supporting role of the Com‐ mission or a MS) under Articles 258–260 TFEU. In this context, also the role of the principles of subsidiarity and proportionality shall be discussed, particularly their (potential) function as promoters of a 'soft law' rather than a hard law intervention *vis-à-vis* the MS. Eventually, the diverging effects of 'soft law' provided for in the above compliance mechanisms and the means of judicial protection the MS may avail themselves of shall be analysed.

Part VI shall be a conclusion, summarising the main findings of this the‐ sis, explaining the resulting legal concerns and pointing at related research questions which this work could not cover.

#### *2. Methodology and embedding in legal scholarship*

In terms of methodology, it is to be remarked that this work is directed towards raising and answering legal questions. It is essentially dedicated to doctrinal legal research, that is to say to an analysis of current law and other authoritative acts, mainly EU law (and here in particular its primary and examples taken from secondary law), EU 'soft law', and the relevant case law of the CJEU. Aspects relating to other scientific fields, in particular political, psychological or economic aspects are referred to, on the basis of pertinent literature, only selectively, where this serves a better understanding of the legal questions underlying this work. When interpreting EU rules, ie when finding out about their meaning pursuant to the expressed will of the respective rule-maker, the traditional methods of legal interpretation, as developed further by the CJEU in the context of EU law, shall be applied. These traditional methods are the verbal, the systematical, the historical and the teleological interpretation.<sup>5</sup> While these methods in principle range on an equal footing, when interpreting EU law (and EU 'soft law', for that matter) the CJEU applies them in its distinctive, nuanced way, thereby taking account of the multi-linguality of EU law. This idiosyncratic approach includes, in particular, legal comparison and – as a variant of the teleological interpretation – the *effet utile*. 6 The distinction between law and 'soft law' brings with it an additional aspect of the mean‐ ing of EU rules as mentioned above, namely whether a concrete 'legal act' is actually legally binding or not. According to the CJEU, either qualification is to be made 'after an analysis, conducted to the requisite legal standard, of the wording, the content and the purpose of the [relevant act], as well as of the context of which it forms part'.<sup>7</sup> Thus, the interpretative approach to be taken here is essentially the same as when it comes to determining the content of EU rules.

In addition to that, the book will proffer a critical engagement with the relevant legal literature. EU 'soft law' has been addressed in legal scholar‐ ship in a variety of contributions, eg the monographs by *Senden*, *Knauff*, *Ştefan* or *Láncos* (in chronological order) or the edited volume of *Elianto‐*

<sup>5</sup> The CJEU essentially follows these methods, even if it may apply a slightly different terminology; see, with further references, case C-621/18 *Wightman*, para 47; similarly, but still expressed differently: case 26/62 *van Gend & Loos*, 12: 'the spirit, the general scheme and the wording of [the respective] provisions'; for the historical interpretation see eg case C-583/11P *Inuit*, para 59; case C-370/12 *Pringle*, para 135; case C-196/21 *SR*, paras 34–41; for the case of EU 'soft law' see case C-16/16 P *Belgium v Commission*, paras 33–38 and 50–52.

<sup>6</sup> See, *ex multis*, Lenaerts/Gutiérrez-Fonds, Law; for the aspect of multi-linguality see also McAuliffe, Language and Baaij, Fifty Years.

<sup>7</sup> Case C-16/16 P *Belgium v Commission*, para 37.

*nio* et al and *Láncos* et al.<sup>8</sup> These contributions, however, do not provide the strong focus on the appropriate legal bases in primary law of this 'soft law', which underlies this thesis. An elaborate account of the use of 'soft law' in the course of compliance mechanisms directed towards the MS – and an account of these compliance mechanisms more generally – cannot, to the best of my knowledge, be found in the literature. In this respect the book will break new ground. As a matter of course, there are scholarly accounts of the Treaty infringement procedure and other infringement procedures laid down in primary law – eg the contributions by *Gil Ibáñez*, *Andersen* or *Prete* (in chronological order)<sup>9</sup> – and there is literature on specific fields of EU law such as energy law or railway law which presents one or the other compliance mechanism. There is also literature on the increasing enforcement of EU law by EU bodies *vis-à-vis* private actors<sup>10</sup> – a matter which shall not be addressed in this work. However, the existing literature does not provide a thorough comparative analysis of these mechanisms, let alone an in-depth consideration of the role 'soft law' plays therein. This book is intended to make a contribution to filling this gap.

### *3. Some technicalities*

The citation of literature in this work largely follows the Oxford University Standard for the Citation of Legal Authorities (OSCOLA) in its 4 th edition. As regards the judgements and other output of the CJEU, it is in particular the indication of the European Case Law Identifier (ECLI) which will allow tracing them. In the footnotes, short citations will be used for both literature and case law.

References to other chapters of this work will indicate, in Roman numer‐ als, the relevant Part and, in Arabic numerals, the relevant (sub-)chapter. Where the reference and the referred (sub-)chapter are contained in one and the same Part, there shall be no indication of the Part.

Abbreviations used in this work are introduced when first used in the *main* text; otherwise, the list of abbreviations at the beginning of this book will serve to explain them.

<sup>8</sup> Senden, Soft Law; Knauff, Regelungsverbund; Ştefan, Soft Law; Láncos, Facets; Eliantonio/Korkea-aho/Ştefan, Soft Law; Láncos/Xanthoulis/Arroyo Jiménez, Legal Effects.

<sup>9</sup> Gil Ibáñez, Supervision; Andersen, Enforcement; Prete, Infringement.

<sup>10</sup> See eg Scholten/Luchtman, Enforcement; Scholten, Trend.

#### *I. INTRODUCTION*

Since the Lisbon reform, the Treaties refer to the Court of Justice of the European Union and the abbreviation CJEU is becoming more and more common. This abbreviation or simply the short term 'Court' are used in this book. These expressions should be understood broadly, so as to encompass the Court of Justice of the European Union and the Court of Justice as referred to in Article 19 para 1 of the Treaty on European Union (TEU), but also their respective predecessors under the pre-Lisbon Treaty versions. The risk of confusion appears to be minor, as the respective footnotes will regularly indicate the concrete cases which are referred to. From the case numbers or the ECLI, the concrete body – in particular the Court of Justice and the General Court (earlier: Court of First Instance) – can be identified.

For the sake of brevity, the term 'EU bodies' is normally used in a broad meaning, so as to encompass what the Treaties refer to as 'institutions, bodies, offices or agencies'.

# II. SOFT LAW: TERMINOLOGY AND LOCALISATION

### *1. Origins and concepts: a theoretical account of 'soft law'*

# 1.1. Origins, ideas and challenges: a *tour d'horizon*

1.1.1. Terminology, recognition and occurrence in practice: an approximation

'THE matter of jurisprudence is positive law: law, simply and strictly so called: or law set by political superiors to political inferiors. But positive law (or law, simply and strictly so called) is often confounded with objects to which it is related by *resemblance*, and with objects to which it is related in the way of *analogy*: with objects which are *also* signified, *properly* and *improperly*, by the large and vague expression *law*' (emphasis in original).<sup>11</sup>

These words of *Austin*, first published in 1832, address well the indetermi‐ nation of the notion of law, which is why they are quoted here. Whoever wants to use it (in an academic context) has to provide his or her own definition in order not to be misunderstood.12 The works on different theories and concepts of law – which fill libraries – bear witness of that fact. Nevertheless, the expediency of the term 'law' in common as well as in specific (scientific) parlance is largely undisputed, and its use – in a more or in a less conscious way – almost inescapable.<sup>13</sup>

'Soft law' describes a vague and malleable concept, as well.14 Having realised that one of its literal components, namely the term 'law', itself is

<sup>11</sup> Austin, Province 18. To avoid misunderstandings which may result from the above quotation, it ought to be stressed that *Austin* understood 'law set by political superiors to political inferiors' as a synonym of 'law, simply and strictly so called', that is positive law; see also Rumble, Positivism 991.

<sup>12</sup> See eg Fastenrath, Normativity 331; Peters, Typology 411; Rill, Fragen 1.

<sup>13</sup> This terminology has existed for many centuries (in various languages), before a truly scientific approach towards law had started only in the 12th century (AD); see Arndt, Sinn 32.

<sup>14</sup> See eg Bast, Handlungsformen 515 f; Shelton, Introduction 2; von Arnauld, Völker‐ recht, para 282; see also *Zeitler*, who pointedly describes 'soft law' as 'geräumige Schublade für alle rechtsähnlichen Erscheinungen' [spacious drawer for all law-like

unclear, this finding hardly comes as a surprise. Considering other (funda‐ mental) ideas used in legal scholarship, 'law' and 'soft law', however, do not appear to be exceptional in this respect. The commonly used terms 'norm', 'separation of powers' or 'accountability' – to name just examples – as such do not convey an entirely clear concept, either. A norm can be written or unwritten, national or international, legal, customary, or moral etc. The separation of powers conveys a picture of State powers being shared between certain branches, normally the legislative, the executive and the judiciary. The term as such, first coined by *Locke* and *Montesquieu*, does not, however, say anything about *how* these powers are (to be) shared. Accountability again can be used as a synonym for responsibility, but can also be used to describe the mechanisms set in place for ensuring this re‐ sponsibility which again may consist of anything between light supervision and strict control. These terms or concepts (intentionally) convey a lot of different meanings, each of which requires further specification when dealt with on a scientific level. Thus, the ubiquity of terminological vagueness in legal discourse in principle is neither new nor inexpedient.<sup>15</sup>

The concept of 'soft law' may be considered unnecessary by those who think that the phenomenon usually addressed by it can be covered entirely by the traditional distinction between law and non-law. <sup>16</sup> This is a concep‐ tual critique. But those who agree that, in whichever legal order, there is a body of (legally non-binding) sovereign rules which needs to be catego‐ rically distinguished from other legally non-binding output of this very sovereign should accept the term for lack of an apparent better alternative.<sup>17</sup> This is why, in my view, the terminological dissatisfaction with the word 'soft law' – which is applied in order to grasp an actual phenomenon, not in order to complicate scholarly terminology – is unjustified. It is a 'trendy

phenomena]; Zeitler, Entwicklungen 1400; for further ascriptions see references in Ştefan, Soft Law 7 f.

<sup>15</sup> For the (possible) expediency of vague terms not only in everyday parlance but also in legal language see Jakab/Kirchmair, Unterscheidung 354.

<sup>16</sup> See eg Klabbers, International Law 38: 'misleading and unhelpful'.

<sup>17</sup> For an attempt to establish an alternative terminology – 'informal law(making)' – see the legal/political science anthology by Pauwelyn/Wessel/Wouters, Informal International Lawmaking; for a definition of the term see Pauwelyn/Wessel/Wouters, Introduction 1–3; for the overlap with the term 'soft law' see Flückiger, Soft Law 409; see also the terminologies of Arndt, Sinn 155: 'alternative Steuerungsinstrumente' [alternative steering instruments] (see 1.2. below) and Martin/Tourard/Loquin/Ravil‐ lon, Chronicle 94: 'supple and blur law'.

phrase',<sup>18</sup> it is true, but in a good art. It is widely used and it immediately conveys an idea of what it says, namely: something 'less' than law.<sup>19</sup> Thus, it may be described – in *Senden's* words – as 'a maybe not perfect, but at least reasonably satisfactory umbrella concept'.<sup>20</sup> Whoever wants to use the term in a *scientific context* first needs to define more closely what he or she understands with 'soft law', ie to explain his or her concept of 'soft law'.<sup>21</sup> This explanation the author will provide below (1.3.4.), and up until then the term shall be understood in its very general meaning just referred to: as describing norms which are something less than law.

Due to the multifacetedness of non-legal regulation, fleshing out one's understanding of 'soft law' is not an easy task.<sup>22</sup> While there are diverging concepts of 'soft law' in place (more often than not, however, even in schol‐ arly literature it is dealt with only cursorily), the first one purportedly<sup>23</sup> stemming from *McNair*, <sup>24</sup> the term as such appears to be the dominant designation for legally non-binding acts.25 In a legal context, it reputedly came in more widespread use in the 1970s.26 A specificity of this term is that it makes use of a fundamental and advanced notion: law – and modifies its meaning by prefixing an adjective: soft. This combination makes it a

<sup>18</sup> Brownlie, Extent 66. It is to be noted that *Brownlie* uses this term in the context of *lex ferenda*, which does not conform to the understanding of 'soft law' applied here. See d'Aspremont, Pluralization 194, with regard to a certain affinity in recent scholarship to term certain norms, institutions, processes or other phenomena 'soft'; Hofmann/Rowe/Türk, Administrative Law 567, to take another example, have descri‐ bed the term as an 'over-simplified (and arguably popular) notion'; Weber, Dichoto‐ my 11, argues that 'the term soft law is now acknowledged as valuable notion'. For the gradual replacement of the term 'quasi-legislation' by 'soft law' in British legal scholarship since the 1980s see Rawlings, Soft law 220. On the mystery the term 'soft law' allegedly carries: Arndt, Sinn 89.

<sup>19</sup> Sceptically as regards an undifferentiated use of the term: Arndt, Sinn 43; against a too restrictive approach: Wellens/Borchardt, Soft Law 273.

<sup>20</sup> Senden, Soft Law 110.

<sup>21</sup> See also Terpan, Soft Law (2013) 5.

<sup>22</sup> With regard to the difficulty to make generally applicable statements on 'soft law' – a fact which strongly affects its definability – see eg Walter, Soft Law 28.

<sup>23</sup> See, each with further references, Hillgenberg, Look 500; Wellens/Borchardt, Soft Law 268 (also on the role of *René-Jean Dupuy*).

<sup>24</sup> McNair, Functions, in particular 110 ff; assuming that with the term 'soft law' *McNair* was actually referring only to acts (still) constituting a *lex ferenda*: d'Aspremont, Softness 1081 (fn 35), with further references; see also Jennings, Lawyer 515 f.

<sup>25</sup> Similarly: Arndt, Sinn 90.

<sup>26</sup> See Arndt, Sinn 36 f; Ştefan, Soft Law 8, both with further references.

catchy word, 'very revealing *precisely because* it is a contradiction in terms' (emphasis added), as *Hillgenberg* put it.<sup>27</sup>

The recognition of 'soft law' (as opposed to law and other categories of norms) requires the consideration of a variety of factors which are 'fluid, cumulative, and interlocking'.<sup>28</sup> It reflects on traditional methods of legal interpretation in an attempt to find out about the (real or at least the demonstrated) will of the rule-makers (in particular the wording of the act and its systematic/contextual assessment<sup>29</sup>), but may also encompass procedural questions such as: Have the procedural requirements – the form and the forum of conclusion – for law-making been met?<sup>30</sup> The assumed importance of the subject matter, on the contrary, can regularly not serve as an indicator, as it tends to be more confusing than enlightening.<sup>31</sup> Often it is

<sup>27</sup> Hillgenberg, Look 500; von Bogdandy/Arndt/Bast, Instruments 111 call it a 'provoca‐ tion' for a traditional concept of law.

<sup>28</sup> Chinkin, Development 37.

<sup>29</sup> For example: Does the originator use the term 'should' or 'shall'? Sceptically: Co‐ man-Kund/Andone, Instruments 182; see also Andone/Greco, Burden 92; Dickschen, Empfehlungen 124; Ruiter/Wessel, Nature 178. The perspectives on the use of the terms 'should'/'shall' may differ, though; see eg the Opinion of the European Econo‐ mic and Social Committee on the proposals for the adoption of Regulation 1092/2010 and Regulations 1093–1095/2010, 2010/C 339/08, 4.2.3.: 'The use of "should" means that these recommendations are more or less compulsory'. With regard to EU law more generally, *Hofmann*, *Rowe* and *Türk* generally identify a decrease of 'mandatory formulations, [EU bodies] preferring a style appearing to aid and persuade'; Hof‐ mann/Rowe/Türk, Administrative Law 566. For other soft phrases see Weil, Norma‐ tivity 414: 'seek to', 'make efforts to', 'promote', 'avoid', 'examine with understanding', 'act as swiftly as possible', 'take all due steps with a view to'; addressing this issue in the context of international accords on environmental and on migration matters: Weismann, Bestimmung 390 ff; with regard to the term 'may' (used in an Association Agreement) see case C-581/11P *Mugraby*, paras 70 f; for the systematic/contextual assessment see Wellens/Borchardt, Soft Law 277–279, with respect to public interna‐ tional law; with regard to EU law see case T-721/14 *Belgium v Commission*, para 18, with many references to the CJEU's case law.

<sup>30</sup> See Wellens/Borchardt, Soft Law 300 f. For the relatively low formal requirements for treaties according to the VCLT 1969 see, eg, its Article 11 entitled 'means of expressing consent to be bound by a treaty'; see Klabbers, Courts 225 and 227 f, with regard to the pertinent case law of the ICJ; see also Pauwelyn, International Law 148 f and 151, who argues that at the level of public international law the legitimacy of law on the one hand and 'soft law' on the other hand may in some cases not differ greatly from each other; discussing the applicability of the VCLT – *per analogiam* – to international 'soft law': Seidl-Hohenveldern, Soft Law 224.

<sup>31</sup> See Chinkin, Development 40; Terpan, Soft Law (2015) 89.

the issues considered important which are regulated in a legally non-bind‐ ing way.<sup>32</sup> For the specific case of public international law see below.

With a view to distinguishing legally binding from legally non-binding international accords, *Klabbers* has proclaimed a 'presumption of legal force' – thereby repudiating the presumption of legal non-bindingness of international agreements proposed by others<sup>33</sup> – which can, as a matter of course, be rebutted.<sup>34</sup> In his view, this presumption is reflected in the Inter‐ national Court of Justice's (ICJ's),<sup>35</sup> but also in the CJEU's jurisprudence.

Having talked about terminological specificities of 'soft law' and the difficulties of its recognition, let us not forget to mention the practical occurrence and the reasons for its adoption, respectively. While it is true that legally non-binding but still authoritative rules in general can be traced far back in legal history – *Klabbers* describes them as 'a phenomenon of all times'36 –, the more widespread use and its systematic appraisal as something related to but not yet law can be perceived only much later. In the specific case of public international law, the broader recognition of the existence of acts being something less than a 'perfect legal act',37 goes back to the 19th century.38 The pertinent scholarly discussion fully unfolded in the course of the 20th century.39 Today its use is more common in some fields of public international law – for example environment, human rights,

<sup>32</sup> See Knauff, Regelungsverbund 259 f.

<sup>33</sup> First of all: Fawcett, Character 386 f; for the adherents of this view see references in Klabbers, Courts 224.

<sup>34</sup> See Klabbers, Courts 224 f; see already Klabbers, Instruments 1019 ff.

<sup>35</sup> This presumption may be facilitated by the general principle of public international law 'pacta sunt servanda' (see also Article 26 of the VCLT 1969).

<sup>36</sup> Klabbers, Courts 223. Take the *senatus consultum* (the Senate's advice) in Ancient Rome as an example, which was legally non-binding, but nevertheless held high au‐ thority, especially in times of the Roman Republic; see Gehrke/Schneider, Geschichte 504; for the mere *auctoritas* of the Senate of Ancient Rome (as opposed to *potestas*) see Goldmann, Gewalt 349, with further references.

<sup>37</sup> Tammes, Decisions 285.

<sup>38</sup> See Klabbers, Courts 222 f; contrasting alternative instruments (mainly of the 20th century) with the modern international treaty: Goldmann, Gewalt 21 ff. For the long-lasting discussion on whether or not public international law is to be called law in the first place see eg J B Scott, Nature, and the references to a variety of scholars made therein. For examples from the early and mid-20th century see Bothe, Norms 71–75; see also Schwarze, Soft Law 231, with a further reference.

<sup>39</sup> For the debate on the legal quality of the Universal Declaration of Human Rights as an early example of the 'soft law' discussion see eg Schwelb, Influence; for its role in paving the way for binding human rights covenants see Brown Weiss, Introduction 5; for its (at least partial) transformation into customary law see Malinverni, Effectivité

labour – than in others, eg trade and arms control.<sup>40</sup> The reasons for the adoption of such acts, initially primarily (bi- or multilateral) accords, have lain in the difficulty to reach the consent of all parties to a legally binding agreement. States have always been hesitant to legally bind themselves with regard to certain issues.<sup>41</sup> At the same time, most political representatives – after often wearing negotiations – consider preferable an agreement on something (eg a legally non-binding document) to no agreement at all or, as it is sometimes ironically described, to a mere agreement to disagree.<sup>42</sup> However, according to the contemporary international law literature, there are more purposes which the adoption of 'soft law' may serve. 'Soft law' acts may also be adopted, to name just a few, as a concretisation of (hard) legal norms,<sup>43</sup> as a preliminary commitment to adopt a legally binding act at a later stage of the negotiations,<sup>44</sup> to serve as a way around difficult ratification processes,45 or to prove the existence of hard (customary) law.<sup>46</sup> With regard to the latter case, the ICJ expressed in its Advisory Opinion on the legality of nuclear weapons that 'soft law' – in this case: a resolution of the UN General Assembly – may 'show the gradual evolution of the

<sup>301;</sup> for the different phases of academic debate on 'soft law', starting in the 1970s, see Mörth, Introduction 4.

<sup>40</sup> See Brown Weiss, Introduction 3.

<sup>41</sup> See Chinkin, Development 21f, with further references.

<sup>42</sup> See Klabbers, Courts 239; Peters, Typology 420.

<sup>43</sup> See Chinkin, Development 30; see examples in Pauwelyn, International Law 155; for different forms of concretisation/interpretation, authentic – non-authentic and authoritative – non-authoritative, see Fastenrath, Normativity 334 ff.

<sup>44</sup> See Brownlie, Extent 66; for the sole purpose of internationalising a matter (by means of 'soft law'), making the adoption of countering rules at the national level less probable: Peters, Typology 411; also the hope that soft provisions may 'develop into something with bite' is an important *stimulus* for the adoption of 'soft law' (where the adoption of hard rules is just politically not feasible); see Hockin, World Trade Organization 256.

<sup>45</sup> See Friedrich, Soft law 136; Pollack/Shaffer, Interaction 246; note § 72 para 1 of the Joint Rules of Procedure of the Federal Ministries of Germany (in their English translation; <https://www.bmi.bund.de/SharedDocs/downloads/EN/themen/mode rne-verwaltung/ggo\_en.pdf?\_\_blob=publicationFile&v=1> accessed 28 March 2023: 'Before drawing up and concluding international treaties (intergovernmental treaties, intergovernmental instruments, interministerial agreements, exchange of notes, and correspondence), the lead Federal Ministry must always verify whether settlement under international law is unavoidable or whether the aim pursued can also be achieved by other means, and in particular by agreements below the level of an international treaty'.

<sup>46</sup> See Chinkin, Development 30 f, with further purposes.

*opinio juris* required for the establishment of a new rule' and thereby 'have normative value'.<sup>47</sup> In other words, where States have made utterances on a certain issue in a legally non-binding way frequently or otherwise author‐ itatively enough, these utterances may be considered as an expression of a conviction that the rules at issue are legally binding. As we shall see, this seemingly paradox conclusion can be drawn also in other cases. Taking account of this multiple use of 'soft law' in international relations, Advocate General (AG) *Cruz Villalón* points out that 'the doctrine on sources of international law has increasingly sought to cover also the acts which, although legally non-binding, none the less exhibit a degree of relevance through references made to them, the reliance placed on them for the purposes of interpreting binding law or their practical effectiveness, all of this under the heading of "*soft law*" (emphasis in original).<sup>48</sup> Against this background, *Dehousse* and *Weiler* remarked, more than 30 years ago, that while '[l]awyers are naturally inclined to minimise the importance of international agreements deprived of binding force [...], agreements of that kind can have a crucial importance'.<sup>49</sup>

### 1.1.2. The challenges of using public international law as a starting point

Having prospered in opposition to a more or less fleshed-out *corpus* of law for well over a hundred years and still prominent in public international law,50 meanwhile the phenomenon of 'soft law' has also gained ground in national legal orders on a larger scale.51 Also in the EU legal order it has been used for a comparatively long time already. Irrespective of the legal

<sup>47</sup> *Nuclear Weapons*, ICJ Reports 1996, para 70. Similarly already in *Nicaragua v United States of America*, ICJ Reports 1986, paras 188 f, and more recently again in *Chagos Archipelago*, ICJ Reports 2019, paras 150–153.

<sup>48</sup> Opinion of AG *Cruz Villalón* in case C-399/12 *Germany v Council*, para 97. To limit the sources of public international law to those mentioned in Article 38 of the Statute of the ICJ is generally refused as overly formalistic and not doing justice to reality; see Koskenniemi, Utopia 181, with further references; Zemanek, Soft law 844; see also *Denmark v Norway*, PCIJ No 53 Ser A/B 1933, 69–71, and *Australia v France*, ICJ Reports 1974, para 51.

<sup>49</sup> Dehousse/Weiler, Single Act 129.

<sup>50</sup> On exemplary categories of 'soft law' in public international law see Knauff, Rege‐ lungsverbund 262 ff and 270 ff; see also Bothe, Norms 70 ff.

<sup>51</sup> For a selection of cases from domestic jurisdictions see Klabbers, Redundancy 174– 177; for the long-lasting reluctance of national constitutions to accept non-binding acts as a legal phenomenon see von Bogdandy/Arndt/Bast, Instruments 111.

order in the context of which 'soft law' is discussed, the issues in many respects remain the same.<sup>52</sup> They address, *inter alia*, the questions how 'soft law' can be distinguished from law on the one hand, and from other legally non-binding behavioural guidance on the other hand; which legal effects 'soft law' may have; whether it is possible to design a categorisation of 'soft law' acts; whether and, if so, which legal protection is available against 'soft law'.<sup>53</sup> Against the background of these manifold questions, and due to the fact that the aspects of the creation, the form and the effects of 'soft law' are – arguably in all international or national legal orders – regulated less intensely than they are, respectively, in the case of law, the systematisation of the body of 'soft law', even if only in one legal order, is a demanding task.

Public international law appears to be the legal order where law and 'soft law' in practice are by tendency most difficult to distinguish. It is its scarce and fragmented regulatory frame which *Chinkin* refers to when she utters that 'the richness and texture of contemporary international law and the broad differences in its form, purpose, style, and participants make illusory attempts to construct any systematic framework for the analysis of soft law that is not interspersed with exceptions, or framed at such a high level of abstraction that its usefulness is diminished'.54 Nevertheless – or maybe even therefore – a preoccupation with public international law promises the most generally applicable findings on 'soft law' as a 'legal' phenomenon. After all, general public international law – in spite of its diversified morphology – geographically speaking is the most universally applicable legal order in place.

Having said that, also the idiosyncrasies of public international law ought to be considered in order not to draw wrongful parallels to the EU or national legal orders.55 This concerns, above all, the strongly heterarchical structure within which the actors of public international law – in particular States and international organisations – together make rules. These two

<sup>52</sup> For the differences see references in Peters, Typology 406 f.

<sup>53</sup> See Klabbers, Reflections 316, with regard to the limited responsibility for making/ap‐ plying 'soft law'; see also von Bogdandy/Dann/Goldmann, Publicness 1389.

<sup>54</sup> Chinkin, Development 25; for these and other particularities of public international law which are relevant here see Griller, Fragmentierungen 246–248, with further references on the diverging views on the character and the categorisation of public international law; for a categorisation of the different approaches towards public international law and an analysis of the way the 'modern international lawyer' (page 157) takes, see Koskenniemi, Utopia 155 ff, and the references made therein.

<sup>55</sup> See generally Klabbers, Redundancy 170.

(inherently linked) elements – the heterarchy and the predominantly bior multilateral law-making – make the law *inter nationes* resemble the mechanisms of the (national) law *inter privatos* much more strongly than national and EU (public) law.<sup>56</sup> This is reflected in the predominant legal instruments of public international law, namely treaties or, more generally: agreements – which are just (in public international law more common) synonyms of the word 'contract'. With these legal instruments, their crea‐ tors most of the time are identical to their addressees. In national and EU (public) law, on the contrary, most of the acts apply to the respective citizens, eg regulations made by the legislator or decisions taken by an administrative authority. What is more, in public international law – simi‐ larly to private law with its fundamental principle of private autonomy – the rule 'Everything which is not prohibited is permitted', as prominently expressed in the *Lotus* case of the Permanent Court of International Justice, applies.57 Arguably it is not only the informality of 'soft law' as such, but also the relative freedom from legal restrictions which *Klabbers* refers to when describing informal international lawmaking58 as manifestation of the popular slogan 'Just Do It'.59 With respect to legally non-binding rules, however, public international law and private law strongly differ from each other: While the broad scope for 'soft law' is made use of abundantly in public international law,60 in private law the creation of non-binding norms appears to be a relatively recent phenomenon61 and in particular non-bind‐ ing contracts among private actors are still a rare exception (see 1.3.3.2. below). Even the so-called *obligatio naturalis –* an only seemingly evident

<sup>56</sup> Note the Austinian distinction between laws set by 'political superiors' and laws set by 'men to men'; Austin, Province 19.

<sup>57</sup> Critically: Hertogen, Lotus, in particular 913; for the similarities between public international law and private law more generally see Lauterpacht, Sources.

<sup>58</sup> This is a specific terminology underlying all contributions to the book Pauwe‐ lyn/Wessel/Wouters, Informal International Lawmaking. It is defined in the contri‐ bution of Pauwelyn, Informal International Lawmaking 22. It does not require acts to be legally non-binding, but it would certainly also encompass such acts ('output informality').

<sup>59</sup> Klabbers, Courts 223.

<sup>60</sup> See Rawlings, Soft law 215, who purports that '*official* business could not sensibly be carried on without [soft law]' (emphasis added).

<sup>61</sup> For legally non-binding acts adopted by private actors take the example of the Corpo‐ rate Governance Codes; with regard to the legal situation in Germany: Arndt, Sinn 60–68; M Weiß, Regulierungsinstrumente 37. According to the definition which will be presented below (1.3.), these Codes regularly do not qualify as 'soft law'.

example in this context – legally obliges the debtor (eg to pay a certain amount of money to the debtee).<sup>62</sup> What is more, its non-enforceability does not root in the contract as such, but is provided for by statutory law as a consequence eg of a lapse of time.

Summing up, public international law and its academic penetration seems to be a promising source of generally applicable findings with regard to the phenomenon of 'soft law'. Notwithstanding, the mentioned differen‐ ces between public international law and national and also EU (public) law should constantly be borne in mind in order to avoid wrongful projections.

### 1.2. Different concepts of soft law

As was set out above (1.1.1.), the notion of 'soft law' is not accepted throughout legal scholarship and even those who accept it attach different meanings to it. A selection of these different meanings, as expressed in the literature, shall be outlined here. For those who outright neglect the idea of 'soft law', exemplarily *Klabbers'* words shall be quoted: 'Our binary law is well capable of handling all kinds of subtleties and sensitivities; within the binary mode, law can be more or less specific, more or less exact, more or less determinate, more or less serious, more or less far-reaching; the only thing it cannot be is more or less binding'.<sup>63</sup> *Klabbers* criticises that the notion of 'soft law' was lacking theoretical groundwork. Even more fundamentally, he claims that the concept is outright unnecessary, as the binary concept of law 'can accommodate various shades of grey without losing its binary character' and thus also the phenomena which are referred to as 'soft law'64 – an understanding which certainly has adherents in the literature.<sup>65</sup> *Klabbers* discusses the scarce State practice on the one hand and on the other hand the international judicial practice, which, in his view, when dealing with 'soft law' recasts it into 'more accepted sources of international law', namely treaties and custom.66 This practice he also finds

<sup>62</sup> See *H René Laurer* and *Wolf-Dieter Arnold* in the discussion of Walter, Soft Law 29 f; see also Bodansky, Character 143, with further references.

<sup>63</sup> Klabbers, Concept 181.

<sup>64</sup> Klabbers, Redundancy 180.

<sup>65</sup> See eg Jabloner/Okresek, Anmerkungen 221; see also references by Arndt, Sinn 39 f, and by Pauwelyn, International Law 128 (fn 16).

<sup>66</sup> Klabbers, Redundancy 174.

confirmed in his selection of cases before domestic courts in which 'soft law' is at issue.<sup>67</sup>

Yet another scholar of public international law, *Weil*, concedes that the line between binding and non-binding provisions is sometimes difficult to draw,<sup>68</sup> and that non-binding provisions may 'create expectations and exert on the conduct of States an influence that in certain cases may be greater than that of [binding rules]'.<sup>69</sup> At the same time, he points out that a legal norm may also contain 'softly' worded provisions, without thereby losing its character as legal norm.<sup>70</sup> He refuses the assumption of different degrees of bindingness which, in his view, does not only affect the distinction between binding and non-binding provisions, but – eg due to the acknowledgement of the concept of *ius cogens* which is said to rank higher than the regular (public international) law – also creates a hierarchy of legally binding rules.71 Thereby a scale 'from nonlaw to superlaw'72 is created which *Weil* considers confusing and, in view of a certain degree of bindingness of all provisions on this scale, useless: 'A normativity subject to unlimited gradation is one doomed to flabbiness […]'.<sup>73</sup>

Those scholars, who are, on the other hand, open to the idea of concep‐ tualising a new category of norms in order to deal with the phenomenon at issue, in part differ significantly in their approach. *Abbott* and *Snidal* perceive 'soft law' as 'legal arrangements' which 'are weakened along one or more of the dimensions of obligation [ie bindingness], precision [of the normative wording], and delegation',74 the latter meaning the latitude to interpret, apply and elaborate the provision at issue.75 On the lower part of the spectrum, that is to say *not even* 'soft law' are 'purely political arrangements in which legalisation is largely absent'.<sup>76</sup> *Abbott* and *Snidal*

<sup>67</sup> Klabbers, Redundancy 174–177.

<sup>68</sup> See also Malinverni, Effectivité 301, speaking of a 'certaine porosité' of this line.

<sup>69</sup> Weil, Normativity 415.

<sup>70</sup> See Weil, Normativity 414.

<sup>71</sup> See Weil, Normativity 421f.

<sup>72</sup> Weil, Normativity 430; see also ibid 427.

<sup>73</sup> Weil, Normativity 429.

<sup>74</sup> Abbott/Snidal, Hard and Soft Law 422; see Terpan, Soft Law (2015) 73, who – in contrast to *Abbott* and *Snidal* – emphasises the criteria 'obligation' and 'enforcement'.

<sup>75</sup> For an in-depth discussion and partly development of these three dimensions see Schelkle, Governance 710 f; see also Mills, Biotechnology 329; critically: Mörth, In‐ troduction 6.

<sup>76</sup> Abbott/Snidal, Hard and Soft Law 422.

emphasise the 'continuous gradations of hardness and softness', a relativity of the normativity of law, and thereby refuse the binary model of law.<sup>77</sup>

Taking up on these three dimensions of law (obligation, precision and delegation), *Kirton* and *Trebilcock* define law as 'an international, or by extension an inter-actor, arrangement that has a very high value on each of those three dimensions'.<sup>78</sup> While law first and foremost relies on the authority and power of the State, 'soft law', by contrast, first and foremost relies 'on the participation and resources of nongovernmental actors in the construction, operation, and implementation of a governance arrange‐ ment', they say.<sup>79</sup> In 'soft law' regimes it is not the formal authority of gov‐ ernments which is relied upon, it is the voluntary participation which stays in the foreground, 'consensus-based decision making' serves 'as a source of institutional binding and legitimacy', and – due to this consensuality of 'soft law' – sanctioning powers of the States are absent.<sup>80</sup>

As a preliminary to his remarks on 'soft law', *Baxter* recalls that rules of public international law often cannot be enforced, but still it is generally accepted that they create rights and obligations. '[M]ore radical' than that, he deems the assertion that even legal norms which do not create rights or duties are part of public international law.<sup>81</sup> *Baxter* favours such an inclu‐ sive understanding of (public international) law. Among other instances of 'soft law', he lists examples of three categories of norms in international agreements which he considers to be soft, as they do not create legal obliga‐ tions: *pacta de contrahendo*, non-self-executing norms (requiring further implementing measures in order to be applicable) and merely hortatory provisions.82 By analysing these examples, he elaborates on the idea that

<sup>77</sup> Abbott/Snidal, Hard and Soft Law 424; for further apologists of such a relativity of normativity see references in Wellens/Borchardt, Soft Law 272.

<sup>78</sup> Kirton/Trebilcock, Introduction 8.

<sup>79</sup> Kirton/Trebilcock, Introduction 9.

<sup>80</sup> Kirton/Trebilcock, Introduction 9, with reference to Ikenberry, Victory. In legal re‐ gimes with more elaborate law-making procedures, eg many national legal orders or that of the EU, such an equivalence of legitimacy cannot be assumed, however. For the effect on the addressees' identity of voluntary compliance as opposed to mandatory compliance see Ahrne/Brunsson, Soft Regulation 188. For the concept of 'consensus' in international decision-making see Schmalenbach/Schreuer, Organisa‐ tionen, paras 1032 f.

<sup>81</sup> Baxter, International Law 549.

<sup>82</sup> See Baxter, International Law 552–554; see also Wirth, Assistance 223.

there are different degrees of bindingness. <sup>83</sup> The traditional binary concept, that norms, legally speaking, are either binding or non-binding, he repudi‐ ates as 'excessively simplistic'.<sup>84</sup>

*Fastenrath*, in turn, essentially argues that the relative normativity inher‐ ent in the concept of 'soft law' can be proven even from a perspective of legal positivism. With regard to the words in which legal norms are expressed, *Fastenrath* highlights that 'the vagueness in content of living languages is indispensable', above all in multilingual norms of public in‐ ternational law.<sup>85</sup> Against this background, international 'soft law' such as resolutions of the UN General Assembly provides an interpretative input just as multilateral treaties not yet in force or even unsuccessful codification conferences may do. Thereby these measures contribute to what *Fastenrath* calls 'the development of law *intra legem*'.<sup>86</sup> In the case of unwritten law – in the given context: customary law and general principles of law – 'soft law' may, *Fastenrath* adds, serve as evidence of practice and thereby contribute to the flexibility of a norm of customary law over time.87 With regard to general principles of law he writes: 'Here, hard law which truly earns its pre-fix "hard", may owe its very existence to soft law'.88 Thus, certain 'soft law' acts may be regarded as 'concretisations of legal ideas'.89 The lacking determination of the evidence required for confirming the existence of cus‐ tomary law or general principles of law leads *Fastenrath* to the conclusion that also in legal positivism the validity of a norm 'is always […] dependent upon contestable claims of varying degrees of authority' and that therefore 'legal positivism is unable to succeed in its attempt to exclude relative

<sup>83</sup> With (unwritten) customary public international law, such a variety of degrees of bindingness is much more accepted among lawyers, he argues; Baxter, International Law 563.

<sup>84</sup> Baxter, International Law 564; see also Chinkin, Development 23, taking up *Ingelse's* word of the 'gliding bindingness'; Peters, Typology 409: 'under-complex and too far away from reality'; critically as well: Thomas Müller, Soft Law 114.

<sup>85</sup> Fastenrath, Normativity 311.

<sup>86</sup> Fastenrath, Normativity 313 f.

<sup>87</sup> See Fastenrath, Normativity 320; see also Friedrich, Soft law 155–157. On the written evidence of customary law see Weil, Normativity 427 f; for the influence the dynamics of customary law may have on (the interpretation of ) international treaties see Article 31 para 3 lit b of the VCLT 1969.

<sup>88</sup> Fastenrath, Normativity 321; emphasising the similarity between principles and 'soft law': Kovács/Tóth/Forgác, Effects 58.

<sup>89</sup> Fastenrath, Normativity 321.

normativity from international law'.<sup>90</sup> Summing up, he argues in favour of accepting 'soft law' also from a legal positivist perspective as 'an instrument which provides, in as positivist a way as possible, understandings on the existence of rules, their formulation and interpretation'.<sup>91</sup> The relativity of normativity is, he concludes, not a problem of 'soft law' alone, but of law in general; and this he confirms for both positivist and natural law perspectives of law.<sup>92</sup>

*Knauff* perceives as constitutive elements of 'soft law' its regulatory char‐ acter and its creation by bodies vested with public authority.<sup>93</sup> With respect to the (non-)bindingness of 'soft law', he advocates a slightly more flexible approach. In view of a trend towards extending law to new forms of regu‐ lation (in particular as regards so-called internal law), he deems impracti‐ cable a strict opposition of law and 'soft law'.<sup>94</sup> According to *Knauff*, 'soft law' unfolds binding effects which are different from legally binding effects traditionally understood. In other words: A legally non-binding norm may nevertheless entail (a different form of ) bindingness. This other form of bindingness allows for a certain decisional leeway of those addressed. How‐ ever, he emphasises, the multitude of 'pathologische[] Fälle' [pathological cases] in the sphere of (hard) law (that is to say the many breaches of law in everyday life as exemplified by the large number of court proceedings) shows that even law is not always applied strictly, which means that also legal bindingness allows for decisional leeway.95 This, in his view, underpins the proximity of law and 'soft law'. In conclusion, *Knauff* defines 'soft law' as behavioural regulation adopted by bodies vested with public authority which is legally non-binding or legally binding only as regards the inner sphere of the regulator, but which has extra-legal steering effects. <sup>96</sup> Similarly and some years earlier, *Senden* has proposed the following definition of soft law: 'Rules of conduct that are laid down in instruments which have not

<sup>90</sup> Fastenrath, Normativity 322.

<sup>91</sup> Fastenrath, Normativity 324.

<sup>92</sup> Fastenrath, Normativity 324 and 330.

<sup>93</sup> Knauff, Regelungsverbund 216–220; *Knauff* thereby repudiates an understanding which also encompasses (soft) regulation by private actors, but accepts some forms of their mere contribution to the adoption of 'soft law'; ibid 218–220; similarly, with reference to *Knauff* and *Senden*: Wörner, Verhaltenssteuerungsformen 9; see also Kirton/Trebilcock, Introduction 10.

<sup>94</sup> See Knauff, Regelungsverbund 223; on the phenomenon of internal law see eg T Schmidt, Geschäftsordnungen.

<sup>95</sup> Knauff, Regelungsverbund 227.

<sup>96</sup> See Knauff, Regelungsverbund 228.

been attributed legally binding force as such, but nevertheless may have certain (indirect) legal effects, and that are aimed at and may produce practical effects'.<sup>97</sup>

With a focus on the international level, *Goldmann* develops the concept of 'international public authority'.<sup>98</sup> It is a comparatively broad term, en‐ compassing both legally binding and legally non-binding rules, but also other public output like information, as long as it qualifies as 'authority', that is to say as long as it has the law-based capacity to limit the addressees' individual or collective self-determination or to determine its use in a simi‐ lar way.<sup>99</sup> The extrinsic motivation to behave in a certain way – by external regulation or introjections (that is to say the unconscious integration of somebody else's view into one's own mental sphere) – must eliminate or at least reduce significantly the risk of dissent,<sup>100</sup> that means in particular: of deviating behaviour. Against the backdrop of this concept, *Goldmann* fleshes out a regime of different forms of actions of international public authority, encompassing normative as well as non-normative output. The distinction in particular between law and 'soft law', according to him, can best be drawn with a view to the respective consequences of non-compli‐ ance.<sup>101</sup>

As well acknowledging the strong proximity of law and 'soft law', *Arndt* proposes his concept of adaptive sources of law.102 In between the realms of law and 'soft law' – or, in his terminology: 'alternative Steuerungsin‐ strumente' [alternative steering instruments] <sup>103</sup> – he conceptualises a third category encompassing, as it were, the overlap of law and 'soft law'. Norms


103 Arndt, Sinn 155 f; for a more specific use made of the term 'steering instrument' see Senden, Soft Law 156, with further references.

<sup>97</sup> Senden, Soft Law 112; see also Snyder, Effectiveness 32; apparently, *Snyder* later, in 2007, added that 'soft law' may not only have practical but also legal effects (reported by Ştefan/Avbelj/Eliantonio/Hartlapp/Korkea-aho/Rubio, Soft Law 10 – footnote 58); Van Vooren, Study 700; see also Bothe, Soft Law 768, referring to 'pararechtliche' [para-legal] norms.

<sup>98</sup> See Goldmann, Soft Law; for the further development of the term see also Gold‐ mann, Gewalt 359 ff; for another legal concept of 'authority', also taking account of 'soft law', see Krisch, Authority.

<sup>102</sup> For the inspiration drawn from *Thomas Möllers* see in particular his works Re‐ chtsquellen 649 and Standards 143.

of this category are not characterised as being both law and 'soft law',<sup>104</sup> but by the fact that they *can be* both (the unclear cases, as it were).<sup>105</sup> Within an adaptive system, he calls for a weighing of different criteria (eg the effective‐ ness or the social usefulness of a norm), according to which the direction and the strength of diffusion ('Diffusionsrichtung', 'Diffusionsstärke') to either law or alternative steering instruments shall be determined.<sup>106</sup> Where it turns out that a norm diffuses in the direction of law, it may trigger – depending on the strength of diffusion – a duty to take note of it, a duty to address it or a duty to comply with it.<sup>107</sup> *Arndt* does not call these effects different degrees of legal bindingness, but 'verschiedenste Nuancen einer rechtlichen Geltungswirkung' [various nuances of legal validity effects].<sup>108</sup>

As was set out above, it is the aim of this sub-chapter to introduce different ways of addressing the phenomenon of 'soft law'. The selection of approaches presented here mainly stems from the field of public interna‐ tional law and legal theory, respectively. There is a large number of further contributions on legally non-binding norms with sometimes very specific *foci*, eg sub-normative (ie sub-legal) regulation by private actors in the field of (national) private law,109 which can only collectively be referred to here.<sup>110</sup>

#### 1.3. Discussion and conclusions

#### 1.3.1. Different schools of thought

Summing up the approaches towards legal bindingness presented above, different schools of thought can be distinguished. One, to which *Klabbers* and *Weil* belong, refuses the concept of 'soft law' as something in between law and non-law, claiming that the binary system distinguishing only (bind‐ ing) law from (non-binding) non-law suffices to cover the phenomena

<sup>104</sup> Note *Arndt's* wide understanding of 'soft law' which includes customs, fashion or the biblical Ten Commandments; Arndt, Sinn 129.

<sup>105</sup> See Arndt, Sinn 127.

<sup>106</sup> See Arndt, Sinn 131f and 156 f.

<sup>107</sup> See Arndt, Sinn 163–167.

<sup>108</sup> Arndt, Sinn 163.

<sup>109</sup> See eg Holliger-Hagmann, Gesetzgeber; Köndgen, Privatisierung; see also Peters, Typology 405.

<sup>110</sup> See eg the approaches in the literature introduced and analysed by Arndt, Sinn 92–116.

usually addressed as 'soft law'.<sup>111</sup> In another approach something like a relativity of normativity, or, more explicitly, different degrees (that is: a scale) of legal normativity is/are accepted. It is, therefore, also referred to as the 'continuum view'.<sup>112</sup> This view is advocated, in detail of course with different arguments, by *Abbott* and *Snidal*, *Kirton* and *Trebilcock*, *Baxter*, and *Fastenrath*. Again different are the approaches of *Knauff* and *Senden*. They separate legal normativity from other forms of normativity, namely 'außerrechtlich[e]' [extra-legal]<sup>113</sup> (*Knauff*) or 'practical'<sup>114</sup> (*Senden*) effects, and thereby avoid the slippery concept of different degrees of legal norma‐ tivity. Nevertheless, they accept that these effects may be legally relevant in one or the other way. With his term 'international public authority' *Goldmann* conflates law, 'soft law' and other output limiting or determining in a similar way the addressee's freedom or self-determination. While he does acknowledge a difference between these categories, he asserts that the distinction is to be drawn through the respective consequences of noncompliance. *Arndt*, who appears to be discontent with the term 'soft law', but not the concept as such, with strong reference to the works of *Thomas Möllers* starts from a binary understanding which he slightly adjusts for the purposes of his adaptive system. By introducing a third category, namely the adaptive sources of law, he allows for a weighing of different criteria and eventually for a 'diffusion' of these acts in the direction of either of the two categories of the binary system. Those diffusing in the direction of law may, depending on the strength of diffusion, cause a duty to take note of them, a duty to address them or, as the strongest effect, a duty to comply with them.

# 1.3.2. On legal (non-)bindingness as distinctive feature

The purpose of a concept of norms, more particularly a concept of law/ 'soft law', is to provide a practicable model which allows to grasp the *corpus* of these norms. While too narrow a definition of this *corpus* would render futile this attempt, as no 'big picture' would be provided, too much

<sup>111</sup> In other respects, in particular as regards the scope of law, these two concepts differ considerably from each other.

<sup>112</sup> See Peters, Typology 408, with further references; see also summary in Ştefan/Av‐ belj/Eliantonio/Hartlapp/Korkea-aho/Rubio, Soft Law 13.

<sup>113</sup> Knauff, Regelungsverbund 227.

<sup>114</sup> Senden, Soft Law 112.

inclusivity, ie a pluralist concept,<sup>115</sup> could be harmful, as well.<sup>116</sup> In other words: the concept, as a model, needs to reduce the complexity of reality – otherwise it would be of no avail –, but must not be simplistic. As we have seen, there are different parameters according to which the phenomenon behind the term 'soft law' is determined by different authors. A wide scale of obligation may be appealing in theory, but its orientation value in practice is very limited. In *Thürer's* words: 'Eine Norm gilt, logisch gesehen, oder sie gilt nicht. Sie kann nicht leicht oder stark, mehr oder weniger gelten' [A norm is valid, logically speaking, or it is not. It cannot be slightly or strongly, more or less valid].<sup>117</sup> Inherent in the validity of a legal norm is its bindingness. The fundamental difference between law and 'soft law' is that law is legally binding and 'soft law' is legally non-binding. This is not to deny the various forms of 'soft law' (and law, for that matter) existing in practice, but it is a statement that the distinction between legal bindingness and legal non-bindingness ought to be upheld and kept in focus when approaching 'soft law'.

In other words, the conceptual coverage of 'soft law' can be provided for best by concentrating on the obligation criterion, and by assigning a kind of normativity to 'soft law' which is different from legal normativity (ie legal bindingness). As is well known, '[t]he universe of norms is larger than the universe of law'.<sup>118</sup> *Verdroß* has famously described this phenomenon as the 'normative[r] Teppich […], in welchem das Recht mit den anderen sozialen Normen verwoben ist' [normative carpet in which the law is interwoven with other social norms].119 Also other regulatory concepts such as customs (eg international *courtoisie*) or morals unfold normativity (see 2.2. below), but not legal normativity. The same is true for 'soft law'. It exists between the unregulated free will on the one hand and legal restrictions on the other hand as a public authority's expression of desire for a certain action

<sup>115</sup> See, as one of the most prominent representatives of pluralist constitutionalism, *Gunther Teubner* and, for example, his work Zivilverfassungen; for a critical account of further examples see d'Aspremont, Pluralization 190–197; see also Somek, Con‐ cept 989, referring to the so-called 'inclusive positivism', a concept of law which also accommodates moral principles.

<sup>116</sup> According to *Arndt*, more conservative approaches – that is: approaches in principle moving within the traditional concept of legal sources – towards tackling the phe‐ nomenon 'soft law' are on the rise these days; Arndt, Sinn 116.

<sup>117</sup> Thürer, Soft Law 440.

<sup>118</sup> Pauwelyn, International Law 125; see already Wengler, Begriff 42.

<sup>119</sup> Verdroß, Völkerrecht 26.

on part of those addressed.120 An act of 'soft law' constitutes a (non-legal) norm of itself, an *argumentum ad verecundiam*, as it were.<sup>121</sup> From a legal perspective, law *must* be complied with, 'soft law' only *should*. It is the similarity with law and its strong interrelation with and even dependence on law which lead *Sereni* to suggest to locate 'soft law' in a 'twilight zone which is not yet law, but in which social and moral considerations are especially persuasive'.<sup>122</sup>

Whether or not we want to call it a 'twilight zone': In my view, the conception of a ('soft law') category of its own is worthwhile. The legal effects<sup>123</sup> of 'soft law', as opposed to the legal effects of law, are regularly related to the application of a separate legal act.<sup>124</sup> Often these effects even depend on this application, in the course of which reference is made to a specific act of 'soft law'.<sup>125</sup> *Fastenrath* takes the example of resolutions of the UN General Assembly which may determine the interpretation of binding acts of (UN) law.126 Such resolutions are – as recommendations127 – legally non-binding. That they may serve as a source of inspiration for the inter‐ pretation of a legally binding norm *qua* their *de facto* authority, and thereby entail legally relevant effects, does not contribute to any degree of legal bindingness of the resolution. Neither can the role assigned to 'soft law' (eg in the *Nuclear Weapons* case referred to under 1.1.1. above) when it comes to providing evidence for the existence of an *opinio iuris* as one constitutive element of customary law relativise its non-bindingness.128 Although also here the legally relevant effects of 'soft law' are considerable,129 it ought to be stressed that the 'soft law' acts at issue do not constitute the *opinio iuris*

<sup>120</sup> See Gramm, Aufklärung 67.

<sup>121</sup> This is well reflected in the words of *Gold* who says that on the part of the addres‐ sees of 'soft law' it is expected that they 'will take [its] content seriously and will give [it] some measure of respect'; Gold, Soft International Law 443.

<sup>122</sup> Angelo P Sereni, quoted in Schwelb, Influence 229; see also Everling, Wirkung 134: '"Grauzone" zwischen Recht und Politik' ['grey zone' between law and politics].

<sup>123</sup> See eg Bothe, Soft Law 768 f; Müller-Graff, Soft Law 22; see also the definitions provided by Knauff, Regelungsverbund 228, and Senden, Soft Law 112.

<sup>124</sup> See d'Aspremont, Softness 1078 f.

<sup>125</sup> See Pauwelyn, International Law 154, with further references.

<sup>126</sup> See Fastenrath, Normativity 313 f.

<sup>127</sup> See terminology in Article 13 of the UN-Charter; see also Shelton, Compliance 126 f, with reference to the case law of the US Court of Appeals.

<sup>128</sup> See Zemanek, Soft law 858 f.

<sup>129</sup> See Peters, Typology 420 f.

themselves<sup>130</sup> ; they only serve as (principally replaceable) evidence of its existence, which is a significant difference. Neither does the role of 'soft law' with regard to the interpretation of law relativise the legal bindingness of law, as *Fastenrath* seems to argue. Either the interpretation is viable, or – to come back to the role of 'soft law' in the context of determining customary law – the attempt to prove the existence of (customary) law is successful, or not. *Tertium non datur*. That the result of such a legal analysis may be contested, does not result in a relativity of bindingness.

While the strict separation between law and 'soft law' ought to be main‐ tained as a matter of conceptual purity, as it were,<sup>131</sup> this is not to be understood as an attempt to make legal reality fit into a concept 'at the price of distorting reality by discarding any variance'.<sup>132</sup> Legal bindingness is a singular form of bindingness, a conception which was created in order to separate legal norms from other norms.133 Legal bindingness (a *must*) is assigned to a norm due to the will expressed by the norm-creators and due to certain procedural requirements being met.134 In case of 'soft law', the will of the norm-creators – as established by applying the vari‐

<sup>130</sup> See Seidl-Hohenveldern, Soft Law 190; see also Knauff, Regelungsverbund 272, with regard to the fact that resolutions of the UN General Assembly (as such) do not become customary law, although *some of their content* may become legally binding. It is possible that the content of 'soft law' over time is transformed into customary law – as it is possible with the content of moral or religious norms; see Arndt, Sinn 45; Dalhuisen, Law 186; taking the example of international *courtoisie* which may become law or law which may degenerate to international *courtoisie:* Schweisfurth, Rechtsnatur 707, with further references.

<sup>131</sup> See also Müller-Graff, Soft Law 20 f; also in the affirmative: Friedrich, Soft law 127 f; Pauwelyn, International Law 159. *Klabbers* has expressed this argument in the following words: 'By creating uncertainty at the edges of legal thinking, the concept of soft law contributes to the crumbling of the entire legal system. Once political or moral concerns are allowed to creep back into the law, the law loses its relative autonomy from politics or morality, and therewith becomes nothing else but a fig leaf for power […]. [W]e need to insist on a degree of formalism, because it is precisely this formalism that protects us from arbitrariness on the part of the powers that be'; Klabbers, Undesirability 387.

<sup>132</sup> Bianchi, Butterfly 207, at 207–209 criticising the 'mainstream positivistic doctrine' in public international law.

<sup>133</sup> *Kelsen* emphasises the 'Eigengesetzlichkeit' of law, according to which legal science shall be limited to positive law, bewaring of incorporating other influences; see Jestaedt, Postulat 7, with further references.

<sup>134</sup> With regard to the procedural requirements see Mertens, Leges; mentioning three kinds of informality in the context of 'soft law' – 'output informality', 'process informality' and 'actor informality': Pauwelyn, Informal International Lawmaking 15–20.

ous methods of legal interpretation – was to create a legally non-binding instrument (a *should*), and normally the (regularly) enhanced procedural requirements for law-making have not been met.<sup>135</sup> This constitutes the decisive difference between the two regimes.<sup>136</sup> The author therefore argues that the formal status of a norm should be taken as a first point of reference. That a norm was created in the course of a law-making procedure (eg a legislative procedure), most strongly reflects the will of its creator to adopt law and justifies a congruous presumption.<sup>137</sup> Where a provision of a 'legal act' (eg a statute) lacks normative content, however, or where it is clear from its wording that the adoption of a *binding* norm was not intended (eg in the case of merely hortatory provisions), that provision does not qualify as (hard) law.<sup>138</sup> It is then either an entirely non-normative or a 'soft law' provision contained in a formal decision, statute, etc.<sup>139</sup> The presumption evoked by the legal form has then been rebutted. Where the

<sup>135</sup> For the normally more limited formal requirements for the adoption of 'soft law' see Gentile, Review 467; Andone/Coman-Kund, EU soft law 4. That is not to say, however, that the creation of 'soft law' may not be subject to a detailed procedure: Pauwelyn, Informal International Lawmaking 17. Where the creation of 'soft law' is subject to a legal procedure, a breach of the procedure amounts to a breach of law – with effects for the status of the act at issue.

<sup>136</sup> Stressing the importance of the will of the actors involved as 'necessary startingpoint': Wellens/Borchardt, Soft Law 270 and 274; A Aust, Theory 787; Friedrich, Soft law 128 f; for the field of public international law see Article 2 para 1 lit a of the VCLT 1969: 'governed by international law', which refers – according to the *travaux préparatoires* of this provision – to the required 'intention [of the parties] to create obligations and rights under international law'; see Schweisfurth, Rechtsnatur 684 ff.

<sup>137</sup> Also *Weil* stresses that from the wording of a provision as such deductions as to its normativity cannot necessarily be made: 'Whether a rule is "hard" or "soft" does not, of course, affect its normative character. A rule of treaty or customary law may be vague, "soft"; but […] it does not thereby cease to be a legal norm. In contrast, however definite the substance of a non-normative provision […] that will not turn it into a legal norm'; Weil, Normativity 414.

<sup>138</sup> See Kanehara, Considerations 81; Schweisfurth, Rechtsnatur 697; Wengler, Re‐ chtsvertrag 195; see also the early joined cases 42 and 49/59 *S.N.U.P.A.T.*, 72, accord‐ ing to which a 'statement [which] does not establish any general rule and does not conclusively affect any individual interest' cannot be called a decision (of the High Authority).

<sup>139</sup> Where the legislator is adopting a statute to wish the emperor a happy birthday (such acts where purportedly adopted bei the *Reichsrat* in honour of the Austrian emperor *Franz Josef*), this act, even if it takes the form of legislation, does, for lack of any normative content, not qualify as law (or 'soft law'); for the question whether such a 'senseless' rule still belongs to the legal order see Thaler, Rechtsordnung.

procedural indicators are weak, that means: where the form of the act does not allow for a (rebuttable) presumption as to its legal bindingness or non-bindingness (that may in particular be the case with certain (informal) agreements concluded under public international law), immediately other indicators of the will of the creator(s) of the norm have to be examined, eg its wording, an express declaration of the norm-creators to the effect that the act is considered non-binding, or the available means of ensuring compliance, etc.<sup>140</sup> Where the latter constitute true enforcement – in partic‐ ular: sanctions – this indicates legal bindingness (see also 2.1.1.1. below).<sup>141</sup> Where these means are weak (eg reporting duties<sup>142</sup>), that suggests the legal non-bindingness of the norm at issue. The line between enforcement and other means of ensuring compliance is sometimes difficult to draw. An example to illustrate this difficulty is what in German is referred to as *Obliegenheit*: a rule which is not binding, but if it is not complied with this may have negative consequences. It is also called a 'legal duty of a lower degree'. Such *Obliegenheiten* often occur in insurance contracts: In case of a damage, the insurance company is to be notified. If that is not done in due time, the insurance company may lawfully refuse to grant (full) coverage. If no notification is done, this is not a breach of contract, but it may have negative consequences (no coverage).<sup>143</sup>

### 1.3.3. The creators of soft law

# 1.3.3.1. On the difference between public and private legal action

When addressing the creators of 'soft law', it is worthwhile to first take a look at the creators of law. The latter shall be considered not only to get a broader picture of what all constitutes law in different legal orders, but also to outline some challenges for addressing non-binding norms, in particular

<sup>140</sup> For further indicators see eg the Joint Practical Guide of the European Parliament, the Council and the Commission for persons involved in the drafting of European Union legislation (2015), 2.3.3.: 'By contrast, in non-binding acts, imperative forms, or a structure or presentation too close to that of a binding act, must not be used'.

<sup>141</sup> See eg Dehousse/Weiler, Single Act 132, describing sanctions as 'the criterion *par excellence* of the existence of legal obligations', while conceding that they are not a necessary condition.

<sup>142</sup> Reporting may also be only voluntary, which further weakens this tool; see eg Friedrich, Soft law 140.

<sup>143</sup> See eg Goldmann, Gewalt 200.

when developing a definition of 'soft law'. In public international law the creators of law are primarily the States and international organisations, in places with the involvement of private bodies, such as non-governmental organisations (NGOs) or multinational undertakings.<sup>144</sup> The widely used term private international law is confusing in that it does not refer to inter‐ national law, but to national law on potential conflicts of norms of private law due to a foreign element of the underlying cases. Norms adopted by private actors – in particular: contracts – are (or may be) the subject of private international law of States, but they do not constitute the private international law themselves. In conclusion, this area of law is not relevant in our context.

EU law again is created by the institutions, bodies, offices and agencies of the EU and exceptionally – in particular: when its founding treaties are amended – by its MS. Private actors may create law on the basis of EU law (eg they may adopt the statutes of a *Societas Europaea* as provided for in Council Regulation 2157/2001), but they do not thereby create EU law themselves. Where exceptionally such power is delegated to them, in a functional perspective the output is – again – to be understood as law created by public bodies.<sup>145</sup>

At the level of national law, it is both the public actors (adopting regu‐ lations, judgements, administrative decisions, etc) and the private actors (concluding contracts, writing last wills etc) who create legal norms. As regards the legal bindingness of these norms, there is no difference between the two spheres.146 There is, however, a difference regarding the scope of norms: Public actors adopt norms in the public interest and in all kinds of policy fields, frequently in a general-abstract way, thereby including all individuals or certain categories of individuals (eg all doctors or all entrepreneurs) of the State concerned. Private actors normally create law only concerning themselves (eg a purchase contract concluded with a car salesman), their family (parents sign an employment contract for their

<sup>144</sup> See Friedrich, Soft law 136–138. For non-private individuals and working groups being empowered to adopt 'soft law' in public international law see Shelton, Com‐ pliance 125 f.

<sup>145</sup> The (historical) example of CESR, CEBS and CEIOPS who adopted EU 'soft law' acts shows that in the EU such delegation is considered problematic: They were considered Union bodies, and only their respective secretariats where established as private persons under French/British/German law; see Weismann, Agencies 86 f.

<sup>146</sup> For *Kelsen's* outright negation of a principal difference between the law emanating from these two spheres see references in Römer, Kritik 88; see also Jabloner, Re‐ chtsetzung 2.

under-aged child), their property (owner puts up a sign barring pedestrians from using a private footpath), their company (adoption of a statute by the shareholders, legally speaking that means: by the company itself ) etc.

As we can see, norms set by private actors regularly concern the creator's own affairs.<sup>147</sup> Only exceptionally the legal order vests these norms with a broader scope. This is the case where collective agreements on wages and other working conditions are concluded by representatives of employers, on the one hand, and of employees on the other hand, and where these agreements are then declared generally binding by the legislator. Another example is the legislator's reference to commercial customs (established by entrepreneurs) or industrial standards (set by private organisations; see in more detail 2.2.3. below). These cases can be explained by the interest-rep‐ resenting character or the expertise of the private actors concerned, and thus the involvement of private actors roots in considerations of contentwise legitimacy.148 It ought to be stressed, though, that it is legislative acts (ie acts adopted by public actors) which declare these rules binding. Again another case is the legislative delegation to private bodies of the power to adopt norms binding on individuals (eg administrative decisions). Since it is the public authority which is conferred here, the acts adopted by the private delegates, functionally speaking, constitute law adopted by public actors.

Without specific legal provisions, the private actors in our examples would not have the power to regulate – *de facto* in the case of mere legis‐ lative references to the respective output of private bodies, *de iure* in the case of legislative delegations of power to private actors – in a binding way with such a broad (in particular territorial and personal) scope.

The so-called *lex sportiva*, eg the rules on sports competitions as set by the Fédération Internationale de Football Association (FIFA) or the International Olympic Committee (IOC), is often cited as another example

<sup>147</sup> See eg Goldmann, Soft Law 374 f; Shelton, Compliance 128 (who does not exclude the use of the term 'soft law' in the context of private rule-making, though). That does not prevent these rules from being taken over as their own by others, if they are convincing; see Cloghesy, Perspective 327, taking the example of the chemical industry's Responsible Care programme, as developed by the Canadian Chemical Producers Association in 2003, which has spread on a worldwide scale; critically with regard to the distinction between a public and a private sphere: Jakab/Kirch‐ mair, Unterscheidung 364.

<sup>148</sup> Such content-wise legitimacy is related to ideas of common good – something in general *public* action is aimed at; see Arndt, Sinn 146–148.

of the enormous power private actors may have in making general-abstract rules.<sup>149</sup> While it is true that these rules constitute law created by private bodies (*in concreto*: associations established under Swiss private law) and while it is true that they are immensely important in the sports business, it ought to be emphasised that all sportspersons submit to these rules of their own accord, thereby making use of their respective private autonomy. That the large amount of power these bodies have makes it literally impossible to succeed as a sportsperson without submitting to the respective rules in a *de facto* perspective certainly relativises the voluntariness of this act. However, the FIFA or the IOC are still only regulating their respective 'own affairs' (see above): the rules related to football competitions or the rules related to the participation in and the performance of Olympic Games, respectively.<sup>150</sup> Thus, however problematic the rule-making power of single actors in the sports business may be, it does not principally challenge the above distinction between public and private rule-making.<sup>151</sup>

In summary, we can conclude, first, that the distinction between legal norms created by private actors and legal norms created by public actors is limited to the level of national law and, second, that the issues dealt with in either category of norms differ. There are important variances to this ideal‐ typical division, but they remain to be exceptions. The principle underlying this division is that private actors in general may not oblige third parties (of age152) without their respective consent (given eg at the end of negotiations on a contract) – and where they may, eg when they bar pedestrians from crossing their own premises, this obligation of third parties is strongly connected to the private actors' own affairs/rights (here: their property). Public actors have the power, more closely defined (and restricted) in the respective constitution and possibly also in sub-constitutional law, to oblige third parties without their consent: They may prohibit the sale of a certain

<sup>149</sup> Also individuals may exceptionally adopt highly influential rules. For the example of the Sullivan Principles adopted in 1977 as a measure against the racist policy of the *Apartheid* regime see Shelton, Compliance 129; see also the new global Sullivan Principles proclaimed in 1999.

<sup>150</sup> For the issues addressed in the context of the *lex sportiva* see Giegerich, Standards 123.

<sup>151</sup> The similarity of the effects of such private rule-making as compared to public rulemaking is acknowledged by the CJEU in particular in its case law on the freedom of workers; see above all case 36/74 *Walrave and Koch*, para 19; case C-415/93 *Bosman*, para 84.

<sup>152</sup> Parents may – to a considerable extent – lawfully oblige their under-aged children, even without the latter's consent.

food product or they may fix its price with effect for the entire territory of the State(s) concerned.

# 1.3.3.2. Non-binding norms: public and private creators

While at least at the level of national law both public and private actors may create – in terms of content and scope very different, but still – law, they may also create non-binding norms: A public authority may publish guide‐ lines to express legally non-binding norms its understanding/application of a certain piece of legislation (concretisation), a private company may adopt a Code of Conduct to suggest ethical decision-making on the part of its managers, a public ombudsman may set out his/her understanding of sound public administration, thereby attempting to steer the performance of the public service in a non-binding way, a private association may pro‐ claim its definition of corporate social responsibility, thereby setting out concrete instructions for undertakings as to how to live up to this responsi‐ bility.153 The idealtypical distinction between norms set by private actors and norms set by public actors appears to be less striking here, because both categories of actors may equally 'not-oblige' third parties without their respective consent. The private owner of property may put up a sign asking pedestrians to smile when crossing her premises. Those addressed may follow this plea, but they may as well not. In trying to influence human be‐ haviour, private actors may also move beyond their own affairs (see 1.3.3.1. above). Hence they may, for example, disseminate their invitation not to buy in supermarket X because its owners are 'not likeable'. Non-binding norms may also be created by private actors at the international level, eg by NGOs *vis-à-vis* States.

There is a wide range of possible non-binding rules private actors may establish (beyond their own affairs), with only some restrictions (eg prohib‐ ition of libel) to be considered.154 In the societal sphere (as opposed to the

<sup>153</sup> For the connection of private corporate social responsibility and non-binding out‐ put of public bodies, namely the OECD Guidelines for Multinational Enterprises, see Friedrich, Soft law 138; Wilkie, Governance 291 and 295 f.

<sup>154</sup> For the distinction between these two categories of non-binding norms adopted by private actors see, as an illustration, § 676 of the German Civil Law Code (BGB): 'Wer einem anderen einen Rat oder eine Empfehlung erteilt, ist, unbeschadet der sich aus einem Vertragsverhältnis oder einer unerlaubten Handlung ergebenden Verantwortlichkeit, zum Ersatz des aus der Befolgung des Rates oder der Empfeh‐

sphere of public authority), as *Lepsius* put it, private power does not have to legally explain itself ('private Macht [muss sich] nicht rechtfertigen').<sup>155</sup> Public actors may attempt to steer human behaviour in a non-binding way, as well, but they are bound by their respective competences, which essentially limit their scope of action to actions in the public interest or to what is referred to as common good.

Considering what was said under 1.3.3.1. above about the power of pri‐ vate actors to regulate their own affairs in a binding way, also non-binding utterances of private actors are particularly authoritative where they con‐ cern their respective own affairs: An example would be the employer's right to give (binding) instructions to his/her employee. It is a right granted on the basis of the employment contract. Where the employer announces that all of his/her employees should follow an open office door policy, this non-binding norm is highly authoritative, not least because the employer (arguably) may give a binding instruction to this effect to his/her employ‐ ees. Or: a farmer may kindly request walkers not to feed his/her cattle. This request carries a certain authority, not least because the farmer may otherwise legally prohibit any feedings by strangers. Thus, non-binding norms set by private actors may be highly authoritative, in particular where they are adopted 'in the shadow' of the power to adopt binding norms. In practice, in particular in oral conversation, the difference between a plea for help, a non-binding request and a binding command may become blurred, but in theory it must be upheld.<sup>156</sup>

lung entstehenden Schadens nicht verpflichtet' [Whoever has provided advice or a recommendation to somebody else is, without prejudice to any responsibility ema‐ nating from a contractual relationship or an unlawful action, not obliged to replace a damage caused by following the advice or the recommendation]. Non-binding norms related to the private actor's own affairs (here: related to a contract) are highly relevant in that they may entail legal responsibility. Beyond that, non-binding norms set by a private actor do not entail any responsibility, as long as they do not emanate from or constitute themselves an unlawful action.

<sup>155</sup> Lepsius, Funktion 54.

<sup>156</sup> A high authority of a non-binding utterance may, of course, also be subject to a power of its originator other than the possibility to ask for the same action in a legally binding way. For example: Where the employer tells his/her employee that he/she should help him/her in a private affair, eg his/her moving house, this (non-binding) plea may be particularly authoritative because the employer has the possibility to act in a legally binding way by dismissing the employee – a right which he/she principally has under the employment contract. Even though, objectively speaking, there is no inherent link between the employer's request and his/her dismissing the employee, the request and the dismissal are arbitrarily linked by

#### *II. SOFT LAW: TERMINOLOGY AND LOCALISATION*

It is a decisive difference between non-binding norms adopted by private actors and those adopted by public actors that the former are characterised by a high volatility of authority – sometimes compliance with them is not more than good manners (eg a teenager leaving the seat for an elderly per‐ son at the train station; a worker saying "good morning" to his colleagues when starting his morning shift; for custom and morals see 2.2.1. and 2.2.2. below), sometimes compliance is highly advisable because the norm is the final 'recommendation' before a legally binding act is adopted (eg dismissal from work or cancellation of membership eg to a sports club). What is more, one and the same non-binding norm may exert a different authority for different addressees: The non-binding command not to buy sweets proclaimed by the grandmother *vis-à-vis* her grandchild A and B, a friend of A, may be highly authoritative for A, but not for B – or *vice versa*. In general, the (oral, written, sign language) communication between private actors is particularly multi-faceted, as *Hart* has well illustrated,157 and only a small portion of it is normative. In many cases it may be unclear whether an utterance is still non-normative or already normative (in a non-binding or in a binding way) – and where the respective addressee complies, this question regularly is not thought of any longer.

With utterances of public actors the situation is different: They are, as an expression of the accorded *imperium*, often normative and always exert a 'minimum authority' in the sense that they are never entirely irrelevant because they are most often linked – in one or the other way – to the power to make law (unilaterally). They are, on a whole, more formalised (than those of private actors) and mostly are provided for in writing. Very frequently they have a general-abstract scope. Exceptionally, public actors also act by means of private norms – eg when they buy office paraphernalia for their officials, thereby concluding a contract with a salesman.158 It does not appear that non-binding output in this field – eg a municipality's plea not to feed the cattle on its premises (see the farmer's example above) – would entail the enhanced authority known from norms bearing public

the employer (or at least that is what the employee may think). This increases the authority of the request.

<sup>157</sup> Hart, Concept 18–20.

<sup>158</sup> It should be stressed that public actors are regularly more restricted in creating private rules (in the German-speaking literature referred to as *Privatwirtschaftsver‐ waltung*) than private actors. Unlike the latter, the former, for example, regularly have to comply with fundamental rights, even when entering the private sphere; see eg G Lehmkuhl, State-Building.

authority. In this case public actors, functionally speaking, act as private actors. Therefore we should specify: It is the public authority involved which increases the authority of non-binding norms.<sup>159</sup>

These particularities establish a principal difference between non-bind‐ ing norms created by actors thereby exercising their public authority and non-binding norms created by actors thereby exercising their private au‐ thority. In my view, they justify detaching non-binding norms as an expres‐ sion of public authority from the very inhomogeneous and legally hardly regulated set of non-binding norms adopted as an expression of private authority, and to classify and subsequently examine them as a category of their own.<sup>160</sup>

Therefore the term 'soft law' shall hereinafter be used to describe legally non-binding norms adopted by a body thereby making use of its public authority (*acta iure imperii*). Public authority is at issue when an entity, via its organs, is (limited to) exercising the powers assigned to it by (public) law.161 It does not comprise the possibilities of legal action granted under private law, eg a labour contract, even if the employer is a public actor. Public authority retains its character also when delegated to other bodies, eg to NGOs or private persons.162 That way also the recommendations of expert groups assembled by and attached to international organisations, eg the recommendations made by the International Law Commission which was created by the UN General Assembly,163 fall within the exercise of pub‐

<sup>159</sup> See Friedrich, Soft law 135, referring to a resulting 'distinct claim to legitimacy and authority', and 379, stressing the prescriptive function of (international) 'soft law' (adopted by international organisations) and its qualification as an act of public authority, each with a further reference; for a broader understanding of the term 'Autorität' [authority] in this context see Schreuer, Haven 69.

<sup>160</sup> See also Knauff, Regelungsverbund 217, with reference to the original use of the term 'soft law' in the context of public international law.

<sup>161</sup> For the term public authority in the context of EU law see Braams, Koordinierung 140 f; for an institutional and functional understanding (definition) of 'public au‐ thority' (in the context of environmental law) see eg Article 2 para 2 of Directive 2003/4/EC.

<sup>162</sup> See Kiss, Commentary 227.

<sup>163</sup> See eg Articles 16 para 2 lit j and 18 para 2 of the Statute of the International Law Commission; for the General Assembly's task to 'initiate studies and make recom‐ mendations' in order to encourage the 'progressive development of international law and its codification' see Article 13 para 1 lit a of the UN-Charter; see Knauff, Regelungsverbund 276 f; with regard to the ILO's non-ratified *conventions*, *Knauff* negates their qualification as 'soft law', ibid 275 f; for similar bodies in the more narrowly-tailored setting of a specific international agreement see the NAAEC and

lic authority.<sup>164</sup> Other forms of involvement of private actors<sup>165</sup> – the case of a delegation of powers has just been addressed – in the decision-making process (eg as experts<sup>166</sup> or informants) does not harm, as long as the entity thereby exercising public authority is or, thereby outweighing private actors, belongs to the formal decision-maker.<sup>167</sup>

# 1.3.4. A concept of 'soft law'

The above elaborations have confirmed that, conceptually speaking, 'soft law' is to be placed in the vicinity of law (adopted by public actors), in most cases more than eg moral or religious norms. This is because 'soft law' is a creature of law. For one thing, because it is limited by the law, as it may only be adopted if and to the extent its respective legal framework so allows.<sup>168</sup> What is more, 'soft law' and law have in common many characteristics, such as their (mostly) written form or a certain degree of publicity, <sup>169</sup> similar (or even the same) norm-creators, at least similar creation processes, and there is a strong relationship of explicit or implicit cross-referencing.170 The legal conditionality, together with the similarities just mentioned, characterises

its Council (within the Commission on Environmental Cooperation) which ought to adopt recommendations on elaborating on the NAAEC; see Article 10 *leg cit*.

<sup>164</sup> See Möllers/Fekonja, Rechtsetzung 803, with regard to the German corporate gov‐ ernance code elaborated by an expert committee established by the government (*Regierungskommission*); critically of such state-centredness: Walker, Pluralism, in particular 321f, with further references.

<sup>165</sup> For the involvement of non-state actors in international decision-making more generally see C Binder, Einfluss; Haslinger, Potential 34–37.

<sup>166</sup> See Friedrich, Soft law 394–397, also with regard to the delegation of decision-mak‐ ing tasks to experts.

<sup>167</sup> Similarly: Knauff, Regelungsverbund 218–220, with examples. However, *Knauff* seems to be stricter when it comes to non-state actors merely mandated by state ac‐ tors; see ibid 245. Doubtfully, but ultimately including 'norms adopted by non-state actors' in her definition: Shelton, Introduction 3 f; see also Boschetti/Poli, Study 27-35.

<sup>168</sup> Against this background, Boyle, Reflections 901 describes 'soft law' as 'simply anoth‐ er tool in the professional lawyer's armoury'.

<sup>169</sup> With regard to 'soft law' see Baxter, International Law 565; Everling, Wirkung 134 and 143; Wellens/Borchardt, Soft Law 271; see also Deumier, Droit souple 249 f.

<sup>170</sup> See Friedrich, Soft law 152 and 182 ff; Knauff, Regelungsverbund 265 f and 295 f, both with further references; for references in the recitals of acts of EU secondary law to the then only soft CFR see Szczekalla, Grundrechte 1020; see also Grund‐ mann, Inter-Instrumental-Interpretation 925 f; Opinion of AG *Cruz Villalón* in case C-399/12 *Germany v Council*, para 93.

'soft law' – and distinguishes it from other sets of norms, eg religious norms ('No meat on Good Friday!'), which in their normativity, if they do not themselves accept the superiority of law,<sup>171</sup> are not dependent on whether they are in compliance with the law. This suggests that we approach 'soft law' in consideration of the underlying legal framework and with the meth‐ odological tools known from legal science, in particular with a view to its interpretation and to cases of collision. <sup>172</sup> Due to this proximity, it may – in spite of the conceptually clearcut distinction between law and 'soft law' – be difficult in places to find out whether a concrete act belongs to the realm of law or to the realm of 'soft law' (see 2.1.2. and 2.1.3. below).<sup>173</sup> All in all, in the scholarly arena lawyers 'stay in business' also with regard to 'soft law',<sup>174</sup> although, of course, this phenomenon may be duly approached from eg a political science or – in the case of public international 'soft law' – an international relations perspective, as well.<sup>175</sup>

175 See Bianchi, Butterfly 214 f; d'Aspremont, Pluralization 197-199; Müller-Graff, Ein‐ führung 142 f. In the political arena this may be different, and there 'soft law' may contribute to a 're-assertion of the political sphere': Dawson, Soft Law 2.

<sup>171</sup> Whether such superiority is acknowledged by a religious regime depends. For the Christian religions see the famous passed-on words of *Jesus Christ*: 'Render therefore unto Caesar the things that are Caesar's; and unto God the things that are God's' (*Matthew* 22, 21).

<sup>172</sup> Pointing at the risk of contradictions between different 'soft law' acts: Senden/van den Brink, Checks 65 f; Opinion of AG *Bobek* in case C-911/19 *FBF*, paras 88 f, pointing at the 'risk of a "crowded soft-law house"' where overlapping soft law mandates of different EU bodies exist, and admitting a certain absurdity ('singular nature') of the concern about a conflict between different legally non-binding acts. The most important collision rules are *lex specialis derogat legi generali*, *lex posterior derogat legi priori*, *lex superior derogat legi inferiori*. The latter rule, in the context of 'soft law', may apply only exceptionally, namely where a formal hierarchy of 'soft law' norms can be established. Different degrees of factual authority are common, though. For the influence factors like mandate, voting procedure etc have on the *au‐ thority* of 'soft law' adopted by international organisations see Shelton, Compliance 128; with regard to EU 'soft law' see V.3.5. below; for the principal applicability of the *lex specialis* and the *lex posterior* rules in the context of 'soft law' see also Goldmann, Gewalt 499.

<sup>173</sup> See Deumier, Droit souple 250; see also Ballreich, Nachdenkliches 387; Hillgenberg, Soft Law 101; Kiegler, Anforderungen 266.

<sup>174</sup> Pauwelyn/Wessel/Wouters, Introduction 5. For a critical account of tendencies to expand the field of public international law so as to provide for sufficient 'legal mate‐ rials' for the lot of international lawyers to be preoccupied with see d'Aspremont, Softness 1088–1093. For the difficulty this entails for traditionally trained lawyers see eg Mertens, Leges 29 f.

The legal framework for the creation both of norms of private authority and of norms of public authority is again set by public actors: in the constitution and in the *corpus* of sub-constitutional legislation. Thus, acts of public authority establish the basis for (further) norms of public authority on the one hand, and for norms of private authority, on the other hand. But while the conduct expressing private authority is principally governed by private autonomy, in some respects restricted by the law<sup>176</sup> and – in case of legal persons – by self-regulation, with entities exercising public authority we need to distinguish: While in public international law (State sovereignty) the rule 'Everything which is not prohibited is permitted' applies as well,<sup>177</sup> in EU law and national public law it is the way round: Public authority may be exercised only where it is allowed. Whichever conduct is not (explicitly or implicitly) allowed is prohibited. This scheme can be applied also to the respective legally non-binding output.

We have discussed above (1.3.3.2.) that the legally non-binding output of private actors has a particularly broad scale and, at its lower end, often merges into everyday communication with no normative content. Legally non-binding emanations of public authority are comparatively more dis‐ tinct, especially where they are intended to be normative (but still legally non-binding). Since they always bear – as expressions of public authority – a minimum degree of authority, it is justified to examine them as a category of their own. The questions which may be raised in the context of this cate‐ gory of acts are often the same as those raised in the context of law adopted as an expression of public authority: Is there a material competence and is there a competence to adopt the act in the way the norm-creator chose to act (ie in a legally binding or in a legally non-binding way)? Are special thresholds met which are set in the context of exercising public authority: eg proportionality or protection of fundamental rights? How can the normcreator make sure its acts are complied with by the respective addressees? Which legal remedies against the act at issue are available? With regard

<sup>176</sup> For an explanation of the different meanings of the term 'private law' in this context see Goldmann, Perspective 57 f. In favour of private autonomy as a gateway for private rule-making: Köndgen, Privatisierung 520 f.

<sup>177</sup> Explicitly: *France v Turkey*, PCIJ No 10 Ser A 1927, paras 96–98 (Dissenting Opin‐ ion by Judge *Loder* with reference to the Court's view). On the assumption that the adoption of 'soft law' in a certain field has implications for the competence to regulate this respective field (in public international law) see Baxter, International Law 565, with reference to pertinent case law; for the similarities of (and differences between) private law and public international law see already 1.1.2. above.

to the creation of private norms some of these questions would not be relevant, and the answers to the relevant questions would be very different to those given in the context of rule-making by entities exercising public authority.<sup>178</sup> Also the rules of interpretation of acts of public (international) law partly differ from those of acts of private law.<sup>179</sup>

This relatively strong similarity between public authority expressed in the form of law and in the form of 'soft law' entails the risk that 'soft law' is increasingly replacing law. The most important reason for such tendencies is that the adoption of 'soft law' is regularly less demanding (in terms of the procedure to be applied) for its creator than the adoption of law.<sup>180</sup> This can be illustrated with regard to the question of competence: While the competence to set norms legally binding upon citizens (or, in case of public international law, States) is essential for the creation of such norms, a suffi‐ cient competence is required also for the adoption of soft rules. However, competences to adopt 'soft law' are regularly granted more generously than competences to adopt law. Where the distinction between these two sets of norms – law and 'soft law' – is blurred, the competence requirements for the adoption of law are at risk of being assimilated to those for the adoption of 'soft law', that is to say of being alleviated. Distorting this limitation would work against the restriction, and also the foreseeability for that matter, of (the exercise of ) public authority.

The unclear cases presented above shall not prevent us from attempting to give a definition of 'soft law'. After all, a resilient definition is a strong contribution to shaping the object of investigation and hence is a *conditio sine qua non* when working with a term as widely and differently used as 'soft law'. Such a definition should be broad enough to encompass the realms of public international law, EU law and arguably also that of national law – in spite of the differences these regulatory levels display in a more detailed perspective. It shall underlie the remainder of this work. In conclu‐ sion, we can define the term soft law (now and henceforth, due to its being

<sup>178</sup> See Pollack/Shaffer, Interaction 246.

<sup>179</sup> Take the *contra proferentem* doctrine as an example which, in many jurisdictions, is applicable in private law, but not in public international law; Articles 31 ff of the VCLT 1969; broadly referring to different effects: Knauff, Regelungsverbund 218, with numerous further references. This is not to say, however, that the language used in different areas of law and hence the methods of interpretation are *principally* different; see eg case 53/81 *Levin*, para 9, in which the Court refers to the 'generally recognized principles of interpretation'.

<sup>180</sup> See eg Wirth, Assistance 222.

defined, without inverted commas) broadly as norms, enacted by entities thereby exercising public authority and thereby aiming at steering human behaviour,<sup>181</sup> which are legally non-binding according to the interpretative‐ ly established will of its creators (or, as an expression of self-obligation, legally binding only upon the creators themselves). The merit of such an encompassing definition or concept is the designation of a phenomenon occurring in different legal orders on the basis of which an exchange of views is facilitated. Establishing this concept, however, is only a first step. It cannot address, let alone bring in order, the idiosyncrasies of the variety of expressions of soft law in all kinds of legal orders. If such classifications are feasible at all, then only with respect to the one selected legal order. In this work, a classification of soft law is attempted with regard to the EU's legal order.

# *2. Delimitation of soft law*

# 2.1. From law

2.1.1. Delimitation with a view to enforceability and effectiveness

# 2.1.1.1. On the issue of enforceability

Having provided for a definition of soft law, we shall now elaborate on some of the issues raised above, thereby making more explicit – but also pointing to the difficulties of – its distinction from other norms and nonnormative output of public bodies. We shall start with a distinction from law.

According to a common (positivist) definition (which shall be taken as a basis here and which was already referred to under 1.3.2. above), law is described, with reference to *Kelsen*, *Hart* or *Raz*, as a system of norms adopted by human beings for human beings which are – *grosso modo* –

<sup>181</sup> That includes 'institutional' behaviour, that is to say the behaviour of an institution in the broadest sense, eg a legal person, because here again it is human beings – as individual or collective organs – determining the legal person's behaviour.

enforceable<sup>182</sup> and effective.<sup>183</sup> Norms are defined as entailing a command, or they should at least be capable of being linked to a command.<sup>184</sup> The latter is the case eg with authorisations or permissions which are granted by a norm.<sup>185</sup> While also soft law, according to the above definition (see 1.3.4.), is composed of norms adopted by human beings for human beings, the *grosso modo* enforceability as the apparent (substantive) differentiator with a view to soft law shall be addressed more closely. A legal norm, where it is not obeyed, can be enforced, that is to say the extrinsic compliance – not: the (intrinsic) agreement with the content of the norm on the part of the person obliged – can be ensured by means of physical compulsion (execution or punishment<sup>186</sup>) which are laid down in advance.<sup>187</sup> Legal norms which by their very nature cannot be physically enforced, such as procedural rights (eg the right to be heard or the right to be represented by a lawyer) or the right to obtain a permission (eg to erect a building on one's premises), are pushed through otherwise, eg by an appeal to a court which repeats the procedure or grants the permission, or orders that this be done by the competent authority. Compliance with legal requirements to be met in order to be awarded a right or an authorisation (eg a concession to undertake a certain business) are ensured in that unless and until they are fulfilled, the authorisation – which is an act of public authority and which is required for the lawful undertaking of a certain business – shall not be granted. Running the business at issue nevertheless would then be considered unlawful and punishable. All in all, the non-compliance with a legal norm can – idealtypically – lead to physical enforcement, sanctions, or

<sup>182</sup> Sceptically as regards the enforcement criterion: D'Amato, International Law 1–6; Klabbers, Instruments 999; Peters, Typology 412. *Peters* also uses the term 'effective‐ ness', but thereby seems to refer to the enforceability of an individual legal norm.

<sup>183</sup> See eg Griller, Grundlagen (2015) 15; see also Conseil d'État, Droit souple 19. Also *Hart* does not outright refuse the importance of legal enforcement, but he takes a more nuanced approach than eg *Austin* (which he criticises): Hart, Concept 39; see also Noonan, Concept 170; Raz, Morality 7; for the three main theses of positivists distinguishing them from naturalists see Raz, Authority 37 ff; see also Engisch, Suche 10 ff and 56 ff.

<sup>184</sup> For the category of 'imperatives' see Larenz/Canaris, Methodenlehre 74–78, with examples.

<sup>185</sup> See Raschauer, Verwaltungsrecht, para 511; Rill, Fragen 7; Walter, Soft Law 23; see also Ruiter/Wessel, Nature 165, who would call such norms 'legally committing' as opposed to the narrower term 'legally obligating'.

<sup>186</sup> See Walter, Soft Law 22.

<sup>187</sup> See Kelsen, Law 76, see also Arroyo Jiménez, Bindingness 18.

at least to an authoritative emanation from a court or another public body that this non-compliance constitutes a violation of law.

Rules of soft law lack such 'enforceability in a broader sense'. This lack of enforceability in a broader sense again reflects their lack of legal bind‐ ingness (leaving the potential self-obligation caused by soft law apart<sup>188</sup>). Whereas law *must* be observed, on the basis of which it is generally enforce‐ able,<sup>189</sup> soft law, according to the will of its creators, *should* be observed and hence is not enforceable. This is, as was already stated above (1.3.2.), the primary (substantial) difference between law and soft law, the *nervus rerum* of the discussion on the distinction between these two categories of norms.

With regard to public international law, its extensive (though by far not all-encompassing) lack of enforceability has been invoked with a view to challenging its legal quality altogether.<sup>190</sup> In a monistic perspective, this argument could be countered by claiming that national legal regimes and public international law form one legal order. In such a holistic perspec‐ tive, non-enforceable rights/obligations of public international law can be counted to the exceptions of non-enforceable law (see in particular 2.1.3. below), but the large majority of rights/obligations of the legal order would remain enforceable (and hence the criterion of *grosso modo* enforceability in the definition above would be fulfilled).191 In a dualistic perspective, public international law is perceived as one system of norms which makes construing non-enforceable rights/obligations as exceptional more difficult. Either way, public international law remains to be a special case. Here the enforceability is much less strongly connected to legal obligation than in national law.192 It is not by chance that the intense discussion of the phe‐ nomenon of soft law has its origin in the literature on public international law.

Nevertheless, it ought to be emphasised that in many instances public international law does provide for enforcement measures, in particular

<sup>188</sup> See III.4.2.2.2.4. and III.4.2.3.2.3. below.

<sup>189</sup> For this unique feature of law see *Kelsen's* word of the 'Zwangsmoment […] [als] das entscheidende Kriterium' [element of force as the decisive criterion] of a legal order; Kelsen, Rechtslehre 78. However, exceptionally there may also be (hard) legal norms which cannot be enforced (see 2.1.3. below); for what *Koskenniemi* calls the 'skeptics', the existence of sanctions *stricto sensu* is a *conditio sine qua non* for the existence of law: Koskenniemi, Utopia 168 f, with further references.

<sup>190</sup> Critically eg Kelsen, Principles 23–25.

<sup>191</sup> Arguing in favour of such a monistic approach: Kelsen, Principles 403 f.

<sup>192</sup> See Bodansky, Character 143.

within the regimes of the UN, the WTO, and other international organisa‐ tions.<sup>193</sup> Enforcement by bodies belonging to other regulatory levels such as national courts additionally has to be taken into account.<sup>194</sup> Also general sanctioning mechanisms of public international law such as reprisal and retortion can be referred to here.<sup>195</sup> That in many cases they are not applied, is a different issue, and may be because frequently they are considered politically unattractive.<sup>196</sup>

Furthermore, it should not be forgotten that also in national private law – where the respective State regularly provides for means of enforcement – legal positions are often not enforced before courts, but pushed through otherwise (eg by means of alternative dispute resolution<sup>197</sup>), or are simply neglected. Civil courts are addressed far less often than they could be, especially where the potential claimant is economically less powerful than the potential defendant, where the risks involved are deemed too high, or where the opponents of a dispute will have to work together or live next to each other further on.<sup>198</sup>

198 For the enforcement of fundamental workers' rights see Canetta/Kaltsouni/Busby, Enforcement 56–61; for private enforcement in the EU in the field of competition law see Waelbroeck/Slater/Even-Shoshan, Study 10 f.

<sup>193</sup> For the (assumed) *grosso modo* effectiveness of public international law see Hongju Koh, Nations 2599 f and 2603, each time with further references; Kirgis, Interna‐ tional Law; with regard to WTO law see Griller/Vranes, EC-Bananas, para 21; with regard to compliance with WTO dispute settlement output see Petersmann, Trends 21; critically: Haas, Hypotheses 23; Zemanek, Soft law 845 f; for the view that the legal regimes of international organisations constitute legal orders separate from the public international legal order see references in Schermers/Blokker, Institutional Law, § 1142; arguing in favour of a broader understanding of enforcement in public international law: Brunnée, Enforcement 3–5; see also Koskenniemi, Utopia 180 f.

<sup>194</sup> See eg Nollkaemper, Role 168 ff.

<sup>195</sup> For further means of ensuring compliance in public international law see Bothe, Norms 88 f; in case of non-compliance with international soft law, only retortion may be used as a means of reaction by another state: Schroeder/Karl, Quellen, para 514; Seidl-Hohenveldern, Soft Law 205; for the estoppel effect as a (contested) argu‐ ment against non-compliance with UN soft law see Schweisfurth, Rechtsnatur 720 f; see also Bothe, Norms 87 and 95; Goldmann, Gewalt 200; Klabbers, Instruments 1003; Wengler, Rechtsvertrag 196 f.

<sup>196</sup> See Hongju Koh, Nations 2635 f.

<sup>197</sup> For the rise of such alternatives even in the framework of court proceedings see eg Roberts, Listing.

### 2.1.1.2. On the issue of effectiveness

Effectiveness is the second element referred to in the above definition of law which we shall examine more closely. The effectiveness of a norm – measured in terms of compliance rates – is essentially a matter of fact (but not always a matter of course). In spite of being measurable only quantita‐ tively, and hence not by applying traditional legal methodology, the degree of effectiveness is not irrelevant for legal norms.<sup>199</sup> The above definition requires that law, as a system of norms, is – *grosso modo* – effective. The system as a whole, not each and every norm of it,<sup>200</sup> needs to be effective in order to qualify as law. The phrase '*grosso modo*' makes clear that not a compliance rate of 100 percent is required, but that less than that may suffice.<sup>201</sup>

Doubtlessly, effectiveness is important also for other sets of norms, not only for law.202 However, unwritten normative regimes such as customs develop in a more flexible way than (written) law. They continuously adapt to humans' actual behaviour. A single custom which is not applied simply ceases to exist. This on-going communication between norm and practice leads to the application of a norm resulting in its existence.203 Where the effectiveness of norms cannot be scrutinised by third parties because compliance does not result in (visible) action,204 eg the religious command 'You shall not covet your neighbour's house', it is not a relevant factor.

Which role does effectiveness play in the case of soft law? While, accord‐ ing to the understanding applied here, it is to be perceived as a regime with a different normativity than law – norms of soft law *should* be applied according to the will of its creators, but it is not a legal *must* –, it is never‐ theless strongly coined by and attached to law. Its existence is conditional

<sup>199</sup> See Arndt, Sinn 122; Bianchi, Butterfly 210 f, with regard to the – inherently related – people's belief that something is law.

<sup>200</sup> See eg Griller, Grundlagen (2012) 15 f; ambiguously: Pöschl/Winkler, Grundlagen 44.

<sup>201</sup> Sceptical of the indetermination, but at the same time conceding the necessity of the effectiveness criterion: Rill, Fragen 6 and 12 f; taking account of the indetermined‐ ness of what (with *Austin*) he refers to as '"a general habit of obedience"': Hart, Concept 23 f.

<sup>202</sup> See Jabloner, Rechtsbegriff 29.

<sup>203</sup> In principle, this applies also to customary law as one branch of a legal order. It is the related *opinio iuris* which converts a custom into customary law.

<sup>204</sup> There may be instruments to disclose (non-)compliance, though, like the hearing of confessions.

on its legality in the respective legal order. This justifies its analysis in legal terms (see 1.3.2. above).<sup>205</sup> The effectiveness of law is the practical counterpart to its claim of legal bindingness, the effectiveness of customs is the practical counterpart to its claim of customary bindingness etc. Soft law – perceived through legal glasses – is non-binding, hence it claims non-bindingness. Thus, requiring a degree of effectiveness for the existence of the soft law regime would be contradictory. Soft law creators hope to reach compliance by convincing the respective addressees, though. This may work out, in particular, because the suggested rules are reasonable and/or because of the addressees' respect of public authority. If the creators of soft law did not hope for compliance, they would not adopt the soft law act in the first place.<sup>206</sup>

While in conclusion law requires a certain degree of effectiveness and while soft law does not, in practice a norm of law can be ineffective, that means not applied regularly, and a norm of soft law can be very effective207 – and *vice versa*. <sup>208</sup> Since there is no general causal link between (non-)effectiveness and (non-)bindingness, this criterion is of no use in distinguishing law from soft law.<sup>209</sup>

# 2.1.2. The recognition of law, soft law and other output of public authority: relevant indicators

Having dwelled on the issue of legal bindingness (as enshrined in the enfor‐ ceability criterion) as the substantial differentiator, we shall now examine potential indicators in this respect. First of all, we need to revive a question which is not, or at least not explicitly, addressed in the above definition of law: the question of form. In complementation of what was brought up in sub-chapter 1 above, the following is to be said: When those addressed by

<sup>205</sup> Note *Hillgenberg's* words, who described soft law as 'Quelle eines normativen Re‐ gimes, das juristischem Denken unterliegt' [source of a normative regime which underlies legal thinking]; Hillgenberg, Soft Law 101.

<sup>206</sup> That law may set a framework which renders compliance with soft law more likely, is a different story, which shall be addressed when specifically dealing with EU soft law in Part III of this study; for the example of monitoring mechanisms in public international law see Kanehara, Considerations 93 f and 96.

<sup>207</sup> On the reasons for compliance with non-legal norms in general see Brown Weiss, Conclusions 539.

<sup>208</sup> See Cannizzaro/Rebasti, Soft law 212; see also case T-113/89 *Nefarma*, para 76.

<sup>209</sup> See Cannizzaro/Rebasti, Soft law 217 f.

rules are entitled to have doubts as to whether a rule is legally binding or not (that is to say whether they *must* or only *should* act in a certain way), one of the core functions of law, the steering of human behaviour, will be hampered by creating or increasing an 'unclear state' of the relationship between rule-makers and the addressees of the rules.<sup>210</sup>

A positivist understanding of law must therefore strongly rely on formal requirements. Where an act of law was adopted in accordance with the formal (procedural) requirements, it qualifies as law even if it contains an unfair or immoral norm. It may later be annulled for non-compliance with higher-ranking law (eg fundamental rights), but up until then it is to be considered law. However, where an act of law contains non-binding norms (eg a recommendation) or where it does not have any determinable norma‐ tive content at all (eg with merely programmatic provisions contained in many national constitutions), the legal form (which it may still maintain under the respective law: statute, regulation, etc) cannot prevent its material qualification as soft law or not even soft law, as non-law.211 Conversely, a substantially speaking obligatory norm only entails (legally) binding effects if it formally qualifies as law (see also 1.3.2. above).

Formal (procedural) rules serve different purposes, among which may be the democratic legitimation of the decision-making process and hence also of the respective output (eg decision-making *quora*), their balancing of different interests (eg the requirement to consult interest groups during the deliberations) or ensuring that the decision-maker takes an informed decision (eg requirement to consult experts). Formal requirements such as signatures and promulgation contribute to informing the people of the existence of certain rules – eg a piece of legislation or a treaty concluded be‐ tween two or more States (objectives of publicity and of legal certainty).<sup>212</sup> The higher these formal requirements to be met by the creator(s), the less

<sup>210</sup> See Somek, Concept 994; see also Craig/de Búrca, EU Law 140; for the importance of 'highly certain normative knowledge' see Terpan, Soft Law (2013) 14, with further references.

<sup>211</sup> See Ingelse, Soft Law 81f, with further references; stressing the meaning of the substance of a provision: Chinkin, Development 25 f. With regard to EU legislation, the Council has held that 'provisions without legislative character should be avoided (wishes, political statements)'; Council Resolution of 8 June 1993 on the quality of drafting of Community legislation, 93/C 166/01. With regard to the normative requirements of decisions according to Article 288 TFEU see Geismann, Art. 288 AEUV, para 58, with many references to the case law.

<sup>212</sup> In case of EU law also the express reference to the legal basis of a legally binding act is such a 'procedural' requirement; see case C-325/91 *France v Commission*, para 30.

likely it is that a norm is adopted 'by mistake' – for example when the creator(s), according to the outward appearance and the wording of the act, clearly set binding rules, even though they (eg the individual parliamentari‐ ans) actually intended to adopt a non-binding act (eg a resolution).<sup>213</sup> Thus, compliance with the respective formal requirements also contributes to disclosing the actual will of the creator(s) to adopt a legally binding (or a legally non-binding) act.<sup>214</sup> This applies in particular to legislation or to constitutional amendments.

Normally, legal orders do not provide for a *numerus clausus* of soft law acts<sup>215</sup> which is why soft law regularly appears in many different forms<sup>216</sup> and may be adopted by a variety of different actors (vested with public authority). As was mentioned above on different occasions, the procedural requirements for its creation (if any) are by tendency lower – less developed – than with law.217 In particular at the EU or at the national level, it may therefore be comparatively easy to distinguish a piece of legislation from a soft law act – due to the strict formal criteria of the former.218 However, an implementing measure or a generally applicable instruction may, with a

218 Providing for counter-examples: Senden, Balance 94; with regard to the role soft law plays in national constitutions see Malinverni, Effectivité 300; for the interaction

<sup>213</sup> In that case the apparent (clear) will trumps the concealed actual will of the normcreator; see Rill, Methodenlehre 465–467. For the limited value of formality criteria in public international law see Pauwelyn, International Law 131–134.

<sup>214</sup> For the presumption of legal bindingness see 1.3.2. above. On the level of the single provisions (contained in a legal act) it is true that the will of the legislator can be less clear, sometimes even after having made use of various methods of legal interpretation.

<sup>215</sup> For the case of public international law see generally Knauff, Regelungsverbund 257; for EU law see III.3.1.2. below.

<sup>216</sup> For an account of the variety of soft law acts at the international, the EU and the national (German) level see Knauff, Regelungsverbund 257 ff; for the heterogeneity of soft law only in international environmental law see D'Amato, Soft Law 56; proposing a list of (in the terminology applied here: also soft law) norms issued by legal instruments see Ruiter/Wessel, Nature 172 f; rating, in the context of EU law, the substance of a measure higher than its form: case T-58/09 *Schemaventotto*, para 86.

<sup>217</sup> Heusel*,* Völkerrecht 302: 'formelle Anspruchslosigkeit' [formal simplicity] of soft law. See, for example, the publication requirements (publication in the OJ) for legislative acts and certain non-legislative acts in the TFEU: Article 297 para 1 subpara 3, and para 2 subpara 2 TFEU, respectively; see also case C-410/09 *Polska Telefonia*, paras 24 f and 30. For soft law acts, no such prescription is laid down in the Treaties (see Article 13 para 2 of Regulation 1049/2001: publication '[a]s far as possible'); see also Dickschen, Empfehlungen 175–180. For the effects of publishing international soft law at the national level see Schreuer, Anwendung.

view to the decision-making procedure, very well be confounded with soft law – and *vice versa*, respectively.

Where procedural requirements for the adoption of a legally binding act are (severely) violated, the output is – depending on the legal order at issue – void or even a so-called non-act. Where at the same time the requirements for the adoption of a soft law act are met (in particular where it turns out that the originator, according to its expressed will, *at least* wanted to adopt a soft law act<sup>219</sup>), the output at issue may be construed as a soft law act, though.<sup>220</sup> Similarly, where there is no legal basis or no competence for adopting a certain legal act, the output may – given the respective competence – be interpreted as a soft law act. Where in such cases the originator's expressed will does not even point in the direction of soft law, the examination of the output can be stopped. Without an appropriate (expressed) will, no soft law act can be created.

A 'soft' wording (aims instead of imperatives, eg 'are invited to', 'shall attempt') makes enforcement impossible and therefore indicates the origi‐ nator's will to adopt only soft law.221 A 'soft' wording is not to be confused with a lack of determination.222 Both legal and soft law norms have to reach a minimum degree of determination in their wording: 'If law is to

222 They may appear in combination, though; see Virally, Valeur 68, uttering that recommendations in public international law display 'une précision souvent très relative'.

between hard and soft law at the national level during the pandemic see Boschet‐ ti/Poli, Study 48-51.

<sup>219</sup> In case of a mental reservation of the originator, the general rules of interpretation apply. In accordance with the so-called declaration theory (as opposed to the will theory), it is the objective declaration ('expressed will') that counts; see eg Armbrüs‐ ter, Vorbemerkung, para 21. With regard to the cases of a threat to or a deception of the legislator: Morlok, Informalisierung 72 (in particular fn 124), with further references. Only in case the addressee of the act knows or can be expected to know that the originator has not had an according will, the appearance does not prevail over the actual will, as then there are no legitimate expectations to be protected on the part of the addressee.

<sup>220</sup> With regard to the will of the norm-creator see Rill, Fragen 9 f. For the importance of the orginator's will, and for the difficulty to establish this will see Pauwelyn, International Law 134–136. For the 'presumptive law' thesis, that is to say a pre‐ sumption of legal bindingness for certain output, see Klabbers, Courts 224 f. For the importance of this will/intention see Ingelse, Soft Law, in particular 79.

<sup>221</sup> See Ingelse, Soft Law 80. See, as an illustrative example, the Commission Commu‐ nication at issue in case C-57/95 *France v Commission*; on the meaning of the wording, see in particular the Opinion of AG *Tesauro* in this case, para 16.

have instrumental value, its content should be reasonably clear'.223 In the context of law, this is one aspect of legal certainty, but also in the context of soft law it applies in principle, because there is simply no normative content without a minimum degree of determination.<sup>224</sup> If necessary, also other indicators, such as the contents and purpose of the relevant act, should be examined,<sup>225</sup> as they may give a hint to the legal quality of the act at issue. Also the available means of ensuring compliance with a norm (if any) may be worth looking at. Means of enforcement characterise law, whereas 'soft sanctions' speak in favour of soft law.<sup>226</sup> A lack of available means of enforcement for a specific norm does not automatically make it non-binding, though (see 2.1.1.1. above and 2.1.3. below).

Other indicators raised in the literature, such as effect – actual changing of the addressees' behaviour – or a certain degree of legitimacy<sup>227</sup> do not appear to require further attention here.228 As we have seen above, the criterion of effectiveness may neither serve to distinguish soft law from law, nor is it a requirement for the existence of soft law, and legitimacy concerns are addressed (already) by procedural requirements, such as the necessity of a legal basis or the required compliance with higher-ranking (positive) norms of the legal order at issue. As regards the extravagant approach proposed on the level of public international law, namely that law is what we believe to be law, the following can be said: While there may be examples in State practice underpinning this approach,229 and while this

227 See d'Aspremont, Pluralization 193.

229 For example the *opinio iuris* as one constitutive element of customary law.

<sup>223</sup> Koskenniemi, Utopia 177; for the allegedly compromise-induced vagueness in the formulation of some pieces of EU legislation see Senden, Soft Law 12.

<sup>224</sup> See Arndt, Sinn 134–136; see Sarmiento, Soft Law 274 f, with regard to the distinc‐ tion between rules and principles which may also be applied in the context of soft law; see also Dworkin, Rights 22-28; for the development of (positivist) legal concepts in 19th century Europe along the lines of (and the quest for) legal certainty see Van Meerbeeck, Principle 279.

<sup>225</sup> See Arndt, Sinn 134–153, proposing further elements; Thomas Müller, Soft Law 119.

<sup>226</sup> See Abbott/Snidal, Hard and Soft Law 422; see also Knauff, Regelungsverbund 246, who – with a view to the possibilities to react to non-compliance – stresses that a strongly 'reduced bindingness' [reduzierte Verbindlichkeit], NB: non-legal binding‐ ness, may not meet the requirements of soft law; for the institutionalised dispute resolution on the basis of soft law provided for by the World Bank's Inspection Panel see ibid 281f.

<sup>228</sup> For a critical account of these criteria see Pauwelyn, International Law 136–139.

belief may have a significant impact on a norm's effectiveness, as a general concept it can be upheld neither conceptually nor practically.<sup>230</sup>

Having listed a number of indicators to distinguish law from soft law, it ought to be stressed that soft law needs to be distinguished not only from law, but – on the other side of the scale – also from output which is not (even) soft law, eg policy papers without any normative content. Here the substantive (rather than formal) criterion of normativity (including non-binding normativity) is highly important. The expressed will of the originator to create a (soft) norm, as determined by using the methods of legal interpretation, is a necessary (but not a sufficient) condition for its actual creation. In this context, first of all, we have to ask whether the act is – either explicitly or implicitly – directed to a certain addressee or group of addressees in order to exclude output of a merely programmatic nature. Where it is, is it apparent that the addressee(s) (can be the creator itself ) shall be committed and, if so, in which way? Here it is again the wording of the act which will not be the only, but normally the most promising source of knowledge (see also 2.3. below).

### 2.1.3. Exemplifying the proximity between law and soft law

### 2.1.3.1. General examples

Above it was attempted to distinguish law and soft law from each other. While this conceptual separation is worthwhile, it should not be concealed that in practice there are cases which, at least *prima facie*, seem to challenge this separation. For instance: A so-called *lex imperfecta* is legally binding, but does not provide for legal consequences of its violation and hence cannot be enforced.231 A declaratory decision or judgement merely states rights or obligations which a certain person has according to the law. It is

<sup>230</sup> See Pauwelyn, International Law 139–141, with further references.

<sup>231</sup> See Rill, Fragen 3, who emphasises that such provisions entail a command, as well, and can be accepted in a system of otherwise enforceable legal norms; see also Austin, Province 32 f, with reference to the 'Roman jurists'. In some cases the abstract possibility to claim damages may render enforceable even a *lex imperfecta*: Where the rule not to eat in the public library is not combined with enforcement measures (eg to expel the non-compliant library-user), non-compliance may still be 'sanctioned' where non-compliance causes a damage to another user of the library, eg when he/she is allergic to the food and falls ill or when he/she is slipping on the banana peel and breaks a leg. He/She can then pursue his/her claim before court, arguing that the damage was caused by another person acting in an unlawful way

binding in that it authoritatively states certain rights or obligations, but – if adopted in accordance with the law – it does not create any new rights or obligations. What about a legal act which has been duly adopted (is valid, that means), but is not yet in force? As it clearly is legally non-binding, the question arises whether it can be qualified as soft law. In light of the above definition (1.3.4.), this question is to be answered in the affirmative. In particular, the norm-creator's intention to steer human behaviour can be confirmed. After all, a *vacatio legis* is regularly foreseen in order to allow those addressed to adapt or get accustomed to the new rules. This implies that already during that time they should comply with these rules, but they are not bound to do so.

Also a provision whose content is expressed in very broad terms (for example: 'The authorities in charge shall take adequate measures to protect the environment')232 raises some questions in this context. The provision may be qualified as legally binding because the wording, at first sight, does not indicate legal non-bindingness. However, a closer look reveals that the terms used ('adequate measures', 'protect the environment') assuming that they are not defined more closely elsewhere in the respective law, leave such a wide margin of action for those addressed ('the authorities in charge') that it is impossible to deduce, by means of a semantic interpretation, a more concrete duty from it. Hence this and similar provisions also cannot be enforced.<sup>233</sup> *D'Aspremont* – because of the leeway the provision grants – would call it a 'soft negotium'.234 But that does not necessarily mean that it constitutes soft law according to the definition applied here. It is

234 D'Aspremont, Softness 1084.

<sup>(</sup>non-compliance with the rule which also constitutes a protection standard for the injured user of the library). That way – although only indirectly – an 'enforcement' may take place; in the context of public international law, see the landmark decision in *Factory at Chorzów*, PCIJ No 9 Ser A 1927.

<sup>232</sup> With regard to the EU, see eg Article 191 para 2 TFEU. On the qualification of this provision ('principle') and on its lacking enforceability see Nettesheim, Art. 191 AEUV, paras 81–83.

<sup>233</sup> See also Bodansky, Character 143. The fact that virtually any legal norm may be subject to (at least slightly) different interpretations is inherent to human language by means of which a legal norm is expressed; see Hart, Positivism 607 f, explaining his famous core-penumbra metaphor, according to which each norm has a penum‐ bra of meaning which requires interpretation to become clear. Only where not even a minimum degree of determination (which may vary from legal order to legal order) is reached, so that the possible results of legal interpretation would be nearly limitless, does a lack of clarity render a norm of law non-enforceable; see also the criterion of 'precision' by *Abbott* and *Snidal* referred to under 1.2. above.

the lack of determination which makes this provision weak, not its legal non-bindingness. It is true, however, that the effect of this provision – allowing for a wide range of different actions – is somehow similar to soft law, which also allows for different actions, meaning that it may lawfully be either complied with or not.<sup>235</sup>

In contrast to legal norms which are not enforceable, in the context of soft law we may come across different modes of pushing compliance, among which the most prominent are the so-called 'comply or explain' 236 and 'naming and shaming' <sup>237</sup> mechanisms.<sup>238</sup> Pointedly, we could say that such rules are legally non-binding but can nevertheless be asserted. This reflects the terminological *contradictio in adiecto*<sup>239</sup> of the concept of soft law mentioned above (1.1.). While some may attribute these means of asser‐ tion to 'a limited normative force'<sup>240</sup> of soft law, in the approach taken here it is explained as a form of normativity which is different from legal bindingness and enforceability. Similarly, the breach of a moral norm, eg adultery, may have (morally) normative consequences, eg a bad reputation among acquaintances or colleagues at work. They may be more individual


<sup>235</sup> Nevertheless, the type of provision at issue here is legally relevant in that it is to be considered by an authority or court eg in the weighing of different interests when deciding on a claim (which is based on a different, legally enforceable provision). Take the preambles of international treaties as an example. They are an integral part of the respective treaty and serve as guidance for its interpretation which is obligatory in the sense that it must be referred to in case the meaning of a treaty provision is unclear; see Article 31 para 2 of the VCLT 1969; similarly with regard to recitals of EU legal acts: Bast, Grundbegriffe 352, with further references. For the place of Recitals in the hierarchical order of rules see case C-345/13 Millen, para 31.

<sup>236</sup> On the theoretical foundation of the 'comply or explain' approach see Horak/Bod‐ iroga-Vukobrat, Experience 184–187; see also Schilchegger, Agenturen 127; Dick‐ schen, Empfehlungen 125–130.

<sup>237</sup> On the 'naming and shaming' strategy in international human rights law see Haf‐ ner-Burton, Sticks; in the context of the PISA Study see von Bogdandy/Goldmann, Ausübung 75; with an example from the field of banking supervision: Müller-Graff, Rechtsschutz 103; with regard to EU law, the Commission in a legislative proposal has described the purpose of 'naming and shaming' as follows: 'for every possible minor offence [it] may be excessive. It remains, however, a useful deterrent in the case of infringement of the Directive's basic requirement […]'; Amended proposal for a Directive establishing a scheme for greenhouse gas emission allowance trading within the Community and amending Council Directive 96/61/EC, COM(2002) 680 final, 3.

<sup>238</sup> For (other) soft means of 'enforcement' see (also) Knauff, Regelungsverbund 294 f; Terpan, Soft Law (2013) 10; Yoshida, Enforcement.

and hence less predictable than those following a violation of (soft) law, but they are still consequences of a breach of a norm.

With regard to the 'enforcement' of soft law, it ought to be stressed that these mechanisms still allow for deviance. In case of the 'naming and shaming' mechanism, the deviator risks his/her name to be published, but there is no means available to actually force him/her to comply. With 'comply or explain', the deviator is asked to provide his/her reasons for deviation. That way, the deviator may even convince the creator of the soft law act (or another body in charge of monitoring compliance) of the necessity to deviate. In case of a soft law act, the 'duty' to comply is only a soft one. The duty to explain in case of deviance may be soft (ie legally non-binding), but may as well be hard.<sup>241</sup> In such a constellation there is a strong intersection of law and soft law. Where the underlying norms are soft, the duty to give good reasons for a deviance, however, is hard, the mechanism practically equals a hard rule-exception clause: A legal norm may oblige its addressees, but at the same time provides for certain exceptions (which justify deviance).242 The addressee may either comply or claim an exception to be applicable. In this case, the legal duty to provide the reasons for deviance seems to 'harden' the soft norm, and the situation is – apart from the fact that simply providing the reasons may allow for a greater leeway than concrete exceptions determined in advance – *practically* equivalent to an entirely hard mechanism.<sup>243</sup>

Another example would be a default rule in the form of *ius dispositivum* which is common in labour law or in tenancy law: Where the contracting parties do not address a certain issue (eg the working hours or the date of payment) in their contract, the law provides for a default rule which is to be applied. If the parties do not want it to apply, they have to agree otherwise.<sup>244</sup>

These cases, which – as examples – do not claim completeness, provide evidence of the strong proximity which may exist between law and soft law. This aggravates the distinction between the two. In my opinion, however, these borderline cases do not *in principle* challenge the separation between

<sup>241</sup> For the latter case see Griller, Übertragung 156, with further references.

<sup>242</sup> See Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 100 f, who even claims that a duty to give reasons for deviation renders the respective rule binding.

<sup>243</sup> For other forms of a 'hardening' of soft law see Andone/Coman-Kund, EU soft law 5–7 and 13; see also Tridimas, Indeterminacy 61.

<sup>244</sup> See eg Fleischer, Gesetz 692.

law and soft law, as proposed here. On the contrary, they facilitate the intellectual grasp of the sometimes only vaguely felt differences between legally binding and legally non-binding norms.

2.1.3.2. Special effects of public international law in EU law – the *Kadi* saga and the case of WTO law

### 2.1.3.2.1. Introduction

In order to enrich the discussion on the proximity between law and soft law specifically with regard to the multi-level regulatory regime, we shall now have a closer look at two further phenomena, which are similar to soft law. They are relating to the effect of public international law in the EU legal order, more precisely the authority of resolutions of the UN Security Council on the one hand, and the enforcement of apparent claims laid down in WTO law, on the other hand. These topics shall be presented and subsequently discussed with a view to deepen the above analysis, fleshing out the difficulty to clearly delineate soft law from law.

It is not the purpose of this chapter to provide an exhaustive account of the two subject matters. Rather, they shall be presented only to the extent necessary to disclose the similarities (and differences) to soft law, which subsequently shall be analysed in more depth.

# 2.1.3.2.2. The effect of UN law in the EU legal order, exemplified in the *Kadi* cases

The question of the effects of resolutions of the UN Security Council in EU law lies at the core of the Court's jurisdiction in the *Kadi* cases. Before dis‐ cussing these cases, a word should be said about the principal relationship between EU law and public international law – from the perspective of the EU legal order. Primary law tends towards a strong consideration of public international law.<sup>245</sup> The EU shall, among other things, 'contribute to peace […] as well as to the strict observance and the development of international

<sup>245</sup> See eg HP Aust, Union 109–111.

law, including respect for the principles of the United Nations Charter'.<sup>246</sup> As regards the abidance by public international law, this provision largely forms a codification of the case law of the CJEU, according to which the EU 'must respect international law in the exercise of its powers'.<sup>247</sup> In par‐ ticular, the Court has decided in a number of cases that public international law concluded by the MS prior to the foundation of or their respective accession to the EU (or one of its predecessors) shall be given preference in principle.<sup>248</sup>

At the same time, EU law with its determinative characteristics such as direct effect, supremacy and its elaborate human rights standard is very different from (general) public international law.<sup>249</sup> These idiosyncrasies are permanent and, according to primary law and also according to the Court, form core principles of the EU legal order.

Against this background, the *Kadi* cases are to be understood. In these cases, requirements following from UN law and their respective implemen‐ tation by the EU – still under the TEU and the Treaty establishing the European Community (TEC; both in their respective Nice version) – were at issue. More concretely, the UN Security Council had issued a number of resolutions requesting States, among other things, to freeze the assets of *Usama Bin Laden* and of organisations associated with *Usama bin Laden*, the *Al-Qaeda* network, and the *Taliban*, as referred to in a list set up and, if needed, to be updated by the UN Security Council Sanctions Committee.<sup>250</sup> The resolutions were adopted under Chapter VII of the UN-Charter which is about 'action with respect to threats to the peace, breaches of the peace, and acts of aggression'.

<sup>246</sup> For the legal situation under the Nice regime – which was relevant in the *Kadi* cases to be addressed below –see in particular Article 11 para 1 TEU (Nice); see also eg Article 177 para 3 TEC (Nice).

<sup>247</sup> Case C-286/90 *Anklagemyndigheden*, para 9; case C-308/06 *Intertanko*, para 51; case C-366/10 *Air Transport Association*, para 123.

<sup>248</sup> See eg case 10/61 *Commission v Italy*, 10; case C-158/91 *Levy*, para 12; case C-324/93 *Evans Medical*, para 27; case C-124/95 *Centro-Com*, para 56.

<sup>249</sup> See generally Weiler, Transformation. Article 3 para 5 TEU broadly expresses that '[i]n its relations with the wider world, the Union shall uphold and promote its values and interests'. Reflecting upon the political dynamic of these values: Lei‐ no/Petrov, Values, with regard to the European Neighbourhood Policy.

<sup>250</sup> Security Council Resolution 1267(1999) of 15 October 1999; Security Council Reso‐ lution 1333(2000) of 19 December 2000; Security Council Resolution 1390(2002) of 16 January 2002; see also Security Council Resolution 1455(2003) of 17 January 2003.

The EU has implemented these resolutions by adopting Council Regula‐ tion 881/2002,<sup>251</sup> taking over the list of persons concerned (to be updated by the Commission), without informing these persons of the reasons for the asset freeze.<sup>252</sup> Therefore Mr *Kadi*, who found himself on the list, among others, filed an action for annulment with the then Court of First Instance, arguing that Council Regulation 881/2002 violated his right to a fair hear‐ ing, his right to property, and his right to effective judicial review.<sup>253</sup>

The Court of First Instance stressed the importance of Article 103 of the UN-Charter, pursuant to which '[i]n the event of a conflict between the ob‐ ligations of the Members of the United Nations under the present Charter and their obligations under any other international agreement, their obliga‐ tions under the present Charter shall prevail'.<sup>254</sup> This includes obligations emanating from the resolutions of the UN Security Council.<sup>255</sup> While this provision and the relevant case law of the ICJ lay down the primacy of the UN-Charter *vis-à-vis* the law of its members from the perspective of public international law, Community law – the Court of First Instance held – acknowledges this principle, as is expressed in particular in Articles 297 and 307 TEC.256 The European Community (EC), not being a member of the UN, is not bound to accept this primacy *qua* the UN-Charter, but it is bound to do so *qua* the TEC.<sup>257</sup>

Already in the original version of the Treaty establishing the European Economic Community (TEEC), namely in its Articles 224 and 234 para 1, the MS have expressed their intention to follow their obligations under the UN-Charter. As a consequence of the subordination to the UN-Charter

<sup>251</sup> This Regulation was based on Articles 60, 301 and 308 TEC. For the adequacy of this combined legal basis see case T-315/01 *Kadi v Council*, paras 89 ff and, partly differently, joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 158 ff; see also Schmalenbach, Kontrollanspruch 37, with further references.

<sup>252</sup> For an account of the relevant events – in particular the various output on the part of the UN and the EU, respectively – see joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 11–45.

<sup>253</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, para 49.

<sup>254</sup> See also *Nicaragua v United States of America*, ICJ Reports 1986, paras 107, to which the Court of First Instance refers in case T-315/01 *Kadi v Council*, para 183.

<sup>255</sup> Article 25 of the UN-Charter; see case T-315/01 *Kadi v Council*, para 184.

<sup>256</sup> See case T-315/01 *Kadi v Council*, paras 185 ff; see now Articles 347 and 351 TFEU.

<sup>257</sup> The motivation for adopting the mentioned TEC provisions lay in general public international law. Accordingly, the MS – when concluding the TEC – 'could not transfer to the Community more powers than they possessed or withdraw from their obligations to third countries under that Charter'; case T-315/01 *Kadi v Coun‐ cil*, para 195.

of the law of the MS and – *qua* primary law – also of Community law, the Court of First Instance eventually refused its general jurisdiction to scrutinise Council Regulation 881/2002, as it is implementing – without the Council thereby having disposed of any discretion – the relevant reso‐ lutions of the Security Council. Scrutinising Council Regulation 881/2002 would mean to indirectly examine these resolutions of the Security Coun‐ cil, and affirming this general jurisdiction would again, according to the Court of First Instance, be incompatible with public international law (in particular with Articles 25, 48 and 103 of the UN-Charter and Article 27 of the Vienna Convention on the Law of Treaties (VCLT) 1969), but also with the Treaties, in particular Articles 5, 10, 297 and 307 para 1 TEC and Article 5 TEU.<sup>258</sup> More generally speaking, it would not be in accordance 'with the principle that the Community's powers and, therefore, those of the Court of First Instance, must be exercised in compliance with international law'.259 Only with regard to *ius cogens* – the body of highest rules of public international law, to which also the UN-Charter has to submit – the Court of First Instance, it held, may scrutinise (indirectly via an examination of Council Regulation 881/2002) the resolutions of the Security Council. It is in particular the 'mandatory provisions concerning the universal protection of human rights' which belong to these supreme rules of public internation‐ al law.260 With regard to the rather loose standard *ius cogens* provides in this context,261 the Court of First Instance concluded – in short – that the asset freeze at issue does not constitute an arbitrary, inappropriate or disproportionate interference with the fundamental right to property of Mr *Kadi* (and others),262 nor have the applicable procedures brought about a breach of the right to be heard or a breach of the right to effective judicial review.263 Consequently, it dismissed the action brought against Council Regulation 881/2002.<sup>264</sup>

<sup>258</sup> See case T-315/01 *Kadi v Council*, paras 222 f. For cases of the European Court of Human Rights dealing with a similar question, namely whether or not certain state action can be assigned to the EU and the UN, respectively (the cases *Bosphorus*, *Behrami* and *Saramati*), see de Búrca, Court 11–17.

<sup>259</sup> Case T-315/01 *Kadi v Council*, para 223.

<sup>260</sup> See case T-315/01 *Kadi v Council*, para 231.

<sup>261</sup> See also Lenaerts, Kadi 709.

<sup>262</sup> Case T-315/01 *Kadi v Council*, para 251.

<sup>263</sup> Case T-315/01 *Kadi v Council*, paras 276 and 291.

<sup>264</sup> For a more detailed account of the reasoning of the Court of First Instance see Schmalenbach, Kontrollanspruch 35–37.

#### *II. SOFT LAW: TERMINOLOGY AND LOCALISATION*

Against the judgement of the Court of First Instance an appeal was filed before the Court of Justice. Remarkably but not entirely unexpectedly,<sup>265</sup> the Court of Justice in this case took a very different view on the question of primacy of UN law *vis-à-vis* Community law. It put the autonomy of the Community legal system and the fundamental rights as 'integral part of the general principles of law whose observance the Court ensures' at the centre of its reasoning.<sup>266</sup> Judicial review is applied to the Council Regulation at issue, not to UN law (in particular: the relevant resolutions of the Security Council), the Court underlined.<sup>267</sup> The Court, it held, may not even perform a scrutiny of UN law which is restricted to the question of compliance with *ius cogens*. <sup>268</sup> It further stresses that the procedure of implementation of resolutions of the UN Security Council is left up to the UN-MS and that therefore UN law does not prohibit any judicial review of the internal lawfulness of an implementing measure in the light of fundamental freedoms.269 The law of the EU submits to public international law in certain cases (eg in Articles 297 and 307 TEC), but these provisions cannot be understood 'to authorise any derogation from the principles of liberty, democracy and respect for human rights and fundamental freedoms enshrined in Article 6(1) [T]EU as a foundation of the Union'.270 Thus, the CJEU is competent to 'ensure the review, in principle the full review, of the lawfulness of all Community acts in the light of the fundamental rights forming an integral part of the general principles of Community law', even if these acts have been adopted in order to implement resolutions of the UN Security Council.<sup>271</sup>

<sup>265</sup> See Schmalenbach, Kontrollanspruch 36.

<sup>266</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 282 f and 316; also note AG *Maduro* in his Opinion in this case where he describes the EU legal order – in contrast to public international law – as a 'municipal legal order of trans-national dimensions' (para 21); for an account of *Maduro's* Opinion more generally see Gattini, Cases 216 f.

<sup>267</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, para 286; see also para 300, where the Court negates the 'immunity from jurisdiction of a Community measure like the contested regulation', stating that such immunity 'cannot find a basis in the EC Treaty'.

<sup>268</sup> Joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, para 287.

<sup>269</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 298 f.

<sup>270</sup> Joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, para 303. AG *Ma‐ duro* in this case puts it this way: 'Yet, in the final analysis, the Community Courts determine the effect of international obligations within the Community legal order by reference to conditions set by Community law' (para 23).

<sup>271</sup> Joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 285 and 326.

In substance, the Court of Justice held that Mr *Kadi's* rights of defence, especially the right to be heard, and the principle of effective judicial pro‐ tection, as well as his fundamental right of respect for his property have been infringed,<sup>272</sup> and subsequently annulled Council Regulation 881/2002 to the extent it concerned the claimants.<sup>273</sup>

Following the judgement of the Court of Justice in *Kadi I*, the Commis‐ sion sent Mr *Kadi* a brief summary of reasons (drafted by the UN Security Council Sanctions Committee), informing him that, on the basis of these reasons, it will adopt a legal act with a view to keeping his name on the list annexed to Council Regulation 881/2002 and giving him the opportunity to comment on these reasons.<sup>274</sup> Mr *Kadi* used this opportunity, request‐ ing the Commission to disclose the evidence supporting the assertions and allegations made in the summary of reasons and also the relevant documents in the Commission's file, requesting an opportunity to make representations on that evidence, once he had received it, and attempting to refute, thereby providing evidence, the allegations made in the summary of reasons.<sup>275</sup>

Subsequently, the Commission listed him again in Annex I to Council Regulation 881/2002 by means of Commission Regulation 1190/2008. Mr *Kadi* then filed an action against this Commission Regulation, as far as it concerned him, arguing – *inter alia* – that it infringed his rights of the defence, to effective judicial protection and to property. The Court of First Instance deemed the named arguments to be justified and annulled Commission Regulation 1190/2008 in so far as it concerned Mr *Kadi*. 276 As regards our focus in this discussion – the relationship between what is now EU law and public international law – the Court of First Instance elaborated on the relevant findings of the Court of Justice in *Kadi I* and concluded that '[s]o far as th[e] principles [of liberty, democracy and re‐ spect for human rights and fundamental freedoms277] are concerned, the Court of Justice […] seems to have regarded the constitutional framework created by the EC Treaty as a wholly autonomous legal order, not subject to the higher rules of international law – in this case the law deriving from the

<sup>272</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, paras 353 and 371.

<sup>273</sup> Similarly in joined cases C‑399/06P and C‑403/06P *Hassan*, paras 69 to 75.

<sup>274</sup> See case T-85/09 *Kadi*, para 53.

<sup>275</sup> See case T-85/09 *Kadi*, para 55.

<sup>276</sup> See case T-85/09 *Kadi*, paras 188 and 193–195.

<sup>277</sup> See joined cases C-402/05P and C-415/05P *Kadi and Al Barakaat*, para 303.

Charter of the United Nations'.278 The Commission, the Council and the UK filed an appeal against this judgement before the Court of Justice.

Against the suggestion of AG *Bot* in *Kadi II*, <sup>279</sup> the Court of Justice dismissed the appeals, arguing – similar to its reasoning in *Kadi I* – that the EU courts must ensure the principally full review of the lawfulness of all EU acts 'in the light of the fundamental rights forming an integral part of the [EU] legal order', including those created to implement resolutions of the UN Security Council adopted under Chapter VII of the UN-Charter.<sup>280</sup> With respect to the tension judicial review of (the Commission Regulation amending) Council Regulation 881/2002 creates with the required respect for UN law, the Court contended and (partly) repeated that '[j]udicial review of the lawfulness of the contested regulation is not equivalent to review of the validity of the resolution which that regulation implements. That review does not challenge either the primary responsibility of the Security Council in the area concerned or the primacy of the Charter of the United Nations over any other international agreement. […] Its purpose is solely to ensure observance of the requirement that Security Council Resolutions are implemented within the European Union in a manner compatible with the fundamental principles of European Union law. More specifically, such review contributes to ensuring that a balance is struck between the requirements of international peace and security, on the one hand, and the protection of fundamental rights, on the other'.<sup>281</sup>

#### 2.1.3.2.3. The effect of WTO law in the EU legal order

Another legal discussion bearing witness of the difficulty to clearly separate law from soft law is the treatment of WTO law in the EU legal order, more precisely the fact that the CJEU in its case law denies the capability of WTO law to have direct effect, wherefrom it concludes that WTO law may

<sup>278</sup> Case T-85/09 *Kadi*, para 119.

<sup>279</sup> Opinion of AG *Bot* in joined cases C‑584/10P, C‑593/10P and C‑595/10P *Kadi*. AG *Bot* supported the Court's approach on the relationship between EU law and UN law as established in *Kadi I*, but in substance it did not deem the claims of Mr *Kadi* to be justified in this case.

<sup>280</sup> Joined cases C‑584/10P, C‑593/10P and C‑595/10P *Kadi*, para 97.

<sup>281</sup> Joined cases C‑584/10P, C‑593/10P and C‑595/10P *Kadi*, para 87.

not serve as a standard of review for the legality of EU secondary law.<sup>282</sup> While in comparison to national legal orders this approach is by no means exceptional, it was most prominently discussed in the context of the case law of the CJEU.<sup>283</sup>

Whether or not a legal norm has direct effect in general depends on the will of the norm-creator.<sup>284</sup> Sometimes the norm-creator does not explicitly utter its will in this respect. In these cases the will of the norm-creator is to be deduced from the wording, the degree of precision and the structure of the norm. Usually, these factors may reasonably be interpreted differently, which is why the question of direct effect is regularly contested – at least where a legal order does not provide for a highest legal authority which could clarify this question once and for all.<sup>285</sup> Against this background, *Klabbers* has described the concept of direct effect as 'little else but a half-hearted doctrine giving courts a free hand in deciding which norms of international law to allow into their legal order'.<sup>286</sup>

Also WTO law does not contain an explicit provision regarding this question. A proposal of Switzerland to incorporate a direct effect clause

Alternatively, like in the US, the (national) legislator could – only for the state at issue, of course – expressly decide on the question of direct effect; see Cottier, Theory 105; Trachtman, Direct Effect 657.

<sup>282</sup> Criticising this strong focus on direct effect: Klabbers, Community Law; see ibid 284 ff, pointing to only *prima facie* exceptions in the Court's case law and relativis‐ ing the Court's role in this context.

<sup>283</sup> See Ruiz Fabri, Case 152. For different nuances of these approaches see ibid 153.

<sup>284</sup> Under CETA, to take this example, the parties to the agreement have explicitly excluded its having direct effect (in their respective domestic legal systems); see Article 30.6.1 *leg cit*. For investors under CETA's Investment Court System (and the respective CETA provisions) the situation is different.

<sup>285</sup> The MS of the WTO and the EU do have a highest legal authority, namely the respective highest courts; see Opinion of AG *Mayras* in cases 21–24/72 *International Fruit Company*, 1234: 'The unity and, it can be said, the very existence of Commun‐ ity law require that the Court is alone empowered to say, with the force of law, whether an agreement binding the Community or all the Member States is or is not directly applicable within the territory of the Community and, if it is, whether or not a measure emanating from a Community institution conforms to that external agreement'.

<sup>286</sup> Klabbers, Community Law 264. Acknowledging this fact, at least implicitly: case C-431/05 *Merck*, para 47. See also case 104/81 *Hauptzollamt Mainz*, para 17, stressing that only if the question of direct effect is not settled by the international agreement at issue, it 'falls for decision by the courts having jurisdiction in the matter, and in particular by the Court of Justice within the framework of its jurisdiction under the Treaty'.

was refused during the Uruguay round.287 This could be understood as the absence of an according will of the members of the WTO, taken as a whole, and hence as an indication of WTO law's lack of direct effect.<sup>288</sup> Nevertheless, it is widely acknowledged that the municipal courts of the WTO members within their respective jurisdiction can decide themselves on whether or not certain provisions of WTO law should be granted direct effect or not.

The CJEU is one of these municipal courts. Its case law on the question of direct effect of what since 1995 is called WTO law, shall stay in the foreground here. In this context, also the Court's more general approach towards the direct effect of international agreements is to be considered.

As early as in 1972, the Court has dealt with the question of direct effect with a view to provisions of the GATT 1947. As in *van Gend & Loos*, it stressed the necessity to consider 'the spirit, the general scheme and the terms' of the legal act at issue, in our case: the GATT.289 In view of the fact that many States – important trading partners of the Community – denied direct effect, reciprocity considerations may also have played a role.290 While accepting that the Community is bound by the GATT,<sup>291</sup> essentially in view of 'the great flexibility of its provisions, in particular those conferring the possibility of derogation, the measures to be taken when confronted with exceptional difficulties and the settlement of con‐ flicts between the contracting parties' the Court has denied the direct effect of GATT 1947/1994 provisions292 (and also of other WTO law293).

<sup>287</sup> See Ruiz Fabri, Case 154, with a further reference.

<sup>288</sup> See Panel Report of 22 December 1999, *United States – Sections 301–310 of the Trade Act of 1974*, WT/DS152/R, paras 7.72 ff, with regard to the 'open question' whether there are certain rights of individuals under WTO law which national courts have to protect, and with further references (fn 661).

<sup>289</sup> Cases 21–24/72 *International Fruit Company*, para 20.

<sup>290</sup> See case C-149/96 *Portugal v Council*, para 43; see also Klabbers, Community Law 278, with further references.

<sup>291</sup> For the GATT's qualification as Community law see case C‑386/08 *Brita*, para 39, with further references; for the questions which arose due to the fact that the EEC was not a 'member of ', but only a 'participant in' the GATT see Constantinesco, Recht 217 f.

<sup>292</sup> Cases 21–24/72 *International Fruit Company*, paras 18 and 21; for further explana‐ tions see case C-280/93 *Germany v Council*, paras 105–110, with further references; for the latter judgement see also Everling, Europe.

<sup>293</sup> See joined cases C-300/98 and C-392/98 *Dior*, para 45, with regard to TRIPS.

In spite of the reform following the Uruguay round, leading to more nuanced rules of the new GATT and other new agreements, and the more effective dispute settlement mechanism of the WTO, the Court upheld this case law.<sup>294</sup> In fact, the Court has not only denied the possibility of individ‐ uals/undertakings to invoke WTO law when claiming the illegality of an EU law act, it has refused to use WTO law as a standard of judicial review of EU law more generally.<sup>295</sup> That is to say that WTO law in principle may not be invoked by the institutions or the MS, either.<sup>296</sup> Neither may a violation of WTO law on the part of EU institutions lead to the EU's noncontractual liability.<sup>297</sup> Even the output of the Dispute Settlement Body, the Court held, principally is to be treated in the same way as the WTO agreements.<sup>298</sup> Only exceptionally, WTO law may be invoked, namely if EU law makes express reference to specific and precise provisions of WTO law299 or if EU law is clearly aimed at implementing WTO law<sup>300</sup> . 301

The far-reaching denial of direct effect of WTO law – but not only of WTO law302 – by the CJEU is in contrast to its case law on some other international agreements, in particular association agreements with

297 See eg case C-104/97P *Atlanta*, para 66.

<sup>294</sup> See case C-149/96 *Portugal v Council*, para 36, acknowledging these novelties as compared to the GATT 1947, and para 47, refusing to review the legality of Com‐ munity acts in the light of the WTO agreements.

<sup>295</sup> Pointing at the conceptual difference between direct effect and the review of legali‐ ty: Klabbers, Community Law 265 and 268.

<sup>296</sup> See case C-149/96 *Portugal v Council*, para 47; with regard to the MS see also Klab‐ bers, Community Law 265 (fn 10); Ruiz Fabri, Case 158, with a further reference; arguing that the Court – implicitly – has used different standards for MS and institutions as privileged claimants under Article 263 TFEU: Holdgaard, External Relations 270, with further references.

<sup>298</sup> See joined cases C-120/06P and C-121/06P *FIAMM*, paras 128 ff; see also references by W Weiß, Art. 207 AEUV, para 203.

<sup>299</sup> See case 70/87 *Fediol*, para 22; for the technique of referencing more generally see 2.2.3. below.

<sup>300</sup> See case C-69/89 *Nakajima*, paras 29–32; see Herrmann/Glöckle, Handelskrieg 482 f, with references to the follow-up case law.

<sup>301</sup> See Ruiz Fabri, Case 158 f, with regard to 'indirect effect' – that is the interpretation of EU law in accordance with WTO law – which may also be seen as a way to increase the effectiveness of WTO law within the EU legal order; see also case C-53/96 *Hermès*, para 28; case C-308/06 *Intertanko*, para 52. Referring to direct effect and consistent interpretation as modes of interaction between two legal orders see de Búrca, Court 39 f, with further references.

<sup>302</sup> See case C-308/06 *Intertanko*, paras 54 ff, with regard to UNCLOS; case C-363/12 *Z.*, paras 85 ff, with regard to the UN Convention on the Rights of Persons with Disabilities.

### *II. SOFT LAW: TERMINOLOGY AND LOCALISATION*

third countries, in which it has confirmed the direct effect of provisions contained in such international agreements,<sup>303</sup> thereby relying on wording, purpose and nature of the agreement at issue. But also in the case of other bilateral free trade agreements concluded between the EC and third countries the Court has confirmed direct effect of selected provisions.<sup>304</sup> In view of this discrepancy, the *principal* denial of direct effect of provisions of WTO law has been much criticised in legal scholarship as a political rather than a legal decision.<sup>305</sup>

### 2.1.3.2.4. Discussion

The two phenomena presented above relate to the discussion on the prac‐ tically often difficult separation between law and soft law in a number of respects.306 Before dwelling on this relationship, it makes sense to provide for a comparison of the two issues *inter se*. In both cases a multilevel legal situation, more concretely the relationship between the EU and a measure of public international law stays in the foreground. The EU courts have played a pivotal role in shaping this relationship, thereby determining the effects public international law has – not in general, but only in a relative manner, namely: to the extent it obliges the EU. This shaping, in *Lenaert's* words, 'is the result of a balancing exercise between safeguarding the EU's constitutional identity and making sure that EU law does not become

<sup>303</sup> See eg case 87/75 *Bresciani*, para 23; case C-162/00 *Land Nordrhein-Westfalen*, paras 19 ff; case C-464/14 *SECIL*, paras 97 ff; case T‑798/14 *DenizBank*, para 144.

<sup>304</sup> See case 104/81 *Hauptzollamt Mainz*, para 27; case C-162/96 *Racke*, para 34. *De Búrca* argues that with regard to international agreements forming an 'integral part' of the EU legal order (this common phrase primarily points at the effects laid down in Article 216 para 2 TFEU) 'the ECJ has almost always declared [with the exception of WTO law] that international agreements entered into by the EC are directly enforceable before domestic courts'; de Búrca, Court 46.

<sup>305</sup> See Klabbers, Community Law 264 and *passim*, with many further references. Note also the words of AG *Cosmas* he uttered in his Opinion in joined cases C-300/98 and C-392/98 *Dior*, para 76, talking about an 'alternative legal framework often marked by a lack of strictness (soft law). That is neither paradoxical nor contradic‐ tory. It is justified by the variable geometry and the still incomplete institutionalisa‐ tion of the coexistence of national, Community and international legal orders. In the context of that institutionalisation, law and politics exchange characteristics: the former imposes its strict and binding nature on the latter and the latter in turn instils its relativity and flexibility in the former'.

<sup>306</sup> As well addressing these two (sets of ) cases together: Nollkaemper, Role 188 ff.

hostile to the international community, but that it is an active part of it'.<sup>307</sup> While the Court has acknowledged in principle that the EU is bound by the relevant acts of international law (resolutions of the UN Security Council and WTO law respectively), it has, in different ways, drawn limits to the ensuing obligations of the EU. In case of the resolutions of the Security Council, the Court has determined that – in spite of the unconditionally drafted primacy clause contained in Article 103 UN-Charter – their imple‐ mentation by means of Union law may not lead to a violation of core principles of the EU such as liberty, democracy and respect for human rights and fundamental freedoms. The compliance with these limits of EU (secondary) law implementing the resolutions of the UN Security Council shall be scrutinised by the CJEU. With regard to WTO law, the Court has confirmed that, as far as the EU is concerned, it forms part of EU law. However, it has repeatedly refused to test the legality of EU (secondary) law against WTO law by negating the latter's direct effect.

In the two cases the alleged obligations of the EU emanating from public international law are 'mitigated' by different techniques. As regards the resolutions of the UN Security Council, the EU will implement them (if it is competent to do so), and the Court will review the EU implementing acts in terms of competence (formal element) but also with regard to material (minimum) requirements. These minimum requirements are notably roo‐ ted in EU law, not in public international law.308 In the case of WTO law the technique is of a formal kind: It is the impossibility for claimants to invoke (violations of ) WTO law in procedures addressing acts of EU secondary law which leads to the immunity of EU secondary law in this respect. This does not materially alter the EU's obligations, but procedurally it prevents violations of these obligations from being decided upon by the CJEU. The underlying reason that WTO law may not be invoked again has a material stance to the extent that it is based on 'the spirit, the general scheme and the terms' of WTO law.309 Pursuant to the Court, the limit is rooted in WTO law, as it is the very nature of its provisions which prevents them from having direct effect. However, this view is contested.

In summary, we can say that, broadly speaking, the effectiveness of public international law is, by the application of different methods, reduced

<sup>307</sup> Lenaerts, Kadi 708.

<sup>308</sup> With the competence as a formal limit this is a matter of course, but with respect to the substantive standard of review it is remarkable.

<sup>309</sup> Cases 21–24/72 *International Fruit Company*, para 20.

in both cases. But what about the bindingness of the norms at issue? As regards the resolutions of the Security Council, the actual result of the *Kadi* cases is that the EU shall not be bound by resolutions going against certain fundamental principles of the EU. Even though the Court avoids expressing this and rather dwells on the fact that the resolutions leave their respective implementation up to the addressees (here: the EU), the Court's approach results in the restriction of the resolutions' legal bindingness to the extent they conflict with said fundamental principles of the EU. To this extent, the primacy of the resolutions – which allegedly is upheld by the Court<sup>310</sup> – in fact is neglected.

In case of WTO law, its lacking capability to serve as a standard of review for EU (secondary) law lies at the core of the issue. The Court, it was said, 'displays a certain sympathy toward international law while nevertheless focusing on fundamental principles of the Community's domestic legal order as the ultimate rule against which the legality of Community action must be judged'.311 In a number of judgements the Court has stressed that this does not alter the fact that the EU is bound by WTO law, and that the latter forms part of EU law.312 It rather camouflages this stated idiosyncrasy of WTO law as an answer to the question of *how* to comply.313 However, if conflicting EU law cannot be reviewed – in that respect – before the EU courts, within the framework of the EU this comes close to a non-enforcea‐ bility314 of WTO law before the CJEU.<sup>315</sup>

There are even more radical ways to perceive the relationship between EU law and WTO law. So far we have addressed the lack of direct effect of WTO law in the jurisdiction of the CJEU. But the legal 'independence' of the EU legal order from WTO law may be depicted in more unorthodox terms. While the above account of the Court's case law was based on the assumption that the EU in principle is bound by WTO law, legal scholar‐

<sup>310</sup> See also case C‑548/09P *Melli*, para 105, with reference to *Kadi I*, stressing the 'primacy of a Security Council resolution at the international level', while insisting on its duty to review the lawfulness of Community measures.

<sup>311</sup> Halberstam/Stein, United Nations 31; see also Lenaerts, Kadi 712.

<sup>312</sup> See Herrmann/Glöckle, Handelskrieg 482, with references to the Court's case law.

<sup>313</sup> See Ruiz Fabri, Case 168.

<sup>314</sup> Emphasising that non-self-executing norms do not entail legal obligations: Baxter, International Law 552–554.

<sup>315</sup> For the enforceability of WTO law *vis-à-vis* the MS, however, see case C-66/88 *Commission v Hungary*. Within the framework of the WTO, a violation of WTO law can be made subject to its dispute settlement procedure.

ship provides for an alternative understanding of the relationship between these two legal orders. According to this view, compliance with WTO law is one possibility, but agreeing on compensation or accepting retaliatory measures is legitimate as well. This would mean that WTO law is not 'categorically binding', but that the WTO regime, apart from compliance with the substantive law by its addressees (which apparently is the preferred behaviour), explicitly provides for (lawful) alternatives.<sup>316</sup> While this view can be applied more generally, that is to say with regard to all members of the WTO regime (States as well as the EU<sup>317</sup>), here it is to be considered with a view to the EU.

Remarkably enough, some of the later Court judgements can indeed be read as supporting this idea,<sup>318</sup> which is also referred to as 'efficient breach theory'.<sup>319</sup> It proposes a new perception of the effects of WTO law (as fleshed out by the courts of the WTO members, eg the CJEU) rather than actually suggesting new effects. While this theory may be used as just another argument in favour of denying the direct effect of WTO law, its entertainment certainly would go beyond that and create a tension with the Court's body of case law on this issue. This is because – the hints just mentioned notwithstanding – the Court in its judgements has explicitly acknowledged the EU's being bound by WTO law (see above). Thus, it ap‐ pears more likely that the Court will stick to its case law, according to which the EU is bound by WTO law (to which it does not accord direct effect), rather than disavow the rule of law by describing the regime of the WTO as 'voluntary, with potential negative effects in case of non-compliance'.

In the perspective of the 'efficient breach theory', WTO law – *qua* being legally non-binding – would come very close to soft law. The severe sanc‐ tions non-compliance may entail in the course of a proceduralised regime – laid down in the Dispute Settlement Understanding – sit oddly with the claim for non-bindingness, though. Charming as the 'efficient breach theory' may be, in my view the legal bindingness of WTO law can be upheld with good reasons. Its widely purported lack of direct effect bars its enforcement (by individuals/undertakings), and hence qualifies it as a special case, but it does not render it legally non-binding. *Trachtman* in this

<sup>316</sup> See Griller/Vranes, EC-Bananas, para 20.

<sup>317</sup> See Article XI para 1 of the WTO Agreement: 'European Communities'.

<sup>318</sup> See eg case C-149/96 *Portugal v Council*, paras 35 ff; see also references in Grill‐ er/Vranes, EC-Bananas, para 22.

<sup>319</sup> See Ruiz Fabri, Case 168 f, with further references.

context said that 'within each society, there exist different kinds of law, with different types and degrees of binding force. There is no "natural" condition of law'.<sup>320</sup> While the author does not follow the idea of degrees of legal bindingness (see 1.3.2. above), it is to be conceded that there are different forms in which legal bindingness may take effect.

In conclusion, both the resolutions of the UN Security Council as dealt with in the *Kadi* cases, and the refusal of direct effect of WTO law by the CJEU relativise the effects of the legal acts at issue. This is brought about by a mitigation of legal authority which is comparable – but not equal – to the phenomenon of soft law. While the latter lacks legal bindingness, in our two examples the scope of legal bindingness is merely *restricted* (resolutions of the UN Security Council) and possibilities of judicial enforcement are ex‐ cluded (WTO law), respectively. By all means, these cases illustrate – once more – the challenge, but also the necessity of drawing a clear (conceptual) line between law and soft law.<sup>321</sup>

### 2.2. From other sets of norms

### 2.2.1. Custom and customary law

Custom describes a set of behavioural patterns habitually performed by a human society.322 Custom may encompass behaviour as diverse as a religious or a profane ceremony, a certain salutation, courtesy rules or a recurrent local sports event. It is a habit which has developed over time and hence can be called traditional. A custom comes into being gradually, by continuous application. Normally also its coming out of practice is subject to an extended period of time in the course of which it is applied less and less frequently, until it vanishes entirely. Exceptionally, it may end abrupt‐ ly due to legal prohibition, eg the annual large paschal bonfire may be

<sup>320</sup> Trachtman, Direct Effect 655; acknowledging a 'natural condition' of law which is, however, 'rough and imperfect, like our society, and like us'; ibid 677.

<sup>321</sup> See also Cannizzaro/Rebasti, Soft law 217.

<sup>322</sup> See the definition of *habitus* provided by Bourdieu, Logic 53: 'systems of durable, transposable dispositions, structured structures predisposed to function as structur‐ ing structures, that is, as principles which generate and organize practices and rep‐ resentations that can be objectively adapted to their outcomes without presupposing a conscious aiming at ends or an express mastery of the operations necessary in order to attain them'.

banned for fire safety concerns or a religious gathering may be interdicted as a means of political repression. A custom regularly is not prescribed in writing. There may be official or semi-official notes on the performance of a custom (eg chronicles), but they are of a merely declarative nature. A custom as such cannot be legally enforced.<sup>323</sup> Should this be possible exceptionally, the custom has become customary law, an unwritten kind of law arising where, in addition to the custom as such (the terms *usus* or *consuetudo* in this context describe the regular performance of the habit), the constitutive requirement *opinio iuris*, that is the general opinion that the performance of the habit actually is required by a legal rule, is met;<sup>324</sup> or, in the words of the ICJ (with regard to public international customary law): the 'belief that this practice is rendered obligatory by the existence of a rule of law requiring it. […] The States concerned must therefore feel that they are conforming to what amounts to a legal obligation. The frequency or even habitual character of the acts is not enough'.325 In public international law, persistent objection by an actor against such a habit can prevent the development of an *opinio iuris* and hence its becoming binding upon the objector.326 Where customary rules are at issue, in principle the general distinction between law and soft law would apply (see 2.1. above). However, while detecting an *opinio iuris* regularly is a demanding task,327 it may be nearly impossible to prove the even more nuanced (mostly implicit) convic‐ tion of the relevant actors that a certain custom actually constitutes soft law. Thus, in practice there does not seem to be much room for customary soft law.

Conceptually speaking, the delimitation of soft law from customary law is relatively easy. Soft law is legally non-binding and non-enforceable, whereas customary law – as law proper – is both legally binding and (regularly) enforceable. Also soft law and custom in theory can be easily distinguished from each other, since they have few things in common, most

<sup>323</sup> For usages in international relations see Bothe, Norms 67.

<sup>324</sup> See eg Article 39 para 1 lit b of the Statute of the ICJ: 'general practice accepted as law'; for the origins of this concept of customary law see J Schröder, Theorie 222 f. On the difficulties to prove the existence of these elements see Knauff, Rege‐ lungsverbund 231f. A custom may also be incorporated in written law and thus be considered law; Walter, Soft Law 30 f.

<sup>325</sup> *Germany v Denmark, Germany v Netherlands*, ICJ Reports 1969, para 77.

<sup>326</sup> See eg Nußberger, Völkerrecht 24.

<sup>327</sup> With regard to the fact that customary law (regularly) is not created intentionally see Klabbers, Community Law 289.

importantly that they both entail a non-legal quality of normativity. The 'creation', ie the development, of a custom is a non-intentional process performed by (parts of ) society most of the time, whereas the adoption of soft law, leaving apart the rather abstract possibility of customary soft law, constitutes a conscious act of an entity vested with public authority. This is reflected in the fact that a custom does not arise in written form, whereas soft law is at least regularly laid down in writing.<sup>328</sup>

While the conceptual difference seems clear, in practice these three nor‐ mative systems – custom, customary law, soft law – may be closely linked to each other. Custom and customary law are separated only by the actors' *opinio iuris*. As a 'psychologisches Element' [psychological element]<sup>329</sup> it is often difficult to prove its existence. Here soft law may come into play. The existence of soft law regulating a certain issue as a 'compromise over time'330 may serve – together with other indicators – as evidence proving the existence of an *opinio iuris* (see 1.3.2. above). These indicators need to be assessed carefully, in particular the assumed elevation of the will to create soft law to the conviction that the underlying rules constitute law.<sup>331</sup> Also with regard to the *usus* – in public international law that is above all State practice and the practice of international organisations – the adoption of soft law acts and the respective compliance may be considered – again: carefully – as evidence.332 Supporters of the idea of 'instant custom' 333 – that is to say customary law which, due to an overwhelming amount of agreement, is not requiring any evidence of *usus* – would maybe qualify certain acts of EU soft law, namely those adopted by consensus,334 as instant custom. While this view – which can be contested with good reasons – is not considered here any further, it can be said upfront that in the given context it would not entail conceptual difficulties. If, according to this view, it qualifies as EU customary law335 it is legally binding, if it qualifies as EU soft law it is not. It cannot be both at the same time.

<sup>328</sup> See Knauff, Regelungsverbund 232, with partly deviating references.

<sup>329</sup> Knauff, Regelungsverbund 232.

<sup>330</sup> Abbott/Snidal, Hard and Soft Law 444.

<sup>331</sup> In a similar context: Arndt, Sinn 49, with further references; Wengler, Rechtsvertrag 194.

<sup>332</sup> See Shelton, Introduction 1; Wittinger, Europarat 208 f; principally in the affirma‐ tive, but – for the time being – sceptically: Arndt, Sinn 46.

<sup>333</sup> For the concept of 'instant custom' see eg Weil, Normativity 435 f.

<sup>334</sup> See Petersen, Customary Law 281, with further references.

<sup>335</sup> For the possibility of customary EU law see Klabbers, Instruments 1015.

# 2.2.2. Morals

The existence of general moral norms is most difficult to prove. This is not only due to the fact that moral norms are very individual – everybody may set moral norms for him- or herself – and may therefore differ from person to person. (Having said that, many human societies claim to have a common moral ground, that is to say fundamental moral convictions which the members of these societies share.) What is more, they are more often than not applied in a private space, beyond public recognition – in a *forum internum*. <sup>336</sup> Even where visible actions are motivated by moral convictions, the content of these convictions cannot always reliably be deduced from the action. The driver of a car may give way to another car at a junction in spite of his own legal right of way for different reasons: He may think it is polite, he may intend to do good to other people, he may want to use the 'gained' time to read a message on his phone, he may want to have a closer look at the person in the other car, he may, as a reactant person, gain pleasure from the fact that he is ignoring a legal rule (even though that rule does not oblige but only entitles him), namely his right of way, etc.337 This difficulty to find out about people's moral convictions – that is their interiority – distinguishes it strongly from other sets of norms.

A custom finds its expression in (exterior) behaviour, customary law does as well,<sup>338</sup> soft law and non-customary law most frequently occur in written form. Morals in general do neither. That it may be moral convictions which – alone or in combination with other considerations – determine the content of a legal or a soft law norm,339 as is most obviously the case eg with human rights provisions, does not have any impact on their normative

<sup>336</sup> See also Bothe, Norms 95.

<sup>337</sup> See Goldmann, Perspective 58, with reference to the 'dual function of law'; see also Goldmann, Gewalt 364 f.

<sup>338</sup> As we have seen, even the (internal) *opinio iuris* is regularly established with refer‐ ence to external acts.

<sup>339</sup> See Habermas, Faktizität 137; see also Bianchi, Butterfly 200, who stresses the structural similarity of law and soft law, also with regard to their respective content. There are various scholars – of different schools – who emphasise that in case of a strong immorality of a law, the latter ceases to exist as a legal provision; see only the famous *Radbruch formula*: Radbruch, Unrecht. Sometimes the addendum 'moral' ought to express that a rule is legally non-binding, hence (potentially) soft law; see eg Hockin, World Trade Organization 258, with regard to Article 1114 para 2 of the NAFTA: 'These are "pure moral imperative" intentional clauses, at least on behalf of the environment and safety'.

quality as law or soft law.<sup>340</sup> Neither does the knowledge that an actor only complies with law or soft law because these rules comply with his/her moral convictions change the fact that his/her – external – actions are in accordance with these legal or soft law rules.<sup>341</sup>

# 2.2.3. Regulation by private actors: the example of standards

Standard-setting by private actors as one example of regulation by private actors,<sup>342</sup> mostly concerning standards of a 'technical' nature, has become more and more common in the past decades – at the national, and to an increasing extent at the EU and on the international level.<sup>343</sup> Partly these private norms<sup>344</sup> have gained considerable momentum, not least due to the increasing economic cooperation between States and, following from that, an outright globalisation of markets.345 These private norms – often referred to as 'standards'346 – are used in a variety of fields, most important‐ ly with regard to product characteristics.347 They are drafted by private actors (without public authority348), mostly organisations, and hence *per*


<sup>340</sup> For the importance of such a conceptual distinction see eg Thaler, Verhältnis.

<sup>341</sup> For the potential 'moral […] effect' of public international soft law see Scherm‐ ers/Blokker, Institutional Law, § 1238.

<sup>342</sup> For the discussion on how to classify private rule-making see overview given by Goldmann, Perspective 48–50, with many further references; as an example for a multiplicity of private bodies operating as a hub for regulatory action take the ENTSO; for its considerable regulatory influence on the elaboration of network codes see in particular Article 58 of Regulation 2019/943.

<sup>343</sup> For the economic developments leading to a drastic increase of EU and interna‐ tional standard-setting since the 80s see Mattli/Büthle, Standards 1–3; C Scott, Government 168 f; see also Türk, Lawmaking 83; for the inclusion of private actors in public policy-making in Europe see Héritier, Modes.

<sup>344</sup> For the specific meaning of the term 'norm' in this context – as opposed to the general understanding in legal science – see Griller, Normung 7 f.

<sup>345</sup> See Schepel, Constitution 2; for the role of private regulation in the pharmaceutical sector see Stenson/Syhakhang/Stålsby Lundborg/Eriksson/Tomson, Pharmacy.

<sup>346</sup> This term is not necessarily indicative of private norm-setting. Also public bodies, eg the ILO, a UN sub-organisation, may adopt standards; for international environ‐ mental standards (and the modalities of their creation) see Parker, Norms 182 f; for the standards set by the ILO (in the form of conventions or recommendations) see Trebilcock, Trade Policy 175.

*se* cannot be qualified as (soft) law.<sup>349</sup> What makes them come close to or even reach the status of public (soft) regulation is their public recognition (which, of course, differs in degree from case to case<sup>350</sup>). This on average high public recognition arguably results from the technical authority these standards bear, but also from the high compliance rates on the part of relevant economic actors they have reached in the past.

With regard to the international level, exemplarily the standards of the International Organization for Standardisation (ISO) ought to be men‐ tioned. This is a 'global network'<sup>351</sup> of (national) standards bodies from cur‐ rently 168 countries, organised as an NGO, which drafts and subsequently sells standards.<sup>352</sup> Another example is the International Electrotechnical Commission (IEC), also comprising a network of National Committees and setting international standards within its field.

Also at the EU level such private standardisation takes place – again by a network of national standardisation bodies, but also with a strong involvement of the European Commission and stakeholder groups.353 The European Standards Organisations (ESOs) – the Comité Européen de Normalisation (CEN), the Comité Européen de Normalisation Eléctrotech‐ nique (CENELEC) and the European Telecommunications Standards Insti‐

<sup>349</sup> See Somek, Concept 988.

<sup>350</sup> See Knauff, Regelungsverbund 243, with further references, who calls for a case by case assessment of technical standards and product certificates, also with regard to their (potential) qualification as soft law.

<sup>351</sup> Mattli/Büthle, Standards 4; for the implementation of standards of the ISO on the national level see Roht-Arriaza, Soft Law 274–279.

<sup>352 &</sup>lt;https://www.iso.org/about-us.html> accessed 28 March 2023. A standard is defined by the ISO as 'a document that provides requirements, specifications, guidelines or characteristics that can be used consistently to ensure that materials, products, processes and services are fit for their purpose'; <http://www.iso.org/iso /home/standards.htm> accessed 28 March 2023 (this definition can now be found eg at <https://carbonnumbers.co.uk/iso-standards-and-certifications/#:~:text =An%20ISO%20standard%20is%20a,are%20fit%20for%20their%20purpose.> ac‐ cessed 28 March 2023); see also Friedrich, Soft law 187–189; Roht-Arriaza, Soft Law 263; Wilkie, Governance 294. For the historical development of selected national and international standard-setting bodies, in particular the ISO, see Mattli/Büthle, Standards 6–8.

<sup>353</sup> See H Hofmann, Normenhierarchien 237; for early contractual relations between the then EEC on the one hand and CEN/CENELEC on the other hand see Erhard, Probleme 27 f; for co-regulation and self-regulation in the EU, and for the limits to the former, for standardisation and the respective EU legislation, for the organ‐ isation of the ESOs, the legitimacy of standardisation, its supervision, and other aspects of standardisation see Hofmann/Rowe/Türk, Administrative Law 587–605.

tute (ETSI) <sup>354</sup> – are private<sup>355</sup> associations established under Belgian and, respectively, French (ETSI) private law – NGOs<sup>356</sup> which draft 'European Standards'.<sup>357</sup> They are connected with both MS and international stand‐ ardisation bodies,<sup>358</sup> but also cooperate strongly with the EU,<sup>359</sup> from which they receive standardisation assignments.<sup>360</sup> They have played an important role in complementing EU internal market law ever since the 1980s.<sup>361</sup>

Also at the national level standards – in general – play an important role. National standardisation bodies cooperate intensely with and largely take over the standards established by the standardisation bodies at the EU and at the international level.<sup>362</sup> The generation procedures, the influence exerted by public bodies and the legal qualification of these standards vary considerably, though.<sup>363</sup>

While a *general* legal qualification of standards is – due to their multi‐ plicity – impossible, an attempt to come to grips with standards from a legal point of view may be the following. Principally, a legal evaluation of

<sup>354</sup> These three bodies mirror the respective international standardisation bodies. The CEN, for example, mirrors the ISO; <https://www.cenelec.eu/aboutcenelec/whowe are/europeanstandardsorganizations/index.html> accessed 28 March 2023; see also Senden, Self-Regulation 13.

<sup>355</sup> They are neither established under public international law nor are their founding members subjects of public international law; see Griller, Normung 49 f.

<sup>356</sup> See Griller, Normungsinstitut 279.

<sup>357</sup> For the role of such standards as technical barriers to trade see Erhard, Probleme 23–29; for an early quantitative account of these 'European standards' see Falke, Standardization 654 f.

<sup>358</sup> See Hofmann/Rowe/Türk, Administrative Law 597; Craig/de Búrca, EU Law 658– 661; Peters, Typology 418 f.

<sup>359</sup> See eg the 2009 Framework Partnership Agreement between CENELEC and the European Commission and EFTA.

<sup>360</sup> For the cooperation between the then European Communities, CEN and CENE‐ LEC, in particular EC requests to CEN/CENELEC for the (paid) elaboration of standards in a certain field see Griller, Normung 23–27; for European agencies as intra-EU bodies setting standards see H Hofmann, Union 460.

<sup>361</sup> See Colombo/Eliantonio, Standards 324; Volpato, Effects 195. For the increased importance the standards of these bodies have gained in the aftermath of the 1985 White Paper 'Completing the Internal Market' of the then newly appointed Delors Commission see Eilmansberger, Binnenmarktprinzipien 261; Snell, Internal market 344.

<sup>362</sup> For the duty of the standardisation bodies of the MS to take over European Stand‐ ards as their own (even if not adopted by unanimity/consensus, but only by a majority) see Rule 2.5. in Part 2 of the Rules of Procedure of CEN/CENELEC.

<sup>363</sup> For an assessment of the situation in selected countries of the EU see Schepel, Constitution 112 ff.

standards can take place on at least two levels.364 First, the way they are drafted and the persons or bodies involved can be considered. Where it is only (private) actors without an according public authority involved, the standards they draft as such have no legal, not even a soft law quality.<sup>365</sup> Where, on the contrary, standardisation bodies are vested with the public authority to adopt such standards, the latter – where they are not to be qualified as law anyway – regularly meet the criteria of soft law.<sup>366</sup> The question of whether or not such public authority exists may, in places, give rise to doubts: eg where State actors (with the respective public authority) are strongly involved in standard-setting undertaken by (formally) private bodies,<sup>367</sup> or where they are even double-hatted in that they participate in 'private' standard-setting, but as well in the creation of laws which again refer to these standards, elevating their content to the level of (soft)

<sup>364</sup> Critically as regards including 'the standardization process in a hierarchy of norms': Türk, Lawmaking 84.

<sup>365</sup> See Braams, Koordinierung 134 f. For the qualification of these standards as facts for example by the Austrian *Verfassungsgerichtshof* see Holoubek/Potacs, Technikrecht 68; see also Griller, Normung 9. Also the CJEU, with regard to a private association of experts in the field of gas and water, appears to refer to a merely factual authority of its output; case C-171/11 *Fra.bo*, para 31.

<sup>366</sup> For the case-by-case analysis required with regard to the criterion of public au‐ thority see Knauff, Regelungsverbund 243 (fn 201). For the EU level see Opinion of AG *Campos Sánchez-Bordona* in case C-613/14 *Elliott*, para 55, with a further reference. At the international level, this question is strongly related to the question whether the norm-setting organisation qualifies as an international organisation. An international organisation (as opposed to an NGO) can only be founded by public actors, in particular states and international organisations; see Schermers/Blokker, Institutional Law, §§ 36 f.

<sup>367</sup> The recommendations or Accords of the BCBS are an example for this problem. The BCBS is composed of representatives of national banking supervisory author‐ ities and central banks. Irrespective of its express lack of 'formal supranational authority' (3. of the BCBS-Charter), self-confidently it describes itself as 'primary global standard setter for the prudential regulation of banks' (1. of the BCBS-Char‐ ter). That the Committee's decisions 'do not have legal force' (3. of the BCBS-Char‐ ter) nearly goes without saying, but their qualification as soft law is questionable. The Committee is situated with the Bank for International Settlements, but without an explicit foundation in public international law. Also, it is not clear whether the members of the BCSB are (self-)bound by the recommendations; see C Möllers, Behördenkooperation 368 f; for the 'European participation' in the BCBS see From‐ age, Articulation; in favour of a qualification of the BCBS's recommendations as soft law: Ho, Compliance; Knauff, Regelungsverbund 289 f and 376; Meyer, Soft Law 888 f; for the legal and factual nature of the Basel Accords see also Arndt, Sinn 83–86. *Köndgen* qualifies them as 'halbstaatliches Expertenrecht' [semi-public expert law]; Köndgen, Privatisierung 493.

law. This addresses the second level which is to be taken into account, ie the way standards are dealt with once they are set. Where they are published by a public body<sup>368</sup> or referred to in a legal provision as a threshold to be met, they are, at least as regards their content,<sup>369</sup> taken over by the (soft) law-maker.<sup>370</sup> Such references can be drafted differently, which is most prominently expressed in the distinction between static and dynamic refer‐ ences.<sup>371</sup> Whereas static references refer to a specific version of a standard or set of standards, dynamic references also allow for a legal incorporation of a future (version of this) standard or set of standards, without the law (ie the reference) having to be adapted.<sup>372</sup> Such a reference can be contained

<sup>368</sup> See eg case C-613/14 *Elliott*, para 43, dealing with private standards published by the Commission; for the related publishing practice of the EU see also Colombo/Elian‐ tonio, Standards 328.

<sup>369</sup> Formally speaking, a non-legal act does not thereby become law, however; see Jabloner, Rechtsetzung 8 f; Korinek, Verbindlichkeit 322.

<sup>370</sup> See Erhard, Probleme 6; see also Peters/Pagotto, Perspective 19 f. See, for example, Article 3 para 2 of Directive 2001/95/EC, according to which '[a] product shall be presumed safe as far as the risks and risk categories covered by relevant national standards are concerned when it conforms to voluntary national standards trans‐ posing European standards, the references of which have been published by the Commission in the [OJ]'. According to para 3 *leg cit*, these standards are on an equal footing with Commission recommendations setting guidelines on product safety as‐ sessment (EU soft law); for the democratic and rule of law concerns the dominance of technical norms in substance set by private actors entails see Holoubek/Potacs, Technikrecht 60 f and *passim*; for examples of soft law containing references to soft law see Ştefan, Soft Law 105.

<sup>371</sup> For the rules of doubt on references contained in acts adopted in the course of an ordinary legislative procedure see EP/Council/Commission, *Joint Handbook for the presentation and drafting of acts subject to the ordinary legislative procedure* (2022) 88 f. It is established practice in the EU to refer to 'essential requirements' related to health, safety and environmental issues in so-called New Approach Directives. Only where these requirements are met may a product be marketed in the internal market. The close determination of the 'essential requirements' is left to be done for the standardisation bodies. Companies may prove that in case of their product the requirements are met even though it does not comply with the pertinent standard, but it is regularly very difficult to provide this evidence; see Colombo/Eliantonio, Standards 334 f, with a further reference.

<sup>372</sup> For the legal technique of references and the legal difficulties it entails see Erhard, Probleme 6–13; Röthel, Normen 46; for different referencing techniques see also Holoubek/Potacs, Technikrecht 66–70; for dynamic references to soft law see Wal‐ ter, Soft Law 27 and, with regard to public international law, Goldmann, Gewalt 56–59; for references to commercial customs see Arndt, Sinn 46; for the case of EU law see case C-613/14 *Elliott*, para 38; for referencing as a way of incorporating provisions of public international law into national law see Shelton, Compliance 131;

in a statute or administrative regulation, but as well in a private contract. Depending on the wording of the reference, the content of the standard or set of standards will then become law or – only where the referring norm constitutes an act of public authority<sup>373</sup> – soft law, at least within the scope of the referring provision.

Especially in the case of dynamic references we may speak of a hidden delegation of (soft or hard, as the case may be) rule-making power to the standard-setters. Whether or not such a delegation is lawful depends on the respective rules on delegation to be applied. As mentioned above, the named European standardisation bodies are strongly involved in the process of rule-making due to the fact that the EU decision-makers (legis‐ lator, Commission), in particular since the Commission has proclaimed the 'New Approach' in its 1985 White Paper 'Completing the Internal Market',374 set out in advance a rule-making work programme including a rough description of the desired content.375 This approach appears to be mitigating the above concerns rather than underpinning them. After all, it involves in particular the Commission – even if only superficially – in shaping the standards and thereby limits the room for manoeuvre of the standardisation bodies. It does not do away with the principal challenge to democratic rule-making a (potential) delegation of rule-making powers to private bodies poses at all levels of law-making, though.<sup>376</sup>

Legal referencing is not the only way to increase the authority of stand‐ ards. Standards also become obligatory where penalties are imposed by law in case of non-compliance with them. Where mere incentives for com‐ pliance with certain standards are created (eg non-fiscal incentives such as technical assistance or fiscal incentives such as subsidies377), room for

for the duty of due consideration as another referencing technique see Müller-Graff, Einführung 152.

<sup>373</sup> For the impossibility of the adoption of soft law by private actors (without public authority) see 1.3.3.2. above.

<sup>374</sup> Commission, 'Completing the Internal Market' (White Paper), COM(85) 310 final; for the importance of such standardisation as part of a 'New Approach' in internal market harmonisation see Craig/de Búrca, EU Law 658-661; see also Griller, Nor‐ mung 18 ff.

<sup>375</sup> See Griller, Normung 23–25.

<sup>376</sup> Some of these issues are addressed in case T‑229/17 *Germany v Commission*, *passim*; for the specific case of harmonised standards and its effects see Volpato, Effects 200-209.

<sup>377</sup> See examples given by Carey/Guttenstein, Governmental Use 22; for a broad con‐ cept of sanctions including incentives as 'positive sanctions': Bittner, Sanktion 31–33.

lawful non-compliance remains, which speaks in favour of their (implic‐ it) elevation to the rank of soft law. From the perspective of the official deciding upon whether or not technical assistance or a subsidy is to be granted, the standard is legally binding to the extent that a positive decision may only be granted where the applicant complies with the standard. The case is less clear where the legislator generally refers to a 'state of the art' standard.<sup>378</sup>

Also other cases of private regulation, eg corporate governance codes, which shall not be dealt with here in detail, may come close to or actually reach the status of soft law due to general recognition by bodies vested with public authority.<sup>379</sup>

### 2.3. From other output of public bodies

Not all output of public bodies can be assigned to a certain set of norms,<sup>380</sup> eg certain letters,381 policy papers, agendas, reports, studies, administrative correspondence or statistical data.382 The purpose of such acts generally is not or only indirectly to steer human behaviour,383 and certainly not to set

<sup>378</sup> With regard to the different degrees of profoundness such reference clauses may require with regard to establishing the 'state of the art' see eg the tiered scheme (three levels) established by the German *Bundesverfassungsgericht* in case BVerfGE 49, 89.

<sup>379</sup> See Knauff, Regelungsverbund 244–247.

<sup>380</sup> For prominent examples of 'informale Kommunikation' [informal communication] of public bodies (and the latter's representatives) in Germany see Croon, Arenen, in particular 50–54.

<sup>381</sup> See case 182/80 *Gauff*, para 18; joined cases 42 and 49/59 *S.N.U.P.A.T.*, 72; for a letter including – on the contrary – an implicit decision see ibid, pages 73 f; case T-116/89 *Prodifarma*, para 84; case C‑701/19P *Pilatus Bank*, paras 31 ff, with regard to an e-mail; for a letter of a Commissioner to the competent ministers in the MS whose content comes close to soft law see <https://agriculture.ec.europa.eu/news/le tter-eu-agriculture-ministers-commissioner-wojciechowski-rural-development-and -covid-19-outbreak-2020-04-08\_en> accessed 28 March 2023.

<sup>382</sup> For the phenomenon of such informal administrative action more generally see, with regard to German law, Schmidt-Aßmann, Verwaltungsrecht 348–350; with regard to soft law and output that does not even qualify as soft law in the context of the ACER see Godin/Polet/Jamar de Bolsée, Analysis 201f.

<sup>383</sup> See in particular Goldmann, Gewalt, providing for an in-depth analysis of public in‐ formation, its steering effects and its relation to normative output of public authori‐ ties; see also von Bogdandy/Goldmann, Ausübung 69; von Graevenitz, Mitteilungen 169 f; for the example of environmental labels as a piece of authoritative information

(soft) rules,<sup>384</sup> but it is to inform about problems and political plans to solve them, certain developments in practice, scientific evidence, to exchange points of view, or to inform about future rules, etc.<sup>385</sup> In case of doubt, the wording of the output at issue and also its usual handling are to be taken into account.<sup>386</sup> The title ('report', 'communication') is only indicative.<sup>387</sup> Also assumedly non-normative output of public bodies may, exceptionally, contain soft law rules or even legal rules.<sup>388</sup>

Legislative or other decision-making proposals are not in any way bind‐ ing upon the future addressees of what maybe will become law, but they


<sup>(</sup>and only indirectly an incentive to adjust consumer decisions, hence only indirectly aimed at steering human behaviour) see Feik, Verwaltungskommunikation 392 ff.

<sup>384</sup> In a non-normative understanding, soft law may have in common with other ad‐ ministrative output its purpose to facilitate communication (to function as threads, that means); see Boehme-Neßler, Unscharfes Recht 535 ff, who describes the law as networks composed of 'Knoten' [knots], that is persons/institutions, legal terms (eg '*culpa in contrahendo*') and certain norms (eg the general part of a civil law code which applies also to its special parts and thereby exerts a connective function), 'Super-Knoten' [super-knots], that is cross-border institutions (such as the EU or international organisations), collision norms and dogmatic constructs (such as the third party effect of fundamental rights) and 'Fäden' [threads], that is communica‐ tion.

<sup>385</sup> Future rules, in principle (for the *vacatio legis* see 2.1.3.1 above), are not soft law, because until they enter into force they do not suggest a certain behaviour, ie they do not command. Sometimes future rules may have effects, though, which transform them into (soft) law; with regard to EU law see eg cases C-129/96 *Inter-Environnement Wallonie*, para 45, and C-144/04 *Mangold*, para 28, according to which already in the course of the implementation period the adoption of measures seriously compromising the (Directive's) result prescribed is prohibited; for the *Mangold* case see also Roth, Mitgliedstaaten 138–140; for the prospective view EU soft law may take see Hofmann/Rowe/Türk, Administrative Law 543; with regard to public international law see Schermers/Blokker, Institutional Law, § 1276.

have an effect on their respective addressee, the legislator/decision-maker. Procedurally, a legislative or other decision-making proposal (mostly ema‐ nating from the executive branch) in many legal orders is binding upon the legislator/decision-maker to the extent that it determines the subject of the act (to be adopted). The legislator/decision-maker can still decide not to adopt an act at all, but if it intends to adopt an act in the course of the procedure initiated by the executive's launching of a relevant proposal it has to stick to the subject. Since the subject of a proposal is regularly malleable, and since the legislator/decision-maker is free to fully change the rules proposed (as long as it sticks to the subject), its leeway is considerable. An executive proposal for adoption normally qualifies as soft law,<sup>389</sup> it is recommended for adoption, without binding the legislator/decision-mak‐ er. Where the body making the proposal disposes of special information and/or expertise and where the final decision-maker lacks these qualities the latter's room for manoeuvre *de facto* is limited more strongly.390 The peculiarities of a proposal as compared to other soft law rules are to be acknowledged: It is legally binding to a very limited extent, and also other‐ wise it disposes of a steering effect (the legislator/decision-maker is asked to adopt the proposal as a legal act), but at the same time deviation by its addressee (amendments to the proposed body of rules in the course of eg legislative negotiations) in practice are highly expected. The life-time of such a proposal is regularly shorter than that of other soft law rules.

<sup>389</sup> For the qualification of a Commission proposal as soft law see III.2.4. below.

<sup>390</sup> See, for the international level, Schmalenbach/Schreuer, Organisationen, para 1050.

# III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW

#### *1. Introduction and overview*

#### 1.1. Introduction

The purpose of Part III of this work is to address selected general issues of EU soft law in order to set the basis for the more specific questions dealt with in Parts IV and V. It is not intended to provide an all-encompassing account of EU soft law here, but above all to discuss matters which are relevant also for the following parts of this work. This is why, for example, no comprehensive taxonomy of EU soft law acts is provided for,391 but only an overview in terms of the potential originators and the potential addressees of EU soft law, thereby referring also to legally non-binding EU acts below the level of soft law (2.). Subsequently, the focus is shifted to the competences to adopt EU soft law (3.) and to the effects of EU soft law (4.). These aspects shall be complemented by chapters on the purposes of (EU) soft law, essentially reflecting upon the reasons for its adoption, and for conferring (EU) soft law powers in the first place (5.), and the judicial review of EU soft law (6.).

Chapter 3 is the most expansive chapter of Part III and already at this stage requires some further remarks on the approach which shall be taken in it. It shall address the meaning of Article 288 TFEU for EU soft law and shall address the question whether the principle of conferral – the primary paradigm when it comes to the EU's competence order – is applicable also in the context of EU soft law. The answer to this question is far from obvious. Having addressed the relevant case law of the CJEU, we shall also take into account the explicit legal bases for the adoption of EU soft law – in particular: recommendations and opinions as those legally non-binding EU acts mentioned in Article 288 TFEU. This exercise serves a number of objectives. First, on a general scale, it is intended to show that the mani‐ fold use of soft law is not only a consequence of everyday administrative practice, but is actually – to some extent at least – explicitly mapped out in the Treaties. Second, it shall allow us to distinguish, in the given context,

<sup>391</sup> For different approaches in the literature to build such a taxonomy see Ştefan/Av‐ belj/Eliantonio/Hartlapp/Korkea-aho/Rubio, Soft Law 17–20.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

different categories of competence clauses. Third, and more specifically, the explicit competences in the Treaties to adopt soft law may be telling with regard to the question of whether the principle of conferral also applies in the context of soft law. *Prima facie*, the multitude of such competences in the Treaties suggests that it does. An in-depth analysis, as we shall see, will lead to more nuanced results. Fourth and fifth, an account of the explicit Treaty competences to adopt recommendations and opinions may allow for insights as to the substantial difference between these two acts and as to the question whether the Treaties provide for a *numerus clausus* of soft law acts.

# 1.2. Overview of the historical and current use of EU soft law

Community law and, since the entry into force of the Treaty of Lisbon: exclusively Union law have/has developed into a highly integrated legal order, in the view of some even into 'the most advanced form of regional integration in the world'.392 It builds a stark contrast to inhomogeneous, decentralised public international law in which – through its founding Treaties – it roots.393 This holds true notwithstanding the incorporation of parts of public international law in the EU legal order.394 Unsurprisingly, also at the level of *soft* law the Community/Union legal order on the one hand, and public international law on the other hand, exert a 'different dynamic'.395 This can be exemplified by comparing the politically often very loaded soft law acts of public international law, eg the UN Universal Declaration of Human Rights,396 with guidelines adopted by the European Securities and Markets Authority (ESMA), a common example of EU soft

<sup>392</sup> Terpan, Soft Law (2015); for the special character of law as object and agent of integration see Dehousse/Weiler, Dimension 234.

<sup>393</sup> See Bianchi, Butterfly 209 f; see also Kelsen, Law 93, emphasising that completely decentralised law is 'primitive law'.

<sup>394</sup> See Article 216 para 2 TFEU and the CJEU's case law, starting with cases 21–24/72 *International Fruit Company*; see also Craig/de Búrca, EU Law 392 f.

<sup>395</sup> Klabbers, Courts 221; for the different framework of EU soft law and public interna‐ tional soft law see also Ferran/Alexander, Soft Law Bodies 759; refusing a transferal of the international law concept of 'soft law' to EU law: Hummer, Interorganverein‐ barungen 97. While *Terpan* emphasises that EU soft law is not 'intrinsically different from soft law in the international realm', he does not refuse to acknowledge existing deviations: Terpan, Soft Law (2013) 4.

<sup>396</sup> For similar examples see Cannizzaro/Rebasti, Soft law 214.

law.397 The former has a broad scope, its provisions are (and have to be) relatively short and hence open for different interpretations. The latter, on the contrary, normally are very specific, detailed and complex.<sup>398</sup> This exemplary comparison shall not suggest that more general EU soft law does not exist,<sup>399</sup> or that public international soft law is always short and fundamental,<sup>400</sup> but it ought to illustrate that due to Community/Union law's higher degree of integration also its soft law instruments tend to be more strongly integrated in everyday administration. More generally, it can be stated that they encompass a larger scale, ie they are more versatile as regards content, form, and purpose.<sup>401</sup>

Soft law – or the 'power to exhort and persuade', as the Court has recently phrased it<sup>402</sup> – has formed part of the ECs', and the EU's respec‐ tively, policy-making tools<sup>403</sup> ever since its foundation.<sup>404</sup> This is reflected upon in the EEC's founding treaty which in its Article 155 provides that the Commission shall 'formulate recommendations or deliver opinions on matters dealt with in this Treaty, if it expressly so provides or if the Com‐

401 Similarly: Terpan, Soft Law (2013) 40.


<sup>397</sup> Eg ESMA, Guidelines on stress test scenarios under the MMF Regulation, ES‐ MA34–49–495 (27 January 2023); see more generally van Rijsbergen, Legitimacy.

<sup>398</sup> While a high degree of detail may be an indicator of legal bindingness, an 'autom‐ atism' in that respect is inappropriate. Also soft law may contain detailed rules; see case T-721/14 *Belgium v Commission*, para 72; Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 128 f; stressing the increasing complexity of EU soft law in general: Korkea-aho, Courts 471. For the general (early) criticism of over-regulation by EU law see – in the context of the common agricultural policy – Opinion of AG *Verloren van Themaat* in case 292/81 *Jean Lion et Cie*, 3913, complaining about '[t]he flood of rules and regulations which [were referred to], quite rightly, as a "labyrinth"'.

<sup>399</sup> See eg the CFR which for the time between 2000 and 2009 is to be qualified as a soft law act or, earlier, the Joint Declaration of the EP, the Council and the Commission of 27th April 1977, C103/1, on fundamental rights; see Österdahl, Soft Law 37; for the effectiveness of this kind of soft law explained with a view to the buzzwords 'visibility' and 'pedagogy' see Sarmiento, Soft Law 280 f.

<sup>400</sup> See eg the non-binding procedures according to Articles 279–285 ('settlement of disputes') and Annex V of the UN Convention on the Law of the Sea (for the non-bindingness of the respective output see Article 7 para 2 of Annex V); see also the example given by Wirth, Assistance 225.

<sup>402</sup> Case C-16/16P *Belgium v Commission*, para 26; case C‑501/18 *Balgarska Narodna Banka*, para 79; case C-911/19 *FBF*, para 48.

mission considers it necessary'.<sup>405</sup> The wording of this provision allows for a wide-spread use of Commission soft law. Similarly, Article 189 TEEC (Rome) and Article 161 of the Treaty establishing the European Atomic Energy Community (TEAC, Rome) stipulate that both the Council and the Commission shall, 'in accordance with the provisions of this Treaty, […] make recommendations or deliver opinions'. While in the original version of the Treaty establishing the European Coal and Steel Community (TECSC) recommendations are – somewhat misleadingly – defined as: 'binding with respect to the objectives which they specify but [they] shall leave to those to whom they are directed the choice of appropriate means for attaining these objectives',<sup>406</sup> the Treaties of Rome apply a linguistically more orthodox (but still rather loose) definition,<sup>407</sup> according to which recommendations (and opinions) 'shall have no binding force'.<sup>408</sup>

In practice, already early in the history of the ECs their institutions, including the European Parliament and the European Council as 'institu‐ tionalised' in 1974,409 have made use of a much wider set of soft law instru‐ ments than that explicitly provided for in the Treaties.410 With regard to the Commission, its soft rule-making in the field of State aid policy may serve the purpose of illustration.411 Starting in the early 70s, there has been a 'gradual increase' of State aid-related soft law acts such as guidelines, frame‐

<sup>405</sup> Similarly: Article 124, 2 nd indent TEAC (Rome); Article 14 para 1 TECSC (Paris). Already at an early stage, Community soft law had been dealt with by the Court: see eg joined cases 1 and 14/57 *Usines à tubes*, in particular 114 f; referring to the so-called 'Christmas Communications' of December 1962 as 'probably first case and still primary example of administrative rules': Peters, Typology 414; similarly: Ştefan, Soft Law 67: 'first soft law instruments ever issued'; with regard to the example of the Commission Communication concerning the *Cassis* judgement see H Hofmann, Normenhierarchien 217.

<sup>406</sup> Article 14 para 3 TECSC (Paris). Note the similarity to the directive as defined now in Article 288 TFEU; emphasising the equivalence of recommendations under the ESCS and directives under the E(E)C: Grunwald, Energierecht 103 (in particular fn 15).

<sup>407</sup> See Schermers/Blokker, Institutional Law, § 1217.

<sup>408</sup> Article 189 TEEC (Rome); Article 161 TEAC (Rome); see Bothe, Soft Law 761.

<sup>409</sup> For this institutionalisation see Nicolaysen, Gemeinschaftsrecht 28; Everling, Wir‐ kung 139. The European Council's institutionalisation *stricto sensu* – that is its elevation to an institution of the EU – was brought about only by the Treaty of Lisbon in 2009.

<sup>410</sup> See Senden, Soft Law 4. For an early categorisation of the different acts and their creators see Bothe, Soft Law 762 ff.

<sup>411</sup> See Aldestam, Soft Law 14–16.

works, communications or codes adopted by the Commission.<sup>412</sup> While the State aid provisions of primary law have provided and still provide for the possibility of Council regulations fleshing out the Treaty provisions,<sup>413</sup> after two unsuccessful Commission proposals in 1966 and 1972 respectively, the legal possibility of a Council regulation practically had for a long time become irrelevant.<sup>414</sup> The State aid regime within which the Commission exercises considerable (and 'hard') powers also makes its soft law output highly authoritative. Accordingly, non-compliance by the MS with the re‐ spective soft law has led to re-evaluations and negative decisions.<sup>415</sup> More generally, in its seminal White Paper of 1985 – designated by *Pelkmans* as an 'exercise in deregulation'<sup>416</sup> – the Commission announced to make increasing use of communications.<sup>417</sup>

The Council has continuously rendered soft law acts in order to utter its opinion and bring in its ideas, often in reaction to a policy initiative set by the Commission. These acts have been named resolutions,418 conclusions, declarations, etc.419 The European Parliament has expressed its opinion in particular where it has been consulted in a decision-making procedure,


<sup>412</sup> See Cini, Soft law approach 198; postulating an increase of Community soft law more generally after 1968: Ştefan, Soft Law 12.

<sup>413</sup> Now: Article 109 TFEU.

<sup>414</sup> See Cini, Soft law approach 199. In the late 90s, eventually two Council Regulations on exemptions from the State aid rules and on procedural issues were adopted (meanwhile replaced by new versions). That the Commission's soft output does not infringe the Council's legislative competences in this context was confirmed by the CJEU *inter alia* in case C-526/14 *Kotnik*, para 59.

<sup>417</sup> See Commission, 'Completing the Internal Market' (White Paper), COM(85) 310 final, para 155.

<sup>418</sup> For the Council's adoption of recommendations which had been gradually superse‐ ded by resolutions see Everling, Wirkung 138. On the important role of resolutions in EU law more generally see von Bogdandy/Arndt/Bast, Instruments 115–117.

<sup>419</sup> Eg Council Resolution on a new approach to technical harmonization and stand‐ ards, C 136/1. While the adoption of these acts are not (explicitly) provided for in the Treaties, the Treaty concerning the Accession of the Republic of Croatia to the European Union (2012), for example, acknowledges their existence and their effects (Article 3 para 3). These designations are not very distinctive: see case 32/79 *Com‐ mission v United Kingdom*, paras 11f, in which the Court uses the terms 'resolution' and 'declaration' as synonyms; with regard to the similarity of 'declarations' and 'conclusions' see Senden, Soft Law 198. On the lack of distinction between different soft law acts (recommendations, declarations, conventions) in public international law see Schermers/Blokker, Institutional Law, § 1216.

either obligatorily according to the Treaties or on a facultative basis.420 But also emanations on its own motion, above all in the field of foreign policy, have had a long history.<sup>421</sup> The latter, however, often do not display any normative content and hence do not qualify as soft law. As regards the European Council, many of the (both legal and soft law) acts – namely those which did not have an explicit Treaty base – have belonged to the realm of public international law rather than that of Community/Union law.<sup>422</sup> This is the case in particular in foreign and security policy which has had a strong intergovernmental character, and – to a lesser degree – still has this character.<sup>423</sup> But also in other policy fields the European Council has expressed itself by means of soft law and other legally non-binding acts, 'in different kinds of *Communiqués*, under different names, like press releases, declarations, conclusions and resolutions',<sup>424</sup> some of which have had a pivotal influence on the political development of the ECs/EU.<sup>425</sup>

While a variety of soft law instruments by name unknown to the Treaties has been used by the mentioned institutions, when it comes to general-ab‐ stract rule-making up until the late 80s the so-called Community method appears to be the predominant approach.426 The Community method des‐ ignates integration by means of supranational – that is by definition: hard – law (Community/Union law), with a supranational executive body inde‐ pendent of the MS pursuing the Community/Union interests (the Commis‐ sion), with the possibility to overrule a minority of MS in the legislative process (qualified majority voting in the Council), and with the possibility of judicial enforcement (before the CJEU) at hand.427 Starting in the 90s,

<sup>420</sup> See Constantinesco, Recht 456–460.

<sup>421</sup> See eg Kreppel/Webb, Resolutions; with regard to the Council conclusions adopted in the context of the CFSP see Council, Comments on the Council's Rules of Procedure (2022) 100 f.

<sup>422</sup> See Constantinesco, Recht 545–547; Nicolaysen, Gemeinschaftsrecht 44; Well‐ ens/Borchardt, Soft Law 296 f, with further references and 299; see, however, the inclusive approach in favour of Community law of the CJEU as expressed eg in case 38/69 *Commission v Italy*.

<sup>423</sup> See Wellens/Borchardt, Soft Law 298 f; note, however, that a number of acts adop‐ ted under CFSP actually display legal bindingness, and only lack enforceability; see Terpan, Soft Law (2015) 80.

<sup>424</sup> Wellens/Borchardt, Soft Law 298.

<sup>425</sup> Eg the Conclusions of the Presidency of the European Council (5 December 1978), *inter alia* referring to the introduction of the European Monetary System (EMS).

<sup>426</sup> See Terpan, Soft Law (2015) 87.

<sup>427</sup> See Commission, European Governance – A White Paper, COM(2001) 428 final, 6; see also Costa, European Parliament 60; for the role of the Court heed its elemental

this has changed and the still dominant Community method in some instances has given way to softer forms of governance – regularly at the cost of Parliament participation and Court control.<sup>428</sup> This transformation can be perceived as the result of a growing discontent with EC legislation since the early 90s.<sup>429</sup> In the late 90s, the number of soft law has seen another boost, in particular in the area of competition and State aid law.<sup>430</sup> While soft law instruments are still firmly situated within the toolkit of EU governance, also with regard to (hard) law as the traditional method of EU regulation some mechanisms have been introduced in order to improve (in different ways) the legislative and other law-making processes.<sup>431</sup> These small reforms reflect the influence emerging new modes of governance have had (also) on traditional regulation.<sup>432</sup>

In 2001, the Commission in one of its White Papers proclaimed a new era of European governance.433 In this programmatic document – which is to be understood as a response to a number of instances of governance failure in the EU434 – the Commission, *inter alia*, called for '*combining* formal rules with other non-binding tools such as recommendations, guide‐ lines, or even self-regulation within a *commonly agreed framework*. This highlights the need for close coherence between the use of different policy instruments and for *more thought to be given to their selection*' (emphases added).435 The Open Method of Coordination (OMC) which has been applied in particular in the field of socio-economic policies since the Spring

434 See Dawson, Waves 210.

expression in joined cases 90 and 91/63 *Commission v Luxembourg*, 631: 'Member States shall not take the law into their own hands'.

<sup>428</sup> See Terpan, Soft Law (2015) 88. NB that *Terpan* applies a wider understanding of soft law, also encompassing hard obligations with only soft or no means of enforce‐ ment; see ibid 74–76; Beckers, Juridification 575 f; Dawson, New Governance 4–6, with further references; Scott/Trubek, Gap 4 f; for the strong increase in legislative activity immediately before this period, starting in the mid-80s under the *Delors* Commission see Stone Sweet, Integration 204–206.

<sup>429</sup> See Senden, Soft Law 11f; see eg Conclusions of the Presidency of the Edinburgh European Council, 11–12 December 1992, SN 456/1/92 REV 1, 33, calling for 'clear‐ er and simpler' Community legislation; Commission, 'Simpler Legislation for the Internal Market (SLIM): A Pilot Project' (Communication), COM(96) 204 final.

<sup>430</sup> See Petit/Rato, Enforcement 202.

<sup>431</sup> See Craig, Administrative Law 220.

<sup>432</sup> See Dawson, Waves 213–216.

<sup>433</sup> European Commission, European Governance – A White Paper, COM(2001) 428.

<sup>435</sup> European Commission, European Governance – A White Paper, COM(2001) 428, 19.

European Council Summit in Lisbon in 2000, and which 'encourages ac‐ tors to make commitment to obligations',<sup>436</sup> is one prominent example of such a soft governance approach.<sup>437</sup> According to the calculations of *von Bogdandy*, *Arndt* and *Bast*, in 2004 the most frequent forms of EU soft law – recommendations, opinions, and resolutions – amounted to 10 per cent of all Community/Union law (in a broader sense) which was in force by then.<sup>438</sup> Since then this share has significantly increased.<sup>439</sup>

The developments bolstered by the White Paper on European Gover‐ nance were not embraced by all political actors. In a 2007 Resolution, for example, the Parliament criticised that 'soft law also tends to create a public perception of a "super bureaucracy" without democratic legitimacy, not just remote from citizens but actually hostile to them, and willing to reach accommodations with powerful lobbies in which the negotiations are nei‐ ther transparent nor comprehensible to citizens, and [that] this may raise legitimate expectations on the part of third parties affected (eg consumers), who then have no way of defending them at law in the face of acts having adverse legal effects for them'.<sup>440</sup>

<sup>436</sup> Dawson, Soft Law 7.

<sup>437</sup> The output of OMC often does not meet the level of concreteness required for soft law according to the definition applied here (see II.1.3.4. above). This may be one of the reasons why it is rather referred to as 'soft governance' by Trubek/Trubek 343 (fn 2); see also Láncos, Facets 38, with further references. For the historical development of the OMC see eg Craig, Administrative Law 199–202; for the Com‐ mission's warning that OMC may 'dilute the achievement of common objectives in the Treaty or the political responsibility of the Institutions' – arguments which the use of soft law more generally may be confronted with – see Commission, European Governance – A White Paper, COM(2001) 428, 21; for a legal qualification of OMC see Dawson, New Governance 51–66; Terpan, Soft Law (2015) 81–84; with regard to the MS as the (potential) creators of soft law in the course of the OMC see Lafarge, Coopération, in particular 78 f; Müller-Graff, Soft Law 26.

<sup>438</sup> See von Bogdandy/Arndt/Bast, Instruments 97.

<sup>439</sup> See the study of Cappellina/Ausfelder/Eick/Mespoulet/Hartlapp/Saurugger/Terpan, Soft law, in particular 7 f, which may not apply the same definition/methods as von Bogdandy/Arndt/Bast, but which nevertheless shows – for the time period 2004-2019 – a clear tendency of an increase of the soft law share in the overall number of EU rule-making acts.

<sup>440</sup> European Parliament, Resolution on institutional and legal implications of the use of 'soft law' instruments, 2007/2028(INI), para Y. It is not by chance that it is the European Parliament arguing against the merits of soft law here. After all, it is its very legislative competence (in particular in the ordinary legislative procedure) which is challenged by alternative modes of regulation; see Knauff, Regelungsver‐ bund 299; with a view to the OMC: Borrás/Jacobsson, Method 200.

In spite of this criticism, perceived holistically, soft law by now appears to be a well-established regulatory instrument in EU governance. While the two non-binding 'legal acts'<sup>441</sup> mentioned in the catalogue of the Treaties (now Article 288 TFEU) are recommendations and opinions, in practice many other designations have been in use. Due to the trend of institutional decentralisation which has strongly transformed the institutional morphol‐ ogy of I E(E)C/EU, in particular since the early 90s, the number of bodies adopting soft law has increased considerably.<sup>442</sup> Apart from the institutions, there is now – arguably *in principle* in accordance with primary law – a large number of other entities adopting EU soft law, eg the European Ombudsman,<sup>443</sup> the European Anti-Fraud Office OLAF, or the variety of European agencies.<sup>444</sup> The EU's varied reaction to the COVID-19 pandemic has again aptly displayed the importance of soft law as a tool allowing for swift regulatory action as well as its partial lack of legitimacy.<sup>445</sup>

<sup>441</sup> See also case C-424/07 *Commission v Germany*, para 75, in which the Court refers to two soft law acts as 'legal instruments'; see also <https://eur-lex.europa.eu/le gal-content/EN/TXT/HTML/?uri=LEGISSUM:l14534&from=DE> accessed 28 March 2023, listing communications, recommendations, white and green papers as 'sources of European Union law'. The terminology the TFEU applies ('legal acts') suggests an incorporation of EU soft law in EU law; see von Bogdandy/Arndt/Bast, Instruments 111; Wörner, Verhaltenssteuerungsformen 216. While this may hold true in a perspective limited to Article 288 TFEU and related provisions, it cannot do away with the fact that legal bindingness is one of the *conditiones sine quis non* of law – also in EU law; see Hofmann/Rowe/Türk, Administrative Law 544, with regard to the 'provocation' to include legally non-binding acts in the category of 'legal acts' (The term 'provocation' in this context has already been used by *von Bogdandy*, *Arndt* and *Bast*, *opus citatum* 111).

<sup>442</sup> See Everson/Joerges, Europeanisation 524–520; with regard to the increase of soft law-making bodies at the international level: d'Aspremont, Pluralization 185.

<sup>443</sup> The European Ombudsman is established by primary law anyway: Article 228 TFEU.

<sup>444</sup> See Görisch, Verwaltung 204–207; Raschauer, Verhaltenssteuerungen 686–688; Ra‐ schauer, Leitlinien.

<sup>445</sup> See Andone/Coman-Kund, EU soft law 1; Eliantonio/Ştefan, Legitimacy; Ştefan, COVID-19 Soft Law 1; Weiß, Pandemic.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

### *2. Different forms of EU soft law: originators and addressees*

### 2.1. Introduction

As we have seen, EU soft law is of ever increasing importance. In the words of the French *Conseil d'État*, it has developed into a 'véritable méthode de gouvernance' of the EU.<sup>446</sup> In the following, an introductory account of EU soft law shall be given, thereby taking the actors concerned as a parameter, that is to say, with a view to its originators and its addressees. As was indicated above, EU soft law is not only adopted by the institutions but also by its bodies, offices and agencies and even – organisationally speaking – non-EU bodies. Conversely, the addressees of EU soft law acts are not only the MS and the citizens/undertakings of the EU, but may be EU institutions or bodies themselves, or even third countries. Through this approach, the multifaceted, 'far from homogeneous'447 nature of EU soft law will become apparent – not only in terms of the actors concerned (of which different groups shall be built), but, due to the examples provided and merely as a collateral effect, also in terms of the different shape EU soft law may take and, in connection therewith, of the different purposes EU soft law may serve.

Eventually, the large body of EU output which neither qualifies as law nor as soft law shall be contoured. This is to acknowledge that besides the conceptual antipodes law and soft law there exists a third category of acts which is of eminent importance for everyday administration as well.<sup>448</sup>

### 2.2. Originators

### 2.2.1. On the question of assignment

The number of creators of EU soft law is quite high, certainly higher than the number of creators of EU law. Most illustrative of this multiplicity of actors involved in the creation of EU soft law is the existence of about 40 European agencies, only some of which have the power to adopt legally

<sup>446</sup> Conseil d'État, Droit souple 28.

<sup>447</sup> Senden, Soft Law 23.

<sup>448</sup> See also II.2.3. above.

binding acts, but the majority of which may adopt soft law.<sup>449</sup> In an attempt to do away with this institutional opaqueness of the creation of EU soft law, its originators shall be divided in three groups: 1) the EU's institutions, 2) its bodies, offices and agencies, and 3) MS and non-EU bodies. With standard acts of EU secondary law, eg a Regulation adopted by the EP and the Council or a Commission decision, their assignment to creator(s) – namely the respective institution(s) (eg in the context of an action for annulment) or the EU (eg in the context of its non-contractual liability<sup>450</sup>) – is a comparatively easy task. In places, this task can be more challenging, for example where a body is empowered to take action *on behalf* of another body to which it does not belong. It is then – legally speaking – the represented body to which the action is normally to be assigned.

In accordance with this institutional assignment, the act at issue regularly can also be allocated to a certain body of law. For example, a Regulation of the EP and the Council constitutes EU law. Also in this context there may be cases where an allocation is more challenging than that. This is the case where bodies from different States or organisations are involved in a certain action; for example, where three States conclude an agreement on the establishment of a power plant in the border region of these three States. States normally cooperate with each other on the basis and by means of public international law. However, a closer look at the agreement may reveal that it shall be subject to the national law of one of the States. It then qualifies as a contract under that respective national law, not under public international law.451 Sometimes the different actors and legal orders involved become so much intertwined that it would be difficult to assign the act at issue to either side. What is more, in such cases an assignation often is of limited value, as, failing to adequately graps the underlying complexity, it hardly provides for orientation. This is because actors from other levels may still be involved on a subordinate scale, and they may be bound by their respective law, etc.

A good example is the European Financial Stability Facility (EFSF).<sup>452</sup> It is a legal person under Luxembourgish law, established by the Euro-MS.

<sup>449</sup> See Chiti, Agencies, in particular 97–100 and 102–106.

<sup>450</sup> For the bodies, offices and agencies see eg Ruffert, Art. 340 AEUV, para 8. For the special rule applying to the ECB – which qualifies as institution of the EU and has legal personality – see Article 340 para 3 TFEU.

<sup>451</sup> See examples provided by Schermers/Blokker, Institutional Law, § 45.

<sup>452</sup> For the following information on the EFSF see Megliani, Sovereign Debt 585 ff. The EFSF which serves as an example here to illustrate the problems underlying the

The EFSF's tasks are subject to a framework agreement concluded between the EFSF and the Euro-MS, which is subject to English law. Disputes between Euro-MS under this agreement are subject to the jurisdiction of the CJEU, disputes between the EFSF and one or more Euro-MS shall be subject to the jurisdiction of the national courts of Luxembourg. The EFSF concludes agreements on loans and other financial instruments to the benefit of one Euro-MS. These agreements are again concluded under English law, subject to the jurisdiction of the courts of Luxembourg. The payment obligations of the EFSF set out in these agreements are subject to compliance of the beneficiary MS with a Memorandum of Understand‐ ing (MoU), an agreement concluded between the beneficiary MS on the one hand and the other Euro-MS on the other hand. This agreement is negotiated and concluded, on behalf of the other Euro-MS, by the Commis‐ sion. When acting in this capacity, the Commission is again bound by EU law, in particular the Charter of Fundamental Rights of the European Union (CFR). The mechanism as a whole cannot be assigned to one level, neither institutionally nor in terms of the applicable law. Such an allocation should be attempted only at a micro-level, that is to say with regard to the single acts involved. But even if an allocation is eventually made, because institutionally or in terms of the applicable law the respective indicators qualitatively outweigh elements pointing to other bodies or legal orders, this allocation – as announced above – may be of limited value, because it cannot do away with the named deviating elements. It is still required to consider additional (subordinate) influences which institutionally or in terms of the applicable law belong to other levels.

These principal considerations on the merits as well as on the pitfalls of assigning an act to an institution and to a legal order are meaningful not only in the context of law, but as well in the context of soft law. They should, therefore, be borne in mind also when addressing, in the following sub-chapters, the originators of EU soft law.

assignment of acts to a certain body and/or legal order, shall be considered in more detail under 2.2.4.1. below.

### 2.2.2. The EU's institutions

The institutions as listed in Article 13 para 1 TEU all have the power – in general or only in specific cases<sup>453</sup> – to adopt soft law. This is explicitly expressed in Article 288 para 1 TFEU – although the respective competence is not thereby conferred – with regard to recommendations and opinions.

The European Council, to begin with, is *qua* the role accorded to it by the Treaties, literally destined to adopt soft law acts. In charge of providing the EU 'with the necessary impetus for its development' and of defining 'the general political directions and priorities thereof ', but at the same time lacking 'legislative functions',<sup>454</sup> soft law appears to be an adequate form of expression for this institution. In practice, the European Council mostly adopts conclusions, resolutions or declarations in order to fulfil its remit. Its conclusions are adopted during each summit.455 They may have an annex containing resolutions, declarations, reports, etc of the Eu‐ ropean Council.456 While the European Council's conclusions, resolutions and declarations may and often do contain soft law provisions – mostly addressed to the Commission and the EU legislator, but also eg to the European Central Bank (ECB), or to the MS (or its ministers) – this assumption needs to be verified case by case. In particular, the requests to other actors uttered therein need to be concrete enough to actually have a concrete steering function.457 The conclusions most of the time also contain other parts in which the European Council – 'without establishing concrete rules or measures' – 'stresses, recalls, notes, agrees, considers, underlines, emphasises, recognises, welcomes, appreciates, etc'.458 The comparatively large quantity of soft law acts of the European Council shall not belie the fact that the European Council disposes of important hard law powers,

<sup>453</sup> The possibilities of the CJEU to adopt soft law, for example, are relatively limited; see eg Nettesheim, Art. 13 EUV, para 85; see also below.

<sup>454</sup> Article 15 para 1 TEU.

<sup>455</sup> For a list of all European Council conclusions since 1975 see <http://www.consilium .europa.eu/en/european-council/conclusions/> accessed 28 March 2023.

<sup>456</sup> See eg Conclusions of the Presidency of the Luxembourg European Council, 12–13 December 1997.

<sup>457</sup> See Senden, Soft Law 194 f. The 'call[ing] on the Commission and the Member States to implement swiftly the priority projects' as laid down in Conclusions EUCO 1/16 of the European Council meeting (18 and 19 February 2016) 3, for example, does not reach the required level of concreteness.

<sup>458</sup> Senden, Soft Law 194; see also Senden, Balance 82.

eg regarding the election of its President,459 its Rules of Procedure,<sup>460</sup> the number of Commissioners<sup>461</sup> or the composition of the European Parlia‐ ment<sup>462</sup> . 463

The Council adopts a wide range of different soft law acts, in particular recommendations,<sup>464</sup> declarations, resolutions<sup>465</sup> or guidelines, the adop‐ tion of some of which are specifically laid down in the Treaties (see 3.5.2.2. below). Generally, these acts are more specific and more detailed than the European Council's (soft law) output.<sup>466</sup> They may be addressed in particular to the MS and the Commission, but also eg to the European Parliament.<sup>467</sup> Where soft law provisions are incorporated in a piece of legislation they still qualify as soft law, even though formally they are part of a legislative act.<sup>468</sup>

The European Parliament adopts recommendations, usually together with the Council.469 Otherwise it makes use in particular of resolutions which cover subjects as diverse as 'European conscience and totalitaria‐


<sup>459</sup> See eg Decision (EU) 2022/492 of the European Council, based on Article 15 para 5 TEU.

<sup>460</sup> See eg Decision 2009/882/EU of the European Council, based on Article 235 para 3 TFEU.

nism'<sup>470</sup>, the avoidance of food waste<sup>471</sup> or the situation in Ukraine<sup>472</sup>. These resolutions are regularly addressed to the Council and the Commission, but – as was the case with the resolution on European conscience and totalitarianism – may also be addressed eg to the parliaments of the MS, the governments and parliaments of the candidate countries, the governments and parliaments of the countries associated with the EU, and the govern‐ ments and parliaments of the members of the Council of Europe. Generally, these resolutions, which often have a foreign policy thrust, if at all, aim at behavioural steering at a very high level of abstraction, and hence in the majority of cases – for lack of concreteness<sup>473</sup> – cannot be called soft law.<sup>474</sup> Against this background, it is to be understood that the CJEU acknowledges a comprehensive power of the EP to adopt resolutions: 'What is more, it must be emphasised that the powers of the Governments of the Member States in the matter do not affect the right inherent in the Parliament to discuss any question concerning the Communities, to adopt resolutions on such questions and to invite the Governments to act'.<sup>475</sup>

The Commission shall render recommendations and opinions according to the Treaties in a number of cases.476 They are addressed in particular to the Council, the European Parliament477 (opinions478) and the MS<sup>479</sup>

<sup>470</sup> European Parliament resolution of 2 April 2009 on European conscience and totali‐ tarianism.

<sup>471</sup> European Parliament resolution of 19 January 2012 on how to avoid food wastage: strategies for a more efficient food chain in the EU.

<sup>472</sup> European Parliament resolution of 6 February 2014 on the situation in Ukraine; European Parliament resolution of 1 March 2022 on the Russian aggression against Ukraine.

<sup>473</sup> For the minimum degree of determination of soft law see II.2.1.2. above.

<sup>474</sup> See Wittinger, Europarat 143; see also joined cases C‑72/10 and C‑77/10 *Costa*, para 74, with further references: 'The principle of legal certainty requires, moreover, that rules of law be clear, precise and predictable as regards their effects, in particular where they may have unfavourable consequences for individuals and undertakings'.

<sup>475</sup> Case 230/81 *Luxembourg v European Parliament*, para 39.

<sup>476</sup> See in particular the Commission's general power to adopt recommendations under Article 292 TFEU. On the dominant role of the Commission in this context see also Nettesheim, Art. 288 AEUV, para 203.

<sup>477</sup> See eg Framework Agreement on relations between the European Parliament and the European Commission (2010), OJ L304/47, paras 21, 37 f.

<sup>478</sup> See Senden, Soft Law 162 f.

<sup>479</sup> See eg Commission Opinion on the Rule of Law in Poland and the Rule of Law Framework (of Article 7 TEU) <http://europa.eu/rapid/press-release\_MEMO-16-2 017\_en.htm> accessed 28 March 2023; Commission Recommendation regarding the rule of law in Poland, C(2016) 5703 final.

(recommendations, opinions). In practice, the Commission adopts further acts which may constitute in their entirety or at least contain soft law, in particular the so-called Communications.<sup>480</sup> Communications are a 'recht schillerndes und facettenreiches Phänomen'481 [rather chatoyant and multifaceted phenomenon] and may serve a variety of (at times overlapping) purposes – informative, explanatory,<sup>482</sup> preparatory (eg Green and White Papers) or concretising (eg Communications rendered in the context of competition and State aid law) purposes. Like with the other institutions, bodies, offices or agencies, also the Commission's legally non-binding out‐ put may have a merely internal scope<sup>483</sup> (eg so-called 'rules of conduct'<sup>484</sup>).

The ECB shall be consulted, and hence may adopt an opinion, with regard to all proposed Union acts and proposals for regulation on the national level '[w]ithin the areas falling within its responsibilities'.<sup>485</sup> In ad‐ dition to that, it adopts recommendations without specified addressees, eg on payment transactions486 and, within the Single Supervisory Mechanism (SSM), recommendations in the field of banking supervision addressed to financial institutions and to national supervisors.487 Most of the time these acts contain soft law.

Also the CJEU and the Court of Auditors may render output which qualifies as soft law. As regards the CJEU, mention should be made of the Opinions which may be rendered by the Advocate General in proceed‐ ings before the CJEU.488 The Opinion shall support (and is intended to influence) the Court in the decision-making process and it is legally non-


<sup>480</sup> These acts may also be named differently, eg 'guidelines', 'notices', 'codes', 'policy frameworks'; see H Adam, Mitteilungen 3; Senden, Soft Law 162; Snyder, Effective‐ ness 33; Raschauer, Leitlinien.

<sup>481</sup> Brohm, Mitteilungen 25.

<sup>482</sup> See eg case C-501/15P *European Union Intellectual Property Office v Cactus*, para 40.

<sup>483</sup> See Pampel, Rechtsnatur 89; see also Hofmann/Rowe/Türk, Administrative Law 572, who consider Commission-internal rules 'always binding, but logically one cannot there speak of an *externally* binding effect' (emphasis in original).

<sup>486</sup> Eg the ECB Recommendations for the security of internet payments and of mobile payments (both adopted in 2013).

<sup>487</sup> Eg the ECB Recommendation on dividend distribution policies, ECB/2019/1.

<sup>488</sup> Since the Advocates General institutionally belong to the Court (Article 252 TFEU), it is justified to also assign their opinions to the Court – in spite of the recognition of the incumbents and their respective output as individual/distinct in the literary debate; see eg Karpenstein, Art. 252 AEUV, paras 16 f.

binding.489 The Court in its judgements often<sup>490</sup> follows the Opinions (with regard to its result or its reasoning, or both<sup>491</sup>), but in places also deviates, sometimes considerably, from them. As another soft law act, the Court has adopted 'Recommendations to national courts and tribunals in relation to the initiation of preliminary ruling proceedings'.<sup>492</sup> The jurisdictional out‐ put of the Court, however, is always binding – a non-binding jurisdiction would 'change the nature of the function of the Court of Justice'.<sup>493</sup> The Court of Auditors examines the accounts of all revenue and expenditure of the Union and principally of all bodies, offices or agencies set up by the Union. An annual report is forwarded to the EU institutions and pub‐ lished in the Official Journal of the European Union (OJ).<sup>494</sup> Apart from an account of the implementation of the respective budget, these reports regularly include conclusions/recommendations to which the addressees may respond. It is in particular these recommendations which regularly constitute soft law. This applies also to the Court of Auditor's special reports.495 The measures taken notably by the Commission in response to recommendations uttered earlier are included in the annual report. In

495 See eg the Special Report 'Single Resolution Board: Work on a challenging Banking Union task started, but still a long way to go' (No 23, 2017).

<sup>489</sup> See, *ex multis*, Hackspiel, Art. 252 AEUV, para 12; for the soft law quality of these opinions see also 3.5.2.5. below.

<sup>490</sup> See Pirrung, Gerichtshof, who speaks – as a rule of thumb – of the Court following the AG in 80 % of the cases; see also de Búrca, Court 23 (fn 121), with further references; with regard to annulment procedures, it is argued that the Court is 67 % more likely to annul (parts of ) an act when the AG so suggests: Arrebola/Mauri‐ cio/Jiménez Portilla, Analysis.

In general, a distinction is to be made between cases which can build on established case law and cases raising new questions of law. With regard to the latter, apparently the Court is less likely to follow the AG's suggestions.

<sup>491</sup> Cases in which the Opinion of the AG had a particularly strong influence on the Court's decision are, eg, the cases C-200/02 *Zhu*, explicit reference in para 20, or C-224/01 *Köbler*, explicit reference in para 48; see more generally Haglund, Advocate General; see also Arrebola/Mauricio/Jiménez Portilla, Analysis 1.

<sup>492</sup> For a discussion of the Court's competence to adopt these recommendations see 3.3.2.2. below.

<sup>493</sup> See Opinion 1/91 *EEA I*, paras 59–62, in particular para 61.

<sup>494</sup> Article 287 para 4 TFEU. Note that there is also a number of specific annual reports on the annual financial audits of the EU's agencies, joint undertakings and other decentralised bodies; see eg <https://www.eca.europa.eu/en/Pages/DocItem.aspx? did={B72375E3-B0E0-467A-AB50-55536ACAC4DE}> accessed 28 March 2023.

addition to that, the Court of Auditors renders observations or requests opinions on matters of accounting, eg in legislative processes.<sup>496</sup>

# 2.2.3. The EU's bodies, offices and agencies

Apart from the institutions, also the bodies, offices and agencies of the EU in many instances adopt soft law acts.<sup>497</sup> This soft law output is manifold. Suffice it to illustrate the variety of acts by selected examples. The European Medicines Agency (EMA) provides guidelines on different aspects of me‐ dicinal products for human use, for example quality or clinical efficacy and safety.<sup>498</sup> They are mainly addressed to applicants for and holders of a market authorisation for medicinal products. The former, in their applications, need to justify deviations from these guidelines.<sup>499</sup> The Euro‐ pean Aviation Safety Agency (EASA), to take another example, publishes so-called certification specifications, soft law rules on different aspects of aviation safety which are first and foremost addressed to those applying for a certification.500 Also the 'Acceptable Means of Compliance' through which the EASA concretises EU aviation law adopted by the legislator or the Com‐ mission ought to be mentioned in this context.501 Within the category of EU bodies, offices and agencies fall also the Economic and Social Committee and the Committee of the Regions, two advisory bodies provided for in the Treaties.502 These two bodies shall submit opinions to the EP, the Council or the Commission when requested to do so or on their own initiative, in particular during legislative or other decision-making procedures.503 Yet another example of an EU body, office or agency adopting soft law is the

<sup>496</sup> Eg Court of Auditors, Opinion 2/2001 on the proposal to recast the Financial Regulation.

<sup>497</sup> See Rocca/Eliantonio, Soft Law, *inter alia* pointing out the fact that 20 European agencies have explicit soft law powers (page 6); with regard to the varied soft law output of European agencies see Senden/van den Brink, Checks 42 ff.

<sup>498</sup> See Fleischfresser, Europäisierung, in particular para 31.

<sup>499 &</sup>lt;http://www.ema.europa.eu/ema/index.jsp?curl=pages/regulation/general/general \_content\_000081.jsp> accessed 28 March 2023.

<sup>500</sup> See Riedel, Gemeinschaftszulassung, in particular 116 f; Simoncini, Regulation 81 ff.

<sup>501 &</sup>lt;https://www.easa.europa.eu/document-library/agency-rules-overview> accessed 28 March 2023.

<sup>502</sup> Articles 300 ff TFEU.

<sup>503</sup> Article 304 para 1 and Article 307 TFEU.

European Ombudsman.504 According to the pertinent EP decision, the Ombudsman shall 'help to uncover maladministration in the activities of the Community institutions and bodies […] and *make recommendations with a view to putting an end to it*' (emphasis added). <sup>505</sup> The *de facto* power of the Ombudsman's soft output is underlined by *Craig* and *de Búrca* who claim that the Ombudsman's office is 'increasingly seen as a source of administrative norms rather than simply a mediation facility for individual complaints'.<sup>506</sup>

In this context, also bodies established by international agreements con‐ cluded between the EU and third parties, in particular third countries, ought to be mentioned. Examples are the Cooperation Council established by Partnership and Cooperation Agreements or the Association Council and the Association Committee established by Association Agreements, which may adopt recommendations relating to the implementation of their respective founding agreements.507 International agreements concluded by the EU are not only part of public international law, but also part of the *acquis communautaire*. <sup>508</sup> Bodies established by them could be qualified as EU bodies. In the perspective of the other party to the agreement – regularly a third country, however, this body constitutes a body set up only by (bilateral) public international law. Therefore also the (soft) output of these bodies has a dichotomic character.

508 See Article 216 para 2 TFEU; this is also settled case law (also eg for mixed associa‐ tion agreements): see eg case 12/86 *Demirel*, para 7, with a further reference.

<sup>504</sup> Stressing the (merely) soft character of the Ombudsman's output: Order in case T-103/99 *Cantine Sociali Venete*, paras 48–50; see also case C-234/02P *European Ombudsman*, para 57.

<sup>505</sup> Article 2 para 1 of EP Decision 94/262/ECSC, EC, Euratom, as amended; see Bonnor, Ombudsman.

<sup>506</sup> Craig/de Búrca, EU Law 85; for a possible move towards hard law see Saurer, Verwaltungsrecht 190 f.

<sup>507</sup> See eg Article 78 of the Partnership and Cooperation Agreement establishing a partnership between the European Communities and their Member States, of the one part, and the Republic of Uzbekistan, of the other part (1999), with regard to the Cooperation Council; Articles 78 ff of the Euro-Mediterranean Agreement establishing an association between the European Communities and their Member States, of the one part, and the Kingdom of Morocco, of the other part (2000), with regard to the Association Council and the Association Committee.

### 2.2.4. MS and non-EU bodies

### 2.2.4.1. Acts relating to different legal orders

In addition to the institutions, bodies, offices and agencies of the EU, there are also bodies which organisationally do not belong to the Union but may nevertheless adopt acts belonging to the EU legal order, or at least contribute to their adoption.<sup>509</sup> This applies to EU soft law<sup>510</sup> as well as to EU law and is possible either due to an express authorisation granted by EU law or, according to some, due to a strong organisational and substantial proximity to the EU and its affairs.<sup>511</sup> With regard to the latter, *Wellens* and *Borchardt*, in an early account of Community soft law, said: 'The more closely the act corresponds to the realisation of the objectives or to the institutional structure of the EEC Treaty, the more the act acquires a community character'.512 For the qualification of an act as EU law or at least partly EU law, this organisational and substantial proximity must be of a certain intensity, and is not reached already where EU institutions 'may play a certain role' in the context of the creation of the act at issue.513 Conversely, the participation of non-EU actors in the creation of an act does not as such prevent its qualification as EU (soft) law.

Taking a more systematic approach, in the context of the assignation of an act to a legal order in my view the first point of reference ought to be the body uttering the norm, that is to say the body creating the act in accordance with its competences. With *agreed* (soft) law these are normally the bodies which shall be committed (softly) by the act.514 Other points of reference – eg contributors or the persons affected by the act (if they are not at the same time the official norm-creators anyway) – may be relevant, as well, but only at a secondary level. While the assignation of an act to

<sup>509</sup> See eg Article 76 lit b or Article 173 para 2 TFEU.

<sup>510</sup> See the example of the Code of Conduct for Mediators described by Korkea-aho, Soft Law 282.

<sup>511</sup> For the principal impossibility of (private) non-EU actors creating *administrative* EU law see Hofmann/Rowe/Türk, Administrative Law 588.

<sup>512</sup> Wellens/Borchardt, Soft Law 304.

<sup>513</sup> See joined cases C-8–10/15P *Ledra*, para 54; Opinion of AG *Wahl* in these joined cases, para 53.

<sup>514</sup> The parties to soft agreements, according to the principle *pacta sunt servanda*, have to benevolently examine whether or not they intend to follow it; see Lorenzmeier, Völkerrecht 76. With regard to agreed soft law more generally Knauff, Regelungsver‐ bund 373–376; see also Georgieva, Soft Law 236 f.

one (exceptionally: two) legal order(s) is certainly important for systematic reasons, it is to be acknowledged that in cases of doubt it is not only the heading – 'act of public international law', 'act of Union law', or 'hybrid act' – that matters, but it is in particular the consequences following from this qualification in the specific case, ie the effects of the involved legal orders, eg in terms of a certain fundamental rights standard to be met. These concrete consequences need to be examined individually in each case.<sup>515</sup> Generalisations are thus to be handled with care.

There is a number of 'cases of doubt', that is to say of (soft law or legal) acts which have a strong proximity to different legal orders. For example: An act *of the MS* – even if concluded in the course of a meeting of the Council or the European Council<sup>516</sup> – *prima facie* appears to be public international law,<sup>517</sup> not EU law.<sup>518</sup> AG *Jacobs* describes these acts as having a 'hybrid character', stressing that 'decisions of the Member States meeting

<sup>515</sup> For the multi-faceted manifestations of 'intergovernmentalism' and the related ty‐ pology see Hinarejos, Crisis 87.

<sup>516</sup> See eg the decision of the Heads of State or Government, meeting within the Euro‐ pean Council, annexed to the Conclusions EUCO 1/16 of the European Council meeting (18 and 19 February 2016); see also the recommendation of the Euro-MS according to Article 140 para 2 subpara 2 TFEU, Article 34 TEU (pre-Lisbon) or Article 220 TEEC; referring to the acts based on the two latter provisions as 'conven‐ tions': Bast/Heesen, Community, para 3. For the various institutional settings in which the MS took and partly still take 'decisions' see Everling, Wirkung 142 (and 135 f for the meaning of such 'decisions' in the earlier history of the EEC); for an ear‐ ly account of the varying views in literature and practice (with regard to agreements concluded between MS) see Schwartz, Übereinkommen 556 ff; for pertinent Court cases see, eg, joined cases C‑59/18 and C‑182/18 *Italy v Council*, paras 100–105; with regard to the Eurogroup see joined cases C-105/15P to C-109/15P *Mallis*, para 61; joined cases C-597/18P, C-598/18P, C-603/18P and C-604/18P *Chrysostomides*, para 87.

<sup>517</sup> That does not necessarily mean that they cannot be considered or even interpreted by the CJEU according to its peculiar methods of interpretation; see Everling, Wirkung 147; for the interpretation of only EU-related law by the CJEU see also case C-53/96 *Hermès*, para 28; joined cases C-300/98 and C-392/98 *Dior*, para 35.

<sup>518</sup> See Senden, Soft Law 56–58, with further references; critically: Kadelbach, Art. 5 EUV, para 14; Ruffert, Art. 288 AEUV, para 107; Tridimas, Indeterminacy 57. The Court held that acts of the Representatives of the Governments of the MS may be camouflaged acts of the Council; joined cases C-181/91 and C-248/91 *European Parliament v Council*, paras 14 and 25. Thereby it challenged the assumption that it actually had been the MS themselves concluding the act, and qualified the act as a Council act ('an almost metaphysical distinction', as *Brown* notes), but it does not seem to exclude in principle the possibility of the MS adopting an act themselves while meeting in the Council; see Brown, Case Law 1355.

in Council do not form part of the Community legal order in the strict sense, but are nevertheless part of the acquis communautaire'.<sup>519</sup> The Court has expressed its willingness to consider such acts to be EU law where in terms of objectives and institutional setting they display strong ties to the EU.<sup>520</sup> In other cases – like that of the 'EU-Turkey agreement' on refugees – the Court eventually refused the EU law quality of the act at issue, explaining that 'the term "EU" must be understood in this journalistic context as referring to the Heads of State or Government of the Member States of the European Union'.<sup>521</sup>

As was mentioned above, the adoption of EU (soft) law by bodies organi‐ sationally not belonging to the EU is possible due to an authorisation gran‐ ted according to EU law or due to a strong organisational and substantial proximity to the EU and its affairs. While the cases mentioned above are examples of the latter scenario, an (historical) example of non-EU bodies adopting EU soft law *qua* authorisation are the Committees of European Banking Supervisors (CEBS), of European Insurance and Occupational Pensions Supervisors (CEIOPS) and of European Securities Regulators (CESR). They were composed of representatives of the respective national supervisory authorities and succeeded by today's European Financial Mar‐ ket Supervisory Authorities (ESAs). To these committees, organised as legal persons according to English, German, and French respectively, private law,

<sup>519</sup> Opinion of AG *Jacobs* in joined cases C-181/91 and C-248/91 *European Parliament v Council*, para 18.

<sup>520</sup> Case 38/69 *Commission v Italy*, para 11: '[A] measure which is in the nature of a Community decision on the basis of its objective and of the institutional frame‐ work within which it has been drawn up cannot be described as "international agreement"'. More generally speaking, the Court seems to consider the substance of the act at issue, rather than its form: '[I]t is not enough', it held, 'that an act should be described as a "decision of the Member States" for it to be excluded from review', but a substantial assessment is required to find out whether it is an act of the EU or of the MS; joined cases C-181/91 and C-248/91 *European Parliament v Council*, para 14. Note that the Court also held that the Eurogroup 'cannot be equated with a configuration of the Council or be classified as a body, office or agency of the European Union within the meaning of Article 263 TFEU'; joined cases C-105/15P to C-109/15P *Mallis*, para 61. For the different interpretative approaches in such cases of doubt see Everling, Wirkung 153; for EU soft law adopted by the MS see also Peters/Pagotto, Soft Law 18.

<sup>521</sup> Case T-192/16 *NF*, paras 57 f; on the underlying question see also case C-11/05 *Friesland Coberco*, paras 37 f.

the Commission522 delegated<sup>523</sup> the power to 'contribute to the common and uniform implementation and consistent application of Community legislation by issuing guidelines, recommendations and standards'<sup>524</sup> and to address opinions to the Commission and to the national supervisory authorities.<sup>525</sup> They were thereby functionally acting as EU bodies.<sup>526</sup> In this example, it was the legal form of the delegates which was private. Their respective output, however, was still decided upon by an assembly of public authority representatives.

A different phenomenon was described as '[p]rivate involvement in EU governance'.<sup>527</sup> The actors addressed here are not only formally but also substantially (that is to say: regarding their professional background) *pri‐ vate* actors. According to the Inter-institutional Agreement on better lawmaking of 2003,<sup>528</sup> private actors may merely participate in EU rule-making in the form of co-regulation,529 or apply self-regulation which will not result in EU (soft) law. Co-regulation is applied in particular in the fields of stand‐ ardisation (see II.2.2.3. above) and social policy, and is a procedure aimed at pooling expertise in EU law-making.530 Self-regulation is to do with agreements among private actors, often – but not necessarily – enacted at

<sup>522</sup> Commission Decisions 2004/5/EC, 2004/6/EC and 2001/527/EC; these acts were – in the course of the reform of the *Lamfalussy* procedure in 2009 – replaced by Commission Decisions 2009/77–79/EC; see also Weismann, Agencies 93–97.

<sup>523</sup> For the delegation of EU powers to private bodies see eg Pawlik, Meroni-Doktrin 147 f.

<sup>524</sup> Article 3 of Commission Decisions 2009/77–79/EC; see also Ottow, Architecture 128.

<sup>525</sup> Article 12 and Article 4 para 1 lit b of Commission Decisions 2009/77–79/EC.

<sup>526</sup> Also organisationally they were considered to be connected to the EU administra‐ tion; see European Ombudsman, case 2497/2010/FOR, confirming the Ombuds‐ man's competence to deal with complaints about the CEBS' alleged maladministra‐ tion (para 10).

<sup>527</sup> Hofmann/Rowe/Türk, Administrative Law 328; in the context of EU soft law see eg van Rijsbergen, Legitimacy 124 ff, with regard to soft law adopted by the ESMA.

<sup>528</sup> European Parliament, Council and Commission, Interinstitutional Agreement on better law-making (2003/C 321/01) [not to be confused by the same-titled Agree‐ ment of 2016 (OJ L123/1)].

<sup>529</sup> For the varying definitions of co-regulation see references in Verbruggen, Co-Regu‐ lation 428–430.

<sup>530</sup> See Hofmann/Rowe/Türk, Administrative Law 605 ff; see also Köndgen, Re‐ chtsquellen, paras 67–70, with examples; for the long history of co-regulation in E(E)C/EU law (starting in the mid-80s) see Verbruggen, Co-Regulation 426 f.

the legislative or other initiative of the EU institutions.<sup>531</sup> Such agreements may relate to and result in EU (soft) law, though.<sup>532</sup> For example: The European Automobile Manufacturers Association (ACEA) has agreed with its Korean and Japanese pendants to reduce CO<sup>2</sup> emissions, based on the requirements laid down in Directive 98/70/EC. This agreement has led to three recommendations in which the Commission recommends to these associations to reduce their CO<sup>2</sup> emissions in accordance with pertinent EU law. Compliance with their respective commitment is monitored by the Commission.<sup>533</sup>


A specific case of 'acts relating to different legal orders' are the MoU conclu‐ ded within the framework of the various European umbrellas set up in reac‐ tion to the State debt crisis in the Eurozone. Here the contested questions

532 See Schwarze, Soft Law 234 f, with a further reference.

<sup>531</sup> See European Parliament, Council and Commission, Interinstitutional Agreement on better law-making (2003/C 321/01) para 22; for example, an agreement of com‐ panies such as Apple, Facebook or Microsoft to ensure children's safe use of the internet; <https://www.reuters.com/article/internet-eu-bullying-idUSLA36235620 090210> accessed 28 March 2023; see also the second MoU on the future common charging solution for smartphones, concluded by major telecommunications firms, which includes a reporting duty *vis-à-vis* the Commission; or the Commission's strengthened 'Code of Practice on Disinformation' with its 34 signatories; <https:/ /ec.europa.eu/commission/presscorner/detail/en/IP\_22\_3664> accessed 28 March 2023; on voluntary cooperation between public and private actors in relation to (self-)regulation more generally see Héritier/Eckert, Modes; for a similar mecha‐ nism see case T-135/96 *UEAPME*, para 9, relating to a framework agreement conclu‐ ded by management and labour organisations which is envisaged to be adopted by the Council as a legislative act.

<sup>533</sup> Commission Recommendations 1999/125/EC, 2000/303/EC and 2000/304/EC. This example is taken from Hofmann/Rowe/Türk, Administrative Law 620. In another case, concerning the safety of pedestrians, voluntary agreements demanded by the Commission and signed by the car industry were considered inappropriate by the EP: '[T]he Union could not abandon its legislative powers to third parties when the protection of citizens was at stake'. Subsequently, an according Directive was adopted; see Commission, Report 'Better Lawmaking 2003', COM(2003) 770 final, 26 f.

to which legal order they (rather) belong and whether or not they qualify as soft law shall be examined.<sup>534</sup> These umbrellas have granted financial assistance to ailing Euro-MS in order to allow them to service their debts and to thereby improve their credit-worthiness. <sup>535</sup> This assistance has been 'strictly conditional' and has, for that purpose, been combined with MoU concluded between the beneficiary MS and the respective facility providing financial assistance (with loans constituting the standard form of financial assistance) to ensure that the MS takes the (presumably) necessary reform measures in order to increase its income and to cut on expenses respective‐ ly (so-called austerity measures<sup>536</sup>) <sup>537</sup>. The MoU, as norms agreed upon between MS and the loan-providing facility, are not self-executing and therefore do not directly affect individuals.<sup>538</sup> Only the national implement‐ ing measures do. While in an early phase of the State debt crisis MS provi‐ ded bilateral (conditional) loans to other MS in trouble,539 here we shall concentrate on the MoU concluded in the context of financial assistance granted by the European Financial Stabilisation Mechanism (EFSM), the EFSF and the European Stability Mechanism (ESM) via financial assistance facility agreements.

The EFSM was created by Council Regulation 407/2010 which was based on Article 122 para 2 TFEU, hence it clearly belonged to the EU legal order. The MoU concluded under the EFSM between the Commission and the beneficiary State shall be 'detailing the general economic policy conditions laid down by the Council' and shall be communicated to the European

<sup>534</sup> With regard to further acts of a dubitable legal quality, adopted by EU institutions in the course of the Euro-crisis, see Beukers, Changes 96; for the *letter of intent* which – in the context of the MoU – results from the negotiations between the beneficiary MS and the IMF see Torsten Müller, Troika 266; for an account of MoU concluded by the ECB see Karatzia/Konstadinides, Nature 450–453.

<sup>535</sup> For other important purposes see, with regard to the ESM, Schwarz, Memorandum of Misunderstanding 415 f. Apart from that, the umbrellas (in particular the ESM) were also aimed at stabilising the Eurozone as a whole.

<sup>536</sup> For decisions of national (constitutional) courts adopted in the context of (nation‐ al) austerity measures see eg the decisions of the Portuguese constitutional court discussed in: Canotilho/Violante/Lanceiro, Austerity.

<sup>537</sup> For the legal framework of austerity measures more generally see Repasi, Protection 1136.

<sup>538</sup> See Repasi, Protection 1137 f, drawing a comparison to directives.

<sup>539</sup> Also the bilateral loans provided to Greece – the predecessor instrument of the umbrellas – were combined with a prescription of austerity measures; see Olivares-Caminal, Architecture 4; Kilpatrick, Bailouts 398; see also case T-541/10 *ADEDY*, paras 12–19.

Parliament and to the Council.540 Hence the decision to grant financial assistance, also laying down the general economic policy conditions, under this regime is taken by the Council, upon a proposal from the Commis‐ sion.<sup>541</sup> The MoU is concretising these general economic policy conditions (*argumentum* 'detailing'). The EFSM which has granted loans in three cases – namely for Ireland, for Portugal and for Greece (short-term assistance) – was operating in coordination with the International Monetary Fund (IMF), which in these cases provided loans as well.<sup>542</sup> In spite of this coordination with the IMF as an international organisation, the EFSM, and also the acts based on Council Regulation 407/2010, are clearly EU law measures.

The EFSF was established as a *société anonyme* incorporated in Luxem‐ bourg, the shareholders of which were the Euro-MS. The operation of the EFSF is subject to a private law agreement between it and its shareholders (the Euro-MS; EFSF Framework Agreement).543 The close links to EU law and EU institutions established by this agreement do not alter its private law nature. According to this agreement, the Commission shall, upon request by a Euro-MS for a Financial Assistance Facility Agreement, and in liaison with the ECB and the IMF, negotiate an MoU with the prospective beneficiary State.544 The MoU shall be in accordance with a Council decision adopted, upon a Commission proposal, pursuant to Arti‐ cle 136 para 1 TFEU, and it shall, upon approval by the Eurogroup Working Group, be signed by the Commission on behalf of the EFSF and by the beneficiary MS. The Financial Assistance Facility Agreement, which shall be compatible with the MoU, shall eventually be signed by the EFSF upon

<sup>540</sup> Article 3 para 5 of Regulation 407/2010.

<sup>541</sup> Article 3 paras 2 f of Regulation 407/2010.

<sup>542</sup> For other (partly) international organisations providing financial assistance and for the conditionality they apply (or do not apply) see Wirth, Assistance 220 f; for the World Bank's approach see Boisson de Chazournes, Guidance 289 f.

<sup>543</sup> See EFSF Framework Agreement between the Euro-MS and the EFSF <https://ww w.sv.uio.no/arena/english/people/guest-researchers/agustinm/crisis-documents -2012/14-efsf-frameworkagreement-consolidated-8sep11.pdf> accessed 28 March 2023. This agreement shall be governed by and shall be construed in accordance with English law; 16. (1) of the EFSF Framework Agreement between the Euro-MS and the EFSF.

<sup>544</sup> For the IMF's role in the 'Troika' and in setting conditionality requirements see Christopherson/Bergthaler, IMF, paras 31.65 – 31.70.

unanimous approval by all Euro-MS and by the beneficiary MS.<sup>545</sup> That also under this regime it is the Commission (on behalf of the Euro-MS, not of the EFSF<sup>546</sup>) which concludes the MoU with the beneficiary MS seems to be effected by the EFSF Framework Agreement, not by EU law. However, the mandate of the Commission to negotiate and to conclude, on behalf of the Euro-MS (and conditional upon approval by the Eurogroup Working Group<sup>547</sup>), an MoU cannot be conferred by a measure of private law (such as the Framework Agreement). After all, the subject of an MoU are in particular matters of public authority, namely issues like how the beneficiary State will adapt its pension law, its unemployment law, the enforcement of its tax law, etc. Therefore it seems appropriate to qualify the mandating of the Commission – even if formally contained in the EFSF Framework Agreement<sup>548</sup> – as a unilateral act of the Euro-MS adopted, for lack of a legal basis in EU law, within the realm of public international law.549 Since it is matters relating to public authority which are addressed, and since there is no appropriate legal basis in EU law, also the MoU in this case is to be qualified as a measure of public international law.550 In


<sup>545</sup> See 2. (1) (a) of the EFSF Framework Agreement between the Euro-MS and the EFSF.

<sup>546</sup> See Recital 2 and 2. (1) (a) of the EFSF Framework Agreement between the Euro-MS and the EFSF; misleadingly: Recital 4 of the Master Financial Assistance Facility Agreement between the EFSF, Ireland and the Central Bank of Ireland.

spite of this qualification, the MoU are in some respects related to the EU and its law: For example, the loans granted under the EFSF (in the case of Ireland and Portugal) – specified in an agreement concluded under English law<sup>551</sup> between the EFSF and the beneficiary State (and its central bank) which makes the provision of the loans conditional upon compliance with the MoU<sup>552</sup> – have been made 'subject to the (EU) EFSM legal regime and sources'.<sup>553</sup>

We shall now turn to the MoU concluded in the context of the ESM by the Commission (with the ECB's and the IMF's participation<sup>554</sup>) on the one hand, and a Euro-MS (and its central bank, respectively) on the other hand. Like under the EFSF, the Commission is not competent to conclude the MoU on the basis of EU law, but – *on behalf of the ESM*<sup>555</sup> – on the basis of the (international) ESM-Treaty (and *in consistency with* EU law556),557 and subject to approval by the ESM.558 At first there shall be a decision of the ESM Board of Governors to grant, in principle, financial assistance to the (presumptive) beneficiary Euro-MS on the basis of an MoU, to be negotiated by the Commission (in liaison with the ECB and, if possible, together with the IMF).559 In parallel, a financial assistance facility agreement on the financial terms and conditions and the choice of instru‐

<sup>551</sup> In the case of Ireland, see 14. (1) of the Master Financial Assistance Facility Agree‐ ment between the EFSF, Ireland and the Central Bank of Ireland.

<sup>552</sup> In the case of Ireland, see Recital 4 of the Master Financial Assistance Facility Agreement between the EFSF, Ireland and the Central Bank of Ireland.

<sup>553</sup> Kilpatrick, Bailouts 401, with further references.

<sup>554</sup> The IMF is concluding a (very similar) MoU with the benficiary country itself. Due to a considerable overlap, the two MoU form one *corpus* of rules; see Kämmerer, Memorandum of Understanding 74.

<sup>555</sup> See Article 5 para 6 lit g and Article 13 para 3 of the ESM-Treaty; see also joined cases C-8–10/15P *Ledra*, para 51; Opinion of AG *Wahl* in these joined cases, paras 100 f.

<sup>556</sup> For the difference between 'consistency' and 'compliance' see Opinion of AG *Wahl* in joined cases C-8–10/15P *Ledra*, para 73.

<sup>557</sup> See Order of the General Court in case T-289/13 *Ledra*, paras 44–46; joined cases C-105/15P to C-109/15P *Mallis*, para 53, both with reference to the CJEU's *Pringle* decision.

<sup>558</sup> Article 13 para 5 of the ESM-Treaty; see also Schwarz, Memorandum of Misunder‐ standing 420.

<sup>559</sup> See Article 13 paras 2 f of the ESM-Treaty. For the (draft) MoU and the involvement of the Commission, the Eurogroup and the ESM see also joined cases C-8–10/15P *Ledra*, paras 14 ff.

ments560 shall be prepared by the ESM Managing Director and adopted by the Board of Governors.<sup>561</sup> Eventually, the MoU, consistent with EU law, including any opinion, warning, recommendation or decision addressed to the beneficiary MS, is signed by the Commission on behalf of the ESM on the one hand and by the beneficiary MS, on the other hand.<sup>562</sup> The Board of Directors shall approve the financial assistance facility agreement and, where applicable, the disbursement of the first tranche of the assistance.<sup>563</sup>

The ESM as the core actor in this context is a legal person established according to public international law.<sup>564</sup> It is the ESM which is bound by its financial agreements, not the EU. With regard to the involvement of the ECB, the following can be said: As a legal person, the ECB is not an organ of the legal person EU, but legally speaking acting for itself.<sup>565</sup> Nevertheless, as an institution of the EU<sup>566</sup> the ECB is vested with EU public authority. In the given context, however, the ECB – like the Commission – is empowered to act on the basis of the ESM-Treaty, not on the basis of the EU Treaties.<sup>567</sup> The fact that these roles are to be exercised *in accordance with* EU law does not change this. Rather, it is a consequence of the MS' and the EU institutions' being bound by EU law also when acting in the field of public

<sup>560</sup> For the most important instrument, the loan, see also the ESM Guideline on Loans <https://www.esm.europa.eu/sites/default/files/esm\_guideline\_on\_loans.pdf> accessed 28 March 2023.

<sup>561</sup> Article 13 para 3 of the ESM-Treaty.

<sup>562</sup> Article 13 paras 4 f of the ESM-Treaty; for complementary (EU law) measures such as the macroeconomic adjustment programmes see Repasi, Protection 1125 f and 1137 f. For the example of Greece see Council Implementing Decision 2015/1411, in particular Recitals 7–9; for the example of Cyprus see European Commission, 'The Economic Adjustment Programme for Cyprus' (2013) 149 European Economy. Occasional Papers, para 57.

<sup>563</sup> Article 13 para 5 of the ESM-Treaty.

<sup>564</sup> See the distinction made by *de Witte* between 'executive agreements', 'complementa‐ ry agreements' and 'autonomous agreements' concluded between the MS. According to this classification, the ESM appears to be an autonomous agreement.

<sup>565</sup> See Dörr, Art. 263 AEUV, para 18. Also an international legal personality of the ECB is accepted: Ott/Vos/Coman-Kund, Agencies 94.

<sup>566</sup> On the hermaphrodite role of the ECB under the Lisbon Treaty see Sáinz de Vicuña, Status 301–304.

<sup>567</sup> See Article 5 para 6 lit g and Article 13 para 3 of the ESM-Treaty; see also case C-370/12 *Pringle*, para 158; joined cases C-8–10/15P *Ledra*, para 52; joined cases C-597/18P, C-598/18P, C-603/18P and C-604/18P *Chrysostomides*, para 131; Craig, Pringle 280.

international law.568 That the ECB and the Commission (and also the IMF, for that matter) merely provide an input to the negotiations and do not exert any formal decision-making power indeed relativises the (formal) influence of these institutions.<sup>569</sup>

Against the background of these findings, it is argued here that the MoU concluded under the ESM – like the ones concluded under the EFSF – clearly belong to the realm of public<sup>570</sup> international law.<sup>571</sup> That the 'borrowing' of EU institutions applied in case of the EFSF and the ESM has raised questions as to its legality is uncontested, but shall not be dwelled on here.<sup>572</sup> Also the question whether the involvement of EU institutions causes the applicability of the CFR via its Article 51 (with regard to the MoU) – which was plausibly argued by *Kilpatrick*<sup>573</sup> and meanwhile confirmed by the Court in the *Ledra* case<sup>574</sup> – shall not be considered here in more detail.

<sup>568</sup> See Article 13 para 3 of the ESM-Treaty; see also page 3 of the Memorandum of Understanding on the working relations between the European Commission and the European Stability Mechanism.

<sup>569</sup> See case C-370/12 *Pringle*, para 161; note the competent (economic and monetary affairs) Commissioner's and the ECB-President's facultative participation in the meetings of the Board of Governors of the ESM as observers, though (Article 5 para 3 of the ESM-Treaty); see also Opinion of AG *Wahl* in joined cases C-8–10/15P *Ledra*, para 42. Critically with a view to the ECB's independence in this context: Opinion of AG *Cruz Villalón* in case C-62/14 *Gauweiler*, paras 143–151; see also Torsten Müller, Troika 269; differently: joined cases C-105/15P to C-109/15P *Mallis*, in particular para 57.

<sup>570</sup> Pointing at the differences between the loan agreements ('public law agreements') at issue and contracts on the delivery of goods or coordination of a project: Kilpatrick, Bailouts 407 f.

<sup>571</sup> See case T-293/13 *Theophilou*, para 46; joined cases C-8–10/15P *Ledra*, para 67; see also Fischer-Lescano, Austeritätspolitik 36; Repasi, Protection 1124. This also seems to be the view of the (other) EU institutions: European Parliament, Report on the enquiry on the role and operations of the Troika (ECB, Commission and IMF) with regard to the euro area programme countries, A7–0149/2014, para 109. That the EP would *wish* the MoU to be placed within the EU law framework, in particular in order to ensure the applicability of the CFR, is a different issue.

<sup>572</sup> See eg Craig, Pringle; Fischer-Lescano/Oberndorfer, Fiskalpakt; Peers, Form.

<sup>573</sup> See Kilpatrick, Bailouts 404 f. See the Explanations relating to the Charter of Funda‐ mental Rights (2007/C 303/02) which suggest – the wording of this provision being unclear in this respect – that the limitation 'when they are implementing Union law' of Article 51 para 1 only applies to the MS, but does not apply to the EU institutions and bodies; see also Schwarz, Memorandum of Misunderstanding 397–400 and 418–421.

<sup>574</sup> Joined cases C-8–10/15P *Ledra*, para 67.

# 2.2.4.2.2. On the question of legal bindingness

As regards the substance of the MoU concluded under the umbrellas, it is dubitable whether or not they constitute soft law. The title 'Memorandum of Understanding' suggests, in accordance with international practice, le‐ gal non-bindingness.<sup>575</sup> From the appearance of the concrete memoranda, however, no general conclusion can be drawn as to their legal quality. Comparatively 'hard' wording ('the Government will by [month/year]';<sup>576</sup> 'the Government commits to'), the partly very high precision of the rules<sup>577</sup> and the strong relationship between financial assistance and compliance with the MoU ('conditionality' <sup>578</sup>), is contrasted with an apparent lack of a classical enforcement regime. The monitoring set in place, however strict it may be, is a standard means of soft 'enforcement'.<sup>579</sup> While legally binding rules may (exceptionally) lack enforceability, an existing legal enforcement mechanism certainly gives proof of the legal quality of the underlying obligation. In the given case, namely that of financial assistance received

<sup>575</sup> See Kämmerer, Memorandum of Understanding 75; Kilpatrick, Bailouts 409 f, with further references; Wengler, Rechtsvertrag 194.

<sup>576</sup> For the legal bindingness this verb suggests see Knauff, Regelungsverbund 290; differently: case C-233/02 *France v Commission*, para 43, contrasting 'will' with the more compelling 'shall'. However, there are also many weaker expressions available than 'will': eg 'should strive for', 'intends to', 'should commit to'; see Opinion of AG *Tesauro* in case C-57/95 *France v Commission*, para 16.

<sup>577</sup> A high level of detail (weakly) points in the direction of legal bindingness; see also case 108/83 *Luxembourg v European Parliament*, para 23. In the context of the MoU see eg the part on tax policy reforms in the MoU concluded between the ESM and the Hellenic Republic and the Bank of Greece of 19 August 2015, 7–9; see also <https://www.imf.org/external/np/loi/2013/irl/060313.pdf> accessed 23 March 2023, pages 22 ff; Torsten Müller, Troika 274–278. Closer scrutiny of their text shows that some (actually few) provisions are drafted more widely, so as to leave a certain leeway for the beneficiary State.

<sup>578</sup> Article 13 para 3 of the ESM-Treaty. For the wide range of requirements this condi‐ tionality may in fact encompass see Schwarz, Memorandum of Misunderstanding 396 f. In *Pringle*, the Court seems to have elevated strict conditionality to a general requirement of financial assistance; case C-370/12 *Pringle*, paras 136 f; see also Ioannidis, Conditionality 62 f. For the role of conditionality in the context of EU funds: Bieber/Maiani, Enforcement 1076 ff; see also Harlow/Rawlings, Process 47; for conditionality in the context of compliance with EU law more generally see Andersen, Enforcement 181 ff; Ioannidis, Members 489.

<sup>579</sup> Monitoring tasks are, pursuant to Article 13 para 7 of the ESM-Treaty, exercised by the Commission, the ECB, but also by the IMF as 'technical assistance'; see eg IMF, Country Report No 14/59 (February 2014) 2 <http://www.imf.org/external/pubs/ft/ scr/2014/cr1459.pdf> accessed 28 March 2023.

from the EFSM/EFSF/ESM, when exploring the 'coerciveness' of the mech‐ anism the legal quality of the respective MoU seems to be of secondary importance.<sup>580</sup> Since compliance with the conditions laid down in the MoU by the beneficiary MS is – according to the respective financial assistance facility agreement which refers to the MoU<sup>581</sup> – mandatory for the grant of financial assistance,<sup>582</sup> there is a strong incentive to comply.<sup>583</sup> Where the beneficiary MS do not comply, they face the severe (but only) consequence of being refused (further) financial assistance.<sup>584</sup> The MS concerned have to weigh – for themselves – the advantages and costs of compliance and non-compliance, respectively.<sup>585</sup> Mere incentives to comply do not form a means of 'legal enforcement', but speak in favour of soft law. On the other hand, the benefits at issue here are not a prize or a subsidy anybody who meets certain predefined requirements can apply for. Rather, they constitute an aid which is calculated and granted upon an individual request.

The conditionality is not set for one specific point in time, but policy objectives are laid down or agreed upon in the MoU which should be reached in the course of an extended period of time. Both the examination of the achievements and, upon a positive result of this examination, the payment of the aid (tranches) are sequenced over a couple of years. These settings cater for a strong synallagmatic relationship between the umbrella at issue and the respective beneficiary. It makes the umbrella's power to refuse payment of the next tranche come very close to a power to punish

<sup>580</sup> See also Beckers, Juridification 576; Knauff, Regelungsverbund 278 f and 339; for the granting of loans by the World Bank see ibid 280 f; similarly: Repasi, Protection 1124.

<sup>581</sup> For the legal technique of referencing see, in the context of private regulation, II.2.2.3. above.

<sup>582</sup> For the ECB's reference to this conditionality in its monetary policy as a further incentive for compliance see eg <http://www.ecb.europa.eu/press/pr/date/2012/htm l/pr120906\_1.en.html> accessed 28 March 2023.

<sup>583</sup> With regard to financial incentives as a means of ensuring compliance see H Adam, Mitteilungen 124 f; Gil Ibáñez, Supervision 288; Knauff, Regelungsverbund 339; with regard to incentives as alternatives to enforcement traditionally understood: Commission, Draft Interinstitutional Agreement on the operating framework for the European regulatory agencies, COM(2005) 59 final, para 7.1.

<sup>584</sup> For similar scenarios in the context of public law agreements in national law: Bauer/Kretschmer, Dogmatik 254.

<sup>585</sup> Addressing the same dynamics in the context of EU neighbourhood policy: Vianel‐ lo, Approach 554 f.

non-compliance.586 The fact that an MoU ought to be consistent with (the pertinent) EU law<sup>587</sup> is a neutral requirement (it neither speaks in favour of nor against its qualification as a soft law instrument) since a legal order, in our case: the EFSF/the ESM, may require consistency with alien rules even for soft law. Neither does the interpretative authority of the CJEU point in either direction.<sup>588</sup> The neutral wording of the MoU *in dubio*<sup>589</sup> speaks in favour of their qualification as law and the particularly strict mode in which the 'incentives' are applied appears to underpin this qualification, both in case of the EFSM (where the MoU qualifies as EU measure) and in the cases of the EFSF and the ESM (where the MoU belongs to public international law).<sup>590</sup>

The ambivalence of the legal quality of MoU is 'made use of ' by different actors in different ways: The IMF perceives its Memoranda of Economic


<sup>586</sup> See Wirth, Assistance 227 f, who accords 'loan covenants' – conditionality contained in loan agreements of the World Bank – a legal status 'similar to that of treaties'. He points to their enforceability, emphasising that 'the Bank could suspend further disbursements, which are customarily made in phases or "tranches"'. Whether en‐ forcement happens in practice is a different story; see ibid 228 f; for the preferability, in *Wirth's* view, of the terms 'mandatory' and 'enforceability' (as compared to 'bind‐ ing') in this context, see ibid 231.

<sup>587</sup> Recital 2 of the EFSF Framework Agreement between the Euro-MS and the EFSF <https://www.sv.uio.no/arena/english/people/guest-researchers/agustinm/crisis-d ocuments-2012/14-efsf-frameworkagreement-consolidated-8sep11.pdf> accessed 28 March 2023; Article 13 para 3 subpara 2 of the ESM-Treaty (which arguably limits its call for compliance to a mere 'consisten[cy] with the measures of economic poli‐ cy coordination provided for in the TFEU, in particular with any act of European Union law, including any opinion, warning, recommendation or decision addressed to the ESM Member concerned'); for similar compliance requirements (eg with international environmental agreements) for projects financially supported by the World Bank see Wirth, Assistance 232.

<sup>588</sup> See Article 38 para 3 of the ESM-Treaty in conjunction with para 2, according to which the CJEU shall decide on any dispute between members of the ESM or be‐ tween them and the ESM 'in connection with the interpretation and application of this Treaty, including any dispute about the compatibility of the decisions adopted by the ESM with this Treaty'. First, it is not clear whether the MoU fall within the CJEU's scope ('decisions adopted by the ESM'; see below). But even if they do, this does not necessarily mean that they are hard law. After all, the Court, in particular according to the preliminary reference procedure, may authoritatively interpret/examine legally non-binding acts (of EU law), as well; see 6.3. below.

and Financial Policies and its Technical Memoranda of Understanding, which – in the given context – are closely linked to the respective MoU concluded under the EFSM, the EFSF or the ESM,<sup>591</sup> as non-binding in‐ struments<sup>592</sup> – a view which eg the Irish Supreme Court seems to support. On the contrary, the Latvian or the Portuguese Constitutional Court, in the perspective of their respective constitution, qualify them as legal obliga‐ tions.<sup>593</sup> The Greek Plenary Assembly, to take another example, did not qualify the first MoU with Greece as an international convention according to the Greek Constitution and therefore denied the necessity of its ratifica‐ tion by a law.<sup>594</sup>

While it was said above that the (lack of ) legal quality of the MoU – due to the strict compliance mechanism attached – does not affect its 'coer‐ civeness' and, against this background, is of secondary importance,<sup>595</sup> with regard to the available remedies the question of legal bindingness is highly important. The ESM-Treaty, as special agreement according to Article 273 TFEU, empowers the CJEU to decide – after the Board of Governors – on 'any dispute arising between an ESM Member and the ESM, or between ESM Members, in connection with the interpretation and application of this Treaty, including any dispute about the compatibility of the decisions adopted by the ESM with [the ESM-Treaty]'.596 While it is not clear wheth‐ er the term 'decision' as used in the ESM-Treaty only encompasses legally binding acts or also eg the guidelines addressed in Article 22 para 1 of the ESM-Treaty,597 in view of the above-mentioned presumption of legal force

<sup>591</sup> For the relationship between the ESM-MoU and the IMF-MoU see Mönning, Staatensanierungsverwaltungsrecht 206.

<sup>592</sup> See Poulou, Grundrechte 147 f.

<sup>593</sup> See O'Donovan, Way 52 f; Fischer-Lescano, Austeritätspolitik 35.

<sup>594</sup> See Tsakiri, Protection 18; for further (different) qualifications see Poulou, Grund‐ rechte 147 f; for similar motives in the case of the USA and the Paris Agreement 2015 see Bodansky, Character 149 f.

<sup>595</sup> For this argument see also, in different contexts, Haas, Hypotheses 23; Wengler, Rechtsvertrag 195.

<sup>596</sup> Article 37 paras 2 f of the ESM-Treaty.

<sup>597</sup> We cannot – without further consideration – take over the terminology of Article 288 TFEU here, because the ESM-Treaty contains no reference or only a mere hint to that effect. In fact, the preceding passage in Article 37 para 2 of the ESM-Treaty – 'any dispute arising between an ESM Member and the ESM' (on the interpretation and application of the ESM-Treaty) – rather suggests a broad understanding of the Court's (and the Board of Governors', respectively) powers.

the MoU can be considered a 'decision adopted by the ESM'.598 That means that Euro-MS can request a decision of the Court (after a decision of the Board of Governors) on the interpretation/application of certain provisions of an MoU. Individuals or undertakings are not entitled to apply to the Court under the ESM-Treaty. Since the MoU does not qualify as EU law, it may not be reviewed by the Court under Article 267 TFEU<sup>599</sup> or Article 263 (para 4) TFEU, either. But even if it did, the latter route would be barred due to the fact that individuals and undertakings are regularly not directly (and individually) concerned by the MoU, but – if at all – only by the national reform measures adopted to implement it. They may be granted standing before the Court on the basis of an action for damages against the EU (non-contractual liability), though. This is because no direct (and individual) concern is required under these proceedings. Rather, and apart from the other requirements, it is sufficient that a *Schutznorm* – a norm protecting the interests of individuals600 – has been violated by the EU. This would be the case, for example, where the Commission – in negotiating the MoU – violates a right enshrined in the CFR.<sup>601</sup>

Summing up, while the MoU concluded under the EFSM qualify as EU law,602 the MoU concluded under the EFSF/ESM are officially agreed upon

<sup>598</sup> More concretely: by the Commission for the ESM, with the approval of the Board of Governors.

<sup>599</sup> See Tuominen, Mechanisms 102 f, with references to the Court's case law.

<sup>600</sup> See eg case C-152/88 *Sofrimport*, para 26; for the term *Schutznorm* and many references to the Court's case law see Steiner, Haftung 162 ff.

<sup>601</sup> The Commission, according to Article 17 para 1 TEU and according to Articles 13 paras 3 f of the ESM-Treaty, has to comply with Union law even when acting outside the EU legal order; see joined cases C-8–10/15P *Ledra*, para 67; joined cases C-597/18 P, C-598/18 P, C-603/18 P and C-604/18 P *Chrysostomides*, para 132. While the Court has not addressed this issue here, when it comes to the required causal link between the damage and the EU action concerned arguably the legal bindingness of the measure at issue is only of subordinate importance. The *de facto* 'coerciveness' of the measure – in our case: the MoU – seems to be sufficient because it strongly pushes for its actual implementation by the beneficiary MS, and that means: the occurrence of the damage. This assumption is supported by the fact that not only legal, but also factual measures may evoke a claim for damages under Article 340 para 2 TFEU; see Gellermann, Art. 340 AEUV, para 16; Jacob/Kottmann, Art. 340 AEUV, para 73; Ruffert, Art. 340 AEUV, para 27, each with references to the Court's case law.

<sup>602</sup> For the reviewability of MoU concluded under the EFSM and alternative EU pro‐ grammes under the preliminary reference procedure see case C‑258/14 *Florescu*. In this case, with regard to an MoU adopted in the context of the Article 143 TFEU-procedure, the Court remained ambivalent. It held that it be 'mandatory',

by the EFSF/ESM and the beneficiary MS (its central bank and possibly financial funds, respectively) and hence are instruments of public interna‐ tional law. From an EU (EFSM) or public international law (EFSF/ESM) perspective, the MoU at least *in dubio* is to be qualified as law. In a national (constitutional) view the qualification may be more clear, though different from constitution to constitution. The fact that the beneficiary State and the other participating States of the EFSF/ESM (not: the EFSF/ESM itself ) are all members of the EU, the call for consistency with pertinent EU law, and the involvement of EU institutions in the negotiation of the MoU, causes 'strong links'<sup>603</sup> of these acts of public international law to EU law,<sup>604</sup> having lead to the term '*Unionsersatzrecht*' for the EFSF and ESM founding acts, and for the measures based upon them.

The purpose of this incursus was to exemplify the challenges which may emerge when we are asked to assign an act to EU law or a different legal order, and when it comes to distinguishing soft law from law in practice. Moreover, it was shown that where there is a strong dependence of one partner (which is called upon to meet certain conditions) to an agreement on the 'delivery' by the other partner, for example where the provision of fi‐ nancial benefits to prevent a State's failure in exchange for national reforms is at issue, the question of whether or not the conditionality provisions are legally binding is sidelined.

#### 2.3. Addressees

The addressees of EU soft law so far have been touched upon here and there. Now they shall be looked at more systematically. The term 'addressee' in this context designates the actor (including natural or legal persons as well as entities without legal personality) the behaviour of which shall be steered by a soft law act. In essence, four groups of addressees can be

but relativised its effects in the same sentence (para 41); for the ambiguity of this particular statement of the Court see Dermine/Markakis, Bailouts 657 f.

<sup>603</sup> Opinion of AG *Wahl* in joined cases C-8–10/15P *Ledra*, para 51, with respect to the ESM.

<sup>604</sup> For the fact that mutual cross-referencing of EU law and public international law as such cannot lead to an incorporation of an act of public international law into EU law resulting in the CJEU's jurisdiction, see case C‑366/10 *Air Transport Association*, para 63, with a further reference; see also Grundmann, Inter-Instrumental-Interpre‐ tation 926–928.

distinguished: the originator-internal addressees, the internal addressees, the external addressees, and the EU-external addressees. While categorising different groups of addressees is worthwhile for systematic reasons, in selected cases the determination of the 'intended or apparent' addressees in practice may turn out to be difficult or even impossible.<sup>605</sup>

Originator-internal (or intra-institutional) addressees are those which form part of the creator of an EU soft law act, ie the staff. An example for a soft law act directed to originator-internal addressees is the Antitrust Manual of Procedures adopted by the Commission's Directorate-General (DG) Competition which clearly sets out that it is 'an internal working tool intended to give practical guidance to staff on how to conduct an investigation applying Articles 101 and 102 TFEU' which 'does not contain binding instructions for staff, and the procedures set out in it may have to be adapted to the circumstances of the case at hand'.606 That also other persons – in this case eg competition lawyers who want to know in detail how the Commission performs its investigations – may find these acts useful as a piece of information and may adapt their own or their clients' actions accordingly, ie that the acts may also have an external effect,607 does not alter the fact that those addressed and hence those (softly) obliged by it in principle are originator-internal persons.608 A publication of such acts to a broader audience, in particular online, may, however, indicate that also an

<sup>605</sup> See Hofmann/Rowe/Türk, Administrative Law 571.

<sup>606</sup> Commission, Antitrust Manual of Procedures. Internal DG Competition working documents on procedures for the application of Articles 101 and 102 TFEU (Novem‐ ber 2019) <https://op.europa.eu/en/publication-detail/-/publication/d7d7e463-ac51 -11ea-bb7a-01aa75ed71a1> accessed 28 March 2023.

<sup>607</sup> See Senden, Soft Law 315 f, with further references; for the necessity of such intrainstitutional soft law see Kovács/Tóth/Forgác, Effects 61; addressing the tension between formal addressee and '*de facto* addressee' with regard to soft law and other acts adopted in the context of EU foreign and neighbourhood policy: Vianello, Approach 553 f.

<sup>608</sup> Stressing this double effect of 'internal' soft law in the context of State aid policy: Aldestam, Soft Law 15, with a further reference; see also Hofmann/Rowe/Türk, Ad‐ ministrative Law 567, who stress the 'indirect external effect' Commission guidelines may have 'through the application of the principles of legal certainty, the protection of legitimate expectations and equal treatment'; doubtfully: Cannizzaro/Rebasti, Soft law 229 f; for a 'measure of internal organization' which nevertheless has 'legal effects vis-à-vis third parties' see case C-58/94 *Netherlands v Council*, para 38, with further references.

(EU-)internal/external steering effect is intended by the originator (here: the Commission).<sup>609</sup>

Internal addressees shall encompass institutions, bodies, offices or agen‐ cies of the EU other than the originator (more precisely: its staff). They are at issue eg in case of an opinion of the Economic and Social Commit‐ tee which is sent to the Commission, or of an opinion of the European Ombudsman sent to a European agency (against which an EU citizen has launched a complaint). This category of acts could be said to belong to the group of inter-institutional soft law acts in a wide sense.<sup>610</sup> Inter-institu‐ tional acts in a narrow sense are generally understood as acts agreed upon by two or more institutions (or bodies, offices, or agencies of the EU), ie on a bi- or multilateral level.<sup>611</sup> These agreements, if not intended to be binding, form soft law with internal addressees (namely: the creators of the agreement).<sup>612</sup>

External addressees are persons or entities affiliated with the EU which are not, at the same time, internal addressees; in particular: MS, MS authorities,613 EU citizens and legal persons seated in the EU (but not belonging to its administration; eg undertakings). Examples for EU soft law acts directed to external addressees are the recommendation of the Eu‐ ropean Banking Authority (EBA) which is sent to the banking supervisory

<sup>609</sup> See case T-496/11 *United Kingdom v European Central Bank*, paras 33 f with regard to a publication 'outside the author itself '; for a communication which has not left 'the internal sphere of th[e] administration' see case C‑619/19 *Baden-Württemberg*, para 52. The Commission's Rules of Procedure in their Article 17 para 4 provide for the publication of Commission acts in the OJ. This general rule may have to be teleologically 'reduced' in places, eg for the sake of data protection; with regard to the reasoned opinion addressed to a MS at the outset of a Treaty infringement procedure see Senden, Soft Law 189.

<sup>610</sup> See Hummer, Interorganvereinbarungen 91 ff, with further references.

<sup>611</sup> See also Stöbener de Mora, Institutionelles; with regard to agreed-upon soft law see Knauff, Regelungsverbund 373–376; see also case C-25/94 *Commission v Council*, para 49.

<sup>612</sup> See Article 295 TFEU, according to which inter-institutional agreements between the EP, the Council and the Commission can be concluded 'which may be of a binding nature'. This means that they may as well be (agreed to be) of a non-binding nature; for the practical effects inter-institutional agreements may have on the competences of the institutions involved see Klamert, Pragmatik 148 f.

<sup>613</sup> For the adoption of acts *vis-à-vis* national authorities, not 'the MS' see Schütze, Rome 1418 f; the different conceptions of the term 'Member State' in EU law are exemplified if we compare Article 258 TFEU (broad conception) and Article 263 para 2 TFEU (narrow conception); with regard to the latter provision see also V.3.6. below.

authority in a certain MS or the Commission's so-called *de minimis* notice (designated as 'Communication')<sup>614</sup> in the field of competition law which does not mention any specific addressees, but *de facto* is first and foremost directed to undertakings within the meaning of Article 101 TFEU and to national competition authorities.<sup>615</sup>

EU-external addressees are actors which do not belong to the EU, eg third countries or international organisations. An example for an act ad‐ dressed to them is the Memorandum of Understanding on reinforcing the EU-China IP Dialogue Mechanism concluded between the EU and China.<sup>616</sup>

2.4. Legally non-binding acts other than soft law

The output of bodies vested with public authority is not bound to be either law or soft law. There is a number of acts (normally) without regulatory content, that is to say without containing a command or with no linkage to a command. They are addressed to actors both within and outside the respective administration: press releases, circulars, surveys, scientific information, certain letters and e-mails, etc.617 More generally, this issue was addressed under II.2.3. above. Here we shall broach it specifically in the context of the EU. The distinction between EU soft law and other legally non-binding acts is relevant because these two categories of acts are different both in terms of requirements and effects. The requirements for a legal basis are more demanding in case of soft law than in case of other legally non-binding acts. Moreover, soft law requires an enhanced degree of determination for its (linkage to) commands so as to give concrete

<sup>614</sup> For the lack of any difference between 'communications' and 'notices' see Senden, Soft Law 142 f. Also the principal difference between 'notices' and 'guidelines', if any, is by far not clear: <https://competition-policy.ec.europa.eu/mergers/legislation/no tices-and-guidelines\_en> accessed 28 March 2023.

<sup>615</sup> Commission, Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1) of the Treaty on the Functioning of the European Union (De Minimis Notice), 2014/C 291/01; see also the Opinion of AG *van Gerven* in case C-234/89 *Delimitis*, para 22 (with regard to the predecessor notice): 'the individuals for whom it is intended'.

<sup>616 &</sup>lt;http://europa.eu/rapid/press-release\_IP-15-5279\_en.htm> accessed 28 March 2023.

<sup>617</sup> See also examples addressed by van Schagen, Regulation; for the potential ambiva‐ lence of a press release of the Commission and the UEFA see case C-117/91 *Bosman*, para 14.

guidance for the addressee's behaviour, whereas in other acts also mere objectives or vague wishes may be uttered. They may as well contain mere information. In terms of effects, it is to be stressed that only soft law contains rules vested with public authority. This will result in more specific steering effects on its addressees than with other non-binding acts.

For other legally non-binding acts it is mostly its content which may (not) be convincing and thus exert (only limited) steering effects. The topic to which the content of the act is dedicated is irrelevant for the categorisation. The scientific finding that a certain substance used in con‐ venient meals significantly enhances the risk of cancer for their consumers may be highly important and demand immediate action on the part of the food safety authorities in charge, but it does not qualify as soft law. On the other hand, an institution-internal guideline on how to use e-mail signatures for professional correspondence may be of limited importance, but it is normative and hence – depending on its legal (non-)bindingness – qualifies as law or soft law. The distinction between soft law and other legally non-binding acts is also reflected upon, to some extent at least, in the EU's publications regime. The EU itself on its EUR-Lex webpage up until recently has applied a three-partite classification of 'EU legislation': 'bind‐ ing legal instruments', 'non-binding instruments',618 'other instruments'. Meanwhile, it distinguishes only between 'legally binding' and 'non-bind‐ ing' 'legal acts'.<sup>619</sup>

For most of the 'other' output it is, for lack of a command, difficult to confound it with soft law, let alone law. In a few cases a doubt may remain, though.620 Due to the relative informality of EU soft law it may be difficult sometimes to distinguish it from mere utterances of opinions. The latter opinions only deserve closer scrutiny as potential soft law where

<sup>618</sup> Note that Article 297 TFEU does not lay down publicity requirements for EU soft law; Knauff, Regelungsverbund 423 f; see also Opinion of AG *La Pergola* in case C-4/96 *NIFPO*, para 56.

<sup>619 &</sup>lt;https://eur-lex.europa.eu/collection/eu-law/legislation/recent.html>; see also the EU's inter-institutional style guide, containing provisions on the structure of the OJ: <http://publications.europa.eu/code/en/en-000500.htm> both accessed 28 March 2023.

<sup>620</sup> See eg the DG Competition Staff Working Document, 'The Application of State Aid Rules to Government Guarantee Schemes covering Bank Debt to be issued after 30 June 2010' which appears to be merely summarising the Commission's new ap‐ proach (as laid down in particular in the Commission's relevant Communications), but which also, eg on its page 7, lays down new rules (qualifying as soft law); see also Opinion of AG *Tesauro* in case C-57/95 *France v Commission*, para 17.

they stem from an actor which is vested by EU law with public authority (as reflected eg in an explicit competence to adopt EU soft law). Where no such authority is provided for, eg in case of an NGO proposing a certain policy approach to the Commission,<sup>621</sup> the question of whether or not this could be EU soft law does not even occur. But also where actors abstractly competent to adopt EU soft law, mainly these are EU institutions, bodies, offices and agencies (see 2.2. above), express their views, these utterances do not necessarily qualify as EU soft law. Even though public bodies, and hence also EU bodies (in a broader sense), are not protected by the freedom of expression as a fundamental right,<sup>622</sup> they may principally express their respective view to each other in a field materially falling within their respective scope of action. This follows from Article 13 para 2 TEU (second sentence: 'The institutions shall practice mutual sincere cooperation'), and is explicitly laid down with regard to the relationship between the EP, the Council and the Commission in Article 295 TFEU which provides, explicitly only since the Treaty of Lisbon,623 that these three institutions 'shall consult each other and […] make arrangements for their cooperation'.624 To assume that all utterances made in the context of such cooperation/consultation need to qualify as soft law would mean to deprive soft law of its specific character. Institutions and bodies of the EU in places need to communicate with each other in an atmosphere of informality, that is to say in expressions other than legal and soft law acts. The provisions mentioned above neither require that each expression in the course of such cooperation/consultation shall be uttered in the form of soft law, nor do they grant a general competence to adopt soft law to the EU bodies at issue.

The ordinary legislative procedure may serve to illustrate the distinction between soft law and other non-binding acts. The EP's position at first reading is an act which is addressed ('communicated') to the Council. It constitutes the final position of this institution at the time of its being launched, and it may (possibly) be accepted by the Council the way it is

<sup>621</sup> See in this context Article 11 para 1 TEU, providing that the institutions shall allow 'citizens and representative associations' to 'make known and publicly exchange their views in all areas of Union action'.

<sup>622</sup> See eg S Augsberg, Art. 11 GRC, para 10; Jarass, Charta, para 19 of Art 11.

<sup>623</sup> See Klamert, Pragmatik 148.

<sup>624</sup> Article 295 TFEU is connected to Article 13 para 2 (second sentence) TEU, but goes beyond that; see Gellermann, Art. 295 AEUV, para 1*;* Kluth, Art. 295 AEUV, para 1; Voet van Vormizeele, Art. 295 AEUV, para 2.

(Article 294 para 4 TFEU). It is the expression of the wish of the EP that the Council agrees to this position and thereby makes its content a legal act. It is one act in a procedure (potentially) containing a number of different acts which are all dedicated to creating a legally binding (more concretely: a legislative) act. In that sense, it does not stand alone but is regularly part of a line of different acts which – in a macro-perspective – serve the same aim, that is to create a legislative act. In a micro-perspective, however, it serves a more specific aim, namely to convince the Council to elevate, by the approval of this institution, the draft norms contained therein to a legislative act. Therefore, in my view and irrespective of its being a special case *qua* being embedded in a whole procedure intended to lead to the adoption of a legislative act, it qualifies as a soft law act. That the Council may as well (partly) disagree with the EP and adopt its own position is not in contradiction to this qualification. Rather, it is a ramification of its legal non-bindingness. While being legally non-binding for others, the EP's position does entail legal effects: The Council cannot proceed without having received the EP's position; furthermore, Article 294 TFEU implicitly requires the Council to consider the EP's position (*argumentum* 'approves'/'not approve'), a duty which addressees of EU soft law often have (see 4.2. below); the EP is bound625 by its position to the extent that the Council can, without further ado (in particular: without asking the EP for its view once more), accept its position, thereby making it a legislative act. Once the Council has approved the EP's position or has adopted its own (different) position, the EP's position – that follows from the system set up by Article 294 TFEU and hence from primary law – ceases to contain a demand, as then it is procedurally impossible to be followed. This is because it has already been followed by the Council or because the Council has decided otherwise, as the case may be. This constitutes a procedural restriction of the soft law effects of the EP's position. For similar reasons, also the EP's view at second reading and the Council's position at first and its view at second reading, respectively, qualify as soft law. The informal discussions, negotiations and other exchanges of views between representatives of the EP and the Council, but also the Commission, taking place in between these more formal steps the TFEU provides, for lack of the above characteristics do not entail soft law acts.

<sup>625</sup> For the bindingness of EU soft law upon its respective creator (self-bindingness) see 4.2.2.2.4. and 4.2.3.2.3. below.

The fact that an act merely serves the purpose of initiating or continuing a process and loses its importance, or untechnically speaking: 'evaporates', once this purpose has been fulfilled,<sup>626</sup> does not *per se* speak against soft law.<sup>627</sup> This is why also the Commission proposal according to Article 293 TFEU may be qualified as an act of soft law. It is not the rules proposed therein which have a (soft) normative effect (because, if eventually adopted in the form of law, they will have different addressees), but the suggested adoption of the proposal by its addressee(s) as a legislative act. It constitutes the Commission's view on what a legislative act regulating (parts of ) a certain policy field shall look like, and it does so with a certain finality (even if the Commission may again alter or withdraw the proposal).<sup>628</sup> The legal effects the proposal entails are: In most cases the legislator may not act without such a proposal; where the Council, pursuant to the Treaties, acts on such a proposal, it may normally amend it only by a unanimous decision629; it can be approved by the legislator and thereby, without further ado (see above), be transformed into a legislative act. Once the legislative act at issue is adopted, the Commission proposal no longer has this effect.<sup>630</sup>

<sup>626</sup> See Schoo, Art. 294 AEUV*,* paras 38 f.

<sup>627</sup> Addressing *travaux préparatoires* as a special case: Rosas, Soft Law 309: 'These can be seen as tools of interpretation which are so directly linked to the adoption of legally binding texts that it seems best to award them separate attention'.

<sup>628</sup> For the underlying purpose of this competence see Krajewski/Rösslein, Art. 293 AEUV, paras 15–17.

<sup>629</sup> Article 293 para 1 TFEU. Note the limits to this power: The Council may not 'depart from the subject matter of the proposal [or] alter its objective', as this would 'deprive [the Commission proposal] of its *raison d'être*' (emphasis in original); case C‑24/20 *Commission v Council*, paras 93 f, with further references.

<sup>630</sup> EUR-Lex even states that in this case its validity has ended; see eg <https://eur-le x.europa.eu/legal-content/EN/ALL/?uri=celex:52012PC0064> accessed 28 March 2023; for the non-binding acts adopted in preparation of a legislative proposal see van Schagen, Regulation 597.

# *3. The legal bases of soft law*

3.1. Preliminary remarks: the meaning of Article 288 TFEU for EU soft law

# 3.1.1. The difference between recommendations and opinions

The list of legal acts laid down in Article 288 TFEU mentions two legal acts<sup>631</sup> with 'no binding force'<sup>632</sup>: recommendations and opinions (whose scope of addressees is not limited in this provision in any way). The fact that their non-bindingness is stressed allows us to conclude that both of them may have normative content, meaning that they may contain or be related to a command. Where an act does not have any normative content, the question of its legal (non-)bindingness does not arise in the first place. That Article 288 TFEU lists two categories of non-binding acts suggests that they are different from each other.633 But in which way do these two types of acts differ from each other?634 The (legally non-binding) normativity of a recommendation – in an exclusively semantic view – appears to be stronger than that of a mere opinion.635 Etymologically, there is a nexus between the term 'command' and the term 'recommendation'.636 Recom‐ mendations often (not always) dispose of general application.637 An opinion

<sup>631</sup> This qualification as a legal act in the terminology of the Treaties is in contrast to the national legal orders of most MS and to the international legal order, respectively; *von* Bogdandy/Bast/Arndt*,* Handlungsformen 114. However, from the terminological elevation of soft law to a source of law as such no specific legal effects can be deduced.

<sup>632</sup> It is clear that only the legal bindingness is addressed here; other legal effects are not thereby excluded; see Knauff, Regelungsverbund 301. Arguing that a 'comply or explain' mechanism excludes the qualification of the act as recommendation or opinion within the meaning of Article 288 TFEU: Wörner, Verhaltenssteuerungsfor‐ men 228.

<sup>633</sup> At the level of international law, no such dual distinction between a 'recommenda‐ tion' and an 'opinion' seems to apply; see Schermers/Blokker, Institutional Law, § 1217, with regard to further designations such as 'advice' or 'resolution'.

<sup>634</sup> Apparently suggesting that there is no difference in substance: Opinion of AG *Te‐ sauro* in case C-303/90 *France v Commission*, para 20: '[…] the measure in question is no more than a recommendation, an opinion addressed to Member States […]'.

<sup>635</sup> Recommendations, it was said, allow their creators 'to make their views known and to suggest a line of action'; <https://european-union.europa.eu/institutions-law-bu dget/law/types-legislation\_en> 28 March 2023; see also Braams, Koordinierung 152; Schwarze, Soft Law 235.

<sup>636</sup> Walter, Soft Law 25.

<sup>637</sup> Cannizzaro/Rebasti, Soft law 221.

is uttered for its addressee to take notice of it – it allows its creator 'to make a statement'.<sup>638</sup> It was remarked that the adoption of a recommenda‐ tion reflects the respective body's own 'Entschlusskraft' [determination],<sup>639</sup> whereas opinions constitute a reaction to another body's initiative,<sup>640</sup> that is another bodies utterance. In practice this holds true for most opinions. However, concluding from the wording of certain competence clauses<sup>641</sup> it cannot be excluded that an opinion is submitted in complete detachment of other acts.

*Von Bogdandy*, *Arndt* and *Bast* conceptualise recommendations as '*nonbinding directives*' (emphasis in original), which may (voluntarily) be trans‐ posed by the MS.<sup>642</sup> With regard to Commission recommendations and opinions *Hofmann*, *Rowe* and *Türk* express that they 'assist [their respective addressee] to evaluate a situation or circumstance and to take appropriate action', but generally they also describe (idealtypical) recommendations as 'active' (initiating) and opinions as 'reactive' (responding).<sup>643</sup> *Senden* contends (with a view to administrative practice) that a recommendation 'is primarily used as a tool or instrument to coordinate or to bring national policies and objectives closer together, without proceeding (yet) to the legislative harmonisation level'.644 Due to this (soft) regulatory character also with regard to their outward appearance they often resemble acts of secondary law.645 It was also said that with recommendations EU bodies

<sup>638 &</sup>lt;https://european-union.europa.eu/institutions-law-budget/law/types-legislation \_en> 28 March 2023.

<sup>639</sup> See Geismann, Art. 288 AEUV, para 63; Nettesheim, Art. 288 AEUV, para 201; Ruffert, Art. 288 AEUV, para 98. *Virally*, with regard to recommendations adopted on the level of public international law, similarly expresses that they imply 'une invitation à adopter un comportement déterminé, action ou abstention'; Virally, Valeur 68.

<sup>640</sup> See Bieber/Epiney/Haag/Kotzur, Europäische Union 211; Härtel, Rechtsetzung 272, with further references; Nettesheim, Art. 288 AEUV, para 201.

<sup>641</sup> See, as one example, Article 304 para 1 TFEU, stipulating the Economic and Social Committee's competence to 'issue an opinion on its own initiative *in cases in which it considers such action appropriate*' (emphasis added).

<sup>642</sup> Von Bogdandy/Arndt/Bast, Instruments 112 f. For the fact that directives may con‐ tain broad provisions which have an effect similar to soft law see Trubek/Trubek, Governance 551, emphasising the coexistance in one directive (here: the EU Water Framework Directive, ie Directive 2000/60/EC) of broad guidance and detailed, binding rules ('hybridity').

<sup>643</sup> Hofmann/Rowe/Türk, Administrative Law 545 f.

<sup>644</sup> Senden, Soft Law 179.

<sup>645</sup> See Knauff, Regelungsverbund 302.

provide their solution to a specific problem and suggest its application,<sup>646</sup> whereas an opinion primarily contains the legal view of the creator, but does not suggest a specific action of the addressee.<sup>647</sup> The author would suppose that also the transmission of the legal view of an EU body, eg the Commission under Article 258 TFEU, may suggest compliance with this view by the addressee, ie a specific action.<sup>648</sup> Opinions are normally not used as alternatives to secondary law,<sup>649</sup> and are generally less homogenous as regards appearance and content than recommendations. However, selec‐ ted examples can be named in which very similar acts adopted by two different bodies are called 'recommendation' in one case, and 'opinion' in the other.<sup>650</sup>

A systematic interpretation reveals a tendency that recommendations (eg of the Commission), if addressed to another institution (eg the Council), are regularly envisaged as a procedural requirement for further (soft law or legal) action.651 The addressee is free not to act, that is to say not to follow the recommendation, though. In these cases recommendations meet a purpose which is comparable to that of proposals. On the contrary, opinions are often asked for ('invited') by their potential recipients. The actor asked is then free to adopt an opinion. In other cases EU bodies are free to adopt opinions on their own initiative.652 The systematic approach applied here is only rudimentary. Having assessed the legal bases for the adoption of recommendations and opinions contained in the Treaties (see 3.4. and 3.5. below), we shall revisit the distinction between recommenda‐ tions and opinions and examine whether new findings have arisen from this assessment (see 3.9. below).

<sup>646</sup> See also case C‑370/07 *Commission v Council*, para 42; case T-496/11 *United King‐ dom v European Central Bank*, para 32.

<sup>647</sup> See Knauff, Regelungsverbund 302 and 304.

<sup>648</sup> See also Knauff, Regelungsverbund 304 f. Opinions are uttered in order to be heard: 'Jede Meinung hat Anspruch, entweder mit Schweigen aufgenommen oder wirksam widerlegt zu werden' (*Franz von Holtzendorff*); unspecifically referring to the differ‐ ence of a Commission opinion under Article 258 TFEU as compared to other Commission opinions: Hofmann/Rowe/Türk, Administrative Law 548.

<sup>649</sup> See Senden, Soft Law 188.

<sup>650</sup> See the example of the ESMA and the Commission given at 3.4.6. below.

<sup>651</sup> Eg Article 121 paras 2 and 4, Article 126 paras 7 and 13, Article 144 paras 2 f or Article 207 para 3 TFEU.

<sup>652</sup> Eg Article 282 para 5, Article 304 para 1, Article 307 paras 1 and 4 TFEU.

# 3.1.2. Is there a *numerus clausus* of EU soft law acts?

The difference between EU law and EU soft law – like in the case of law and soft law more generally – is established by scrutinising the act and its context (see II.2.1.2. above).<sup>653</sup> But also within the realm of EU soft law different acts can be perceived.<sup>654</sup> In practice EU soft law acts do not only occur under the titles 'recommendation' and 'opinion' respectively, but they bear many different names – an 'unsystematic, indeed unpredicta‐ ble, nomenclature'<sup>655</sup> – such as 'Communication', 'Resolution', 'Guidelines', 'Questions and answers' ('Q&A'),<sup>656</sup> 'Vademecum' or 'Standards'.<sup>657</sup> This does not *per se* contradict the assumption that Article 288 TFEU contains an exhaustive list, a (very short) catalogue of EU soft law acts<sup>658</sup> (see also 3.6. below); neither does the *prima facie* more differentiated terminology applied elsewhere in the Treaties, in particular 'conclusions' (eg in Article 135 or Article 148 TFEU) or 'guidelines' (eg in Article 121659, Article 148 or

<sup>653</sup> See Opinion of AG *Kokott* in case C-226/11 *Expedia*, paras 26 ff, with reference to the three indicators 'wording', 'purpose' and 'context'; see also Hofmann/Rowe/ Türk, Administrative Law 552, with further references.

<sup>654</sup> For the diversification of the legally non-binding output of public actors on a global level ('global governance') see Goldmann, Perspective 61.

<sup>655</sup> Hofmann/Rowe/Türk, Administrative Law 537; for the many (partly: soft law) acts adopted as part of EU external action under different names see Vianello, Approach 551.

<sup>656</sup> Sometimes also referred to as 'Frequently Asked Questions'; see also Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 10 f.

<sup>657</sup> See, for further examples, Majone, Agencies 269; van Rijsbergen/Rogge, Changes, with regard to the ESAs; see also von Bogdandy/Arndt/Bast, Instruments 114, assu‐ ming that the Commission is 'eager to reserve opinions [as opposed to the other acts just mentioned] for specific and, probably, important measures'.

<sup>658</sup> Refusing the lament of those claiming there to be a '"proliferation of instruments" [Laeken Declaration on the future of the European Union (2001) 4] […], implying an uncontrolled and dangerous multiplication of instruments' also in the field of legally binding instruments (after the Treaty of Nice and in the run-up to the draft Constitutional Treaty): von Bogdandy/Arndt/Bast, Instruments 91f: 'The structure of the legal instruments is complex and only partially determined by the Treaties, but it is not chaotic'.

<sup>659</sup> Article 121 para 2 TFEU mentions the 'broad guidelines of the economic policies of the Member States and of the Union' which are briefly referred to as 'broad economic policy guidelines' (see Article 139 para 2 lit a TFEU).

Article 156 TFEU).<sup>660</sup> 'What's in a name?', one is tempted to ask.<sup>661</sup> Since the terms recommendation and opinion are sufficiently broad, not least be‐ cause the Treaties fail to flesh them out (see 3.1.1. above), all these acts could be assigned to either group. Thus, the *numerus clausus* claim uttered with regard to Article 288 TFEU (limited to recommendations and opinions<sup>662</sup>) could be upheld.<sup>663</sup> A Communication regularly contains a certain (legal)

For accounts of this diversity in the literature see Knauff, Regelungsverbund 320 for the partly interchangeable denomination of soft law acts of the Commission; Cosma/Whish, Soft Law 25 f with regard to the field of competition law; Pampel, Rechtsnatur 33 f; von Bogdandy/Arndt/Bast, Instruments 120 for the example of a legally non-binding 'directive'; von Graevenitz, Mitteilungen 169 f for 'recommen‐ dations' and, in particular, 'opinions' which do not contain a behaviour-steering element concrete enough to (non-legally) commit somebody else.


<sup>660</sup> The Commission even has referred to its non-binding comments as 'decision' which the then Court of First Instance considered irrelevant in case T-295/06 *Base*, para 97.

<sup>661</sup> For examples in the CJEU's case law: case 147/83 *Binderer*, para 11: 'the choice of form cannot alter the nature of a measure'. In *Grimaldi* the Court has explicitly extended this finding to the case of legally binding and legally non-binding acts; case C‑322/88 *Grimaldi*, para 14; joined cases C‑463/10P and C‑475/10P *Deutsche Post*, para 58, in which the CJEU qualified a 'request' of the Commission as a decision; stressing the 'more formal status' of a request: Opinion of AG *Geelhoed* in case C-304/02 *Commission v France*, para 22; case T-671/15 *E-Control*, para 83: '[T]he fact […] that the contested opinion contains a "decision" […] does not bind this Court in its assessment for the purposes of determining whether the contested opinion is an act that is capable of forming the subject matter of an action for the purposes of Article 263 TFEU'. Conversely, acts whose name does not suggest that they have a normative character may contain soft law provisions: case T-190/00 *Regione Siciliana*, para 100, with reference to a Commission report containing guidelines; with regard to a 'letter' turning out to be a legally binding decision: joined cases 7/56 and 3–7/57 *Algera*, 54 f; similarly: joined cases C-189, 202, 205, 208 and 213/02P *Dansk Rørindustri*, para 211; case C-526/14 *Kotnik*, para 45.

opinion of the Commission,664 a Memorandum of Understanding conclu‐ ded between two institutions could be qualified as a joint decision or a joint recommendation, as the case may be, a Resolution, eg of the European Council, could be understood as recommendation or opinion (regularly drafted in broad terms<sup>665</sup>),<sup>666</sup> Guidelines give guidance on how to proceed on a certain matter and hence could be called recommendations,<sup>667</sup> etc.<sup>668</sup>

With regard to Article 263 para 1 TFEU, the following is to be noted: This provision excludes recommendations and opinions of the Council, the Commission, and the ECB. These acts therefore cannot be annulled by the Court. With regard to the EP, the European Council, and bodies, offices or agencies of the Union, the *lex citata* includes all acts 'intended to produce legal effects *vis-à-vis* third parties'. Thereby the Treaty seems to acknowledge that the latter bodies may adopt legally non-binding acts other than recommendations and opinions – and that would mean: that the Treaties principally allow for such other acts to be adopted. Therefore, the wording of Article 263 para 1 TFEU speaks against the *numerus clau‐ sus* argument. In judicial practice the Court does not make a difference between the exclusion of 'recommendations and opinions' on the one hand, and the inclusion only of 'acts intended to produce legal effects *vis-à-vis* third parties', on the other hand.669 It held that 'an action for annulment must be available in the case of all measures adopted by the institutions, whatever their nature or form, which are intended to have legal effects'.<sup>670</sup>

<sup>664</sup> Describing Communications as '"verwaltungsvollzugsbezogene" Empfehlung' [rec‐ ommendation concerning administrative execution]: Brohm, Mitteilungen 67.

<sup>665</sup> As in case of international agreements, this broadness is regularly required to keep all of the participating (Member) States on board; see Dawson, Governance 405.

<sup>666</sup> For the EP's own view that its resolutions constitute '"opinions" or recommenda‐ tions': case 230/81 *Luxembourg v European Parliament*, 269. The Court itself is not explicit on this question (eg para 39).

<sup>667</sup> See also case C-911/19 *FBF*, paras 42–45. On the legal bindingness of guidelines of the ECB issued under Article 12.1 of the ESCB/ECB-Statute see Hofmann/Rowe/ Türk, Administrative Law 548.

<sup>668</sup> This is reflected in the Commission's combination of the terms 'recommendation' on the one hand and 'rules of conduct', 'guidelines' etc, on the other hand; see Senden, Soft Law 162 and 173 f, with further references.

<sup>669</sup> See eg case T-154/10 *France v Commission*, paras 37 f, with further references. Remarkably, the wording of Article 265 TFEU excludes only recommendations and opinions, and only with regard to actions filed by natural or legal persons; see W Cremer, Art. 265 AEUV, para 6; Dörr, Art. 265 AEUV, para 14.

<sup>670</sup> Case C-114/12 *Commission v Council*, para 39, with further references; see also case T-496/11 *United Kingdom v European Central Bank*, para 32.

This interpretation does away with the differentiated wording of Article 263 para 1 TFEU, and it acknowledges the existence of EU soft law beyond recommendations and opinions (even for the Council, the Commission and the ECB) – by excluding it from judicial review (see also 6.2. below).

Also apart from the Court, the institutions do not appear to have applied the *numerus clausus* concept. This is reflected in their publication policy. As *von Bogdandy*, *Bast* and *Arndt* have noted in 2002, the Commission, for example, has published (most) recommendations and opinions in the L-series (Legislation; originally: *legislatio*) of the OJ, whereas other soft law acts were published in the C-series (Information and Notices; originally: *communicatio*).<sup>671</sup> Today the L-series contains four headings, the C-series five.<sup>672</sup> Certain Council, Commission or ECB recommendations are pub‐ lished in L II (non-legislative acts),<sup>673</sup> whereas other recommendations of these institutions and EP recommendations for the attention of the Council are published in C I (resolutions, recommendations and opinions).674 ECB recommendations adopted in accordance with Article 129 or Article 219 TFEU are published in C III (preparatory acts).675 Opinions are published either in C I (if non-compulsory opinions) or in C III (if compulsory

674 See <http://publications.europa.eu/code/en/en-110303.htm> accessed 28 March 2023. For Commission recommendations addressed to only one or a small number of addressee(s) this may be different; see Senden, Soft Law 173. With regard to the C-series more generally see case C-428/14 *DHL*, para 34; case C-410/09 *Polska Tele‐ fonia*, para 35; for the less strict distinction applied with regard to rules published on the respective institution's website see von Graevenitz, Verrechtlichung 76.

<sup>671</sup> See von Bogdandy/Bast/Arndt*,* Handlungsformen 118; see case C-226/11 *Expedia*, according to which 'the "C" series of the *Official Journal of the European Union* […], by contrast with the "L" series of the Official Journal, is not intended for the publi‐ cation of legally binding measures, but only of information, recommendations and opinions concerning the European Union' (para 30); see also case T-721/14 *Belgium v Commission*, para 40; similarly: Opinion of AG *Kokott* of 6 September 2012 in this case, para 32; cautious as regards the explanatory power of the publication series in which an act appears: von Graevenitz, Mitteilungen 170; Senden, Soft Law 101. For the meaning of publication in the OJ more generally: Sarmiento, Soft Law 275.

<sup>672</sup> See Council, Comments on the Council's Rules of Procedure (2022) 101f.

<sup>673</sup> See <http://publications.europa.eu/code/en/en-110203.htm> accessed 28 March 2023. That publications in the L-series are not necessarily intended to have legal effects was confirmed by the Court in case T-721/14 *Belgium v Commission*, para 39.

<sup>675</sup> See <http://publications.europa.eu/code/en/en-130800-tab.htm> accessed 28 March 2023.

opinions). Guidelines are published in L II,<sup>676</sup> resolutions in C I.<sup>677</sup> Com‐ mission communications are regularly published in the category C II (in‐ formation).<sup>678</sup> White Papers of the Commission are sometimes published in the C-series, sometimes they are not published in the OJ. Green Papers of the Commission – which contain soft law rules even less often than White Papers – are regularly not published in the OJ.<sup>679</sup> This practice is coined by the EU's Publications Office and by the institutions.<sup>680</sup>

Overall, a *numerus clausus* conception of Article 288 TFEU with regard to soft law acts seems to be feasible, but it is not in compliance with what already early legal scholarship<sup>681</sup> and administrative practice<sup>682</sup> suggest: namely, a more diversified morphology of EU soft law, extending the num‐ ber of categories of EU soft law beyond recommendations and opinions. Also the European Convention in 2002 argued in favour of a certain flexibility in this respect.683 Eventually, though, Article 288 TFEU and its predecessors have never been adapted to the rank growth of non-binding acts in practice. This may lead one to assume that the MS have approved of a limitation of non-binding acts to two categories.

<sup>676</sup> See <http://publications.europa.eu/code/en/en-110200.htm> accessed 28 March 2023.

<sup>677</sup> See <http://publications.europa.eu/code/en/en-110303.htm> accessed 28 March 2023.

<sup>678</sup> Critically: Schweda, Principles, para 30.

<sup>679</sup> For the character and purpose of Green Papers see Senden, Soft Law 124–126.

<sup>680</sup> With regard, for example, to the Council's publication preferences see Council, Comments on the Council's Rules of Procedure (2022) 101-104. Critically of the inconsistency of the soft law publication regime: Eliantonio, Soft Law 497, with a further reference; see della Cananea, Administration 63, with regard to the Com‐ mission's soft law in the field of State aid policy; for the – related – translation regime of EU soft law see case C-410/09 *Polska Telefonia*, in particular para 37.

<sup>681</sup> See Braams, Koordinierung 156; Gärditz, Unionsrecht, para 56; Ştefan, Soft Law 11; see also the references in von Bogdandy/Arndt/Bast, Instruments 96 (fn 17).

<sup>682</sup> See Turgis, Communications 52, with further references.

<sup>683</sup> See European Convention, Report of 29 November 2002 from the Chairman of the Working Group IX on Simplification, CONV 424/02, 6 f; for legally binding acts see ibid 4–6. Also the history of the Treaty of Lisbon in the context of the revision of the list of legal acts now contained in Article 288 TFEU does not allow for the conclusion that the Masters of the Treaties intended to limit the number of soft law acts available; see Schwarze, Soft Law 247 f.

### 3.2. The applicability of the principle of conferred powers

#### 3.2.1. Introduction

With institutions, bodies, offices and agencies of the EU rendering a variety of soft law acts, the question arises on which legal foundation these acts are adopted. 'Unlike international soft law,' *Cannizzaro* and *Rebasti* argue, 'European soft law does not operate in a normative vacuum but rather within the framework of the Treaties'.<sup>684</sup> While the Treaties leave it 'obscure how soft law is to be anchored in the Community legal system',<sup>685</sup> this observation cannot lead to the conclusion that the question of competence for the adoption of EU soft law does not deserve further consideration.

It could be argued that soft law, *qua* being legally non-binding, cannot possibly violate EU law, which is why compliance with the latter – and as a consequence this also means: with the competence order of the EU – is not required.686 However, this assumption is to be refused.687 Due to its potentially strong steering effects, soft law may very well interfere with the

<sup>684</sup> Cannizzaro/Rebasti, Soft law 231; note also the words of Knauff, Regelungsver‐ bund 296: 'gleichsam "natürliche" Daseinsberechtigung' [quasi 'natural' right to exist] of soft law in public international law as opposed to the EU legal order; see also Schermers/Blokker, Institutional Law, § 1218: 'All international organisations are empowered to issue recommendations'. Insisting on the applicability of the prin‐ ciple of attributed powers also with regard to international soft law: Sands/Klein, Bowett's Law, para 11–054; von Bogdandy, Principles 1933.

<sup>685</sup> Senden, Soft Law 24.

<sup>686</sup> See references in Senden/van den Brink, Checks 21.

<sup>687</sup> See case 230/81 *Luxembourg v European Parliament*, para 30, stressing the impor‐ tance of the respective 'content and observance of the rules on competence'; see also case C-42/99 *Queijo Eru*, para 20, with regard to the requirement that EU soft law be in accordance with EU law; Opinion of the Legal Service of the Council on the Commission's Communication on a new EU Framework to strengthen the Rule of Law, 10296/14, paras 18 f. Only exceptionally may soft law deviate from secondary law, and arguably only were it does not thereby impose obligations: see eg case T-87/05 *EDP*, paras 161–163; also the non-adoption of EU soft law can violate EU law, as is suggested by Article 265 para 1 TFEU which – deducing *e contrario* from its para 3 – allows the institutions and the MS to bring before the Court any failure to act (including the failure to adopt a legally non-binding act) on the part of an(other) institution; also the Court's Opinion in case 2/13 *ECHR II*, paras 196–200, is to be considered, in which it stresses the risk for the EU's autonomy of requests of national courts for (non-binding) advisory opinions from the European Court of Human Rights. For the non-bindingness of these advisory opinions see also <https://www.echr.coe.int/Documents/Press\_Q\_A\_Advisory\_opinion\_ENG.PDF> accessed 28 March 2023.

competences of the MS or of EU actors688 (and, as regards the rights of individual actors, also with fundamental rights<sup>689</sup>).

The EU's competences are subject to the principle of conferral.<sup>690</sup> Whether this principle also applies to the EU's soft law powers will be examined in the subsequent sub-chapters. When talking about competen‐ ces in this context, conceptually we have to distinguish between first the EU's competence to adopt soft law acts (competence of the Union; in German-speaking scholarship referred to as *Verbandskompetenz*), second the originator's general/specific power to adopt certain EU soft law acts (competence of the EU actor at issue; *Organkompetenz*<sup>691</sup>), and third the legal basis for the concrete soft law act (substantive legal basis).

While in practice it is possible that two or all of the three parts of the legal foundation – *Verbandskompetenz*, *Organkompetenz* and substantive legal basis – fall within one legal provision,692 for the sake of theoretical clarity they should be kept separate. *Verbandskompetenz* and *Organkompe‐ tenz* must be laid down in primary law,693 if only implicitly, whereas the legal basis for a concrete soft law act can also be provided for in acts based on the Treaties (secondary law).

# 3.2.2. The principle of conferral – an interpretation of the relevant terms

The principle of conferral is the core principle on the distribution of pow‐ ers between the EU and its MS.694 The starting point for answering the question whether or not it applies also to soft law powers is therefore an interpretation of Article 5 para 2 TEU. According to this provision, the EU shall 'act only within the limits conferred upon it […] in the Treaties' (*Verbandskompetenz*). This also applies to the *Organkompetenz*: Each insti‐

<sup>688</sup> See also Andone/Greco, Burden 88–90; Kadelbach, Art. 5 EUV, para 12; Wörner, Verhaltenssteuerungsformen 280 f, with further references.

<sup>689</sup> See eg Fischer-Lescano, Austeritätspolitik 37 ff.

<sup>690</sup> Article 5 para 1 TEU reads: 'The limits of Union competences are governed by the principle of conferral'. For the roots of this principle in public international law see Engström, Powers 45 ff.

<sup>691</sup> For the wide understanding of this term see eg Nettesheim, Art. 13 EUV, paras 7 f.

<sup>692</sup> A specific *Organkompetenz* and the substantive legal basis of an act always fall within one provision, see eg Article 36 para 2 TEU.

<sup>693</sup> See Nettesheim, Art. 13 EUV, para 85.

<sup>694</sup> For the different functions of the principle of conferral see Senden, Soft Law 291, with references to the literature.

tution, body, office or agency can act only 1) if an according competence is conferred on the EU, and 2) 'within the limits of the powers conferred on it [the institution, body, office or agency] in the Treaties' (Article 13 para 2 TEU<sup>695</sup>).<sup>696</sup> In this context, the Court has unequivocally held that 'the choice of the legal basis for a measure may not depend simply on an institution's conviction as to the objective pursued but must be based on objective factors which are amenable to judicial review'.<sup>697</sup>

The term 'competence' is very general. Its verbal interpretation does not suggest in any way that only the possibility to adopt legally binding acts shall be subsumed under the term.<sup>698</sup> Also from a systematic point of view, 'competences' within the meaning of the Treaties seem to be more than just the possibility to adopt legally binding acts.<sup>699</sup> The main argument in favour of such a systematic interpretation is Article 292 TFEU, a remainder

<sup>695</sup> While Article 13 para 2 TEU refers only to 'institutions', a wide reading of this term so as to include bodies, offices and agencies ought to be applied. The Masters of the Treaties – the MS – have certainly not intended to exclude these other EU actors from this conditionality of powers; in a similar vein: Nettesheim, Art. 13 EUV, paras 7 f.

<sup>696</sup> See Streinz, Art. 5 EUV, para 8. The fact that the wording of Article 249 TEC ('In order to carry out their task and in accordance with the provisions of this Treaty […]') in its successor provision, Article 288 TFEU, has been changed ('To exercise the Union's competences, the institutions shall adopt […]') does not rel‐ ativise the principle of conferral. The term 'competences' in Article 288 TFEU, correctly understood, means 'powers laid down in the Treaties'. In this respect the meaning has not changed as a consequence of the Lisbon reform. The new wording, however, indicates more clearly than its predecessor provision that there shall be no *Verbandskompetenz* which is not covered by an according *Organkompetenz* (*ar‐ gumentum* 'exercise the Union's competences' as opposed to 'their task'); pointing in a different direction: case C-16/16P *Belgium v Commission*, para 26.

<sup>697</sup> Case 45/86 *Commission v Council*, para 11; see also case C-70/88 *European Parlia‐ ment v Council*, para 9. The Court has acknowledged that the legislator 'legitimately' may have doubts as regards the appropriateness of a certain competence clause; case 8/73 *Hauptzollamt Bremerhaven*, paras 4–6.

<sup>698</sup> For the comparatively wide meaning of the term 'competence' in EU law see Braams, Koordinierung 218 ff; see also Nettesheim, Art. 288 AEUV, para 200; Sen‐ den, Soft Law 319 f; Schroeder, Art. 288 AEUV, paras 129 and 132; von Bogdan‐ dy/Bast, Competences 232 ff; against this view: Bieber, Art. 7 EG, para 55; Biervert, Mißbrauch 89 f; Nicolaysen, Gemeinschaftsrecht 60; Kraußer, Prinzip 88 and 94.

<sup>699</sup> While Article 2 para 2 TFEU – with regard to exclusive and shared powers of the EU – stresses the competence to adopt legally binding acts, it is to be noted that also coordinating powers of the EU – which do not necessarily entail the power to adopt legally binding acts (see Articles 5 f TFEU) – are a competence category of their own.

of Article I-35 of the (draft) Constitutional Treaty: The explicit vesting of the Council and the Commission<sup>700</sup> with the competence to adopt recom‐ mendations in Article 292 TFEU serves as a strong argument in favour of the applicability of the principle of conferral in the context of soft law pow‐ ers,<sup>701</sup> especially if it is contrasted with the more limited power of the ECB to adopt recommendations only 'in the specific cases provided for in the Treaties'.<sup>702</sup> While the former constitutes a general empowerment to adopt recommendations, the latter emphasises that with regard to the ECB no such general empowerment applies.<sup>703</sup> It is not perceivable why the TFEU would expressly mention these powers in a separate provision, thereby apparently distinguishing between the Council and the Commission on the one hand, and the ECB on the other hand, if it did not intend to grant the respective powers to the Council and the Commission (as regards the ECB, admittedly, the provision has a merely declaratory character).704 Article 292 TFEU generally indicates the applicability and importance of the principle of conferral in the context of the power to adopt recommendations, but it does not comprehensively regulate this power for all EU institutions,

<sup>700</sup> The Commission's general power to adopt recommendations conforms to Article 211 (second indent) TEC; with regard to the predecessor of Article 211 TEC, Article 155 TEEC, its conferral of powers on the Commission was clearly confirmed by the Court in case T-113/89 *Nefarma*, para 79; case C-303/90 *France v Commission*, para 30; differently: Nettesheim, Art. 292 AEUV, para 5.

<sup>701</sup> See European Parliament, 'Better Regulation and the Improvement of EU Regula‐ tory Environment. Institutional and Legal Implications of the Use of "Soft Law" Instruments', Note of March 2007, PE 378.290, 10 f. That Article 292 TFEU is a competence clause (not a declaratory provision) is, at least implicitly, confirmed in case T-721/14 *Belgium v Commission*, para 20.

<sup>702</sup> See Gellermann, Art. 292 AEUV, paras 2 f; Ruffert, Art. 292 AEUV, para 2; see also Nettesheim, Art. 292 AEUV, paras 1–4 and 7, doubting the qualification of Article 292 TFEU as a competence clause, but addressing the question of whether or not the principle of conferral applies in the context of recommendations.

<sup>703</sup> See Geismann, Art. 292 AEUV, para 3; Gellermann, Art. 292 AEUV, paras 3 and 5; Ruffert, Art. 292 AEUV, para 2. For the ECB's far-reaching power to adopt opinions see Article 127 para 4 TFEU.

<sup>704</sup> Neither is the argument convincing that Article 292 TFEU must be declaratory (in its entirety) due to its specific location in the Treaties. It is true that the respective Treaty section is entitled 'The legal acts of the Union' and that Article 288 TFEU only lists the most common acts of EU law, but does not grant the power to adopt them. Other provisions of this Section do grant powers, though: Articles 290 and 291 TFEU, for example, grant the power to confer certain powers, mainly to the Commission; see also Hofmann/Rowe/Türk, Administrative Law 237. Thus, it is by no means non-system that Article 292 TFEU grants powers.

bodies, offices and agencies. Thus, it cannot be held to have exclusionary effect in the sense that it excludes the power to adopt recommendations of other institutions, bodies, offices or agencies.<sup>705</sup>

Apart from Article 292 TFEU, there are further provisions which may shed light on the applicability of Article 5 para 2 TEU to soft law powers. Article 7 TFEU, for example, very broadly – ie not distinguishing between the power to adopt EU law and the power to adopt EU soft law – stipulates that the EU 'shall ensure consistency between its policies and activities […] in accordance with the principle of conferral of powers'. According to Arti‐ cle 2 para 5 TFEU, a further point of reference, 'in certain areas and under the conditions laid down in the Treaties, the Union shall have the compe‐ tence to carry out actions to support, coordinate or supplement the actions of the Member States […]'. We can assume that supporting, coordinating and supplementing MS' actions may take place (also,706 or even mainly) by means of adopting legally non-binding acts.707 These non-binding acts do not necessarily need to be soft law acts (for non-normative acts like reports, registers or work programmes in the EU context see 2.4. above), but they may *as well* be soft law acts. This is reflected eg in the provision of the OMC for some of these supporting, coordinating and supplementing competences, for example Article 148 (Employment), Article 153 (Social policy) and Article 173 para 2 TFEU (Industry). If the competence to adopt non-binding acts were not affected by the principle of conferral, why would the Masters of the Treaties have expressly provided for such delegation of power?708 It could be argued that the cited provisions are about regulating the procedure of support, coordination and supplementation, not about granting the – already existing – power to adopt EU soft law. However, the wording of the respective provisions does not in any way support this argument. Rather, it does not principally differ from the – uncontested – granting of the power to adopt legally binding acts: It stipulates that the institutions 'shall' or 'may' adopt legally non-binding acts just like, to take examples, Article 18 TFEU provides that they 'may' or Article 114 TFEU prescribes that they 'shall' adopt legislation. The Articles providing for OMC do not say, for example, that the institutions shall make use of their power to adopt legally non-binding acts in this or that way – which

<sup>705</sup> See Dickschen, Empfehlungen 27 f.

<sup>706</sup> The second sentence of the *lex citata* refers to 'legally binding acts of the Union'.

<sup>707</sup> See Tallberg, Paths 615.

<sup>708</sup> See also Kadelbach, Art. 5 EUV, paras 8.

would at least suggest that the powers at issue already exist.<sup>709</sup> What is more: The use of the term 'competences' in Article 2 para 5 TFEU – which, in this provision at least, clearly encompasses the power to adopt soft law – supports the view that the principle of conferral, which itself refers to 'competences', is applicable also with regard to EU soft law powers.<sup>710</sup> This is repeated in Article 6 TFEU, listing the concrete policies which fall within this competence category. Also here the term 'competences' is to be understood as including the adoption of soft measures.<sup>711</sup>

The pendant, as it were, of the term 'competences' in Article 5 TEU is the term 'powers' in Article 13 para 2 TEU. Both terms – *competences* and *powers* – are broad and imprecise. Like in case of the term 'competence', a certain 'power' can, also in a genuinely legal understanding, entitle to actions of many different categories: the power to rule, to adopt, to propose, to coordinate, etc. In general, it seems that in the Treaties' terminology competences are assigned to the EU (or remain with the MS), whereas powers are assigned to the institutions, bodies, offices or agencies of the EU. In places, however, these words are used as synonyms – both in the Treaties712 and even more so in legal literature.713 The principle of conferral according to Article 5 TEU – the EU's particular principle of legality –

<sup>709</sup> That in other fields a competence on the part of the Commission to engage in OMC may be implied is a different issue (which is not principally contested here); see Bast, Art. 5 EUV, para 48a.

<sup>710</sup> It is true that the Treaties do not always apply the term 'competences' in exactly the same way. With 'fields of competence' (Article 127 para 4 and Article 160 TFEU) the TFEU refers to the *Organkompetenzen* of institutions (ECB, Council), the term 'spheres of competence' (eg Article 191 para 4 or Article 211 TFEU) is intended to describe the EU's set of powers and the MS' powers. Obviously in an entirely different meaning, namely as 'personal/professional capacity', the term competence is to be understood when it is used to describe the qualities required for a certain post: 'general competence' (Article 17 para 3 TEU), 'recognised competence' (Article 255 para 2 TFEU). It is evident that here another kind of competence is referred to.

<sup>711</sup> See Nettesheim, Art. 6 AEUV, para 16.

<sup>712</sup> See Article 348 TFEU, speaking of MS' powers, Article 14 TFEU which refers to the 'powers' of the Union and the MS, or Article 207 para 6 TFEU which addresses the institutions' 'competences'; see also Bradley, Legislating 104.

<sup>713</sup> See references in Goldmann, Gewalt 495 f. Pursuant to search queries on www.cur ia.europa.eu, in (the English version of the) case law of the CJEU the word compe‐ tence (as applied in this context) seems to be in regular use only since the late 80s. Before that the term 'powers' ('Community powers') was much more common, the term 'competences' being used only sparingly. Also the TEEC in (the English translation of ) its original version used this *terminus technicus* only in its Article 173 ('lack of competence').

is equally referred to as 'principle of conferred *powers*' <sup>714</sup> or 'principle of conferral of *powers*' <sup>715</sup> (emphases added).

That good arguments speak in favour of generally including the capacity to adopt soft law acts in the term 'power' shall be exemplified as follows: Article 130 TFEU *inter alia* prescribes the independence of the ECB. The provision says that when exercising 'the powers […] conferred upon [it] by the Treaties', it shall neither seek nor take instructions from any body. The TFEU in its Article 132 specifies that the ECB shall 'make recommendations and deliver opinions'. It would run counter to the objective of the ECB's independence<sup>716</sup> to apply Article 130 only with regard to the ECB's power to adopt legally binding acts. In order to apply this provision also with regard to the making of recommendations and the delivery of opinions, they ought to be subsumed under the term 'powers'. Article 130 TFEU illustrates the broad understanding of the term 'powers' as used in the Treaties, in partic‐ ular that it may also include the possibility to adopt legally non-binding acts.717 A similar understanding seems to be applied by the CJEU, eg when dwelling on the Commission's 'express conferral of the power to adopt acts with no binding force'.718 While these examples do not actually give proof of the applicability of the principle of conferral with respect to soft law powers, they do suggest a wide scope of the term, so as to include the power to adopt soft law acts.

Another, more pragmatic argument brought forward in favour of apply‐ ing the principle of conferral to the adoption of soft law is the following. Its exclusion could lead to the institutions, and – to a more limited extent – also bodies, offices and agencies having recourse to soft law increasingly, thereby extending their scope of action 'softly' and by stealth – at the cost of the MS or of other EU actors.<sup>719</sup> *Peters* argues that '*ultra vires*-soft

<sup>714</sup> See eg Goucha Soares, Principle.

<sup>715</sup> Article 7 TFEU.

<sup>716</sup> For the different dimensions of the ECB's independence see Repasi, Limits 7.

<sup>717</sup> Coming to the same result: Senden, Balance 88.

<sup>718</sup> Case T-113/89 *Nefarma*, para 79; see also case C-370/12 *Pringle*, para 113, in which the Court refers to the Council's power to adopt recommendations under Article 126 paras 7 f TFEU.

<sup>719</sup> See also D Lehmkuhl, Government 157, with regard to competition law. For the famous phrase 'integration by stealth' see Majone, Dilemmas. Sceptically also Gold, Soft International Law 443, who remarks in this context: 'It is easy to be too condescending toward soft law'; see also Dawson, Waves 212; Simoncini, Regulation 20; Ştefan, Developments 882, with further references. For this phenomenon in public international law see Friedrich, Soft law 386: 'mission creep' by international

law can […] in practical terms pave the way to a formal extension of the competences'.<sup>720</sup> She exemplifies this by the environmental policy, research and technological development, culture and public health matters which, on the then Community level, had long been addressed by means of legally non-binding measures (including soft law), before pertinent competences were introduced in primary law by the Single European Act (SEA) and the Treaty of Maastricht, respectively.<sup>721</sup> These cases illustrate the strong steering effects of soft law.<sup>722</sup> In view of these effects which are sometimes very similar to those of law, applying a sustainable competence regime with respect to soft law powers is necessary in order to protect its effectiveness (*effet utile*).<sup>723</sup> Otherwise – that is to say: where no or only an overly lax regime on soft law powers is applied – the risk of soft law being abused as a substitute of legal rules (for which there is no competence) would be significant. In spite of the increasing importance and the sometimes remarkably strong steering effects of EU soft law, a number of scholars negate the applicability of the principle of conferral in this context.<sup>724</sup>

724 See eg Biervert, Mißbrauch 89 f, with a further reference; Calliess, Art. 5 EUV, para 9; Eekhoff, Verbundaufsicht 176 (fn 246), with a further reference; Rossi, Soft

organisations making increasing use of soft law; with regard to recommendations see Kotzur, Art. 292 AEUV, para 2.

<sup>720</sup> Peters, Typology 420.

<sup>721</sup> See Peters, Typology 423, with further references.

<sup>722</sup> For the possibility that soft law 'erober[t]' [conquers] new fields of EU action see von Bogdandy/Bast/Arndt, Handlungsformen 117; see also Opinion of AG *Sharpston* in case C-660/13 *Council v Commission*, para 62, who – on the question of whether an action against a non-binding agreement concluded by the Commission without an according authorisation by the Council is admissible under Article 263 TFEU – explained that '[a]ctions under Article 263 TFEU can be brought on grounds of, *inter alia*, lack of competence. Regardless of whether or not an act itself has legal effects, the fact that one institution has taken it whereas the Treaties give powers to do so to another institution means that the act of taking the decision has legal effects (by usurping the powers of the second institution). In the present case, applying that method would mean that where the Commission has taken a decision whereas, based on the substance of the pleas, the Treaties provide that this decision fell within the powers of the Council, the challenged act of the Commission has legal effects within the meaning of Article 263 TFEU'.

<sup>723</sup> See European Parliament, Resolution of 4 September 2007 on institutional and legal implications of the use of 'soft law' instruments, 2007/2028(INI), recitals I and L; Van Vooren/Wessel, Relations 37; confirming the applicability of the principle of conferral in an adapted form: Senden/van den Brink, Checks 22; see also Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 93–95, also with regard to the EU's institutional balance and – related to the former – to the separation of powers within the EU.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

A mediating understanding of the EU's competence regime with regard to EU soft law would suggest that soft law powers can be affirmed where the EU has (any) competence in the policy field at issue – at least *in dubio*, that is to say where the Treaties do not contain a clear indication to the contrary. In other words: A soft law competence can be assumed, unless it follows (explicitly or implicitly) from the Treaties that in a certain case there shall be no such competence. As we shall see in the next sub-chapter, some judgements of the Court seem to follow this approach.

# 3.2.3. The case law of the CJEU

Let us now take a look at the pertinent case law of the CJEU. The Court in general has made clear early on that EU soft law is subordinate to EU law,725 hence basically also to fundamental EU principles such as the principle of conferral. In its famous *Grimaldi* judgement the Court held, at first sight, quite differently: 'Recommendations […] are generally adopted by the institutions of the Community when they do not have the power under the Treaty to adopt binding measures or when they consider that it is not appropriate to adopt more mandatory rules'.726 However, the fact that soft law may be adopted for lack of a competence to adopt binding rules logically does not allow for the conclusion that *therefore* no competence is required for the adoption of soft law. The Court later on has explicitly

Law 15-17; see also references by Kadelbach, Art. 5 EUV, para 12 and by Knauff, Regelungsverbund 405; see also the considerations of Griller, Übertragung 155–158, uttered in a slightly different context, but worthwhile also here. For those confirm‐ ing the applicability of the principle of conferral in the context of soft law powers see Braams, Koordinierung 136–138, with many further references; see also Raschauer, Leitlinien 38; Ruffert, Art. 288 AEUV, para 99, with regard to recommendations and opinions, and with further references; apparently in favour (with regard to soft international commitments): Viterbo, Arena 216 f.

<sup>725</sup> See case 43/75 *Defrenne*, para 57; case 59/75 *Pubblico Ministero*, para 21, both stressing that the time-scale laid down in a resolution may not modify the pertinent time-scale prescribed in a Treaty provision. With regard to secondary law explicitly: case 149/73 *Witt*, para 3; see also joined cases 69–70/76 *Dittmeyer*, para 4; case 798/79 *Hauptzollamt Köln-Rheinau*, paras 11f; case 190/82 *Blomefield*, para 21 (with regard to internal soft law); case 310/85 *Deufil*, par 22; case C-266/90 *Soba*, para 19; case C-35/93 *Develop Dr. Eisbein*, para 21; case C-226/94 *Albigeois*, para 21; case T-9/92 *Peugeot*, para 44; for an example of soft law deviating from law (in the field of climate protection regulation) see J Scott, Limbo 336.

<sup>726</sup> Case C‑322/88 *Grimaldi*, para 13; with regard to this passage see Opinion of AG *Tesauro* in case C-325/91 *France v Commission*, para 22.

refused an all-encompassing competence to adopt soft law on the part of the institutions: '[T]he fact that a measure such as the Guidelines is not binding is [not] sufficient to confer on [the Commission] the competence to adopt it. Determining the conditions under which such a measure may be adopted requires that the division of powers and the institutional balance established by the Treaty […] be duly taken into account'. <sup>727</sup> With regard to the Commission's competence to lay down detailed rules by means of a soft law act, eg in the field of State aid or competition law, the Court held that 'in the exercise of the powers conferred on it by Articles 87 EC and 88 EC, the Commission may adopt guidelines designed to indicate how it intends, under those articles, to exercise its discretion in regard to new aid or in regard to existing systems of aid'. <sup>728</sup> With regard to a similar act, it held that the Commission adopted it 'in accordance with the powers thus vested in it by Article [107 – after Lisbon] et seq. of the Treaty'.729 As a preliminary result, we can state that the Court for the adoption of an EU soft law act deems necessary the existence of an appropriate competence.

This finding still does not answer the question whether the Court – with regard to soft law powers – follows a positive approach, the principle of conferral, or a negative (or: *in dubio*) approach, according to which the respective power is to be confirmed if primary law does not contain indicators to the contrary (see 3.2.2. above).730 Both approaches require

<sup>727</sup> Case C-233/02 *France v Commission*, para 40, with regard to (non-binding) guide‐ lines negotiated between the Commission and the United States Trade Representa‐ tive and the Department of Commerce; similarly: case C-660/13 *Council v Commis‐ sion*, para 43. With regard to the relationship between the principle of conferred powers and the duty to maintain the Treaties' institutional balance see Senden, Soft Law 74–76; see also case T-327/13 *Mallis*, para 43; case C-62/14 *Gauweiler*, para 12, in which the Court negates its own power to adopt a purely advisory act in an Article 267 TFEU procedure: 'the Court does not have jurisdiction to provide […] answers which are purely advisory'. As authoritative interpreter of EU law, the Court plays a special role in the EU. Hence it does not come as a surprise that it considers its rulings binding.

<sup>728</sup> Case C-242/00 *Germany v Commission*, para 27; similar in case C-526/14 *Kotnik*, para 39; see also Opinion of AG *Léger* in case C-382/99 *Netherlands v Commission*, para 47, in which the Commission's power to adopt guidelines in the field of State aid law is deduced from the principle of good administration; stressing the self-lim‐ iting effect of such acts, but at the same time highly critical of the Commission practice at issue here: Cannizzaro/Rebasti, Soft law 226–229.

<sup>729</sup> Case C-169/95 *Spain v Commission*, para 19; see also case T-149/95 *Ducros*, para 61, with references to further case law.

<sup>730</sup> Unclear also in case C-16/16P *Belgium v Commission*, paras 12 and 28, in which the claimant argues that the recommendation at issue is challengeable under Article 263

a competence to adopt soft law. It is the method by which they come to confirm or refuse a competence in which they differ from each other. In general, it is more difficult to confirm a positive competence than to find no rule to the contrary. The Court's reference to the division of powers and the institutional balance established by the Treaty in the case *France v Commission* cited above, in my view, is at least to be qualified as a hint at the applicability of the principle of conferral.<sup>731</sup> In the *Nefarma* case, the Court has referred to the 'express conferral of the power to adopt acts with no binding force' in what is now Article 292 TFEU, deducing therefrom that voluntary compliance with these measures is 'an essential element in the achievement of the goals of the Treaty'.<sup>732</sup> While not being an express confirmation of the applicability of the principle of conferral, the Court's approach in these cases rather speaks in favour of its applicability than against it.

Admittedly, there are also cases more or less vaguely hinting in a different direction.733 In the case *Commission v McBride* and others it held 'that the requirement for legal certainty means that the binding nature of any act intended to have legal effects must be derived from a provision of EU law which prescribes the legal form to be taken by that act and which must be expressly indicated therein as its legal basis'.734 The Court also assumed that the indication of legal bases may create, on the part of the addressee, the impression that the act is legally binding.735 This indicates that the *referral to* a legal basis, not the legal basis itself, may not be required in case of a soft law act.

TFEU for violation of – *inter alia* – the principle of conferral, to which the Court only generally replies that 'it is not therefore sufficient that an institution adopts a recommendation which allegedly disregards certain principles or procedural rules in order for that recommendation to be amenable to an action for annulment, although it does not produce binding legal effects'.

<sup>731</sup> Even more determined: Van Vooren/Wessel, Relations 38, who deduce from that passage that '[t]he application of the principles of conferral (Article 5 TEU) and institutional balance (Article 13 TEU) continue to apply and must be respected'.

<sup>732</sup> Case T-113/89 *Nefarma*, para 79.

<sup>733</sup> Case T‑320/09 *Planet AE*, para 57, with further references; case C-501/11P *Schindler*, para 68, where the Court, with regard to Commission Guidelines on the method of setting fines in competition law, generously said: 'No provision of the Treaties pro‐ hibits an institution from adopting such rules of practice'. This seems to constitute the negative or *in dubio* approach referred to above.

<sup>734</sup> Case C‑361/14P *Commission v McBride*, paras 47.

<sup>735</sup> Case T-496/11 *United Kingdom v European Central Bank*, para 47; see also case C-687/15 *Commission v Council*, para 54.

All in all, the Court's case law, for lack of a uniform line of argumenta‐ tion on the issue of soft law powers, remains unclear. We can only deduce that the Court demands an according competence for the adoption of soft law, and that – when it comes to the method by means of which this competence is to be established – it does not outright refuse the application of the principle of conferral.

### 3.2.4. Résumé

Following this verbal, systematic and teleological interpretation of Article 5 TEU and Article 13 para 2 TEU respectively, and having considered the relevant case law of the Court, it appears that EU law requires an according competence for the adoption of EU soft law. As regards the required charac‐ teristics of this competence – positive competence (positive approach) or no rule to the contrary (negative or *in dubio* approach) – we may argue as follows: Given, first, that the principle of conferral is the legal foundation of the EU's activity and, second, that it does not explicitly exclude soft law acts (which are provided for in many provisions throughout the Treaties) from its scope and, third, given that neither the Treaties nor the Court's case law expressly confirm a negative or *in dubio* approach, the better reasons speak in favour of the applicability of the principle of conferral.<sup>736</sup>

In practice, the originators of EU soft law only sometimes indicate the legal basis for their soft law acts (within the acts themselves).737 This may be interpreted as a preference for the negative or *in dubio* approach, or simply as reflecting a lack of awareness of the underlying problem.<sup>738</sup>

<sup>736</sup> See eg von Alemann, Einordnung 124. See Senden, Soft Law 294 f and 479 f, arguing in favour of applying to soft law 'one or more of [the] other functions' of the principle of conferred powers, but not its requiring 'the establishment of a legal basis in the Treaty or in secondary legislation', with a view to 'ensuring that the [respective EU body] acts within the boundaries of the powers and tasks assigned to it'; see also Council, Comments on the Council's Rules of Procedure (2022) 100 f.

<sup>737</sup> See Andone/Greco, Burden 89. See also the exemplary list of soft law acts adopted by the Commission in Meijers Committee, Note 4 f. As one of the rare exceptions see Article 18 para 2 of Council Regulation 1/2003 in which the legislator insists on the indication of the legal basis also for a 'simple request for information' (as opposed to a decision according to para 3), apparently a soft law act of the Commis‐ sion; see Lenaerts/Maselis/Gutman, Procedural Law 275 f.

<sup>738</sup> See also Andone/Greco, Burden 89 f, who assert that 'not mentioning the legal basis on which the recommendations have been enacted amounts to an evasion of the

We can broadly conclude that any Union action, be it binding or not, must rest on a competence conferred upon the EU and must be in compli‐ ance with EU primary – and possibly secondary – law more generally, since otherwise it would infringe upon the competences of the MS or the powers of the respective other EU actors, the protection of which is the main purpose of the principle of conferral. In view of the unclear case law of the Court, and in particular in view of an administrative practice which seems to be inattentive to this question, some doubts remain. In the subsequent chapters the competence clauses for the adoption of soft law acts are examined in more detail. With a comprehensive account of the pertinent structure of the Treaties at hand, we shall revisit our preliminary result – that the principle of conferral applies also to EU soft law powers (see 3.9. above).


As was set out above, the principle of conferral is the central legal reference point when it comes to delimiting the EU's and its institutions', bodies', offices' and agencies' competences. It is coined by the rather permissive case law of the CJEU, which is often based on *effet utile* considerations.739 The implied powers doctrine allows, under certain conditions, for a particularly extensive interpretation (or even: development of law740) of competence clauses. It may be somehow at odds with the principle of conferral,741 but it constitutes – pursuant to the Court's case law – an established part of the

burden of proof: the Commission should argue that it has a power as conferred in the Treaties, but instead falls sometimes short of doing so'.

<sup>739</sup> See Calliess, Art. 5 EUV, para 18; Schima, Art. 5 EUV, para 17. For the localisation of the *effet utile* both in the field of interpretation and actual development of law (often referred to in its German translation: *Rechtsfortbildung*) see Potacs, Auslegung 92–95; for the CJEU's general openness to legal developments see summary in Pechstein/Drechsler, Auslegung, paras 56–61.

<sup>740</sup> For this term (*Rechtsfortbildung*) see the preceding fn and, more generally, Larenz/ Canaris, Methodenlehre 191 ff.

<sup>741</sup> This tension is explicitly acknowledged in case T‑143/06 *MTZ*, para 47, with further references; case T-496/11 *United Kingdom v European Central Bank*, para 105.

EU's competence regime.742 Hence both approaches have to be applied in combination, if need be. Already in one of its earliest judgements, the Court held that 'it is possible […] to apply a rule of interpretation, according to which the rules laid down by […] a law presuppose the rules without which that […] law would have no meaning or could not be reasonably or usefully applied'.<sup>743</sup>

The concept of implied powers allegedly stems from early 19th century case law of the US Supreme Court<sup>744</sup> and has become an established in‐ terpretative tool in national jurisdictions as well as in international case law.<sup>745</sup> The ICJ, for example, in its Advisory Opinion in the famous *Count Bernadotte* case held that the UN 'must be deemed to have those powers which, though not expressly provided in the Charter, are conferred upon it by necessary implication as being essential to the performance of its duties'.746 This passage suggests that essential powers can be implied to the UN-Charter as a whole, not (only) to express powers. Judge *Hackworth* in its Dissenting Opinion favoured a stricter approach, arguing that '[t]here can be no gainsaying the fact that the Organization is one of delegated and enumerated powers. It has to be presumed that such powers as the Member States desired to confer upon it are stated either in the Charter or in complementary agreements concluded by them. Powers not expressed

<sup>742</sup> In favour of a harmonious conception of the co-existence of the principle of con‐ ferral and the implied powers doctrine: Bast, Art. 5 EUV, para 21; see also case 2/94 *ECHR*, para 29: '[…] where no specific provisions of the Treaty *confer* on the Community institutions express or *implied powers* to act' (emphasis added); similarly with regard to the principle of attribution in public international law: Friedrich, Soft law 382 f.

<sup>743</sup> Case 8/55 *FEDECHAR*, 299. Here the Court seems to qualify implied powers as following from legal interpretation, not as a development of law. See references to other legal authorities referring to this principle, which allegedly is a 'Selbstver‐ ständlichkeit' [matter of course] in: Nicolaysen, Theorie 131f and 134. *Stadlmeier* contends that the CJEU already in the *FEDECHAR* case has applied the resulting powers doctrine. The Court's finding reasonably could have been based on different legal arguments (in particular the competence-based implied powers doctrine), though. Also later the Court has referred to the *FEDECHAR* case in the context of competence-accessory implied powers; see arguments and references in Stadlmeier, Implied Powers 376 f; differently: Senden, Soft Law 71 (fn 43).

<sup>744</sup> See Stadlmeier, Implied Powers 354 f, with further references.

<sup>745</sup> For the German and the international legal order see Kruse, Implied powers; point‐ ing at the importance not only of the implied powers doctrine but also of customs in order to legitimise powers going beyond (express) attribution: Friedrich, Soft law 387.

<sup>746</sup> *Reparation for injuries*, ICJ Reports 1949, 174, 182.

cannot be freely implied. Implied powers flow from a grant of expressed powers, and are limited to those that are "necessary" to the exercise of powers expressly granted. No necessity for the exercise of the power here in question has been shown to exist'. He criticised the ICJ's 'generosity': 'The results of this liberality of judicial construction transcend, by far, anything to be found in the Charter, as well as any known purpose entertained by the drafters of the Charter'<sup>747</sup> . 748

This disagreement reflects the two categories in which implied powers scholarly can be divided, namely competence-accessory implied powers and objective-accessory implied powers (or resulting powers<sup>749</sup>).<sup>750</sup> Coming back to the case of the EU, we can state the following: Implied powers of the first kind, that is competence-accessory implied powers, can be assumed where the Union and its bodies respectively, has/have a related express competence (regularly in the Treaties). Only on the basis of this compe‐ tence additional powers can be implied which are 'necessary' for exercising this competence. Implied powers of the second kind – objective-accessory implied powers or resulting powers – can be assumed ('implied') already if the political objectives of the Treaties so 'require'.<sup>751</sup>

The CJEU has been cautious in applying the unorthodox and highly problematic interpretative tool called resulting powers doctrine, especially in the context of internal competences.752 While this approach principally could be in accordance with the CJEU's *effet utile* doctrine,753 the compe‐ tences which could – due to the EU's broad scope of objectives – possibly

751 See Stadlmeier, Implied Powers 376.

753 See Šadl, Role 33 ff.

<sup>747</sup> Dissenting Opinion by Judge *Hackworth* in *Reparation for injuries*, ICJ Reports 1949, 198 f.

<sup>748</sup> It appears that the ICJ has taken a more restrictive approach in later case law; see eg *Nuclear Weapons*, ICJ Reports 1996, para 21; see also Klabbers, Introduction 80.

<sup>749</sup> See Nicolaysen, Theorie 140, who arguably applies a slightly wider understanding of the term 'resulting powers'.

<sup>750</sup> On this established differentiation see Stadlmeier, Implied Powers 361f, with further references.

<sup>752</sup> In the context of external competences the Court has taken a more permissive approach, which today is reflected upon in Article 216 para 1 (second alternative) TFEU. Article 216 para 1 TFEU – together with Article 3 para 2 TFEU – is to be understood as a codification of the up to then case law; see also Nowak/Masuhr, EU only 203 f. However, also in other fields the Court exceptionally refers to the result‐ ing powers doctrine. The rhetoric applied in case T-240/04 *France v Commission*, para 36, with many further references, for example, suggests that it is the 'objectives of the Treaties' which matter.

be implied via this deduction of means from ends appears to be limitless.<sup>754</sup> Thus, this approach would create too strong a tension with the principle of conferral. Since the Treaty of Rome, the Treaties have contained a provi‐ sion which allows for an objective-based extension of competences by the legislator, namely the so-called 'flexibility clause'. Under the Lisbon regime this provision is contained in Article 352 TFEU, which says: 'If action by the Union should prove *necessary*, within the framework of the policies defined in the Treaties, *to attain one of the objectives set out in the Treaties*, and the Treaties have not provided the necessary powers' (emphases added), the Council (with the consent of the Parliament) shall unanimously adopt the 'appropriate measures'.<sup>755</sup> As opposed to the resulting powers doctrine which essentially is a (very extensive) form of interpretation (or rather: a development of law beyond legal interpretation), Article 352 TFEU allows for an extension of the EU's *Verbands-* and *Organkompetenzen* under consideration of the political objectives *by the legislator*, which ensures democratic legitimacy and, above all, legal certainty.

With the resulting powers doctrine being applied by the CJEU only very restrictively,756 and for obvious lack of a legal act based on Article 352 TFEU which may serve as a general legal basis for the adoption of soft law acts, it is the competence-accessory implied powers doctrine which remains to be discussed in the context of the power to adopt soft law acts.

The case law of the CJEU on competence-accessory implied powers shows that the Court so far has implied correspondent/complementary competences as regards both internal and external competences. The CJEU's dogma in this context is: '[W]hen an article of the Treaty confers a specific task on an institution, it must be accepted, if that provision is not to

<sup>754</sup> See also Senden, Soft Law 313, with further references.

<sup>755</sup> For the limits of the flexibility clause see eg case 2/94 *ECHR*, paras 30 and 35; for the term 'appropriate measures' in the context of Article 108 para 1 TFEU see Opinion of AG *Darmon* in joined cases 166 and 220/86 *Irish Cement*, para 24.

<sup>756</sup> When stating that what is now Article 352 TFEU may apply only if there is no corre‐ sponding express or implied competence, it seems to confirm that objective-based competences may only be created on the basis of Article 352 (and not implied by the Court); see case 2/94 *ECHR*, para 29: '[The flexibility clause] is designed to fill the gap where no specific provisions of the Treaty confer on the Community institutions express or implied powers to act, if such powers appear none the less to be necessary to enable the Community to carry out its functions with a view to attaining one of the objectives laid down by the Treaty'; see also case C‑166/07 *Eu‐ ropean Parliament v Council*, para 41; see also case C-295/90 *European Parliament v Council*, para 20.

be rendered wholly ineffective, that it confers on that institution necessarily and per se the powers which are indispensable in order to carry out that task'.<sup>757</sup> Along this line of argumentation, the Court in its case law implied the competence of the High Authority of the European Coal and Steel Community (ECSC) to recover overpayments ('necessary corollary') to the competence to make equalisation payments;<sup>758</sup> the competence of the Com‐ mission to require the MS to notify certain information to its competence (duty) to arrange consultations between the MS and the Commission;<sup>759</sup> the competence of the European Economic Community (EEC) to enter into international agreements to a legislative competence of the EEC in the same field.<sup>760</sup> On the contrary, the Court refused to confirm implementing powers of the Commission by means of comparison with other (in this respect more explicit) Treaty provisions and the Treaty's 'general structure' – arguably also for lack of necessity.<sup>761</sup>

Whilst the acknowledgement of certain implied powers by the CJEU allegedly is 'exceptional',762 it is nevertheless an important asset when iden‐ tifying *Verbandskompetenz* as well as *Organkompetenz.* With the much clearer competence regime after the Lisbon Treaty the importance of the

759 See joined cases 281, 283, 284, 285 and 287/85 *Germany v Commission*, para 28.

<sup>757</sup> See eg case T-496/11 *United Kingdom v European Central Bank*, para 104, with further references. Similarly already joined cases 281, 283, 284, 285 and 287/85 *Ger‐ many v Commission*, para 28. For cases not applying this wording see case 242/87 *Commission v Council*, in particular paras 12 f; case C-106/96 *United Kingdom v Commission*, para 19.

<sup>758</sup> See joined cases 4–13/59 *Mannesmann*, 130 f.

<sup>760</sup> For the case law on implied powers allowing the Community to conclude interna‐ tional agreements see case 22/70 *Commission v Council*, para 28; for the early follow-up cases see references in Cremona, Relations 433–435; see Klamert, Loyal‐ ty 73–75, with regard to the role the legal notion of loyalty played in this case, and 105 f, with regard to its effects and the Court's rationale to affirm implied powers here; for the exclusion of reverse implied powers, that is to deduce internal competences from external competences of the EU, in the case of the Common Commercial Policy see Article 207 para 6 TFEU; see also Streinz, Europarecht (10th edn) para 1271.

<sup>761</sup> See case 25/59 *Netherlands v High Authority*, 371f; case 20/59 *Italy v High Authority*, 335–338.

<sup>762</sup> See case T-240/04 *France v Commission*, para 37; case T-143/06 *MTZ*, para 47; case T-496/11 *United Kingdom v European Central Bank*, para 105.

implied powers doctrine arguably has decreased, but not lost its impor‐ tance entirely.<sup>763</sup>

But may implied powers also be used to 'determine' a legal basis for soft law acts which cannot be based on an express Treaty competence? Since implied powers assumed by the Court are annexed to a certain competence (accessoriness), they may not serve as a *general* legal basis for the adoption of soft law, but – if at all – be relevant when identifying specific compe‐ tences of the EU and its bodies, respectively.<sup>764</sup> On the presumption that the principle of conferral applies to soft law powers, it appears reasonable to apply the CJEU's complementary implied powers doctrine also in this context. The then Court of First Instance in case T-240/04 *France v Com‐ mission* suggests so when – in the context of implied powers – it generally states: 'Not only the substantial provisions, but also the form and *binding nature* of the regulation, must fulfil that condition of necessity' (emphasis added).765 In the following paragraph, it expresses: 'To consider that the Commission was implicitly empowered to adopt the contested regulation, it is necessary, not only that the Commission could adopt measures organ‐ ising details of procedure for the examination of investment projects that are communicated to it […], but also that it requires the adoption of those measures in the form of a regulation, binding in its entirety and directly applicable in all Member States'.766 If the bindingness of a measure must be necessary in order for the competence to adopt such measure to be implied to an express competence, we may conclude *e contrario* that also (mere) soft law measures may be deemed necessary.

Since its necessity (for the effective exercise of an express competence) is the core condition for a competence to be implied, it requires special atten‐ tion. The predominance of this criterion in practice boils down the ques‐ tion whether or not competences can be implied to the question: necessary or not? In that sense, the examples of competences implied by the Court in its case law reflect necessary competences, competences not implied were deemed 'not necessary'. The Court assigns to this term the meaning it has in everyday language. An interpretation as strict indispensability today

<sup>763</sup> See Borchardt, Grundlagen, para 481; see also case C-600/14 *Germany v Council*, para 45. For the long-lasting claim for a more precise competence regime of the EU see eg Steindorff, Grenzen 26 ff.

<sup>764</sup> See Knauff, Regelungsverbund 402–404, with many further references.

<sup>765</sup> Case T-240/04 *France v Commission*, para 38.

<sup>766</sup> Case T-240/04 *France v Commission*, para 39.

does not seem to be intended by the Court767 and such a high threshold would be impossible to apply in each and every case, simply for lack of data. Instead, the Court assumes a considerable leeway in assessing whether or not a certain competence is necessary.<sup>768</sup> In terms of importance and malleability, the term 'necessity' as used in the context of implied powers is comparable to the 'necessity to reach the aim' which forms the final step of the scheme for the evaluation of eg fundamental rights infringements (proportionality test<sup>769</sup>). Applying this margin of appreciation, the Court has, for example, considered necessary the competence of the Commission to 'adopt guidelines requiring compliance, not only with criteria pertaining exclusively to competition policy, but also with those applicable in relation to the common fisheries policy, even if the Council had not expressly authorised it to do so'<sup>770</sup> and considered not necessary the recommendation to suspend investment projects for organising the communication, exami‐ nation and discussion procedure for certain investment projects.<sup>771</sup>

With regard to the Commission's power to adopt 'interpretative and decisional instruments'772 (eg the Commission's communications in the field of State aid law) in the literature it is argued – and in the CJEU's case law it was decided – that it should be implied in the executive powers of the Commission in the respective field.773 This is to say that where the Com‐ mission has to apply a certain provision it may explicate in a soft law act how it interprets this provision,774 or announce a shift in its interpretation resulting in a new legal situation.775 The European Parliament (politically) affirms this practice in the interest of legal certainty, but in this context also

<sup>767</sup> For the change of the Court's wording from 'indispensable' to 'necessary' and for an interpretation of this change in parlance see Chamon, Agencies 141.

<sup>768</sup> Also the term 'necessary powers' in Article 352 TFEU allows for a wide discretion of the legislator; see Rossi, Art. 352 AEUV, paras 51–53.

<sup>769</sup> See eg Ehlers, Principles, paras 48 f, with further references.

<sup>770</sup> Case C-311/94 *IJssel-Vliet*, para 34.

<sup>771</sup> See case T-240/04 *France v Commission*, para 41.

<sup>772</sup> Senden, Soft Law 138; see also Braams, Koordinierung 154 f.

<sup>773</sup> See Nettesheim, Art. 291 AEUV, para 27; see also references in Senden, Soft Law 313–318; critically in the context of State aid (soft) law: Cini, Soft law approach 200 f; with regard to interpretative Commission communications more generally see Hofmann/Rowe/Türk, Administrative Law 552–555; Turgis, Communications 51.

<sup>774</sup> See case C-146/91 *KYDEP*, para 30; case C-169/95 *Spain v Commission*, para 19; case C-387/97 *Commission v Greece*, para 84; see also case T-374/04 *Germany v Commission*, para 110.

<sup>775</sup> See Ştefan, Soft Law 62 f, with references to case law.

warns of 'ambiguous and pernicious' instruments leading to an 'inadmissi‐ ble extension of law-making by soft law'.<sup>776</sup> According to the Court, the Commission needs to have a specific executive power in the respective field, though. The mere possibility that it may make the violation of a certain provision subject to a Treaty infringement procedure – which could apply to nearly any rule of EU law – does not appear to be sufficient,<sup>777</sup> also against the background of Article 290 and Article 291 TFEU. From the benevolent case law allowing for soft powers of the Commission implied in (hard) decision-making power, the *principal* possibility of implying soft powers also to the (hard) powers of other institutions, bodies, offices or agencies of the EU can be deduced.<sup>778</sup> On the whole, arguably the power to adopt a soft law measure is more likely to be 'necessary' and hence to be implied than the power to adopt hard law measures (in addition to those to which they should be implied), as it appears to be less intrusive to other bodies' and the MS' competences, and therefore also more likely to be in accordance with the principle of subsidiarity.<sup>779</sup>

<sup>776</sup> European Parliament, Resolution on institutional and legal implications of the use of 'soft law' instruments, 2007/2028(INI), para A and para 10. The EP further asserts that '[s]oft law tends to create a public perception of a "super-bureaucracy" without democratic legitimacy, not just remote from citizens, but actually hostile to them, and willing to reach accommodations with powerful lobbies which are neither transparent, nor comprehensible to citizens' (para Y). In its view, 'the distinction between *dura lex/mollis lex*, being conceptually aberrant, should not be accepted or recognised' (para B); see also Senden/van den Brink, Checks 16; for earlier criticism of the EP see Senden/Prechal, Differentiation 181; Ştefan, Soft Law 21; see also Résolution du Parlement européen du 8 mai 1969, sur les actes de la collectivité des États membres de la Communauté ainsi que les actes du Conseil non prévus par les traités adoptée à la suite du rapport fait au nom de la Commission juridique par M. Burger, C63/18 (1969), in which the EP took an (early) critical account of the adoption of Community acts not provided for in the Treaty, in particular by the Council; for the EU institutions' view on soft law more generally see Frykman/ Mörth, Soft Law 155.

<sup>777</sup> See case C-146/91 *KYDEP*, para 30: *argumentum* 'in the context of its collaboration with the national authorities'.

<sup>778</sup> See Senden, Soft Law 480, referring to a Council competence to adopt recommen‐ dations which may be implied in its decision-making power according to Article 202 para 2 TEC.

<sup>779</sup> See Knauff, Regelungsverbund 411f; Senden, Soft Law 179 f and 206 f; see also V.3.4.2. below.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

### 3.3.2. *Argumentum a maiore ad minus*

#### 3.3.2.1. The *argumentum a maiore ad minus* in EU law

When analysing the competence to adopt soft law, the question arises whether or not the competence to adopt law regularly implies a competence to adopt a (content-wise comparable) soft law act directed to the same addressees. In a regime based on the rule of law, considering a legally non-binding act a weaker form of exercising power (a *minus*, as it were), this deduction *a maiore ad minus* seems to be viable.<sup>780</sup> This is because a non-binding act principally allows for a lawful deviation from the deman‐ ded behaviour. This approach may be applied whenever in a certain case further reaching powers undoubtedly exist. We may illustrate this with an example from the field of public international law: Apart from the peacemaking measures explicitly laid down in Articles 39 ff of the UN-Charter, also the adoption of mere peace-keeping measures is deemed lawful. That the adoption of the latter, something *less* than peace-making measures, is in accordance with the UN-Charter is argued *a maiore ad minus*. <sup>781</sup> This interpretative tool, in principle, has been accepted also by the CJEU.<sup>782</sup>

The *argumentum a maiore ad minus* applied in the context of competen‐ ces is to be perceived as a sub-category of the implied powers doctrine.<sup>783</sup> It implies powers, not primarily under consideration of the criterion of necessity, but with regard to the *amount* of existing powers. These existing powers may imply the competence to apply less intrusive means. The con‐ sideration to adopt the least intrusive act available in order to set in place a certain policy is a general quest which, in the context of EU law and to the extent it benefits the MS' room for manoeuvre, may be deduced from the principle of subsidiarity.<sup>784</sup> After all, such action facilitates an important

<sup>780</sup> See also Nettesheim, Art. 292 AEUV, para 7.

<sup>781</sup> See Lorenzmeier, Völkerrecht 100.

<sup>782</sup> See eg case T‑469/07 *Philips*, paras 71 and 85.

<sup>783</sup> Traditionally, the *argumentum a maiore ad minus* is understood as an analogy-like tool; see Larenz/Canaris, Methodenlehre 208. Since the Court perceives implied powers as a method of interpretation, consequently it must also qualify the *argu‐ mentum a maiore* that way. After all, it seems to be the least intrusive variant of the implied powers doctrine because it allows to imply powers only to (related) more far-reaching competences.

<sup>784</sup> Note that this principle is guiding the *exercise of* competences, not the competences themselves; see Knauff, Regelungsverbund 411–415; Lienbacher, Art. 5 EUV, para 18; Raschauer, Leitlinien 34; Senden, Soft Law 90, each with further references;

objective of the subsidiarity principle, that is to 'promote […] local owner‐ ship over policies and regulation'.<sup>785</sup> Some also refer to the principle of proportionality in this context.<sup>786</sup> Also the European Council in 1992 has held that under the principle of proportionality '[n]on-binding measures such as recommendations should be preferred where appropriate'.<sup>787</sup> In my opinion, this view is to be refused. The principle of proportionality cannot be understood as suggesting the use of soft law as this would mean suggest‐ ing the disproportionality of law (in certain cases at least).<sup>788</sup> The rule of law, one of the core principles the EU legal order is based upon,<sup>789</sup> impedes a view according to which a legally binding act is disproportionate *qua* be‐ ing legally binding.<sup>790</sup> The act may be unlawful (even: *ultra vires*) because higher-ranking law prescribes the adoption of a legally non-binding act, but that is a different scenario. It is the content of a legal act or its classification (eg a Regulation instead of a Directive) which may render it disproportion‐ ate, not its legally binding character.791 Also the express reference to the requirement of a 'satisfactory achievement of the objective of the measure' and in particular 'the need for effective enforcement' in para 6 of the (old) Protocol (No 30) on the application of the principles of subsidiarity and proportionality annexed to the TEC (1997) may be brought forward against the view that the principle of proportionality suggested the use of EU soft law instead of law.<sup>792</sup>

Generally, the application of the *a maiore ad minus* approach in the giv‐ en context seems to be plausible. Also the CJEU in its *Grimaldi* judgement held that recommendations 'are generally adopted by the institutions of the Community when they do not have the power under the Treaty to adopt

similarly, but in the context of the principle of proportionality: Schima, Art. 5 EUV, para 73.

<sup>785</sup> Stoa, Subsidiarity 31.

<sup>786</sup> See Hetmeier, Art. 296 AEUV, para 2.

<sup>787</sup> Conclusions of the Presidency of the Edinburgh European Council, 11–12 December 1992, SN 456/1/92 REV 1, 21.

<sup>788</sup> Unclear: case C-643/15 *Slovakia and Hungary v Council*, paras 245 f.

<sup>789</sup> See Article 2 TEU; see also eg case 294/83 *Les Verts*, para 23.

<sup>790</sup> See Schmidt-Aßmann, Verwaltungsrecht 350, with further references; Knauff, Re‐ gelungsverbund 412 f (fn 88); differently: Senden, Soft Law 90, with a further reference; Senden, Rulemaking 64.

<sup>791</sup> The reference to the principle of proportionality in Article 296 para 1 TFEU does not contradict such an interpretation.

<sup>792</sup> That this old Protocol may be relevant for fleshing out the principle of proportional‐ ity even today is confirmed eg by Calliess, Art. 296 AEUV, para 7.

binding measures or *when they consider that it is not appropriate*[ <sup>793</sup>] *to adopt more mandatory rules*' <sup>794</sup> (emphasis added). But there are certainly limits to this approach. One impediment could be Article 296 para 1 TFEU. It provides that '[w]here the Treaties do not specify the type of act to be adopted, the institutions shall select it on a case-by-case basis'.<sup>795</sup> *E contrario* it could be concluded that where the type of act to be adopted *is* specified for a concrete case, the institutions do not have the choice to opt for other acts, not even soft law acts.<sup>796</sup> In this context, the Court held: 'The fact that an institution of the European Union derogates from the legal form laid down by the Treaties constitutes an infringement of essential procedural requirements that is such as to require the annulment of the act concerned, since that derogation is likely to create uncertainty as to the nature of that act or as to the procedure to be followed for its adoption, thereby undermining legal certainty'.<sup>797</sup> *Prima vista*, this speaks against a general application of the *argumentum a maiore ad minus* in our context. As regards the personal scope of Article 296 para 1 TFEU, it is to be assumed that it does not only address 'the institutions' *stricto sensu*, but that it also applies to bodies, offices and agencies. Such a wide understanding is underpinned by the title of the Treaty section under which Article 296 falls: 'Procedures for the adoption of acts and other provisions'.<sup>798</sup>

<sup>793</sup> The question is whether mandatory rules would be 'appropriate', not whether they would be 'proportionate'. The principle of proportionality can, as opined above, not command the use of soft law instead of law. The Court in *Grimaldi* arguably is not referring to the proportionality principle here. Although the principle had been applied by the CJEU even before its explicit incorporation in the TEC and the new TEU (with the Treaty of Maastricht), eg the German ('*kein Anlaß* zu einer zwingenderen Regelung'; emphasis added), French ('*il n'y a pas lieu* d'édicter des règles plus contraignantes'; emphasis added) and Spanish ('*no es oportuno* dictar disposiciones más vinculantes'; emphasis added) versions of this judgement suggest that 'appropriate' is meant in a more general way, pointing at the discretion the institutions have when acting under (then) Community law.

<sup>794</sup> Case C-322/88 *Grimaldi*, para 13.

<sup>795</sup> This selection needs to be taken with care, not arbitrarily. It is the 'necessary' act which is to be taken; see joined cases 8–11/66 *Cimenteries*, 92.

<sup>796</sup> See also Senden, Soft Law 327; critically, but on the basis of a now out-dated version of the Treaties: von Bogdandy/Bast/Arndt, Handlungsformen 115; comparing the Nice and the Lisbon versions of this provision: de Witte, Instruments 96 f.

<sup>797</sup> Case C-687/15 *Commission v Council*, para 44.

<sup>798</sup> This is also argued by some in the case eg of Article 288; see Nettesheim, Art. 288 AEUV, para 72; against such an inclusive view (still with regard to Article 249 TEC): Vogt, Entscheidung 24 f; see also 3.4.2. below.

The exclusion of the *argumentum a maiore ad minus* is made explicit eg in the case of Article 296 para 3 TFEU, which states<sup>799</sup> that the EP and the Council '[w]hen considering draft legislative acts' may not adopt 'acts not provided for by the relevant legislative procedure in the area in question'. This means that acts other than legislative acts<sup>800</sup> may not be concluded in a *legislative* procedure.<sup>801</sup> In the following sub-chapter, we shall examine practically highly important Treaty provisions containing competences of EU actors with a view to whether soft law powers may be implied to them, arguing (mainly) *a maiore ad minus*.

# 3.3.2.2. The (lack of an) *argumentum a maiore* (law) *ad minus* (soft law) in selected Treaty provisions

The first provision which shall be discussed here is Article 294 TFEU. In my view, Article 296 para 3 TFEU does not exclude the adoption of soft law acts pursuant to the procedure laid down in Article 294 TFEU when it is clear from the beginning (and proposed by the Commission) that no (draft) legislative act is negotiated (see in more detail 3.4.3. below). This argument is underpinned by Article 292 TFEU which suggests that the Council may adopt recommendations also where eg a special legislative procedure is prescribed: *argumentum* 'It shall act unanimously in those areas in which unanimity is required for the adoption of a Union act'.<sup>802</sup> If by analogy we apply this finding to the ordinary legislative procedure, the adoption of soft law seems to be allowed also under Article 294 TFEU. The analogous application could, however, be refused with reference to the general power of the Council to adopt recommendations which the EP – the second legislator in the ordinary legislative procedure – lacks. Thus, it could be argued that there is no regulatory gap. Whether the Council would then, in view of its competence granted under Article 292 TFEU, be entitled to adopt a recommendation on its own in areas in which the

<sup>799</sup> On the merely declaratory character of Article 296 para 3 TFEU see Krajewski/Rös‐ slein, Art. 296 AEUV, para 52.

<sup>800</sup> See definition in Article 289 para 3 TFEU.

<sup>801</sup> See Geismann, Art. 296 AEUV, para 5; Krajewski/Rösslein, Art. 296 AEUV, para 52; Vcelouch, Art. 296 AEUV, para 77. Differently: Schoo, Art. 296 AEUV, para 6, arguing that this provision merely repeats the principle of conferral.

<sup>802</sup> For the risks this competence entails (with a view to the draft Constitutional Trea‐ ty): von Bogdandy/Arndt/Bast, Instruments 114.

ordinary legislative procedure is applicable, is unclear. In terms of one of the aims of the Treaty of Lisbon, that is to promote the role of the EP, this would certainly be an odd result. In practice, the EP and the Council have adopted eg recommendations<sup>803</sup> in the course of the co-decision procedure (since the Treaty of Lisbon called 'ordinary legislative procedure').<sup>804</sup>

But it is not always the *argumentum a maiore* which brings to light a competence to adopt soft law where a pertinent competence to adopt hard law is provided for. A material competence to be examined here is Article 106 para 3 TFEU. Here the Commission's task to 'ensure the application of the provisions of this Article' is coupled with its power to 'where necessary, address appropriate directives or decisions to Member States'. The words 'where necessary' indicate that also less intrusive means may be taken, eg the adoption of soft law addressed to the MS.<sup>805</sup> In this case there does not seem to be a regulatory gap, as already the wording hints at a soft law power. When Article 105 para 1 TFEU stipulates that the Commission 'shall ensure the application of the principles laid down in Articles 101 and 102' and, in case of an infringement, 'shall propose appropriate measures to bring it to an end', this leaves open the question of which measures the Commission ought to take.806 Para 2, according to which the Commission shall record the infringement in a reasoned decision '[i]f the infringement is not brought to an end', however, suggests that the Commission should try with less intensive means before. Such less intensive means certainly include soft law measures. Here the Commission's competence to adopt soft law can be deduced *e contrario*: Where a decision constitutes the *ultima ratio*, in principle any less intensive measures may be taken before that. In this context, only soft law acts are available as less intensive measures. Also in this case the (mere) interpretation of the provision results in the confirmation of a soft law power. Thus, there is no room for the application of the *argumentum a maiore ad minus*.

Another example is Article 114 para 1 TFEU. This provision allows the European Parliament and the Council, 'acting in accordance with the ordi‐

<sup>803</sup> See eg Recommendation of the European Parliament and of the Council of 10 July 2001 on mobility within the Community for students, persons undergoing training, volunteers, teachers, and trainers, OJ 2001 L 215/30.

<sup>804</sup> For these questions see 3.4.3.1. below.

<sup>805</sup> Only *prima facie* speaking against this view: case T-116/89 *Prodifarma*, paras 81f.

<sup>806</sup> *Senden* calls this an '"in-between" legal basis', as it obliges to act, but does not determine in which form; Senden, Soft Law 327 f and 337 (with regard to the principle of effectiveness).

nary legislative procedure', to 'adopt the measures for the approximation of the [laws of the MS] which have as their object the establishment and functioning of the internal market'. While the malleable term 'measure' speaks in favour of discretion as regards the choice of form,<sup>807</sup> and while the prescribed compliance with the ordinary legislative procedure can, after what was said above, *a maiore ad minus* also be performed in order to adopt a soft law act (as long as it is not negotiated as 'legislative' act), it seems that the purpose of Article 114 TFEU, the approximation of laws, can hardly be reached by non-binding acts.<sup>808</sup> In spite of these apparent restrictions, the CJEU held that the European Parliament and the Council may, by means of a legislative act, set up a new body entrusted with the power to adopt soft law measures in order to facilitate an approximation of laws.<sup>809</sup> So while the legislator arguably may not itself adopt soft law measures based on Article 114 para 1 TFEU, it may on this basis, confer this power upon the Commission.<sup>810</sup>

This seemingly odd result could be justified by considering the distribu‐ tion of powers among the EU institutions: The EP and the Council have the power to adopt suitable, generally applicable, measures. In order to reach the aim of approximation, they must be legally binding (here: legislative) measures. The Commission and – to a limited extent – the Council or European agencies are in charge of executing legislation, that is to say to ensure compliance with its rules. This task may also be fulfilled by means of the adoption of soft law, be it individual or general in application. This soft law can ensure that the existing legislative rules (approximating the laws of the MS) are correctly applied. It is a consequence of the principal

<sup>807</sup> See case C-217/04 *United Kingdom v European Parliament/Council*, para 43; see also Biervert, Mißbrauch 104 f; Knauff, Regelungsverbund 413, with regard to the principally 'neutral' term 'measure'; Tietje, Art. 114 AEUV, para 115.

<sup>808</sup> See Articles 289 para 1 and 296 para 3 TFEU; see also case C-376/98 *Germany v European Parliament*, para 83; case C-58/08 *Vodafone*, para 35: '[T]he authors of the Treaty intended to confer on the Community *legislature* a discretion' (emphasis added); unclear: M Schröder, Art. 114 AEUV, para 57; Korte, Art. 114 AEUV, paras 74–76.

<sup>809</sup> Case C-217/04 *United Kingdom v European Parliament/Council*, para 44; see also para 28, in which the European Parliament expressly and *in eventu* argues that implied powers conferred by what is now Article 114 TFEU allow it, together with the Council, to create an agency (vested with the power to adopt soft law); see references by Weismann, Agencies 64 (fn 388).

<sup>810</sup> See Goldmann, Gewalt 501, who refuses to qualify such a scenario as 'delegation'; pointing to this problem in a different context: Steiblyté, Delegation 69.

distribution of powers between the institutions – the EU's institutional balance – that the EU's executive branch is exercising powers different from those of its legislative branch. Therefore, to take an important example, it is the Commission (or, exceptionally,<sup>811</sup> the Council in an executive capacity) which must be vested with implementing powers according to Article 291 para 2 TFEU, where uniform conditions for implementing legally binding acts are needed. The legislator is competent to regulate within the respec‐ tive policy field and also to provide for the related powers of the executive, but it may not exercise itself the executive (implementing) powers just mentioned. Therefore we can conclude that primary law provides for a distribution of powers between the institutions, according to which the legislator may vest primarily the Commission (or European agencies<sup>812</sup>) with powers it may not exercise itself.<sup>813</sup>

Another question is whether the power to adopt legally binding (execu‐ tive) acts implies the power to adopt soft law acts. In this context, the two main general provisions in the TFEU are Articles 290 and 291 which allow for the delegation of the power to adopt delegated/implementing acts. While the *telos* ('supplement or amend […] elements of the legislative act') of Article 290 TFEU clearly exclude legally non-binding acts, Article 291 para 1 TFEU non-specifically refers to 'implementing powers'. The latter acts may have a general-abstract or an individual-concrete scope.814 The wording of Article 291 para 2 TFEU allows for a reading, according to which these powers also encompass the power to adopt legally non-binding acts.815 In the literature such a wide interpretation is largely affirmed.816 Un‐

<sup>811</sup> For the Council's general restrictions, pursuant to the Court, on reserving the powers to implement for itself see case 16/88 *Commission v Council*, para 10; case C-257/01 *Commission v Council*, para 51.

<sup>812</sup> See Opinion of AG *Kokott* in case C-217/04 *United Kingdom v European Parlia‐ ment/Council*, fn 30, in which she refers to implied powers in the context of Euro‐ pean agencies and the competences now contained in Article 114 TFEU.

<sup>813</sup> See also Schütze, Rome 1398.

<sup>814</sup> See Bast, Hierarchy 161; Ilgner, Durchführung 276, with further references.

<sup>815</sup> Regulation 182/2011 laying down the rules and general principles concerning mech‐ anisms for control by Member States of the Commission's exercise of implementing powers only refers to 'draft implementing acts' and hence as well leaves open the question whether or not it shall apply only to legally binding acts. For practical examples of soft law acts adopted in the course of comitology procedures see J Scott, Limbo 347.

<sup>816</sup> In favour of an inclusive interpretation: Nettesheim, Art. 291 AEUV, para 27; Ruffert, Art. 288 AEUV, para 11; F Schmidt, Art. 291 AEUV, para 15; Senden/van den Brink, Checks 38; unclear: Kröll, Artikel 290 und 291 AEUV 205 f; see also von Bogdan‐

der the (unofficial) predecessor provisions of Article 291 (and Article 290) TFEU – Article 202 (third indent) and Article 211 (fourth indent) TEC, which would also have allowed for such a wide interpretation – purportedly hardly any legally non-binding acts were adopted.<sup>817</sup> Assuming a certain continuity of Article 291 TFEU, this (historical) practice is an argument against the possibility to delegate to the Commission (or the Council) the power to adopt legally non-binding acts under Article 291 TFEU.

Implementing powers shall enable the Commission (the Council) to set 'uniform conditions for implementing legally binding Union acts'. A need for uniform conditions may be given where there are considerable differen‐ ces in the national law which is adopted/applied in the implementation of the act.<sup>818</sup> It can be doubted that there are cases in which legally non-bind‐ ing acts would be suited best to serve the aim of *unifying* (ie not merely 'coordinating') the conditions for implementation, as allowing for the possi‐ bility to (lawfully) deviate – however effective the respective soft law meas‐ ures may be expected to be – does not serve the aim of unification.819 These doubts also apply in a situation where the Commission (or the Council), within the ambit of its implementing powers, adopts recommendations not on the basis of its respective implementing powers (*lex specialis*), but on the basis of its general competence to adopt recommendations pursuant to Article 292 TFEU (*lex generalis*; for that competence clause see in more detail 3.4.3. below).

The Court in principle seems to allow for the adoption of general-ab‐ stract soft law as an expression of implementing powers.820 If this is ac‐

dy/Bast/Arndt, Handlungsformen 115; emphasising the similarity as 'funktionelles Pendant' [also referred to as 'funktionales Pendant'; functional pendant] of the ESAs' guidelines and recommendations as compared to acts pursuant to Articles 290 and 291 TFEU, but emphasising that when ESA guidelines and recommenda‐ tions are adopted the procedural requirements of these provisions are not met: Wörner, Verhaltenssteuerungsformen 233 ff; against the inclusion of soft law acts: Craig, Comitology 199; Möstl, Rechtsetzungen 1082.

<sup>817</sup> See Ilgner, Durchführung 29.

<sup>818</sup> See Kröll, Artikel 290 und 921 AEUV 205.

<sup>819</sup> In its earlier case law the Court seems to have argued that way: case 74/69 *Hauptzol‐ lamt Bremen-Freihafen*, para 9; see also Eliantonio, Soft Law 497, with a further reference; but see case C-35/93 *Develop Dr. Eisbein*, para 21; case C-259/97 *Clees*, para 12; case C-396/02 *DFDS*, para 28, each with regard to the Explanatory Notes to the nomenclature of the Customs Cooperation Council; for a similar discussion in the context of harmonisation see 5.1. below.

<sup>820</sup> Case C-355/10 *European Parliament v Council*, paras 80–82. Even though the Court eventually qualifies the rules at isse as 'intended to produce binding legal effects',

cepted, then also the principal lawfulness of adopting individual-concrete legally binding measures on the basis of Article 291 TFEU (see above) must be extended to (individualised) soft law acts.

One of the rare recommendations of the CJEU – and this shall form our last example here – is its output called 'Recommendations to national courts and tribunals in relation to the initiation of preliminary ruling proceedings' (2019). These recommendations contain 'practical guidance', among other things, on the national court's or tribunal's decision to make a reference, on the communication with the CJEU, and on urgent referen‐ ces.<sup>821</sup> While the recommendations do not explicitly refer to a legal basis for their adoption, in para 2 it is stated that they are intended 'to clarify the provisions of the rules of procedure' of the CJEU which are, in particular, based on Article 19 TEU and Article 253 TFEU. While the former provision remains silent on that issue, Article 253 TFEU expressly states that the Court shall establish its Rules of Procedure. It could be argued that the Court's competence to provide soft guidance on preliminary reference procedures (which are regulated in Part III of its Rules of Procedure) can – *a maiore ad minus* – be deduced from its competence to adopt Rules of Procedure. In the view of some, the Rules of Procedure of the CJEU con‐ stitute internal law,822 following which the *argumentum a maiore* cannot convincingly be applied to deduce the power to adopt external soft law (for the distinction between internal and external soft law see 3.3.3.1. below). While soft law is a *minus* as compared to a legally binding act, externality constitutes a *maius* over a merely internal act. In my view, the Court's Rules of Procedure are not entirely internal, but some provisions are intended to and actually have a strong external radiance, in particular those concerning party rights, and also those concerning references for a preliminary ruling (Articles 93 ff *leg cit*). On the basis of this assumption, the deduction of the Court's power to render the said recommendations *a maiore ad minus* appears to be legally flawless.

An alternative legal solution could be to deduce the Court's soft law pow‐ er at issue, again *a maiore ad minus*, from its power to decide in a legally

it appears to approve of the possibility to adopt implementing soft law; in the affirmative also Gärditz, Unionsrecht, para 14.

<sup>821</sup> Note that the predecessor recommendations from 2012 in their para I.6. were more explicit in terms of their non-bindingness ('in no way binding') and its purposes ('to supplement […] the Rules of Procedure').

<sup>822</sup> See Senden, Soft Law 53.

binding way on preliminary references – not only substantially,<sup>823</sup> but also eg as regards their admissibility. Exceptionally, the Court may even – in the course of a preliminary reference procedure initiated in different national procedures, in particular in the context of State liability proceedings, or in the course of a Treaty infringement procedure – come to declare unlawful the omission (withdrawal) of a preliminary reference by a national court or tribunal.<sup>824</sup> It is to be noted, though, that the Court is allowed, under Article 267 TFEU, to answer questions related to individual cases, but not to rule in a general-abstract way. This situation is comparable to the Commission's adoption of general-abstract soft law in order to facilitate its (individual-concrete) execution of State aid policy – in which case, as was set out under 3.2.3. above, the Court refused claims of unlawfulness.

3.3.3. Internal soft law

3.3.3.1. The phenomenon of internal soft law

So far we have mainly dealt with external EU soft law, which shall – due to its eminent importance – stay in the foreground in this work. In this sub-chapter, however, special attention shall be drawn to *internal* EU soft law. Literally no modern bureaucracy of a certain size, be it based on the rule of law or not, can forgo the possibility to harmonise its decision-mak‐ ing practice by internal (soft) regulation.825 Internal acts may be adopted at different levels of the internal hierarchy of an EU body, eg the Commission. Whereas 'decisions of principle'826 can only be adopted by the college of Commissioners or, exceptionally, by the competent Commissioner alone, the adoption of management or administrative measures may be delegated to the competent Directors-General and Heads of Department.827 These measures can also take the form of soft law.

<sup>823</sup> Note, however, that the Court has refused its competence to adopt a legally nonbinding reply to a *concrete* preliminary reference request by a national court or tribunal; case C-62/14 *Gauweiler*, para 12, with a further reference.

<sup>824</sup> See case C-224/01 *Köbler*, paras 117 f.

<sup>825</sup> See Rawlings, Soft law 217 ff, also pointing at the downside of an overboarding use of (internal) soft law; for an example of national – German – internal soft law, its theoretical classification and its application in practice see Arndt, Sinn 54–60.

<sup>826</sup> Case 5/85 *AKZO*, para 37.

<sup>827</sup> See overview in Article 4 of the Rules of Procedure of the Commission, C(2000) 3614, and the provisions referred to therein. For an example of a 'management

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

While external soft law is mainly addressed to persons external to the originator(s), internal soft law is *meant* for internal use only. However, the fact that an act is internal does not mean that it may not have, indirectly, also some external effects. The application of an internal act regularly *affects* external actors in one or the other way.<sup>828</sup> Sometimes classic internal acts, like rules of procedure, may even contain provisions which are addressed not only to the officials of its originator but also to third parties, thereby developing an external dimension, eg provisions concretising the rights of parties to an administrative procedure.<sup>829</sup> Therefore, *prima facie* internal soft law may move, as *Rawlings* has put it, 'along an[] axis, from internal operational advice to guidance for the regulated and the public'.<sup>830</sup> The (external) publication of these acts hints at a broader than a merely internal audience.<sup>831</sup> Internal soft law acts – belonging to the 'interne Verwaltung' [internal administration]832 – regularly provide for an internal procedure, eg with regard to applications for access to documents according to Regu‐ lation 1049/2001.833 Some of these instruments are 'governing the exercise of the discretion conferred on the Commission' (or other EU bodies).<sup>834</sup> Often they merely summarise the approach to be taken by the body's staff and hence recapitulate, or restate, the respective EU law (including its case law).835 If (parts of ) these documents entirely lack any normative content,

832 Priebe, Aufgaben 75.

measure' see case 5/85 *AKZO*, para 38; see also Eekhoff, Verbundaufsicht 123; Röhl, Entscheidung 340, both with references to further case law.

<sup>828</sup> See H Adam, Mitteilungen 155; Goldmann, Gewalt 364 f; Hofmann/Rowe/Türk, Administrative Law 97, 538 and 550; Senden, Soft Law 315 f, with further references.

<sup>829</sup> For the example of the CJEU's Rules of Procedure see 3.3.2.2. above. For the officials as addressees of such acts on the one hand, and market participants who are only exceptionally directly addressed, but for whom these acts are of pivotal importance and who largely comply with them, on the other hand see Arndt, Sinn 56–60.

<sup>830</sup> Rawlings, Soft law 224; see also case T-339/04 *France Télécom*, para 83.

<sup>831</sup> In the context of the World Bank's 'internal' guidance see Boisson de Chazournes, Guidance 284.

<sup>833</sup> Note, however, that (possibly in addition to that) also a legally binding act with an also external scope may be adopted: see eg Decision of the Steering Committee of the Research Executive Agency on the Implementation of Regulation (EC) N° 1049/2001 of the European Parliament and the Council regarding Public Access to Documents REA/SC(2008)4 rev.1; see also Schwarze, Soft Law 244 f.

<sup>834</sup> Senden, Balance 89, with reference to the pertinent case law (on staff matters).

<sup>835</sup> For the effect such 'summaries' may have see Georgieva, Soft Law 244; Knauff, Regelungsverbund 325 f; Ştefan, Soft Law 101f (see also 103, with regard to the Court's reference to such acts), each with further references; for the importance of such restatements see Jansen, Methoden 48; even confirming a *quasi*-normative

they cannot be called soft law (descriptive acts or parts of acts).<sup>836</sup> Where additional 'fine-tuning' is provided – in the sense that an autonomous administrative practice, of course on the basis of and (regularly<sup>837</sup>) in accordance with the relevant rules, is laid down in a legally non-binding way – the soft law character of such an act (or part of it) can be affirmed (prescriptive acts or parts of acts).<sup>838</sup> Where internal rules turn out to be (general) instructions proper, they are legally binding upon their addressees and hence cannot be called soft law.<sup>839</sup> Where the 'explanation' provided in such rules goes beyond the regulatory content of the underlying act, the Court has to annul this 'explanation' – thereby at least implicitly confirming its legal effects.<sup>840</sup> In practice, it can be very difficult to determine whether a *prima facie* rule actually is a rule or whether it is merely the repetition of a rule laid down elsewhere – or whether it contains elements of both.<sup>841</sup> Some

character of such acts: Meier, Mitteilung 1307; for the factual compliance with these 'communications' see Köndgen, Rechtsquellen, para 64.

<sup>836</sup> See eg para 44 of the internal Antitrust Manual of Procedures of the Commission DG Competition (November 2019) <https://op.europa.eu/en/publication-detail/-/ publication/d7d7e463-ac51-11ea-bb7a-01aa75ed71a1> accessed 28 March 2023; see also della Cananea, Administration 69; Hofmann/Rowe/Türk, Administrative Law 539 f; for the Court's approach see case C‑362/08P *Hilfsfonds*, para 34, with further references.

<sup>837</sup> For the Commission's 'proposal' of a legal understanding deviating from that of the Court see the example of Commission Communication 'Guidance on the Com‐ mission's enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings', 2009/C 45/02; for a critique of this Guidance see Gormsen, Commission.

<sup>838</sup> See Raschauer, Verhaltenssteuerungen 699; Weigt, Rechtsetzung 49, with regard to the Commission Interpretative Communication on the Community law applicable to contract awards not or not fully subject to the provisions of the Public Procure‐ ment Directives. An action for annulment against this Communication was later declared inadmissible by the then Court of First Instance; case T-258/06 *Germany v Commission*.

<sup>839</sup> See Hofmann/Rowe/Türk, Administrative Law 572; for the effect of instructions see also 4.2.3.1.

<sup>840</sup> See case C-366/88 *France v Commission*, paras 23 f; critically pointing at the possi‐ bility of expanding regulation by soft law: Korkea-aho, Soft Law 276.

<sup>841</sup> For the frequent overlap between summary and actual (soft) rule-making see Sen‐ den, Soft Law 140 f; see also case T-81/97 *Regione Toscana*, para 22. The problem of repetition of norms has raised the attention of the legislator. With regard to Reg‐ ulations 1093–1095/2010, the legislator introduced a new provision, Article 16 para 2a, obliging the ESAs in the following way: 'Guidelines and recommendations shall not merely refer to, or reproduce, elements of legislative acts. Before issuing a new guideline or recommendation, the Authority shall first review existing guidelines and recommendations, in order to avoid any duplication'.

of these (soft) internal rules are published online,<sup>842</sup> but many of them are not.

While it was said that soft law may entail a self-obligation (upon its originator),843 the addressees of soft law cannot be legally bound. The selfobliging potential of *internal* soft law requires some further considerations. This effect requires a certain publicity of the act at issue, ie some external outreach. If no person outside the body at issue can become aware of the rules, their trust in the actual application of these rules cannot possibly be disappointed. It is the originator which addresses the soft law rules to its staff. As addressees, the staff cannot be legally bound directly by this soft law. Where internal rules are – failing publication – not accessible to third parties, deviance from them, for lack of information, cannot possibly be invoked by the latter. In case of such truly internal soft law, the staff (as addressees) are not bound by it, but the originator is bound to the extent that it may not – eg in disciplinary proceedings – reproach one of its officials with having complied with the act at issue (eg because it has later turned out to be unlawful).844 To the extent the internal soft law act has an external outreach, the originator may be obliged to comply with its soft law *vis-à-vis* a third party. The staff are then obliged to comply

<sup>842</sup> Eg the Internal Guidelines on the new impact assessment procedure developed for the Commission services <http://ec.europa.eu/governance/docs/comm\_impact \_en.pdf>, or the internal Antitrust Manual of Procedures of the Commission DG Competition (November 2019) <https://op.europa.eu/en/publication-detail/-/pu blication/d7d7e463-ac51-11ea-bb7a-01aa75ed71a1>, both accessed 28 March 2023. The Manual is characterised as an 'internal working tool intended to give practical guidance to staff on how to conduct an investigation applying Articles 101 and 102 TFEU'. It 'does not contain binding instructions for staff'. In the context of the publication of EU soft law, *Snyder* has coined the term 'regulation by publication': Snyder, Practice 2; for the meaning of the publication of norms in legal history see Jansen, Methoden 37 f; with regard to EU State aid law see Schweda, Principles, para 30.

<sup>843</sup> See 4.2.3.1. below. However, such self-obligation can also be excluded; see the Com‐ mission's *Handbook on Implementation of the Services Directive* – which explicitly excludes binding effects on the Commission (page 1) – as an example <https://f ve.org/cms/wp-content/uploads/handbook-on-the-impl-of-the-Services-Direct ive\_en.pdf> accessed 28 March 2023; for the Court's case law see case T‑185/05 *Italy v Commission*, para 47; see also Pampel, Rechtsnatur 55–64; von Graevenitz, Mitteilungen 172, both with further references; for the self-obliging effect of soft law in public international law see Schermers/Blokker, Institutional Law, §§ 1241 and 1261.

<sup>844</sup> For the invocation of illegality by members of staff themselves see case C-171/00P *Libéros*, para 35.

with the (affected parts of the) act indirectly, due to their acting on behalf of their employer and due to their duty to consider, when doing their job, the obligations of their employer.<sup>845</sup> Consequentially, an 'unmotivated' deviation from such soft law act by the staff (legally acting on behalf of the originator) may be qualified as an infringement of legitimate expectations or the principle of equal treatment.<sup>846</sup>

# 3.3.3.2. The competence to adopt internal soft law

The EU's *Verbandskompetenz* to organise itself, ie in particular to change its current internal organisation in a certain way, is subject to the princi‐ ple of conferred powers.<sup>847</sup> In general, this applies also to the respective *Organkompetenz*. An institution, body, office or agency of the EU may organise itself on the basis of an express competence, eg the Council's competence to 'decide on the organisation of [its] General Secretariat' (Ar‐ ticle 240 para 2, second sentence TFEU), the Commission's (President's) competences to adopt its Rules of Procedure 'so as to ensure that both it and its departments operate' (Article 249 para 1 TFEU), to 'lay down guidelines within which the Commission is to work'848 and to 'decide on the internal organisation of the Commission' (Article 17 para 6 lit a and b TEU), or the competence of the Management Board of the EASA to adopt its (the Board's) Rules of Procedure (Article 98 para 2 lit j of Regulation 2018/1139).849 That the CJEU refers to the power of 'the Bureau [of the

<sup>845</sup> See Article 11 para 1 of the Staff Regulations of EU Officials, according to which an official shall 'carry out the duties assigned to him objectively, impartially and in keeping with his *duty of loyalty* to the Communities' (emphasis added); for the case that soft law, internally, conveys 'internal administrative instructions under the principle of hierarchy' see Arroyo Jiménez, Bindingness 30 f.

<sup>846</sup> See Hofmann/Rowe/Türk, Administrative Law 542; see also 4.2.3.2.3. below.

<sup>847</sup> See Streinz, Art. 13 EUV, para 28.

<sup>848</sup> See Hofmann/Rowe/Türk, Administrative Law 565, according to whom this provi‐ sion covers 'a wide range of internal organizational measures'.

<sup>849</sup> See Priebe, Entscheidungsbefugnisse 88, according to whom the competence to adopt Rules of Procedure only allows for a regulation of issues internal to the respective administration; see – on the contrary – the widely-drafted Article 20 of the Commission's Rules of Procedure, according to which '[t]he Commission may, in special cases, set up specific structures to deal with particular matters and shall determine their responsibilities and method of operation'; for the competence of European agencies to adopt their respective rules of procedure see Orator, Möglich‐ keiten 408 f.

General Assembly, the predecessor of the EP] […] to organise its Secretariat as it wishes and in the interests of the service, and […] [to its acting] in the full exercise of its powers in abolishing a post which it considered unneces‐ sary'<sup>850</sup> without mentioning an express Treaty base speaks in favour of a generally broad discretion of EU institutions, bodies, offices and agencies as concerns their respective self-organisation.<sup>851</sup> The Court's judgement in the *Macevicius* case points in a similar direction when it stresses the principal 'freedom of the Community institutions to organise their internal work in the best interests of the service'.<sup>852</sup> The Court in the above and other cases remains silent on the question of legal basis.<sup>853</sup> An explanation for this would be that it applies the implied powers doctrine in these cases (see 3.3.1. above). The Assembly's/European Parliament's<sup>854</sup> right to self-organisation, for example, could be perceived as power implied to its (express) right to adopt its own Rules of Procedure according to Article 25 TECSC (in the *Kergall* case) and Article 142 TEEC (in the *Macevicius* case), respectively. This approach seems to be confirmed by the later case law of the Court.<sup>855</sup>

854 Up until the Treaty of Maastricht, the Treaty name was 'Assembly'. However, start‐ ing already in the 50s, the Assembly referred to itself as 'European Parliament', and so have done – to an increasing extent – the other institutions, the MS etc.

855 See eg case 230/81 *Luxembourg* v *European Parliament*, para 38. In this case the Court bases the EP's right to determine its internal organisation on its express Treaty competence to adopt its Rules of Procedure; confirmed in case 149/85 *Wybot*,

<sup>850</sup> Case 1/55 *Kergall*, para 7 lit b.

<sup>851</sup> See joined cases C-237/11 and C-238/11 *France v European Parliament*, para 42, with regard to the duty of other actors to respect this discretion; see also case C-301/02P *Tralli*, para 39, with regard to institutions and bodies established under primary law, and para 42 with reference to para 34 of the judgement in case C-409/02P *Pflugradt*. In the latter judgement the Court stated that the ECB is 'a Community body, entrusted with public interest responsibilities and authorised to lay down, by regulation, provisions applicable to its staff'. In *Tralli*, the Court converted this into the following legal deduction: '[A] Community body entrusted with public interest responsibilities *is* authorised to lay down, by regulation, provisions applicable to its staff' (emphasis added); see also Hummer, Interorganvereinbarungen 85.

<sup>852</sup> Case 66/75 *Macevicius*, para 17.

<sup>853</sup> Without reference to a specific competence clause, the CJEU has furthermore con‐ firmed the Court of Auditors' right to lay down 'in a general internal decision rules governing the exercise of the discretion which it has under the Staff Regulations'; case 146/84 *De Santis*, para 11; see also case C-443/97 *Spain v Commission*, para 31, with regard to the Commission; similarly, and with further references, case T-2/90 *Ferreira de Freitas*, para 61; case T-185/05 *Italy v Commission*, para 45, both with regard to the Commission; see also Gärditz, Unionsrecht, para 57, with further references.

In view of this wide discretion on the way in which self-organisation takes place,<sup>856</sup> the power to adopt (soft) internal guidelines on the interpre‐ tation of applicable legal provisions or on internal procedures – which are also forms of (internal) self-organisation – can be implied to the general right eg of Commissioners to give instructions to the departments assigned to them (which is laid down in the Commission's Rules of Procedure)<sup>857</sup> and of superior officials to give instructions to subordinates according to the Staff Regulations.<sup>858</sup> The power to give binding instructions encompass‐ es, arguing *a maiore ad minus*, also the – less intrusive – power to adopt (general) soft law instructions,<sup>859</sup> such as the above mentioned Antitrust Manual of Procedures of the Commission.<sup>860</sup>

In conclusion, we can say that the competence requirements for the adoption of internal soft law are not principally different from those for the adoption of external soft law. It appears, however, that with the right to self-organisation EU bodies dispose of a general right to regulate their internal organisation, *a maiore ad minus* also by means of soft law, and hence finding a competence to adopt internal soft law regularly is less demanding than searching for a competence to adopt external soft law.

para 16; see also joined cases C-237/11 and C-238/11 *France v European Parliament*, para 42, in which the Court stresses 'the Parliament's power to determine its own internal organisation'; see also the case law referred to by H Hofmann, Rule-Making 162 f; see Nettesheim, Art. 13 EUV, para 67, who acknowledges an *unwritten* power of self-organisation where an explicit competence is lacking in primary/secondary law.

<sup>856</sup> Similarly: Hofmann/Rowe/Türk, Administrative Law 538 f; critically: ibid 543 f.

<sup>857</sup> Article 16 para 2 of the Rules of Procedure of the Commission.

<sup>858</sup> Article 21 para 2 of Regulation 31 (EEC), 11 (EAEC) (Staff Regulations of Officials and the Conditions of Employment of Other Servants of the EEC and the EAEC).

<sup>859</sup> See Hofmann/Rowe/Türk, Administrative Law 542.

<sup>860</sup> The legislator appears to share this view when in Council Regulation 139/2004 (EC Merger Regulation) it mentions the need for the Commission to 'publish guidance' – in my view a clear reference to soft law – only in its Recitals (namely Recitals 28 f ), without at the same time conferring an according competence.

# 3.4. General competence clauses in the Treaties

# 3.4.1. Introduction

Recommendations and opinions are the two (potential<sup>861</sup>) EU soft law acts laid down in the Treaties' catalogue of legal acts, ie Article 288 TFEU. In this sub-chapter, general competence clauses to adopt either recommenda‐ tions or opinions shall be addressed. Competences to adopt EU soft law bearing other names than 'recommendation' or 'opinion' are dealt with under 3.6. below. A general competence here is understood as a competence which is not limited to concrete (specific) situations ('special competences'), but which applies across the board. As to be expected in a legal regime based on the rule of law, general competences are not limitless, though, but subject to varying restrictions. In view of this, the term 'general competen‐ ce' is not to be understood as all-encompassing, but as 'general competence subject to limitations'. The 'special competences' to adopt recommenda‐ tions or opinions are subject to sub-chapter 3.5. below.

In the following, the *prima facie* general competence clauses Article 288 TFEU, Article 292 TFEU, Article 127 para 4 and Article 132 para 1 (third indent) TFEU, and various Treaty provisions empowering committees shall be analysed, also with a view to their (potential) limits (3.4.6. below).

# 3.4.2. Article 288 TFEU – a general competence clause?

When talking about *prima facie* general competences of EU institutions, bodies, offices or agencies to adopt soft law, we ought to have a closer look at Article 288 TFEU which lays down the 'legal acts of the Union' which its 'institutions shall adopt'. Its predecessor, Article 249 TEC, began with the words: '[i]n order to carry out their task and in accordance with the provi‐ sions of this Treaty'. In Article 288 TFEU this wording is replaced by: '[t]o exercise the Union's competences'. This is interpreted as a way to expressly include the second Treaty, the TEU, and the competences provided for therein.<sup>862</sup> Moreover, emphasis is laid on the Union's *Verbandskompetenz*, rather than on the competences of the single institutions. This is to indicate

<sup>861</sup> As was noted under 3.1.1. above, acts bearing either of these names do not necessar‐ ily contain (soft) norms.

<sup>862</sup> See Geismann, Art. 288 AEUV, para 16; Nettesheim, Art. 288 AEUV, para 3; see also Biervert, Art. 288 AEUV, paras 2 f.

that all the competences of the EU shall be exercised by the institutions, a mismatch of competences and tasks<sup>863</sup> being excluded. Article 288 TFEU cannot be regarded as a competence clause, though.<sup>864</sup> It neither provides a legal competence of the EU as a legal person (*Verbandskompetenz*) nor of any of its bodies (*Organkompetenz*) to adopt certain legal acts. It rather has a (non-exhaustive<sup>865</sup>) declaratory and systematising character and, in addi‐ tion to that, fleshes out Article 296 para 1 TFEU which grants discretion to the institutions: where they have an unspecific competence to act, they can choose the most appropriate type of act (in accordance in particular with the principle of proportionality).<sup>866</sup> From Article 288 TFEU, a body's power to adopt a certain act can only be deduced in conjunction with a (substantive) competence clause (*argumentum* '[t]o exercise the Union's competences'). Thereby Article 288 TFEU may also have to be teleologically reduced, for instance in order to explain why the Court does not have a power to adopt directives.

Article 288 TFEU in its para 1 stipulates that 'the institutions' of the EU shall 'adopt regulations, directives, decisions, recommendations and opinions'. The institutions of the EU are taxatively enumerated in Article 13 para 1 TEU. The question is whether also here the term 'institutions' is to be understood widely so as to encompass bodies, offices and agencies of the EU, as well.867 In view of the long-lasting practice to equip, mostly by means of secondary law, other EU bodies than institutions with the power to adopt (certain868) acts listed in Article 288,869 it is to be affirmed.870 In conclusion, we can state that this provision lists – in a non-exhaustive way – the acts (some or all of ) which may in principle be adopted, in accordance

<sup>863</sup> See Articles 2 f TEC (Nice); see also Nettesheim, Art. 288 AEUV, para 3.

<sup>864</sup> See Nettesheim, Art. 288 AEUV, para 15; Ruffert, Art. 288 AEUV, para 6; Schroeder, Art. 288 AEUV, para 10; Senden, Soft Law 295.

<sup>865</sup> For further, so-called *sui generis* acts in EU law see de Witte, Instruments 81–83 (Nice) and 100–102 (Lisbon); Pampel, Rechtsnatur 85–90; Ruffert, Art. 288 AEUV, paras 100 ff.

<sup>866</sup> See Ruffert, Art. 288 AEUV, para 15; see also case C‑322/88 *Grimaldi*, para 13.

<sup>867</sup> See 3.2.2. above.

<sup>868</sup> Bodies, offices and agencies cannot be entrusted with the power to adopt *all* of the acts mentioned in Article 288 TFEU. Especially their (partial) empowerment to adopt acts with a general-abstract scope ('rule-making') is highly contested; for this debate see Weismann, Agencies 34 f, with further references.

<sup>869</sup> This is acknowledged also elsewhere in primary law: eg Articles 267 lit b and 277 TFEU.

<sup>870</sup> Applying such an inclusive understanding as well: Fischer-Appelt, Agenturen 118 f; Schroeder, Art. 288 AEUV, para 14.

with other primary law, by the institutions, bodies, offices and agencies of the EU, and that it cannot be qualified as a competence clause.

### 3.4.3. Article 292 TFEU

# 3.4.3.1. The power to adopt recommendations of the Council and of the Council and the EP, respectively

The first sentence of Article 292 TFEU reads: 'The Council shall adopt recommendations'. This means a general competence of the Council to adopt acts belonging to one important category of EU soft law, namely recommendations.<sup>871</sup> The two subsequent sentences also regard the Coun‐ cil's competence to adopt recommendations. These are general procedural requirements which are mentioned in Article 292 TFEU in order to clarify that they apply also to the adoption of Council recommendations, namely the requirements of a Commission proposal or the requirement of unanim‐ ity872 in the Council.873 These procedural rules also have a competence impact in that they suggest – in spite of the more liberal wording – that the Council may, according to Article 292 TFEU,874 adopt recommendations only in the field of the Council's responsibilities.875 The requirements for the degree of determination of competence clauses which may be deduced from the principle of conferred powers and the rule of law, respectively, are comparatively low to the extent that they concern the adoption of soft law.<sup>876</sup> Affirming a competence to adopt soft law beyond the actors' responsibilities laid down in the Treaties would run counter to the principle of conferred powers, though.877 It is to be noted, as well, that Article 292

<sup>871</sup> Against an extension of this competence to all kinds of soft law acts: Wörner, Verhaltenssteuerungsformen 282.

<sup>872</sup> For lack of an explicit rule, the Council decides by qualified majority (Article 16 para 3 TFEU). This applies also to recommendations. The Council's decision-mak‐ ing by simple majority, under primary law, does not seem to play a role in the context of recommendations.

<sup>873</sup> For examples in the Treaties see Ruffert, Art. 292 AEUV, para 3.

<sup>874</sup> Elsewhere in the Treaties the Council may be granted a specific power to adopt recommendations; see eg Article 319 para 1 TFEU; see also Nettesheim, Art. 292 AEUV, para 14, with further examples.

<sup>875</sup> See also Geismann, Art. 292 AEUV, para 2 with further references.

<sup>876</sup> See Braams, Koordinierung 194 f, with further references.

<sup>877</sup> Too wide an understanding of the Council's and, with a view to Article 292 TFEU (last sentence), the Commission's competence to adopt recommendations would

TFEU does not as such exclude the possibility of the mentioned institutions or other EU bodies disposing of soft law powers on other legal bases (no exclusionary effect).<sup>878</sup>

The Council may adopt recommendations where it is competent to act in an unspecified way (eg 'the Council shall adopt measures […]').<sup>879</sup> This is clear from the wording of Article 292 in conjunction with Article 296 para 1 TFEU. In addition to that, the Council may – concluding *a maiore ad mi‐ nus* – in principle adopt recommendations where it has the power to adopt law. For *Nettesheim*, where the Treaties do not provide for a specific compe‐ tence to do so, the Council's power to adopt (external) recommendations seems to be limited to cases in which the Council ought to act according to the ordinary or a special legislative procedure. <sup>880</sup> This, he argues, follows from sentences two and three of Article 292. In my view, the wording of Article 292 does not exclude other cases<sup>881</sup> – after all, it generally demands compliance with procedural requirements 'where the Treaties [so] provide'. In practice, the author would agree, though, that it is mainly competence clauses providing for legislative procedures which are applicable in this context, namely special legislative procedures (for the ordinary legislative procedure see below). In most cases in which the Council may act other than according to a (special) legislative procedure, a recommendation ap‐ pears to be inadequate.882 One exception is Article 46 para 6 TEU, pursuant to which '[t]he decisions and recommendations of the Council within the framework of permanent structured cooperation, other than those provided for in paragraphs 2 to 5, shall be adopted by unanimity'. Article 46 in its other paragraphs only mentions Council decisions to be adopted in a procedure other than a legislative procedure, but does not mention any

distort the allocation of powers to the different actors of the EU and hence would also jeopardise the orderly application of the principle of mutual sincere coopera‐ tion between the institutions, bodies, offices and agencies of the EU (see 3.4.6. below).

<sup>878</sup> See Dickschen, Empfehlungen 26–28.

<sup>879</sup> Eg in Article 322 para 2 or in Article 331 para 1 TFEU. See eg the Council Rec‐ ommendation on a Quality Framework for Traineeships, 2014/C 88/01, which is expressly based on Article 292 TFEU in conjunction with Articles 153 and 166 TFEU.

<sup>880</sup> See Nettesheim, Art. 292 AEUV, para 9.

<sup>881</sup> See also Biervert, Art. 292 AEUV, para 2.

<sup>882</sup> See eg Articles 155 para 2 or 207 para 4 subpara 2 TFEU, according to which the Council shall conclude agreements, or Article 108 para 2 subpara 3 TFEU, according to which the Council shall decide upon a concrete case.

(specific) recommendations the Council may adopt; neither does Article 42 para 6 TEU which refers to this permanent structured cooperation. Against this background, it can be concluded that Article 46 para 6 TEU is to be read in conjunction with Article 292 TFEU, with the result that the Council under Article 46 TEU may adopt recommendations whenever it is competent to adopt decisions. This view is also compatible with Article 3 of the pertinent Protocol on permanent structured cooperation, according to which the assessments of the European Defence Agency (EDA) **'**may serve as a basis for Council recommendations and decisions adopted in accordance with Article 46 [TEU]'.<sup>883</sup>

As regards the competence to act according to the ordinary legislative procedure, on the basis of which acts are adopted jointly by the Council and the European Parliament, we need to clarify whether the EP has a competence to adopt recommendations in the first place. From Article 292 TFEU such a competence of the EP cannot be deduced. Following *e contrario* from this provision (which also empowers the Commission and at least refers to the powers of the ECB), it could even be argued that the EP shall not have a general power to adopt recommendations. However, this argument does not appear to be convincing for the following reasons. The Court has confirmed a 'right inherent in the Parliament to discuss any question concerning the Communities, to adopt resolutions on such questions and to invite the Governments to act'.884 It is dubitable, however, whether the Court here was referring to concrete rule-making by means of soft law or whether it alluded to more broadly drafted expressions of opin‐ ion (the typical content of resolutions) which – for lack of determination or for an outright lack of normativity – most of the time do not qualify as soft law (see II.2.1.2. above). In view of the subordinate, largely only consultative role the EP played in EEC decision-making at the time the quoted judge‐ ment was rendered (ie in 1983),885 it is more likely that the Court had in mind the latter kind of acts. While this judgement does not seem to confirm the EP's power to adopt recommendations, it is to be emphasised that the EP today, apart from the Council, is the most important legislative body of the EU, legislation being the highest-ranking and most far-reaching nor‐

<sup>883</sup> Protocol No 10 on permanent structured cooperation established by Article 42 of the Treaty on European Union.

<sup>884</sup> Case 230/81 *Luxembourg v European Parliament*, para 39.

<sup>885</sup> For the only gradual empowerment of the EP since the 70s see Hix/Høyland, Empowerment 172 f.

mative output conferred under the regime of the EU Treaties.<sup>886</sup> Excluding the EP here would also mean impeding the Council from adopting recom‐ mendations following the procedure laid down in Article 294 TFEU (the 'ordinary legislative procedure'). On the contrary, allowing the Council to adopt recommendations on its own regarding a question which the EP and the Council have the power to legislate on, would constitute an unlawful ousting of the EP and its legislative power.<sup>887</sup> It should also be mentioned – as a piece of information on rule-making in practice – that in the past the Council and the EP have adopted a number of recommendations according to Article 294 TFEU (and its predecessors).<sup>888</sup>

On the basis of the above legal considerations, the author concludes that also in case of the EP a competence to adopt recommendations may be deduced *a maiore ad minus* from its ordinary legislative competences. This applies where the act to be adopted is specified and, even more so, where this is not the case.889 In other words: Where the EP (together with the Council) is allowed to adopt law following the ordinary legislative procedure, it may – *ceteris paribus*, in particular that means: together with the Council, on a proposal from the Commission and by meeting the respective majority requirements890 – also adopt less, namely non-law (more specifically: a recommendation). For the Council the applicability of the procedural requirements is clear because it is explicitly obliged by Article 292 TFEU, but – *per analogiam*891 – also the EP has to apply the

<sup>886</sup> The MS' competence to amend this regime, ie to adopt primary law, is superior, it is true. However, it is not conferred by the Treaties but roots in public international law and is only specified in Article 48 TEU.

<sup>887</sup> For the risks of the adoption of soft law ousting law-making competences in the EU legal order see Brunessen, Effets 29.

<sup>888</sup> See eg the Recommendation of 18 June 2009 on the establishment of a European Quality Assurance Reference Framework for Vocational Education and Training, 2009/C 155/01.

<sup>889</sup> Note that also then the acts available to the EP and the Council – strictly speaking – are limited to legislative acts, namely regulations, directives and decisions (*argu‐ mentum* 'ordinary *legislative* procedure'; Article 289 paras 1f TFEU). This is why also in these cases the *argumentum a maiore ad minus* needs to be applied in order to confirm the legislators' soft law power; sceptically with regard to this practice: Härtel, Rechtsetzung 274.

<sup>890</sup> See Article 11 paras 4 f of the Council's Rules of Procedure; also in the EP's Rules of Procedure a specific rule for the vote on recommendations is missing.

<sup>891</sup> Note that Article 292 TFEU is applied *per analogiam* only with regard to the procedural requirements, not to create the EP's power to adopt recommendations in the first place.

process underlying the 'ordinary legislative procedure'. A draft legislative act may not – in the course of a legislative procedure – be transformed into a (draft) soft law act (see Article 296 para 3 TFEU). This provision prohibits the adoption of other than legislative acts following a *legislative* procedure. Arguably, it is the determination 'legislative' which Article 296 para 3 – for reasons of clarity – intends to prevent in the context of the adoption of acts other than legislative acts, not the procedure as such.<sup>892</sup> Where it is clear from the beginning (eg from the Commission adopting its proposal) that the intended final output is a legally non-binding act, the procedure laid down in Article 294 TFEU may be applied.

Even where it is stipulated that the European Parliament and the Council shall adopt a specified act according to the ordinary legislative procedure (for example a directive according to Article 50 para 1 TFEU), in principle the EP and the Council may also adopt a recommendation on this basis. Sometimes the Treaties make it clear, however, that no legally non-binding acts shall be adopted on the basis of a certain competence. In Article 165 para 4 TFEU (first indent), for example, the adoption of 'incentive measures' by the EP and the Council *in the ordinary legislative procedure* is provided for. Its second indent allows for the Council to adopt recommen‐ dations. This makes it clear that, unlike under the pre-Lisbon regime,893 the EP and the Council shall not adopt legally non-binding acts under Article 165 para 4 TFEU,894 but that only the Council shall adopt recommenda‐ tions (which, arguing *e silentio legis*, shall not contain incentive measures, though). Also where the Council is competent to adopt a specified act in the course of a special legislative procedure, it may not adopt a soft law act where this clearly goes against the purpose of the respective provision; see eg Article 86 para 1 TFEU, according to which the Council, 'by means of

<sup>892</sup> See Gellermann, Art. 296 AEUV, para 18, with reference to European Convention, Report of 29 November 2002 from the Chairman of the Working Group IX on Simplification, CONV 424/02, 6 f; see also Council, Comments on the Council's Rules of Procedure (2022) 100; with regard to the respective provision in the Consti‐ tutional Treaty see Senden, Soft Law 481f.

<sup>893</sup> See eg Recommendation 2006/962/EC of the European Parliament and of the Council of 18 December 2006 on key competences for lifelong learning which is based ('in particular') on Article 149 para 4 and Article 150 para 4 TEC; for the discussion on the non-bindingness of such incentive measures prior to the Lisbon Treaty see Niedobitek, Art. 165 AEUV, paras 59 f, with further references.

<sup>894</sup> See reference to the CJEU's case law in Blanke, Art. 165 AEUV, para 100. Differently: Rosas, Soft Law 311; see also Simm, Art. 165 AEUV, para 24, with further references.

regulations adopted in accordance with a special legislative procedure, may establish a European Public Prosecutor's Office from Eurojust'.

The confirmation of a general competence of the EP to adopt recommen‐ dations under the ordinary legislative procedure was mainly due to the Council's involvement in this procedure and the Council's firm and express soft law power laid down in Article 292 TFEU. As regards the legislative powers of the EP outside the ordinary legislative procedure (which are scarce anyway<sup>895</sup>), for lack of a link to the Council no *general* power to adopt recommendations can be deduced (indirectly) from Article 292 TFEU. This does not exclude the confirmation of an according competence *a maiore ad minus* in a concrete case (see 3.3.2.1. above).

# 3.4.3.2. The power to adopt recommendations of the Commission and of the ECB, respectively

In its last sentence Article 292 TFEU stipulates that the Commission shall adopt recommendations and that the ECB shall do so 'in the specific cases provided for in the Treaties'. The Commission thereby is vested with a comprehensive competence to adopt recommendations.896 This reflects the legal situation under the TEC,897 according to which the Commission could 'formulate recommendations […] on matters dealt with in this Treaty, if it expressly so provides or if the Commission considers it necessary'.<sup>898</sup> Where both the Commission and the Council (or the Council together with the EP) have the power to adopt recommendations on a certain issue,

<sup>895</sup> See, for example, Article 223 para 2 and Article 228 para 4 TFEU which both provide for the EP's power to adopt regulations in the course of a special legislative procedure.

<sup>896</sup> Note that under Article 211 (2nd indent) TEC (Nice) the Commission was vested with a comprehensive power not only to adopt recommendations, but also opin‐ ions; note that Belgium in case C-16/16P Belgium v Commission, para 19, argued that the Commission, in addition to 'the procedural legal basis' of Article 292 TFEU, would also require a material competence for the adoption of a recommendation. This argumentation was not taken up by the Court. For the meaning of Article 17 TEU for the Commission's competence to adopt soft law of all sort see Georgieva, Soft Law 226 f; see also case C-660/13 *Council v Commission*.

<sup>897</sup> See Gellermann, Art. 292 AEUV, para 5; Hofmann/Rowe/Türk, Administrative Law 547.

<sup>898</sup> Article 211 (2nd indent) TEC (Nice). Under the Nice regime, the Commission's competence did not extend into the former second and third pillar of the EU, though; see Nettesheim, Art. 292 AEUV, para 14.

political considerations may be determinative for either option.899 Like the Council, the Commission – in spite of the wording of Article 292 – does not dispose of an all-encompassing power to adopt recommendations. Rather, its power to adopt recommendations is limited to its (wide) field of responsibilities<sup>900</sup> and furthermore may be restricted by specific Treaty pro‐ visions.<sup>901</sup> In other words: The Commission may adopt recommendations if and to the extent to which this is not implicitly or explicitly barred by the Treaties, the outer limit of the Commission's power being the EU's *Verbandskompetenz*. With regard to the quorum and the majority required for the adoption of a Commission recommendation, the general decisionmaking rules apply.<sup>902</sup>

The ECB, on the contrary, according to Article 292 TFEU may adopt recommendations only in the specific cases in which the Treaties provide for this competence.903 As regards the ECB, Article 292 TFEU cannot be considered as providing for a competence. In the given context, it merely states a legal matter of course: that the ECB may exercise its competence (to adopt recommendations) where the Treaties so provide. This constitutes a reflection in particular of Article 132 para 1 (3rd indent) TFEU, according to which the ECB shall '[i]n order to carry out the tasks entrusted to the [European System of Central Banks; ESCB]' and 'in accordance with' the Treaties and the Statute of the ESCB and the ECB 'make recommenda‐

<sup>899</sup> See Commission Proposal COM(2008) 179 final, 7, according to which the Com‐ mission opted for proposing a Recommendation of the EP and the Council instead of adopting a Recommendation itself for the following reasons: 'A Commission Rec‐ ommendation would be a statement of Commission views, but neither the Member States nor the European Parliament would be involved in its formulation and it would not generate the political commitment needed to ensure implementation at the national level, which is crucial to successful European cooperation in this area. Use of a Commission Recommendation would carry the risk of being seen as a development which runs counter to subsidiarity'.

<sup>900</sup> See Merli, Art. 292 AEUV, para 15; Nettesheim, Art. 292 AEUV, para 13. For an example of a Commission soft law act which went beyond what the legislator intended as the scope of its policy see Weigt, Rechtsetzung 49.

<sup>901</sup> See eg Article 24 para 1 subpara 2 TEU for the (limited) role of the Commission in the area of CFSP; see also Merli, Art. 292 AEUV, para 15.

<sup>902</sup> See Article 250 TFEU and Articles 7 f of the Commission's Rules of Procedure, C(2000) 3614, as amended.

<sup>903</sup> Critically as regards the misleading wording of Article 292 TFEU: Nettesheim, Art. 292 AEUV, para 5.

tions'.904 As regards the quorum and the majority required for the adoption of an ECB recommendation, the institution's general decision-making rules apply.<sup>905</sup>

# 3.4.4. Article 127 para 4 and Article 132 para 1 (3rd indent) TFEU

The ECB's competence to adopt opinions seems to be more encompassing than its competence to adopt recommendations. According to Article 127 para 4, the ECB 'may submit opinions to the appropriate Union institu‐ tions, bodies, offices or agencies or to national authorities on matters in its fields of competence'.<sup>906</sup> This is – with a different wording<sup>907</sup> – also laid down in Article 282 para 5 TFEU: 'Within the areas falling within its responsibilities, the [ECB] […] may give an opinion' on 'all proposed Union acts, and all proposals for regulation at national level'.908 Thus, the ECB has an extensive right to address opinions not only to other EU bodies, but also to national authorities.909 This right is limited by the ECB's fields of competence, in which 'by virtue of the high degree of expertise

<sup>904</sup> See also the repetitive Article 34 para 1 (3rd indent) of the Statute of the ESCB and of the ECB.

<sup>905</sup> See Article 10 of the Statute of the ESCB and of the ECB; Article 4 of the ECB's Rules of Procedure, ECB/2004/2, as amended.

<sup>906</sup> See also the repetitive Article 4 lit b of the Statute of the ESCB and of the ECB; negating the soft law quality of these opinions: Knauff, Regelungsverbund 304.

<sup>907</sup> The wording of Article 282 para 5 goes back to the draft of the Constitutional Treaty and has not been adapted. The intended scope therefore seems to be the same as that of Article 127 para 4 TFEU; see Griller, Art. 127 AEUV, para 64, with a further reference.

<sup>908</sup> For the material scope see Griller, Art. 127 AEUV, paras 65–67; see case C-11/00 *Commission v European Central Bank*, paras 110 f, with reference to the scope of the duties to consult the ECB (a counterpart to its right to adopt opinions) under the predecessor provision of Article 127 para 4 TFEU, Article 105 para 4 TEC. The scope of these consultation rights was and now (under the TFEU) is the same – 'in its fields of competence' – as that of the ECB's right to adopt opinions.

<sup>909</sup> For the respective duty of the national authorities to consult the ECB see Article 127 para 4 (2nd indent) TFEU; Council Decision 98/415/EC on the consultation of the European Central Bank by national authorities regarding draft legislative provisions, which was based on the predecessors of Articles 127 para 4 and 129 para 4 TFEU; see also ECB/Eurosystem, Guide to consultation of the European Central Bank by national authorities regarding draft legislative provisions (October 2015) <https://www.ecb.europa.eu/pub/pdf/other/consultationguide201510.en.pdf> accessed 28 March 2023.

that it enjoys, [it] is particularly well placed to play a useful role'<sup>910</sup> in the respective decision-making processes. This is not the case in 'an area in which the ECB has not been assigned any specific tasks', even if it would be affected by the decision to be adopted.<sup>911</sup> However, it may – arguably due to the close inter-relation between monetary and economic policy<sup>912</sup> and hence due to the ECB's expertise also in this field – encompass measures falling within the realm of economic policy.<sup>913</sup> Asked to provide an opinion in a field in which it does not have sufficient expertise, in light of Article 127 para 4 TFEU the ECB has to refuse to provide an opinion.

Article 132 para 1 (3rd indent) TFEU lists the competences of the ECB.<sup>914</sup> While Article 288 TFEU lists the legal acts to be adopted by the institutions '[t]o exercise the Union's competences', Article 132 para 1 TFEU is more specific in this respect: 'In order to carry out the tasks entrusted to the ESCB, the European Central Bank shall, in accordance with the provisions of the Treaties and under the conditions laid down in the [E(S)CB-Statute] […] make recommendations and deliver opinions'. As regards the ECB's power to adopt recommendations, the meaning of Article 132 para 1 TFEU does not go beyond that of Article 292 TFEU. As regards the ECB's power to adopt opinions, Article 127 para 4 TFEU is certainly the more pertinent (general) provision.

<sup>910</sup> Case C-11/00 *Commission v European Central Bank*, para 110.

<sup>911</sup> Case C-11/00 *Commission v European Central Bank*, para 111; for the opinions launched by the ECB in the preparation of the Constitutional Treaty, and the Treaty of Lisbon respectively, see Sáinz de Vicuña, Status 299–301.

<sup>912</sup> For this inter-relation see eg Thiele, Akteur.

<sup>913</sup> See case C-370/12 *Pringle*, para 61, according to which a *facultative* consultation of the ECB is not unlawful; for the supporting role the ECB/Eurosystem has in economic policy see in particular Article 127 para 1 TFEU.

<sup>914</sup> It is largely affirmed that this provision actually confers competences which, in my view, is not entirely clear. If it were a competence-conferring norm, its second case (opinions) would still, as compared to Article 127 para 4 TFEU, be *lex generalis* – *argumentum* 'in accordance with the provisions of the Treaties and under the conditions laid down in the [E(S)CB-Statute]'. It would therefore be derogated by Article 127 para 4 TFEU; for the similarity to Article 288 TFEU see Griller, Art. 132 AEUV, para 1; for its competence-conferring character see also Häde, Art. 132 AEUV, para 1; Wutscher, Art. 132 AEUV, paras 1f.

# 3.4.5. Opinions of committees

The EU disposes of a large number of committees of various form and dif‐ ferent fields of expertise.<sup>915</sup> Their assessments feed into different decisionmaking processes within the EU legal order. In the majority of procedures, the committees' respective output is not binding,<sup>916</sup> but regularly has an influence on the final decision, the degree of which is varying.<sup>917</sup> Here we shall take a look at provisions laid down in the Treaties, according to which few of these committees dispose of an – at least *prima facie* – general competence to adopt opinions.

At first we address the Political and Security Committee (PSC) which, according to Article 38 TEU, shall 'monitor the international situation in the areas covered by the [Common Foreign and Security Policy; CFSP] and contribute to the definition of policies by delivering opinions'. These opin‐ ions shall be addressed to the Council at its request or at the request of the High Representative for Foreign Affairs and Security Policy, or on the PSC's own initiative. The PSC shall furthermore monitor the implementation of agreed policies (without prejudice to the powers of the High Representa‐ tive). Apart from entirely non-normative PSC output this monitoring may entail, arguably also in this specific role it may adopt soft law in the form of opinions.918 Also its tasks referred to in Article 38 para 2 TEU (exercising the political control and strategic direction of crisis management operations according to Article 43 TEU) may include the power to adopt soft law in the form of opinions. However, the exercise of these tasks is subject to con‐ cretisation by the Council.919 A true decision-making power of the PSC is

<sup>915</sup> For a categorisation of the variety of EU committees see Harcourt, Governance 10.

<sup>916</sup> For an exception see the status of opinions of the comitology committees in the examination procedure; see 3.7.2.1. below.

<sup>917</sup> See case C-11/05 *Friesland Coberco*, para 39; see also C‑572/18P *thyssenkrupp*.

<sup>918</sup> The wording 'monitor the international situation in the areas covered by the [CFSP]' of Article 38 para 1 TEU (first sentence) is wide enough to also cover the monitoring of the implementation of agreed policies. Therefore the power to adopt opinions extends to the latter task; for the similarity of these tasks see also Marquardt/Gaedtke, Art. 38 EUV, para 4.

<sup>919</sup> Article 43 para 2 TEU; see eg (meanwhile repealed) Council Decision (CFSP) 2015/778, in which the Committee was vested with different decision-making pow‐ ers (Articles 6 para 1 and 9 para 5).

subject to (in practice regularly granted)<sup>920</sup> authorisation by the Council.<sup>921</sup> This power generally – *a maiore ad minus* – may imply the power to adopt soft law, in particular: to adopt recommendations.<sup>922</sup> Whether this can actually be affirmed in a certain situation needs to be assessed case by case, taking into account in particular teleological considerations. The power to take 'relevant decisions on the setting-up of a [committee]',<sup>923</sup> for example, cannot reasonably be exercised by adopting a recommendation.

Another soft law competence of a committee which may be called 'gen‐ eral' in that it is not limited to concrete (specified) situations,<sup>924</sup> is laid down in Article 134 para 2 TFEU. This provision vests the Economic and Financial Committee (EFC) with the power to deliver opinions to the Council or to the Commission at their respective request, or on its own initiative. Listed as one of the Committee's tasks<sup>925</sup> – *pari passu* with other tasks – in Article 134 para 2 TFEU, it cannot be assumed that the power to submit opinions on its own initiative may only be used when exercising the other tasks, eg keeping under review the economic and financial situation of the MS and of the EU (2nd indent *leg cit*); this is to be concluded *a fortiori*, as the other tasks partly go together with a specific tool (explicitly: report in the 2 nd and 4 th indent, also in para 4 *leg cit*; implicitly: draft/opinion in the 3 rd indent). It can be assumed, however, that the Committee's competence to adopt opinions is limited, apart from the EU's broad *Verbandskompetenz*, by the EFC's – admittedly wide – scope of responsibilities which is hinted at in Article 134 para 1 TFEU.926 In this provision the Committee's *raison d'être* is formulated as the promotion of

<sup>920</sup> See Marquardt/Gaedtke, Art. 38 EUV, para 7; see also H-J Cremer, Art. 38 EUV, para 6; references by Terhechte, Art. 38 EUV, para 4.

<sup>921</sup> Article 38 para 3 TEU.

<sup>922</sup> For the resemblance between recommendations and secondary law see 3.1.1. above.

<sup>923</sup> See Article 9 para 5 of (meanwhile repealed) Council Decision (CFSP) 2015/778.

<sup>924</sup> In specified situations soft law powers of the EFC are laid down in Article 126 para 4 TFEU (opinion) and in Articles 143 para 1 and 144 para 3 TFEU (right to be consulted, which equals a right to render an opinion upon request).

<sup>925</sup> Whether the adoption of soft law is named 'task' or 'competence'/'power' depends on the perspective (which may also be influenced by the wording of the relevant provision; see Article 134 para 2, 1 st indent, TFEU). The task to adopt a soft law act certainly always also entails the respective competence. On the other hand: Whether a competence to adopt a soft law act ought to be used in a certain situation is to be examined with a view to the respective actor's tasks.

<sup>926</sup> See Palm, Art. 134 AEUV, para 15; similarly, with reference to the Statute of the EFC, Wutscher, Art. 134 AEUV, para 8.

the 'coordination of the policies of Member States to the full extent needed for the functioning of the internal market' (within which field it shall focus on economic and financial issues<sup>927</sup>). Where the EFC launches an opinion in order to contribute to the preparation of the work of the Council or in exercising other advisory and preparatory tasks assigned to it by the Council in accordance with Article 134 para 2 (3rd indent in conjunction with 1 st indent), it is furthermore limited ('without prejudice to') by the tasks of the Committee of Permanent Representatives (COREPER) and of the Council's General Secretariat according to Article 240 TFEU.

Article 150 TFEU lays down that the Council shall establish an Employ‐ ment Committee with advisory status 'to promote coordination between Member States on employment and labour market policies'. One of the tasks of the Employment Committee is to formulate opinions at the request of the Council or the Commission, or on its own initiative.928 This gen‐ eral competence is, similar to the case of the EFC, limited by the EU's *Verbandskompetenz*, the Employment Committee's (wide) scope of respon‐ sibilities – the promotion of the coordination on employment and labour market policies between MS – and both the COREPER's and the tasks of the Council's General Secretariat according to Article 240 TFEU.<sup>929</sup>

The legal situation is similar in case of the Social Protection Committee (SPC). According to Article 160 TFEU, it has 'advisory status' and shall 'promote cooperation on social protection policies between Member States and with the Commission' – again a relatively wide scope which can work only as a blurry (material) limit to the SPC's soft law power. It may formu‐ late opinions at the request of the Council or the Commission, or on its own initiative. The SPC may also 'undertake other work within its fields of

<sup>927</sup> The Council Decision 2012/245/EU on a revision of the Statutes of the Economic and Financial Committee specifies the Committee's powers with the exemplary list ('may, inter alia') enshrined in Article 2 of the Statutes annexed to the Decision. According to Article 3 of the Statutes, the Committee shall be guided by the 'general interests of the Union'. This can be interpreted as suggesting that the Committee shall not adopt opinions beyond its expertise, as this would not be in the interest of the EU.

<sup>928</sup> Article 150 (2nd indent) TFEU. That also the second task mentioned in the 2 nd indent, namely the contribution to the preparation of the Council proceedings referred to in Article 148, implies the power to adopt an opinion is underpinned by Article 148 para 4: 'having received the views of the Employment Committee'; see Simon, Art. 150 AEUV, para 6.

<sup>929</sup> See Simon, Art. 150 AEUV, para 5.

competence'.930 The EU's *Verbandskompetenz*, the tasks of the COREPER and of the Council's General Secretariat again form further limits to this power.

The Economic and Social Committee (ESC) and the Committee of the Regions (CoR) are dealt with in the TFEU under the heading 'The Union's advisory bodies'. Among the EU's committees the ESC and the CoR have a distinguished position.<sup>931</sup> Also their high (maximum) number of members – 350 each<sup>932</sup> – contributes to an enhanced legitimacy of their output. The ESC shall be consulted by the EP, the Council or the Commission where the Treaties so provide (obligatory consultation) <sup>933</sup> and otherwise where they consider it appropriate (facultative consultation). An obligatory consultation shall not be replaced by a merely facultative consultation.<sup>934</sup> Conversely, the ESC may address an opinion (together with a record of the proceedings935) to these institutions on its own initiative where it considers this appropriate.936 The EP, the Council or the Commission may set a time limit of not less than one month within which the ESC shall render its opinion (both in case of an obligatory or a facultative consultation937), oth‐ erwise the institutions may proceed in the respective decision-making proc‐ ess. These rules apply, *mutatis mutandis*, also to the CoR.938 With regard to the facultative consultation of the CoR which shall take place, as in the case of the ESC, whenever the named institutions consider it appropriate, the Treaty emphasises the cases concerning cross-border cooperation.<sup>939</sup> Another peculiarity of the CoR is that it is to be informed by the named

<sup>930</sup> In this 'other work' the SPC may not go beyond its consultative role; see Benecke, Art. 160 AEUV, para 2.

<sup>931</sup> See Blanke, Art. 300 AEUV, para 5.

<sup>932</sup> This maximum currently is exploited neither by the ESC nor by the CoR; see <http://www.eesc.europa.eu/?i=portal.en.about-the-committee> and <https://cor.e uropa.eu/en/about/

Pages/default.aspx>, both accessed 28 March 2023.

<sup>933</sup> For the Treaty Articles providing for such an obligatory consultation see Jaeckel, Art. 304 AEUV, para 9.

<sup>934</sup> See, with regard to the EP's right to be consulted, case C-316/91 *European Parlia‐ ment v Council*, paras 16 f, with references to further case law.

<sup>935</sup> This includes in particular information on the turnout of the vote; Brinker, Art. 304 AEUV, para 12, with references to the ESC's Rules of Procedure.

<sup>936</sup> Article 304 paras 1 and 3 TFEU.

<sup>937</sup> See Jaeckel, Art. 304 AEUV, para 22.

<sup>938</sup> Article 307 TFEU; for a list of the Treaty Articles providing for an obligatory consultation see Blanke, Art. 307 AEUV, para 4.

<sup>939</sup> Article 307 para 1 TFEU.

institutions when they have requested an opinion from the ESC.<sup>940</sup> Where the CoR then 'considers that specific regional interests are involved', it may render an opinion on this issue, as well (accessory consultation).<sup>941</sup> While this information may be important for the CoR, the collateral 'entitlement' is redundant (and hence only declaratory), as the CoR according to Article 307 para 4 TFEU is granted a comprehensive right to address *sua sponte* an opinion to the EP, the Council and the Commission.

The ESC and the CoR have a general right to render opinions to the EP, the Council and the Commission. Apart from the *Verbandskompetenz*, there are no explicit legal limits to the material scope of the opinions, as the scope of the committees' tasks is not specified. It can be concluded from their respective names and composition, and from the case law of the CJEU,<sup>942</sup> however, that they shall focus on economic and social questions and, respectively, on questions regarding the EU's regions. Leaving ques‐ tions of legal competence apart, the following statement may be appropri‐ ate: Where the committees move within their respective field of expertise, their opinions arguably reach high levels of factual authority – which is essential for bodies limited to render legally non-binding measures.

# 3.4.6. Limits to a 'general' competence to adopt soft law

Having examined the (general) competence clauses for the adoption of rec‐ ommendations and opinions, and having addressed some of the potential restrictions to the scope of these competence clauses, the question arises whether more generally the soft law power of an EU body may conflict with the competence of another EU body to regulate in the same policy field. The assumption that also a soft law act may interfere with other bodies' competences cannot be countered by referring to the fact that soft law acts are – *qua* being legally non-binding – ousted by legal acts anyway. This would mean a strong underrating of the factual steering effects of soft

<sup>940</sup> That there is no such duty to inform in case of the ESC issuing an opinion on its own initiative is clear from the wording (*argumentum* 'is consulted'); see also Blanke, Art. 307 AEUV, para 8.

<sup>941</sup> Article 307 para 3 TFEU; critically: Blanke, Art. 307 AEUV, para 8; Hönle, Art. 307 AEUV, paras 10–13.

<sup>942</sup> With regard to the ESC see joined cases C-281, 283–285, 287/85 *Germany v Com‐ mission*, para 38: The ESC 'is to advise the Council and Commission on the solu‐ tions to be adopted with regard to practical problems of an economic and social nature and to deliver opinions based on its specific competence and knowledge'.

law – rules do not merely steer human behaviour, '[r]ules [also] influence norms, rules facilitate coordination, and rules create differentiation and status orders'<sup>943</sup> – and of the challenges to legal certainty such a normative conflict may entail. It would furthermore relativise the purpose of the EU's competence regime. What is more: Soft law may not only interfere with other bodies' hard law power, but also with their respective power to adopt *soft* law. Which competence prevails in a certain case is to be examined by applying legal methodology (eg the principle *lex specialis derogat legi generali*).<sup>944</sup> A conflict may also be dissolved by considering the actual tasks of the bodies at issue. Where – in the context of soft law – a cleavage between competences and tasks exists, more generally the following can be said: Where an EU body is competent to adopt soft law acts (eg opinions) in general, but – in a specific case – does not have the task (responsibility) to interfere, it shall not make use of its competence.945 In this context, it is said that the actor *can* do more than it *shall*. 946

An example for an interference with another EU body's competence can be found in State aid law in which the Commission has adopted a number of guidelines,947 notices, codes, etc.948 Under Article 109 TFEU, the Council may, on a proposal from the Commission and after consulting the EP, make 'any appropriate regulations for the application of Articles 107 and 108'.949 This regulatory competence of two institutions in one policy field – 'concurrent powers' – holds the problem of parallel rule-making.<sup>950</sup> While the Court did not consider this a problem in terms of the principle of 'institutional equilibrium',951 arguably also because in this case there was

<sup>943</sup> Ahrne/Brunsson, Soft Regulation 187.

<sup>944</sup> See Senden, Soft Law 481, arguing that a specific legal basis for the adoption of a Council recommendation prevents the Commission from making use of its general competence to adopt recommendations according to Article 211 para 2 TEC.

<sup>945</sup> See Ruffert, Zuständigkeitsgrenzen 161 ff, with regard to the international level.

<sup>946</sup> On this phenomenon in public law more generally see Brandt, Umweltaufklärung 88–90.

<sup>947</sup> For Commission guidelines in general see Hofmann/Rowe/Türk, Administrative Law 548–551.

<sup>948</sup> For these and further designations see H Hofmann, Rule-Making 158.

<sup>949</sup> For how the Council made use of its competence see Koenig/Paul, Art. 109 AEUV, paras 4 f; for the political difficulties of Council State aid legislation up until 2000 see Cini, Soft Law 8: '[T]he absence of state aid legislation led […] to the construc‐ tion of a body of informal Commission rules which served as a substitute for "hard" legislation'; see also Cini, Soft law approach 200 f; Ştefan, Soft Law 48 and 53 f.

<sup>950</sup> See Senden, Balance 91f.

<sup>951</sup> See joined cases T-132/96 and T-143/96 *Sachsen/Volkswagen*, para 241.

no relevant (countervailing) Council act,<sup>952</sup> it was made clear, by AG *Wahl*, that these soft law acts cannot be considered 'binding – not even *de facto* – on pain of eluding the legislative procedure set out in the FEU Treaty'. Were the Commission's measures binding, he adds with reference to the Court's case law,<sup>953</sup> they 'would be null and void'.<sup>954</sup>

The duty of each EU body to consider the competences of its peers roots in the principle of mutual sincere cooperation, as laid down in Article 13 para 2 TEU. Accordingly, an EU body when acting 'in turn must have regard to the power' of other EU bodies.<sup>955</sup> This applies irrespective of whether the action is legally binding or not.

Such duty of mutual sincere cooperation may be concretised by sec‐ ondary law: Where the ESMA, to take an example, deems a competent supervisory authority in a MS to have breached pertinent Union law, it may address a recommendation to the respective authority, thereby 'set‐ ting out the action necessary to comply with Union law'.956 Where the authority addressed has not complied with Union law as laid down in the recommendation within one month from its receipt, the Commission 'may, after having been informed by the [ESMA] or on its own initiative, issue a formal opinion requiring the competent authority to take the action necessary to comply with Union law'.957 According to Regulation 1095/2010 this procedure may be further extended, but this is of no relevance in the given context. What is relevant is that here the principle of mutual sincere cooperation arguably restricts the Commission's power to adopt soft law. In that sense, the Commission should abstain from adopting a reasoned opin‐ ion according to Article 258 TFEU to the MS of the respective competent supervisory authority, once the ESMA has addressed a recommendation to the national authority according to the above procedure. Rather, it should

<sup>952</sup> On the very limited use the Council makes of its competence according to Article 109 TFEU see von Wallenberg/Schütte, Art. 109 AEUV, paras 1f.

<sup>953</sup> See eg case C-57/95 *France v Commission*.

<sup>954</sup> Opinion of AG *Wahl* in case C-526/14 *Kotnik*, paras 36–38. A legal act invalidated (by the Court) that way loses its legal quality, ie it ceases to legally exist, and therefore – on a normative level – ranks lower than soft law.

<sup>955</sup> See case 230/81 *Luxembourg v European Parliament*, para 38; case C-65/93 *Euro‐ pean Parliament v Council*, para 27; case C-94/00 *Roquette Frères*, para 31; case C-73/14 *Council v Commission*, para 61, with further references; for the connection between inter-institutional loyalty and the EU's institutional balance see Klamert, Loyalty 28.

<sup>956</sup> Article 17 para 3 subpara 1 of Regulation 1095/2010.

<sup>957</sup> Article 17 para 4 subpara 1 of Regulation 1095/2010.

follow the procedure initiated by the ESMA and adopt, where necessary, a formal opinion according to Article 17 para 4 subpara 1 of Regulation 1095/2010. The Commission, beyond doubt, is competent to adopt a rea‐ soned opinion based on Article 258 TFEU, but it would have to initiate a new procedure aiming at ensuring MS' compliance with EU law and would thereby diminish the thrust of the ESMA's action – and it would do so without a compelling, or only a sensible reason. After all, both procedures have essentially the same aim – MS' compliance with EU law – and the Commission has a role to play also in the Article 17-procedure. Therefore, the principle of mutual sincere cooperation suggests that the Commission should follow the path of Article 17 already trodden by the ESMA instead of initiating a separate procedure aiming at the same result.<sup>958</sup>

A further limit to the adoption of soft law is that an according compe‐ tence (eg established *a maiore ad minus*) may not be used in order to *circumvent* the adoption of legal norms provided for in a competence clause959 or where this would constitute an explicit disregard of the tasks of the actors involved. A circumvention would be at issue where a recommen‐ dation, regulating only the basics, leaving (important) details for the MS to be decided (a 'soft directive', as it were), is adopted instead of the required

959 See Senden, Soft Law 327 f, with regard to '"obligating" legal bases' which oblige the legislator to actually *legislate*; see, with regard to Commission recommendations, Andone/Greco, Burden 92; from the European Parliament's perspective: Europe‐ an Parliament (Committee on Legal Affairs), Working Document on institutional and legal implications of the use of 'soft law' instruments (14 February 2007), PE 384.581v02–00, 6; with regard to public international law see Bodansky, Character 145.

<sup>958</sup> Note that only Article 17 para 6 of Regulation 1095/2010 – according to which the ESMA may, its own and the Commission's subsequent soft law measures being unsuccessful, address a binding decision directly to a financial market participant – is explicitly '[w]ithout prejudice to the powers of the Commission under Article 258 TFEU', so as to dispel the impression that the ESMA's competence ousts the Commission's role as guardian of the Treaties in this respect. With regard to the preceding steps of the procedure this question is – according to the letter of the law – left open. It could be argued that since Article 258 TFEU constitutes primary law, it is clear that the competences of the Commission laid down in this provision cannot be reduced by an act of secondary law, here: the ESMA-Regulation. But if the relevant provisions of Regulation 1095/2010 are understood as a concretisation of what the principle of sincere cooperation demands anyway, the conflicting rules would have equal rank. *Storr* suggests that if the national authority has failed to comply with the Commission's formal opinion, apart from the ESMA decision directly addressed to the financial market participant, the Commission may initiate a Treaty infringement procedure; Storr, Agenturen 81f.

decision (eg in case of Article 155 para 2 TFEU).<sup>960</sup> Thereby the prohibition to adopt a directive (*argumentum* 'decision') would be circumvented by a recommendation (in this case functioning like a directive).<sup>961</sup> While the power to adopt a recommendation on the basis of Article 155 TFEU may in principle be deduced *a maiore ad minus*, the directive-like shape of the recommendation<sup>962</sup> suggests a circumvention of one of the objectives of Article 155 TFEU, namely a definitive implementation of the agreements concluded between management and labour according to Article 155 para 1 TFEU.<sup>963</sup>

Where the legislator is required to act in a legally binding way and it is clear that a legally non-binding act would not serve the purpose of the respective norm, the adoption of a soft law act would be in disregard of the respective tasks. Article 24 TFEU may serve as an example. According to this provision, the Council and the European Parliament shall adopt regulations in accordance with the ordinary legislative procedure to lay down the procedures and conditions required for a citizens' initiative. <sup>964</sup> A soft law act instead of a regulation would definitely not serve the purpose of Article 24 TFEU; it would run counter to the aim of legal certainty which is imperative especially in this case. It could be argued that here the legislator lacks a soft law competence in the first place, because – for the reasons just mentioned – the *argumentum a maiore ad minus* is not valid.

More generally, the excessive adoption of recommendations by the EP and the Council on a proposal from the Commission in the course of the

<sup>960</sup> See Senden, Soft Law 489. For a circumvention disclosed by the Court see joined cases 8–11/66 *Cimenteries*, 92; see also Alberti, Evolution 647, with regard to the ECB's powers; critically with regard to 'legislative or administrative activism' by means of soft law: Christianos, Effectiveness 329; for the unlawfulness of 'evading a procedure specifically prescribed by the Treaty' ('misuse of power' in Article 263 para 2 TFEU) see case C-180/96 *United Kingdom v Commission*, para 64, with a further reference.

<sup>961</sup> See case T-258/06 *Germany v Commission*, para 28; case C-687/15 *Commission v Council*, paras 40–44, with further references; see also more generally Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, para 93: 'What is perhaps the greatest strength of recommendations may also then be the greatest danger. They could be used as more than just tools for advancing policies that are politically (lack of consensus) or legally (no specific powers to that effect) gridlocked. They could also potentially be used as a tool to circumvent the same legislative processes'; critically as well: Arnull, Recommendations 617.

<sup>962</sup> For the term 'directive-like recommendation' see also Láncos, Facets 24 f.

<sup>963</sup> See also Hesse, Art. 155 AEUV, paras 17 f; Rebhahn, Art. 155 AEUV, para 5.

<sup>964</sup> See also Article 11 para 4 TEU.

procedure laid down in Article 294 TFEU would qualify as a disregard of the tasks of these institutions. It may appear opportune in selected cases in which, for political reasons, the respective majorities for legislative acts cannot be achieved, but it must remain the exception. Where it is difficult to find majorities for certain legislative projects, in general they should be amended accordingly, or – where this is not possible for whichever reason – they should simply fail. A replacement of the legislative activity of the EP and of the Council on a large scale would constitute an unlawful contempt of the principal legislative task of these institutions as generally laid down in Article 14 para 1 TEU ('exercise legislative […] functions').<sup>965</sup>

A historical case is the (attempted) harmonisation of the maximum blood alcohol content for drivers of motorised vehicles. After an according Commission proposal sent to the Council in 1988 failed for reasons of subsidiarity concerns,966 the Commission in 2001 essentially adopted the proposal as a recommendation.967 In view of the extended time span be‐ tween the failure of the proposal and the adoption of the recommendation – 13 years – and in view of the Commission's acknowledgement of the MS' subsidiarity concerns,968 it can be doubted that the latter actually constitutes a circumvention of a legislative procedure. Still, the (other) reasons why the Commission has adopted the recommendation are evident, as well: the negative experiences with the legislative procedure in the past and the little chance of an outcome that is satisfying for the Commission in a new procedure.969 The Commission adopted a recommendation in spite

<sup>965</sup> Note the words of AG *Bobek* in case C-911/19 *FBF*, who demands sufficient judicial review of soft law, as otherwise there would be 'a further spread of "crypto-legisla‐ tion" in the form of soft law in the Union. […] EU bodies are able, through soft law, to create parallel sets of rules which bypass the legislative process and which might have an impact on institutional balance' (para 89).

<sup>966</sup> Commission, Proposal for a Council Directive relating to the maximum permitted blood alcohol concentration for vehicle drivers, COM(88) 707 final.

<sup>967</sup> Commission, Recommendation of 17 January 2001 on the maximum permitted blood alcohol content (BAC) for drivers of motorised vehicles, 2001/C 48/02. For the effectiveness of this recommendation compare the BACs provided for in the MS in 2001 (as listed in the recommendation) with those from 2016; <http://etsc.eu/bl ood-alcohol-content-bac-drink-driving-limits-across-europe/> accessed 28 March 2023.

<sup>968</sup> See Commission, Recommendation of 17 January 2001 on the maximum permitted blood alcohol content (BAC) for drivers of motorised vehicles, 2001/C 48/02, 1.1.4.

<sup>969</sup> Critically with regard to 'parallel means of legislation', namely Council declarations: Opinion of AG *Darmon* in case C-292/89 *Antonissen*, para 26; see also: Senden, Soft Law 28.

of the request of the EP's Transport Committee of 1999 to renew its pro‐ posal.<sup>970</sup> In a similar case – a legislative proposal failed and was followed by a Commission Communication with essentially the same content<sup>971</sup> – the Court, considering the parallels between these two documents, held the Communication to be legally binding, and annulled it for lack of the Commission's competence.<sup>972</sup>

The CJEU gave an example in this context which is related to the Court itself. In a preliminary ruling procedure it refused to provide its interpretation of a legal act, since this interpretation would – due to the

971 The Commission Communication on an internal market for pension funds, 94/C 360/08, was adopted shortly after an according proposal for a directive failed ('Re‐ grettably, the Commission has felt itself obliged to withdraw its proposal because of a deadlock in the negotiations with Member States in the Council'; 1.4. of the Communication); for similar examples see Gundel, Rechtsschutz 595 (fn 8).

972 See case C-57/95 *France v Commission*, paras 12 and 25. AG *Tesauro* in para 17 of his Opinion in this case criticised the Commission's behaviour as 'camouflaging the proposal for directive as a communication'; this political matter and its judicial aftermath are considered in detail by C Adam, Politics 25–28. See also case C-325/85 *Ireland v Commission*, paras 16 ff, in which the Court considers the illegality of Commission decisions (based on its proposals) attempting to fill the decisional vacuum left by the Council – which, for political reasons, was unable to act; see also other cases cited by Schmidt/Schmitt von Sydow, Art. 17 EUV, para 70 (fn 99); see also Brohm, Mitteilungen 198. For a case in which the Court refused to annul a Communication for lack of new rules see case T-258/06 *Germany v Commission*, in particular para 162; for the 'Maulkorbfunktion' [muzzle function] the action underlying this case was intended to have with regard to Commission action: U Stelkens, Rechtsetzungen 408.

Less problematic is certainly the reverse order: a recommendation setting the scene and presenting a certain set of rules, which is then – some years later and, for exam‐ ple, for lack of effectiveness or in order to actually create legal claims for EU citizens after MS have got accustomed to the common rules – followed by proposal and subsequently a legislative act to that effect; see eg Commission Recommendation of 22 December 1986 concerning the introduction of deposit-guarantee schemes in the Community, 87/63/EEC, which was followed by Directive 94/19/EC, adopted upon a proposal of the Commission. The reason for this 'hardening' of the rules was that the Recommendation 'has not fully achieved the desired result' (Recital of Directive 94/19/EC). This Directive was later replaced by Directive 2014/49/EU.

Sometimes the Commission in the respective soft law act even 'threatens' to propose a legislative act in case compliance with the former is dissatisfactory; see eg Recital 12 of Commission Recommendation of 30 July 1997 concerning transactions by electronic payment instruments and in particular the relationship between issuer and holder, 97/489/EC.

<sup>970</sup> See Commission, Recommendation of 17 January 2001 on the maximum permitted blood alcohol content (BAC) for drivers of motorised vehicles, 2001/C 48/02, 1.1.4.

inapplicability of this act in the national procedure – not be binding on the referring (national) court: 'It cannot be accepted that the replies given by the Court to the courts of the contracting States are to be purely advisory and without binding effect'.<sup>973</sup>

# 3.5. Special competence clauses in the Treaties

# 3.5.1. Introduction

Having considered the general competence clauses for the adoption of recommendations and opinions in the Treaties, we shall now take a look at specific competence clauses, allowing for the adoption of recommendations and opinions eg in the course of certain decision-making procedures. An analysis of these special competence clauses shall allow for a more compre‐ hensive view on the conferral of soft law powers by means of EU primary law. For each institution, the purpose of the respective soft law powers shall be categorised as either 'support of decision-making/rule-making', 'initiation of (soft) decision-making/rule-making', or 'soft decision-mak‐ ing/rule-making'. The term 'rule' is to be understood in a non-specific way as generally applicable norms, eg provisions of international agreements directly applicable in the EU/the MS. With regard to the competences to adopt recommendations and opinions of bodies, offices and agencies laid down in primary law (see 3.5.3.1. below), due to their scarcity such a categorisation does not seem to be worthwhile. It is not the purpose of this chapter to explicate the respective competence clauses in a comprehensive, commentary-like fashion, but to highlight and discuss the adoption of recommendations and opinions – the two main (potential) expressions of soft law, as listed in Article 288 TFEU – set out therein. A few words on the respective provision as such may be necessary, though, in order to put the concrete recommendation/opinion in context. Primary law competences to adopt soft law acts other than recommendations and opinions, such as pro‐ posals, requests or initiatives,974 shall be presented and analysed collectively under 3.6. below.

<sup>973</sup> Case C-346/93 *Kleinwort Benson*, paras 23 f; more recently referred to in case C-62/14 *Gauweiler*, para 12.

<sup>974</sup> Exceptionally, ie where they are to be mentioned in the context of the adoption of recommendations/opinions on the basis of a specific competence clause, also these

Furthermore, it is to be noted upfront that also provisions not expressly mentioning the power to adopt a 'recommendation' or an 'opinion' *may* disclose – at a 'second look', as it were – a competence to adopt either of these soft law acts. The Treaties contain provisions in which they allow EU bodies, in particular the institutions, to take 'the necessary measures',<sup>975</sup> 'incentive measures'<sup>976</sup> or simply 'measures',<sup>977</sup> to 'provide incentives and support',<sup>978</sup> to 'support and strengthen'<sup>979</sup> or to 'facilitate'<sup>980</sup>. The list of such malleable competences and/or tasks could be further extended. These terms in their literal understanding (verbal interpretation) may all be read to contain a soft law power.<sup>981</sup> Often it is the procedure to be applied – regularly: the ordinary or a special legislative procedure – which indicates that these measures are envisaged as legislative (ie legally binding) acts.<sup>982</sup> Where no such procedure is provided for, the question whether or not a soft law power is to be confirmed for the respective body can only be answered after further analysis, namely systematical, teleological and – where the information available so allows – historical interpretation. With regard to the competence of the standing committee according to Article 71 TFEU ('internal security' of the Council), for example, its power to 'facilitate coordination' is read as including the power to adopt recommen‐ dations.983 Article 156 TFEU, to take another example, explains itself what it understands with the Commission's power to 'encourage cooperation' and to 'facilitate the coordination', respectively, between MS and their actions, respectively: *inter alia* the delivery of opinions. With Article 165 para 4 (1st indent) TFEU, on the contrary, an interpretation of the Article in its entirety discloses that no soft law power is conferred (see 3.4.3.1. above).

acts shall be taken into account: eg the proposals and their effects according to Article 293 TFEU.

<sup>975</sup> Eg Article 189 para 2, Article 197 para 2, Article 215 para 1, Article 325 para 4 TFEU.

<sup>976</sup> Eg Article 19 para 2, Article 149, Article 165 para 4 (1st indent) TFEU.

<sup>977</sup> Eg Article 5 paras 1 and 2 (see also Article 121), Article 21 para 3, Article 214 para 3 TFEU.

<sup>978</sup> Article 79 para 4 TFEU.

<sup>979</sup> Article 85 para 1, Article 88 para 1 TFEU.

<sup>980</sup> Eg Article 71, Article 156 TFEU.

<sup>981</sup> See von Bogdandy/Arndt/Bast, Instruments 112; see also Classen, Art. 166 AEUV, para 42, with regard to the term 'measures'; Senden, Soft Law 327 f, with regard to such '"enabling" legal bases'.

<sup>982</sup> In such cases still implied soft powers may be affirmed (see 3.3.1. above).

<sup>983</sup> See Breitenmoser/Weyeneth, Art. 71 AEUV, para 14; Röben, Art. 71 AEUV, para 6; only referring to opinions: Dannecker, Art. 71 AEUV, para 5.

### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

Further *general* remarks on the meaning of the phrases addressed above cannot be made, but a concrete analysis of these terms and their respective meanings – and in particular of whether they include soft law powers – is to be undertaken case by case.<sup>984</sup>

### 3.5.2. Institutions

### 3.5.2.1. Commission

# 3.5.2.1.1. Support of decision-making/rule-making

According to the Treaties, the Commission contributes to the adoption of EU soft law in various ways. One of these ways is laid down in Article 223 para 2 TFEU. While para 1 *leg cit* is the legal basis for regulating the EP elections procedure, para 2 empowers the EP to adopt regulations on its own initiative on the 'regulations and general conditions governing the performance of the duties of its Members'.985 It shall do so in the course of a special legislative procedure after seeking an opinion from the Commission and with the consent of the Council. While the Commission is free to adopt or not to adopt an opinion986 (*argumentum* 'seeking') and, where the Commission in fact has delivered an opinion, the EP is free to follow or not to follow it, the EP has to ask for an opinion by all means. The same applies, *mutatis mutandis*, to Article 228 para 4 TFEU, according to which the EP shall lay down the regulations and general conditions governing the performance of the duties of the European Ombudsman 'after seeking an opinion from the Commission and with the consent of the Council'.<sup>987</sup>

<sup>984</sup> See Hofmann/Rowe/Türk, Administrative Law 549, with further examples and referring to the respective unclear wording.

<sup>985</sup> See in particular Decision 2005/684/EC, Euratom of the European Parliament of 28 September 2005 adopting the Statute for Members of the European Parliament.

<sup>986</sup> Note that the German version of this provision allows for an *Anhörung* of the Commission, not for a *Stellungnahme* (which is the translation of the word 'opinion' as used in Article 288 TFEU). In other language versions of the TFEU ('avis' in French, 'dictamen' in Spanish) the same term is used in both provisions. This indicates that also Article 223 para 2 TFEU is referring to an opinion within the meaning of Article 288 TFEU, which may – the term *Anhörung* so suggests – also be delivered orally by the Commission. The oral delivery of an EU soft law act is exceptional, but not excluded in principle.

<sup>987</sup> Note that in this – as compared to Article 223 – similarly worded provision the term *Stellungnahme* (corresponding to Article 288 TFEU) is used in the German version.

Also in the ordinary legislative procedure the Commission, apart from its initiating proposal, has means at its hands to influence the procedure. The Commission may render an opinion on the amendments (possibly) sugges‐ ted by the EP at first and/or at second reading, and/or by the Council at first reading.<sup>988</sup> While the EP is not legally bound by these opinions, the Council's quorum is raised to unanimity if it intends to adopt amendments on which the Commission has delivered a negative opinion;<sup>989</sup> where the opinion is positive, a qualified majority in the Council suffices. Against this background, also the question whether or not the EP considers the Commission's position (uttered in the proposal or in an opinion) is likely to have a significant influence on the course of the legislative proceedings at issue. The unanimity requirement in the Council is the procedural per‐ petuation, as it were, of Article 293 para 1 TFEU, according to which the Council, when acting on a proposal from the Commission, generally may amend it only by unanimous action.990 In the ordinary legislative procedure this rule does not apply in the conciliation phase.991 Where exceptionally it is not the Commission initiating the ordinary legislative procedure, but a group of MS, the ECB or the CJEU, a negative opinion of the Commission does not change the majority requirements in the Council.<sup>992</sup>

Another instance of the Commission's soft law power in the context of rule-making is its competence to give, in the context of the request of a group of MS to establish enhanced cooperation between themselves within the framework of CFSP, its opinion 'in particular on whether the enhanced cooperation proposed is consistent with other Union policies'. This opinion shall be forwarded to the EP 'for information'.993 Where enhanced cooper‐ ation is proposed in other policy fields, the Commission may submit a proposal to the Council to that effect.994 It may, via this measure, bring in its views on the project. It is also free not to submit a proposal at all when informing the MS of the reasons for not doing so.<sup>995</sup>

According to Article 19 para 2 of the Statute of the European Investment Bank (EIB), the Commission shall submit an opinion on applications to

<sup>988</sup> Article 294 paras 6, 7c and 9 TFEU; see graph in Schütze, Constitutional Law 172.

<sup>989</sup> Article 294 para 9 TFEU.

<sup>990</sup> See Krajewski/Rösslein, Art. 294 AEUV, para 55.

<sup>991</sup> Article 293 para 1 TFEU, with references to further exceptions.

<sup>992</sup> Article 294 para 15 subpara 1 TFEU.

<sup>993</sup> Article 329 para 2 TFEU.

<sup>994</sup> Article 329 para 1 subpara 1 TFEU.

<sup>995</sup> Article 329 para 1 subpara 1 TFEU.

the EIB for an investment made by a MS or by an undertaking. Where the Commission delivers an unfavourable opinion, it is legally non-binding, but the Board of Directors – the competent decision-making body of the EIB – may only grant the financial means at issue by unanimous decision (instead of the majority according to Article 10 para 2 of the Statute), the director nominated by the Commission abstaining.<sup>996</sup> The effect of a nega‐ tive Commission opinion on decision-making by the Board of Directors is equivalent to that on decision-making by the Council in the ordinary legislative procedure (see above). Also the Management Committee of the EIB shall deliver an opinion (see 3.5.3.1. below).

3.5.2.1.2. Initiation of (soft) decision-making/rule-making

In the field of economic policy coordination, the Commission plays an important role in initiating soft regulation by the Council. According to Article 121 para 2 TFEU, the Council shall formulate a draft for the broad economic policy guidelines (BEPG) of the MS and of the EU on a rec‐ ommendation by the Commission. In practice, the Commission regularly submits a draft recommendation to the Council for discussion on the basis of a report by the EFC.997 Eventually, and considering the discussion, the Commission issues a (revised) recommendation for draft BEPG which is then handled by the (European) Council in accordance with Article 121 para 2 subparas 2 f TFEU.

Where the economic policies of a MS do not comply with its respective BEPG or where they risk 'jeopardising the proper functioning of [EMU]', the Commission may address a warning to the MS concerned.998 The

<sup>996</sup> Article 19 para 6 of the Statute of the EIB.

<sup>997</sup> See Häde, Art. 121 AEUV, para 3.

<sup>998</sup> Article 121 para 4 TFEU; for the legal irrelevance of such a warning see Kempen, Art. 121 AEUV, para 22; see, however, Bandilla, Art. 121 AEUV, para 28, who classifies the Commission's warning as a 'recommendation', but stresses that the Commission with this tool cannot demand more from the MS addressed than that it takes the 'notwendigen Maßnahmen' [necessary measures]. It is reserved to the Council to give more detailed advice. In that sense, it is doubtful whether the Commission's warnings reach the degree of concreteness required for a soft law act in the first place (see II.2.1.2. above); for the regularly stricter wording of warnings as compared to recommendations (in the context of national law) see Feik, Verwaltungskom‐ munikation 427; for administrative warnings, and the different content they may display (also: merely factual statements) more generally see ibid 28 f.

Council may address the 'necessary recommendations' to the respective MS on a recommendation from the Commission.<sup>999</sup> The Council has a discretion as to whether or not it launches recommendations (*argumentum* 'may'),<sup>1000</sup> but if it decides in the affirmative a Commission recommenda‐ tion is required first. The Council may make its recommendations public on a proposal from the Commission. Since the Council cannot act under para 4 without an according recommendation/proposal by the Commis‐ sion, Article 135 TFEU provides that in case of the Commission's inaction the Council or a MS may request the Commission to make a recommen‐ dation or a proposal, respectively. The Commission is not bound by this request,<sup>1001</sup> but shall examine it and submit its conclusions to the Council without delay.<sup>1002</sup>

With regard to the excessive deficit procedure according to Article 126 TFEU, the Commission shall address an opinion to a MS where it considers that an excessive deficit in this MS exists or may occur. It shall inform the Council accordingly.1003 With this opinion, the 'politische Entscheidung' [political decision] to open an excessive deficit procedure is taken.1004 That this opinion is a requirement for the Council's decision (taken upon a Commission proposal) according to para 6, can be deduced from the order of paragraphs which – in Article 126 – suggests a chronological order of actions. That the Council decides upon a Commission proposal – under the predecessor provision in the TEC (Nice) it decided upon a Commission recommendation1005 – means that it may now deviate from

1004 Häde, Art. 126 AEUV, para 37.

<sup>999</sup> For the example of Ireland's pro-cyclical fiscal policies see Hodson/Maher, Soft law 804. The Treaty leaves it open whether the Council recommendations are an alternative to the Commission warning or whether they may (or can only) be adopted in addition to the warning. Article 6 para 2 subpara 2 of Council Regulation 1466/97 provides that the Council shall adopt its recommendations 'within 1 month of the date of adoption of the warning'. This suggests that a Commission warning is a necessary procedural requirement for the (subsequent) adoption of Council recommendations.

<sup>1000</sup> See Häde, Art. 121 AEUV, para 13; Thomas Müller, Art. 121 AEUV, para 35.

<sup>1001</sup> See Kempen, Art. 121 AEUV, para 21; Zahradnik, Art. 135 AEUV, para 3.

<sup>1002</sup> Article 135 (2nd sentence) TFEU.

<sup>1003</sup> Article 126 para 5 TFEU; for the – as compared to the standard infringement procedure under EU law, namely the Treaty infringement procedure – subordinate role of the Commission see Gil Ibáñez, Supervision 109.

<sup>1005</sup> See Seidel, Economic and Monetary Union 34.

the Commission's suggestion only by unanimous decision.1006 Where the Council has taken a decision determining the existence of an excessive defi‐ cit, it shall, on a recommendation from the Commission, but without being bound by it,<sup>1007</sup> adopt recommendations addressed to the respective MS with the aim to remedy the excessive deficit.<sup>1008</sup> These recommendations may be made public by the Council, again upon a recommendation by the Commission,<sup>1009</sup> according to para 8. Also where the Council decides to give notice to an inactive MS according to para 9, where it decides to apply or intensify measures according to para 11 or where it abrogates decisions or recommendations referred to in paras 6 to 9 and 11, it shall act only upon a recommendation from the Commission.<sup>1010</sup> Once adopted, the Council is not entitled to amend the recommendations without a 'fresh recommendation' from the Commission.<sup>1011</sup> Also with regard to Article 126 (except for para 14), Article 135 TFEU applies: The Council or a MS may request the (up to then inactive) Commission to make a recommendation or a proposal, respectively (see above).

Article 143 TFEU deals with the scenario that a non-euro MS – a MS 'with a derogation' – is, for certain reasons, in difficulties or is seriously threatened with difficulties as regards its balance of payments. Where these difficulties 'are liable in particular to jeopardise the functioning of the internal market or the implementation of the common commercial policy', the Commission shall investigate the position of the respective MS and the actions taken or to be taken by it. Where necessary, the Commission shall recommend to the MS which measures to take. This recommendation – as a piece of soft law – is neither legally binding upon the MS nor is it a

<sup>1006</sup> Article 293 para 1 TFEU; see Häde, Wirtschafts- und Währungsunion 203. The increased authority of Commission proposals is also highlighted by the fact that the requirements for a qualified majority in the Council are lower when it decides upon a Commission proposal: see Article 238 paras 2 f TFEU. It is due to this difference between proposal and recommendation that the latter was described as a 'abgeschwächte Variante des Initiativmonopols der Kommission' [lessened form of the Commission's monopoly of initiative]; Simon, Art. 148 AEUV, para 19.

<sup>1007</sup> See case C-27/04 *Commission v Council*, para 91.

<sup>1008</sup> Article 126 para 7 TFEU.

<sup>1009</sup> Article 126 para 13 subpara 1 TFEU. According to Article 121 para 4 TFEU, a Commission *proposal* is required for the publication of Council recommendations.

<sup>1010</sup> Article 126 para 13 subpara 1 TFEU.

<sup>1011</sup> Case C-27/04 *Commission v Council*, para 92, with regard to the predecessor pro‐ vision of Article 126 para 7 TFEU. This finding can also be applied to other cases in which the Council decides/recommends upon a Commission recommendation.

requirement for the MS to act.1012 However, as subpara 2 *leg cit* suggests, compliance by the MS addressed with the Commission recommendation, if it has adopted it at all,<sup>1013</sup> is required – *argumentum* '[i]f […] the measures suggested by the Commission do not prove sufficient' – for the Commis‐ sion to be allowed to recommend to the Council the granting of mutual assistance and appropriate methods therefor.<sup>1014</sup> The Council again may grant mutual assistance only upon a recommendation by the Commission (*argumentum* 'mutual assistance recommended by the Commission' in para 3 *leg cit*).<sup>1015</sup> It may as well refuse to grant mutual assistance. The Commis‐ sion recommendation, in other words, is a requirement for Council action, but does not prevent the Council from omitting to act, ie from not granting mutual assistance.

Where a sudden crisis in the balance of payments occurs and the Council does not immediately decide to grant mutual assistance, a non-euro MS may take the necessary protective measures as a precaution, of which the Commission and the other MS shall be informed.1016 Also in this case the Commission may recommend to the Council to grant mutual assistance according to Article 143 TFEU (see above). As well upon a Commission recommendation, the Council may decide that the MS concerned shall amend, suspend or abolish the protective measures taken.1017 Here again the

1016 Article 144 paras 1f TFEU.

<sup>1012</sup> See Bandilla, Art. 143 AEUV, para 13. This power of the Commission is rather a soft decision-making/rule-making power, a category of powers which is addressed under 3.5.2.1.3. below. Here it is mentioned as well only for the sake of contextuali‐ sation.

<sup>1013</sup> Where further MS measures would not be sufficient or where, after an examina‐ tion pursuant to subpara 1, there are no (further) adequate measures available, the Commission may immediately (in the first case: after hearing the EFC) recom‐ mend mutual assistance to the Council; see Häde, Art. 143 AEUV, para 6.

<sup>1014</sup> See Council Regulation 332/2002. A Commission proposal for a new regulation**,** COM(2012) 336 final/2, did not succeed.

<sup>1015</sup> For the MoU to be concluded between the Commission and the MS concerned in this case, and for the practical experiences with Article 143 TFEU see Flynn, Article 143 TFEU, paras 3 f; Dermine/Markakis, Bailouts.

<sup>1017</sup> Article 144 para 3 TFEU. Since the MS's protective measures must be limited to what is 'strictly necessary to remedy the sudden difficulties' (Article 144 para 1 TFEU), the protective measures become undue once the Council has decided to grant mutual assistance, because it thereby does away with the strict necessity of the protective measures. Nevertheless, provision is made for an explicit procedure for the Council to request the MS concerned to amend, suspend or abolish these measures. This is because the Council may apply para 3 without having granted

recommendation is a requirement for the Council to act, but the Council is also free not to act.

According to Article 148 para 4 TFEU, the Council, upon a recommenda‐ tion by the Commission, may make recommendations to MS with a view to facilitating compliance of their employment policies with the Council's guidelines for employment.<sup>1018</sup> The wording 'if it considers it appropriate' emphasises the Council's discretion, but even without this phrase it would be clear that the Council may, in spite of the Commission recommendation, decide not to act.<sup>1019</sup>

Another policy field in which the Commission disposes of soft law power to initiate a decision-making/rule-making process is the Common Commercial Policy (CCP). The opening of negotiations of agreements in this policy field with third countries or international organisations shall be authorised by the Council upon a recommendation by the Commission.<sup>1020</sup> This is in compliance with Article 218 TFEU, according to which the Commission or, where the agreement envisaged relates at least principally to CFSP, the High Representative of the Union for Foreign Affairs and Security Policy shall submit recommendations to the Council, which shall then authorise1021 the opening of negotiations and nominate either of them as Union negotiator or the head of the Union's negotiating team.1022 Special provisions apply in the context of agreements concerning monetary or foreign exchange regime matters. The Council, on a recommendation from the Commission (and only the Commission1023), shall decide the arrange‐ ments for the negotiation and for the conclusion of such agreements.1024 It


mutual assistance before; see Bandilla, Art. 144 AEUV, paras 11f; Häde, Art. 144 AEUV, para 2.

<sup>1018</sup> For the EU's traditionally weak, but – due to the intersection with many other policy fields – meaningful competence in the field of employment policy see Garben, Article 148 TFEU, para 2.

<sup>1021</sup> With regard to earlier (pre-Lisbon) Council practice, according to which it manda‐ ted the Commission by mere conclusions or in the minutes of its meetings see Lorenzmeier, Art. 218 AEUV, para 25a.

<sup>1022</sup> Article 218 para 3 TFEU. The Union negotiator may either be the Commission or the High Representative. Where the envisaged agreement affects the competences of both, they build a negotiating team with the more concerned of them being its head; see Schmalenbach, Art. 218 AEUV, para 10; see also Kaddous, Role 213.

<sup>1023</sup> See Palm, Art. 219 AEUV, para 55.

<sup>1024</sup> Article 219 para 3 TFEU.

does not necessarily have to negotiate itself, but may as well delegate this competence.<sup>1025</sup> The conclusion of formal agreements on an exchange-rate system for the euro in relation to the currencies of third States, and the adoption, adjustment or abandonment of the central rates of the euro within an exchange-rate system may all be done only by the Council – on a recommendation from the Commission (or the ECB).<sup>1026</sup> In the absence of exchange-rate systems in relation to third State currencies the Council, on a recommendation from the Commission (or the ECB), may formulate gen‐ eral orientations for exchange-rate policy in relation to these currencies.<sup>1027</sup> Also with regard to Article 219 TFEU, the Council or a MS may request the Commission to make a recommendation.<sup>1028</sup>

The Commission may also initiate the amendment of a selected part of the Statute of the ESCB and of the ECB. According to Article 40 para 2 of the Statute, the voting modalities for the Governing Council of the ECB as laid down in Article 10 para 2 *leg cit* may be amended by a unanimous decision of the European Council and subsequent approval by all MS, either on a recommendation of the ECB or on a recommendation of the Commission.

### 3.5.2.1.3. Soft decision-making/rule-making

The Commission's soft rule-making power is reflected, for example, in Arti‐ cle 60 TFEU. The liberalisation of services beyond the respective directives, in MS in which the general economic situation and the situation of the respective economic sector so permit, shall be facilitated by Commission recommendations addressed to these MS. These recommendations are le‐ gally non-binding, but have – in conjunction with the MS' obligation to 'endeavour to undertake the [further] liberalisation of services' laid down in Article 60 para 1 – a strong factual authority.1029 However, the significance

<sup>1025</sup> See Häde, Art. 219 AEUV, para 14; Kempen, Art. 219 AEUV, para 12; arguing in favour even of the possibility to delegate the competence to conclude such agree‐ ments: Palm, Art. 219 AEUV, para 56, with a further reference.

<sup>1026</sup> Article 219 para 1 TFEU: The unanimity requirement (subpara 1) and the endeav‐ our to reach a consensus consistent with the objective of price stability (subpara 2) required for a Council decision do not seem to leave any room for delegations.

<sup>1027</sup> Article 219 para 2 TFEU.

<sup>1028</sup> Article 135 TFEU.

<sup>1029</sup> See Müller-Graff, Art. 60 AEUV, para 3.

of Article 60 in practice is, due to the direct applicability of Articles 56 f TFEU, remarkably low.<sup>1030</sup>

Another example is taken from transport policy. According to Article 97 TFEU, MS shall endeavour to reduce progressively the costs charged by a carrier in respect of the crossing of frontiers and in addition to the transport rates. In order to facilitate that reduction, the Commission shall address recommendations to the MS. To date the Commission has not adopted any such recommendations. While, similar to Article 60, the obli‐ gation of the MS may increase the effectiveness of (potential) Commission recommendations, at the same time it is to be noted that the wording of this provision is vague ('reasonable level', 'taking […] into account', 'reduce […] progressively') and leaves the MS with a considerable latitude.<sup>1031</sup>

Where a MS intends to adopt or amend a measure which may cause a distortion of the conditions of competition in the internal market within the meaning of Article 116 TFEU,1032 it shall consult the Commission which shall recommend to the MS concerned measures appropriate to avoid the distortion in question.1033 In case of non-compliance with the recommenda‐ tion, other MS shall not be required according to Article 116 to amend their own provisions in order to eliminate such distortion. Where the MS concerned by ignoring the recommendation causes distortion detrimental only to itself, Article 116 shall not apply.1034 Where the above distortion has already materialised (*argumentum* 'is distorting'), the Commission shall, in accordance with Article 116 para 1, consult the MS concerned.1035 The practical relevance of Articles 116 f TFEU is low.<sup>1036</sup>

In the situation described in Article 143 para 1 TFEU – that a non-euro MS is in or seriously threatened with difficulties as regards its balance of payments and where such difficulties are liable in particular to jeopardise the functioning of the internal market or the implementation of the CCP – the Commission shall investigate the position of that MS and, as a form of soft decision-making, address a recommendation to it, setting out the

<sup>1030</sup> For the (now: historical) relevance prior to the end of the transitional period by 1 January 1970 see Randelzhofer/Forsthoff, Art. 60 AEUV, paras 1f.

<sup>1031</sup> See Rusche/Kotthaus, Art. 97 AEUV, paras 3 f.

<sup>1032</sup> In addition to that, it is required that the distortion 'needs to be eliminated'; see Article 116 para 1 TFEU.

<sup>1033</sup> Article 117 para 1 TFEU.

<sup>1034</sup> Article 117 para 2 TFEU.

<sup>1035</sup> For the legislative action that may follow see Article 116 para 2 TFEU.

<sup>1036</sup> See Korte, Art. 117 AEUV, para 14.

measures to be taken. For the granting of mutual assistance by the Council see 3.5.2.1.2. above.

Also in the field of social policy, the Commission, according to Article 156 TFEU, shall – in its facilitation of the coordination of MS action in all social policy fields mentioned in the respective chapter of the TFEU – deliver opinions after consulting the ESC.<sup>1037</sup>

Also in the Treaty infringement procedure the Commission disposes of soft decision-making powers. Where the Commission considers that a MS has failed to comply with an obligation under the Treaties it shall, after an informal procedure and after giving the MS concerned – by means of a so-called warning letter – the opportunity to submit its observations,<sup>1038</sup> deliver a reasoned opinion.<sup>1039</sup> Where this opinion is not complied with, the Commission may bring the case before the CJEU.<sup>1040</sup> Until the Court ac‐ tually hears the case, the Commission may, without explanation, 'withdraw or stop proceedings at any time'.1041 Where a MS considers that another

1038 See W Cremer, Art. 258 AEUV, para 6.


<sup>1037</sup> Note that the CJEU affirmed that the Commission may also issue binding meas‐ ures based on one of the predecessors of Article 156, namely Article 118 TEEC (Rome). It argued that Article 118 TEEC is to be read in conjunction with Article 117 TEEC (now: Article 151 TFEU) and that the approximation of provisions mentioned here 'gives the Commission the task of promoting close cooperation between Member States in the social field'; joined cases 281, 283 to 285 and 287/85 *Germany v Commission*, paras 11–13; see also Declaration 31 on Article 156 of the Treaty on the Functioning of the European Union in which the MS stress the 'complementary nature' of the EU measures laid down in this provision, but in which no explicit reference to this Court decision is made.

<sup>1039</sup> For the reasoning requirements for reasoned opinions see case C-223/96 *Commis‐ sion v France*, para 12: 'coherent and detailed statement'. It is questionable whether from the fact that this opinion is called 'reasoned opinion' enhanced reasoning requirements as compared to other opinions can be deduced. According to Article 296 para 2 TFEU, all legal acts shall 'state the reasons on which they are based'. According to the Court's case law the purpose of the reasoning of (individual-con‐ crete) acts is 'to enable the Court to review the legality of the decision and to provide the person concerned with sufficient information to make it possible to ascertain whether the decision is well founded or whether it is vitiated by a defect which may permit its legality to be contested'; case C-199/99P *Corus*, para 145, with further references. While Article 263 TFEU does not allow for the review of a true soft law act, there are possibilities to reach judicial review also in this context (see 6.2. below). In addition to that, it must be stressed that a sound reasoning may increase the factual authority of any act, which is particularly important for soft law acts.

MS is in breach of its obligations under the Treaties it may bring the case before the CJEU. Before doing so, however, it shall contact the Commission which may, after each of the MS concerned was given the chance to make submissions to the Commission, deliver a reasoned opinion. Where the Commission has not delivered an opinion within three months after the matter was brought before it, the MS itself may bring the matter before the CJEU.

One more competence of the Commission is to be mentioned under the heading 'Soft decision-making/rule-making', namely the Commission's power to adopt recommendations under Protocol No 31 of the Treaties. This Protocol is about the tariff preferences for certain petroleum products resulting from the association of the Netherlands Antilles with the EU. Pursuant to its Article 7, the Commission shall recommend administrative conditions, according to which – *inter alia* – the MS shall provide the Commission with the information necessary for the implementation of this Protocol.

### 3.5.2.2. Council

### 3.5.2.2.1. Support of decision-making

The Council's role as supporter of decision-making procedures by means of soft law is less comprehensive in the Treaties than that of the Commis‐ sion. Nevertheless, there are examples of procedures in which the Council contributes to decision-making by means of a soft law act. According to Article 140 para 2 TFEU, the Council shall, on a proposal from the Com‐ mission, decide which non-euro MS fulfils the necessary conditions to join the Eurozone,1042 and abrogate the derogations of the MS concerned. The Council shall act having received, within six months of receipt of the Commission proposal, a recommendation to decide in a certain way. This recommendation shall be adopted by the Council members constituting, untechnically speaking, the Eurogroup,1043 by a qualified majority of their

<sup>1042</sup> For these conditions see Article 140 para 1 TFEU.

<sup>1043</sup> See Protocol No 14 ('On the Euro Group') and Article 137 TFEU, respectively; see Palm, Art. 140 AEUV, para 13; note the Court's judgement in joined cases C-105/15P to C-109/15P *Mallis*, in which it held that 'the Eurogroup cannot be equated with a configuration of the Council or be classified as a body, office or

votes. Since they act chronologically after a Commission proposal, but not procedurally *on* a Commission proposal, and since they cannot in this case be called 'the Council', they are not, according to Article 293 para 1 TFEU, in any way bound by the Commission proposal.<sup>1044</sup> Nevertheless, the sup‐ portive role exercised by the euro-MS here shall – due to the institutional proximity – be listed under the heading 'Council'. Since the Council then still decides 'on a proposal from the Commission', it may deviate from it only by unanimity (Article 293 para 1 TFEU). A formal refusal of the proposal is not subject to the unanimity requirement, though.<sup>1045</sup>

# 3.5.2.2.2. Initiation of decision-making

In the field of EU external action the European Council shall adopt deci‐ sions on the strategic interests and objectives of the Union on the basis of the principles and objectives set out in Article 21 TEU.1046 To that end, the Council shall adopt a recommendation with the required majority (in CFSP matters principally that is unanimity, otherwise principally quali‐ fied majority1047), upon which the European Council shall unanimously adopt its decision. In this context, the High Representative (for the area of CFSP) and the Commission (for other areas of external action) may submit joint1048 proposals to the Council.1049 The Council recommendation is not a necessary requirement for the European Council to adopt its decision. The latter may as well decide on its own.1050 Where the Council has adopted

agency of the European Union within the meaning of Article 263 TFEU' (para 61); see also joined cases C-597/18P, C-598/18P, C-603/18P and C-604/18P *Chrys‐ ostomides*, para 87.

<sup>1044</sup> See also Häde, Art. 140 AEUV, para 44.

<sup>1045</sup> See Krajewski/Rösslein, Art. 293 AEUV, para 8, with further references and em‐ phasising the practical irrelevance of this case.

<sup>1046</sup> Article 22 para 1 TFEU. That no legislative acts may be adopted under the CFSP is expressly laid down in Article 31 para 1 (last sentence) TEU.

<sup>1047</sup> See H-J Cremer, Art. 22 EUV, para 11; Kaufmann-Bühler, Art. 22 EUV, para 7.

<sup>1048</sup> For the joint action of the Commission and the High Representative in the field of EU external action more generally see Ramopoulos/Wouters, Landscape 20, with examples <https://ghum.kuleuven.be/ggs/publications/working\_papers/new\_se ries/wp151-160/wp156-ramopoulos-wouters.pdf> accessed 28 March 2023; for the general regime of the initiation of decision-making in CFSP see Article 30 TEU.

<sup>1049</sup> Article 22 para 2 TFEU.

<sup>1050</sup> See Kaufmann-Bühler, Art. 22 EUV, para 7; Marquardt/Gaedtke, Art. 22, para 6.

a recommendation, the European Council shall not be bound, but shall retain the flexibility which is characteristic of this institution.<sup>1051</sup>

The Council's participation laid down in Article 283 TFEU concerns the selection and appointment of the members of the Executive Board of the ECB. According to para 2 *leg cit*, <sup>1052</sup> they shall be appointed by the Europe‐ an Council with a qualified majority from among persons 'of recognised standing and professional experience in monetary or banking matters' on a recommendation from the Council, after it has consulted the EP and the Governing Council of the ECB.<sup>1053</sup>

Also in budgetary matters the Council initiates, with its recommenda‐ tion, a decision, namely the decision of the EP to give a discharge to the Commission in respect of the implementation of the budget.<sup>1054</sup> While the recommendation is legally not binding upon the EP, the Commission shall 'take all appropriate steps' to act on, among other acts, the comments<sup>1055</sup> accompanying the recommendations on discharge adopted by the Council and, at the request of the EP or the Council, report on the measures taken in this context.<sup>1056</sup>

### 3.5.2.2.3. Soft decision-making/rule-making

The Council disposes of a soft rule-making competence in the framework of Article 7 TEU, laying down the procedure against a MS (about to) breaching the values referred to in Article 2 TEU: human dignity, freedom, democracy, etc. Under this 'last resort'1057 regime the Council may, follow‐

<sup>1051</sup> See Kaufmann-Bühler, Art. 22 EUV, para 7; with regard to the European Council's role in EU external relations: Wouters/Coppens/De Meester, Relations 149 f.

<sup>1052</sup> See Article 11 para 2 of the Statute of the ESCB and of the ECB, referring to Article 283 para 2 TFEU.

<sup>1053</sup> For the required 'common accord' in the European Council under the TEC and its merits see Endler, Zentralbank 425–430.

<sup>1054</sup> Not only the Council's recommendation, also the EP's decision not to give a discharge to the Commission lacks legal consequences; see Waldhoff, Art. 319 AEUV, para 3.

<sup>1055</sup> On these comments see Niedobitek, Art. 319 AEUV, para 19.

<sup>1056</sup> Article 319 para 3 TFEU; on documents accompanying soft law instruments more generally see Coman-Kund/Andone, Instruments 188.

<sup>1057</sup> Commission, 'A new EU Framework to strengthen the Rule of Law' (Communica‐ tion), COM(2014) 158 final/2, 6. For the original version of Article 7 TEU pursuant to the Treaty of Amsterdam see Pernice, Constitutionalism 735–738.

ing the procedure laid down in para 1 *leg cit*, determine that there is a 'clear risk of a serious breach'<sup>1058</sup> of these values by a MS. Before that, the Council shall, however, hear the respective MS and it may address recom‐ mendations to it in order to remedy the situation upfront.<sup>1059</sup> These recom‐ mendations shall be adopted in accordance with the same procedure as that provided for the determination of a risk of a breach of values. Hence recom‐ mendations shall be proposed, thereby providing the reasons, by one-third of the MS, the EP or the Commission, and shall be adopted by the Council with a four-fifths majority of its members after obtaining the consent of the EP.<sup>1060</sup> While the recommendations are legally non-binding,<sup>1061</sup> they dispose of high political authority.<sup>1062</sup> The procedure pursuant to Article 7 TEU, although initiated by the Commission and the EP, respectively, in the case of Poland and Hungary, has not been advanced so far.<sup>1063</sup> The literature in particular on the Council's competence to adopt recommendations under Article 7 TEU is scarce.<sup>1064</sup>

<sup>1058</sup> For this term see Commission, Communication on Article 7 of the Treaty on European Union. Respect for and promotion of the values on which the Union is based, COM(2003) 606 final, 7 f.

<sup>1059</sup> On the political risks of such recommendations see Voet van Vormizeele, Art. 7 EUV, para 8; see also Commission, Communication on Article 7 of the Treaty on European Union. Respect for and promotion of the values on which the Union is based, COM(2003) 606 final, 7.

<sup>1060</sup> The EP's consent is not redundant in case it has itself proposed the recommenda‐ tion; after all, the Council may have drafted its recommendations differently. For the exclusion of the MS concerned from the vote and from the calculation, and for the two-thirds majority of the votes cast required in the EP respectively, see Article 354 para 4 TFEU.

<sup>1061</sup> See Ruffert, Art. 7 EUV, para 12, who qualifies these recommendations as recom‐ mendations within the meaning of Article 288 TFEU.

<sup>1062</sup> See former President of the Commission *José Manuel Barroso* who laments the missing bridge between 'political persuasion and targeted infringement' on the one hand, and 'the nuclear option of Article 7 of the Treaty', Speech before the EP on 11 September 2013, Strasbourg <http://europa.eu/rapid/press-release\_SPEECH-13-68 4\_en.htm> accessed 28 March 2023.

<sup>1063</sup> See Kochenov, Article 7, 127. For the 'pre-Article 7 procedure' (*Viviane Reding*; <http://www.politico.eu/article/hungary-eu-news-article-7-vote-poland-rul e-of-law/> accessed 28 March 2023) envisaged by the Commission see Commis‐ sion, 'A new EU Framework to strengthen the Rule of Law' (Communication), COM(2014) 158 final/2; see also Halmai, Possibility 4 f.

<sup>1064</sup> See Schorkopf, Art. 7 EUV, paras 22–24, where these recommendations do not seem to be mentioned; in para 26 Council recommendations are referred to which may be adopted in conjunction with (not: prior to) the decision according to

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

Article 121 TFEU provides for the power of the Council to adopt rec‐ ommendations in two instances. According to para 2, the Council shall – on the basis of the European Council's conclusion on the draft BEPG submitted by the Council – adopt a recommendation setting out these broad guidelines and shall inform the EP thereof. The conclusion of the European Council does not require a full approval, but the European Council may suggest certain amendments.<sup>1065</sup> While the BEPG are referred to as guidelines, in fact they constitute – according to the explicit wording of the TFEU – a recommendation.<sup>1066</sup> With regard to the parts of the BEPG which only concern the Euro area generally, non-euro MS shall not vote in the Council.<sup>1067</sup> According to para 4 *leg cit*, where the economic policies of a MS are not consistent with the BEPG or where they risk jeopardising the proper functioning of the EMU, the Commission may address a warning to the MS concerned. At the same time, the Council may address the necessa‐ ry recommendations to this MS.1068 It shall do so upon a recommendation from the Commission.1069 Where the Council adopts a recommendation in reaction to non-compliance by a MS with the BEPG, it has the character of a soft reaction to non-compliance with non-binding norms.

According to Article 126 para 7 TFEU, in the context of the excessive deficit procedure, where the Council determines that an excessive deficit exists, it shall, on a recommendation from the Commission, address recom‐ mendations to the MS concerned 'with a view to bringing that situation to an end within a given period'.1070 Where there has been 'no effective action

Article 7 para 1 TEU. For a brief account of these recommendations see Besselink, Bite 133 f.

<sup>1065</sup> See Bandilla, Art. 121 AEUV, paras 11f; Hattenberger, Art. 121 AEUV, para 10, with further references.

<sup>1066</sup> *Argumentum* 'the Council shall adopt a recommendation setting out these broad guidelines' (Article 121 para 2 TFEU). This example could be brought forward in favour of the argument that Article 288 TFEU provides a *numerus clausus* of legal acts, and that acts named otherwise – in particular in the realm of soft law – may be assigned to one of the five 'legal acts' set out therein; see 3.1.2. above.

<sup>1067</sup> Article 139 para 2 lit a TFEU; see also Article 136 para 1 TFEU.

<sup>1068</sup> On the rare use the Council makes of this power see Schulte, Art. 121 AEUV, paras 49 f.

<sup>1069</sup> On a *proposal* from the Commission the Council may decide to make its recom‐ mendations public. For the suspended voting rights for non-euro MS in case of recommendations made *vis-à-vis* members of the Eurozone see Article 139 para 4 lit a TFEU.

<sup>1070</sup> For the Council's discretion to deviate from the Commission's recommendation see case C-27/04 *Commission v Council*, para 80.

in response' to this recommendation on the part of the MS, the Council may publish its recommendations.<sup>1071</sup> While these recommendations are legally non-binding themselves, their purpose – to reach a sound treasury – coincides with the obligation of MS to 'avoid excessive government deficits' (Article 126 para 1 TFEU).<sup>1072</sup>

According to Article 148 para 4 TFEU, the Council shall examine the implementation of the employment policies of the MS in view of the Council's guidelines for employment,<sup>1073</sup> the result of negotiations also referred to as OMC.<sup>1074</sup> In this context, it may, on a recommendation from the Commission, make recommendations to MS. A joint annual report which the Council and the Commission shall, on the basis of the results of that examination, submit to the European Council may be perceived as a follow-up measure which does not constitute a sanction, but increases awareness also of non-compliance (with the guidelines for employment, but possibly also with the Council recommendations) by certain MS.<sup>1075</sup>

Further competences of the Council to adopt recommendations are laid down in Articles 165 para 4, 166 para 4 (education, vocational training, youth and sport), 167 para 5 (culture), 168 para 6 (public health) TFEU. In these cases the Council *shall* (according to Article 168 para 6: *may*) adopt recommendations on a proposal from the Commission. These recommen‐ dations have a (soft) decision-making/rule-making purpose.<sup>1076</sup>

### 3.5.2.3. European Council – soft rule-making

In the field of Common Security and Defence Policy (CSDP), the Europe‐ an Council may decide on the progressive framing of a common Union

<sup>1071</sup> Article 126 para 8 TFEU.

<sup>1072</sup> See Häde, Aussetzung 758, with regard to case C-27/04 *Commission v Council*.

<sup>1073</sup> For the legal qualification of these guidelines outside the framework of Article 288 TFEU see Knauff, Regelungsverbund 311.

<sup>1074</sup> See Article 148 para 2 TFEU; see also Garben, Article 148 TFEU, para 1.

<sup>1075</sup> On the 'Gruppendruck' [peer pressure] and the '"politisch-psychologische" Wir‐ kung' [politico-psychological effect] these recommendations may entail see Steinle, Beschäftigungspolitik 371; for the multi-step peer review procedure applied by the ESMA as another example see van Rijsbergen, Legitimacy 228 ff; for the largely positive perception of peer review on the part of the MS in general see Dawson, Soft Law 15.

<sup>1076</sup> Suggesting a 'besonderen Stellenwert' [particular importance] of this measure in the given context: Niedobitek, Art. 165 AEUV, para 62.

defence policy. It shall in this case recommend to the MS the adoption 'of such a decision in accordance with their respective constitutional requir‐ ements'.<sup>1077</sup> In the literature, this is seen as the 'ratification' of a *de facto* simplified Treaty revision procedure.<sup>1078</sup> The recommendation is a mere hint at the necessity of 'ratification' at the national level. The bindingness of this recommendation is unclear, and arguably is to be examined case by case, so as to ensure certain flexibility (also for the MS) in CSDP.<sup>1079</sup>

### 3.5.2.4. European Parliament – support of decision-making/rule-making

In the field of CFSP, the EP may not only address questions,<sup>1080</sup> but may also make recommendations to the Council and to the High Representative. These recommendations are regularly contained in the EP's resolutions.<sup>1081</sup>

In the context of international agreements concluded, for one part, by the EU, the EP may have a right to be consulted.1082 In this case, the EP shall deliver its opinion, within an appropriate time limit set by the Council or otherwise, in accordance with the principle of sincere cooperation,<sup>1083</sup> within a reasonable time. The Council which shall conclude the respective agreement has to consider the opinion. Where the EP does not deliver an opinion in due time, the Council shall proceed. In case of agreements relating exclusively to the CFSP no involvement of the EP is provided for.<sup>1084</sup>

<sup>1077</sup> Article 42 para 2 subpara 1 TEU.

<sup>1078</sup> H-J Cremer, Art. 42 EUV, para 10; Marquardt/Gaedtke, Art. 42 EUV, para 7.

<sup>1079</sup> See Isak, Art. 42 EUV, para 69; Kaufmann-Bühler, Art. 42 EUV, paras 27 f.

<sup>1080</sup> This right to pose questions entails a duty of those addressed to respond accord‐ ingly; see H-J Cremer, Art. 36 EUV, para 9.

<sup>1081</sup> See Marquardt/Gaedtke, Art. 42 EUV, para 6, also with regard to the follow-up measures of the Council.

<sup>1082</sup> The EP shall be consulted in case of international agreements other than those for the conclusion of which the EP's consent is required; Article 218 para 6 lit a and b TFEU.

<sup>1083</sup> See case C-65/93 *European Parliament v Council*, para 23.

<sup>1084</sup> Article 218 para 6 TFEU.

3.5.2.5. Court of Justice of the European Union (and its components) – support and initiation of and actual (soft) decision-making/rulemaking?

Article 218 para 11 TFEU provides for the possibility for a MS, the EP, the Council or the Commission to request from the CJEU an opinion on whether an international agreement envisaged by the EU is compatible with the Treaties.<sup>1085</sup> Thereby the MS or the named institutions may clarify *ex ante* – where 'purpose and broad outline of the agreement' are clear: even before negotiations are taken up<sup>1086</sup> – whether or not the agreement to be concluded is in accordance with primary law.<sup>1087</sup> The opinion of the Court is legally binding in the sense that where it is adverse, the agreement 'may not enter into force unless it is amended or the Treaties are revised'.<sup>1088</sup> This makes it clear that the Court's opinion here is not to be qualified as an opinion within the meaning of Article 288 TFEU, but as an opinion *sui generis*. This is palpable when having a look at the German version of the TFEU in which – unlike eg in the French, the Italian or the Spanish version – different terms are used for these two kinds of output: *Gutachten* (Article 218 para 11 TFEU) and *Stellungnahme* (Article 288 TFEU). Thus, the Court's opinion according to Article 218 para 11 TFEU does not require further analysis in this context, as it is clear that a legally binding effect on those involved in the negotiation/conclusion of the agreement at issue is intended.

Also the 'unanimous opinion of the Judges and Advocates-General of the Court of Justice' referred to in Article 6 para 1 of Protocol No 3 on the Statute of the CJEU does not count as an opinion according to Article 288 TFEU. The 'unanimous opinion' rather is an expression of the quota required for a decision of the Court to deprive a judge/AG of his/her office or of his/her right to a pension or other benefits in its stead, namely: unanimity of all judges and AG of the CJEU apart from the judge/AG<sup>1089</sup>

<sup>1085</sup> That means that requesting an opinion from the Court under Article 218 para 11 TFEU is not obligatory; see Lenaerts/Maselis/Gutman, Procedural Law 555.

<sup>1086</sup> Opinion 1/03 *Lugano Convention*, para 111, with references to further case law.

<sup>1087</sup> For the comprehensive scope of judicial review in this context see Lorenzmeier, Art. 218 AEUV, paras 72 and 75.

<sup>1088</sup> Article 218 para 11 TFEU.

<sup>1089</sup> Article 8 of the Statute of the CJEU.

concerned. This is again underpinned when having a look at other language versions of the provision.<sup>1090</sup>

The following example is (partly) of historical importance only, as the former Civil Service Tribunal was dissolved on 1 September 2016.<sup>1091</sup> In Article 3 para 2 of Annex I of the Statute of the CJEU, dealing with the Civil Service Tribunal (now repealed by Article 2 para 3 of Regulation 2016/1192), reference is made to a recommendation by the CJEU, upon which the Council should lay down the conditions and the arrangements governing the submission and processing of applications for the position of a judge of the Civil Service Tribunal. In the context of this procedure a committee of seven former judges of the Court of Justice/General Court or (other) lawyers of recognised competence was established, the membership and operating rules of which were determined by the Council upon a rec‐ ommendation by the President of the Court of Justice. This exemplifies the influence the Court has/had when its own affairs are/were regulated. The Council was not legally bound by the recommendations, but – stemming from the Court, the highest judicial authority of the EU, and concerning in the first place a part of the Court, namely the Civil Service Tribunal<sup>1092</sup> – these recommendations arguably carried a high degree of authority.

The committee provided an opinion on the candidates' suitability to perform the duties of a judge of the Civil Service Tribunal and proposed at least twice as many candidates as there were judges to be appointed, namely those with the 'most suitable high-level experience'.1093 Here argua‐ bly the Council was – *de facto*, not legally – bound by the committee's opinion,1094 which is an opinion according to Article 288 TFEU.1095 This

<sup>1090</sup> German: *Stellungnahme* (Article 288 TFEU) – *Urteil* (Protocol No 3); French: *avis* – *jugement*; Italian: *parere* – *giudizio*; for the authenticity of all official language versions of an act of EU law, and the potential need for comparison, see case C-561/19 *Consorzio Italian Management*, paras 42-44; for the dismissal of AG *Sharpston* in the context of *Brexit* see Kochenov/Butler, Independence.

<sup>1091</sup> For the (political) background to this transformation see <http://curia.europa.eu/j cms/jcms/T5\_5230/en/> accessed 28 March 2023.

<sup>1092</sup> The Civil Service Tribunal was attached to the General Court; see Lenaerts/Mase‐ lis/Gutman, Procedural Law 34.

<sup>1093</sup> Article 3 para 4 of the Annex to the Statute of the CJEU.

<sup>1094</sup> This is also expressed by the manual on the *Civil Service Tribunal* (2014) issued by the CJEU's press and information department which at page 2 says that the Council appointed the judges 'on the proposal of the committee'.

<sup>1095</sup> This follows also from other language versions of Article 3 para 4 of the Annex to the Statute of the CJEU: *Stellungnahme*, *avis*, *parere*.

was a result of the committee's authoritative composition, also in terms of independence.<sup>1096</sup> The fact that the committee had to name twice as many candidates as there were posts (which means that the Council could – with‐ in a certain frame – choose) made the opinion's factual bindingness appear less compelling. It could be argued that the committee did not form part of the CJEU and that hence its opinion is to be listed below under the heading 'EU-external actors' (3.5.3.2.). However, due to the institutional (and also personal) proximity to the CJEU this committee had, its consideration in the context of the CJEU (and its components) appears appropriate.

A similar procedure is applied – and to that extent the above remarks are not only of historical interest – for the selection of the judges of the Court of Justice and the General Court, respectively.<sup>1097</sup> Also here the opinion of the panel bears a high authority. So far MS, in spite of numerous negative votes, have always followed the panel's opinions.1098 However, since the MS can nominate the candidates – with no possibility for candidates to submit direct applications, as was the case with the Civil Service Tribunal – the MS' influence on the selection of judges is much stronger.<sup>1099</sup>

Another instance of legally non-binding output issued by (a member of ) the CJEU1100 are the submissions delivered by the AG. <sup>1101</sup> The submissions are also referred to as Opinion1102 or View,1103 sometimes also as propos‐

<sup>1096</sup> It ought to be mentioned, however, that the committee has close links to the Council in terms of appointment, secretarial support and finances; see above and the Annex to Council Decision 2005/49/EC, Euratom concerning the operating rules of the committee provided for in Article 3(3) of Annex I to the Statute of the CJEU.

<sup>1097</sup> See Article 255 TFEU and Council Decision 2021/2232 on the composition of the current panel.

<sup>1098</sup> See Seventh Activity Report of the panel provided for by Article 255 of the Treaty on the Functioning of the European Union (2022) 9 <https://curia.europa.eu/jcms/upload/docs/application/pdf/2022-07/2022.2597 qcar22002enn\_002.pdf> accessed 28 March 2023.

<sup>1099</sup> Critically: Burgstaller, Art. 255 AEUV, paras 11–13, with further references.

<sup>1100</sup> For the AG' forming part of the CJEU in spite of the wording of Article 252 TFEU see Jacobs, Advocates General 18, with further references; Lenz, Amt 721.

<sup>1101</sup> Article 252 TFEU; Article 20 para 4 of the Statute of the CJEU; see para 5 *leg cit* for cases in which the AG shall not deliver submissions. For further tasks of the AG see eg Hackspiel, Art. 252 AEUV, para 9.

<sup>1102</sup> This is the name which the written submissions regularly bear.

<sup>1103</sup> Eg: View of AG *Tizzano* in case C-27/04 *Commission v Council*; View of AG *Kokott* in case C-370/12 *Pringle*.

al.1104 The AG ought to 'assist' the judges.<sup>1105</sup> They may convince the Court as a result of their legal argumentation,<sup>1106</sup> but they cannot bind the Court in any way.<sup>1107</sup> In *Jacobs'* words: '[U]nlike a judgment, the Opinion does not decide the case, even provisionally: its purpose, according to the Treaties, is to assist the Court in the performance of its task'.<sup>1108</sup> Or, as *Léger*, another AG, has put it: 'The Advocate General is impartial, independent, influen‐ tial, yet at no point does the AG usurp the most fundamental judicial prerogative of deciding cases. No matter how eloquent, how persuasive an Opinion may be, it may be disregarded for, after all, Judges are grown-ups capable of making up their own minds'.<sup>1109</sup> While the AG seeks to influence the Court with his/her Opinion, an according duty to consider<sup>1110</sup> on the part of the Court is not provided for and arguably would be incompatible with the Court's independence and its authoritative answering of legal


<sup>1104</sup> See Lenaerts/Maselis/Gutman, Procedural Law 776.

<sup>1105</sup> Articles 19 TEU and 252 TFEU; with regard to the General Court see Article 254 para 1 TFEU; see also Article 49 para 2 of the Statute of the CJEU, according to which – in the context of procedures before the General Court – the 'reasoned submissions on certain cases […] [shall] assist the General Court in the perform‐ ance of its task'. This relationship of assistance is also reflected in Declaration 38 attached to the Treaties, according to which the Court may request an increase in the number of AG by three. The Court has actually done so which lead the Council to increase, by its Decision 2013/336, the number of AG to nine by 1 July 2013 and to eleven by 7 October 2015. That the term 'assist' is used here cannot alter or relativise the AG' impartiality and independence (also from judges and other AG); Article 252 TFEU; see also Lenz, Amt 721.

questions.<sup>1111</sup> That the Court is obliged (Article 296 TFEU and Article 36 of the Statute of the CJEU) to provide a comprehensible reasoning for its decisions – and that it may thereby also take *into account* the arguments brought forward in the AG's Opinion<sup>1112</sup> (if only to rebut them<sup>1113</sup>) – is a different issue.<sup>1114</sup>

In conclusion, it is to be noted that Opinions of AG qualify as soft law, because they – according to their role as laid down in the Treaties – present a certain legal solution, as the final legal view of the AG, of a case and there‐ by are designed to have a steering effect. The actual strength of this effect hardly follows from the general authority of the act, but the AG's Opinion – stemming from a monocratic organ, a highly personalised act – may be (and regularly is) influential on the judgement only due to the persuasive power of its individual arguments. Therefore the effects are highly volatile. They necessarily vary from Opinion to Opinion, from case to case. The (varying) steering effect the Opinions of the AG have is entrenched in pri‐ mary law and hence it does not conflict with the independence of the CJEU which is, as well, laid down in primary law. The fact that the Court may not follow an Opinion in a certain case does not constitute a contradiction to the assisting role of the AG's submissions, it does not even relativise them. Also arguments which the Court decides to refuse may assist the Court in its work, in that also (explicitly or implicitly) refused legal arguments may increase the credibility/quality of the judgement. Thereby the AG contribute to 'ensur[ing] that in the interpretation and application of the Treaty, the law is observed'.1115 Also apart from the proposed legal solution

<sup>1111</sup> This lack of formal impact on the Court's decision is also reflected in the fact that parties do not have a right to respond to the AG's Opinion; see Lenaerts/Ma‐ selis/Gutman, Procedural Law 776, with reference to the case law. Also with regard to national courts, (soft) unsolicited interferences by EU bodies are generally considered problematic; see eg the Commission opinion rendered – upon request by the national court – according to Article 29 of Council Regulation 2015/1589; for the capacity of soft law to pose a risk to the Court's independence see Peters, Soft law 41.

<sup>1112</sup> For an empirical study of how often the Court has followed the Opinion of the AG see Tridimas, Role 1362–1365.

<sup>1113</sup> Note, however, that disagreement with the AG's Opinion may be expressed by the Court in many ways, also by not mentioning it at all; see Tridimas, Role 1371–1373.

<sup>1114</sup> For the importance of the Court's reasoning, and on the diverging views on it expressed in the literature, see in general Dawson, Court, on its 'reckonability' in particular 426 f; on its 'coherence' see Bengoetxea/MacCormick/Moral Soriano, Integration 64–81.

<sup>1115</sup> Order in case C-17/98 *Emesa Sugar*, para 13.

### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

(ideally to new points of law<sup>1116</sup>), namely with regard to the condensed presentation of the facts of the case and the elaboration of relevant legal questions (to be solved by the Court), the Opinion does have a supporting function.<sup>1117</sup>

3.5.2.6. European Central Bank

3.5.2.6.1. Support and initiation of rule-making

The ECB supports rule-making falling within its field of competence by providing its expertise on a large scale. This is reflected upon in particular in Article 127 para 4 TFEU. According to this provision, it shall be actively consulted, but it may also submit opinions to the relevant EU (and nation‐ al) actors on its own motion.

The ECB may act as the initiator of a rule-making procedure. This is confirmed (but not laid down) in Articles 289 para 4 and 294 para 15 TFEU with regard to legislative acts. According to Article 129 para 3 TFEU, certain provisions of the Statute of the ESCB and of the ECB may, as an exception to the regular procedure required for the amendment of primary law, be amended by the EP and the Council according to the ordinary legislative procedure. They shall act either on a recommendation from the ECB and after consulting the Commission or on a proposal from the Commission and after consulting the ECB. Other provisions ('[c]omplementary legislation'1118) of the Statute may be amended by the Council on a proposal from the Commission and after consulting the EP and the ECB, or on a recommendation from the ECB and after consulting the EP and the Commission (Article 129 para 4 TFEU).1119 The provisions of the Statute which may be amended in either of the two procedures concern 'technical' issues (sometimes of high relevance: eg Article 18 on open market and credit operations, Article 33 para 1 lit a on the allocation of net profits and losses of the ECB, or the sanctioning power of the ECB

<sup>1116</sup> See Article 20 para 5 of the Statute of the CJEU; see also Lenaerts/Maselis/Gut‐ man, Procedural Law 776.

<sup>1117</sup> See Jacobs, Advocates General 21.

<sup>1118</sup> Article 41 of the Statute of the ESCB and the ECB.

<sup>1119</sup> See also Article 40 para 1 and Article 41 of the Statue of the ESCB and the ECB; for the reduction of safeguards for the ECB brought about by the Treaty of Lisbon see Häde, Art. 129 AEUV, para 5.

under Article 34 para 3), not highly political issues such as the composition of the Governing Council or the calculation of the respective majorities required for decision-making. The latter is regulated by Article 10 para 2 of the Statute, which may be amended according to Article 40 para 2 of the Statute: by a unanimous decision of the European Council either on a recommendation by the ECB and after consulting the Commission and the EP, or on a recommendation by the Commission and after consulting the ECB and the EP. These amendments shall be approved by the MS in accordance with their respective constitutional requirements. In all these cases matters immediately concerning the ECB are at stake. Therefore the MS have provided for its competence to initiate the amendment of the respective rules. In this role, however, the ECB is not always on an equal footing with the Commission, since the Commission – according to Article 129 paras 3 f TFEU – may *propose* amendments, with the effect that the Council may, apart from the exceptions listed in Article 293 para 1 TFEU, only amend that proposal by a unanimous decision. While thereby at least formally the Commission has a stronger tool at hand (which is, with a view to other Treaty amendment procedures, and the ordinary legislative proce‐ dure respectively, systemically coherent1120), the ECB's recommendations still *de facto* bear considerable authority. In the amendment procedure laid down in Article 40 para 2 of the Statute, the Commission and the ECB may *both* initiate rule-making (only) by means of a recommendation.

Article 219 TFEU provides for a special procedure for the conclusion of formal agreements setting up exchange-rate systems for the euro in relation to the currencies of third States (see 3.5.2.1.2. above). In this con‐ text, the Council acts as the decision-maker, but may do so only on a recommendation by the ECB (or by the Commission after consulting the ECB). The Council shall 'endeavour to reach a consensus consistent with the objective of price stability',1121 an objective the realisation of which is first and foremost the task of the ECB.1122 The ECB's expertise with regard to this question is the reason why it is involved in the decision-making process here. The Council may, following the same procedure, adopt, adjust or abandon the central rates of the euro within the exchange-rate system. In the absence of an exchange-rate system in relation to certain third States the

<sup>1120</sup> See in particular Article 48 paras 2 and 6 TEU, and Article 294 para 2 TFEU, respectively.

<sup>1121</sup> Article 219 para 1 TFEU.

<sup>1122</sup> See Article 127 para 1 TFEU.

Council may, again following this procedure, formulate general orientations for the exchange-rate policy in relation to the respective currencies. Due to its relevant expertise, the ECB's viewpoints carry considerable weight.<sup>1123</sup>

According to Article 27 para 1 of the Statute of the ESCB and of the ECB, the Governing Council as the ECB's main decision-making organ shall rec‐ ommend independent external auditors for the auditing of the ECB and the national central banks (NCBs). The Council shall approve these auditors. The Governing Council's recommendations here affect the E(S)CB's own management and their adoption is subject to pre-defined procedures.<sup>1124</sup> This suggests that the Governing Council's recommendations (or rather: the ECB's recommendations<sup>1125</sup>) here bear a high degree of authority.

# 3.5.2.6.2. Soft decision-making

The ECB (via its Governing Council) may also act as a soft decision-maker itself, namely according to Article 35 para 6 of the Statute of the ESCB and of the ECB in conjunction with Article 271 lit d TFEU.1126 Where a NCB has, in the view of the ECB, failed to comply with its obligations under the Statute it shall 'deliver a reasoned opinion on the matter after giving the national central bank concerned the opportunity to submit its observations'. Where the NCB does not comply with this opinion within the period set by the ECB, the latter may bring the case before the CJEU. This procedure is a *lex specialis* of Article 258 TFEU.1127 Where the Court determines that the NCB has violated obligations under the Statute, the

<sup>1123</sup> Differentiated: Thiele, Operations, para 23.15, with further references.

<sup>1124</sup> ECB/Eurosystem, Good Practices for the selection and mandate of External Audi‐ tors according to Article 27.1 of the ESCB/ECB Statute (2017).

<sup>1125</sup> As an organ of the legal person ECB, the Governing Council acts for the ECB.

<sup>1126</sup> For one of the rare applications of the procedure laid down in Article 14.2. of the Statute of the ESCB and of the ECB see the ECB's invocation of the Court in the case of the Governor of the Latvian central bank being barred from holding his office at this central bank and from exercising his functions as a member of the ECB's Governing Council; see ECB, Press Release of 6 April 2018 <https:// www.ecb.europa.eu/press/pr/date/2018/html/ecb.pr180406.en.html> accessed 28 March 2023; see joined cases C‑202/18 and C‑238/18 *Ilmārs Rimšēvičs and European Central Bank v Republic of Latvia*, ECLI:EU:*C*:2019:139.

<sup>1127</sup> This arguably follows from the NCBs' independence and the special role the ECB plays within the ESCB; see also ECB/Eurosystem, Guide to consultation of the European Central Bank by national authorities regarding draft legislative provisions (October 2015) 27 (fn 22).

latter shall be required to take the necessary measures to comply with the Court's judgement.<sup>1128</sup> The ECB's reasoned opinion has the same effect as the Commission's reasoned opinion under Article 258 TFEU. For a more detailed analysis of the procedure see IV.2.1. below.

### 3.5.3. Other actors

#### 3.5.3.1. EU-internal actors

Not only the institutions, but also other EU bodies may adopt recommen‐ dations or opinions according to special competence clauses laid down in EU primary law. An example for this is the High Representative's compe‐ tence to address a recommendation to the Council where the conclusion of an agreement between the EU and third countries or international organi‐ sations which exclusively or principally relates to the CFSP is envisaged. The Council shall then adopt a decision authorising the opening of nego‐ tiations (see 3.5.2.1.2. above).

In the context of enhanced cooperation according to Part Six, Title III of the TFEU, the High Representative shall submit an opinion, namely where MS wishing to establish enhanced cooperation within the framework of CFSP have addressed a respective request to the Council. Whereas the High Representative shall provide an opinion on whether the enhanced cooperation proposed is consistent with the EU's CFSP, the Commission shall give its opinion in particular on whether it is consistent with the other Union policies (see 3.5.2.1.1. above). The Commission may also utter its point of view on the consistency with the CFSP and on other issues, though (*argumentum* 'in particular').1129 In that sense, the scope of issues possibly to be addressed by the Commission in its opinion is much wider than that of the High Representative. In terms of effects, however, no difference between the two opinions is intended by the Treaty.

According to Article 222 TFEU, the so-called solidarity clause, the Union and the MS shall act jointly in a spirit of solidarity if a MS is the object of a terrorist attack or the victim of a natural or man-made disaster. The arrangements for the implementation by the EU of the solidarity clause shall be laid down in a Council decision. The Council shall act on a joint

<sup>1128</sup> Article 271 lit d TFEU.

<sup>1129</sup> Article 329 para 2 TFEU; see Pechstein, Art. 329 AEUV, para 7.

proposal by the Commission and the High Representative.<sup>1130</sup> In this con‐ text, and explicitly without prejudice to the preparatory work of COREP‐ ER, the Political and Security Committee and the Standing Committee on Operational Cooperation on Internal Security referred to in Article 71 TFEU shall – 'if necessary' – adopt joint opinions.<sup>1131</sup> The necessity of a joint opinion arguably is to be determined by the two committees.

According to Article 271 lit a TFEU, the EIB is vested with the powers of the Commission under Article 258 TFEU with regard to MS' (non-)compli‐ ance with their obligations under the Statute of the EIB.<sup>1132</sup> This provision which confers on the EIB the competence to adopt a reasoned opinion – an EU soft law act – is the pendant of Article 271 lit d TFEU, vesting 'guardian powers' according to Article 258 TFEU upon the ECB. What was said in this context *mutatis mutandis* applies here, as well – in particular the *lex specialis* argument1133 (see 3.5.2.6.2. above).

The Statute of the EIB provides for soft law powers of the EIB, and its organs respectively, in a number of cases. The Management Committee, an organ of the EIB, shall, acting by a majority,1134 submit an opinion on proposals for raising loans or granting of finance.1135 Where the opinion

<sup>1130</sup> See Council Decision 2014/415/EU on the arrangements for the implementation by the Union of the solidarity clause; for the respective joint proposal from the Commission and the High Representative – JOIN(2012) 39 final – see Blockmans, L'Union 125–132. Where the decision has defence implications, the Council shall furthermore act in accordance with Article 31 para 1 TEU.

<sup>1131</sup> In Council Decision 2014/415/EU no reference is made to such a joint opinion which arguably means that no such opinion was delivered (see Article 296 para 2 TFEU). That the committees shall be involved also when the Council Decision based on Article 222 para 3 TFEU is amended is clear from the Treaty, but also explicated in Article 9 para 2 of this Decision.

<sup>1132</sup> Note what the Court said with regard to the double nature of the EIB and the applicability of then Community law: 'The position of the Bank is therefore am‐ bivalent inasmuch as it is characterized on the one hand by independence in the management of its affairs, in particular in the sphere of financial operations, and on the other hand by a close link with the Community as regards its objectives. It is entirely compatible with the ambivalent nature of the Bank that the provisions generally applicable to the taxation of staff at the Community level should also apply to the staff of the Bank'; case 85/86 *Commission v European Investment Bank*, para 30.

<sup>1133</sup> See Karpenstein, Art. 271 AEUV, para 4.

<sup>1134</sup> Article 11 para 4 of the Statute of the EIB.

<sup>1135</sup> Article 19 para 4 of the Statute of the EIB. For the financing activity of the EIB in practice see Becker, Investitionsbank, 'II. Tätigkeiten'; for its activities in a historical perspective see Skiadas, Court 216 f.

of the Management Committee is negative, the Board of Directors may grant the finance concerned only by unanimous decision.<sup>1136</sup> This opinion – adopted by and addressed to one of the organs of the EIB – is a body-in‐ ternal soft law act with no (direct) body-external effect. Where both the Commission and the Management Committee have launched a negative opinion, the Board of Directors may not grant the finance concerned.<sup>1137</sup> Hence, in combination, two unfavourable opinions – legally non-binding individually – have a prohibitive effect. While it could be argued that, taken together, the two negative opinions are – against the express wording of Article 288 TFEU – legally binding, the fact that these effects are laid down in primary law, as well, boils down this conflict of norms to a *lex generalis*-*lex specialis* relationship, in which the *lex specialis* prevails within its scope of application.<sup>1138</sup>

The ordinary Treaty revision procedure according to Article 48 TEU provides for the Convention, composed of representatives of the national parliaments, of the Heads of State or Government of the MS, the EP and the Commission.1139 This Convention shall examine the proposals for amendments and shall 'adopt by consensus a recommendation to a conference of representatives of the governments of the Member States'.<sup>1140</sup> This conference determines by common accord, after considering but not being legally bound by the recommendation of the Convention,1141 the amendments to be made to the Treaties.1142 Failure to reach consensus – and hence failure to adopt a recommendation – cannot prevent the MS from convening a conference of representatives of the governments of the MS, and hence from proceeding with the revision procedure.1143 Is this recommendation EU soft law at all, or is it – for the MS act within the realm of public international law when amending the EU Treaties – public international soft law? While the EU Treaties are Treaties of public

<sup>1136</sup> Article 19 para 5 of the Statute of the EIB.

<sup>1137</sup> Article 19 para 7 of the Statute of the EIB.

<sup>1138</sup> On the in principle equal rank of norms of EU primary law see H Hofmann, Normenhierarchien 84–86; for the problem of soft law acts contradicting each other: Dawson, Soft Law 7.

<sup>1139</sup> For this Convention more generally see Klinger, Konvent.

<sup>1140</sup> Article 48 para 3 TEU.

<sup>1141</sup> See Meng, Art. 48 EUV, para 15.

<sup>1142</sup> Article 48 para 4 TEU.

<sup>1143</sup> See Meng, Art. 48 EUV, paras 9 f.

international law,1144 it is widely argued that the EU established by these Treaties constitutes a separate legal order.<sup>1145</sup> Therefore, when EU primary law specifies the procedure for its amendment this is, together with all the acts provided for in this procedure, to be considered EU law. Also the use of the Treaty terminology ('recommendation'; see Article 288 TFEU) supports the assumption that it shall be an act of EU law.<sup>1146</sup> Whether also the originator of this act, the Convention, can be considered an EU body is not apparent.<sup>1147</sup> Its composition is only relatively vaguely regulated by the TEU and its concrete composition is mainly for the MS to decide.<sup>1148</sup> The decision to convene (ie to temporarily establish) a Convention is taken by the President of the European Council upon a decision by the European Council. The dominance of the MS, the only temporary existence of the Convention and its single purpose to facilitate (or not to facilitate) a Treaty amendment cannot alter the fact that this procedure shall move within an inter-governmental part *of EU law*. <sup>1149</sup> Therefore also the Convention rather is to be qualified as an EU-internal actor. The consequence of this qualification is that the Convention in its action is limited to the powers accorded to it by the Treaties which boil down to the competence to adopt a recommendation. The Convention could not, for example, adopt a binding agreement according to EU law, even if the representatives assembled in the Convention would be authorised accordingly by their respective MS. Hence a deviation from the procedure laid down in Article 48 TEU on the basis of public international law does not appear to be lawfully possible.1150 This does, of course, not affect its competence (and its task) to recommend the adoption of rules deviating from (current) EU law.

<sup>1144</sup> See eg Griller, Constitution 24.

<sup>1145</sup> See in particular case 26/62 *van Gend & Loos*, 12: 'the Community constitutes a new legal order of international law for the benefit of which the states have limited their sovereign rights'.

<sup>1146</sup> See Ohler, Art. 48 EUV, para 35.

<sup>1147</sup> In the affirmative: Ohler, Art. 48 EUV, para 36; see also C Möllers, Gewalt 274 f.

<sup>1148</sup> On the merely consultative purpose of the involvement of representatives of the EP and the Commission see Meng, Art. 48 EUV, para 10. The highly political issue of the concrete composition of the Convention is, according to *Ohler*, for the European Council as a whole (not: its President) to decide; see Ohler, Art. 48 EUV, para 35.

<sup>1149</sup> See de Witte, International Law 268–270; referring to the mitigation of the inter‐ governemental character of this procedure due to the inclusion of the EP: Ohler, Art. 48 EUV, para 29.

<sup>1150</sup> See case 43/75 *Defrenne*, paras 57 f.

### 3.5.3.2. EU-external actors

Apart from EU-internal actors, there are a number of actors not institution‐ ally belonging to the EU, which are nevertheless empowered by EU primary law to adopt soft law.

In the field of social policy, the Commission shall promote the consulta‐ tion of 'management and labour' (also referred to as 'social partners') at Union level. Therefore the Commission shall, before submitting proposals in this field, consult management and labour on 'the possible direction of Union action'.<sup>1151</sup> If, after that consultation, the Commission considers Union action advisable, it shall again consult management and labour, this time on the content of the envisaged proposal. They shall address an opin‐ ion or – 'where appropriate' – a recommendation to the Commission.<sup>1152</sup> Management and labour – on a whole – are the representatives of the interests of their respective clientele. They may encompass chambers of commerce, trade unions and other interest groups. Their degree of institu‐ tional formality can, due to this variety of actors, not be determined. When selecting out of those parts of management and labour ('European social partners'), the Commission shall ensure that they are 'truly representati‐ ve'.1153 While social partners are not EU bodies, but only bodies referred to in EU law and – as 'European social partners' – selected by an EU institu‐ tion, the Commission, they (ie the concrete group of actors in the concrete consultation procedure as composed according to the concrete selection by the Commission) are vested with the power to adopt opinions and recommendations, hence EU legal acts according to Article 288 TFEU.<sup>1154</sup>

<sup>1151</sup> Article 154 para 2 TFEU.

<sup>1152</sup> Article 154 para 3 TFEU. For the possibility of contractual relations between the EU and management and labour see Article 155 TFEU; see also Korkea-aho, Soft Law 284.

<sup>1153</sup> Case T-135/96 *UEAPME*, para 89; for so-called 'representativeness studies' see Commission, A Practical Guide for European Social Partner Organisations and their National Affiliates (Vademecum, July 2017) 7 f; see also the 'List of consulted organisations' under Article 154 TFEU' as published (and updated) by the Com‐ mission at <http://ec.europa.eu/social/main.jsp?catId=329&langId=en> accessed 28 March 2023.

<sup>1154</sup> It was certainly not necessary to vest the 'European social partners' with such a competence – after all, they could also express their views in other ways. The terminology 'opinion' and 'recommendation' in a Treaty provision speaks in favour of the assumption that it is the legal acts mentioned in Article 288 TFEU which are referred to.

In practice, the social partners may also enter into negotiations with the Commission. Where they do produce a written document, it appears to be individual or joint opinions most of the time.<sup>1155</sup>

Article 302 TFEU refers to the composition of the ESC. The Council shall adopt the list of members drawn up in accordance with the proposals made by each MS. Prior to that, the Council shall consult the Commission, and it may also obtain the opinion of 'European bodies which are represen‐ tative of the various economic and social actors and of civil society to which the Union's activities are of concern'. In practice, the Council very rarely makes use of this possibility.<sup>1156</sup> The opinion referred to in this consultation procedure is not a legal act according to Article 288 TFEU. Unlike in the case of the Commission's consultation of the 'European social partners', the 'European bodies' referred to here can be chosen *ad hoc* by the Council which is not in any way restricted (eg by a list of bodies set up in advance) and hence are even less homogenous than the former. This complete lack of institutionalisation and, even more so, the fact that the terminology used in other language versions1157 does not reflect Article 288 TFEU, speak against EU soft law in this case.

Protocol No 1 on the role of national parliaments in the EU provides for the possibility of national parliaments to submit a reasoned opinion on whether a draft legislative act complies with the principle of subsidiari‐ ty.1158 This competence is specified in Article 6 of Protocol No 2 on the application of the principles of subsidiarity and proportionality, according to which a national parliament or a chamber of it may, within eight weeks from the date of transmission of a draft legislative act in the official languag‐ es of the Union, send to the Presidents of the European Parliament, the Council and the Commission a reasoned opinion on the act's non-compli‐ ance with the principle of subsidiarity.1159 Where the draft legislative acts

<sup>1155</sup> Commission, 'Consulting European social partners: Understanding how it works' (2011) 8 f <https://op.europa.eu/en/publication-detail/-/publication/5208f68c-3db 1-405e-9b4a-51316aeacc03> accessed 28 March 2023.

<sup>1156</sup> See Jaeckel, Art. 302 AEUV, para 17, with further references.

<sup>1157</sup> The distinction made becomes clear when having a look eg at the German, French or Spanish version of Article 302 TFEU.

<sup>1158</sup> Article 3 of Protocol No 1 on the role of national parliaments in the European Union.

<sup>1159</sup> Article 3 para 1 of Protocol No 1 on the role of national parliaments in the European Union; Article 6 of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

exceptionally originate from another entity, eg the ECB, the President of the Council shall forward the reasoned opinion to it. The EP, the Council, the Commission and possibly other originators of the draft act shall 'take account' of the reasoned opinions.<sup>1160</sup>

Each national parliament shall have two votes, in case of a bicameral parliament each chamber shall have one vote.<sup>1161</sup> Where the reasoned opin‐ ions stating non-compliance of a certain draft act represent at least one third<sup>1162</sup> of the votes allocated to the national parliaments, the Commission (or exceptionally another originator of the draft) has to review the act and to decide, thereby giving the reasons, to maintain, amend or withdraw the draft.<sup>1163</sup> Where, for a draft act initiating the ordinary legislative procedure, the reasoned opinions launched represent a simple majority of the allocated votes, the Commission has to review the act and may decide to maintain, amend or withdraw the draft. In the former case – maintenance of the act – the Commission has to justify in a reasoned opinion why it considers the act to be compatible with the principle of subsidiarity. If, by a majority of 55 percent of the members of the Council or a majority of the votes cast in the European Parliament, the legislator opines that the draft is not in compliance with the subsidiarity principle, it shall not be given further consideration.1164 Hence in this case (ordinary legislative procedure) the effects of the opinion depend on whether the support it has got from among the group of its (potential) originators, the national parliaments that is, reaches the critical threshold of one third or of a simple majority of the allocated votes.

In the context of financing of undertakings or other public or private entities, Article 19 of the Statute of the EIB provides that they may apply for financing directly to the EIB. Applications may also be made through the Commission or through the MS on whose territory the investment will be

<sup>1160</sup> Article 7 para 1 subpara 1 of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

<sup>1161</sup> Article 7 para 1 subpara 2 of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

<sup>1162</sup> Where the draft is based on Article 76 TFEU it shall be one quarter; Article 7 para 2 subpara 1 of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

<sup>1163</sup> Article 7 para 2 subpara 2 of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

<sup>1164</sup> Article 7 para 3 subpara 2 lit b of Protocol No 2 on the application of the principles of subsidiarity and proportionality.

carried out. Where the application is not made through the MS concerned, it shall be asked to submit its opinion. Where the MS does not do so within two months, the EIB may assume that there is no objection to the respective investment.<sup>1165</sup> In this case, the Statute determines – similar to the case of the national parliaments' concerns about compatibility of an EU legislative act with the principle of subsidiarity, as discussed above – the form the MS' (and the Commission's) views shall take. In this procedure it is qualified as an opinion according to Article 288 TFEU.

As we have seen above, not every 'opinion' mentioned in the Treaties actually is an opinion within the meaning of Article 288 TFEU. Here is another example: According to Article 25 of the Statute of the CJEU, the Court may 'entrust any individual, body, authority, committee or other organisation it chooses with the task of giving an expert opinion'. The expert opinions referred to here are not opinions according to Article 288 TFEU. They are intended to clarify facts, not to set (soft) legal norms.<sup>1166</sup> Therefore they can be proven wrong,1167 whereas (soft) legal norms can only be proven illegal. The Court may consider them as evidence of certain statements in the course of a judicial procedure.1168 If and to the extent that scientific opinions of an EU body (eg the European Food Safety Authority EFSA) express a normative content (eg the sentence: 'The food ingredient shall not be certified for marketing') it constitutes – to that extent – soft law. Where it merely lists properties and risks of the ingredient at issue, as a piece of evidence it needs to be duly considered. Where the opinion *must* be rendered in the course of the underlying procedure, its adoption is to be considered as a procedural requirement, just as the adoption of a soft law act may be a procedural requirement. In that sense, the effects of soft law and expert opinions may in certain cases be very similar, both procedurally

<sup>1165</sup> The same applies for the Commission, accordingly; see Article 19 para 2 of the Statute of the EIB.

<sup>1166</sup> For the role of expert opinions in the context of individual-concrete administrative decisions in a democratic *Rechtsstaat* more generally see Nußberger, Sachverstän‐ digenwissen.

<sup>1167</sup> In places, the legislator even addresses the case of expert opinions which differ from each other; see eg Article 59 paras 3 f of Regulation 726/2004.

<sup>1168</sup> On the legal status of expert opinions – and the reasons for consideration – see Weismann, Agencies 71–74; see also Mills, Biotechnology 331f, taking the example of the Codex Alimentarius Commission and its output; on the *quasi*-legal authori‐ ty of scientific output in legal history see Jansen, Methoden 45 f.

(requirement of adoption) and substantially (duty to consider<sup>1169</sup> ; steering effect).<sup>1170</sup> This does not alter the fact that they are, qualitatively speaking, of a different kind (for the legally non-binding acts other than soft law see also 2.4. above).

# 3.6. Competences to adopt EU soft law other than recommendations and opinions

In the above sub-chapters emphasis was laid on recommendations and opinions as the two legally non-binding acts mentioned in the catalogue of 'legal acts of the Union' contained in Article 288 TFEU. In addition to that, however, there is a number of further soft law acts mentioned in the Treaties,1171 and again further soft law acts not mentioned in the Treaties but used in EU administrative practice.1172 In the following, these two issues – other soft law acts than those referred to in Article 288 TFEU 1) in the

<sup>1169</sup> For the example of EMA committees' opinions and the 'detailed explanation of the reasons for the differences' the Commission has to provide in case its draft de‐ cision deviates therefrom see Article 10 para 1 of Regulation 726/2004. In practice, this has resulted in a mere 'rubber-stamping' on the part of the Commission; see Orator, Möglichkeiten 145; for the role of committee expertise in the then Europe‐ an Medicines Evaluation Agency more generally see Gehring/Krapohl, Regulation.

<sup>1170</sup> This similarity is one aspect of the concept of 'öffentliche Gewalt' [public authori‐ ty] as coined specifically by *Goldmann* and *von Bogdandy*; see in particular von Bogdandy/Goldmann, Ausübung; Goldmann, Gewalt.

<sup>1171</sup> That soft law may not only be referred to, but also be contained in primary law is exemplified by the Explanations relating to the Charter of Fundamental Rights, OJ 2007/C 303/02, which – according to their self-description – 'do not as such have the status of law, [but which] are a valuable tool of interpretation intended to clarify the provisions of the Charter' and, pursuant to Article 52 para 7 CFR, 'shall be given due regard by the courts of the Union and of the Member States'; for the soft law character of the CFR itself until its entry into force as primary law on 1 December 2009 see Opinion of AG *Colomer* in case C-553/07 *College van burgemeester en wethouders van Rotterdam*, para 22 (fn 23); see also Knauff, Re‐ gelungsverbund 315–318, with many further references; Müller-Graff, Einführung 156; Ştefan, Soft Law 19 f, with further references; for another soft law act relating to the EU's fundamental rights see Commission Recommendation 2017/761 on the European Pillar of Social Rights, the substance of which is reflected upon in the Interinstitutional Proclamation on the European Pillar of Social Rights of the European Parliament, the Council, and the Commission, 2017/C 428/09.

<sup>1172</sup> *Láncos* distinguishes recommendations and opinions – which she refers to as 'for‐ mal [soft law] measures' – from other soft law measures such as communications or white papers ('non-formal [soft law] measures'); Láncos, Facets 16 f.

Treaties and 2) only in practice – shall be addressed exemplarily, that is to say with no claim for completeness.

Apart from recommendations and opinions, the Treaties mention further acts which are legally non-binding and also otherwise fulfil (or rather: *may* – in their concrete form in a specific case – fulfil) the criteria of EU soft law (see in particular II.1.3.4. above). These are eg guidelines,<sup>1173</sup> warnings<sup>1174</sup> or conclusions<sup>1175</sup>. While it is difficult to clearly define recommendations and opinions (see 3.1.1. above), it is even more difficult to assign a specific meaning to those acts which would allow us to clearly separate them from each other on the one hand, and from recommendations and opinions on the other hand. Rather, in terms of shape and general effects they appear to be very similar to each other. Proposals,<sup>1176</sup> which could be added to the above (non-exhaustive) enumeration, can be defined more closely. Already from their name it can be concluded that they aim at initiating a decisionmaking process in a wider sense. Applying a systematic interpretation, it becomes clear that it is the Commission,1177 exceptionally together with the High Representative,1178 and, above all on the basis of the TEU, the High Representative on its own, which are entitled to make proposals.1179 Article 293 TFEU provides for specific effects of Commission proposals: Where the Council acts on a Commission proposal, it may amend this proposal – exceptions apart – only by acting unanimously (para 1)1180; as long as the Council has not acted, the Commission may amend its proposal at any time during the decision-making procedure in a wider sense (para 2). The similarity between proposals and recommendations – both acts suggest a


1179 Generally: Article 30 para 1 TEU (according to this provision, also MS are entitled to submit proposals to the Council); Articles 27 para 3, 33 TEU; exceptionally also in the TFEU: eg Article 218 para 9; on the High Representative's role in this context see also Marquardt/Gaedtke, Art. 27 EUV, para 3.

1180 With regard to proposals issued by the High Representative, it is to be noted that in the field of CFSP – in which the High Representative is acting predominantly – most Council decisions shall be adopted unanimously anyway (Article 31 para 1 TEU).

<sup>1173</sup> Eg Article 25 lit a, Article 50 para 2 TEU; Articles 26, 156, 171 para 1 TFEU (in conjunction with Article 172 TFEU).

<sup>1174</sup> Eg Articles 121 para 4, 168 para 1 TFEU.

<sup>1175</sup> Eg Articles 135, 148 para 1 TFEU.

<sup>1176</sup> Eg Articles 95 para 3, 103 para 1 TFEU: see also case C-301/03 *Italy v Commission*, paras 21f with regard to the character of a proposal.

certain action – is underpinned by the fact that, eg according to Article 129 TFEU, the legislator may act on a proposal from the Commission or (alternatively) on a recommendation from the ECB.<sup>1181</sup> According to Article 281 TFEU, to take another example, the European Parliament and the Council shall act 'at the *request* of the [CJEU]' (emphasis added) or on a proposal from the Commission<sup>1182</sup> which suggests that also the request may be<sup>1183</sup> very similar to a proposal and hence to a recommendation.<sup>1184</sup>

<sup>1181</sup> Article 129 paras 3 f TFEU. When the Treaty provides for legislative action either on a *recommendation* from the Commission or on a recommendation from the ECB, it is clear that in this case it intends to exclude the effects of a Commission *proposal* according to Article 293 TFEU (eg Article 219 para 1 TFEU). When considering the important role the Council shall play in the context of the creation of exchange-rate systems for the euro in relation to the currencies of third states, the reduction of Commission power seems to be plausible. What is more, with regard to a matter of monetary policy, it would seem inappropriate to assign to the Commission a more important role than to the ECB; see also Häde, Art. 219 AEUV, para 1.

<sup>1182</sup> Similarly: Article 308 para 3 TFEU with regard to the EIB.

<sup>1183</sup> See also Article 153 para 3 TFEU (*argumentum* 'may'). With other requests, how‐ ever, it is to be assumed that they are legally binding and hence do not qualify as soft law; see eg Article 48 para 2 TFEU: Where a member of the Council requests the matter to be referred to the European Council, this seems to grant a right to this member – and hence the request must be considered legally binding. This is confirmed by the wording in which the consequences are laid down: 'the ordinary legislative procedure *shall* be suspended' (emphasis added) and 'the European Council *shall*' (emphasis added) choose between two alternatives of reaction. Also the competence of the European Council to request the Commission, according to lit a *leg cit*, to submit a new proposal is legally binding; arguably confirming the legal bindingness: Langer, Art. 48 AEUV, para 87. There are a number of other provisions in which a request for referral is provided for; eg Articles 82 para 3, 83 para 3, 86 para 1 subpara 2 TFEU. Also these requests are legally binding. As a general rule, it can be said that requests are legally binding where procedural action (in particular a referral) or action which is content-wise not predetermined (submission of a report, delivery of an opinion, undertaking of studies etc) is asked for. Where a content-wise predetermined action is asked for by means of a request, the request is rather legally non-binding: eg an increase of the number of AG by the Council at the request of the Court (Article 252 TFEU; *argumentum* 'may') or the dismissal of the European Ombudsman if he/she no longer fulfils the conditions required for the performance of his/her duties or if he/she is guilty of serious misconduct by the CJEU at the request of the EP (Article 228 para 2 subpara 2 TFEU; *argumentum* 'may'). Having a look at other language versions of the Treaties, the more differentiated terminology (the German version, for example, distinguishes between a by tendency legally binding *Antrag* and a by tendency legally non-binding *Ersuchen*) may be indicative; but it may as well be misleading: eg in the case of Article 319 para 3 TFEU in which the *Ersuchen* of the

Also Article 296 para 2 TFEU generally stresses the similar character of proposals, initiatives, recommendations and requests,<sup>1185</sup> and so does – with regard to the latter three acts – Article 289 para 4 TFEU.

Apart from the – considering Article 288 TFEU a canon of EU legal acts – extra-canonical soft law acts mentioned (elsewhere) in the Treaties, eg the Council guidelines and conditions referred to in Article 26 para 3 TFEU,<sup>1186</sup> there are a number of (potential<sup>1187</sup>) EU soft law acts in prac‐ tice which are not mentioned in the Treaties, such as communications,<sup>1188</sup> standards or codes of conduct<sup>1189</sup> . <sup>1190</sup> These acts are adopted by all kinds of EU bodies in all kinds of policy fields. That they are not (expressly)

EP or the Council is certainly legally binding; see Niedobitek, Art. 319 AEUV, para 19. In case of a request based on secondary law, its qualification also depends on its form: see joined cases C-293/13P and C-294/13P *Del Monte*, para 183.

<sup>1184</sup> For a recent request of the CJEU on the reform of the preliminary reference proce‐ dure see <https://curia.europa.eu/jcms/upload/docs/application/pdf/2022-12/de mande\_transfert\_ddp\_tribunal\_en.pdf> accessed 28 March 2023. A request may also be used in order to evoke a recommendation or a proposal; see Article 135 TFEU. That this provision, *inter alia*, refers to Article 140 para 1 TFEU, in which Commission reports are mentioned, but not recommendations or proposals, does not mean that reports are to be equated with recommendations/proposals, but rather that also these reports may be requested by the Council or a MS on the basis of Article 135 TFEU; see Häde, Art. 135 AEUV, para 5, with a further reference.

<sup>1185</sup> See Calliess, Art. 296 AEUV, para 34. Opinions – in spite of their being listed in Article 296 para 2 TFEU, as well – have a slightly different character in that here the focus is laid on uttering one's point of view, and less on instigating certain action of others. These two purposes, however, may as well overlap (see 3.1.1. above).

<sup>1186</sup> Critically as regards the exclusion of the EP and stressing the practical irrelevance of this provision: Korte, Art. 26 AEUV, para 42; sceptically as regards the intended self-binding effect of these acts: M Schröder, Art. 26 AEUV, paras 37 f.

<sup>1187</sup> As was noted above (see II.2.1.2.), it may be that such acts lack any normative content and are limited, for example, to a mere summary of the relevant legal provisions. In this case, of course, the act does not constitute soft law and none of the scrutiny applied here is required. It may happen, though, that the provision of information complemented by a mere wish is understood as an (implicit) com‐ mand to act; see the case T-193/04 *Tillack*, para 79. For the notion of 'regulation by information' see also Hofmann/Rowe/Türk, Administrative Law 551.

<sup>1188</sup> Exceptionally (and inadequately so), communications may contain legally binding provisions; see case C-135/93 *Spain v Commission*, paras 3, 10, 18; for the excep‐ tionality of these circumstances see case C-292/95 *Spain v Commission*, paras 28 ff; see also Aldestam, Soft Law 22 ff.

<sup>1189</sup> Eg the Code of Conduct for business taxation (1997), adopted by the Council and the representatives of the governments of the MS meeting within it; see Gribnau, Code 67; for examples of codes of conduct in public international law see Bothe, Norms 81f.

mentioned in the Treaties does not *per se* mean that their adoption is unlawful. A comprehensive interpretation – or a (justifiable) development of law – of competence clauses may disclose a primary law power to adopt a certain soft law act (for soft law powers laid down in secondary law see 3.7. below). When examining whether or not a competence to adopt an extra-canonical soft law act not expressly mentioned in the Treaties exists, it is advisable to 'convert' – conceptionally at least – the respective soft law act into a soft law act the body at issue is expressly empowered to enact in primary law (if any). This may be in particular a recommendation or an opinion, and – in the case of the Commission – a proposal, respectively. Where this conversion is possible and an according competence (eg to adopt a recommendation) exists, the power to adopt the act at issue can be affirmed. Where this conversion is not possible, the interpretation of the relevant Treaty provisions may still reveal that the power to adopt the act at issue is actually conferred, but the interpretative 'exercise' is certainly more demanding then. The malleability of the terms 'recommendation' and 'opinion' makes it difficult to think of a soft law act which does not allow to be 'materially' assigned to ('converted into') either category, though.1191 That these acts bear varying names, and may, if at all, be published in sections of the OJ different from those of 'regular' recommendations or opinions (or proposals) does not harm in the given context (see 3.1.2. above). It is their normative substance which matters. In these considerations again the *numerus clausus* concept shimmers through.

<sup>1190</sup> As regards 'guidelines', it is to be noted that also the Treaties provide for their adoption, eg in Article 5 TFEU or – as was mentioned before – Article 26 para 3 TFEU. The guidelines of the Governing Council of the ECB are even consid‐ ered to be legally binding for the ECB's Executive Board; see Article 12.1 of the Statute of the ESCB and of the ECB; see also C-355/10 *European Parliament v Council*, para 82, with regard to the binding effects of guidelines contained in Regulation 562/2006 (a meanwhile outdated version of the so-called 'Schengen Borders Code'). It follows that the term guidelines is used, in the context of EU law, both within and outside the Treaties, to describe both legally binding and legally non-binding rules.

<sup>1191</sup> See Knauff, Regelungsverbund 372, emphasising the fact that new soft law acts in general conceptually resemble soft law acts already in being.

3.7. Special competence clauses in EU secondary law and in public international law

### 3.7.1. Introduction

The focus of this chapter is the competences to adopt EU recommendations and opinions. In the preceding sub-chapters we took a look at the Treaty competences to adopt recommendations and opinions as well as Treaty competences to adopt extra-canonical soft law in practice. In order to sup‐ plement this view, we shall now address examples of cases where the Treaty competences to adopt recommendations and opinions are 'concretised' in secondary law on the one hand, and where soft law powers are granted to EU bodies in public international law, on the other hand.

The provision of soft law powers in EU secondary law must have a sufficient legal basis in primary law. Secondary law often sets out soft law powers of EU institutions, bodies, offices and agencies in the framework of certain procedures which are not explicitly mentioned in the Treaties. With the methodological toolkit referred to above, it is then to be examined whether or not the Treaties allow for the provision of these competences. Listing all relevant secondary legislation, let alone depict it, would not be a viable approach to address this phenomenon. Rather, two examples – namely Regulation 182/2011 (the so-called Comitology Regulation) and Council Regulation 168/2007 (the founding regulation of the Fundamental Rights Agency [FRA]) – shall be analysed with a view to the soft law powers they confer on comitology committees and the FRA, respectively. Thereby two important groups of actors not empowered in the Treaties – comitology committees and European agencies – shall be exemplarily ad‐ dressed. In Part IV of this work, further examples of secondary law, vesting in particular the Commission and European agencies with soft law powers, shall be presented and analysed, and their legality shall be examined in particular under V.3. below.

Also public international law may serve as a source of EU soft law powers. The respective acts of public international law which are in places concluded by a number of MS and which often 'substitute'<sup>1192</sup> – for different reasons – EU law proper must not contradict EU primary law. Therefore also in this case the legality of soft law must be examined with a view to the Treaties. In this context, the example of the Treaty on Stability, Coordina‐

<sup>1192</sup> For the notion of '*Unionsersatzrecht*' see 2.2.4.2.2. above.

tion and Governance in the Economic and Monetary Union (TSCG) shall be presented and analysed with a view to its making use of EU institutions and vesting them with, or rather: concretising, their respective EU soft law powers.

# 3.7.2. Special competence clauses in EU secondary law

3.7.2.1. Regulation 182/2011

The Comitology Regulation vests committees composed of representatives of the MS' bureaucracies with the competence to adopt an opinion on a (draft) implementing act proposed by the Commission.<sup>1193</sup> Depending on the procedure to be applied, the opinions exert a different effect.<sup>1194</sup> In the 'official view'1195 the Regulation provides for two procedures: the advisory and the examination procedure.1196 Where the advisory procedure applies, the Commission shall take the 'utmost account' of the delivered opinion.<sup>1197</sup> Where the examination procedure applies, the opinion may have different effects. In case it is positive, the Commission may adopt the act. The same is true – with some exceptions1198 – where no opinion is delivered. In case it is negative, the Commission shall not adopt the act.1199 In this latter case the Commission may either submit to the committee an amended version of the draft or refer to the appeal committee.1200 The appeal committee shall replace the appealed opinion by its own opinion. If it is positive or if no opinion is delivered, the Commission may adopt the act. If it is

1195 Craig, Administrative Law 134.


<sup>1193</sup> While it appears that MS can freely choose the persons who ought to represent them (Article 3 para 2 and Article 10 para 1 lit c of Regulation 182/2011), the usual practice seems to have been that MS send officials from their respective ministries rather than independent experts; see Egeberg/Trondal, Agencies 871, with a further reference.

<sup>1194</sup> On the Commission's endeavour to convince the committees content-wise see F Schmidt, Art. 291 AEUV, para 39.

<sup>1196</sup> For the respective scope of these procedures see Article 2 of Regulation 182/2011; for the procedural variations see Craig, Administrative Law 135 f.

<sup>1200</sup> See Article 5 para 3 of Regulation 182/2011; for the composition and practice of appeal committees see Volpato, Delegation, in particular 179-181.

negative, the Commission shall not adopt the act.1201 That these opinions constitute EU legal acts can be deduced from different characteristics of the comitology regime: 1) The committees' composition is laid down in the Regulation, even if not in much detail.<sup>1202</sup> Although mainly composed of MS' representatives, the committees are created and institutionally belong to the EU,<sup>1203</sup> more particularly they are attached to (but not actually part of ) the Commission. This is exemplified by the fact that they are chaired by a Commission representative.<sup>1204</sup> 2) The opinions provided by the com‐ mittees are not exclusively, not even predominantly factual ('objective') expert opinions, but they also have a normative ('political') thrust.<sup>1205</sup> The national bureaucrats in the committees dispose of relevant knowledge and experience, but they are embedded in a necessarily political (national) administration.<sup>1206</sup> 3) The legal effects of the committees' opinions are laid down in detail in EU law.

While the qualification of the committee opinion in the advisory proce‐ dure as EU soft law is clear,1207 this is dubitable in case of the examination procedure. Since in the examination procedure negative opinions – leaving

1205 See case T-13/99 *Pfizer*, paras 283 and 285.

<sup>1201</sup> See Article 6 para 3 of Regulation 182/2011.

<sup>1202</sup> See Article 3 para 2 of Regulation 182/2011.

<sup>1203</sup> See Türk, Comitology 347 f; see also W Weiß, Verwaltungsverbund 52 f: 'europäi‐ sche Exekutivstruktur' [European executive structure]. The wording 'control by Member States' in Article 291 para 3 TFEU is not to be understood institutionally, but substantially. In substance, it is the MS controlling the Commission under the comitology regime; for the alleged political dominance of EU institutions within comitology see Craig, Administrative Law 122–126, with further references; for the 'ultimately autonomous decision-making powers of the Member States': Everson/Joerges, Europeanisation 526.

<sup>1204</sup> For the qualification of the committees see also case T-188/97 *Rothmans*, paras 56 ff, in which the then Court of First Instance concluded 'that, [at least] for the purposes of the Community rules on access to documents, "comitology" commit‐ tees come under the Commission itself ' (para 62).

<sup>1206</sup> On the fact that the committees do not have to provide the reasons for their opinions see Weismann, Agencies 73.

<sup>1207</sup> See the qualification of the opinion of a comitology committee in an advisory procedure by AG *Alber* as binding in a 'relaxed manner' by which he means that the Commission 'could not simply disregard such an opinion but was obliged to provide reasons for any divergences from it': Opinion of AG *Alber* in case C-248/99P *France v Monsanto*, paras 133 f; see F Schmidt, Art. 291 AEUV, para 29, according to whom the opinion has the character of a recommendation – a statement which is slightly confusing, but which stresses the soft law quality of the committee opinion.

the exceptional cases pursuant to Article 7 apart – prohibit the adoption of the respective act by the Commission, and positive opinions arguably oblige the Commission to adopt the act,<sup>1208</sup> the question arises whether opinions adopted in the examination procedure can be qualified as opin‐ ions according to Article 288 TFEU, which *per definitionem* 'shall have no binding force'. Unlike with the opinions of the Commission and the EIB's Management Committee which – if negative – in combination have a prohibitive effect according to Article 19 para 5 on the EIB-Statute (see 3.5.3.1. above), here the obligatory (prohibitive/requesting) effect is laid down in secondary law only. Non-compliance with Article 288 TFEU can therefore not be explained as a *lex specialis*-*lex generalis* relationship. The committees' opinions cannot be interpreted as an act of self-obligation of the Commission. As was mentioned above, the committees are attached to, but are not themselves (part of ) 'the Commission'.1209 What is more, the obligatory effect of opinions (in the examination procedure) is provided for by a Regulation adopted by the EP and the Council. Rather than as an act of self-obligation the kind of opinion addressed here is – due to its legal bindingness – to be qualified as an opinion *sui generis*. Conceptually, it lies somewhere between an opinion proper and a decision, because it is binding but obviously shall be excluded from the jurisdiction of the CJEU (*argumentum* 'opinion'). In case of Regulation 2023/1114 the situation is different: The legislator empowers ia the ECB to adopt binding 'opinions' (see eg Article 24 para 2), but suggests in Recital 46 that the Court may review them. This is reflected in the special legal remedy against negative opinions provided in Article 6 of Regulation 182/2011 (referral to appeal committee). Concerns with regard to Article 288 TFEU, which does not provide for such an act,1210 and with regard to a distortion of the EU's institutional balance1211 – a committee addresses a legally binding opinion *sui generis* to the Commission – can be countered by referring to Article

<sup>1208</sup> *Argumentum* 'shall' in Article 5 para 2 of Regulation 182/2011.

<sup>1209</sup> In view of the long-lasting political battle the Commission has fought against comitology, alleging an act of self-obligation by the Commission carries a certain absurdity; see Craig, Administrative Law 127, with further references.

<sup>1210</sup> With regard to the 'deficien[cy]' of Article 249 TEC, the predecessor of Article 288 TFEU, see Senden, Soft Law 53.

<sup>1211</sup> See in particular the case 25/70 *Einfuhr- und Vorratsstelle*; see Craig, Administra‐ tive Law 116 f, with references to the case law and, with regard to the Lisbon regime, 136 f; see Ponzano, Acts 140 f, with regard to improvements brought about by the Treaty of Lisbon.

291 para 3 TFEU which expressly stipulates that the EP and the Council 'shall lay down in advance the rules and general principles concerning mechanisms for control by Member States of the Commission's exercise of implementing powers'.<sup>1212</sup> From a purely literal perspective, this wording appears to be wide enough so as to encompass legally binding forms of control. The Court in a number of cases has confirmed the lawfulness of comitology and hence also of the binding effects of committee opinions in certain procedures.<sup>1213</sup>

In view of the fact that also prior to the Treaty of Lisbon comitology was regulated by means of secondary law, and in view of the fact that also in these respective legal acts comitology committees were vested with the power to adopt such opinions, it can be assumed that the Masters of the Treaties by adopting Article 291 TFEU intended to address not only the non-obliging opinions,1214 but – due the apparent limitation of the Council's role in favour of 'control by Member States'1215 – also the *sui generis* opinions. While the Treaty of Lisbon has created a new system of executive legal acts to be adopted by the Commission,1216 with regard to implementing acts according to Article 291 TFEU comitology was intended to live on.<sup>1217</sup>

<sup>1212</sup> See also Kröll, Artikel 290 und 291 AEUV 210 f.

<sup>1213</sup> Note in particular the Court's decision in case 5/77 *Tedeschi*, para 55, with regard to the Commission's power 'to issue, in accordance [with the respective comitolo‐ gy procedure], any other measure which it considers appropriate' in case its (first) proposal evoked a negative committee opinion; for the 'astonishment in the legal literature' about this statement of the Court (which acknowledges but does not dwell on the crucial fact that also a new proposal by the Commission could not be adopted in case of a negative committee opinion): Bergström, Comitology 148 f.

<sup>1214</sup> For the effects of (negative) committee opinions under earlier comitology regimes see eg Mensching, Komitologie-Beschluss.

<sup>1215</sup> Compare the in this respect different wording of Article 202 (3rd indent) TEC on the one hand, and Article 291 para 3 TFEU on the other hand; see also Ilgner, Durchführung 242 f; critically: Craig, Administrative Law 136 f.

<sup>1216</sup> See eg Ilgner, Durchführung 197 ff; see also Working Group IX on Simplification, Report CONV 424/02 (29 November 2002), which – as it turned out: wrongly so – announced that '*any* change would not come under the Treaty directly but under secondary legislation' (emphasis added).

<sup>1217</sup> This is also reflected in the application of the old Comitology Decision (Council Decision 1999/468/EC, as amended in 2006) even after the Treaty of Lisbon entered into force (until the adoption of Council Regulation 182/2011); see Ruffert, Art. 291 AEUV, para 12, with further references.

### 3.7.2.2. Council Regulation 168/2007

According to Article 4 para 1 lit d of the founding regulation of the FRA, it may 'formulate and publish conclusions and opinions on specific thematic topics, for the Union institutions and the [MS] when implementing Union law, either on its own initiative or at the request of the [EP], the Council or the Commission'. In Recital 13 of the Regulation it is added that this shall take place 'without interference with the legislative and judicial proce‐ dures established in the Treaty'. FRA opinions on Commission proposals or positions of the legislator in the course of legislative procedures shall, however, only be adopted upon a request by the respective institution.<sup>1218</sup> The opinions of the FRA clearly are legally non-binding. It is to be noted that in the Treaties no explicit competence for the FRA, or other European agencies for that matter (with exception of the EDA1219), to adopt opinions is laid down.1220 However, primary law – implicitly1221 – acknowledges that there are or at least may be agencies with a power to adopt legal acts.<sup>1222</sup>

Regulation 168/2007 was based on the so-called flexibility clause, now Article 352 TFEU. Hence in this case the competence to adopt opinions of a European agency, the FRA, hardly seems problematic. It was acknowl‐ edged that the Treaties (then the TEU and the TEC) did not provide the respective competences, and – for this reason – the Regulation was based on Article 308 TEC (now: Article 352 TFEU).1223 In that sense, Regulation 168/2007 is, in the opinion of the legislator, an 'appropriate measure' within the meaning of Article 352 TFEU.<sup>1224</sup>

Also apart from the FRA, European agencies are vested with soft law powers (some of which shall be addressed in Part IV of this work) – on

<sup>1218</sup> Article 4 para 2 of Council Regulation 168/2007; see von Bogdandy/von Bernstorff, Agentur 155 f.

<sup>1219</sup> Article 3 of Protocol No 10 on permanent structured cooperation established by Article 42 of the Treaty on European Union.

<sup>1220</sup> For explicit competences other than to adopt opinions see Article 88 TFEU (Euro‐ pol, Eurojust).

<sup>1221</sup> For the lack of an explicit provision see case T-510/17 *Del Valle Ruíz*, para 207.

<sup>1222</sup> See in particular Articles 263 para 1 and 267 para 1 lit b TFEU; also referred to by the CJEU in case C-270/12 *United Kingdom v European Parliament/Council*, para 80.

<sup>1223</sup> For the merits of Article 352 TFEU as a legal basis of European agencies see Kirste, System 273 f.

<sup>1224</sup> For the role of the predecessors of Article 352 TFEU as a legal basis for EU soft law see Senden, Soft Law 178 and 184.

various legal bases.1225 Since most European agencies are not empowered by primary law directly, but only through secondary law, the question of competence is, first, whether the EU actually disposes of the power its legislator intends to delegate to the agency at issue, and – in the affirmative – second, whether the legislator is allowed to delegate it according to the so-called *Meroni* doctrine.<sup>1226</sup>

While these questions are to be answered case by case, on a general scale the following can be remarked: Generally speaking, it is to be noted that the Court appears to be permissive as regards the competence of the legislator to vest a new EU body with the power to adopt an opinion. With regard to what is now Article 114 TFEU, it not only allowed for the creation of a new agency, but also for vesting it with the competence to adopt opinions,<sup>1227</sup> more recently even for creating an agency with (hard) regulatory powers.1228 Here the *Meroni* case law on the delegation of powers also to EU administrative bodies comes into play.1229 In my view, this case law should not only be applicable to the delegation of hard but also of soft law powers.1230 This does not seem to be what a strict reading of *Meroni* suggests.1231 The delegation of 'a discretionary power, implying a wide mar‐

<sup>1225</sup> For the legal bases of the founding acts of European agencies (which regularly also are the legal bases of the empowerment of these agencies) see table in: Griller/Orator, Everything 32 ff.

<sup>1226</sup> For the case of a delegation of powers the legislator may not exercise itself see 3.3.2.2. above.

<sup>1227</sup> See case C-217/04 *United Kingdom v European Parliament/Council*, para 64. The Court is unclear about the methodological foundation of its finding; with regard to the legal basis for the establishment of the agency at issue in this case, the ENISA, see Ohler, Gemeinschaftsagentur 374. See also case C-380/03 *Germany v Council*, para 42; case C‑358/14 *Poland v European Parliament/Council*, para 37, both referring to the legislator's discretion under what is now Article 114 TFEU.

<sup>1228</sup> See case C-270/12 *United Kingdom v European Parliament/Council*, paras 88 ff; joined cases C-584/20P and C-621/20P *Commission v Landesbank Baden-Würt‐ temberg and SRB*, paras 105 f; for the compliance of the SRB's powers with *Meroni* see also case T-481/17 *SFL*, paras 126–132.

<sup>1229</sup> Exemplarily for the large amount of literature published on the *Meroni* doctrine: Pawlik, Meroni-Doktrin, and the references made in this book; Simoncini, Regu‐ lation 14 ff.

<sup>1230</sup> Appraising *pro* and *contra*: van Rijsbergen, Enforceability 117; see also Busuioc, Rule-Making 123 and *passim*; Ştefan/Petri, Review 531f, 549 and *passim*.

<sup>1231</sup> See case C-270/12 *United Kingdom v European Parliament/Council*, paras 63–68 (with reference to the *Romano* case). However, also these passages do not outright exclude the application of the *Meroni* criteria to (the delegation of ) soft law powers; more restrictive is case T-755/17 *Germany v ECHA*, para 139, in which the

gin of discretion which may, according to the use which is made of it, make possible the execution of actual economic policy'<sup>1232</sup> or, in the wording of more recent case law, 'discretion which may, according to the use which is made of it, make it possible to take political decisions in the true sense, by substituting the choices of the delegator by those of the delegatee, and thus bring about an "actual transfer of responsibility"'<sup>1233</sup> may be somewhat less likely where only soft law powers are delegated. However, it is by far not improbable, and hence compliance with the *Meroni* limits ought to be scrutinised also in these cases.<sup>1234</sup> For example: The delegation of the power of the Commission to propose legislative acts to a different body would certainly distort the EU's institutional balance (which underlies the *Meroni* criteria) and thus be unlawful, even though the Commission's right to initiate legislative processes does not entail hard law powers.<sup>1235</sup> Whether the 'execution of actual economic policy' is possible by means of a soft law act is to be examined with a view to the concrete regulatory regime as a whole. The BEPG referred to in Article 121 TFEU (see 3.5.2.1.2. and 3.5.2.2.3. above), for example, are legally non-binding measures, but they 'lay down the scope and the direction of policy coordination of EU

1233 Case C-718/18 *Commission v Germany*, para 131.

General Court held that 'a grant of powers to such an entity is compatible with the requirements of the Treaties, if it does not concern acts having the force of law *and* if the powers granted are precisely delineated and amenable to judicial review' (emphasis added). While the wording of this passage suggests that the *Meroni* criteria in principle are to be applied also to soft law powers (*argumentum* 'and'), the ensuing contradiction of this statement (judicial review of EU soft law) renders it more likely that the General Court intended to say 'or', making these two elements alternatives. If that is the understanding of the General Court, in my view this understanding is not supported by the in this respect more open judgement in the case C-270/12 mentioned above, to which the General Court expressly refers.

<sup>1232</sup> Cases 9–10/56 *Meroni*, 152.

<sup>1234</sup> Apparently in favour of the applicability of *Meroni*: Opinion of AG *Bobek* in case C-911/19 *FBF*, para 86; Colombo/Eliantonio, Standards 334 f (with regard to standards); see also Senden/van den Brink, Checks 65 and 84, pointing at the risk of the *Meroni* criteria being circumvented by soft law, denying their applicability with regard to soft law and recommending, *de lege ferenda*, their adaptation so as to apply also to soft law powers; similarly: Rocca/Eliantonio, Soft Law 6; against the applicability of *Meroni* (with regard to the ESAs' powers to adopt guidelines and recommendations): Dickschen, Empfehlungen 215.

<sup>1235</sup> A Commission proposal is binding only to the extent that it determines the subject of the (to be adopted) act; see also II.2.3. above.

Member States'1236 and hence, due to their principal role in the procedure laid down in Article 121 TFEU,<sup>1237</sup> bring about the execution of economic policy. Even though it is clear that the BEPG are eventually 'recommended' by the Council directly on the basis of primary law, and hence the *Meroni* criteria, for lack of a delegation of powers *by* an EU body (on the basis of secondary law), are not applicable, this example illustrates that soft law may very well be a relevant instrument for the execution of economic policy.

# 3.7.3. Special competence clauses in public international law

Apart from primary and secondary EU law, EU bodies may also be vested with soft law powers on the basis of public international law. In order to exemplify this possibility, an act of public international law which shows a particularly strong proximity to EU law shall be drawn on: the TSCG. 1238 Article 7 of this Treaty provides that the MS of the Eurozone shall 'commit to supporting the proposals or recommendations submitted by the Europe‐ an Commission where it considers that a [Eurozone MS] is in breach of the deficit criterion in the framework of an excessive deficit procedure' (unless a qualified majority of these MS is against it). In other words, the Eurozone MS are obliged to act in accordance/decide in favour of such a Commission proposal or recommendation, unless a qualified majority of the Eurozone MS votes against it.1239 The Treaty does not, strictly speaking, *transfer* a soft law power upon the Commission, but it enhances the requirements for acting against the Commission's soft law by the MS of the Eurozone.

1239 With regard to this reverse qualified majority voting see Palmstorfer, Majority, in particular 192 f (with regard to the TSCG).

<sup>1236</sup> European Parliament, 'Broad Economic Policy Guidelines and Employment Guidelines' (2015) 1 <http://www.europarl.europa.eu/RegData/etudes/ATAG/ 2015/542652/IPOL\_ATA(2015)542652\_EN.pdf> accessed 28 March 2023.

<sup>1237</sup> See Schulte, Art. 121 AEUV, para 13.

<sup>1238</sup> With regard to this proximity see Article 16 TSCG, according to which this Treaty materially shall be incorporated in EU law within five years of its entry into force; for the current – not at all promising – stage of implementation of this plan see <http://www.europarl.europa.eu/legislative-train/theme-deeper-and-fairer-econ omic-and-monetary-union/file-integration-of-the-fiscal-compact-into-secondary -eu-law> accessed 28 March 2023; for the political reasons that the TSCG has not been adopted in the form of EU law in the first place see Fischer-Lescano/Obern‐ dorfer, Fiskalpakt 9 f; with regard to the relationship of complementation and/or proximity of public international law with EU law see decision of the German *Bundesverfassungsgericht* in case 2 BvE 4/11, para 100.

Thereby it creates a power of the Commission which is stronger than that provided for in the EU Treaties.<sup>1240</sup>

The possibilities for EU bodies to be granted competences by public in‐ ternational law are limited. The 'lending' of EU institutions (*Organleihe*<sup>1241</sup>) by public international law regularly requires the approval of all MS,<sup>1242</sup> in case of an *Organleihe* against current primary law it requires a formal amendment of the EU Treaties.<sup>1243</sup> The TSCG originally was not approved by all MS (the UK and the Czech Republic did not sign it<sup>1244</sup>), let alone accompanied by a Treaty amendment.<sup>1245</sup> Substantially, the conferral of powers shall, in particular, 'not alter the essential character of the powers conferred on those institutions by the EU and FEU Treaties'<sup>1246</sup> – a require‐ ment which it appears to be relatively easy to meet.<sup>1247</sup>

<sup>1240</sup> That the conferral of a proper decision-making power by an act of public interna‐ tional law would be problematic is at least suggested by the Court; case C-370/12 *Pringle*, para 161. For a somewhat similar example in the ESM-Treaty see Article 13 para 3 subpara 2 *leg cit*, according to which the Commission shall only sign a MoU (on behalf of the ESM) where it is in compliance with EU law, 'including any opinion, warning, recommendation or decision addressed to the ESM Member concerned'. Again, the effects of EU soft law (here: on public international law) are strengthened by an act not belonging to the EU legal order; see also 2.2.4.2.1. above. Note that in case of non-compliance of a provision of the TSCG with Union law, its Article 2 para 2 provides that this provision shall not apply; sceptically with regard to legal certainty: Fischer-Lescano/Oberndorfer, Fiskalpakt 13.

<sup>1241</sup> For an explanation of this term (in a different context) see van Hoek/Luchtman, Convention 494.

<sup>1242</sup> See Opinion of AG *Jacobs* in joined cases C-181/91 and C-248/91 *European Parlia‐ ment v Council*, para 13, also referring to the Commission's voluntariness; see also Craig, Pringle 268.

<sup>1243</sup> On the obligatory character of Article 48 TEU see Ohler, Art. 48 EUV, para 26. For the argument that 'almost everything' in the TSCG could have been enacted pursuant to the Treaties see Craig, Pringle 276.

<sup>1244</sup> The Czech Republic acceded to the TSCG in 2019 and the UK ceased to be a MS in 2020. Croatia joined the EU in 2013 and the TSCG in 2018.

<sup>1245</sup> For an account of the in this respect critical approaches of the government and parliament respectively of the UK: House of Commons, Treaty 15–18; sceptically as regards the evasion of the requirement of agreement among all MS brought about by international treaties facilitating what is called 'differentiated integration': Peers, Form 40.

<sup>1246</sup> Case C-370/12 *Pringle*, para 163.

<sup>1247</sup> See Craig, Pringle 278.

### 3.8. The effects of a lack of a legal basis

Having discussed the legal bases for EU soft law acts – in particular: recom‐ mendations and opinions – laid down in primary law, and selected exam‐ ples of such legal bases laid down in secondary law and, exceptionally, in public international law respectively, we may now consider the consequen‐ ces, if any, of a lack of an adequate legal basis for an EU soft law act.<sup>1248</sup> This issue, in practice, is handled with much leniency or – due to its (alleged) subordinate importance in the discussion about soft law – even ignorance, and it benefits from the Court's presumption of lawfulness which arguably also applies to EU soft law acts.<sup>1249</sup> These benevolent circumstances of answering the question whether there is a competence for the adoption of a certain soft law act should not be misunderstood as arbitrariness, though. It goes without saying that where EU law requires a legal basis for an act to be lawfully adopted, it must also provide for consequences where no such legal basis is available. While, procedurally speaking, the CJEU may not be called upon to review the legality of a (true) EU soft law act according to Article 263 TFEU, it has confirmed its competence to do so in a preliminary reference procedure. Furthermore it may, in the course of whichever procedure, incidentally evaluate such an act (for the judicial review of soft law see Chapter 6 below).

From a constitutional/administrative law perspective, the lack of an ad‐ equate legal basis leads to the voidness of the act at issue. Relative voidness and absolute voidness (nullity) are to be distinguished from each other.<sup>1250</sup> Relatively void acts of EU (soft) law apply 'until such time as they are annulled or withdrawn'.1251 This would be the regular case for EU soft law lacking an adequate (at least implicit) legal basis.1252 Where, however, exceptionally an act is 'tainted by an irregularity whose gravity is so obvious that it cannot be tolerated by the Community legal order [it] must be trea‐

<sup>1248</sup> Briefly addressing the entailing questions: Goldmann, Gewalt 502 f.

<sup>1249</sup> See case C-475/01 *Commission v Greece*, para 18, with further references.

<sup>1250</sup> Differenty with regard to soft law: Müller/Scholz, Banken 488; for the exceptional case of an *ab initio* nullity of an act, in which case it does not exert legal effects, see Dörr, Art. 263 AEUV, para 38.

<sup>1251</sup> Case C-137/92P *Commission v BASF*, para 48. With regard to the *ex tunc* effect of the Court's stating the nullity of an act in the course of a preliminary reference procedure see Müller/Scholz, Banken 489, with a further reference; with regard to an only partial annulment of apparent soft law see Pampel, Rechtsnatur 128 f.

<sup>1252</sup> See also Eliantonio, Soft Law 498.

ted as having no legal effect'.<sup>1253</sup> Such acts are absolutely void; legally they do not even start to exist. They are – in the form of soft law – non-existent, non-acts.<sup>1254</sup> For a soft law act this would mean that it does not entail those effects which the EU legal order usually provides for such act, that is to say it does not (softly) demand compliance, and not even consideration is required. The 'irregularity' referred to above must comprise severe mis‐ takes<sup>1255</sup> in the creation of the presumptive soft law act which may be of a procedural or of a substantive kind.<sup>1256</sup>

On a whole, the procedural requirements for the adoption of EU soft law may be lower than those for the adoption of EU law, but they do – in varying complexity – exist. In terms of substance, the complete lack of a normative content may cause the presumptive soft law act to be absolutely void. But for this qualification to be made, it must be clear that the creator of the norm actually intended to create a soft law act and not, for exam‐ ple, a paper issued for informative purposes only.1257 Due to the generally decreased procedural requirements for the creation of soft law, it is much more difficult to shed light on the intention of the creator of the act. This makes it harder to distinguish (intended) soft law from non-normative acts than to distinguish (intended) law from soft law or non-normative acts. Acts which, in the form of soft law, would be absolutely void, may easily be (re)interpreted as policy papers with no (intended) normative effect. This is why the absolute nullity of an EU soft law act – and, in connection therewith, the question whether it can then be reinterpreted as something else – is rather a theoretical problem.

Practically more relevant is the relative voidness of EU soft law. In the case of the halted excessive deficit procedures against Germany and France1258 the Court, *inter alia*, declared Council recommendations unlaw‐

<sup>1253</sup> Case C-137/92P *Commission v BASF*, para 49; see also case C-235/92P *Montecatini*, paras 96–98.

<sup>1254</sup> See case C-235/92P *Montecatini*, para 77; see also Dörr, Art. 263 AEUV, para 38.

<sup>1255</sup> For the reasons for (relative) nullity of a legally binding act see Article 263 para 2 TFEU; denying the existence of a *Fehlerkalkül* for a certain type of soft law output, namely the guidelines of European agencies: Raschauer, Leitlinien 42.

<sup>1256</sup> See Hofmann/Rowe/Türk, Administrative Law 641, stating (with further referen‐ ces) that the criteria for absolute voidness, according to the CJEU's case law, have become 'clearer', but at the same time 'more difficult to fulfil than was the case in the early days of the E(E)C; critically: Senden, Soft Law 288 f.

<sup>1257</sup> For the case of a 'purely informative' Eurogroup statement see case T-327/13 *Mallis*, paras 60 f.

<sup>1258</sup> For the background to these cases see Hodson/Maher, Soft law 801f.

ful for their wrong legal basis and for a procedural shortcoming. The Court, arguably aware of its incompetence – according to the then preva‐ lent view of the Court – to annul a legally non-binding act of EC law, referred to '[t]he *decision* to adopt […] recommendations' (emphasis added) and annulled the Council's conclusions 'in so far as they contain […] a *deci‐ sion* modifying the recommendations previously adopted by the Council' (emphasis added).<sup>1259</sup> This misleading terminology cannot, however, alter the fact that the Court annulled – as far as the modification of recommen‐ dations (and not the decision to hold the excessive deficit procedure in abeyance) is concerned – an EU soft law act, namely a recommendation of the Council.<sup>1260</sup> With the case law on the Court's powers under Article 267 TFEU having become more generous (see 6.3. below), the annulment of soft law by the Court may become a more frequent occasion. In the case *Balgarska Narodna Banka*, '[h]istory was made'1261 and a soft law act was declared invalid by the Court.<sup>1262</sup>

It is also possible that, for the sake of legal certainty1263 (in the context of legally non-binding acts admittedly an ambivalent expression1264) – alter‐ natively, and more broadly, we could say: for the sake of legal hygiene – the creator of the norm itself is called upon reacting to legality concerns (eg raised by the Court in the course of a preliminary reference procedure) and, if need be, to repeal/modify illegal soft law. Soft law, as is the case

<sup>1259</sup> Case C-27/04 *Commission v Council*, paras 95–97 (with regard to the annulment of the modification of the 'recommendations previously adopted by the Council under Article 104(7) EC'; differently: View of AG *Tizzano* in this case, paras 133–137. Note that Article 104 para 13 TEC actually referred to 'decisions' when addressing – *inter alia* – the recommendations of the Council according to para 7 *leg cit*.

<sup>1260</sup> See Häde, Aussetzung 757 f, with further references. For the legal non-binding‐ ness of the Council recommendations addressed here see also Streinz/Ohler/Herr‐ mann, Stabilitäts- und Wachstumspakt 1557.

<sup>1261</sup> Marjosola/van Rijsbergen/Scholten, Force 1523.

<sup>1262</sup> Case C‑501/18 *Balgarska Narodna Banka*.

<sup>1263</sup> For examples of cases in which soft law may actually increase legal certainty see references in Senden, Soft Law 333 (but see also 339).

<sup>1264</sup> See Jansen, Methoden 1f.

with law, may normally be repealed/modified<sup>1265</sup> at any time.<sup>1266</sup> This is to be done by a *contrarius actus*, by another soft law act *of the same kind* that is.<sup>1267</sup> That means, for example, that a soft law act with general application cannot be repealed/modified by a soft law act directed to an individual addressee. This is reflected in the *CIRFS* case in which the Court, although dealing with an act which eventually turned out to be legally binding, very generally held: 'A measure of general application can‐ not be impliedly amended by an individual decision'.<sup>1268</sup> In another case the Court required the *formal* amendment of internal (only self-obliging) Commission rules.<sup>1269</sup> In this context, AG *Tizziano* may be quoted who said: 'I observe first that, in accordance with a general principle, the power of the institutions to adopt a particular act necessarily also includes the power to amend that act, on condition that the provisions on the exercise of the relevant power are complied with. […] The opposite conclusion would have to be drawn, in my opinion, only if it were shown that the act being amended had been adopted as part of a rigidly regulated procedure which carried an obligation for the competent institution to adopt the subsequent act in the procedural chain by a set deadline, after which the institution lost the power to take a decision'.1270 Hence for the amendment of a certain soft law act in principle the same procedure as for its adoption is to be applied. Therefore a recommendation, for example, cannot be modified by a decision. The decision may contain norms constituting a modified version of the norms contained in the recommendation, which would thereby –

<sup>1265</sup> See the examples of the Commission's first Banking Communication which was 'withdrawn', as of 31 July 2013, by its second Banking Communication, 2013/C 216/01 (para 94); ECB Recommendation ECB/2014/2 amending Recommendation ECB/2011/24 on the statistical reporting requirements of the European Central Bank in the field of external statistics; see also Article 126 para 12 TFEU which stipulates that the 'Council shall abrogate some or all of its […] recommendations […] to the extent that the excessive deficit in the Member State concerned has […] been corrected'.

<sup>1266</sup> See von Graevenitz, Mitteilungen 172, with further references.

<sup>1267</sup> See case T-251/00 *Lagardère*, para 130: '[A] body which has power to adopt a particular legal measure also has power to abrogate or amend it by adopting a *contrarius actus*, unless such power is expressly conferred upon another body'; see also Braams, Koordinierung 163; Häde, Aussetzung 757.

<sup>1268</sup> Case C-313/90 *CIRFS*, para 44; see also Ştefan, Soft Law 170.

<sup>1269</sup> See case T‑185/05 *Italy v Commission*, paras 43 and 49.

<sup>1270</sup> View of AG *Tizzano* in case C-27/04 *Commission v Council*, paras 134 f.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

*qua* legal bindingness – oust the recommendation,<sup>1271</sup> but it cannot modify the recommendation itself. As regards the repeal of soft law, it appears that not only a *contrarius actus*, but also a hard law act with the same scope – individual-concrete or general-abstract, as the case may be – can, *qua* superiority, 'eliminate' a soft law act.

3.9. The revisitation of the above approaches on the difference between recommendations and opinions, on whether there is a *numerus clausus* of EU soft law acts, and on the principle of conferral

Having analysed in more detail the explicit primary law competences to adopt soft law, in particular recommendations and opinions, and having addressed selected examples of relevant competences laid down in secon‐ dary law, we shall now reconsider three key questions with a view to whether the approaches taken at the beginning of this chapter require an adjustment: whether primary law provides for a difference between recommendations and opinions, whether it provides for a *numerus clausus* of EU soft law acts, and whether the principle of conferral also applies in the context of EU soft law.

As regards the difference between recommendations and opinions in our study of the competence clauses laid down in particular in the Treaties, the initial characterisation of the two EU soft law acts mentioned in Article 288 TFEU appears to have been confirmed. While recommendations are rather prescribed where suggestions or the initiation of a decision-making process, or the adoption of general (soft) rules are at issue, opinions rather consti‐ tute the output of consultation on (draft) measures or on actions already taken. Opinions may contain an expression of view ('draft measure X goes against our interests', 'situation Y is unlawful'), whereas recommendations rather suggest specific action ('measure X should be adopted and situation Y should be addressed in this or that way').1272 The recommendation rather is an expression of *actio*, the opinion rather of *reactio*.

<sup>1271</sup> For the collision between law and soft law more generally see Klabbers, Redundan‐ cy 177. For the repeal of (part of ) a Directive by a Regulation see eg Regulation 1907/2006, 'amending' Directive 1999/45/EC by deleting its Article 14.

<sup>1272</sup> Article 154 para 3 TFEU provides that management and labour shall address an opinion or – 'where appropriate' – a recommendation to the Commission. This suggests a hierarchy of these two acts, meaning that the adoption of a recom‐

In practice, these subtle distinctions may easily blur. After all, and in particular due to the malleability of the terms at issue, the above charac‐ teristics do not apply in a compelling or absolute manner, but only by tendency. What is more, in some cases the characteristics of the act at issue may come close both to a recommendation and to an opinion, so that it appears to be little more than a matter of taste whether – *de lege ferenda* – in a certain procedure the power to adopt a recommendation or the power to adopt an opinion should be granted. Nevertheless, the final decision of the Masters of the Treaties or the legislator – in favour of a recommendation (and against an opinion) or in favour of an opinion (and against a recommendation), as the case may be – shall not generally be underrated as arbitrariness, because it must be assumed that most of the time it is the result of a conscious choice.<sup>1273</sup> Depending on this choice, the above understanding of recommendations and opinions shall feed into the interpretation of the respective act and the underlying competence clause, respectively.

As regards the question of whether Article 288 TFEU entails a *numerus clausus* of EU soft law acts, it is to be acknowledged that not only acts of secondary law but also the Treaties themselves provide for a variety of acts bearing different names, such as guidelines, warnings or conclusions.<sup>1274</sup> Only exceptionally, they have distinct effects, like the proposal according to Article 293 para 1 TFEU. Against this background, the *numerus clausus* argument does not seem to be supported by the Treaties other than by Article 288 TFEU itself. This confirms the preliminary conclusion reached above under 3.1.2.

mendation is more demanding, in terms of its adoption and/or in terms of its consideration by the addressee.

<sup>1273</sup> Take this as an example for a conscious choice on the part of the legislator: The Commission Proposal COM(2013) 27 final (Article 21) for what has become Article 25 of Regulation 2016/796 has provided for an ERA *recommendation* to be adopted. Article 25 of Regulation 2016/796, however, due to the Council's position at first reading in the ordinary legislative procedure, provides for the ERA to adopt an *opinion*. Another example is Directive 2002/21/EC which in its Article 15 laid down a competence of the Commission to render recommendations on relevant product and service markets. The Commission Proposal COM(2000) 393 final (in its Article 14) still stipulated a competence to adopt binding decisions. This competence was downgraded to a mere power to adopt recommendations in the course of the legislative procedure.

<sup>1274</sup> In addition to that, the institutions, bodies, offices and agencies of the EU have adopted various kinds of soft law acts, which are not explicitly (by name) provided for in EU law – neither in primary nor in secondary law.

The preliminary finding made under 3.2. above was that the main prin‐ ciple on the distribution of powers between the MS and the EU – the principle of conferral – does not only apply with regard to law-making powers, but also, for lack of an apparent alternative in the Treaties, with regard to soft law powers. From a legal point of view, the applicability can be confirmed with good reasons. But it must be acknowledged that there are also arguments against its applicability in the given context, and that also the case law of the Court leaves room for doubt here.

The existence of the competence category of supporting, coordinating and supplementing competences as laid down in Article 6 TFEU (which to a large extent entails the adoption of soft law acts<sup>1275</sup>), but also the existence of various explicit soft law powers, as analysed in more detail above, clearly are pro-arguments. If no conferral of powers were required for the adoption of EU soft law, why would the Treaties set up a competence category which is mostly concerned with the adoption of soft law acts, and why would they explicitly provide for specific soft law competences in so many different provisions? A heretic answer would be: Most of the time, these powers are explicated in order to structure the respective decision-making proce‐ dures. The Treaties clearly lay down eg the Commission's power to send a recommendation to the Council, for this institution to take a decision (see eg Article 143 para 3 TFEU), because this is how the MS – as parties to the Treaties – wanted the institutions to draw up the final decision. In other words: The intention of the provisions at issue is not to grant the Commission a power to send a recommendation to the Council, but to make clear how the procedure should go along. From my point of view, the procedural character certainly plays a role,1276 but it cannot do away with the competence-conferring nature of these provisions.1277 Both elements coexist. Where the Commission's power to adopt recommendations is at is‐ sue, it is easy to refer to the general competence clause of Article 292 TFEU, arguing that any provision mentioning this competence in a specific context is merely declaratory. <sup>1278</sup> But what about other institutions such as the ECB which are entitled to adopt recommendations only by special competence clauses? For them what was referred to above as special competence clauses is constitutive, not merely declaratory.

<sup>1275</sup> See Klamert, Article 6 TFEU, paras 11–13; Senden/van den Brink, Checks 22.

<sup>1276</sup> See also Article 292 (2nd sentence) TFEU.

<sup>1277</sup> See also Senden/van den Brink, Checks 21f.

<sup>1278</sup> Here it is the existence of Article 292 TFEU itself which supports the argument that the principle of conferral applies to soft law powers.

A middle way between the application of the principle of conferral on the one hand and the view that no competence at all is required for the adoption of EU soft law, on the other hand, is the *in dubio* approach.<sup>1279</sup> Also pursuant to this approach, EU soft law acts may only be adopted if an according competence exists. However, this competence can be assumed as long as there is no counter-indication, eg no legislative power of the Council excluding the power of the Commission to adopt general recom‐ mendations in this field. As the author contended above, while parts of the case law seem to support this view, the applicability of the principle of conferral – from a legal point of view – is more convincing.

From a practical perspective, the differences between the application of the principle of conferral and the *in dubio* approach seem to be marginal. For two institutions – the Commission and the Council – Article 292 TFEU, with regard to recommendations, provides for a generous regime anyway. Apart from that, it is in particular the application of the implied powers doctrine which allows the institutions, bodies, offices and agencies to adopt soft law far beyond their explicit empowerment in the Treaties or – on the basis of the latter – in secondary law. This concessive approach seems to reflect the system of the Treaties.<sup>1280</sup>

### 3.10. Résumé and transition

Before turning to the effects of soft law, let us briefly summarise the findings made in this chapter. Once the applicability of the principle of conferral with regard to soft law was discussed and preliminarily confirmed at the outset of this chapter, its focus was laid on the characteristics and the legal bases for the adoption of recommendations and opinions (the

<sup>1279</sup> Arguing against the applicability of the principle of conferral, but in favour of a '"lite" competence test' in respect of soft law powers: Korkea-aho, Courts 489 f.

<sup>1280</sup> If bodies of subordinate importance such as the ESC or the CoR have a far-reach‐ ing power to adopt opinions, it would be paradoxical to be overly strict with insti‐ tutions such as the European Parliament. Still, the EP's legally non-binding output is to be assessed critically where there does not seem to exist a relevant competence (not even an implied one), eg its manifold resolutions relating to foreign policy – a field (CFSP) where the EP, with only few exceptions, is excluded from taking action. In our context, it is to be conceded that many of these resolutions do not have normative content, but are merely 'political'. Nevertheless, in some cases the result of a diligent application of the principle of conferral would probably exclude this kind of interference on the part of the EP.

two legally non-binding 'legal acts' laid down in Article 288 TFEU), as provided for, explicitly or only implicitly, in primary law. On the basis of selected examples, a complementary view was taken, first, on the power to adopt soft law acts provided for in primary law which are not designated 'recommendation' or 'opinion' and, second, on legal bases for the adoption of recommendations and opinions laid down in secondary and in public international law. Thereafter, the effects on soft law of a lack of a legal basis were analysed. Eventually, the findings on the conceptual difference between recommendations and opinions, on the question whether the Treaties provide for a *numerus clausus* of soft law acts and on the applica‐ bility of the principle of conferral were revisited against the background of the broader picture of soft law competences in the Treaties, as provided for in Chapter 3.

In order to allow for a more nuanced conclusion, we shall concentrate on three main issues: 1) the applicability of the principle of conferral and its ramifications; 2) the different categories of legal bases; 3) the different functions of soft law as laid down in primary law. Subsequently, a transition to the next chapter shall be attempted.

1) It has been found that the better reasons speak in favour of the applicability of the principle of conferral in the context of soft law acts. The rule of law, one of the fundamental principles on which the EU is based, cannot allow for a normative system (with, in part, highly significant effects) emanating from it, which in respect of the fundamental question of competences is detached from the requirements of law.1281 This result – that the principle of conferral applies also to soft law powers – is by no means mundane. The consequence, namely that each and every soft law act needs to be set up in law, indirectly at least also in primary law, is remarkable. The comparably generous case law of the Court on implied powers, which can be made use of also in the context of soft law, allows for affirming implicit competences, especially were a related hard law power already exists. This makes it considerably easier to confirm the existence of a legal basis for soft law, but by no means does it render the question of competence pointless, nor should the examination of this question be understood as a mere formality.

In the course of analysing the legal framework, in particular the case law of the Court, indicators of an alternative approach with regard to soft law powers could be found: Where no rule to the contrary exists, a soft

<sup>1281</sup> See case 294/83 *Les Verts*, para 23.

law power can be presumed (*in dubio* approach). While the methodological techniques applied in these two regimes – a positive and a negative one – vary, it is to be conceded that in practice, in particular due to the concessive effect of the implied powers doctrine, their respective results may be very similar.

2) As regards competences laid down in primary law, a distinction was made between general competence clauses on the one hand and special competence clauses on the other hand. Whereas general competence clau‐ ses allow for an encompassing power to adopt – in our case – recommen‐ dations or opinions, special competence clauses provide for such power only in a certain policy field or, even more restrictively, in a certain decision-/rule-making procedure. While the limits of special competence clauses are inherent to the clauses themselves, also general competence clauses are, in one way or the other, restricted in order to make sure that law-making procedures are not being evaded and that the powers of other institutions, bodies, offices or agencies – in a spirit of loyalty among these actors – are not being thwarted.

The respective competence clauses laid down in secondary law must all root in primary law, and an empowerment of EU bodies *qua* public international law shall, at least, 'not alter the essential character of the powers conferred on those institutions by the EU and FEU Treaties'.<sup>1282</sup>

3) The use of soft law may serve different political functions. Law being 'geronnene Politik', coagulated politics that is, it does not come as a surprise that at least some of these functions are also reflected in the Treaties.<sup>1283</sup> From the general competence clauses three main functions of soft law could be extracted: the support of decision-/rule-making, the initiation of (soft) decision-/rule-making, and soft decision-/rule-making. This underlines the versatile character of soft law which allows for flexibility not only because of its legal non-bindingness, but also because of the different forms it may take and hence the various contexts in which it may be used; for a more encompassing picture of the purposes of soft law see Chapter 5 below.

One final point is to be mentioned here, which did not constitute the focus of this chapter, but which emerged as a collateral finding, as it were. It is the special effects explicitly laid down in law, which some of the soft law acts dealt with here have. Recapitulating some of these effects shall form a transition to the subsequent chapter which addresses the different

<sup>1282</sup> Case C-370/12 *Pringle*, para 163.

<sup>1283</sup> See Goldstein/Kahler/Keohane/Slaughter, Introduction 387.

categories of effects of soft law. Three examples (discussed above) shall serve as an illustration. First example: Where the Council acts on a propos‐ al from the Commission, it may – exceptions apart – amend that proposal only by acting unanimously. Second example: According to the Statute of the EIB, the Management Committee of the EIB and the Commission shall submit an opinion on proposals for granting of finance. Where they both have launched a negative opinion, the Board of Directors of the EIB may not grant the finance concerned. Third example: Article 7 of the TSCG obliges the MS of the Eurozone to 'commit to supporting the proposals or recommendations submitted by the European Commission where it considers that a [Eurozone MS] is in breach of the deficit criterion in the framework of an excessive deficit procedure', unless a qualified majority of these MS votes against it.

In these examples the effects of the respective acts are increased as compared to the regular effects of soft law. While the first one makes it more difficult to deviate from it, the second one – in case of two negative opinions from different bodies – obliges the addressee to comply, and in the third one again the commitment to soft law is increased. This shows that not only may soft law reach different degrees of (factual) effectiveness (ie different compliance rates), but its effects may also be reinforced *by means of law*. While in the first and in the third example the respective acts remain legally non-binding, in the second example the two acts – in combination – entail a legally binding (prohibitive) effect. These and other effects of soft law shall be dealt with in the following chapter.

### *4. Legal, factual and mixed effects of soft law*

#### 4.1. Introduction

Having discussed the morphology of EU soft law and its legal bases, we shall now turn to its effects, that is to say the effects it has on its addressees. Compliance with rules can have many reasons, only one of which is the motive of norm-abidingness. Other motives may be fear from 'sanctions' (in case of soft law this may be eg peer pressure), the conviction that the (compliant) behaviour is morally right, reciprocity of the norm at issue, convenience or politeness. These motives may take effect consciously or subconsciously. Either way, they reach far into the personality of the indi‐ vidual/collegiate addressee, which is why the actual number of motives is in fact indeterminable. A necessary prerequisite for the decision whether or not to comply is consideration, that is to say the taking into account of a rule. Without consideration, the addressee cannot know about the norma‐ tive content of the respective soft law act and hence (non-)compliance is entirely incidental. Whether or not soft law is considered or even complied with again depends on its respective effects, and these effects should be divided here in legal, factual and mixed effects.

The motives mentioned above have been evoked by the factual effects, but they may be underpinned by legal effects, ie effects laid down in law. In spite of the legal non-bindingness of soft law rules, there are legal rules or principles suggesting the consideration of or even the compliance with them. Schematically speaking, to ask for the factual effects means to ask why soft law *is* considered/complied with, whereas the legal effects tell us why soft law *should be* considered/complied with. While the 'voluntary compliance with the non-binding acts of the institutions is an essential element in the achievement of the goals of the Treaty',1284 as the Court put it, the EU's legal order itself contributes to obtaining this 'voluntary' compliance. This appears to be necessary but not always sufficient, because the compliance rates with soft law, in the respective norm-creator's perspec‐ tive, are sometimes hardly satisfactory, as the example of Commission Rec‐ ommendation 2011/442/EU on access to a basic payment account shows. Eventually, this recommendation was, for lack of satisfactory compliance rates, replaced by Directive 2014/92/EU (the so-called 'Payment Accounts Directive').1285 In the respective legislative proposal the Commission stated: **'**Compliance with the Commission's Recommendation on access to a basic payment account was also largely inadequate […]. The introduction of a binding measure is the most effective and efficient way of achieving the set objectives. Only a binding legislative instrument can guarantee that the policy options are introduced in all 27 Member States and that the rules are enforceable'. <sup>1286</sup> This is an example of an outright replacement of soft law by law.1287 In other cases the norm-creator contents itself with

<sup>1284</sup> Case T-113/89 *Nefarma*, para 79.

<sup>1285</sup> Directive 2014/92/EU on the comparability of fees related to payment accounts, payment account switching and access to payment accounts with basic features; see Lengauer/Weismann, Zahlungskonten-Richtlinie.

<sup>1286</sup> Commission proposal COM(2013) 266 final, 10.

<sup>1287</sup> See also, for example, Article 6 para 4 of Regulation 2019/942, which carries a similar *telos*: '[The ACER] shall promote cooperation between the [national] regulatory authorities and between regulatory authorities at regional and Union

measures making non-compliance with soft law less attractive, or at least less easy, eg by increasing, in the case of a soft law act addressed to the Council, the majority required for a decision not to follow it. The majority requirement in combination with the factual (increased) difficulty to reach this (larger) majority promotes compliance. Since here the legal aspect, the majority requirement, without itself containing a legal obligation to abide, is clearly linked to the factual reason – the increased ('political') difficulty to reach the larger majority – this and similar constellations shall be addressed under the heading 'mixed effects'. The factual effects in practice are highly important, but they – and the factual aspects of the mixed effects – can be better assessed applying quantitative, not genuinely legal methodology. This is why here they are addressed only cursorily to broaden the view, and, more particularly, to acknowledge the close inter-relation between legal and factual effects when it comes to the application of soft law in practice. This inter-relation is expressed not only in what the author refers to as mixed effects, but in practice exists also beyond this category.<sup>1288</sup>

As regards the legal effects, one sub-chapter shall be dedicated to the effects on MS, and one to the effects on EU institutions, bodies, offices and agencies. In both sub-chapters first the effects as fleshed out by the Court in its case law shall be presented. In this context it is to be emphasised, first, that the Court's jurisprudence is (constantly developing) case law, and, second, that each soft law act is individual. While the legal effects of two different soft law acts may be the same in a certain instance and while we can assume that, in principle, the results of a Court case on a specific soft law act are relevant also for other soft law acts, these two factors shall remind us not to generalise the effects of soft law uncritically and ignorant of the specificities of the individual case.1289 Subsequently, the actual or potential legal reasons for these effects – namely certain principles

level and shall take into account the outcome of such cooperation when formu‐ lating its opinions, recommendations and decisions. Where the ACER *considers that binding rules on such cooperation are required*, it shall make the appropriate recommendations to the Commission' (emphasis added).

<sup>1288</sup> For example: The addressee of soft law may partly comply with it because it deems it ought to (legal effect), partly because it is convinced by its content (factual effect); for the (in the practice of EU bodies neglected) role of argumentation in soft law acts see Coman-Kund/Andone, Instruments 183; Andone/Coman-Kund, EU soft law 12 ff.

<sup>1289</sup> Arguing for a case by case assessment of whether the Court's case law is applicable: Wörner, Verhaltenssteuerungsformen 190 f, with references to different views in the literature; see also Hofmann/Rowe/Türk, Administrative Law 568.

or provisions of primary law – shall be addressed. This is necessary because the Court does not always mention the reason *why* the soft law act at issue in a certain case entails the legal effects established in the judgement. Sometimes it merely states the effects.

The case of individuals/undertakings shall not be addressed here under a separate heading, but only to the extent that their being addressed may cause duties of the MS or of EU institutions, bodies, offices or agencies. The effects soft law may have on individuals/undertakings are not relevant for the compliance mechanisms presented and discussed in the following parts of this work. Beyond its legal non-bindingness,<sup>1290</sup> the Court so far has ad‐ dressed the legal effects of EU soft law on individuals/undertakings mainly in the context of legitimate expectations/equality.<sup>1291</sup> **T**he general<sup>1292</sup> factual effects apply, *mutatis mutandis*, also to individuals/undertakings, the most important one being the fear from disadvantages, eg a fine imposed by the Commission.<sup>1293</sup>

### 4.2. Legal effects

### 4.2.1. Introduction

It can hardly come as a surprise that in a legal study the first and most intensively dealt with effects of soft law shall be its legal effects. But what are 'legal effects'? There is no uniform use of the term in legal discourse. When discussing this issue, we have to bear in mind two general theorems: 1) From the fact that an act is complied with (is effective, that means), no deductions can be made as to its legal nature (see also II.1.3.2. above). 2) Soft law is hierarchically subordinate to law and may not be contrary to law.

According to Article 263 TFEU, the Court shall review the legality of acts adopted by EU institutions, bodies, offices or agencies 'intended to produce

<sup>1290</sup> Case C-322/88 *Grimaldi*, para 16: '[Recommendations] cannot create rights upon which individuals may rely before a national court'.

<sup>1291</sup> For the legitimate expectations of individuals which soft law may create see 4.2.2.2.4. and 4.2.3.2.3. below; for further (potential) legal effects see H Adam, Mitteilungen 124, with regard to Commission communications.

<sup>1292</sup> These do not include the special scenario addressed under 4.3.2.2. below.

<sup>1293</sup> Case 60/81 *IBM*, para 19.

legal effects'.<sup>1294</sup> With 'legal effects' the Court primarily understands 'legal bindingness' *vis-à-vis* third parties.<sup>1295</sup> This is in conformity with Article 288 TFEU which states that recommendations and opinions – two acts which are excluded from judicial review under Article 263 TFEU<sup>1296</sup> – 'shall have no binding force'. To which extent the Court is ready to acknowledge 'legal effects' other than legal bindingness is unclear. That it does in princi‐ ple, it has famously expressed in the *Grimaldi* judgement: '[T]here is no reason to doubt that the measures in question are true recommendations, that is to say measures which, even as regards the persons to whom they are addressed, are not intended to produce binding effects […]. However, […] the measures in question cannot […] be regarded as having no legal effect'.<sup>1297</sup>

A broader understanding of 'legal effects' could be the following: The legal effects of soft law are determined exclusively by law – above all by the requirement of an according competence to adopt soft law. As was mentioned above (II.1.3.), soft law is strongly connected to and exists only on the basis of law – it is a 'legal product', as it were.1298 The decisive difference between law and soft law is that the former is legally binding, the latter is not. Hence the legal effects of soft law may not be legal bindingness (otherwise it would be law), but they are always legal – determined by law that is. This is the understanding which is applied in this chapter for the purpose of a categorisation of the effects of soft law. The effects of soft law are also called 'legal' to distinguish them from factual effects. That way, the (potential) legally prescribed effects of soft law – eg that it ought to be considered – are conceptually distinguished from eg the human desire to rely on official rules, as evoked by a soft law act. Exceptionally, EU soft law may be granted legally binding force by other acts, eg the ESM-Treaty, according to which '[t]he MoU shall be fully consistent with the measures

<sup>1294</sup> For the inspiration this phrase has drawn from case 294/83 *Les Verts* see Senden, Soft Law 237.

<sup>1295</sup> See eg case C-562/12 *Liivimaa Lihaveis*, para 46.

<sup>1296</sup> It is only the recommendations and opinions of the Council, the Commission and the ECB which are *explicitly* excluded; see also 6.2. below.

<sup>1297</sup> Case C-322/88 *Grimaldi*, paras 16 and 18; for the Court's original term 'legal effects' gradually changing to 'binding legal effects', indicating a larger variety of different legal effects, see Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 69–71.

<sup>1298</sup> See already case 310/85 *Deufil*, para 22; for later case law referring to this decision see eg case T-110/97 *Kneissl*, para 51.

of economic policy coordination provided for in the TFEU, in particular with any act of European Union law, including any opinion, warning, recommendation or decision addressed to the ESM Member concerned' (Article 13 para 3 *leg cit*).<sup>1299</sup> Such measures elevating the legal status of soft law may also be acts of EU secondary law explicitly incorporating as legally binding a certain soft law act.<sup>1300</sup> In this case, the content of soft law is rendered legally binding by means of law.

Sometimes also the expression 'indirect legal effects' of soft law is chosen in legal literature.<sup>1301</sup> What it means is, abstractly speaking, that soft law applied in combination with law indirectly becomes legally effective. For example: Where soft law is used to determine the content of fundamental principles of law or to interpret a Treaty provision, its content is applied 'as law' because both fundamental principles of law and the Treaty provi‐ sion are legally binding.1302 In my understanding, the practice that law is interpreted in accordance with pertinent soft law regularly<sup>1303</sup> reflects a factual effect: Due to the factual authority of soft law it is referred to in the interpretation of law. Where a legal norm makes reference to a soft law act, the content of the latter may become binding due to this reference (see also II.2.2.3. above).

As was mentioned above, the discussion of the legal effects of soft law shall be split in effects on MS on the one hand, and on institutions, bodies, offices and agencies of the EU, on the other hand. It comprises not only the Court's case law, but also legal principles of EU law which may (potentially) be the legal bases for the legal effects of soft law. The effects of EU soft law have been considered by the CJEU in a number of cases in which it

1303 The effects are legal only where law prescribes such interpretation; see eg the (legal) effects of the (non-binding) Explanations relating to the CFR, as provided for in Article 6 para 1 subpara 3 TEU.

<sup>1299</sup> For the interlinkage of economic policy soft law with financial support under the Recovery and Resilience Facility, thereby increasing its authority, see Bekker, Recovery.

<sup>1300</sup> See Georgieva, Soft Law 257, with reference to the pertinent cases *Mangold* and *Kücükdeveci*; for the technique of referencing in EU law see Sarmiento, Soft Law 271–273; for the technique of referencing more generally see II.2.2.3. above.

<sup>1301</sup> See H Adam, Mitteilungen 118 f; Georgieva, Soft Law 237 ff; see also Senden, Soft Law, in particular chapters 8–10. *Senden* takes a more differentiated approach, though; ibid 242 f. The terminology arguably is inspired by the 'indirect effect' of law used in the context of directives; see eg Craig/de Búrca, EU Law 244 ff.

<sup>1302</sup> See Snyder, Agreements 463, who speaks of 'legal effects' to describe, for example, that an act may serve 'as source of information and […] aid in judicial interpreta‐ tion'.

has described – in a multi-variant way – the location of soft law in the conceptual space between legal insignificance and legal bindingness.

### 4.2.2. Member States

4.2.2.1. The effects of soft law according to the Court's case law

### 4.2.2.1.1. Introduction

Already in its earlier case law the Court confirmed its competence and its principal obligation to interpret legally non-binding acts of then Commun‐ ity bodies.<sup>1304</sup> It has unmasked both (*prima facie*) soft law to be law and *vice versa*. <sup>1305</sup> In its judgement in the famous *Grimaldi* case, rendered in 1989, it approached the issue of Community soft law more comprehensively. In this case, the *tribunal du travail* of Brussels referred to the Court*, inter alia*, a question on the interpretation of a Commission recommendation addressed to the MS which concerned the adoption of a European schedule of occupational diseases. *Salvatore Grimaldi*, an Italian having worked in Belgium for about 30 years, contested a decision of the *Fonds des maladies professionnelles* which did not recognise the *Dupuytren's contracture*, from which *Grimaldi* suffered, as an occupational disease. The reason for the negation was that the Belgian schedule of occupational diseases, unlike the Commission recommendation mentioned above, did not list this disease.

With regard to the above Commission recommendation and with regard to a Commission recommendation on the conditions for compensation of persons suffering from occupational diseases, the Court concluded that 'there is no reason to doubt that the measures in question are true recom‐ mendations, that is to say measures which, even as regards the persons to whom they are addressed, are not intended to produce binding effects. Consequently, they cannot create rights upon which individuals may rely

<sup>1304</sup> See case 113/75 *Frecassetti*, paras 8 f. In this case the Court does not deny the possibility to ask it for the interpretation of a legally non-binding Community act, even though it concludes that the recommendation at issue is not applicable in the given case; see also case 90/76 *van Ameyde*, para 15.

<sup>1305</sup> See joined cases 8–11/66 Cimenteries, 92; case 19/67 Bestuur der Sociale Verze‐ keringsbank, 355; case 98/80 Romano, para 20. For an implicit reference of the legislator to this case law see Recital 46 of Regulation 2023/1114, with regard to binding opinions the ECB may adopt under this act.

before a national court'.1306 However, the Court adds, this does not divest such acts of any 'legal effect'.<sup>1307</sup> They may 'cast light on the interpretation of national measures' and in particular then national courts are obliged to at least consider them.<sup>1308</sup>

The contested<sup>1309</sup> *Grimaldi* judgement – which arguably can be applied also in the context of other forms of EU soft law<sup>1310</sup> – is only one (impor‐ tant) piece in the large puzzle of Court cases dealing with soft law. It addresses predominantly the legal effects of Commission recommendations on national courts, and – indirectly – also on the national executive.<sup>1311</sup> After this introduction through *Grimaldi*, we shall take a more systematic approach, covering the effects of various soft law acts on the *trias politica* of the MS: the legislative, the executive, and the judiciary.

# 4.2.2.1.2. The effects on the legislative, the executive, and the judiciary of the MS

In respect of the effects of EU soft law on the MS, the Court generally stated that 'voluntary compliance with the non-binding acts of the institutions is an essential element in the achievement of the goals of the Treaty' and that 'express conferral [by the MS] of the power to adopt [such acts]' shows that.1312 While compliance with EU soft law by the national legislator is indeed welcomed and certainly in the spirit of the Treaties, as it were, this cannot do away with the leeway MS have when confronted with EU soft law.1313 This leeway excludes a duty to adopt national legislation in

<sup>1306</sup> Case C‑322/88 *Grimaldi*, para 16.

<sup>1307</sup> Case C‑322/88 *Grimaldi*, para 18.

<sup>1308</sup> Case C‑322/88 *Grimaldi*, paras 18 f.

<sup>1309</sup> See eg Ruffert, Art. 288 AEUV, para 97, who describes the duty of national courts to consider recommendations as determined by the Court in *Grimaldi* as a bind‐ ing effect, and hence as 'Rechtsfortbildung gegen den Vertragstext' [legal develop‐ ment against the text of the Treaty].

<sup>1310</sup> See Kovács/Tóth/Forgác, Effects 65, with a further reference.

<sup>1311</sup> See Sarmiento, Soft Law 267: 'nothing stops [the said effects on national courts] from being extended to national administrations as well'; supporting this view: case C-274/07 *Commission v Lithuania*, para 50.

<sup>1312</sup> Case T-113/89 *Nefarma*, paras 79 and 85.

<sup>1313</sup> See case 229/86 *Brother Industries*, 3763; see also Cannizzaro/Rebasti, Soft law 222 f, with references to further literature.

accordance with soft law, let alone a duty to actually transpose the content of soft law into national law.<sup>1314</sup>

This leeway appears also to apply to the MS' executive. When the Court in the *Expedia* case states that 'the competition authority of a Member State may take into account the thresholds established in paragraph 7 of the *de minimis* notice [a Commission soft law act] but is not required to do so',<sup>1315</sup> *prima vista* this appears to be in contradiction with earlier judgements insisting on a duty to consider. However, the wording in the language of the case, ie French, is, in this respect, slightly different: 'l'autorité de con‐ currence d'un État membre peut prendre en considération les seuils établis au point 7 de la communication de minimis sans pour autant être obligée de s'y tenir'. This means that the authorities may take into account the *de minimis* notice, but are not obliged *to follow* it. Also the (other) translations considered here1316 seem to reflect this meaning. However, also in these language versions consideration is voluntary.1317 The *Cour de cassation* in *Expedia* – a preliminary reference case – did not specifically ask whether such a duty to consider exists,1318 which is probably why the Court touches upon this issue only briefly.

AG *Kokott* in her Opinion in the *Expedia* case takes a much more pronounced stance on this question, claiming that the national authorities and courts cannot 'simply ignore[]' but 'must take due account of the Com‐ mission's competition policy notices, such as the *de minimis* notice, when exercising their powers under Regulation No 1/2003'.1319 Since the Court did not make any reference to this passage of the AG's Opinion – neither to confirm nor to refuse it – the AG's approach, while being of interest as a legal opinion, does not help to elucidate the Court's view in this respect.

With regard to a Council recommendation on electricity tariff structures in the Community, the Court simply confirmed that it is not binding upon

<sup>1314</sup> See eg Opinion of AG *Geelhoed* in case C-478/99 *Commission v Sweden*, paras 37–40.

<sup>1315</sup> Case C-226/11 *Expedia*, para 31.

<sup>1316</sup> These are the German, the Italian and the Spanish translation of the judgement.

<sup>1317</sup> Similarly: Láncos, Core 775; see also case 37/79 *Marty*, para 10, and – likewise on the Commission's comfort letters which were in use in competition law – joined cases 253/78 and 1–3/79 *Procureur de la République*, para 18 (*argumentum* 'in no way').

<sup>1318</sup> See case C-226/11 *Expedia*, para 13.

<sup>1319</sup> Opinion of AG *Kokott* in case C-226/11 *Expedia*, para 38. For the application of the *de minimis* notice in Dutch case law see Senden, Soft Law 139.

MS1320 and in respect of Explanatory Notes of the Commission it stated that they are an 'important aid to the interpretation of the scope of the various tariff headings but do not have legally binding force'.<sup>1321</sup> In a recent case on one of the Commission's Communications on State aid for banks,<sup>1322</sup> the Court confirmed that it is not binding upon MS (in general), without thereby dwelling on whether or not MS are obliged to at least consider it.<sup>1323</sup> In this case again it is the AG who elaborates this issue, claiming CJEU case law to lay down a duty of the MS authorities to take into account EU soft law.<sup>1324</sup>

As regards national courts, the CJEU in *Grimaldi* requested them to 'take into consideration' Commission recommendations (see above).<sup>1325</sup> In the aftermath of *Grimaldi*, the Court has, with different nuances in different


<sup>1320</sup> See case C-207/01 *Altair Chimica*, para 43.

<sup>1321</sup> Case C-666/13 *Rohm*, para 25, with references to further case law. For the case of a code of concuct of the Commission which went beyond the secondary law it intended to explain (and was therefore, as a 'measure intended to have legal effects of its own', annulled) see case C-303/90 *France v Commission*, paras 18–26.

cases, roughly followed this path.1326 The Court has reiterated many times that recommendations are 'not without any legal effect'.<sup>1327</sup> Rather, national courts are '*bound* to take recommendations into consideration […], in particular where such recommendations cast light on the interpretation of national measures adopted in order to implement them or where they are designed to supplement binding provisions of EU law' (emphasis add‐ ed).<sup>1328</sup>

In the *Kreussler* case the Court held, with regard to Commission guide‐ lines, that they 'may provide useful information for the interpretation of the relevant provisions of European Union law and therefore contribute to ensuring that they are applied uniformly' which is why the national court 'may […] take account of that document'.<sup>1329</sup> It is unclear whether the Court – not referring to *Grimaldi* – wanted to apply a more lenient approach (*argumentum* 'may'1330) or whether this is an insignificant change of word‐ ing. Markedly different was the Court's wording in the case *Koninklijke* in which it held – with regard to the effects a general-abstract Commission recommendation has on the national courts, thereby referring to the *Arcor* case (which again makes reference to *Grimaldi*) – that 'a national court may depart from [this recommendation] only where [...] it considers that this is required on grounds related to the facts of the individual case, in particular the specific characteristics of the market of the Member State

1327 Joined cases C-317–320/08 *Alassini*, para 40.


<sup>1326</sup> See case C-188/91 *Deutsche Shell*, para 18; case C-207/01 *Altair Chimica*, para 41; case C-410/13 *Baltlanta*, para 64, with regard to Commission guidelines; see also case C-89/95P *D*, paras 8 f, in which the Court held – with regard to a code of practice of the Commission, annexed to a recommendation – '[w]hoever is called on to judge the facts alleged has absolute discretion in assessing the probative value of the evidence adduced', the code not prohibiting the taking into account of specific evidence. Since the Court holds that the code 'cannot be interpreted' as prohibiting certain evidence in the relevant proceedings, it does not feel required to say anything about the legal bindingness of the code.

<sup>1328</sup> Joined cases C-317–320/08 *Alassini*, para 40; see also Opinion of AG of *Ruiz-Jarabo Colomer* in case C-415/07 *Gennaro*, para 34; case C-911/19 *FBF*, para 71. In the *Bal‐ garska Narodna Banka* case, the Court confirmed the obligation of national courts to take a certain EU recommendation 'into consideration', thereby referring to a recital of the underlying act of secondary law which placed the recommendation within a regime aimed at '[e]nsuring the correct and full application of Union law'; case C‑501/18 *Balgarska Narodna Banka*, para 81; less strict in case C-62/20 *Vogel*, para 31: 'should be […] consulted'.

in question'.1331 This finding was later confirmed in another case where a duty to give reasons for deviating from a recommendation was explicitly provided for in secondary law.<sup>1332</sup>

As set out above (4.2.2.1.1.), it can be assumed that where the Court confirms a duty of national courts to take into account certain soft law, this duty applies also to administrative authorities.<sup>1333</sup> After all, these courts regularly (and among other things) decide on the legality of the decisions of the executive. It would therefore be inconsistent in this context not to apply – in principle – the benchmark applicable to the judiciary also to the executive.

In view of the Court's heterogeneous case law, it appears that also the specific creator of soft law and the broader legal context within which it was adopted are relevant for the legal effects it entails. This is underpinned by case law on acts produced by other actors than the EU institutions, in which the Court has considered this question only cursorily – with no clear outcome.<sup>1334</sup>

In conclusion, as regards the legal effects of soft law on the MS the Court does not appear to distinguish clearly between the three branches of public authority of the MS. Also as regards the (assumed) obligation to consider EU soft law, the Court's case law is fragmented. Nevertheless, there is a considerable body of case law in which a duty, on the part of the MS, to take the EU soft law act at issue into account is confirmed.

### 4.2.2.1.3. *Prima facie* soft law which turns out to be legally binding

In the cases dealt with so far the Court has constantly upheld, even if not with an entirely homogeneous wording, the legal non-bindingness of the respective soft law act. On the contrary, in a number of other cases the Court stated that a *prima facie* soft law act was 'intended to produce binding legal effects'.1335 The Court, for example, considered a Commission

<sup>1331</sup> Case C-28/15 *Koninklijke*, para 42.

<sup>1332</sup> Case C-277/16 *Polkomtel*, para 37.

<sup>1333</sup> Sceptical: Arroyo Jiménez, Bindingness 21.

<sup>1334</sup> See case C-106/14 *FCD*, para 28 (with regard to ECHA guidance); case C-613/14 *Elliott*, para 43 (with regard to private standards published by the Commission).

<sup>1335</sup> Case C-355/10 *European Parliament v Council*, para 82; joined cases C‑463/10P and C‑475/10P *Deutsche Post*, paras 30 and 45; case T‑96/10 *Rütgers*, para 34; case T‑676/13 *Italian International Film*, para 34; for an early case in this context see

Communication to exert binding effects due to its 'imperative wording' and regardless of its not having been notified to the MS.<sup>1336</sup> In its analysis, the Court also took account of the concrete circumstances of the case, namely that the Commission published its Communication after a directive on the same matter was withdrawn. A comparison of these two texts 'discloses parallels, in particular as regards the definitions, scope and content of those texts'.<sup>1337</sup> This 'indicates that the Commission was seeking, by means of the Communication, to secure the application of rules identical or similar to those contained in the proposal for a directive'.<sup>1338</sup>

With regard to a Commission act referred to as 'Code of Conduct', the Court held that it 'constitutes a measure intended to have legal effects of its own, distinct from those created by [the Regulation it is supposed to concretise], and that it is therefore a measure against which an action for annulment may be brought'.<sup>1339</sup>

Another case refers to Commission guidelines concerning State aid in the fisheries sector. The Court qualified these guidelines as 'one element of [the] obligation of regular, periodic cooperation [in the constant review of existing State aid by the Commission1340] from which neither the Com‐ mission nor a Member State can release itself '.1341 To justify this mutual bindingness, the Court in particular brings forward that the MS at issue,

joined cases 8–11/66 *Cimenteries*, 91; see also Scholz, Integration 53, with further references.

<sup>1336</sup> Case C-57/95 *France v Commission*, paras 18 and 23; for an in-depth consideration of the wording of an act when examining its legal (non-)bindingness see case T-721/14 *Belgium v Commission*, paras 21–28; for the 'imperative, compelling lan‐ guage' which also true soft law acts may contain see Georgieva, Soft Law 227.

<sup>1337</sup> Case C-57/95 *France v Commission*, para 12.

<sup>1338</sup> Case C-57/95 *France v Commission*, para 12; see Hofmann/Rowe/Türk, Admin‐ istrative Law 552; with regard to legally binding communciations see also case C-325/91 *France v Commission*.

<sup>1339</sup> Case C-303/90 *France v Commission*, para 25; on this case see also Pampel, Rechtsnatur 119; see also von Bogdandy/Arndt/Bast, Instruments 96, who said: 'The borderline to an unlawful creation of new types of acts is crossed if the legal effects of an act become obscured'.

<sup>1340</sup> This must be considered a specific framework of cooperation. The mere existence of the principle of sincere cooperation pursuant to Article 4 para 3 TEU would not be sufficient to trigger a binding effect; see also Senden, Soft Law 465. For an(other) increased form of cooperation apparently not reaching the required intensity see case C-428/14 *DHL*, para 43.

<sup>1341</sup> Case C-311/94 *IJssel-Vliet*, para 37; see also case C-288/96 *Germany v Commission*, para 65; case C-313/90 *CIRFS*, paras 44 f; on these three cases see Láncos, Core 770–774.

the Netherlands, has confirmed that State aid granted to the Dutch fisheries sector is in conformity with said guidelines.<sup>1342</sup> From this the Court dedu‐ ces an approval of the guidelines on the part of the Netherlands. The Court thereby elevates the guidelines to an administrative agreement between the Commission and the Netherlands, which makes these rules different from regular soft law.<sup>1343</sup>

# 4.2.2.2. The actual or potential legal reasons for these effects

4.2.2.2.1. Administrative cooperation according to Article 197 TFEU

The unprecedented<sup>1344</sup> provision of Article 197 para 1 TFEU stipulates that '[e]ffective implementation of Union law by the Member States […] shall be regarded as a matter of common interest'; its para 2 addresses the possibility of Union support for MS' efforts to improve their administrative capacity to implement Union law.1345 Since para 3 provides that Article 197 shall be 'without prejudice to the obligations of the Member States to implement Union law', it is not clear whether a legal reason to comply with EU soft law can be deduced from para 1. It could be argued that there is no obligation of MS to comply with EU soft law anyway, hence no such obligation can be touched upon in the first place.1346 However,

<sup>1342</sup> See also Hofmann/Rowe/Türk, Administrative Law 574. This element of MS con‐ sent is also referred to – *ex negativo* – in joined cases T-132/96 and T-143/96 *Sachsen/Volkswagen*, para 209, with a further reference; for lack of MS consent, the 'appropriate measures' proposed by the Commission according to what is now Article 108 para 1 TFEU did not entail binding effect in case T-330/94 *Salt Union*, para 35. The cooperation element may not be downplayed – mere acceptance by a MS does not appear to be sufficient; see Eliantonio, Soft Law 506 f.

<sup>1343</sup> Similarly in this respect: case C-313/90 *CIRFS*, paras 35 f; see Georgieva, Soft Law 236 f.

<sup>1344</sup> See Chiti, Governance 53 f.

<sup>1345</sup> See Schütze, Rome 1407 f, with regard to Article 197 para 2 TFEU and its (poten‐ tial) consequences for the procedural autonomy of the MS; for the meaning of this provision with regard to 'horizontal soft law' see Lafarge, Coopération 80. For lacking capacity as a reason for MS' violations of EU law see Tallberg, Paths 630 ff; for the ambivalent role of capacity in the context of compliance see also Börzel, Noncompliance 41–43.

<sup>1346</sup> The national law of a MS may provide for a binding effect of certain EU soft law. The national authorities are then legally bound by it pursuant to national, not EU law; see von Graevenitz, Mitteilungen 173, with further references.

we may still ask: Does Article 197 TFEU in any way increase the *stimulus* to comply with, or at least to consider, EU soft law? Does it increase the authority of EU soft law? First, we need to find out what all the phrase 'effective implementation of Union law' referred to in para 1 encompasses. 'Effective' has a lot of different meanings in the context of EU law.<sup>1347</sup> In the given context, it is of only subordinate importance and therefore requires no further analysis. More important appears the term 'implementation'. In EU law,<sup>1348</sup> the term 'implementation' necessarily implies compliance – 'compliant implementation', strictly speaking, constitutes a pleonasm.<sup>1349</sup> 'Implementation' is also used in the context of EU soft law,<sup>1350</sup> and since the term 'Union law', used in an unspecified way, appears to encompass also Union soft law,<sup>1351</sup> the better reasons here speak in favour of a subsumption of EU soft law under the phrase 'effective implementation of Union law'.<sup>1352</sup> This means that EU soft law is encompassed by Article 197 para 1 TFEU.

In a next step it is to be clarified whether from the order that the '[e]ffective implementation of Union law by the [MS] […] shall be regarded as a matter of common interest' a 'legal plea' to the MS for consideration of or compliance with EU soft law can be deduced. It is argued by some

1352 See, in contrast, Article 291 para 1 TFEU, according to which MS 'shall adopt all measures of national law necessary to implement *legally binding* Union acts' (emphasis added); see also case C-360/09 *Pfleiderer*, para 24.

<sup>1347</sup> See Ohler, Art. 197 AEUV, para 2.

<sup>1348</sup> In the parlance of international law the terminology may be more differentiated: Haas, Hypotheses 21.

<sup>1349</sup> Nevertheless, the *oxymora* 'incorrect implementation' or 'wrongful implementa‐ tion' are common in so-called EU speak; see eg Commission, 'Monitoring the application of European Union Law. 2015 Annual Report', COM(2016) 463 final, 12; see also Andersen, Enforcement 113.

<sup>1350</sup> See eg Commission, 'Cultural Heritage. Digitalisation, online accessibility and dig‐ ital preservation. Report on the Implementation of Commission Recommendation 2011/711/EU 2013–2015' (2016). For the wide interpretation of the term 'implemen‐ tation' see Budischowsky, Art. 197 AEUV, para 16, with references to the case law of the CJEU; Ohler, Art. 197 AEUV, para 4; Senden, Soft Law 322, with further references; see also case 16/88 *Commission v Council*, para 11.

<sup>1351</sup> This is suggested by the facts that, first, EU primary law does not know the term 'soft law' and, second, Article 288 TFEU, also listing recommendations and opinions, is entitled 'The legal acts of the Union'. Also the judges of the CJEU do not seem to use the term 'soft law'; see Láncos, Core 759; Andone/Coman-Kund, EU soft law 9, who refer to the exceptional use of this term in legal acts, eg selected inter-institutional agreements; for the AG see references in Christianos, Effectiveness 327; more recently: Opinion of AG *Bobek* in case C-911/19 *FBF*, eg para 54.

that para 1 is of a merely programmatic nature<sup>1353</sup> which would exclude a concrete command. Others interpret the meaning of 'common interest' as expressing a (low) level of integration and addressing a duty to avoid, as far as possible, actions which have a negative effect on other MS.<sup>1354</sup> Failure to consider, or to comply with respectively, EU soft law can have a negative effect on other MS, in particular where, in case of an act applicable to all MS, the other MS comply with it.

In my view, the generality of Article 197 para 1 TFEU renders it a provi‐ sion with no concrete regulatory content.<sup>1355</sup> It is not entirely declaratory in nature, as it stipulates that the effective implementation of EU law by the MS 'shall be regarded as a matter of common interest'. From this command, however, it is difficult to deduce concrete obligations. A mere duty to *consider* EU soft law may possibly be read into the elevation of EU law implementation to a 'matter of common interest'. As was stated above, this argument cannot be countered by para 3, according to which '[t]his Article shall be without prejudice to the obligations of the Member States to implement Union law', because there is no general obligation of MS to implement (ie to comply with) EU soft law. However, as we shall see, there are Treaty provisions more strongly hinting at a MS' duty to consider EU soft law. Against this background, when arguing in favour of a MS' duty to consider EU soft law (or even duties going beyond that), more pertinent provisions or principles of EU law should be invoked (see below). Thus, in the given context Article 197 TFEU has little more than cosmetic effect.

### 4.2.2.2.2. Sincere cooperation ('loyalty') according to Article 4 para 3 TEU

### 4.2.2.2.2.1. Overview

Article 4 para 3 TEU lays down the principle of sincere cooperation, ac‐ cording to which 'the Union and the Member States shall, in full mutual re‐ spect, assist each other in carrying out tasks which flow from the Treaties'.

<sup>1353</sup> See Frenz, Verwaltungskooperation; Ohler, Art. 197 AEUV, para 1; Terhechte, Art. 197 AEUV, para 3; Weerth, Art. 197 AEUV, para 6.

<sup>1354</sup> See M Schröder, Vollzug 672 f, describing the character of Article 197 para 1 TFEU as 'imperativ' [imperative]; Vedder, Art. 197 AEUV, para 4, both with further references.

<sup>1355</sup> See Klamert, Loyalty 31, who argues that the regulatory content of Article 197 para 1 TFEU already follows from Article 4 para 3 TEU.

The provision furthermore stipulates that MS shall take 'any appropriate measure, general or particular, to ensure fulfilment of the obligations aris‐ ing out of the Treaties or resulting from the acts of the institutions' and that they shall 'facilitate the achievement of the Union's tasks and refrain from any measure which could jeopardise the attainment of the Union's objectives'.

The duty of loyal cooperation of the MS with the Union has been con‐ tained in the Treaties ever since the establishment of the EEC<sup>1356</sup> and the Court has stressed its 'particular importance vis-à-vis the judicial authori‐ ties of the Member States, who are responsible for ensuring that Communi‐ ty law is applied and respected in the national legal system'.<sup>1357</sup> With regard to the pertinent Article 5 TEEC which expressly enshrined obligations only of the MS,<sup>1358</sup> the Court held 'that there are legal obligations imposed on Member States which do not result from any specific action by the Council or the Commission, but which arise from their general obligation to act in a way consistent with the objectives and spirit of the Treaty. A resolution of the Council, adopted on the proposal of the Commission on an important issue of considerable difficulty for the Community, thus might[1359] give rise

<sup>1356</sup> See Article 5 TEEC. For the early understanding of the principle of sincere cooper‐ ation in general as a mere 'statement of principle and political intent' see Klamert, Loyalty 234; comparing loyalty in EU law with the principle of good faith in public international law: Georgieva, Soft Law 245; for the role of good faith in public international law as a *stimulus* to consider legally non-binding acts see Miehsler, Autorität 40, with further references; for the practical experiences before national courts (Hungarian courts) see Kovács/Tóth/Forgác, Effects.

<sup>1357</sup> Case C-2/88 Imm. *Zwartveld*, para 18. For the importance of the loyalty principle both for administrative authorities and courts see Pernice, Constitutionalism 724 f. For the relationship with the principle of substantive coherence of EU policies (enshrined in Article 7 TFEU) see Braams, Koordinierung 206 f.

<sup>1358</sup> For the recognition of a general principle of E(E)C/EU law of mutual sincere cooperation by the Court see eg case 230/81 *Luxembourg v European Parliament*, para 37; case C-2/88 Imm. *Zwartveld*, para 10; joined cases C-36/97 and C-37/97 *Kellinghusen*, para 30. This principle also includes the duty of sincere cooperation of the institutions *vis-à-vis* the MS, which mutuality is now codified in Article 4 para 3 TEU, and between MS (often referred to as 'solidarity'). In the given con‐ text, however, only the MS' obligations *vis-à-vis* the institutions shall be addressed; for the difficulty (and the possible dispensability) of clearly separating loyalty and solidarity see Isak, Loyalität 309; see, however, case C‑848/19P *Germany v Poland*, paras 37 ff, in which the Court fleshes out, in the context of energy policy, solidarity as a 'fundamental principle of EU law'.

<sup>1359</sup> It is clearly indicated that the Court does not intend to generally elevate Coun‐ cil resolutions to legally binding acts, but that it pays tribute to the specific cir‐

to legal obligations'.1360 More recently, the principle of sincere cooperation was interpreted as also obliging the MS to 'ensure, in their respective territories, the application of and respect for EU law'.<sup>1361</sup> That the duty of cooperation 'has no binding legal effect on the national legal [ie judicial] authorities', as the then Court of First Instance held in the *Tillack* case,<sup>1362</sup> cannot be confirmed in this generality.<sup>1363</sup>

The second sentence of Article 4 para 3 TEU expresses the MS' duty to 'take any appropriate measure […] to ensure fulfilment of the obligations arising out of the Treaties or resulting from the acts of the institutions'. This duty only refers to the 'obligations' arising out of the Treaties or resulting from the acts of the institutions and hence does not appear to include legally non-binding EU soft law.<sup>1364</sup>

The third sentence of Article 4 para 3 TEU expresses the MS' duty to 'facilitate the achievement of the Union's tasks and [to] refrain from any measure which could jeopardise the attainment of the Union's objectives'. Its first alternative – to facilitate the achievement of the Union's tasks – cannot be understood as an obligation to comply with EU soft law. In the past, specific duties to inform – not a general duty to provide information – have been deduced from this provision.1365 There is an exemplary evidence

cumstances of the case (*argumentum* 'might'); in a related case, the Court stressed that the resolution was eventually 'formally approved' by the MS some days after its adoption; case 61/77 *Commission v Ireland*, 420.

<sup>1360</sup> Case 141/78 *France v United Kingdom*, 2933. For an example of soft law allegedly concretising the MS' duties under Article 4 para 3 TEU and partly entailing legal bindingness see Müller-Graff, Art. 60 AEUV, paras 2 f, with a further reference.

<sup>1361</sup> Opinion 2/13 *ECHR II*, para 173.

<sup>1362</sup> Case T-193/04 *Tillack*, para 49; see H Hofmann, Decision making procedures 155– 157.

<sup>1363</sup> See Schill/Krenn, Art. 4 EUV, paras 88 ff. With regard to the compliance with agreements concluded by international organisations on the part of their respec‐ tive member states which – for lack of an according provision (eg Article 216 para 2 TFEU) – are not binding on them, *Schermers* and *Blokker* said: '[T]he obligation of loyalty to the organization offers another ground for accepting such a provision, even where not expressly incorporated in the constitution'; Schermers/Blokker, Institutional Law, § 1787.

<sup>1364</sup> When the violation of concrete obligations is at issue, the Court – in the context of a Treaty infringement procedure – sees no point in additionally scrutinising a violation of the principle of sincere cooperation; case C-392/02 *Commission v Denmark*, para 69; differently: Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 102 f.

<sup>1365</sup> See Eekhoff, Verbundaufsicht 51; Obwexer, Art. 4 EUV, para 130, with further references to the CJEU's case law; see also Article 337 TFEU which, according

that the legislator now explicitly invokes Article 4 para 3 TEU in cases in which it has earlier remained silent in that respect.<sup>1366</sup>

The second alternative prohibits measures which could jeopardise the attainment of the Union's objectives, irrespective of whether they go against (other) obligations emanating from EU law. Hence the wording of this pro‐ vision would encompass a case in which a MS does not apply EU soft law, thereby jeopardising the attainment of the Union's objectives.<sup>1367</sup> Applying a wide interpretation of the term 'measure', this does not only encompass measures but also omissions (instances of failure to act) in contradiction with EU soft law. The EU undertakes to reach its objectives *also* by means of soft law. Cases in which the mere non-application of soft law (which does not – at the same time – result in the violation of a legally binding provision of EU law) could jeopardise the attainment of these objectives are exceptional (see also 4.2.2.2.2.2. below).

### 4.2.2.2.2.2. The consequences for soft law

From the many concrete duties of the MS the Court has deduced from the principle of sincere cooperation, only the duty to apply national law in conformity with Union law1368 appears to be promising when searching for a duty of or, at least, a 'legal *stimulus*' for MS to apply EU soft law. In view of the Court's case law, which essentially is located in the context of legally binding acts,1369 its analogous application to EU soft law in general appears to be questionable, though (see in more detail 4.2.2.2.3. below).1370 What

to the literature, does not itself entitle the Commission, but which is subject to concretisation through secondary law; see Jaeckel, Art. 337 AEUV, paras 1f; Ladenburger, Art. 337 AEUV, para 2; Wegener, Art. 337 AEUV, para 1, each with further references.

<sup>1366</sup> See eg Recital 5 of Directive 2015/1535, with regard to the MS' duty to notify the Commission of certain projects, as opposed to the respective Recital of the predecessor directive (Recital 5 of Directive 98/34/EC).

<sup>1367</sup> See case T-113/89 *Nefarma*, para 79. For the malleability of these objectives, which should, according to *Klamert*, only entail binding effects for the MS where they have been concretised by a Union act see Klamert, Loyalty 291.

<sup>1368</sup> See case C-306/12 *Welter*, para 30.

<sup>1369</sup> See Schill/Krenn, Art. 4 EUV, paras 67 ff, with references in particular to the Court's case law.

<sup>1370</sup> The Court has invoked the principle of sincere cooperation in the context of the question whether national courts have to accept the findings of an only provision‐ al Commission decision; case C-574/14 *PGE Górnictwo*, paras 33 and 41. With

we may deduce from this duty, however, is an increased authority (below the level of legal bindingness) of EU soft law.<sup>1371</sup> A MS is free not to apply it, but its application is still the expected case, hence a duty to justify, to give the reasons for deviation that is, could be read into the principle of sincere cooperation.<sup>1372</sup> It could also be referred to as an extended duty to inform – a duty which, admittedly in other contexts, has been read into the principle of (then) 'close cooperation' by the Court.<sup>1373</sup> It has stressed this duty to inform specifically in order 'to facilitate the achievement of the Commission's task of ensuring compliance with the Treaty'.<sup>1374</sup>

regard to 'regular' law, however, the Court held that given the infringement of an obligation laid down in secondary law 'no purpose [is] served by considering the question whether it had thereby also failed to fulfil its obligations [according to the principle of sincere cooperation]': case C-374/89 *Commission v Belgium*, para 13, with reference to case C-48/89 *Commission v Italy*; similarly with regard to a primary law *lex specialis*: case C-18/93 *Corsica Ferries*, para 18; in other cases the Court is less clear in this respect; see Schill/Krenn, Art. 4 EUV, para 74.

<sup>1371</sup> See also Brohm, Mitteilungen 111 and 178-183; Eekhoff, Verbundaufsicht 240, with regard to a duty to react to non-binding requests for information, inspection or redress; Hofmann/Rowe/Türk, Administrative Law 575; Pampel, Rechtsnatur 98 f, with further references; Raschauer, Leitlinien 37; Thürer, Role 134, who, as early as in 1990, argued that 'the principle of community loyalty gives rise to certain legal obligations, such as the duty to consider and make an effort to comply with soft law and not to act against it unless good reasons for doing so are set out'; van Rijsbergen, Legitimacy 61f. For the link between a duty to consider and a factual duty to comply see Arndt, Sinn 165.

<sup>1372</sup> See Schaller, Intensivierung 425; Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 100 f; see also H Adam, Mitteilungen 83 (with regard to recom‐ mendations and opinions): duty 'ernsthaft [zu] prüfen' [to seriously consider] and to refuse compliance only 'mit ausreichender Erklärung' [with sufficient reason‐ ing]; Everling, Wirkung 151, stressing the need for 'wichtige Gründe' [important reasons] for deviating behaviour; see also case C‑28/15 *Koninklijke*, para 52, on the requirements of judicial review by a national court in this context; Recital 16 of Commission Proposal for a Regulation of the European Parliament and of the Council establishing a framework for screening of foreign direct investments into the European Union, COM(2017) 487 final (which was not incorporated in Regulation 2019/452, which was eventually adopted on the basis of this proposal): 'The Member States should take utmost account of the opinion and provide an explanation to the Commission if they do not follow this opinion, in compliance with their duty of sincere cooperation under Article 4(3) TEU'.

<sup>1373</sup> See case 804/79 *Commission v United Kingdom*, 1062; case C-285/96 *Commission v Italy*, paras 19 f; case C-459/03 *Commission v Ireland*, para 179. Arguing in favour of a generalisation of this duty where the interests of the Union are at risk: Klamert, Loyalty 235 f; differently: Schill/Krenn, Art. 4 EUV, para 68; for exceptions to a duty to provide information see eg Opinion of AG *Kokott* in case C-550/07P *Akzo Nobel*, para 137.

Whereas such a duty to give the reasons for non-compliance with a soft law act, which in a number of cases is explicitly provided for in secondary law,<sup>1375</sup> so far has not been confirmed by the Court as a general obligation, it has, repeatedly and for a long time now, held that (certain) EU soft law acts are to be taken 'into consideration'.<sup>1376</sup> In this context, *Senden* referred to an 'obligation of effort, as opposed to an obligation of result'.<sup>1377</sup> AG *Kokott* has explicitly based her argument that 'national authorities and courts must take due account of the Commission's competition policy notices', EU soft law that is, on the principle of sincere cooperation.<sup>1378</sup> The Court appears to have fleshed out the ensuing duties – or at least to have attempted to do so – by emphasising the necessity 'to weigh on a case-by-case basis' the various Union policy interests at stake when deciding on whether or not to follow EU soft law.<sup>1379</sup>

In view of the dynamic approach the Court has been taking on close/ sincere cooperation,1380 it does not seem to be far-fetched that the Court may in the future come to confirm a duty to provide the reasons for non-compliance with EU soft law on the basis of Article 4 para 3 TEU.<sup>1381</sup> After all, also the case law on the pre-effect of directives imposes duties on the MS prior to the expiry of the implementation period, ie at a time when non-compliance with the directive is, according to Union law, in principle still lawful.1382 The Court in this context held that 'according to

<sup>1374</sup> Case C‑372/05 *Commission v Germany*, para 76.

<sup>1375</sup> Eg Article 16 para 3 of Regulation 1094/2010.

<sup>1376</sup> See again case C‑322/88 *Grimaldi*, paras 18 f; for the meaning of these words see also Láncos, Facets 88 f.

<sup>1377</sup> Senden, Soft Law 350; for different readings of this MS' obligation see references in Korkea-aho, Courts 477.

<sup>1378</sup> Opinion of AG *Kokott* in case C-226/11 *Expedia*, para 38; see also Opinion of AG *Wahl* in case C-526/14 *Kotnik*, para 38.

<sup>1379</sup> Case C-360/09 *Pfleiderer*, para 31; see also Polley, Access 453, pointing at the limited competence of the EU in procedural matters which may have lead to the Court's reserved approach. Duties of information eg on a change of administrative practice – which could be relevant also in the case of non-application of soft law – are also known *between MS*; see case 42/82 *Commission v France*, para 36; see also Schill/Krenn, Art. 4 EUV, paras 100 f.

<sup>1380</sup> For the changing understanding of 'loyalty' as expressed in literature and case law see Brohm, Mitteilungen 112.

<sup>1381</sup> See A Geiger, Leitlinien; Nettesheim, Art. 292 AEUV, para 4; doubtful, but open: Georgieva, Soft Law 239 and 245.

<sup>1382</sup> See case C-129/96 *Inter-Environnement Wallonie*, para 45; see also Klamert, Loyal‐ ty 107 f.

the case-law of the Court, although the Member States are not obliged to adopt measures to transpose a directive before the end of the period prescribed for transposition, it follows from the second paragraph of Article 10 EC [now Article 4 para 3 subpara 3 TEU] in conjunction with the third paragraph of Article 249 EC [now Article 288 para 3 TFEU] and from that directive itself that during that period they must refrain from taking any measures liable seriously to compromise the result prescribed by that directive'.<sup>1383</sup> The strong effects the Court has deduced from Article 4 para 3 TEU in this context – the pre-effect of directives – may serve to underpin the approach suggested here in the context of soft law.

In the field of cartel law it is argued by some that national authorities and courts shall be bound by the Commission's soft law output.<sup>1384</sup> How‐ ever, only exceptionally, where otherwise the attainment of the Union's objectives would be jeopardised, this argument is plausible. The Court in two cases concerning common fisheries policy has affirmed the possibility of implying such duty to the principle of sincere cooperation.1385 In the case 804/79 *Commission v UK* it had to deal with a situation in which the Com‐ mission has proposed conservation measures in the field of fisheries policy to the Council, which the Council itself could not agree upon in due time. Instead, the Council took over part of the Commission proposal in the form of 'guidelines' which, according to the Court, reflected the 'Council's intention to reinforce the authority of the Commission's proposals and, on the other hand, its intention to prevent the conservation measures in force from being amended by the Member States without any acknowledged need'.1386 In essence, the Court declared these guidelines binding to the effect that their breach would constitute a violation of the principle of sincere cooperation. It held that Article 5 TEEC, a predecessor provision

<sup>1383</sup> Case C-422/05 *Commission v Belgium*, para 62, with references to the Court's further case law. For the even more far-reaching duty of a MS to approximate its legislation during an (exceptionally) extended implementation period see case C-144/04 *Mangold*, para 72. This duty was affirmed without express reference to the principle of sincere cooperation, though.

<sup>1384</sup> See eg Schweda, Bindungswirkung 1141f, thereby conceding that the Commission and the German Federal Ministry for Economic Affairs do not share this view; arguing against *Schweda's* position: Pohlmann, Bindungswirkung.

<sup>1385</sup> Case 32/79 *Commission v United Kingdom*, 2420 f and 2427 (with regard to a recommendation belonging to public international law); case 804/79 *Commission v United Kingdom*, para 28.

<sup>1386</sup> Case 804/79 *Commission v United Kingdom*, para 25; for the facts of the case see in particular its paras 19–28.

of Article 4 para 3 TFEU, 'imposes on Member States special duties of action and abstention in a situation in which the Commission, in order to meet urgent needs of conservation, has submitted to the Council proposals which, although they have not been adopted by the Council, represent the point of departure for concerted Community action'.<sup>1387</sup> Against this background, *Senden* argued that a duty of the MS and their respective authorities to comply with EU soft law 'only exists when the soft law act at issue can be considered a specific expression of that principle; that is to say, when the soft law act establishes what the duty of cooperation actually entails as regards the matter in question'.<sup>1388</sup> In *Senden's* view, such specific expressions of cooperation can be found in the field of State aid law (soft law acts taken on the basis of what is now Article 108 para 1 TFEU<sup>1389</sup>) and competition law (particularly in the framework of what is now Article 101 TFEU1390), even though she stresses that 'institutional clarification' to that end would be desirable 'for the sake of legal certainty'.<sup>1391</sup>

This assessment was confirmed by the Court: With regard to Chapter IV of Regulation 1/2003, it held '[t]hat cooperation is part of the general prin‐ ciple of sincere cooperation, referred to in Article 10 EC, which governs the relationships between the Member States and the Community institutions. As the Court has held, the duty of sincere cooperation imposed on the Community institutions is of particular importance where that cooperation involves the judicial authorities of a Member State who are responsible for ensuring that Community law is applied and respected in the national legal system'.1392 Whether this boils down to the bindingness of or at least a duty of the MS to consider the respective soft law acts, is left open by the court in this case.

1391 Senden, Soft Law 446.

<sup>1387</sup> Case 804/79 *Commission v United Kingdom*, para 28. For the exceptional role of this and related judgements see Everling, Wirkung 151f; Riedel, Gemeinschaftszu‐ lassung 127, with further references; von Bogdandy/Arndt/Bast, Instruments 116.

<sup>1388</sup> Senden, Soft Law 356 and 443 f; see also ibid 356 f, with regard to a potential duty to transpose soft law into national law.

<sup>1389</sup> See H Adam, Mitteilungen 107–113.

<sup>1390</sup> With regard to the authorisations of national authorities by the Commission according to Article 105 para 2 TFEU, and the alleged legal bindingness of this authorisation in light of the MS' loyalty obligations, see Ludwigs, Art. 105 AEUV, para 10.

<sup>1392</sup> Case C-429/07 *Inspecteur van de Belastingdienst*, para 21; reiterated in case C-428/14 *DHL*, para 30, with references to further case law.

In a later case the Court – with a view to legally non-binding1393 state‐ ments of position made by the Commission in the context of the execution of State aid-related Commission Decision 2007/374/EC – held that 'it must be borne in mind that application of the European Union competition rules is based on an obligation of cooperation in good faith between the national courts, on the one hand, and the Commission and the European Union Courts, on the other […]'. In this context, 'national courts must take all the necessary measures, whether general or specific, to ensure fulfilment of the obligations under European Union law and refrain from those which may jeopardise the attainment of the objectives of the Treaty, as follows from Article 4(3) TEU'.<sup>1394</sup> Where a national court has 'doubts or […] difficulties' as regards the application of pertinent EU law, they 'must take [the Commission statements of position] into account as a factor in the assessment of the dispute before it and must state reasons having regard to all the documents in the file submitted to it'.1395 While here the Court (once again) affirms a duty to consider on the part of the national courts, while from the Court's wording its eagerness to stress that the national court really *should follow* the Commission statements of position becomes clear, and while the Court seems to request the national court to include the Commission statements in its reasoning, at the same time it leaves no doubt as to their legal non-bindingness – also in the framework of competition (including State aid) law, a policy field which is determined by strong cooperation. As AG *Wathelet* said with reference to this judgement: Having to '"[…] take into account the guidance provided by the Council" (which the Commission has acknowledged in this case[]) is not the same thing as "being legally bound to follow it"'.<sup>1396</sup>

As mentioned above, cases of an increased authority of EU soft law, beyond a mere duty to consider, remain exceptional. They should not be made subject to lightheaded analogous application. In his Opinion in the *Kotnik* case, AG *Wahl* correctly pointed out: 'Although the Court has held that the provisions of such acts of "soft law" are, by virtue of the duty of sincere cooperation enshrined in Article 4(3) TEU, to be taken into due account by the Member States' authorities,[] that duty cannot be understood as making those rules binding – not even *de facto* – on pain of

<sup>1393</sup> See case C-69/13 *Mediaset*, para 27.

<sup>1394</sup> Case C-69/13 *Mediaset*, para 29.

<sup>1395</sup> Case C-69/13 *Mediaset*, para 31.

<sup>1396</sup> Opinion of AG *Wathelet* in case C-425/13 *Commission v Council*, para 181.

eluding the legislative procedure set out in the FEU Treaty' (emphasis in original).<sup>1397</sup> What is more, due to the mutuality of this principle, it could be argued that sincere cooperation prevents EU institutions from insisting on MS' compliance with EU soft law, and to rely on the MS' assessment (ie the decision not to comply, which may follow from the MS' consideration of the respective EU soft law act).<sup>1398</sup>

# 4.2.2.2.3. The principle of interpretation of national law in line with Union law

The requirement that national law is to be interpreted in line with Union law is another rule from which a duty to consider (or another expression of an increased authority of ) EU soft law may be deduced.1399 The Court held that this rule is 'inherent in the system of the Treaty, since it permits the national court, for the matters within its jurisdiction, to ensure the full effectiveness of Community law when it determines the dispute before it'.1400 Sometimes it is also explicitly traced to the principle of sincere coop‐ eration.1401 Thus, it can also be understood as an add-on to the elaborations under 4.2.2.2.2.2. above. The relevant case law predominantly refers to the interpretation of national law 'in the light of the wording and the purpose of [directives]'.1402 An analogous application to the case of EU soft law,

<sup>1397</sup> Opinion of AG *Wahl* in case C-526/14 *Kotnik*, para 38; see also case T-721/14 *Belgium v Commission*, paras 43 ff; Selmayr, Art. 282 AEUV, para 129, with regard to the ECB's opinions.

<sup>1398</sup> See eg joined cases T‑208/11 and T‑508/11 *LTTE*, para 97, in which the General Court stresses the mutual trust between EU institutions and MS resulting from the principle of sincere cooperation.

<sup>1399</sup> The practically most frequent case is that EU soft law is concretising EU law. This may also reflect on national law where the latter is implementing EU law; for the reverse case – the interpretation of EU soft law in line with EU law – see case C-410/13 *Baltlanta*, para 65.

<sup>1400</sup> Joined cases C-397–403/01 *Pfeiffer*, para 114.

<sup>1401</sup> See Calliess/Kahl, Art. 4 EUV, para 98; Hatje, Art. 4 EUV, para 52, with reference to the Court's case law.

<sup>1402</sup> Case 14/83 *von Colson*, para 26; drawing a line to *Grimaldi*: Arnull, Status; see also case C-106/89 *Marleasing*, para 8. A similar rule applies with regard to the relationship between public international law and EU legislation: 'Community legislation must, so far as possible, be interpreted in a manner that is consistent with international law, in particular where its provisions are intended specifically to give effect to an international agreement concluded by the Community'; case

whereby a duty to interpret national law in the light of the wording and purpose of EU soft law may be construed, is to be refused.<sup>1403</sup> The said obligation of MS results from the legal bindingness upon MS of directives or EU law more generally. Such a legal bindingness in the context of EU soft law is, by definition, excluded.<sup>1404</sup> Therefore no general duty of an interpre‐ tation of national law in accordance with EU soft law may be inferred.<sup>1405</sup>

However, from this principle something less, namely a duty of considera‐ tion of EU soft law,<sup>1406</sup> may be construed.<sup>1407</sup> The Court has, in a number of cases, affirmed that 'national courts are bound to take the recommenda‐ tions into consideration in order to decide disputes submitted to them, in particular where they cast light on the interpretation of national measures adopted in order to implement them or where they are designed to supple‐ ment binding Community provisions' (see 4.2.2.1.2. above).<sup>1408</sup>


C-341/95 *Bettati*, para 20; see also Nollkaemper, Role 183–185, for the interpreta‐ tion of national law in conformity with public international law.

# 4.2.2.2.4. Legal certainty, legitimate expectations, equality and effectiveness

The principle of legitimate expectations is strongly connected to the notion of legal certainty which both are 'part of the legal order of the Commu‐ nity'.<sup>1409</sup> It even appears that the protection of legitimate expectations is one prong of the principle of legal certainty.<sup>1410</sup> In that context, very gen‐ erally the Court held that 'Community legislation must be clear and its application foreseeable for all interested parties'.<sup>1411</sup> The latter requirement ('calculability') cannot be applied (*per analogiam*) to EU soft law, even though legal certainty is considered a 'fundamental principle' of Union law.<sup>1412</sup>

It may be argued that – being an 'official act' of an EU body – an EU soft law act evokes a legitimate expectation of those concerned (eg individuals or undertakings) that it is complied with by its addressees.<sup>1413</sup> Thereby – indirectly – the addressees of EU soft law, which often are MS authorities, would be obliged to conform with EU soft law *vis-à-vis* individuals/undertakings.1414 However, it is to be emphasised that only 'pre‐ cise, unconditional and consistent assurances, originating from authorised,

<sup>1409</sup> Case 205/82 *Milchkontor*, para 30; see also case 169/80 *Administration des douanes*, para 17; for the interwovenness of the two principles see Georgieva, Soft Law 252, with further references, also to the CJEU's case law; proposing an application in conjunction with Article 4 para 3 TEU in the context of EU soft law: ibid 255; for the principle of legitimate expectations in EU law more generally see Craig/de Búrca, EU Law 593–597, with reference to the pertinent case law; Sharpston, Expectations.

<sup>1410</sup> See also Wörner, Verhaltenssteuerungsformen 189.

<sup>1411</sup> Case C-325/91 *France v Commission*, para 26; see also case C-483/99 *Commission v France*, para 50; case T‑358/11 *Italy v Commission*, para 123.

<sup>1412</sup> Case C-143/93 *van Es*, para 27; case C-110/03 *Belgium v Commission*, para 30. For the nevertheless existing relationship between this principle and soft law see H Hofmann, Rule-Making 165.

<sup>1413</sup> For the wide-spread presumption that compliance with a soft law act constitutes a 'safe harbour', constituting compliance also with (hard) law see eg Dickschen, Empfehlungen 146–148; T Möllers, Standards 159 and 162; Wörner, Verhaltens‐ steuerungsformen 188–190; with regard to a similar presumption in the context of European standards: Griller, Normung 29; with regard to public international soft law see Brown Weiss, Introduction 4; Friedrich, Soft law 181 and 185; Miehsler, Autorität 40, with further references.

<sup>1414</sup> See Opinion of AG *Mazák* in case C-360/09 *Pfleiderer*, para 45; this view was not supported by the ensuing judgement of the Court; see Láncos, Core 768; critically with regard to the EU body's lack of responsibility *vis-à-vis* the individual/under‐ taking: Kühling, Telekommunikationsrecht, para 73; Schaller, Intensivierung 427 f.

reliable sources'1415 provided for in an EU soft law act in accordance with the applicable rules<sup>1416</sup> may create 'justified hopes',<sup>1417</sup> and that – on the part of those potentially protected – the threshold applied is that of a 'prudent and discriminating' actor.<sup>1418</sup> Such an actor can regularly be expected to recognise EU soft law as a legally non-binding act, which is why it cannot take compliance for granted,<sup>1419</sup> ie it cannot legitimately expect compliance by others.<sup>1420</sup> There is no 'soft law certainty', we could say.

There is one important exception to this unpredictability of application. Soft law in principle is binding upon its creators. Thus, their respective


<sup>1415</sup> Case C-526/14 *Kotnik*, para 62; see also case T-72/99 *Meyer*, para 53, both with further references. While the Court in the *Kotnik* case is dealing with a soft law act, a Commission Communication on State aid for banks, the question in this case was not whether such an act may produce legitimate expectations, but whether its content infringes upon legitimate expectations. The Court did, as far as the question of legitimate expectations was concerned, not dwell on the legal non-bindingness of the act; AG *Wahl* did: Opinion in case C-526/14 *Kotnik*, para 48.

compliance with soft law is (in principle) obligatory and hence can be (reasonably) foreseen. Due to the ever increasing intertwinement of EU and national administration,<sup>1421</sup> it could be argued that individuals/under‐ takings (*vis-à-vis* the MS implementing EU law) may, in specific cases, be entitled to have such expectation also *vis-à-vis* a MS authority.<sup>1422</sup> This expectation would then be caused by an act of EU soft law. That is to say that individuals/undertakings could take for granted that they will not be subject to corrective action by a MS authority where they act in compliance with EU soft law. The bindingness of EU soft law upon its creator would then – via the tool of legitimate expectations – be extended to the MS and their respective authorities. While this does not seem to apply to national courts, for national authorities involved in the adoption of the EU soft law act at issue – eg by taking part in the relevant decision-making procedure of a European agency as the creator of soft law – a 'Kollektivbindung' [collective bindingness] was proposed.1423 In my view, this proposal is to be refused. As expectations in this direction would not normally come up in what was described above as 'prudent and discriminating' actor, they are not worthy of protection.1424 Only where the MS authority at issue has itself taken action justifying certain expectations of individuals/undertakings – eg by continuously complying with a certain EU soft law act – a sudden deviating action by the authority may qualify as *venire contra factum prop‐ rium* and hence violate the principle of legal certainty (in combination with the legitimate expectations; see above).1425 In this case the 'bindingness of

<sup>1421</sup> See W Weiß, Solidarität 424 f, with examples and further references.

<sup>1422</sup> See von Graevenitz, Mitteilungen 173, with reference to the case C-226/11 *Expedia*.

<sup>1423</sup> See references in Dickschen, Empfehlungen 147 f, with regard to national courts, and with regard to the argument of 'collective bindingness'; for the possibility of factual effects in this constellation see 4.3.2.2. below.

<sup>1424</sup> It may appear that the Court applied – in the specific field of competition law – a requirement that 'the administration may not depart in an individual case without giving reasons' not only to the Commission as the creator of the soft law act at issue, but also to the MS authorities (*argumentum* 'administration'); joined cases C-189, 202, 205, 208 and 213/02P *Dansk Rørindustri*, para 209; case T-73/04 *Le Carbone-Lorraine*, para 70. However, such an extended 'self-bindingness' would stretch the effects of soft law too far and also the indications in the above case law are too weak to confirm that assumption; see also Schwarze, Soft Law 243.

<sup>1425</sup> See Beckers, Juridification 576, with references to CJEU case law, and 581 ff, with respect to the Code of Conduct for Business Taxation.

EU soft law' upon the MS authority would follow from its own action, not from EU soft law as such.<sup>1426</sup>

A specific question of soft law to be dealt with in the context of legal cer‐ tainty is retroactivity. It is to be said that soft law may principally exert ret‐ roactive effect under the same conditions as law, namely only exceptionally, when it is ensured that 'the change [ie the adoption of new soft law rules] was reasonably foreseeable at the time when the infringements concerned were committed'.<sup>1427</sup> For example: Persons 'involved in an administrative procedure in which fines may be imposed cannot acquire a legitimate expectation in the fact that the Commission will not exceed the level of fines previously imposed or in a method of calculating the fines'.<sup>1428</sup>

Guarantees of equality – another potential basis for an indirect binding effect of EU soft law on MS authorities (apart from legitimate expectations and/or legal certainty) – may facilitate not only the creation1429 but also the actual application of soft law.1430 EU soft law addressed (also) to the MS limits – on a principally voluntary basis – their respective room for manoeuvre. It is the suggestion of a certain political, legislative, etc action or a certain interpretation of law. In principle, these acts cannot oblige MS, not even under consideration of equality rights.1431 However, where MS have decided to follow the suggested approach once, they may be obliged

<sup>1426</sup> Conditions set by EU law may make it easier to trace a MS authority's behaviour in this respect; see eg Article 16 para 3 of Regulations 1093–1095/2010, providing for a comply or explain mechanism with regard to the ESAs' guidelines and recommendations and demanding the publication of non-compliant authorities, possibly including the reasons they have indicated; see also Simoncini, Regulation 155; differentiated: Tridimas, Indeterminacy 61.

<sup>1427</sup> Joined cases C-189, 202, 205, 208 and 213/02P *Dansk Rørindustri*, paras 223 f; see also case C-63/93 *Duff*, para 20; for the application in the context of soft law see Ştefan, Soft Law 118 and 130.

<sup>1428</sup> Case C-397/03P *Archer Daniels Midland*, para 22. These calculation methods are often specified in soft law acts.

<sup>1429</sup> See case C-443/97 *Spain v Commission*, para 32; see more generally in this context Senden/van den Brink, Checks 32 f.

<sup>1430</sup> With regard to equal treatment see case C-443/97 *Spain v Commission*, para 32; case T-2/90 *Ferreira de Freitas*, para 61, with many further references; case T-214/95 *Vlaamse Gewest*, para 89; see also Opinion of AG *Campos Sánchez-Bor‐ dona* in case C-625/15P *Schniga*, para 67; Senden, Balance 92; references in W Weiß, Leitlinien(un)wesen 258.

<sup>1431</sup> See Senden, Soft Law 442.

to consistently apply it also to similar cases in order to comply with the non-discrimination rules laid down in EU or the respective national law.<sup>1432</sup>

In conclusion, it cannot be excluded entirely that from the principles of legal certainty, of protection of legitimate expectations and of equality obligations of MS to apply EU soft law may arise, in particular where the MS themselves have taken certain action justifying these claims. However, these scenarios are rather exceptional.

Finally, the EU's effectiveness principle is to be examined. However, also from this principle as such no obligation of MS to apply EU soft law can be deduced. The *effet utile* may, however, be brought forward to strengthen an argument that a certain effect of soft law – eg a duty to justify non-compliance with EU soft law – follows from Article 4 para 3 TEU (see 4.2.2.2.2. above).<sup>1433</sup>

With regard to unlawful soft law acts, the above principles, if at all, only apply to a very limited extent. Where the illegality of soft law is blatantly obvious, there is no room eg for a protection of expectations or for equalitybased claims on the part of those concerned.1434 Where the illegality – as later on determined by an authoritative body such as the Court – could not reasonably be foreseen, considering the specificities of the case at issue the trust in the durability of the soft law act may very well be protected.<sup>1435</sup>

4.2.3. Institutions, bodies, offices and agencies of the EU

4.2.3.1. The effects of soft law according to the Court's case law

Having dealt with the effects of EU soft law on MS, we shall now address its effects on the EU institutions, bodies, offices and agencies, as reflected upon by the Court, which shall again be dealt with in the three categories legislative, executive, and judiciary.

<sup>1432</sup> For soft law which content-wise promotes equality see Knauff, Regelungsver‐ bund 479 f, 483, and 549, with regard to a duty to justify.

<sup>1433</sup> See Brohm, Mitteilungen 116; Senden, Soft Law 85 and 442.

<sup>1434</sup> See case C-313/90 *CIRFS*, para 45. Clearly unlawful action in principle may not give rise to a claim for equal treatment; for this principle see eg Reimer, Gleichheit.

<sup>1435</sup> See von Graevenitz, Mitteilungen 172; Wörner, Verhaltenssteuerungsformen 203 f; see also joined cases 40 to 48, 50, 54 to 56, 111, 113 and 114/73 *Suiker Unie*, para 556, relating to a Commission Communication which was drafted in a misleading way and which was subject to (as the Court concluded: legitimate) expectations; joined cases C-75/05P and C-80/05P *Glunz*, para 65.

As regards the legislative, the wide discretion this branch of the EU disposes of in its decision-making is to be mentioned first.<sup>1436</sup> This is reflec‐ ted in particular in the leeway accorded to the Council when acting on a proposal/recommendation<sup>1437</sup> and, in particular in the ordinary legislative procedure, also to the EP (acting on a Commission proposal). Unless the legislator does not react at all,<sup>1438</sup> the consideration by the legislator of Commission soft law acts initiating rule-making is a necessary requirement for its reaction. Hence a 'necessity to consider' is inherent in the legislator's institutional position, according to which it can (in the majority of cases) only act upon an appropriate Commission proposal/recommendation.

Also the EU's executive principally seems to be bound to take into con‐ sideration EU soft law of other EU bodies.<sup>1439</sup> Since the EU's executive is by far the most important creator of EU soft law, the Court's case law is mainly about the – enhanced – effects of soft law *upon its creator*. Initially, this effect was affirmed by the Court in a number of staff cases. But which kind of bindingness is at issue here? In one of the first cases on this matter, the

<sup>1436</sup> As a consequence of the legislator's discretion, the Court has even refused the Council's being bound by its prior announcements; see case C-4/96 *NIFPO*, para 31: 'Annex VII [to a Council Resolution], which expresses essentially the Council's political will to take account, in applying the future common fisheries policy, of the special needs of regions in which the populations are particularly dependent on fishing and related activities, cannot produce legal effects capable of limiting the Council's legislative powers'; see also Cannizzaro/Rebasti, Soft law 222; for the (principal) self-bindingness of EU soft law see 4.2.3.2.3. below.

<sup>1437</sup> According to Article 293 para 1 TFEU, the Council may amend a Commission proposal by unanimity; concluding *e contrario*, Commission recommendations may (normally) be amended by qualified majority; see also (more in principle): Opinion of AG *Jacobs* in joined cases C-13–16/92 *Driessen en Zonen*, para 36.

<sup>1438</sup> Normally the legislator is not bound to act; note eg that there is no deadline for the legislator's action during the first reading of the ordinary legislative procedure (Article 294 paras 3–6); see Schoo, Art. 293 AEUV, para 17; for an exceptional duty to legislate see case 13/83 *European Parliament v Council*.

<sup>1439</sup> With regard to the Commission's duty to consider the soft law of the Council: Opinion of AG *Wathelet* in case C-425/13 *Commission v Council*, paras 176–178, with further references. Ambivalently with regard to the Commission's duty to consider ECHA guidance: case C-106/14 *FCD*, para 28. For the legal effects of *international* soft law on the Commission see case T-481/11 *Spain v Commission*, paras 78–81. While the Court stresses the freedom of the Commission to follow or not to follow standard recommendations of the UN Economic Commission for Europe addressed (also) to the MS of the EU, it confirms a duty of the Commission to at least take these standard recommendations into account (paras 80 and 85); more reluctantly with regard to MS' obligations in this respect: case C-279/12 *Shirley*, para 38.

Court emphasised that the Council 'should have regarded itself as under a moral obligation to comply with [its internal staff rules]'.<sup>1440</sup> Also in an early staff case concerning the Commission, it held that '[a]lthough an internal directive has not the character of a rule of law which the administration is always bound to observe, it nevertheless sets forth a rule of conduct indicating the practice to be followed, from which the administration may not depart without giving the reasons which have led it to do so'.<sup>1441</sup> This '*Louwage* formula' since then has become settled case law,<sup>1442</sup> also with a view to other EU bodies<sup>1443</sup> and also beyond staff cases<sup>1444</sup> – in particular in the areas of competition and State aid law.<sup>1445</sup>

The Court seems to have taken a stricter approach when it held and confirmed on several occasions (in particular in the context of Commission output) that the creator of soft law 'cannot depart from those rules under pain of being found, where appropriate, to be in breach of general princi‐ ples of law'.1446 However, it is to be noted that the self-bindingness of soft


<sup>1440</sup> Case 105/75 *Giuffrida*, para 17. Indeed, the Court intentionally left open the ques‐ tion whether or not the internal staff rules were to be considered 'a decision' (ie a legally binding act). The fact that the Court here spoke of a 'moral obligation', however, suggests that it implied the act was legally non-binding.

<sup>1441</sup> Case 148/73 *Louwage*, para 12; see further references in: H Adam, Mitteilungen 119; See also the Opinion of AG *Ruiz-Jarabo Colomer* in case C-415/07 *Gennaro*, para 36, in which he stresses that the Commission may not depart from its soft law 'without good reason'. This seems to correspond to a more general understanding of soft law; *Laurent Cytermann*, *rapporteur géneral adjoint* with the French *Conseil d'État*, with regard to (French) soft law said: 'L'utilisateur peut s'en écarter s'il a de bonnes raisons de le faire mais il ne peut pas complètement l'ignorer'; <http://ww w.dalloz-actualite.fr/interview/droit-souple-quelle-efficacite-quelle-legitimite-quell e-normativite> accessed 28 March 2023; see also Virally, Décennie 29.

law is not absolute and very much depending on the case at issue.<sup>1447</sup> The Court has indeed suggested that a soft law regime does not always have to be applied slavishly by its creator, but that – in the context of State aid and only 'in exceptional circumstances' – the Commission may grant an authorisation also where the criteria set out in its respective soft law act are not met.<sup>1448</sup> Such a deviance arguably is only allowed where it is not to the detriment of the party concerned,<sup>1449</sup> and where it applies to all (future) parties alike.<sup>1450</sup> In addition to that, soft law is not binding upon its creator when it has explicitly excluded any such effects,<sup>1451</sup> or when the soft law act provides for a wrongful interpretation of EU law.<sup>1452</sup>

As regards the legal effects on the judiciary, it is first to be noted that, since the Court is the highest authority in matters of EU law (in its own words having 'exclusive jurisdiction over the definitive interpretation of EU


For the importance of the Commission's compliance with its soft law beyond the legal argument of self-bindingness see Opinion of AG *Geelhoed* in case C-125/05 *VW-Audi*, paras 34 ff.

<sup>1447</sup> See eg the case of Commission Communication 'European agencies – The way forward', COM(2008) 135 final, in which the Commission uttered that it 'intends to […] [p]ropose no new regulatory agencies until the work of the evaluation is complete (end of 2009)' (pages 9 f ). This intention was counteracted in particular by the legislative proposals to create the ESAs, as launched by the Commission in September 2009. It is questionable, though, whether the Commission has legally bound itself by this statement. The expression 'intends to' seems to leave some lati‐ tude for the Commission. Against the background of the then on-going banking crisis, from a political perspective it was comparatively easy to justify the proposed establishment of the ESAs as necessary deviation from the Commission's earlier expressed 'intention'. The legislative proposals do not mention this deviation, let alone the reasons for it. The Commission may even have tried to conceal the connex to European agencies. After all, apart from Annex II *inter alia* on the ESAs' financial model, no reference to European agencies is made in the proposal and, as is well known, the ESAs are referred to as 'authorities' (even though, according to scholarly and even the Commission's own classifications, they qualify as European agencies). Considering a more generally binding effect: Hofmann/Rowe/Türk, Administrative Law 575 f.

law'1453), the situation is different from the EU's legislative and executive branches. According to Article 19 TEU, the Court 'shall ensure that in the interpretation and application of the Treaties *the law* is observed' (emphasis added).<sup>1454</sup> This position of the Court would not be incompatible, but somewhat at odds with an obligation to consider non-law – unless this obli‐ gation is itself laid down in law. This seems to be reflected in the Court's case law: 'the Court *may* […] take [recommendations] into consideration where they provide useful guidance for the interpretation of the relevant provisions of EU law' (emphasis added).<sup>1455</sup> But it is also apparent that it does not principally seem to exclude *any* legal effect of EU soft law on itself.<sup>1456</sup> It is not for nothing that the Court has declared soft law to be part of the *acquis communautaire*<sup>1457</sup> and that, when presenting the 'legal framework' or the 'legal background' of a case, it regularly includes the relevant soft law acts.1458 In the context of the Commission's suggestions for the calculation of a fine according to what is now Article 260 TFEU, it

<sup>1453</sup> Opinion 2/13 *ECHR II*, paras 245 f.

<sup>1454</sup> The Commission, at times, stresses the Court's superior role; see eg Commission Communication 'Freedom to provide services and the general good in the insur‐ ance sector', 2000/C 43/03, 6: 'It goes without saying that the Commission's interpretations do not prejudge the interpretation that the Court of Justice of the European Communities, which is responsible in the final instance for interpreting the Treaty and secondary legislation, might place on the matters at issue'. A closer analysis of the Court cases – in the field of competition and State aid law – leaves it unclear whether the Court of First Instance/General Court is actually, as has been argued in the literature, more willing than the Court of Justice to take soft law into account; see Ştefan, Soft Law 71f, with further references.

<sup>1455</sup> Joined Cases C-422/19 and C-423/19 *Dietrich*, para 48. See also case C-409/00 *Spain v Commission*, in which the Court downplayed its role as (potential) addres‐ see: '[T]hose notices and guidelines apply primarily to the Commission itself ' (para 69, with a further reference). The assumption that EU soft law has the same effects on the CJEU as it has on national courts may be 'coherent', but it cannot be deduced from the Court's case law; stressing, in this context, the Court's role in scrutinising the validity of EU soft law: Sarmiento, Soft Law 267.

<sup>1456</sup> For the AG' wariness in earlier cases to recommend to the Court to refer to the Commission's *de minimis* notice see references in Senden, Soft Law 367 f.

<sup>1457</sup> See references to the case law in Ştefan, Soft Law 118–120.

<sup>1458</sup> See Senden, Soft Law 361; for the Court taking inspiration from soft law see eg joined cases C-223/99 and C-260/99 *Agorà*, para 41; case C-310/99 *Italy v Commis‐ sion*, para 52, both with further references. It is to be stressed that the Court and the AG mostly refer to soft law in order to support a certain argument (of their respective own); see Ştefan, Soft Law 61, with regard to the field of competition and State aid law.

held that 'these suggestions of the Commission cannot bind the Court […]. However, [they] are a useful point of reference'.<sup>1459</sup>

More than a useful point of reference soft law may be where it is explicit‐ ly referred to in a legislative act,<sup>1460</sup> eg as a set of rules non-compliance with which is unlawful.<sup>1461</sup> Also in other cases the Court has acknowledged EU soft law, in particular of the Commission, eg on the interpretation of Treaty provisions<sup>1462</sup> or on the nomenclature of the Common Customs Tariff.<sup>1463</sup> With regard to the latter it held, without distinguishing between different addressees, that they 'are an important aid to the interpretation of the scope of the various tariff headings but do not have legally binding force'.<sup>1464</sup> Having said that, there are also cases in which the Court does not make an effort to consider soft law (or at least such an effort is not reflected in its reasoning), sometimes it even interprets the law contrary to what the wording of soft law suggests.1465 Overall, it appears that the Court does not follow a clear line in considering EU soft law1466 and that a duty to consider on behalf of the Court cannot generally be deduced from its case law.

<sup>1459</sup> Case C-387/97 *Commission v Greece*, para 89; case C-310/99 *Italy v Commission*, para 52. In case T-52/16 *Crédit mutuel Arkéa*, paras 73–77, the General Court states that certain guidelines of the European Committee of Banking Supervisors 'may be taken into account'. After considering them, it held – arguably with a view to the concrete case – that 'they cannot be accorded any particular weight'; see also references in Ştefan, Soft Law 158–161.

<sup>1460</sup> See case C-292/89 *Antonissen*, paras 17 f; case C-203/96 *Chemische Afvalstoffen*, para 31. With regard to 'dynamic references' to soft law and their risks in terms of democracy see Friedrich, Soft law 402.

<sup>1461</sup> See Korkea-aho, Soft Law 286–289, with examples.

<sup>1462</sup> See eg cases C-367/98 *Commission v Portugal*, para 47; C-483/99 *Commission v France*, para 43; C-503/99 *Commission v Belgium*, para 43.

<sup>1463</sup> See case C-203/96 *Chemische Afvalstoffen*, in particular paras 30 f, with regard to the Council Resolution of 7 May 1990 on waste policy; with regard to the Commis‐ sion's so-called 'Golden Share Communication' (97/C 220/06) see Senden, Soft Law 472, with references to case law.

<sup>1464</sup> Case C-124/15 *Salutas*, para 31; for the frequency of the EU courts' references to soft law in the field of competition and State aid law see Ştefan, Soft Law 87 ff.

<sup>1465</sup> It is to be noted, however, that the Court did not acknowledge to act contrary to soft law – Commission guidelines – in this case, but that the Commission has 'misconstrued [its] scope'; case T-304/08 *Smurfit*, para 97; critically: Storr, Wirtschaftslenkung 39 f.

<sup>1466</sup> See Eliantonio/Ştefan, Soft Law 459 f; Korkea-aho, Courts 476, with further refer‐ ences. In the context of Commission soft law, *Senden* argued that the Court is 'not very willing to take account' of it; Senden, Soft Law 373. The AG generally have been more explicit in this context; see eg Opinion of AG *Kokott* in case C-383/09

In connection with the self-bindingness of soft law discussed above, we may ask whether the Court – in reviewing the legality of a Commission decision – is bound by the Commission's pertinent soft law, as well. The answer to this question is contested.<sup>1467</sup> The Court held and has repeatedly confirmed that the Commission, eg in competition law, disposes of a wide discretion, and that its soft law often serves the purpose of concretising this discretion.<sup>1468</sup> In the course of judicial review, the Court may not substitute the Commission's exercise of discretion (as expressed in the act under review) by its own view.<sup>1469</sup> In light of this discretion, and depending on its width, it appears appropriate that the Court at least exercises deference to the Commission's discretion-concretising soft law. This applies also where the Court ought to protect legitimate expectations created by soft law. In specific cases this may boil down to an actual obligation of the Court to apply the Commission's soft law (see 4.2.3.2.3. below). Where the Commis‐ sion in its decision has unlawfully deviated from its soft law, the Court shall state the illegality of the former.<sup>1470</sup>

### 4.2.3.2. The actual or potential legal reasons for these effects

### 4.2.3.2.1. Sincere cooperation ('loyalty') according to Article 13 para 2 TEU

The sincere cooperation between the EU institutions (*Organtreue*), which the Court held to be part of the general principle of E(E)C/EU law of mu‐ tual sincere cooperation, is now codified in Article 13 para 2 TEU (second

*Commission v France*, para 28; Opinion of AG *Kokott* in case C-127/02 *Landelijke Vereniging*, para 95.

<sup>1467</sup> Unclear: case C-125/07P *Erste Group*, para 143; critically: W Weiß, Leitli‐ nien(un)wesen 258.

<sup>1468</sup> For the case law on the Commission's discretionary policy choices more generally see Craig/de Búrca, EU Law 610-612; for the case of competition law see ibid 1145 f; for the duty to exercise the allowed discretion see case T-122/15 *Landeskreditbank Baden-Württemberg*, para 139, with references to further case law. Describing this discretion, procedurally, as the legal basis of soft law: Hofmann/Rowe/Türk, Ad‐ ministrative Law 570.

<sup>1469</sup> See eg case T-35/99 *Keller*, para 77; see also Storr, Wirtschaftslenkung 29 f; Opin‐ ion of AG *Cosmas* in case C-83/98P *Ladbroke Racing*, paras 15 f, pointing to the Court's tendency to widen the scope of its review; for an exceptional higher intensity of review of discretionary acts see case C-501/11P *Schindler*, paras 36 f, with regard to Article 31 of Council Regulation 1/2003.

<sup>1470</sup> See eg case T‑185/05 *Italy v Commission*, paras 43 and 49.

sentence).1471 The concrete mutual duties of the institutions following from the principle of mutual sincere cooperation are to be fleshed out by the Court which so far has had few opportunities to do so.<sup>1472</sup> The lapidary wording of the provision – '[t]he institutions shall practice mutual sincere cooperation' – bears witness of the kinship between *Organtreue* on the one hand, and the sincere cooperation between the institutions and the MS, and between the MS themselves, on the other hand. This suggests that the duties between the institutions *principally* correspond to those existing between the institutions and the MS, and among the MS respectively.<sup>1473</sup> In general, it is to be noted that sincere cooperation 'is exercised within the limits of the powers conferred by the Treaties on each institution. The obligation[s] resulting from Article 13(2) TEU [are] therefore not such as to change those powers'.<sup>1474</sup> When applying this provision in the context of soft law, this quotation may remind us not to overrate the effect of soft law. Increased requirements for the 'treatment' of soft law by the EU bodies addressed by it may at the same time reduce the competences of these bodies, eg their prerogative of assessment. After all, it is one of the purposes of Article 13 para 2 TEU to ensure that EU bodies do not impede each other in exercising their respective powers.1475 Thus, the EP, for example, may not delay the issuing of its opinion on a legislative act without an objective reason where the Council has emphasised and reasoned the need for a timely adoption of this legislative act.<sup>1476</sup>

As was said above in the context of Article 4 para 3 TEU, also from Article 13 para 2 (second sentence) TEU no duty to apply the soft law adop‐ ted by other EU bodies can be deduced. A duty to merely consider other

<sup>1471</sup> See Obwexer, Art. 4 EUV, para 66, with reference to the CJEU's case law; see also Klamert, Loyalty 12 f, arguing that the Court has interpreted Article 4 para 3 TEU (and its predecessors, respectively) very differently with regard to inter-institution‐ al loyalty (as compared to the MS-EU relationship); for the inter-institutional dimension see also Cremona, Interest 157 f.

<sup>1472</sup> See Nettesheim, Art. 13 EUV, para 79.

<sup>1473</sup> See case 204/86 *Greece v Council*, para 16: '[The] dialogue [between the institu‐ tions; here: between the Commission, the Council and the EP] is subject to the same mutual duties of sincere cooperation which, as the Court has held, govern relations between the Member States and the Community institutions'; see also case C-65/93 *European Parliament v Council*, para 22.

<sup>1474</sup> Case C-73/14 *Council v Commission*, para 84, with a further reference; see also case C-660/13 *Council v Commission*, para 32, with further references.

<sup>1475</sup> See Jacqué, Art. 13 EUV, para 15.

<sup>1476</sup> See case C-65/93 *European Parliament v Council*, paras 23–28.

bodies' respective soft law may, however, be implied to the principle of mutual sincere cooperation between the institutions.<sup>1477</sup> It could be argued that this would also include a general duty to provide information on the reasons for non-compliance which could serve, as it were, as proof that the act at issue has actually been considered. Against this line of argumentation speaks Article 296 para 2 TFEU (see 4.2.3.2.2. below). Since this provision for legal acts only requires a reference to 'proposals, initiatives, recommen‐ dations, requests or opinions required by the Treaties', we may conclude *e contrario*: first, that there is no such duty for other acts of EU soft law – in particular those laid down only in secondary law – and, second, that there is no general duty to (substantially) justify deviance from an EU soft law act. Since this is not required for legal acts, it cannot be required in cases where non-compliance with an EU soft law act finds its expression otherwise, eg in an omission (non-action), either.

An example for the 'fleshing out' of the (assumed) obligations following from the principle of *Organtreue* in administrative practice is the Com‐ mission's approach towards Article 11 of the Comitology Regulation,<sup>1478</sup> pursuant to which the EP or the Council may express their view that a draft implementing act exceeds the implementing powers provided for in the basic act. The Commission, in a declaration annexed to the Comitology Regulation, stated that it will '*immediately* review the draft implementing act taking into account the positions expressed by the European Parliament or the Council', thereby '*duly* [taking] into account the urgency of the mat‐ ter' (emphases added). 'Before deciding whether the draft implementing act shall be adopted, amended or withdrawn, the Commission will inform the European Parliament or the Council of the action it intends to take *and of its reasons for doing so*' (emphasis added).1479 The italicised parts constitute concessions which go beyond the Comitology Regulation and may, apart from simply expressing a good will, be considered (by the Commission) as following from the principle of sincere cooperation according to Article 13 para 2 TEU. Whether the Commission, in addition to that, extends the EP's and the Council's means of control beyond cases in which the

<sup>1477</sup> See Selmayr, Art. 282 AEUV, para 80, with regard to the ECB's opinions. See eg the Commission's view paraphrased in case C-119/97P *Ufex*, para 9.

<sup>1478</sup> Regulation 182/2011/EU.

<sup>1479</sup> Statements by the Commission, OJ L 55/19 of 28 February 2011 <http://eur-lex.eu ropa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52011XC0228(01)&from=EN> accessed 28 March 2023; see also Ilgner, Durchführung 223.

basic act was adopted under the ordinary legislative procedure1480 is – in a literal understanding of the Commission's statements – possible, but still unclear.<sup>1481</sup>

### 4.2.3.2.2. Article 296 para 2 TFEU

Article 296 para 2 TFEU stipulates that '[l]egal acts shall state the reasons on which they are based and shall refer to any proposals, initiatives, recom‐ mendations, requests or opinions required by the Treaties'. This provision aims at ensuring the reviewability of the act on the one hand (obligation to state the reasons),<sup>1482</sup> and at ensuring the contributory rights of EU bodies, in particular the Commission, in the various decision-making procedures provided for in the Treaties, on the other hand (obligation of reference).<sup>1483</sup> The required reference shall officially prove that the contributory rights were made use of.1484 Failure to grant the contributory rights in a decisionmaking procedure renders void the act in question,1485 the mere failure to refer to such a contributory act does not.

It is contested whether a mere reference to the act at issue suffices, or whether a presentation of the arguments brought forward, a list of those which eventually were refused or even, more generally, a discussion of the respective arguments is required.1486 In general, individual-concrete

<sup>1480</sup> Article 11 of Regulation 182/2011/EU.

<sup>1481</sup> See Kröll, Rechtsetzung 294.

<sup>1482</sup> An essential tool in this context is the reasoning of soft law acts: Therefore the reasoning serves in particular the interests of the addressees of the measure or other persons for whom the act is of immediate concern. Thus, the Court decided that a MS, having participated in a legislative process leading to the adoption of an act, 'cannot validly complain that the Parliament and the Council, the authors of the [act], did not place it in a position to know the reasons for the choice of measures which they intended to implement'; case C-508/13 *Estonia v European Parliament/Council*, para 62.

<sup>1483</sup> See Krajewski/Rösslein, Art. 296 AEUV, para 44, with further references; for a wide interpretation of the term 'opinions' so as to include all kinds of contributions (required by the Treaties): Calliess, Art. 296 AEUV, para 34; similarly: Geismann, Art. 296 AEUV, para 24.

<sup>1484</sup> See Krajewski/Rösslein, Art. 296 AEUV, para 45, with further references; Vcelouch, Art. 296 AEUV, para 18.

<sup>1485</sup> See case 828/79 *Adam*, paras 15–17.

<sup>1486</sup> Affirming a duty to provide a (substantive) discussion in accordance also with the duty to state the reasons stipulated in Article 296 para 2 TFEU: Calliess, Art. 296

measures have to provide a more detailed reasoning than general-abstract measures.<sup>1487</sup> In particular in the context of generally applicable acts, no general duty to explicitly affirm/refuse or even substantially discuss all the arguments brought forward in the said contributory acts can be deduced from Article 296 para 2 TFEU, neither from the duty to state reasons nor from the duty of reference contained therein. This is underpinned by the Court's case law,<sup>1488</sup> determining that while the statement of reasons according to Article 296 TFEU 'must show clearly and unequivocally the reasoning of the EU authority which adopted the contested measure, so as to enable the persons concerned to ascertain the reasons for it and to enable the Court to exercise judicial review, it is not required to go into every relevant point of fact and law'.<sup>1489</sup> The quality of the reasoning also depends on 'the circumstances of each case, in particular the content of the measure in question, the nature of the reasons given and the interest which the addressees of the measure, or other parties […], may have in obtaining explanations' and 'must be assessed with regard not only to its wording but also to its context and to all the legal rules governing the matter in question'.<sup>1490</sup>

Apart from the enhanced requirements for the reasoning of individualconcrete acts, as opposed to general-abstract acts, another differentiation can be carried out: the more contested a measure (normally this is reflec‐ ted also in the arguments of the contributing EU bodies), the higher the threshold for the reasoning to be given. Thus, a substantive discussion of the counter-arguments *may* be necessary in order to satisfactorily provide

AEUV, para 34; Vedder, Art. 296 AEUV, para 10; against them: Geismann, Art. 296 AEUV, para 22; Krajewski/Rösslein, Art. 296 AEUV, para 45; and arguably also Gellermann, Art. 296 AEUV, para 17.

<sup>1487</sup> See joined cases C-78/16 and C-79/16 *Pesce*, paras 88-90; case T-122/15 *Landeskre‐ ditbank Baden-Württemberg*, para 123, with a further reference; see also Türk, Oversight 133, with references to the Court's case law; note Article 41 para 2 lit c CFR.

<sup>1488</sup> See references in Hofmann/Rowe/Türk, Administrative Law 199.

<sup>1489</sup> Joined cases C-78/16 and C-79/16 *Pesce*, para 88, with further references; see also case C-519/15P *Trafilerie Meridionali*, paras 40 f; case C-493/17 *Weiss*, para 31, with further references; note also joined cases T-74/00, T-76/00, T-83/00, T-84/00, T-85/00, T-132/00, T-137/00 and T-141/00 *Artegodan*, para 42.

<sup>1490</sup> Case T-63/16 *E-Control*, para 68, with further references; see also case T-122/15 *Landeskreditbank Baden-Württemberg*, para 124, with a further reference. The 'legal rules governing the matter in question' may also include soft law acts; see para 125.

reasons in accordance with Article 296 para 2.<sup>1491</sup> This may also require a consideration of soft law containing such arguments (irrespective of whether this soft law is 'required by the Treaties' or not). However, the consideration of soft law is not necessarily a burden for the decision-maker. The reasoning contained in a soft law measure may be taken over by other acts, eg when the Commission in an individual-concrete decision on a State aid case refers to its (earlier adopted) guidelines for the examination of State aid in the relevant sector, thereby (partially) substituting the decision's own provision of reasons.<sup>1492</sup> This practice may also be applied with regard to soft law acts stemming from other EU bodies. In these cases the consid‐ eration of soft law facilitates the reasoning of an act.

In conclusion, it can be said that in Article 296 para 2 TFEU it is the duty to state reasons which effects a duty to consider contributory (soft law) acts, not the duty to refer to these acts. A provision of reasons, limited as it may be, allows the addressee(s) to be informed about the main reasons leading to the adoption of a measure, which again presupposes that the decision-maker took account of the contributory acts mentioned in Article 296 TFEU and – since the limitation to the contributory acts required by the Treaties only relates to the duty of reference – beyond.

### 4.2.3.2.3. Legal certainty, legitimate expectations, equality and effectiveness

In complementation to what was said above under 4.2.2.2.4. about the legal foundation of the principles of legal certainty, protection of legitimate expectations, equality and effectiveness in EU law (which *mutatis mutandis* applies here, as well), we shall now address the relevant effects of EU soft law specifically on EU institutions, bodies, offices and agencies.

<sup>1491</sup> See case 158/80 *Rewe-Handelsgesellschaft Nord*, paras 25–27, in which the Court decided that due to a 'contradiction in the statement of reasons' and a lack of 'legal justification' the Regulation in question 'does not contain a statement of the reasons on which it is based as required by Article [296 TFEU]'; see also case C-304/01 *Spain v Commission*, para 50, with further references.

<sup>1492</sup> For hard law acts referring to soft law as a form of '"ready-made" reasons'; Ştefan, Soft Law 127, with references to the Court's case law; see also case T-576/18 *Crédit agricole*, para 138, with further references.

As regards the binding effect of EU soft law upon its creator(s),<sup>1493</sup> the Court has repeatedly considered legal certainty, legitimate expectations and the equality principle as laid down, *inter alia*, in Articles 20 f CFR,<sup>1494</sup> as legal bases for this effect.<sup>1495</sup> Whether the principles of legal certainty and legitimate expectations and of equality cause the self-binding effect on their respective own or only in combination is unclear. In a judgement on the first Banking Communication the Court mentioned both legitimate expectations and equality in this context.<sup>1496</sup>

Individuals/undertakings concerned by EU soft law can rely on this effect, meaning that – given their legal standing – they can request annul‐ ment of a legal act which does not comply with relevant soft law adopted by the creator of this act.<sup>1497</sup> Whether this is also possible for MS or (other)


That EU soft law addressed to another EU body gives rise to legitimate expecta‐ tions of individuals that the addressee will comply with it appears to be, if at all, a rare exception. With regard to non-binding EFSA output addressed to the Commission: case T-177/13 *TestBioTech*, para 114. The Court negates the existence of legitimate expectations of individuals here, but it does not in general exclude the possibility that an EU soft law act addressed to an(other) EU body may give

<sup>1493</sup> EU bodies may explicitly exclude this effect; see eg Commission Communication 'Towards a reinforced culture of consultation and dialogue – General principles and minimum standards for consultation of interested parties by the Commission', COM(2002) 704 final, 10 and 15; arguing against an automatic bindingness of soft law upon its creators: Arndt, Sinn 49; with regard to inter-institutional agreements, the following is to be said: If the institutions are allowed to enter into binding agreements with each other (see von Alemann, Einordnung 131–135), they are – as an exception to the principal self-obligation brought about by soft law – also free to agree on legally non-binding terms; see case C-25/94 *Commission v Council*, para 49; see also Hummer, Interorganvereinbarungen 99–102, analysing the Court's case law.

EU bodies is questionable. In my view this should be confirmed, as it is not perceivable that MS in principle are in need of the above protection to a lesser extent than individuals/undertakings.<sup>1498</sup> Also EU bodies should in principle be entitled to rely on the self-binding effect of the soft law created by other EU bodies.

While soft law, as the Court stated with regard to State aid guidelines, 'certainly help[s] to ensure that [the Commission] acts in a manner which is transparent, foreseeable and consistent with legal certainty',<sup>1499</sup> the Court has not requested strict compliance with this soft law by its respective creator at all times (see 4.2.3.1. above).<sup>1500</sup> In case of deviation, the above principles require the respective EU body to provide the reasons therefor<sup>1501</sup> (so as to ideally make clear that due to the specificities of the given case there is no violation of these principles).

Also with regard to the amendment of soft law – a scenario which is somewhat related to the non-application of (a certain version of ) soft law by its creator – these principles play a role.1502 While amendments must, as in the case of law, be possible in principle,1503 an abrupt and far-reaching amendment – without any transition period – may well be contrary to the protection of legitimate expectations.1504 An appropriate reasoning has a key function in justifying an amendment, thereby possibly outdoing the protection of certain expectations.1505 The (potential) legitimate trust that soft law acts are not amended tends to be higher in case of individual-con‐ crete soft law acts than in case of general-abstract ones.<sup>1506</sup>

cause to legitimate expectations in individuals; after all, it lists a number of further reasons for refusing legitimate expectations here.

<sup>1498</sup> See Wörner, Verhaltenssteuerungsformen 192–194, in particular 193, with further references; for the applicability of the principle of equality also in the relationship EU-MS see case C-387/97 *Commission v Greece*, para 84.

<sup>1499</sup> Case C-310/99 *Italy v Commission*, para 52; see also Opinion of AG *Ruiz-Jarabo Colomer* in case C-415/07 *Gennaro*, para 35.

<sup>1500</sup> See eg case C-520/09 *Arkema*, para 93; for an example in which strict compliance was required see case T-210/01 *General Electric*, para 516.

<sup>1501</sup> See Opinion of AG *Tesauro* in case C-58/94 *Netherlands v Council*, para 20.

<sup>1502</sup> See Walzel, Bindungswirkungen 110.

<sup>1503</sup> See case C-1/98P *British Steel*, para 52.

<sup>1504</sup> See case C-63/93 *Duff*, para 20. This passage informs us more generally that where 'constant adjustments to meet changes in the economic situation' are necessary, the retroactivity of rules – and *a fortiori* arguably also the abrupt change of these rules – may be lawful.

<sup>1505</sup> See also Thomas, Bindungswirkung 427.

<sup>1506</sup> See Wörner, Verhaltenssteuerungsformen 197.

### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

The broad principle of effectiveness or *effet utile* of EU law may suggest the consideration of EU soft law by other EU bodies, and also its self-bind‐ ing effect.<sup>1507</sup> While the adoption of soft law has been praised as increasing the effectiveness of EU law,<sup>1508</sup> as a follow-up this effectiveness also requires the application of soft law. However, due to its broadness, it does not appear that from the principle of effectiveness specific effects of EU soft law *vis-à-vis* EU institutions, bodies, offices and agencies can be deduced. Therefore the principle of effectiveness may only be brought forward as a principle in the light of which more concrete norms, eg Article 13 para 2 TEU, are to be interpreted.

# 4.3. Factual effects

### 4.3.1. Introduction

It was aptly remarked by *Schermers* and *Blokker* that '[t]he existence of a legal obligation provides merely one of many reasons for observing a rule'.1509 Here we shall address the factual effects of a category of norms which do not create legal obligations. The factual effects of soft law depend on its addressees, and hence they are as individual and multi-faceted as life itself. People – as citizens or as monocratic/collegiate decision-makers of an authority, of a public or of a private undertaking – may *feel* the necessity to follow soft law for a variety of different reasons: they may feel morally obliged, they may feel it is more opportune, they may simply want to obey what they consider a command, they may be persuaded by the content of the rule,1510 etc. While these reasons, and the related effects of soft law respectively, principally apply to private persons, MS, and EU actors as addressees of EU soft law alike, which is why here no differentiation between different groups of addressees is required, it is impossible to list

<sup>1507</sup> For the effectivity aspect of sincere cooperation see Senden, Soft Law 95; see also Opinion AG *Kokott* in case C‑43/10 *Aitoloakarnanias*, para 226, where effective‐ ness considerations seem to play a role.

<sup>1508</sup> See eg Brohm, Mitteilungen 80 f, with further references.

<sup>1509</sup> Schermers/Blokker, Institutional Law, § 1220; for the argument that a rule's legiti‐ macy (in the view of its addressees) is decisive for compliance see Tyler, People.

<sup>1510</sup> See case C-16/16P *Belgium v Commission*, para 26, which mentions – in the context of recommendations – 'the power to exhort and to persuade'.

them all.1511 Therefore we shall confine ourselves to three aspects of the factual effects of soft law, one very generally referring to human nature (4.3.2.), the second one dealing with the phenomenon of 'nudging' which has found resonance in the EU's toolkit for ensuring compliance (4.3.2.1.), and the third one consisting of a systemic idiosyncrasy of the European *Verwaltungsverbund* (4.3.2.2.). The second and the third factor are subsets of the first one, as they are both to be understood against and based on the specificity of human nature. This is why they shall be dealt with under sub-headings to 'human nature'. As mentioned above, this sub-chapter in no way claims completeness, but is expressly limited to selected thoughts which should help understand the factual effects of EU soft law.

### 4.3.2. Human nature

The general aspect of soft law's factual effects in this context is human propensity to obey the commands of what is perceived as an 'authority' on the one hand, and to adapt one's behaviour to that of others on the other hand.1512 This was, among others, impressively demonstrated in the world-famous experiments of *Asch* and *Milgram*. <sup>1513</sup> In the setting of these experiments no legally binding rules were provided. Rather, the test persons adapted their behaviour to the (felt) command of others (peers and a presumed expert, respectively). That a human propensity to obey even un‐ lawful commands exists, has been legally acknowledged, for example, in the form of the legal construct *Befehlsnotstand*, which allows for delinquents to be subject to milder punishment, or even to be acquitted, where they were following an order from a superior body/person when committing the incriminated act. Whether lawful or not, it is often felt to be conven‐ ient to follow an existing rule, because it appears to reduce one's own responsibility (in psychology/sociology this phenomenon is referred to as

<sup>1511</sup> For different theories towards explaining compliance in the realm of public inter‐ national law see Kingsbury, Concept. With regard to MS' compliance with EU soft law, *Härtel* mentions practicability, understanding and opportunity as possible motives; Härtel, Rechtsetzung 273.

<sup>1512</sup> With regard to EU (soft) law see U Stelkens, Rechtsetzungen 407; with regard to public international law see Koops, Compliance 33. For a broader picture of intrinsic and extrinsic motivation see Ryan/Deci, Motivations; Rupp/Williams, Efficacy.

<sup>1513</sup> See Asch, Studies; Milgram, Obedience 371, with further references, also to his own work.

'diffusion of responsibility').<sup>1514</sup> An existing rule may also be deemed to spare (intellectual) resources because the addressee does not have to think himself or herself of the adequate behaviour in a certain situation.<sup>1515</sup> The question whether this rule is legally binding or not, may then – to some extent at least – lose its importance.<sup>1516</sup>

Both behavioural mechanisms addressed here – obedience to a higher authority and peer pressure – are at issue when it comes to the application of rules of soft law. First, they originate with an authority and, second, where they are addressed to more than one actor, the respective other actors' behaviour may have an influence on individual compliance. It may now be countered that EU soft law is entirely different from the commands given in the *Milgram* experiment, and that *Asch's* experiment was about individuals succumbing to an objectively incorrect opinion of the majority, not about compliance or non-compliance with legally non-binding rules. But it is not the purpose of these lines to enter into an intricate psychologi‐ cal discourse and to scientifically apply the findings of these experiments to the case of EU soft law anyway. This would, for lack of existing empirical data, require new experiments. Rather, the author intends to point to two fundamental *stimuli* of human behaviour which assumedly have, in a num‐ ber of cases, an influence on the application of soft law by those addressed. The differences between EU soft law and the commands or pseudo-com‐ mands in the mentioned experiments – EU soft law is more complex and therefore less easy to be followed *spontaneously*, it is maybe less authorita‐ tive, its addressees are regularly more self-confident and knowledgeable than the test persons (they know about soft law's legal non-bindingness), etc – cannot do away with the general relevance of these findings for the topic 'factual effects of EU soft law'.

<sup>1514</sup> The counterpart of this reduction of responsibility on the part of the addressee is the exercise of authority *qua* the adoption of rules on the part of their respective creator; see Jabloner, Rechtsetzung 16; for the increased effectiveness of soft law in case of emergencies see Feik, Verwaltungskommunikation 387.

<sup>1515</sup> See Kovács/Tóth/Forgác, Effects 59.

<sup>1516</sup> See Jabloner, Richterrecht 29, who, in a different context, namely that of case law, describes the phenomenon of a blurring of 'voluntary' consideration and 'obligatory' application of case law.

### 4.3.2.1. The steering effects of 'nudging'

Research on the effect of so-called 'nudging' has shed light on the effects of steering measures which are neither order nor ban, neither financial in‐ centive nor penalty. Nudges are described as 'liberty-preserving approaches that steer people in particular directions, but that also allow them to go their own way'.<sup>1517</sup> This actual freedom to act in one or the other way is a necessary condition of nudging. It is based on an empirical approach towards human behaviour in decision-making. The possibilities which nudging entails, by now studied in academia for about 20 years, are being considered by an increasing number of political actors as a welcome sup‐ plement to more traditional methods of governing human behaviour.<sup>1518</sup>

*Sunstein* distinguishes four tendencies of human behaviour which make human decision-making accessible to nudges: Inertia and procrastination; framing and presentation; social influences; difficulties in assessing proba‐ bility.1519 Here the author would like to dwell on selected findings with regard to two of these tendencies. In respect of inertia and procrastination, one of *Sunstein's* findings is that 'default rules have a large effect on social outcomes'.1520 Default rules determine the choice between at least two al‐ ternatives where the person concerned does not (for whichever reason) actively choose. In the field of retirement savings, *Sunstein* exemplifies, the content of the default rule is highly important. If the question reads 'Do you want to opt in to a retirement plan?' the number of participants is substantially lower than if the addressees are asked 'Do you want to opt out of a retirement plan?', making clear that in the latter alternative in case of inaction they would be enrolled in the programme. Such a 'go with the flow'1521 approach 'may well be the most effective [group of ] nudges'.<sup>1522</sup> With regard to procrastination *Sunstein* claims, *inter alia*, that 'the identifi‐ cation of a specific, clear, unambiguous path or plan has an important effect on social outcomes. Complexity or vagueness can ensure inaction […]'.<sup>1523</sup>

<sup>1517</sup> Sunstein, Nudging 583.

<sup>1518</sup> See Reisch/Sandrini, Nudging 20 f.

<sup>1519</sup> See Sunstein, Regulation 1350 ff.

<sup>1520</sup> Sunstein, Regulation 1350.

<sup>1521</sup> Dolan/Hallsworth/Halpern/King/Vlaev, Mindspace.

<sup>1522</sup> Sunstein, Nudging 585.

<sup>1523</sup> Sunstein, Regulation 1352 f.

Pertaining to the heading 'social influences', he emphasises the depend‐ ence of human behaviour on the behaviour of the respective peers.<sup>1524</sup> What may be considered a commonplace is of salient importance when it comes to steering human behaviour. The behaviour of others in a person's envi‐ ronment coins the image of 'what ought to be done', of what is necessary to maintain a good reputation.<sup>1525</sup> *Sunstein* adds that '[i]n some contexts, social norms can help create a phenomenon of *compliance without enforce‐ ment* – as, for example, when people comply with laws forbidding indoor smoking or requiring the buckling of seat belts, in part because of social norms or the expressive function of those laws' (emphasis in original).<sup>1526</sup>

While the focus of nudging traditionally has been laid on consumer policy, we may consider the respective findings also from a different per‐ spective, namely from the perspective of soft regulation both addressed to citizens/undertakings and to public authorities.1527 Compliance with EU soft law is certainly not an automatised act like switching off the light when leaving a room,1528 but a conscious, reflected-upon decision. Nevertheless, it may be influenced by nudges.1529 What was paraphrased above about the importance of the default rule also seems to underpin the assumption that if there is a rule, people are – for the reason of inertia – more likely to act in a way corresponding to that rule than if there is no such rule. This is a human idiosyncracy which works in favour of (compliance with) soft law. The effect of soft law may be enhanced, for example, by a 'comply or explain' requirement which also avails itself of people's inertia. Also the finding on people's dependence on their respective peers is relevant, in particular when considering the 'naming and shaming' practice in the EU, especially *vis-à-vis* national authorities.1530 The concept of 'naming and shaming' or 'naming, blaming and shaming' rests on people's dependence

<sup>1524</sup> See Sunstein, Regulation 1356.

<sup>1525</sup> Sunstein, Regulation 1357.

<sup>1526</sup> Sunstein, Regulation 1357.

<sup>1527</sup> Because also the action of public authorities is governed by human beings, in principle there is no difference in the decision-making (motivation etc) of citizens on the one hand, and of public authorities on the other hand; see also Goldmann, Gewalt 338 f.

<sup>1528</sup> For nudges facilitating the switching off of lights see Reisch/Sandrini, Nudging 112 f.

<sup>1529</sup> For the distinction between these two kinds of nudges – those focusing on automatised behaviour and those aiming at conscious decision-making – see Re‐ isch/Sandrini, Nudging 33, with further references.

<sup>1530</sup> See Armstrong, Character 198.

on others, on their fear of losing their (good) reputation. Thus, where deviance from (soft) rules is disclosed to peers, this may have a chilling effect for potential deviators. Of course, this works only as long as the deviators constitute a minority. Where non-compliance with certain soft law acts becomes more and more common among peers, the 'naming' loses its 'blaming' and 'shaming' effect, because then apparently deviance does not come together with a loss of reputation. In accordance with this logic, there would even be an increased risk of peers adapting their behaviour to that of their surroundings, ie to deviate themselves.

Such nudges need not necessarily be systemically connected to the soft law measure at issue, in the sense that they are, for example, laid down in the very provision which constitutes the (eg secondary law) legal basis for the adoption of the respective soft law act. The incentive to comply may equally well stem from a very distant source. This is the case, for example, with the threat of the Commission initiating a Treaty infringement proce‐ dure. A MS body may comply with EU soft law – possibly following an according instruction from a superior national body – in order not to draw the Commission's attention to the respective MS, and in order to thereby reduce the likelihood of the Commission initiating a Treaty infringement procedure for a different reason.1531 The Commission has already made use of this power in order to facilitate MS support in a different scenario.<sup>1532</sup>

The EU law incentives to comply with EU soft law addressed here are all nudges according to the above definition: They have a steering effect, but allow their addressees to 'go their own way' (*Sunstein*), that is to say that these addressees are still legally free to comply or not to comply.

<sup>1531</sup> Where the soft law at issue is concretising EU law, non-compliance with the former may be interpreted as a violation of the latter. A Treaty infringement procedure may then be launched also in this case; see Thomas Müller, Ziele 13; see also case T-258/06 *Germany v Commission*, para 151, with further references.

<sup>1532</sup> See Blauberger/Weiss, Commission 1123; Pollak/Slominski, Energy Market 100 f. On the other hand, the Commission has also suspended proceedings for reasons not related to the question of whether the MS concerned has complied with Union law; for the Commission's putting on hold of pending infringement proceedings against Greece during the financial crisis see Gormley, Infringement 68.

### 4.3.2.2. EU soft law created by MS officials

A structural element of the European *Verwaltungsverbund* is the strong involvement of MS representatives (including representatives of national authorities) in the decision-making at EU level. Among other things, this involvement or participation increases the acceptance of the resulting deci‐ sions and makes it, where they are addressed to a MS or to a national authority, more difficult to oppose them.<sup>1533</sup>

The described scenario is typical in particular of European agencies/net‐ work bodies.<sup>1534</sup> Their respective main decision-making bodies are normal‐ ly composed of one representative per MS/competent national authority.<sup>1535</sup> A number of these agencies/network bodies is competent to address (soft law) acts to its national counterparts. For example: According to Article 6 para 5 of Regulation 2019/942, the Agency for the Cooperation of Ener‐ gy Regulators (ACER) 'shall provide a factual opinion at the request of one or more regulatory authorities or of the Commission, on whether a decision taken by a regulatory authority complies with the network codes and guidelines referred to in [the pertinent EU law]'.<sup>1536</sup> These opinions are adopted by the ACER's Director, but only upon a favourable opinion of the Board of Regulators.1537 The Board of Regulators is composed of senior representatives of the competent authorities in the MS. Each representative may once be in the situation that it co-adopted a favourable view on a draft opinion which was then adopted and addressed to his/her respective national authority. The fact that he/she has engaged in *institutional cooper‐*

<sup>1533</sup> The Court refers to this mechanism in case C-311/94 *IJssel-Vliet*, para 39. With regard to the cooperation between Commission and MS under Article 108 para 1 TFEU see Georgieva, Soft Law 236 f and 244 f; Schweda, Principles, para 32; J Scott, Limbo 341f; see also H Adam, Mitteilungen 107–113.

<sup>1534</sup> See Lafarge, Coopération 68 and, specifically with regard to Eurojust, Europol and Frontex, 80–82; see also C Scott, Government 167.

<sup>1535</sup> See Analytical *Fiche* Nr<sup>o</sup> 5 ('Composition and designation of the Management Board') 1 <https://docplayer.net/19569067-Analytical-fiche-nr-5-composition-and -designation-of-the-management-board-1.html> accessed 28 March 2023.

<sup>1536</sup> Explaining the insistence of the Council on the formulation 'opinion, based on matters of fact' under Article 7 para 4 of Regulation 713/2009 (the predecessor of Regulation 2019/942): Ermacora, Agency 268.

<sup>1537</sup> Article 22 para 5 lit a of Regulation 2019/942; for the dominant role of the Director in ACER decision-making see Ştefan/Petri, Review 528 f.

*ation*, <sup>1538</sup> more precisely that he/she has contributed to the adoption of the opinion in his/her role as a member of the ACER's Board of Regulators makes it difficult for him/her to oppose such an opinion in his/her func‐ tion as a representative of the national authority,<sup>1539</sup> not least for reasons of credibility.<sup>1540</sup> A self-obliging effect *vis-à-vis* the respective EU body – according to which the senior official of the national authority has to make the latter comply with a soft law act which he/she has favoured within the Board of Regulators, and which was then addressed to his/her respective authority – is to be denied, though.<sup>1541</sup> It is to be emphasised that these effects, if any, are factual. The described constellation cannot be qualified as legal 'agreement' between the representatives of national authorities<sup>1542</sup>; for potential legal effects in such constellations see in particular 4.2.2.2.4. above.

These loyalty effects cannot be triggered where the agency's main deci‐ sion-making body is not composed of MS representatives. This was promi‐ nently shown in the dispute on the pesticide-ingredient *glyphosate* in which the EFSA – whose Management Board by then was composed of experts, whose respective nationality played an only marginal role1543 – and the MS uttered opposing opinions.<sup>1544</sup>


<sup>1538</sup> For different categories of cooperation, institutional cooperation being one of them, see Schmidt-Aßmann, Verwaltung 1380.

#### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

The loyalty effects may also be comparatively low in case of Commission soft law addressed to a national authority. The Commission is composed of one representative per MS, it is true, but the bonds to the national administrations are much weaker, in particular because here there is no double-hatting (as in the case of most agencies/network bodies) linking the EU and the respective national administration.<sup>1545</sup> However, this is only one aspect of the loyalty corset, as it were. The 'consociational model of interest intermediation'<sup>1546</sup> may enhance the factual propensity to comply with EU soft law also in other constellations.

### 4.4. Mixed effects

With soft law as a set of legally non-binding norms, it is possible for its addressees to lawfully refuse compliance. This non-application may be aggravated in different ways, though.1547 In addition to the above reasons facilitating compliance with EU soft law, Union law may make non-com‐ pliance with certain acts of EU soft law subject to enhanced conditions, or simply set an enhanced threshold for its consideration – for both MS and EU bodies. Examples for such a differentiation in EU (primary and secondary) law are multiple. Some of these examples shall be mentioned here with a focus on the 'treatment' of soft law they entail, not on the legal field in which they are domiciled.

Article 143 TFEU provides for a case in which the Council may act only upon a recommendation by the Commission. Where the Council does not follow a Commission recommendation (ie does not grant the assistance at issue), the Commission attains regulatory power in lieu of the Council (para 3). The Council may then, however, revoke/adapt the Commission's actions without being dependent on a respective Commission recommen‐ dation.<sup>1548</sup>

<sup>/</sup>2016/mar/04/eu-states-rebel-against-plans-to-relicense-weedkiller-glyphosate>, both accessed 28 March 2023.

<sup>1545</sup> See Döring, Composition 225, who stresses the importance of the Commissioners' 'socialization in office'; pointing at these dynamics in the context of the solidarity principle: W Weiß, Solidarität 415 f.

<sup>1546</sup> Hix, System 223–225, with further references.

<sup>1547</sup> In the context of Article 126 para 9 TFEU, for example, *Hofmann*, *Rowe* and *Türk* speak of 'quasi-coercive steps [that] may follow'; Hofmann/Rowe/Türk, Ad‐ ministrative Law 546 (fn 58).

Pursuant to Article 7 of Protocol No 2 on the Application of the Princi‐ ples of Subsidiarity and Proportionality, the effect of national parliaments' reasoned opinions on the non-compliance with the principle of subsidiarity of a draft EU legislative act depends on quantity: Where they represent at least one third of all votes allocated to national parliaments, the draft must be reviewed. The originator of the draft legislative act may then maintain, amend or withdraw the draft, thereby giving the reasons for its decision (para 2). Where, under the ordinary legislative procedure, the reasoned opinions respecting the Commission proposal for a legislative act represent at least a simple majority of these votes, the proposal must be reviewed. The Commission may then maintain, amend or withdraw the proposal. If it decides to maintain it, it shall justify its choice in a reasoned opinion (para 3).

According to Article 19 paras 4–6 of Protocol No 5 on the Statute of the European Investment Bank, the Management Committee of the EIB shall examine whether financing operations submitted to it are in accordance with this Protocol. After this examination, it shall forward the case to the Board of Directors for a decision, together with an opinion. Where this opinion is unfavourable, the Board may grant the finance concerned only by a unanimous decision. The effect of an unfavourable opinion – ie an enhanced degree of approval, namely unanimity, required for a decision not complying with this opinion – is comparable to that of a negative Commission opinion in the ordinary legislative procedure (Article 294 para 9 TFEU): Amendments on which the Commission has delivered a negative opinion may be adopted only by a unanimous decision of the Council (see 3.5.2.1.1. above). It goes without saying that it is politically more difficult to reach unanimity within a collegiate body than to reach a (simple or qualified) majority.

An example laid down in secondary law is the so-called reverse (quali‐ fied) majority voting which is, for example, provided for in Regulation 1466/97, as amended by the so-called 'Six Pack'. Article 6 para 2 provides for the following procedure which shall be dealt with here in relative isola‐ tion: 'the Commission […] shall recommend to the Council to adopt the decision establishing that no effective action has been taken. The decision shall be deemed to be adopted by the Council unless it decides, by simple

<sup>1548</sup> See Bandilla, Art. 143 AEUV, para 33, who stresses the Council's role as a 'Beru‐ fungsinstanz' [appeal body].

majority, to reject the recommendation within 10 days of its adoption by the Commission'.<sup>1549</sup> In other words: A decision not to follow the Commis‐ sion recommendation requires a majority in the Council, whereas the lack of such opposing majority suffices for the recommendation to become a decision. This substantially increases the authority of the Commission's rec‐ ommendation and suggests a 'go with the flow' approach on the part of the Council (see 4.3.2.1. above). The recommendation may, of course, still be refused, but only under enhanced conditions. It was argued that this regime actually provides for a *de facto* decision-making power of the Commission with a right to object on the part of the Council. This image reflects well the practical effects, but from a legal point of view it ought to be emphasised that the law provides for a Commission *recommendation* (addressed to the Council) which may be transformed into a Council decision (addressed to a MS) due to the Council's non-objection, that is to say: its inaction. Therefore, reverse (qualified) majority voting is an example – admittedly: an extraordinary example – of an increased authority of Commission soft law, not of the Commission's power to adopt a binding decision.

Another example taken from secondary law is the procedure laid down in Article 17 of the Regulation 1093/2010. Where a breach of Union law by a competent authority in the MS (or, respectively, the ECB as a banking supervisor) is suspected by the EBA, it may – after some preliminaries – address a recommendation to this authority, aimed at ensuring compliance by the latter. Where this recommendation is not followed, the Commission may adopt a formal opinion (with a similar content) as a follow-up to the EBA's recommendation. Where the competent authority does not comply with the formal opinion in due time, either, the EBA may, 'where the relevant requirements of [the pertinent EU law] are directly applicable to financial institutions […], adopt an individual decision addressed to a financial institution […] requiring it to take all necessary action to com‐ ply with its obligations under Union law […]'.1550 Here it is the increased authority of the following acts which arguably vests the recommendation of the EBA with a higher caliber than usual recommendations. After all, a recalcitrant competent authority – if it does not manage to convince the EBA (the Commission) – has to fear a legally binding EBA decision

<sup>1549</sup> Article 6 para 2 subpara 5 of Regulation 1466/97. For further examples of reverse (qualified) majority voting see Palmstorfer, Majority 191–193.

<sup>1550</sup> Article 17 paras 3–6 of Regulation 1093/2010.

prevailing over any previous decision it has adopted on this matter.1551 Also potential reputational losses in the peer group (here: the EBA's Board of Supervisors) are to be taken into account. As *Peters* said (more generally): 'If there is a continuing long-term relationship among the participants in which they must interact, they are likely to comply'.<sup>1552</sup> In this (potential) mix of motives for compliance also loyalty considerations may come into play.

The reasons for compliance mentioned here do not root in any legal ob‐ ligation. The Council is (legally) free to refuse the Commission recommen‐ dation in the case of Article 143 TFEU and also where reverse (qualified) majority voting is applied, eg according to Article 6 para 2 of Regulation 1466/97, the Commission is (legally) free not to follow the national par‐ liaments' uttered view, the Board of Directors is (legally) free to decide against the opinion of the Management Committee, and the competent authority is (legally) free to refuse to follow an EBA recommendation and a formal opinion of the Commission, respectively. However, not to follow these acts is 'more difficult' than in a regular case. This is due to factual reasons: the difficulty to find an enhanced majority/unanimity against the recommendation, the (political) effort to justify non-compliance, the unreasonableness of disobeying the recommendation of the EBA which – in legal terms – holds the upper hand anyway.

The effects addressed here are factual, but to an extent also 'legal', be‐ cause they are provided for by law and hence intended by the Masters of the Treaties and the legislator, respectively, to work as deterrent against non-compliance. Due to this factual-legal ambiguity, they are referred to as 'mixed effects'.

<sup>1551</sup> Article 17 para 7 of Regulation 1093/2010.

<sup>1552</sup> Peters, Typology 426; see also Coen/Thatcher, Network 67; Haas, Hypotheses 34; for the involvement of national bodies in ensuring compliance with soft law on the international level see Goldmann, Gewalt 63 f.

### *5. The purposes of soft law*

### 5.1. On the categorisation of soft law in general

The primary purpose of soft law is to regulate.<sup>1553</sup> To aim at a high degree of compliance on the part of the addressees of regulation is inherent in this purpose. It cannot be assumed that a norm-creator intends to create a norm entirely without effect.<sup>1554</sup> In that respect there is no difference between law and soft law.<sup>1555</sup> The objectives of regulation are, in particular, to provide orientation and to ensure order and peace.<sup>1556</sup> A more nuanced approach may disclose a number of purposes, such as concretisation of law or harmonisation, <sup>1557</sup> which all reflect aspects of the primary purpose, ie to regulate. The French *Conseil d'État* in its comprehensive study 'Le droit souple'1558 has split this purpose in four functions of soft law: substitution (*substitut*), preparation (*préparation*), company (*accompagnement*) or per‐ manent alternative (*alternative pérenne*).<sup>1559</sup>

<sup>1553</sup> Considering the role and purpose of EU soft law on an international scale: Hop‐ kins/McNeill, Hard Law 115; with a view specifically to Commission recommenda‐ tions see Andone/Greco, Burden 84 f.

<sup>1554</sup> See Potacs, Auslegung 93: '[…] weil einem Rechtsetzer eine völlig wirkungslose Vorschrift prinzipiell nur schwer als von ihm gewollt zugesonnen werden kann' [because in principle it is difficult to imagine that a norm-creator has intended a provision entirely without effect].

<sup>1555</sup> This principal similarity is also reflected in Article 296 para 1 TFEU, according to which the institutions *principally* may choose from among the means of regulation – different forms of law and soft law – available; note the words of *Möllers* who said that 'as long as certain goals are achieved, it is irrelevant if this happens by use of legal forms or by informal means': C Möllers, Governance 316 f.

<sup>1556</sup> See, *ex multis*, Griller, Grundlagen (2015) 3–5.

<sup>1557</sup> That legally non-binding acts may cater for harmonisation – under EU law – is contested (see 5.2. below).

<sup>1558</sup> This study also refers to EU soft law. Apparently, it does not refer to the genuinely French 'droit souple' which actually is a specific kind of (hard) law; see Ballreich, Nachdenkliches 383 (fn 4); for the shift in the case law of the *Conseil d'État* this study may have brought about see Gundel, Rechtsschutz 600 f; taking up the categorisation of the *Conseil d'État* and applying it to soft law adopted by MS during the COVID-19 pandemic: Boschetti/Poli 40-44.

<sup>1559</sup> See Knauff, Regelungsverbund 378, distinguishing the following functions of soft law: preparation of law, company of law, replacement of law; Peters, Typology 421, who refers to EU soft law's function of 'complementing, supporting and interpreting primary and secondary Community law' as 'law-plus function'; for concrete examples from public international law see Shelton, Compliance 120 ff.

Substitution applies where, for legal or factual (eg political) reasons, proper law cannot be adopted,<sup>1560</sup> but a rule is still (deemed to be) nee‐ ded.<sup>1561</sup> The EU agency Eurofound has expressed this on its webpage in the following terms: 'In reality […] soft law in the EU tends to be used in situations where Member States are unable to agree on the use of a measure which is legally binding,[<sup>1562</sup>] or where the EU lacks competence to enact such a "hard law" measure. The Member States and EU institutions are thus able to adopt EU policy proposals, while leaving their implementation optional for those Member States who do not wish to be bound by manda‐ tory conditions. They are thus an option for the Commission to use when faced with resistance from some Member States, which could block policy proposals'.<sup>1563</sup>

An example for a preparatory function being fulfilled is where emergent phenomena which ought to be regulated (in the future) cannot yet be defined precisely enough to be regulated by law, which is why they are regulated by soft law (in preparation for law).1564 Alternatively, soft law can be used as a precursor for future legislation which – for the time being – is not possible, in the EU eg for lack of MS support.<sup>1565</sup>

<sup>1560</sup> For the adoption of soft law in order to circumvent the legislative or the ratification process see Bothe, Norms 94; Knauff, Regelungsverbund 251 and 254; Rossi, Soft Law 11.

<sup>1561</sup> See Bayne, Hard and Soft Law 348, with examples from EU law, 349 with an example from WTO law; Schermers/Blokker, Institutional Law, §§ 1233–1236, with examples from public international law.

<sup>1562</sup> See also Müller-Graff, Einführung 147, taking the example of the EU's European employment strategy and stressing that a soft approach may avoid differentiated integration.

<sup>1563 &</sup>lt;https://www.eurofound.europa.eu/observatories/eurwork/industrial-relatio ns-dictionary/soft-law> accessed 28 March 2023. At the international level, the establishment of the CSCE (and since 1994 the OSCE) on the basis of a soft law act exemplifies this phenomenon; see Zemanek, Soft law 858, with further references; see also Goldmann, Gewalt 393 f; for the particularly high authority of the CSCE, measured in terms of compliance with its output, see Shelton, Compliance 128 and 142. The renaming to OSCE did not lead to the emergence of the OSCE as an international organisation (ie a body with at least limited international legal personality); see Peters, Compact.

<sup>1564</sup> See Bayne, Hard and Soft Law 350, on the influence OECD soft law had on WTO negotiation rounds (and eventually on hard law).

<sup>1565</sup> See eg Council Recommendation 86/665/EEC, according to which the Council limits itself to a recommendation because 'it has so far proved difficult to elaborate a hotel grading system at Community level but it would nevertheless be desirable to consider the possibility of doing so in future'; for the development of the law on

Company describes essentially the concretisation of law, and the term '*alternative pérenne*' addresses cases for which regulation by means of soft law is thought to be the *better* – not the more opportune (see 'substitution' above) – alternative to regulation by law, eg for reasons of flexibility.<sup>1566</sup>

While this list of soft law functions – which, as was indicated above, can all be assigned to the purpose of regulation – in principle appears appropriate, it is difficult in practice to clearly distinguish 'substitution' from the 'permanent alternative', because the political conviction that soft regulation is the better option may often be nourished by the (presumed) impossibility to find a political accord on a binding measure or by the consideration of legal (constitutional) constraints. So while the theoretical difference between these two functions is apparent, in practice arguably they overlap considerably.

Another approach towards categorisation is to contrast the legal purpo‐ ses to regulate by means of soft law with the factual purposes. The purpo‐ ses which follow from the law are legal purposes. All other purposes are factual purposes. The range of factual purposes is broad and determined by motives as diverse as: the intention to exclude the EP from the deci‐ sion-making process (which would not be possible in a specific legislative procedure); the intention to react fast (faster than a legislative procedure would allow1567) to a certain problem in order to reduce pressure from the media; the assumption that a hard measure would give rise to criticism from political actors (eg in the MS). Also the legal purposes are manifold and shall only be exemplified here: compliance with the principle of subsid‐ iarity which may suggest the use of soft law instead of hard law (see V.3.4.2. below); considerations of effectiveness which may suggest, for example and in the short run, the adoption of a recommendation instead of a directive, the adoption of which would presumably take years; a competence to act, but a prohibition to harmonise laid down in primary law, may as well

access to (Commission) documents see Schwarze, Soft Law 244 f, with references to the relevant case law.

<sup>1566</sup> See Conseil d'État, Droit souple 136 f.

<sup>1567</sup> See eg Commission MEMO/14/484 <https://europa.eu/rapid/press-release\_ME MO-14-484\_en.htm> accessed 28 March 2023: 'a Commission Recommendation can be adopted immediately whereas proposals for legislation would have to be adopted by the EU's Council of Ministers and the European Parliament which can take time'.

suggest the use of soft law, eg a recommendation, to coordinate MS action 'without proceeding, explicitly at least, to harmonisation'.<sup>1568</sup>

It is to be noted that sometimes these two kinds of purposes – legal and factual ones – which are to be kept separate for conceptual reasons, intersect; eg when the prohibition to harmonise meets with the political conviction of the norm-creator that harmonisation is not opportune, and that therefore soft law is to be adopted, or in the above example related to considerations of the effectiveness of rule-making, which, without their legal edging, can as well be perceived as factual.

5.2. On the case of EU soft law in particular

5.2.1. Accepted purposes of EU soft law

The main purpose of soft law – that is to set rules and to achieve (volunta‐ ry) compliance with these rules – is most strongly dependent on its persua‐ siveness1569 in the concrete case.1570 The main instrument in that respect is the reasoning contained in each soft law act. Thereby also an account of the often highly complex scientific or technical facts underlying a certain matter is provided. Against this background, the informative function of the output of public administration in general and of soft law in particular is not to be underestimated ('regulation by information').<sup>1571</sup>

<sup>1568</sup> Senden, Soft Law 177; for examples see ibid 177 f; see also von Bogdandy/Arndt/ Bast, Instruments 113; suggesting that harmonization may also be brought about by soft law: case T-109/06 *Vodafone España*, paras 90 f.

<sup>1569</sup> For the distinction between command and persuasion, and the role soft law plays in between see Majone, Agencies 267–269.

<sup>1570</sup> See Rosas, Soft Law 318: 'the persuasive weight of different soft law sources cannot be determined on the basis of pre-determined lists but depends more on the nature and content of each instrument and the context in which it is being used'. An essential tool in this context is the reasoning of soft law acts. If the reasoning is not convincing (eg because it does not sufficiently address counter-arguments offered prior to the adoption of the act), it will be less successful in reaching compliance. With this established purpose of soft law in mind, the legal reasoning requirements should not be interpreted too laxly; see Andone/Greco, Burden 79.

<sup>1571</sup> For the importance of this notion, illustrated with regard to the ECB's announce‐ ment of OMT, see Tridimas/Xanthoulis, Analysis 18; critically: von Bogdan‐ dy/Goldmann, Ausübung 71 ff; with regard to EU competition law: H Hofmann, Rule-Making 169 f. Information may, without containing any norms, still influence human behaviour. This causes *Goldmann* to include information within its con‐

Whereas an EU soft law act is legally non-binding for third parties, it regularly is legally binding upon its creator(s). This self-obligation may also be listed as a (possible) purpose of EU soft law, and as a reason for its adoption respectively, as it ensures legal certainty and equal treatment of third parties.<sup>1572</sup> In this context, the then Court of First Instance held: 'Thus, simple guidelines or a simple communication – the Commission's compliance with which can be reviewed by the Community judicature – would have sufficed to guarantee the necessary transparency and legal certainty relating to the Commission's compliance with the obligations which it intended to impose on itself '.<sup>1573</sup> This self-binding effect, clearly a legal, not a factual purpose, allows soft law to fulfil a law-like function, which is particularly important where soft law is used as an alternative to law, either because the adoption of law is not possible (in the terminology of the *Conseil d'État* presented under 5.1. above: substitution) or because soft law is deemed to be more desirable (permanent alternative). But also where soft law is complementing EU law (see below), its self-bindingness is highly expedient. Less important this characteristic seems where soft law serves as a preparation of law, as a *lex ferenda* or *droit vert* and, as such, is used to facilitate a dialogue between the actors involved.1574 Here it is rather the sometimes educative function of soft law which comes into play.1575 This purpose has proved itself in practice (factual purpose) and at the same time it is reflected upon in all kinds of procedures laid down in EU law (legal purpose).<sup>1576</sup>

When soft law complements EU law, it mostly does so in the form of a concretisation or, what is similar, an interpretation of law. In the categorisa‐ tion of the *Conseil d'État*, this falls under the heading 'company', whereby

cept of 'authority' (to be distinguished from the concepts of 'law' and 'soft law', respectively); see Goldmann, Perspective 61 ff; similarly, for the possible overlap between information and policy-making: Majone, Agencies 264 f.

<sup>1572</sup> See 4.2.3.2.3. above.

<sup>1573</sup> Case T-240/04 *France v Commission*, para 42.

<sup>1574</sup> See Ingelse, Soft Law 77 f and 89 (for this terminology); Senden, Balance 92; Snyder, Effectiveness 33. For the 'separate life as a form of soft law' preliminary draft legal texts may develop: Chinkin, Development 26 f; Frykman/Mörth, Soft Law 155; Hofmann/Rowe/Türk, Administrative Law 543; for an example of draft soft law and its effects see Ştefan, Soft Law 124 f, with regard to a draft leniency notice of the Commission.

<sup>1575</sup> See Knauff, Regelungsverbund 250.

<sup>1576</sup> For the preparatory purpose of EU soft law see the initiation of (soft) decisionmaking/rule-making addressed in particular under 3.5.2. above.

an authentic or an executive (soft) interpretation is produced.<sup>1577</sup> Authentic it is where law creator and concretising/interpreting body are identical. Ex‐ ecutive it is where there is no such identity, but the concretising/interpret‐ ing body is, as an executive authority, involved in monitoring compliance with the concretised/interpreted law. Such concretisation – which often is undertaken only once the first experiences with the application of a legal act in administrative practice have been made – is often 'indispensable for national enforcers, especially where formal […] decisions do not sufficiently inform national decisional practice'.<sup>1578</sup> It allows the EU (and the MS) to go easy on law enforcement resources and 'individuals and Member States will be able to minimize [the risk] that their activities could be regarded as violation of [EU law] at a later stage in a Commission investigation'.<sup>1579</sup> *Hofmann* has, in this context, referred to Commission communications whose adoption has increased significantly since the mid-80s1580 and which may, *inter alia*, lead to a national administrative practice which is in com‐ pliance with EU law. This reduces the likelihood of the Commission having to initiate long-winded Treaty infringement procedures. <sup>1581</sup> While it is true that the vast majority of the infringement procedures is resolved prior to the case being referred to the CJEU, also the administrative phase of the Treaty infringement procedure can take a long time.<sup>1582</sup>

While the Commission is most active in the soft concretisation/interpre‐ tation of EU law, also other institutions, bodies, offices or agencies within their respective field of activity have complemented EU law in this way.<sup>1583</sup>

<sup>1577</sup> See the cautious early case law: case 74/69 *Hauptzollamt Bremen-Freihafen*, para 9: 'An unofficial interpretation of a regulation by an informal document of the Com‐ mission is not enough to confer on that interpretation an authentic Community character'; for the interpretation of an international agreement by means of vari‐ ous soft law acts adopted by the Commission and the Council see Müller/Slomin‐ ski, Role 877–881.

<sup>1578</sup> Georgieva, Soft Law 227.

<sup>1579</sup> Hofmann/Rowe/Türk, Administrative Law 567.

<sup>1580</sup> See Turgis, Communications 54.

<sup>1581</sup> See H Hofmann, Normenhierarchien 217, with further references; see eg the Commission Communication concerning the *Cassis* judgement; for other early examples of communications fulfilling a guidance function or an information function see Meier, Mitteilung 1303–1307; with regard to public procurement law see Lutz, Vergaberegime 897; see also case C-69/05 *Commission v Luxemburg*.

<sup>1582</sup> See Koops, Compliance 158.

<sup>1583</sup> See eg the CPVO's Guidelines with Explanatory Notes on Article 63 of Council Regulation (EC) 2100/94 of 27 July 1994 on Community plant variety rights. For one of the rare recommendations of the CJEU see Recommendations to national

In doing so, each EU actor 'doit trouver un juste équilibre entre l'obligati‐ on de ne pas créer d'obligations nouvelles sous peine d'annulation et la nécessité d'adopter un instrument ayant une valeur ajoutée'.<sup>1584</sup> If an inter‐ pretation of law suggested in a soft law act is followed, soft law – as a matter of course: only indirectly – 'participates' in the pivotal characteristics of EU law, that is direct effect and supremacy.<sup>1585</sup> It is needless to say that the CJEU has the final authority in matters of interpretation of EU law. This applies also with regard to EU soft law.<sup>1586</sup>

So far we have addressed the purposes of voluntarily adopted EU soft law. However, we also need to take into account that EU actors may be under an obligation to adopt soft law. Hence, it may also be one of the (legal) purposes of the adoption of soft law to meet an underlying obliga‐ tion. Such an obligation may be deduced from the principle of sincere cooperation pursuant to Article 4 para 3 TEU or Article 13 para 2 TEU.<sup>1587</sup> According to this principle, an institution may not impede another insti‐ tution in exercising its respective competences. For example: Where the Commission fails to adopt a recommendation in accordance with Article 126 para 7 TFEU, even though the Council has decided that an excessive deficit exists, the Council cannot adopt recommendations addressed to the MS concerned. It can only act upon a recommendation by the Commission which in the categorisation of the *Conseil d'État* has a preparatory function. The Commission, like all other institutions, bodies, offices and agencies, shall act in accordance with EU law. The Commission may therefore – in certain circumstances and acknowledging its discretion – be obliged to adopt a recommendation to the Council so as to enable it to adopt recommendations to the MS.1588 The Commission's (potential) obligation

courts and tribunals in relation to the initiation of preliminary ruling proceedings (2019).

<sup>1584</sup> Turgis, Communications 55 f.

<sup>1585</sup> See Sarmiento, Soft Law 270 f; see also Rosas, Soft Law 308.

<sup>1586</sup> See eg case T-73/98 *Prayon-Rupel*, para 71, in which the Court interpreted Com‐ mission soft law in a different way than the Commission itself; see also case C-526/14 *Kotnik*, paras 95 ff; case T-27/02 *Kronofrance*, para 79; differently, namely indicating the high authority of soft law for the interpretation of secondary law *in casu*, in case C-393/16 *Vin de Champagne*, para 45.

<sup>1587</sup> Stressing the Commission's duty to give guidance to MS authorities, eg in State aid law, on the basis of Article 4 para 3 TEU and (potentially) in the form of soft law: Opinion of AG *Wahl* in case C-526/14 *Kotnik*, para 83.

<sup>1588</sup> Critically with regard to the Commission's failure to act where there was need for a modification of the Council's recommendations: Häde, Aussetzung 763.

to adopt a recommendation is reflected in the competence of the Council to request the Commission to make a recommendation according to Article 135 TFEU.<sup>1589</sup> The Commission is not obliged to follow this request, but it shall examine it and 'submit its conclusions to the Council without delay'. What is more, the Council may launch an action with the CJEU, accusing the Commission of failure to act in infringement of the Treaties pursuant to Article 265 para 1 TFEU. That this provision also encompasses the failure to adopt a recommendation can be deduced *e contrario* from its para 3 which – unlike para 1, and only with regard to actions of natural or legal persons – explicitly excludes recommendations and opinions.<sup>1590</sup>

In conclusion, the categorisation of purposes of soft law presented under 5.1. above – the scheme of the *Conseil d'État* and the more basic distinction between legal and factual purposes – can reasonably be applied also in the given context, thereby revealing another method of structuring the large *corpus* of EU soft law.

# 5.2.2. Avoiding law as a purpose of EU soft law

Having addressed a variety of purposes of EU soft law, we shall now dwell on the purpose to avoid or even to evade law. The purposes of soft law and its 'hohe politische Attraktivität' [high political attractiveness] <sup>1591</sup> strongly rely on the experience that its use often facilitates an 'agreement' between parties which would not otherwise – in the form of law – have been conclu‐ ded.1592 Parties may not want to be legally bound for various reasons.<sup>1593</sup> Against this background, soft law may also serve as a way of regulation whereby an often complex and, from a political point of view, difficult law-making process is evaded.1594 The procedural complexity – which, of

<sup>1589</sup> For the EP's and the Council's self-standing and general power to request the Commission to make a proposal see Articles 225 and 241 TFEU; for the Commis‐ sion's power to adopt and withdraw a legislative proposal see also case C-409/13 *Council v Commission*, paras 70–74.

<sup>1590</sup> See also Dörr, Art. 265 AEUV, para 14.

<sup>1591</sup> Knauff, Regelungsverbund 248.

<sup>1592</sup> See Thomas Müller, Ziele 13.

<sup>1593</sup> See Ingelse, Soft Law 77; Terpan, Soft Law (2015) 89; Weismann, Bestimmung 381f.

<sup>1594</sup> See the introduction of the *Lamfalussy* regime in financial market law which does not only – at its level 2 – accelerate the creation of binding decisions, but also – at level 3 – caters for the adoption of soft law; for the procedural innovations brought

course, is not an end in itself1595 – in conjunction with the difficulty to ensure the agreement of the required majority in the legislative bodies often makes soft law, as *Gold* said in the context of public international law, 'the only alternative to anarchy'.<sup>1596</sup> With a view to a merely national or EU context we could, less dramatically, say: the only alternative to an unregulated situation.

It need, however, not always be the complexity of the legislative proce‐ dure, but it is often also its exclusivity which causes the norm-creator to opt for the soft law road. Where the involvement of private actors in the decision-making process is – for whichever reason – (deemed to be) required, a procedure leading to the adoption of soft law may be chosen due to its less strictly regulated and hence *potentially* more inclusive, open, participatory character.<sup>1597</sup> In these specific cases soft law is actually consid‐ ered to be the better option than law (permanent alternative). This entails a certain ambivalence which *Georgieva* has aptly expressed in the context of competition soft law: 'Thus, competition soft law portrays an intriguing dichotomy. While attempting to provide democratic values such as clarity, certainty, and participation,[] competition soft law in an increasingly com‐ plex policy setup[] simultaneously erodes those same values because of its non-justiciability'.<sup>1598</sup>

Another reason why soft law may in places be deemed more opportune than law is the outright lack of or at least uncertainty about the existence of a competence to adopt a legal act.1599 It is then adopted 'faute de mieux', as *Schwarze* put it.1600 While, as was clarified above, also the adoption of an EU soft law act must rest on an adequate legal basis, the requirements for such a legal basis are regularly lower than in the case of law and, what is more, soft law is less likely to be brought before and scrutinised by the

1596 Gold, Soft International Law 444.

about by the *Lamfalussy* regime see eg Weismann, Agencies 81–97; for the legal and factual qualification of the level 3 output see Arndt, Sinn 74–78.

<sup>1595</sup> For the impact the chosen procedure (ie the chosen legal basis) has on its outcome see eg case C-62/88 *Greece v Council*, para 10; Opinion of AG *Tesauro* in case C-300/89 *Commission v Council*, para 2.

<sup>1597</sup> See Knauff, Regelungsverbund 251; for the importance of inclusivity in the drafting of private standards see Roht-Arriaza, Compliance 209 f.

<sup>1598</sup> Georgieva, Soft Law 228.

<sup>1599</sup> See Lafarge, Coopération 75 f; Senden, Soft Law 169, with regard to Commission recommendations; differently with regard to Council recommendations: ibid 177. This reason is acknowledged also by the Court in case C-322/88 *Grimaldi*, para 13.

<sup>1600</sup> Schwarze, Soft Law 238.

Court (see in particular 6.2. below). That way, soft law may effectively regulate – and survive – even if it is unlawful, thereby contributing to what *Majone* called 'integration by stealth'.<sup>1601</sup>

The avoidance of law, be it in order to make use of a more simple and/or inclusive procedure, be it in order to limit the risk of an act being challenged before the Court, may constitute an abuse of the chosen type of action.<sup>1602</sup> Such moves also have a democratic thrust, insofar as the avoided or even circumvented legislative procedures often carry more democratic participation (even if in a more formalised way) and hence also more dem‐ ocratic legitimacy than (many of ) the procedures leading to the adoption of soft law.<sup>1603</sup> Having said that, there may also be legitimate reasons for the avoidance of law, such as preventing a political deadlock in legislative negotiations or allowing for a more inclusive decision-making procedure. Where these legitimate reasons are particularly strong, taking the soft law route may be justified. However, the decision in favour of soft law should not be taken too lightheadedly, in particular where law-making procedures laid down in primary law are avoided.

While soft law may be a welcome alternative to law in places, suffice it to briefly mention here that in other cases it may turn out to be politically inappropriate and to require 'hardening'.1604 While the reasons for this preference given to law over soft law lie in the concrete circumstances of

<sup>1601</sup> Majone, Dilemmas.

<sup>1602</sup> See generally on this topic Biervert, Mißbrauch; see also joined cases 8–11/66 *Cimenteries*, 92; with regard to non-binding public international 'law' see Bothe, Norms 94.

<sup>1603</sup> See Frykman/Mörth, Soft Law 155; Meijers Committee, Note 1; see also Arndt, Sinn 185; Dawson, Soft Law 8; Senden, Soft Law 172. For early complaints by the EP and the French *Conseil d'État* see Ştefan, Enforcement 215; for substitution dynamics in EU company law see Lutter, Empfehlungen 799, stressing the 'hohe[n] Charme' [great charm] for the Commission that it does not need to involve the legislator when adopting the recommendations at issue; critically: D Lehmkuhl, Government 150.

<sup>1604</sup> See the words of Commissioner *Charlie McCreevy* on potential follow-up action to a Commission recommendation on the cross-border management of copyright for legitimate online music services: '[I]f I am not satisfied that sufficient progress is being made, I will take tougher action'; quoted in European Parliament (Commit‐ tee on Legal Affairs), Working Document on institutional and legal implications of the use of 'soft law' instruments (14 February 2007), PE 384.581v02–00, 5, including critical remarks on this case.

### *III. FUNDAMENTAL QUESTIONS OF EU SOFT LAW*

the cases at issue,1605 more generally speaking, and in spite of the increasing importance of soft law, it appears that the implementation of a given EU policy by means of law still in the majority of cases is more effective than soft regulation.

### *6. Judicial review of soft law*

# 6.1. Introduction

The question whether and, if so, to which extent a soft law act can be made subject to legal review is of eminent importance for the legal protection of those (negatively) affected by the act, and – more generally speaking – the question whether a soft law act can be scrutinised by a Court is highly rele‐ vant from a rule of law and from a democracy perspective.1606 While there is a number of possibilities that a soft law act be examined (eg by the EP according to Article 226 TFEU, by the European Ombudsman according to Article 228 para 1 TFEU,1607 or by an agency-internal Board of Appeal1608), the focus here shall be on judicial review (in a broad sense) by the CJEU. In this context, essentially two procedures are to be mentioned: the annulment procedure (Article 263 TFEU) and the preliminary reference procedure (Article 267 TFEU). Other procedures will be briefly addressed thereafter.

<sup>1605</sup> With regard to the originally soft, and subsequently 'hardened', regulation of credit rating agencies see Ferran/Alexander, Soft Law Bodies 760. Another example is Article 13 para 3 subpara 2 of the ESM-Treaty which declares binding for its MoU certain opinions, warnings and recommendations addressed to the respective MS. For those who argue in favour of a general legalisation see references in Knauff, Regelungsverbund 19 f.

<sup>1606</sup> See eg Thomas Müller, Soft Law 114.

<sup>1607</sup> See eg Senden/van den Brink, Checks 58 ff; Vianello, Approach.

<sup>1608</sup> Soft law cannot normally be challenged before Boards of Appeal. However, if the binding act under review was preceded/prepared by a soft law act, an indirect consideration of soft law by the respective Board of Appeal may be feasible; with regard to the ESAs' Joint Board of Appeal and the EP's suggestion to broaden the scope of its review powers so as to include soft law see Chamon/Fromage, Added Value 21; for some kind of scrutiny which the Commission may exercise with regard to the ESAs' soft law see Article 60a of Regulations 1093–1095/2010; see on this ibid 30 f.

### 6.2. The annulment procedure

As regards the annulment procedure, Article 263 para 1 TFEU sets out that it encompasses the review of the legality of 'legislative acts, of acts of the Council, of the Commission and of the European Central Bank, other than recommendations and opinions, and of acts of the European Parliament and of the European Council intended to produce legal effects *vis-à-vis* third parties'. The latter limitation applies also to acts of bodies, offices or agencies of the EU, which the Court is competent to review according to Article 263 TFEU, as well. The Court does not appear to make a difference between these two groups of excluded acts – recommendations and opin‐ ions on the one hand, and acts not producing legal effects *vis-à-vis* third parties, on the other hand. Instead – and with regard to all acts adopted by the mentioned EU actors which are brought before it – it concentrates on the (intended1609) legal effects of an act *vis-à-vis* third parties – 'whatever their nature and form'.1610 This only means that nature and form may be trumped by the (otherwise) established intention of the creator of the act. Overall, it is in particular the wording and the context which are to be included in the assessment.<sup>1611</sup>

In the context of Article 263 TFEU, 'legal effects *vis-à-vis* third parties' appear to be understood as synonymous with external legal bindingness.<sup>1612</sup>

<sup>1609</sup> The phrase 'intended to produce legal effects *vis-à-vis* third parties' was introduced by the Treaty of Maastricht. While the Treaty has been using it only in the context of the EP (and, since the Treaty of Lisbon, the European Council and other EU bodies), the Court has increasingly applied the alternatives – (actual) 'legal effects' or 'intented to produce legal effects' (*telos*) – to examine whether or not an action for annulment against an act is admissible; see references in Dörr, Art. 263 AEUV, para 41.

<sup>1610</sup> The landmark decision in this context is case 22/70 *Commission v Council*, para 42 (from where the quote is taken); see also later case law: case C-362/08P *Hilfsfonds*, para 55, with a further reference; case T-258/06 *Germany v Commission*, paras 29–31, referring to form, nature and wording (which may all indicate the existence or lack of legal effects), but stressing that these factors are only to be considered among others.

<sup>1611</sup> See case C-443/97 *Spain v Commission*, paras 34–36; case T-496/11 *United King‐ dom v European Central Bank*, para 31, both with further references; for the focus on the appearance for the addressee rather than the intention of the creator in this case see Türk, Liability 45.

<sup>1612</sup> See case C-31/13P *Hungary v Commission*, paras 54 f, with further references; see also Thomas Müller, Soft Law 118. Merely internal acts may not be subject to an action for annulment; see case T-236/00R *Stauner*, para 43, with further references. One important exception are decisions in staff matters which normally

In the words of the Court: This 'legal effect' required under Article 263 TFEU by no means 'relates to any legal effect, irrespective of its nature', but it is the 'binding nature' which matters.<sup>1613</sup> Where, on the contrary, an act is 'only proposing a course of conduct […] [and hence] similar to a mere recommendation within the meaning of Article 288 TFEU […] it should be concluded that the act does not have legal effects that are such as to render an action for annulment brought against it admissible'.1614 That nonbinding acts may affect the interests of individuals, for example in the form of what the Court refers to as 'purely implementing measures', does not render them reviewable under Article 263 TFEU. Such implementing meas‐ ures are, in particular, 'measures which, without giving rise to any rights or obligations for third parties, are designed merely to put into practical effect an earlier measure, or measures adopted in order to implement earlier deci‐ sions which produce only internal legal effects within the administration and do not affect the interests of third parties'.1615 The Court also explicitly determined that 'preparatory act[s] or intermediate measure[s]'1616 as well as 'confirmatory measures, […] mere recommendations and opinions and,


also concern only the internal organisation of a body, but which may be reviewed, nevertheless; see further references in W Cremer, Art. 263 AEUV, para 20. Also in case of (intended) legal effects *vis-à-vis* third countries, the Court in principle confirms the admissibility of an action under Article 263 TFEU against the respec‐ tive act; for the refusal of admissibility see eg case T-670/14 *Milchindustrie*, with regard to Commission Guidelines on state aid for environmental protection and energy 2014–2020. For a critique of the focus on legal bindingness and arguments in favour of a wider understanding of 'legal effects' see Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, paras 94 and 109 ff.

<sup>1613</sup> Case C‑689/19P *VodafoneZiggo*, paras 56–58; also consider the wording in case C‑431/20P *Tognoli*, para 33: 'any measures […] which are intended to have bind‐ ing legal effects, are regarded as acts open to challenge, within the meaning of Article 263 TFEU'. Legal effects below this threshold do not suffice; see eg ibid, para 53, and case C-687/15 *Commission v Council*, para 54; with regard to legally non-binding 'intermediate measures whose purpose is to prepare for the final decision' see joined cases C‑551/19P and C‑552/19P *ABLV*, para 39, with further references. *De lege ferenda* arguing, along the lines of liberal constitutionalism, in favour of a more open understanding of Article 263 TFEU: Gentile, Review; see also Ştefan/Petri, Review 533; see also the differentiated approach of AG *Hogan* in his Opinion in case C-572/18P *thyssenkrupp*, para 70.

<sup>1614</sup> Case T-496/11 *United Kingdom v European Central Bank*, para 32; see also case C‑370/07 *Commission v Council*, para 42, with further references; case C‑689/19P *VodafoneZiggo*, paras 51–53.

in principle, internal instructions'<sup>1617</sup> fall outside the scope of Article 263 TFEU.

A peculiar issue are acts which appear to be soft law at first sight, but turn out to be law (*prima facie* soft law). The general hazardousness of *prima facie* soft law acts was referred to by AG *Tesauro*. According to him, they 'give rise to confusion and uncertainties amongst its addressees, be they Member States or individuals, as to whether the conduct contemplated by it is obligatory. Manifestly, this is detrimental, not only to individuals, but also to the administration'.<sup>1618</sup> Confronted with such acts, the Court has first unmasked them as acts intended to have or actually having 'legal effects' and eventually annulled them (eg for not having been adopted on an adequate legal basis).<sup>1619</sup>

Due to the rather material question of whether an act has 'legal effects', the examination of admissibility thereby becomes intertwined with ques‐ tions of substance.1620 It is the substance of an act which is decisive also in the context of admissibility, not its name.1621 The Court expressed this in the following way: 'In order to determine whether the contested act produces binding legal effects, it is necessary to examine the substance of that act and to assess those effects on the basis of objective criteria, such as the content of that act, taking into account, as appropriate, the context in

<sup>1617</sup> Case T-721/14 *Belgium v Commission*, para 17.

<sup>1618</sup> Opinion of AG *Tesauro* in case C-325/91 *France v Commission*, para 21.

<sup>1619</sup> See eg case C-57/95 *France v Commission*; case C-16/16P *Belgium v Commission*, paras 29–31; case T-561/14 *One of Us*, para 83 (confirmed in appeal case C-418/18P *Puppinck*); see also Opinion of AG *La Pergola* in case C-443/97 *Spain v Com‐ mission*, paras 22 and 27, proposing the annulment of 'internal guidelines' of the Commission (The Court did not follow this proposal.); see H Hofmann, Rule-Making 176, with further references; Senden, Soft Law 149, referring to a case in which the Commission refrained from making the legal non-bindingness of its Communication explicit – contrary to a plea of the EP. For the possibility of only partial annulment of *prima facie* soft law acts see arguments by Pampel, Rechtsnatur 128 f.

<sup>1620</sup> See eg case C-325/91 *France v Commission*, para 11; for further references see H Hofmann, Rule-Making 176.

<sup>1621</sup> See case C‑322/88 *Grimaldi*, para 8, in which the Court, with regard to recom‐ mendations, stresses 'that acts *in the nature of* recommendations' (emphasis add‐ ed) are excluded from scrutiny under what is now Article 263 TFEU; see also case T-721/14 *Belgium v Commission*, para 66; see Opinion of AG *Tesauro* in case C-366/88 *France v Commission*, para 6. Still stressing the importance of 'style [ie title; see the German version of the judgement: 'Bezeichnung'] and form' of an act: joined cases 90 and 91/63 *Commission v Luxembourg*, 631; see also Dörr, Art. 263 AEUV, para 42; Raschauer, Leitlinien 37 f, both with further references.

which it was adopted and the powers of the institution which adopted the act'.<sup>1622</sup>

It may occur that a soft law act is indirectly scrutinised in the course of an annulment procedure. Where a soft law act *content-wise* is the main basis of an act having legal effects (eg because a regulation essentially takes over the content of the recommendation which has initiated the decisionmaking procedure or because a decision applies an interpretation of EU law which is suggested by guidelines), it is likely that the Court, in the course of an annulment procedure against the latter act, indirectly also considers, ie scrutinises, the soft law act.<sup>1623</sup> In the context of a scientific opinion adopted by an EU committee, the Court held that '[a]lthough the opinion does not bind the Commission, it is none the less extremely important so that any unlawfulness of that opinion must be regarded as a breach of essential procedural requirements rendering the Commission's decision unlawful'.1624 While a scientific opinion does not necessarily contain rules and thus may not qualify as soft law, we may still conclude from these words that preparatory acts (including soft law) will be taken into account when the resulting act is examined.

The case law also contains contested judgements which, against the backdrop of the above case law, seem to be non-system. One example is the decision of the General Court, taken on the basis of an annulment action filed by a privileged claimant, namely the UK as a MS, in which it (partially) annulled the ECB's Eurosystem Oversight Policy Framework, according to conventional wisdom a legally non-binding act.<sup>1625</sup>

In conclusion we can say that Article 263 TFEU, as interpreted by the Court, does not allow for the annulment of a true soft law act – for lack of (intended) legal effects and irrespective of whether it is called recommenda‐

<sup>1622</sup> Case C-16/16P *Belgium v Commission*, para 32; case C‑310/21P *Aquind*, para 50; see also case C-911/19 *FBF*, para 38; joined cases C‑551/19P and C‑552/19P *ABLV*, para 41, both with further references.

<sup>1623</sup> See case 60/81 *IBM*, para 12; case T-326/99 *Olivieri*, paras 50 and 55; case T-671/15 *E-Control*, para 81.

<sup>1624</sup> Joined cases T-74/00, T-76/00, T-83/00, T-84/00, T-85/00, T-132/00, T-137/00 and T-141/00 *Artegodan*, para 197.

<sup>1625</sup> Case T-496/11 *United Kingdom v European Central Bank*, in particular para 84; see discussion by Alberti, Evolution 644–647; for the exceptional annulment pro‐ cedure against a recommendation in case C-27/04 *Commission v Council* see Häde, Aussetzung 757 f.

tion, opinion or otherwise.<sup>1626</sup> More progressive approaches, as uttered eg by AG *Bobek*, <sup>1627</sup> so far have not been taken up by the Court.<sup>1628</sup>

# 6.3. The preliminary reference procedure

Quantitatively speaking, the preliminary reference procedure is the most important procedure for the assessment of EU law by the CJEU. The pred‐ ecessor provisions of Article 267 TFEU, apart from the interpretation of the Treaty, only referred to the validity check and interpretation of acts of the institutions of the EC (and of the ECB), and to the interpretation of the statutes of certain bodies established by the Council respectively.<sup>1629</sup> Considering this exclusive wording, the Court hesitated to deal with acts from bodies not explicitly mentioned in Article 177/234 of the Treaty.<sup>1630</sup> Starting in the early 90s, the Court has gradually developed a more gen‐ erous approach in this respect.1631 With the Lisbon reform, the scope of creators of eligible acts under the preliminary reference procedure was extended to 'bodies, offices or agencies' of the EU. In the *Elliott* case, the Court broadened the ambit of Article 267 TFEU further by even accepting to interpret acts 'which, while indeed adopted by bodies which cannot be described as "institutions, bodies, offices or agencies of the Union", are by their nature measures implementing or applying an act of EU law'.<sup>1632</sup>

When it comes to the required effects of eligible acts, Article 267 TFEU and its predecessors have been more encompassing. Unlike Article 263 TFEU and its predecessors, the preliminary reference regime does not provide for a limitation to acts producing legal effects *vis-à-vis* third parties.

<sup>1626</sup> For actions for annulment against 'explanatory' acts which go beyond the norma‐ tive content of the legal act to be 'explained' see 3.3.3.1. above.

<sup>1627</sup> Opinion of AG *Bobek* in case C-16/16P *Belgium v Commission*, para 4.

<sup>1628</sup> At the national level, courts show an increasing willingness to review soft law; see Eliantonio, Review 292–299.

<sup>1629</sup> Article 177 TEEC; Article 234 TEC.

<sup>1630</sup> See case C‑322/88 *Grimaldi*, para 8; case C-11/05 *Friesland Coberco*, paras 36 ff; with regard to the Court's approach towards soft law adopted by the Commission together with the MS see Eliantonio, Soft Law 505, with a further reference.

<sup>1631</sup> See eg case C-188/91 *Deutsche Shell*.

<sup>1632</sup> Case C-613/14 *Elliott*, para 34, with further references. This wording seems to be excessively broad, as also MS acts could implement or apply an act of EU law; with regard to output adopted in the framework of the OMC: Knauff, Regelungs‐ verbund 512.

Accordingly, the Court has for a long time interpreted soft law acts adopted by the institutions under the preliminary reference procedure.<sup>1633</sup> With regard to the interpretation of acts, in general the Court appears to take a liberal, a 'flexible approach'.<sup>1634</sup> The case of the ECB's (mere) press release announcing the Outright Monetary Transactions Programme (OMT) is a good example for this.<sup>1635</sup> While an action of private parties under Article 263 TFEU was refused as inadmissible,<sup>1636</sup> in the *Gauweiler* case the Court has accepted to provide a preliminary ruling on the press release, namely on its interpretation.<sup>1637</sup> While the Court's interpretation of a soft law act uttered in the course of a preliminary reference procedure is binding,<sup>1638</sup> the act itself remains to be non-binding.<sup>1639</sup> Where the national court or tribunal decides to apply it, it is bound by the CJEU's interpretation.<sup>1640</sup>

With regard to the second prong of the preliminary reference procedure, the validity of acts other than the Treaties, the Court's jurisdiction with regard to soft law has not received broad attention in the case law for quite some time.1641 In view of Article 263 TFEU, as interpreted by the Court, and

<sup>1633</sup> See eg cases 113/75 *Frecassetti*, para 8; C-188/91 *Deutsche Shell*, para 18; C-42/99 *Queijo Eru*, paras 20 ff; C-101/08 *Audiolux*, among others para 46; C-526/14 *Kot‐ nik*, para 33; T-109/06 *Vodafone España*, para 102.

<sup>1634</sup> Opinion of AG *Campos Sánchez-Bordona* in case C‑613/14 *Elliott*, para 61, with references to the Court's case law. For recent indications of a rebound effect in the Court's approach see Wahl/Prete, Gatekeepers.

<sup>1635</sup> Whether this press release constitutes a soft law act is disputed. For lack of a rule, ie of normative content, the author would say it is not; differently: Alberti, Evolu‐ tion 632, thereby pointing to the specificities of this press release as compared to usual EU soft law; for the question of liability see Türk, Liability 43; see also case T‑192/16 *NF*, para 42, specifically referring to press releases.

<sup>1636</sup> See case C-64/14P *European Central Bank v von Storch*.

<sup>1637</sup> See case C-62/14 *Gauweiler*, para 28. This concessive approach follows from the Court's reliance on the 'direct knowledge of the facts giving rise to the dispute and […] [its] responsibility for the subsequent judicial decision' of the referring court; see also case C-112/00 *Schmidberger*, para 31.

<sup>1638</sup> See eg case C-346/93 *Kleinwort Benson*, para 24.

<sup>1639</sup> See Georgieva, Soft Law 231–233, with further references. For the possibly success‐ ful approach for individuals to make EU soft law subject to national proceedings and then – via Article 267 TFEU (requiring a request of the national court) – to bring it to the attention of the CJEU see Eliantonio, Soft Law 513.

<sup>1640</sup> See eg joined cases C-120/06P and C-121/06P *FIAMM*, paras 123 f.

<sup>1641</sup> Case C-322/88 *Grimaldi*, para 8, according to which, in very general terms, it may 'give a preliminary ruling on the validity and interpretation of all acts of the institutions of the Community without exception'; more specific (and denying the scrutiny of validity of a soft law act not stemming from an institution): case C-11/05 *Friesland Coberco*, paras 36-41; see Opinion of AG *Bobek* in case C-16/16P

its exclusion of true soft law acts, ruling on the validity of EU soft law under Article 267 TFEU could be considered as an (unlawful) evasion of the requirements laid down for the annulment procedure.<sup>1642</sup> This may be one of the reasons why the Court's approach had been rather reserved in this respect,<sup>1643</sup> even if it did not outright refuse the possibility of examining the validity of soft law.<sup>1644</sup> AG *Bobek* rather observed a tendency to convert questions on the validity of soft law into questions of interpretation.<sup>1645</sup> In recent judgements, however, the Court was quite explicit in this respect. In *Belgium v Commission* the Court clearly held that 'even though Article 263 TFEU excludes the review, by the Court, of acts which are in the form of recommendations, Article 267 TFEU confers on the Court jurisdiction to deliver a preliminary ruling on the validity and interpretation of all acts of the EU institutions without exception'.<sup>1646</sup> This line of argumentation was confirmed by subsequent case law, among other things leading to the partial


*Belgium v Commission*, paras 106–108, suggesting the possibility of a preliminary reference to assess also the *validity* of a recommendation; for the lack of clarity of the concept of validity in the context of EU soft law see Knauff, Soft Law 738 f.

invalidation of an EBA recommendation for wrongful legal conclusions made therein.<sup>1647</sup> The standard of review applied for the examination of soft law does not appear to be principally different from that referred to in the examination of law.<sup>1648</sup> The effect of the invalidity of an (interpretative) soft law provision is that the actor in charge 'must not take [this soft law] into consideration when interpreting EU law'.<sup>1649</sup> That is how AG *Campos Sánchez-Bordona* put it in the *Balgarska Narodna Banka* case, and the Court does not seem to have contradicted him.

### 6.4. Other procedures

EU soft law may also be (indirectly) reviewed by the CJEU in the course of other procedures. According to Article 340 para 2 TFEU, the Union shall make good any damage caused 'by its institutions or by its servants in the performance of their duties'.1650 The illegality of the damaging behaviour must be 'sufficiently serious' for there to be a claim under Article 340 paras 2 f TFEU.1651 Where the EU body concerned disposes of discretion, this means that only a 'manifest[] and grave[] disregard[] [of ] the limits on [this] discretion' can lead to a damages claim.<sup>1652</sup>

Already in light of this requirement, it can be doubted that non-compli‐ ance with EU soft law by the named actors may lead to a successful claim

<sup>1647</sup> See case C‑501/18 *Balgarska Narodna Banka*, paras 98-101.

<sup>1648</sup> See case C‑501/18 *Balgarska Narodna Banka*, para 83; case C-911/19 *FBF*, para 57 (examination of validity) and 67 (standard of review); for the relationship between the two judgements, and for a discussion of the latter, see Chamon/de Arriba-Sellier, Justiciability; for a critique of the Court not actually sticking to the standard of review it has announced to apply see ibid 308–313; for the question of whether the (exceptional) duty to make a preliminary reference pursuant to the *Foto-Frost* case law also applies in the context of soft law see Scholz, Soft law 457, with further references.

<sup>1649</sup> Opinion of AG *Campos Sánchez-Bordona* in case C‑501/18 *Balgarska Narodna Banka*, para 120.

<sup>1650</sup> For the ECB see para 3 *leg cit*; for European agencies see, for example, Articles 3 and 39 of Regulation 2019/1149 establishing a European Labour Authority.

<sup>1651</sup> For the original limitation of this qualification to normative illegality and its grad‐ ual extension to executive and judicial illegality see I Augsberg, Art. 340 AEUV, paras 51–53; for exceptional cases not requiring illegality see ibid, paras 82 ff.

<sup>1652</sup> Case C-352/98P *Bergaderm*, para 43.

for damages.<sup>1653</sup> Since soft law in principle, leaving the binding effect on its respective creator apart,<sup>1654</sup> is legally non-binding, a deviation from the prescribed behaviour can by definition – and given that this behaviour is not, in addition to that, legally prescribed elsewhere<sup>1655</sup> – not be illegal, let alone constitute a sufficiently serious breach of Union law. This holds true also with regard to the Treaty infringement procedure according to Articles 258–260 TFEU, which does not concern the behaviour of EU bodies, but in which a MS' alleged failure 'to fulfil an *obligation* under the Treaties' (em‐ phasis added) is at issue.<sup>1656</sup> The situation is different – and generally more complicated – when compliance with a soft law act leads to a damage (and an action for damages) or is at issue in a Treaty infringement procedure.<sup>1657</sup>

As regards Article 340 paras 2 f TFEU, relevant scenarios may not only be (non-)compliance with (unlawful) soft law, but the damaging behaviour may also itself take the form of soft law.1658 In the *Arizmendi* case the then Court of First Instance, in the context of a reasoned opinion adopted by the Commission in a Treaty infringement procedure, stated that 'it cannot be precluded that in very exceptional circumstances a person may be able to demonstrate that such a reasoned opinion is vitiated with illegality con‐

<sup>1653</sup> Similarly: Arroyo Jiménez, Bindingness 26 f; Ştefan, Soft Law 241, with regard to State liability (which also requires there to be a 'sufficiently serious breach'). In other legal orders, non-compliance with non-binding norms may – exceptionally – lead to damages claims; see eg Arndt, Sinn 174–176.

<sup>1654</sup> See 4.2.3.2.3. above. For the legitimate expectations soft law may create see 4.2.2.2.4. above. The violation of legitimate expectations may give rise to a damages claim.

<sup>1655</sup> See also Sarmiento, Soft Law 278–280. If a legal provision, as interpreted by a soft law act, is deemed to be violated, an action for damages may be successful. That way, the violation of soft law may – indirectly – result in the confirmation of a damages claim; see (in the context of State liability) case C‑501/18 *Balgarska Narodna Banka*, para 81; see also Arroyo Jiménez, Bindingness 20.

<sup>1656</sup> See also case C-69/05 *Commission v Luxemburg*. For the 'different purposes' and 'different conditions' of the annulment procedure on the one hand, and the Treaty infringement procedure on the other hand see case C-16/16P *Belgium v Commis‐ sion*, para 40; see also Wörner, Verhaltenssteuerungsformen 510 f.

<sup>1657</sup> The Court's rather restrictive approach can be deduced from its general case law (which is not related to damages claims, though); see case 133/79 *Sucrimex*, paras 20–23; case T-54/96 *Oleifici Italiani*, para 67; case T‑585/14 *Slovenia v Commis‐ sion*, para 44. See also J Hofmann, Protection 464 f, with further references. With regard to private addressees of EU soft law see von Graevenitz, Mitteilungen 173.

<sup>1658</sup> With regard to acts of public international (soft) law, namely the MoU adopted by the Commission on behalf of the ESM, see joined cases C-8–10/15P *Ledra*, para 55.

stituting a sufficiently serious breach of a rule of law that is likely to cause damage to him',<sup>1659</sup> and subsequently held:

'The fact that a reasoned opinion adopted by the Commission under the first paragraph of Article 226 EC is not a measure intended to produce binding legal effects with respect to third parties and that, accordingly, that opinion is not a measure capable of forming the subject matter of an action for annulment […] does not affect the preceding assessment. A reasoned opinion may, owing to its unlawful content, cause harm to third parties. Thus, for example, it cannot be precluded that the Commission should cause harm to persons who have entrusted it with confidential information by disclosing that information in a reasoned opinion. Like‐ wise, it cannot be precluded that a reasoned opinion should contain inaccurate information about certain persons likely to cause them harm' (emphasis in original).<sup>1660</sup>

The wrongful disclosure of information is not the only way in which soft law can cause a damage. It may also be the (soft) legal substance of the act which can be made subject to a damages claim.<sup>1661</sup>

In our context, the main question is whether a claim for damages or a Treaty infringement procedure may lead to a (direct or indirect) *review of* (the legality of ) *soft law* by the Court. From the above we can conclude that in the course of the procedure following an action for damages, an EU soft law act may in fact be scrutinised. In the course of a Treaty infringement procedure this may principally be the case, but – since an EU soft law act does not impose obligations on the MS – in practice is less likely.1662 There is, however, one soft law act which will always be considered (indirectly), if the Court renders a judgement: the content of the Commission's reasoned

<sup>1659</sup> Joined cases T-440/03, T-121/04, T-171/04, T-208/04, T-365/04 and T-484/04 *Arizmendi*, para 68.

<sup>1660</sup> Joined cases T-440/03, T-121/04, T-171/04, T-208/04, T-365/04 and T-484/04 *Arizmendi*, para 69; case T-107/17 *Steinhoff*, paras 55-57, with many further refer‐ ences; case T-868/16 *QI*, para 71; see also Senden, Soft Law 465, with reference to case law.

<sup>1661</sup> See case T-107/17 *Steinhoff*, para 57.

<sup>1662</sup> Soft law may be invoked by a MS to justify its behaviour, though, which could lead to a consideration of this act by the Court; see case C-342/05 *Commission v Finland*, paras 29 ff.

opinion pursuant to Article 258 TFEU, to the extent that it is repeated in the Commission's action.<sup>1663</sup>

Apart from that, Article 277 TFEU allows for an *incidenter* review of acts of general application, which may lead to a declaration of inapplicability in the case at issue. This review may be performed in other procedures such as the aforementioned annulment procedure, the preliminary reference proce‐ dure, or the procedure on an (alleged) failure to act (Article 265 TFEU).<sup>1664</sup> The Court has specified that under Article 277 TFEU not only regulations, but all 'acts of the institutions [in a broader sense] which, although they are not in the form of a regulation, nevertheless produce similar effects' can be challenged.<sup>1665</sup> Thus, also soft law acts of general application seem to fall within this category.<sup>1666</sup>

1664 Article 277 TFEU applies to all procedures before the CJEU; see eg Stoll/Rigod, Art. 277 AEUV, para 6. For the possibilities to act against an EU body's failure to adopt a soft law act see Article 265 TFEU (privileged claimants). A non-privileged claimant, however, is limited to complain about the failure to address to him/her an act other than a recommendation or an opinion; critically of the Court's strict approach: J Scott, Limbo 344 ff and 349 ff (with reform proposals); see also references in Eliantonio/Ştefan, Soft Law 464.

1666 In its original version, this provision explicitly excluded recommendations and opinions; see Article 173 TEEC (Rome). The current version – Article 277 TFEU – refers to Article 263 TFEU instead, which also speaks against the inclusion of soft law – or rather: in favour of an inclusion of (unlawful) soft law (of a general application) only to the extent it is challengeable under Article 263 TFEU. However, the term 'similar effects' in the Court's case law appears to be rather malleable. The Court seems to be more concerned about a 'direct legal connection' of the soft law act to the act challenged in the main procedure, which it was ready to confirm in a number of cases; see eg joined cases C-189, 202, 205, 208 and 213/02P *Dansk Rørindustri*, para 237; case T-23/99 *LR AF 1998*, paras 274–276; cases T‑394/08, T‑408/08, T‑453/08 and T‑454/08 *Sardegna*, paras 206–210, all with regard to Commission guidelines; for the CJEU's readiness to include acts adopted by European agencies see Ehricke, Art. 277 AEUV, para 11, with further references.

<sup>1663</sup> After all, it is the action ('application' according to Article 120 of the Rules of Procedure of the CJEU) initiating the Court procedure, not the reasoned opinion.

<sup>1665</sup> Case 92/78 *Simmenthal*, para 40.

https://doi.org/10.5771/9783748935865, am 02.04.2024, 08:01:05 Open Access – - https://www.nomos-elibrary.de/agb

# IV. MECHANISMS IN EU LAW TO ENSURE LEGAL COMPLIANCE OF MEMBER STATES

#### *1. Introduction*

'The European Union is a community of law, based on common values shared by Member States. Applying and enforcing EU law, and respect for the rule of law are at its very foundation'.<sup>1667</sup> These words of the Commis‐ sion address the application and enforcement of EU law which lead us to a new focus of this book. Having presented, analysed and more generally discussed fundamental questions of EU soft law, we shall now explore various mechanisms which are aimed at ensuring MS' compliance with EU law. The term 'Member States' is to be understood broadly here, as used eg in the context of the Treaty infringement procedure.1668 Therefore, also pro‐ cedures performed *vis-à-vis* certain national authorities shall fall within the ambit of this Part.1669 As regards the term (non-)compliance, *Young's* rather intuitive definition can be applied: 'Compliance can be said to occur when the actual behavior of a given subject conforms to prescribed behavior, and non-compliance or violation occurs when actual behavior departs sig‐ nificantly from prescribed behavior'.1670 It is an objective understanding of compliance which shall be applied here. A cause-effect relationship between the norm and the behaviour is not required.<sup>1671</sup>

There is a vast body of literature on the question why States comply with their (international) obligations.1672 From a theoretical perspective, es‐

1671 See Nollkaemper, Role 160.

<sup>1667</sup> Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 1.

<sup>1668</sup> See Andersen, Enforcement 60 f, with references to the case law; see Lenaerts/Ma‐ selis/Gutman, Procedural Law 143 f, with further references.

<sup>1669</sup> This comprehensive understanding of the term 'MS' does not render irrelevant whether a certain MS or a certain national authority is addressed by an EU body in the course of a compliance mechanism; see also V.3.6. below.

<sup>1670</sup> Young, Compliance 3; for a similar definition as behaviour 'consistent with (inter‐ national) norms and rules' see Börzel, Noncompliance 14, with further references; in case of specific legal acts, eg directives, the question of compliance may encom‐ pass various aspects: see eg Falkner/Treib/Hartlapp/Leiber, Europe 11–14.

<sup>1672</sup> See also the theoretical overview by Conant, Compliance.

sentially we can distinguish three schools of thought on this question which are reflected upon in realist, institutionalist and normative theories. Realist theories are based on the conviction that States follow their obligations if they have an interest (national interest) in doing so (eg liberal intergovern‐ mentalism). Institutionalist theories stress the importance of institutions established by a group of States (eg neo-functionalism). These institutions can affect State behaviour – not only by the norms they create as such, but in particular by the institutions implementing them, thereby increasing or reducing the incentives of States to comply or not to comply by referring to State reputation, monitoring schemes, cooperation etc. Normative theories put more emphasis on compliance as a result of States (State actors) sens‐ ing a moral obligation to comply.<sup>1673</sup> With regard to the EU, interestingly, research on the implementation of Union law by the MS has begun only in the mid-80s.1674 Even the Commission as the 'guardian of the Treaties' for a long time has sidelined the implementation dimension of Community and later Union regulation, laying emphasis on this issue only since the 90s.<sup>1675</sup> The intention here is not to find out why MS comply with EU law,1676 but to present and analyse how EU law ensures – by means of certain procedures (compliance mechanisms) – that MS comply with it. Thus, Parts IV and V shall focus on some of the legal tools intended to ensure MS compliance and on their legality, respectively.

While also general-abstract measures – such as generally applicable guidelines or instructions – may aim at ensuring compliance with EU law, namely by concretising it, the following chapter is dedicated to acts which are addressed to one or more MS, covering a concrete case (individ‐ ual-concrete measures). The former generally are intended to define their addressees' (including the MS) behaviour *ex ante*, <sup>1677</sup> thereby determining the law to be applied in the future, the latter take effect *ex post*, that is to say they provide for a *reaction* to MS action. The Commission as the core actor in this context is often vested with both tasks – to concretise EU law via general-abstract (legally binding or legally non-binding) measures

<sup>1673</sup> For these theories see Burgstaller, Theories 95 ff.

<sup>1674</sup> See Falkner/Treib/Hartlapp/Leiber, Europe 14 f.

<sup>1675</sup> See Mastenbroek, Compliance 1104. *Mastenbroek* in this piece also provides for an overview of research on compliance with Community/Union law.

<sup>1676</sup> For an overview of the (mainly political science) literature on implementation of and compliance with EU law and its historical development see Treib, Governance.

<sup>1677</sup> Only exceptionally – that is where they have retroactive effect – they function also *ex post*.

on the one hand, and to ensure compliance with EU law (possibly as interpreted in its general-abstract measures), on the other hand.<sup>1678</sup> In the context of the latter procedures, a number of different terms are used (not uniformly), most importantly 'control', 'enforcement', 'implementation', 'su‐ pervision'.<sup>1679</sup> The Commission, for example, mainly seems to use the term 'enforcement' when referring to infringement procedures to be initiated by it under the Treaties.<sup>1680</sup> While acknowledging the partly different, partly overlapping meanings of these terms, here the broader term 'compliance' shall be used. The necessary distinction between 'implementation' and 'enforcement' (two categories substantially rooting in the Treaties) shall follow only after the presentation of the mechanisms (see V.3.1. below).

Many procedures laid down in EU law directly or indirectly aim at MS compliance,<sup>1681</sup> but the focus here shall be narrowed to mechanisms which are (at least partly) governed by *administrative* EU actors and which are aimed at ensuring compliance by the MS – by means of EU law or EU soft law – with concrete rules or objectives, as laid down in law or soft law (compliance mechanisms).1682 While in Part III all kinds of EU soft law acts – those with a general-abstract and those with an individual-con‐ crete scope – have been considered, in the given context it is acts with an individual-concrete scope which are at issue. This is why in particular the following procedures shall not be addressed here: the instrument of State liability which is enforced before national courts; the mechanism laid down in Article 7 TEU or the three-stage process proclaimed by

<sup>1678</sup> See J Scott, Limbo 331f, pointing at 'the Commission's dual role in elaborating guidance and in identifying and prosecuting a breach of the underlying framework norm'.

<sup>1679</sup> For the terms 'supervision' and 'control' see Audretsch, Supervision 4; for these terms in German scholarship see Eekhoff, Verbundaufsicht 5–7, with further references; see also Weiler, Supranational Law 413 f, addressing – as a third cat‐ egory next to implementation and enforcement – 'application'; with regard to 'enforcement' and 'management' as two alternative perspectives on compliance see Tallberg, Paths 609 f.

<sup>1680</sup> See eg Commission, 'Completing the Internal Market', COM(85) 310 final, para 125, or <https://ec.europa.eu/environment/legal/law/14/Introduction\_EU\_Enviro nmental\_Law.htm> accessed 28 March 2023; applying a broader understanding of 'enforcement': Scholten, Trend.

<sup>1681</sup> See the different lines of development relating to ensuring compliance with EU law drawn by Chiti, Governance.

<sup>1682</sup> For different categories of compliance mechanisms in public international law see Kingsbury, Concept 64.

the Commission1683 (highly political rather than 'technical' and applied only exceptionally) which are about compliance with values rather than concrete rules;<sup>1684</sup> the preliminary reference procedure pursuant to Article 267 TFEU which is entirely judicial,<sup>1685</sup> on the part of the EU performed only by the CJEU;<sup>1686</sup> or temporary mechanisms such as the Cooperation


<sup>1683</sup> See Commission, 'A new EU Framework to strengthen the Rule of Law' (Commu‐ nication), COM(2014) 158 final, 7 ff; see also III.3.5.2.2.3. above.

and Verification Mechanism for Bulgaria and Romania<sup>1687</sup> . <sup>1688</sup> Neither shall the procedure laid down in Article 37 of the ESM-Treaty (which is not laid down in EU law<sup>1689</sup>) or the procedure pursuant to Article 105 TFEU, according to which the Commission shall – in cooperation with the MS authorities – investigate cases of suspected infringements (by undertakings) of the principles laid down in Articles 101f TFEU, be presented here.<sup>1690</sup> Also tools not providing for the creation of law or soft law do not fall within the research focus at issue here. An example is the Internal Market Scoreboard (IMS) which contributes – *inter alia*, but not primarily – to improving compliance with EU law by the MS by publishing reports on the implementation of EU law in the MS.<sup>1691</sup> Procedures by means of which EU law is enforced *vis-à-vis* individuals or undertakings (eg EU competition law as enforced by the Commission) in general fall outside the scope of this work.<sup>1692</sup>

mechanism which shows parallels to the preliminary reference procedure see Article 5 of Decision ECB/2004/3 on public access to ECB documents, according to which certain documents 'may be disclosed by the NCB only subject to prior consultation of the ECB concerning the scope of access, unless it is clear that the document shall or shall not be disclosed. Alternatively the NCB may refer the request to the ECB'.

<sup>1687</sup> See <https://commission.europa.eu/strategy-and-policy/policies/justice-and-fun damental-rights/upholding-rule-law/rule-law/assistance-bulgaria-and-romania-u nder-cvm/cooperation-and-verification-mechanism-bulgaria-and-romania\_en> accessed 28 March 2023; see also von Bogdandy**/**Ioannidis, Deficiency 85–87.

<sup>1688</sup> Excluded from this survey is also the mechanism laid down in Article 38 TEU in which a Political and Security Committee shall, among other things, monitor the implementation of agreed policies. The decisions it may take upon an authori‐ sation pursuant to para 3 *leg cit* also concern a broadly formulated matter, namely political control and strategic direction of the operation at issue.

<sup>1689</sup> Heed the (meanwhile failed) Commission Proposal COM(2017)827, which sug‐ gests to transform the ESM into a European Monetary Fund (EMF) and which – if it were adopted – would have done away with this mechanism; for another compli‐ ance mechanism based on public international law, but to be performed by an EU body, see Article 10 para 2 of the Agreement on the transfer and mutualisation of contributions to the Single Resolution Fund.

<sup>1690</sup> The appropriate measures according to para 1 *leg cit* are regularly proposed to un‐ dertakings. The authorisation of MS by the Commission is, technically speaking, not primarily aimed at ensuring compliance of MS authorities, but at making it possible for MS to act in the first place; see Ludwigs, Art. 105 AEUV, paras 1–4 and 9 f, with regard to the low practical importance of this procedure and claiming a *de facto* obligatory effect on MS authorities of the Commission authorisation just mentioned.

<sup>1691</sup> See Koops, Compliance 160–164.

The most prominent of the compliance mechanisms at issue here is cer‐ tainly the Treaty infringement procedure, but there is also a large number of further mechanisms essentially serving the same aim (within specific policy fields), particularly, but not only in secondary law. These mecha‐ nisms have mainly been introduced as an addition to the Treaty infringe‐ ment procedure. The latter procedure has been considered dysfunctional in many cases as 'not always adequate to ensure compliance with the relevant Community provisions particularly at a stage when infringements can be corrected'.<sup>1693</sup> It is due to the large number of compliance mechanisms laid down in secondary law that the subsequent presentation cannot be exhaustive.

In this Part, selected compliance mechanisms shall be sorted in terms of their rough legal origin (primary law or secondary law), and in terms of their use of law and/or soft law *vis-à-vis* the MS, respectively. For this reason, the compliance mechanisms are assigned to either of these three categories: hard mechanisms (exclusively providing for acts of law), mixed mechanisms (providing for both acts of law and acts of soft law) and soft mechanisms (exclusively providing for soft law acts). Since the procedures shall be perceived from a MS perspective, the indicator for either category is the acts *addressed to MS*, not, for example, a Commission recommenda‐ tion addressed to the Council which is a requirement for the adoption of a Council decision addressed to a MS. These acts addressed to EU bodies will be mentioned for the sake of completeness, though.

The structure shall be as follows: After an account of the general compli‐ ance mechanism of the EU – the Treaty infringement procedure – with a focus on the acts addressed to a MS in the course of this procedure, special

<sup>1692</sup> Only where such a procedure is contained in a compliance mechanism as defined above – in particular as an *ultima ratio* in case the competent national authorities do not comply with EU law even upon request by the Union body in charge – it shall be included in the discussion. Contrasting these indirect mechanisms ('mon‐ itoring the enforcement efforts of national authorities') with 'direct enforcement', that is to say ensuring compliance with EU law by private actors: Scholten/Lucht‐ man/Schmidt, Proliferation 1 and 5 ff.

<sup>1693</sup> Preamble to Council Directive 89/665/EEC; see also, with reference to this pas‐ sage, case C-433/93 *Commission v Germany*, para 23. For the room for improve‐ ment in the resolution of alleged infringements on the part of the MS see also the Report to the EEC Commission by the High Level Group on the Operation of Internal Market, presided over by then former Commissioner *Peter Sutherland* (1992; so-called 'Sutherland-Report'), in particular Recommendations 20, 27 and 31; for the variation of non-compliance with EU law per policy field see Börzel, Noncompliance 140 ff.

compliance mechanisms – that is to say mechanisms aimed at ensuring compliance within a specific field of EU law – shall be addressed in accord‐ ance with the above categorisation. Following a condensed presentation of each mechanism, the sequence and the legal quality of acts addressed to a MS in the course of the respective procedure shall be recapitulated and analysed briefly. The findings with respect to each category shall be conflated in summaries.

Part IV as such aims at providing a broad, but still exemplary account of the different compliance mechanisms laid down in EU law, thereby providing, as it were, the raw material for the comprehensive analysis of the selected compliance mechanisms which is proffered in Part V under the heading 'classification and legal assessment of compliance mechanisms'.

### *2. The mechanisms in detail*

2.1. The general compliance mechanism: the Treaty infringement procedure

### 2.1.1. Introduction

The central and general (ie not restricted to specific fields of EU law) mechanism for ensuring compliance of MS with EU law is the Treaty infringement procedure. Apart from the sanctions regime now provided for in Article 260 para 2 TFEU and the fast-track procedure according to para 3 *leg cit*, it has been laid down in the founding treaty of the E(E)C (and now: the EU) ever since its inception.1694 Exceptions apart,1695 it now also applies to the EAC. <sup>1696</sup> The ECSC had its own procedure essentially governed by the High Authority.1697 With the Treaty of Maastricht, the sanctioning

<sup>1694</sup> See Articles 169–171 TEEC.

<sup>1695</sup> Eg Article 38 para 3 TEAC which constitutes a fast-track procedure comparable to Article 260 para 3 TFEU, but without the immediate possibility to impose sanctions.

<sup>1696</sup> Article 106a TEAC. For specific infringement procedures laid down in the histori‐ cal versions of the TEAC see Schermers/Waelbroeck, Protection, § 1215.

<sup>1697</sup> Articles 86–88 TECSC. A decision of the High Authority could be made subject to review by the Court. Counter-measures could be applied by the High Authority 'with the concurrence of the Council acting by a 2/3 majority'. For a comparison between this procedure and the Treaty infringement procedure see Ionescu, Wir‐ kungen 271–273. In the past 20 years, the Commission has repeatedly demanded

regime just mentioned was introduced in the then renamed TEC.1698 In addition to that, since the Treaty of Lisbon the fast-track procedure for failure to notify MS measures transposing a (legislative) directive according to Article 260 para 3 TFEU (as already laid down in the Constitutional Treaty) applies (see in more detail 2.1.2. below).<sup>1699</sup> The Commission, as 'guardian of the Treaties'<sup>1700</sup> the EU authority responsible for ensuring com‐ pliance under Articles 258 and 260 TFEU (together with the Court), has a wide discretion on whether or not it launches an infringement procedure, be it on the basis of Article 258, be it (at the next procedural level) on the basis of Article 260 TFEU.<sup>1701</sup> Up until the early 2000s, it had not published any guidance as to which cases it would concentrate on.<sup>1702</sup> Only in its 2002 Communication entitled 'Better Monitoring of the Application of Community Law' it slightly limited its leeway by setting out its priorities for the initiation of infringement proceedings.<sup>1703</sup>

Up until the mid-70s, the meaning of the Treaty infringement procedure in the EEC had been limited (around 30 procedures per year). From 1977 onwards, it had strongly increased, with over 500 initiated proce‐ dures in 1985, and reached its peak in 2004 with nearly 3,000 procedures


1702 See Smith, Dialogue 553.

similar powers in the EC/EU context; see Andersen, Enforcement 124 f, with further references; for support in the literature see Prete, Infringement 369-378.

<sup>1698</sup> For the Court's proposal to introduce a financial sanctions regime and for the strong backing it got from the UK see Kilbey, Penalties 744–746.

<sup>1701</sup> The High Authority under the TECSC, on the contrary, was obliged to pursue infringements; see Andersen, Enforcement 131.

<sup>1703</sup> See Commission, 'Better Monitoring of the Application of Community Law' (Communication), COM(2002) 725 final, 11f. Starting in 2017, the Commission now applies a new prioritisation policy with regard to MS' infringements of EU law; see Commission Communication 'EU law: Better results through better application', 2017/C 18/02; Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 20. For the handling of complaints from civil society and businesses see Commission, 'Updating the handling of relations with the complainant in respect of the application of Union law' (Communication), COM(2012) 154 final; also note the existence of related regimes such as SOLVIT; for the performance of SOLVIT in practice see Hobolth/Sindbjerg Martinsen, Networks; Börzel, Noncompliance 117; for the Commission's Communications related to its powers under what is now Article 260 para 2 TFEU see Prete, Infringement 257–259; see also the recent Communication 'Financial sanctions in infringement proceedings', 2023/C 2/01, in which the Commission reviews its previous Communications on this matter.

launched.<sup>1704</sup> Since then, the number has decreased considerably, not least due to the introduction of EU Pilot in 2008 (see 2.1.2. below). In 2022, the Commission initiated only 551 procedures.<sup>1705</sup>

The sanctioning regime was applied for the first time in a case against Greece in which the Court rendered its judgement in 2000, seven years after the introduction of the scheme.<sup>1706</sup> For several years, the sanctioning of MS has been rather exceptional<sup>1707</sup> and critics have perceived 'a lack of enthusiasm on the part of the Commission and the Court to make effective use of Article 260(2) TFEU'.<sup>1708</sup> In recent years, it has been, if at all, an annual handful of cases – seven in 2020,<sup>1709</sup> none in 2021,<sup>1710</sup> and one in 2022,<sup>1711</sup> respectively – in which the Court has imposed financial sanctions on MS.<sup>1712</sup> On the whole, the Treaty infringement procedure has been criticised as cumbersome, time-consuming and unfit to solve compliance problems in everday administration.1713 Having said that, with 'about 90 %' the compliance rate reached in the pre-litigation phase of infringement procedures actually initiated is very high.<sup>1714</sup>


<sup>1704</sup> See Karpenstein, Art. 258 AEUV, para 8; see graph by Börzel/Knoll, Non-compli‐ ance 10.

<sup>1705</sup> See <https://commission.europa.eu/law/law-making-process/applying-eu-law/i nfringement-procedure/2022-annual-report-monitoring-application-eu-law\_en> accessed 28 March 2023.

# 2.1.2. The procedure in short

Where the Commission 'considers that a [MS] has failed to fulfil an obli‐ gation under the Treaties,[<sup>1715</sup>] it shall deliver a reasoned opinion on the matter', after it has given the respective MS the 'opportunity to submit its observations'.<sup>1716</sup> This opportunity in practice is given by a so-called letter of formal notice – in the words of the Commission a 'request for information'<sup>1717</sup> – which 'comprises an initial succinct résumé of the alleged infringement'.<sup>1718</sup> The MS is normally given two months to react to the letter. The letter of formal notice and the reasoned opinion, together 'de‐ limit the subject matter of the dispute',<sup>1719</sup> but the formal notice already circumscribes the charges contained in a (possible<sup>1720</sup>) reasoned opinion, and the Commission may normally neither extend nor alter the scope of the complaint in its reasoned opinion.1721 While the Commission output

<sup>1715</sup> A list of what all the term 'obligation under the Treaties' encompasses is, with references to the relevant case law, provided by Andersen, Enforcement 46.

<sup>1716</sup> Article 258 para 1 TFEU.

<sup>1717 &</sup>lt;http://europa.eu/rapid/press-release\_MEMO-12-12\_en.htm> accessed 28 March 2023.

<sup>1718</sup> Lenaerts/Maselis/Gutman, Procedural Law 189. Exceptionally, the Commission may issue a second, supplementary letter of formal notice; see eg case C-10/10 *Commission v Austria*, para 11. For the purpose of the formal notice see also case 211/81 *Commission v Denmark*, para 8: '[A] letter giving formal notice is intended to delimit the subject-matter of the dispute and to indicate to the Member State which is invited to submit its observations the factors enabling it to prepare its defence'.

<sup>1719</sup> Case C-280/02 *Commission v France*, para 29. With regard to the reasoned opin‐ ion, the Court held that '[i]f a charge was not included in [there], it is inadmissible at the stage of proceedings before the Court'; case C-305/03 *Commission v United Kingdom*, para 22, with further references.

<sup>1720</sup> The Commission – in spite of the wording of Article 258 para 1 TFEU: 'shall' – is not obliged to adopt a reasoned opinion after a non-satisfactory reaction to a letter of formal notice on the part of the MS concerned; see Andersen, Enforcement 49; Gormley, Infringement 65, both with further references; see also Müller/Slomin‐ ski, Role 874 f, with further references; for the (limits of ) the Commission's discre‐ tion see European Ombudsman, Decision on complaint 995/98/OV concerning the Macedonian Metro Joint Venture (2001) paras 1.6 – 1.9 <https://www.ombuds man.europa.eu/mt/decision/en/1088> accessed 28 March 2023.

<sup>1721</sup> See case C-280/02 *Commission v France*, paras 29 f, with further references. This applies with one exception: The Commission may very well limit the scope: see Craig/de Búrca, EU Law 473, with references to the case law. New evidence to underpin the charges may be brought forward also later. For the possible extension of the scope of the reasoned opinion to events occurring after the reasoned opin‐ ion was adopted see Andersen, Enforcement 50, with further references; see also

adopted in the course of a Treaty infringement remains confidential until a judgement is handed down by the CJEU,<sup>1722</sup> the fact that a reasoned opinion was issued (not: its exact content) is normally disseminated to the public early on in the form of a press release, also to increase the pressure on the MS concerned.<sup>1723</sup> After all, the pre-litigation phase is coined by the 'quasi-diplomatischen' [quasi-diplomatic]<sup>1724</sup> efforts to convince the other side.<sup>1725</sup> It reflects the respectful and considerate approach<sup>1726</sup> towards sov‐ ereign States which is known from traditional public international law.<sup>1727</sup> Most Treaty infringement cases are settled in the pre-litigation phase (see 2.1.1. above).<sup>1728</sup> In this context, also 'EU Pilot' is to be mentioned. Operative

1724 Wegener, Art. 271 AEUV, para 5.

1728 Each year, several hundreds of new infringement procedures face up to only low two-digit numbers of Court judgements under Article 258 TFEU; see, for example, the general statistical overview of the Commission's Annual Report for 2021 on

case T-258/06 *Germany v Commission*, para 152, with a further reference. For an example of a complementary reasoned opinion launched by the Commission see the procedure against Hungary relating to its Higher Education Law: <http://euro pa.eu/rapid/press-release\_MEMO-17-3494\_en.htm> accessed 28 March 2023.

<sup>1722</sup> See Craig/de Búrca, EU Law 474 f, with references to the case law; see also Article 4 para 2 (second and third indent) of Regulation 1049/2001; for the duration of confidentiality in case of the third indent see case T-194/04 *Bavarian Lager*, para 148.

<sup>1723</sup> See eg <http://europa.eu/rapid/press-release\_IP-17-3186\_en.htm> accessed 28 March 2023; see also Falkner/Treib/Hartlapp/Leiber, Europe 207.

<sup>1725</sup> In the words of the Court, this phase 'is to enable the Member State to comply of its own accord with the requirements of the Treaty or, if appropriate, to justify its position'; case C-159/94 *Commission v France*, para 103, with reference to further case law.

<sup>1726</sup> Note the words of AG *Roemer* in his Opinion in case 7/71 *Commission v France*, 1026: 'Moreover, this procedure naturally puts in issue to a certain extent the prestige of the Member State concerned, even though it is merely an objective procedure intended to clarify the legal situation, without any moral judgment. For these reasons it seems proper to rule out any automatic application, any compul‐ sion to initiate it and instead to leave to the Commission a discretionary power to decide whether and when the procedure should be initiated. Many different factors may come into play in this respect, for example, attempts to reach an amicable settlement (which may take time) or the fact that [the alleged infringement] had only relatively slight effects that did not justify judicial proceedings'; for the 'more ad hoc than systematic' contacts between Commission and MS in the pre-litiga‐ tion phase until the late 80s see Gormley, Infringement 66; see also Audretsch, Supervision 17.

<sup>1727</sup> For the differences as compared to traditional public international law and for a critique of the pre-Maastricht set-up of the Treaty infringement procedure see Weiler, Transformation 2419 f.

since 2008, it is not reflected in any way in the Treaties. It provides for a procedure prior to the pre-litigation phase ('"pre-pre-litigation" phase'<sup>1729</sup>), which is divided in different steps with different deadlines, thereby using an online database and communication tool, with a view to solving the (suspected) non-compliance.<sup>1730</sup> In spite of a positive evaluation of 'EU Pilot', the *Juncker* Commission decided to largely abolish it and avail itself of this 'lengthy step' only where 'recourse to EU Pilot is seen as useful in a given case'.<sup>1731</sup> A broader application of this tool was again envisaged under President *von der Leyen*. 1732

Coming back to the procedure pursuant to primary law, the reasoned opinion deserves closer attention. It is an act of EU soft law, as, without being legally binding,<sup>1733</sup> it 'formal[ly] request[s] to comply with [it, and thereby also with] EU law'.<sup>1734</sup> If the MS concerned does not comply with the reasoned opinion within a certain period laid down by the Commission (normally it is again two months1735), the Commission 'may bring the

<sup>&#</sup>x27;Monitoring the application of European Union law', COM(2022) 344 final, 27 and 31; see also Andersen, Enforcement 52.

<sup>1729</sup> Andersen, Enforcement 47. For the pre-litigation tool 'SOLVIT' aimed at support‐ ing EU citizens/undertakings which are denied certain mobility-related rights in another MS, and its reform respectively, see Commission, 'EU law: Better results through better application' (Communication), 2017/C 18/02, 16 f; see also Koops, Compliance 164–168.

<sup>1730</sup> See <http://ec.europa.eu/internal\_market/scoreboard/\_archives/2014/07/perform ance\_by\_governance\_tool/eu\_pilot/index\_en.htm> accessed 28 March 2023.

<sup>1731</sup> Commission, 'EU law: Better results through better application' (Communica‐ tion), 2017/C 18/02, 12; critically: Wendenburg/Reichert, Vertragsverletzungsver‐ fahren 1344 f.

<sup>1732</sup> See Prete/Smulders, Age 300; see also V.2.5.3. below.

<sup>1733</sup> The legal non-bindingness of the Commission's reasoned opinion has been con‐ firmed by the Court on a number of occasions; see eg case C-191/95 *Commission v Germany*, paras 44–46. In para 46 of this judgement the Court specifies that '[t]he reasoned opinion therefore has legal effect only in relation to the commencement of proceedings before the Court […] so that where a Member State does not comply with that opinion within the period allowed, the Commission has the right, but not the duty, to commence proceedings before the Court'; see Audretsch, Supervision 25, referring to literature arguing in favour of a binding effect of the Commission's reasoned opinion.

<sup>1734 &</sup>lt;http://europa.eu/rapid/press-release\_MEMO-12-12\_en.htm> accessed 28 March 2023; see also Gil Ibáñez, Supervision 95 f.

<sup>1735 &</sup>lt;https://ec.europa.eu/info/law/law-making-process/applying-eu-law/infringemen t-procedure\_en> accessed 28 March 2023.

matter before the Court'.1736 With these scarce words, Article 258 TFEU – which has essentially remained unchanged over the past 60 years<sup>1737</sup> – sets out the transition from the so-called pre-litigation phase to the litigation phase. The action must be based on 'the same objections' as the Commis‐ sion's reasoned opinion.<sup>1738</sup> This does not render the reasoned opinion enforceable, though, as it is the (alleged) failure to fulfil an obligation under the Treaties which is at issue in the Court proceedings (not: the failure to comply with the reasoned opinion), and, in connection therewith, it may turn out that, in the view of the CJEU, the MS has in fact complied with its obligations under the Treaties, and that therefore the reasoned opinion is legally wrong.<sup>1739</sup>

Excursus<sup>1740</sup>: Article 259 TFEU addresses the competence of a MS to bring an action before the CJEU against another MS for failure to fulfil an obligation under the Treaties. This is a remnant of the public interna‐ tional legal origin of what is now the EU: that parties to an agreement control each other with a view to compliance with that agreement. Before doing so, under Article 259 TFEU a MS shall 'bring the matter before the Commission'. Also in this procedure, the Commission shall, having given

<sup>1736</sup> The Commission is not obliged to take the next step – here: filing an action – immediately after the period set has expired. In a case in which the Commission filed an action only eleven months after expiry of the two months period laid down in its reasoned opinion, the Court held that 'it is for the Commission to choose when it will bring an action for failure to fulfil obligations before the Court'; case C-350/08 *Commission v Lithuania*, paras 29 and 33.

<sup>1737</sup> See Smith, Evolution 352.

<sup>1738</sup> Case C-11/95 *Commission v Belgium*, para 73; case C-422/05 *Commission v Bel‐ gium*, para 25, both with references to further case law.

<sup>1739</sup> The Court's superiority *vis-à-vis* the Commission under what is now Article 258 TFEU is reflected in settled case law: '[T]he Commission is not empowered to determine conclusively, by opinions formulated pursuant to Article 169 or by other statements of its attitude under that procedure, the rights and duties of a Member State or to afford that State guarantees concerning the compatibility of a given line of conduct with the Treaty. According to the system embodied in Articles 169 to 171 of the Treaty, the rights and duties of Member States may be determined and their conduct appraised only by a judgment of the Court'; joined cases 142/80 and 143/80 *Amministrazione delle Finanze dello Stato*, para 16. This was confirmed by the Court on many occasions, eg in case C-135/01 *Commission v Germany*, para 24.

<sup>1740</sup> Since the topic of this chapter are procedures allowing EU actors to ensure compli‐ ance *vis-à-vis* MS, at first sight Article 259 TFEU does not appear to belong here. Since it may lead to the Commission initiating a procedure pursuant to Article 258 TFEU, though, it shall be outlined here as a brief excursus. For the reasons for the low importance of Article 259 TFEU in practice see Tallberg, Paths 615 f.

the MS concerned the opportunity to express their view ('orally and in writing'<sup>1741</sup>), deliver a reasoned opinion. The reasoned opinion here plays a different role than under Article 258. Since the Commission is not a party to the procedure, it is – as an impartial actor bound only by the law – not obliged to support the view of the suspicious MS.<sup>1742</sup> It may as well adopt a negative opinion where it deems the allegations unjustified.<sup>1743</sup> Following the reasoned opinion, and regardless of its content, the suspicious MS may file an action with the CJEU.<sup>1744</sup> Where the Commission fails to deliver such an opinion within three months from being approached by the MS suspecting an infringement of EU law, the latter may file an action with the Court, as well. To put it short: The MS suspecting an infringement of EU law must first address the Commission. After it has delivered a reasoned opinion or where it has failed to do so within three months, this MS may approach the CJEU.

The action of the Commission pursuant to Article 258 or of a MS pur‐ suant to Article 259 is binding to the extent that it specifies the matter of the proceedings before the Court. However, an action does not commit the Court to take a certain decision, in view of the general contradictory character of Court proceedings not even in a legally non-binding way. Thus, its content does not qualify as soft law. Once an action is filed, with the Commission (or the MS) bearing the burden of proof for the infringement,1745 the Court considers the case. Where it deems the action to be at least partly justified, the MS 'shall be required to take the necessary measures to comply with the judgment of the Court'.<sup>1746</sup>

<sup>1741</sup> Note that Article 258 TFEU does not explicitly provide for the possibility of an oral utterance.

<sup>1742</sup> See Eberhard, Art. 259 AEUV, para 33; Karpenstein, Art. 259 AEUV, paras 11f, both with further references

<sup>1743</sup> See eg case C-364/10 *Hungary v Slovakia*, para 19. Thus, under Article 259 TFEU, the reasoned opinion constitutes a legal assessment of the MS's allegations which may be positive or negative – with regard to the allegations raised by the MS. Where it is positive, it softly requires the other MS to adapt its behaviour ac‐ cordingly. Where the opinion is negative, it (softly) confirms the legality of the (non-)action at issue and the other MS is not asked to change its behaviour. In either case it states – in a legally non-binding way – what the law is in a concrete case, and hence qualifies as soft law; see also Karpenstein, Art. 259 AEUV, para 15.

<sup>1744</sup> See Eberhard, Art. 259 AEUV, para 23; Wunderlich, Art. 259 AEUV, para 11, both with further references.

<sup>1745</sup> See case C-494/01 *Commission v Ireland*, paras 41f, for the standard of proof and for the MS' duty to cooperate, each with further references.

<sup>1746</sup> Article 260 para 1 TFEU.

Where the Commission – not: a MS suspecting an infringement of EU law within the meaning of Article 259<sup>1747</sup> – considers that the MS has not done so, it may, having provided the MS with the opportunity to utter its view, bring the case before the Court according to Article 260 para 2 TFEU. This is the (potentially applied) follow-up part of the Treaty infringement procedure. It is essentially about the infringement of a Court judgement launched in the course of a procedure according to Article 258 or Article 259 TFEU. Since the Treaty – more precisely: Article 260 para 1 TFEU – obliges the MS addressed to comply with the judgement, it is not wrong to also call Article 260 para 2 TFEU (part of ) the Treaty infringement proce‐ dure.<sup>1748</sup> As under Article 258 TFEU, the pre-litigation procedure officially starts with a letter of formal notice by means of which the Commission must give the MS concerned an opportunity to submit its observations.<sup>1749</sup> Since the Treaty of Lisbon, pursuant to Article 260 para 2 TFEU no prior reasoned opinion is required thereafter.1750 In its action the Commission shall qualitatively (lump sum 'or' – in the interpretation of the Commission and the Court that means: *and/or*1751 – penalty payment) and quantitatively (amount) specify the financial sanction it considers appropriate.1752 Where the Court finds that the MS has not complied with its judgement, it may – thereby disposing of a wide margin of appreciation1753 – impose a financial sanction.<sup>1754</sup>

According to Article 260 para 3 TFEU, a special (shortened) procedure applies where a MS has failed to fulfil its obligation to notify measures

<sup>1747</sup> See Posch/Riedl, Art. 260 AEUV, para 32; Wunderlich, Art. 260 AEUV, para 13, both with further references.

<sup>1748</sup> For the distinction between 'first order compliance' (with an obligation under the Treaties) and 'second order compliance' (with the Court's judgement according to Article 260 para 1 TFEU) see Andersen, Enforcement 44, with a further reference.

<sup>1749</sup> See Commission, 'Implementation of Article 260(3) TFEU' (Communication), SEC(2010) 1371 final, 2.

<sup>1750</sup> See Andersen, Enforcement 102 f; Grohs, Article 258/260, 70; Wennerås, Use 80.

<sup>1751</sup> Case C-304/02 *Commission v France*, paras 82 f.

<sup>1752</sup> For the Commission's recent adaptation of its calculation method in this respect see Commission Communication 'Financial sanctions in infringement procee‐ dings', 2023/C 2/01.

<sup>1753</sup> See W Cremer, Art. 260 AEUV, para 17; Posch/Riedl, Art. 260 AEUV, para 66.

<sup>1754</sup> The Court may even impose a financial penalty where the Commission has not requested one; see case C-304/02 *Commission v France*, para 90. For the debate on the introduction of sanctions which has started long before the negotiations on what became the Treaty of Maastricht see Tesauro, Remedies 19 f; for the application of this regime more generally see Wennerås, Sanctions.

transposing a directive adopted under a legislative procedure.1755 The non-communication of the transposition of directives – or rather: the non-transposition which, most of the time, is the reason for the non-com‐ munication<sup>1756</sup> – has been a severe problem and its combatting is one of the priorities of the Commission's current Treaty infringement 'policy'.<sup>1757</sup> Under this procedure the Commission may, when submitting an action according to Article 258 and in order to increase the pressure on the MS,<sup>1758</sup> immediately specify the financial sanction to be paid by the MS.<sup>1759</sup> Where the Court (partly) follows the action of the Commission, it may impose a financial sanction not exceeding the amount specified by the Commission.<sup>1760</sup> This shortened procedure applies to the Commission, but not to a MS suspecting the non-communication of the transposition of a (legislative) directive according to Article 259 TFEU.

### 2.1.3. Soft and hard elements of the procedure

In terms of the hard law-soft law dichotomy of compliance mechanisms, the Treaty infringement procedure can be described as a 'mixed mechanism'. In the pre-litigation phase under Article 258 TFEU, it is regularly the exchange of views and the endeavour to convince the respective opposite which stays in the foreground on both sides. This phase is coined by the absence of legally binding acts, also on the part of the Commission. The (potential)

1760 In the regular procedure according to Article 260 para 2 TFEU, the Court is not restricted in that way; see Posch/Riedl, Art. 260 AEUV, para 66; Wunderlich, Art. 260 AEUV, para 36.

<sup>1755</sup> For this procedure and the new approach taken in this context since President *Juncker* see Wendenburg/Reichert, Vertragsverletzungsverfahren 1339; for MS' practice to notify to the Commission inappropriate 'transposition measures' in order to gain time for proper transposition see Falkner/Treib/Hartlapp/Leiber, Europe 220 f.

<sup>1756</sup> See Wunderlich, Art. 260 AEUV, para 33, also with regard to the issues of wrongful and partial transposition; see also case C-543/17 *Commission v Belgium*, paras 50 ff.

<sup>1757</sup> See Commission, 'Better Monitoring of the Application of Community Law' (Communication), COM(2002) 725 final, 17 f; see, more recently, Commission, 'EU law: Better results through better application' (Communication), 2017/C 18/02, 15; for data from practice see Falkner/Treib/Hartlapp/Leiber, Europe 220; with regard to environmental law see Grohs, Article 258/260, 58.

<sup>1758</sup> See Karpenstein, Art. 260 AEUV, para 57.

<sup>1759</sup> Arguing that this would eventually slow down adjudication: Peers in House of Lords Select Committee on European Union, para 4.161.

preliminary output the Commission may forward to the MS concerned in the course of this correspondence normally cannot be qualified as EU soft law. Even the letter of formal notice, by name the most formal act of these preliminary expressions of view of the Commission, 'cannot, of necessity, contain anything more than an initial brief summary of the complaints'.<sup>1761</sup> While the letter of formal notice is of eminent importance as the act initiat‐ ing the pre-litigation phase of the Treaty infringement procedure pursuant to Article 258 TFEU and providing the MS addressed with the opportunity to utter its view, it does rather not qualify as EU soft law. This is for its lack of a command or at least a capability of being linked to a command<sup>1762</sup> which suggests that its primary purpose is not to ensure compliance (with the letter), but to start off a (formalised) dialogue. The Commission itself has expressed that the purpose of this letter is to request information, not to utter the Commission's legal view1763 – and that means: not to request a change in the MS' behaviour.<sup>1764</sup>

The legal qualification of the reasoned opinion is different.1765 While it may not contain new charges as compared to the letter of formal notice (see 2.1.2. above), it must describe the infringement 'in detail and prescribes the time within which the Member State must put an end to it'.1766 This is a clear – legally non-binding – command addressed to the MS.1767 The fact that it is the Court and not the Commission which may authoritatively state that there is an infringement of an obligation under the Treaties<sup>1768</sup> cannot do away with the fact that by adopting this opinion the Commission intends to ensure compliance with its own legal point of view (as expressed

<sup>1761</sup> Andersen, Enforcement 48.

<sup>1762</sup> See II.2.1.1.1. above.

<sup>1763</sup> See reference in Horspool/Humphreys, European Union Law 228.

<sup>1764</sup> *A fortiori*, this holds true for other letters sent in the course of this dialogue and for utterances given by the Commission prior to the pre-litigation phase, notably during the so-called EU Pilot procedure (see above and also V.2.5.3. below).

<sup>1765</sup> For the reasoned opinion adopted under Article 259 TFEU see above.

<sup>1766</sup> Koen Lenaerts/Maselis/Gutman, Procedural Law 188.

<sup>1767</sup> Lenaerts/Maselis/Gutman, Procedural Law 189 underpin this by expressing that the reasoned opinion is adopted 'in order to ensure that the Member State in question is accurately apprised of the grounds of complaint maintained against it by the Commission and can thus bring an end to the alleged infringement […]'.

<sup>1768</sup> See expressly joined cases 142/80 and 143/80 *Amministrazione delle Finanze dello Stato*, para 16; case C-393/98 *Valente*, para 18; joined cases T‑440/03, T‑121/04, T‑171/04, T‑208/04, T‑365/04 and T‑484/04 *Arizmendi*, para 69.

therein).1769 Where this opinion (which may also propose measures to rem‐ edy the infringement,<sup>1770</sup> thereby exhibiting elements of a recommendation) is complied with by the MS addressed, the Commission will not file an action with the Court. Then the only risk for the MS concerned is that the Court may later, in a different procedure, come to the conclusion that the MS, by complying with the reasoned opinion of the Commission, has actually infringed upon its obligations under the Treaties – a possible, but in practice highly unlikely scenario.<sup>1771</sup> It is nevertheless 'significant that [the adoption of the reasoned opinion] takes place in the shadow of formal adjudication'.<sup>1772</sup> The localisation of the reasoned opinion in this particular procedure significantly increases its factual authority,<sup>1773</sup> as does the fact that at the time of its adoption an express *audiatur et altera pars* has been exercised already.<sup>1774</sup> Also the decisive influence of the Commission on whether or not the litigation phase is entered contributes to its authority.

The judgement of the Court rendered according to an action under Article 258 (or Article 259) TFEU is legally binding.1775 It states in a legally binding way whether or not the MS concerned has infringed an obligation under the Treaties and, in the former case, the MS 'shall be required to take the necessary measures to comply with [it]'.1776 A judgement establishing a Treaty infringement may form the basis of a State liability claim.1777 The qualifications made or referred to here *mutatis mutandis* also apply to the respective acts adopted under Article 260 para 2 (and para 3) TFEU.

In summary, the sequence of acts is the following: The Commission ad‐ dresses at least one legally non-binding act (which does not qualify as soft law) to the MS concerned, which may be underpinned by a subsequent soft law act of the Commission. In case of non-compliance with this soft law act, the Commission may address a largely non-binding non-soft law act to the CJEU, upon which the latter may render a legally binding act addressed

<sup>1769</sup> See case T-194/04 *Bavarian Lager*, para 149, with a further reference.

<sup>1770</sup> See Lenaerts/Maselis/Gutman, Procedural Law 189 f.

<sup>1771</sup> Note the Court's statement that 'except where such powers are expressly conferred upon it, the Commission may not give guarantees concerning the compatibility of specific practices with the Treaty'; case C-415/93 *Bosman*, para 136.

<sup>1772</sup> Andersen, Enforcement 90.

<sup>1773</sup> See, with many further references, P Stelkens, Verwaltungsrecht, para 82.

<sup>1774</sup> See also case T-258/06 *Germany v Commission*, para 152.

<sup>1775</sup> For the legal qualification of the submissions of the AG which may precede the judgement see III.3.5.2.5. above.

<sup>1776</sup> Article 260 para 1 TFEU.

<sup>1777</sup> See Lenaerts/Maselis/Gutman, Procedural Law 207 f.

to the Commission and the MS concerned.1778 Under Article 260 para 2 TFEU, the Commission may address another legally non-binding non-soft law act to the MS concerned, which may be immediately followed by a largely non-binding non-soft law act to the CJEU, upon which the latter addresses a legally binding act to the Commission and the MS concerned.

The sequences of acts under the Treaty infringement regime are plau‐ sible. The Commission step by step increases the pressure on the MS concerned. At the beginning, it is just an exchange of views, the letter of formal notice marking the beginning of the pre-litigation phase. Where the MS does not budge, the Commission, under Article 258 TFEU, adopts a soft law act – the reasoned opinion – which clearly suggests that the MS ought to comply with it. If the MS refuses to comply (which it is, legally speaking, free to do), the Commission may turn to the CJEU, asking for a legally binding decision, a judgement. Under the sanctions regime a reasoned opinion is not envisaged (any more).

While the Court, if it holds the Commission's action to be admissible, may state that there is no infringement, empirical data shows that the risk for the MS to be condemned by the Court is considerable. <sup>1779</sup> If we perceive Articles 258 and 260 TFEU as one procedure, the full Treaty infringement procedure, it is – for the occurrence of acts of both EU soft law and EU law which are addressed to MS – to be qualified as a mixed compliance mechanism.

<sup>1778</sup> See Article 88 para 2 of the Rules of Procedure of the Court of Justice. This sequence also applies to the fast-track procedure laid down in Article 258 in conjunction with Article 260 para 3 TFEU.

<sup>1779</sup> In 2021, 90 % (18 out of 20; in 2020: 27 out of 28, that is 96 %) of the actions filed with the CJEU resulted in a judgement in the Commission's favour. Commission, 'General Statistical Overview', SWD(2022) 194 final, 30 f; Commission, 'General Statistical Overview', SWD(2021) 212, 24; see also Prete, Evolution 72.


2.2.1.1. In primary law

2.2.1.1.1. Article 106 para 3 TFEU

Article 106 TFEU defines the role of undertakings owned or privileged by the MS, and of the MS themselves respectively, in EU-wide competition. Para 1 stipulates that with regard to public undertakings and with regard to undertakings to which MS grant special or exclusive rights, MS are bound by the rules contained in the Treaties, in particular<sup>1780</sup> in Article 18 and Articles 101–109 TFEU. Para 2 lays down special rules with regard to undertakings entrusted with the operation of services of general economic interest1781 or having the character of a revenue-producing monopoly. Para 3 provides that the Commission shall ensure the application of the provi‐ sions of this Article, *inter alia* by addressing, where necessary, appropriate directives or decisions to MS.<sup>1782</sup>

The mechanism laid down in Article 106 para 3 TFEU is hard, as it provides for two, pursuant to Article 288 TFEU legally binding, categories of acts the Commission may adopt alternatively: directives or decisions.<sup>1783</sup> A decision on the basis of Article 106 para 3 TFEU is regularly addressed to one MS, obliging it to refrain from, or to take respectively, certain action. In the given context, the Commission's power to adopt decisions is most relevant, as it is a means of enforcing EU law individually *vis-à-vis* one or more MS. By means of a directive, on the contrary, the Commission

<sup>1780</sup> The Court has also applied Article 106 TFEU in other areas, for example in the context of fundamental freedoms; see eg case C-157/94 *Commission v Netherlands*, para 32 (freedom of goods).

<sup>1781</sup> For a monographic account of this concept see Melcher, Dienstleistungen, in particular 70 ff.

<sup>1782</sup> Other measures against the MS, eg Commission measures under Article 114 TFEU or the Treaty infringement procedure, are not thereby excluded; see Wernicke, Art. 106, para 78, who also deems lawful the adoption of Commission opinions and recommendations under Article 106 para 3 (para 82); note, in this context, case C-325/91 *France v Commission*; for the relationship of what is now Article 106 para 3 TFEU and the Treaty infringement procedure see Gil Ibáñez, Supervision 107–109.

<sup>1783</sup> See also explanations in joined cases 188–190/80 *France v Commission*, paras 11–14.

may set *general* standards on the relationship between MS and state-owned and/or privileged undertakings according to Article 106 para 1 TFEU.<sup>1784</sup> If the Commission intends to launch a Treaty infringement procedure after the above procedure, no short-cut – as provided for in Article 108 TFEU – can be used.<sup>1785</sup>

# 2.2.1.1.2. Article 108 TFEU

This provision belongs to the TFEU's State aid regime which is presented in most textbooks on EU law. Here Article 108 para 1, para 2 subparas 1 and 2, and para 3 TFEU shall be focused on. The procedure laid down in paras 1 and 2 applies to existing aids, the one laid down in paras 3 and 2 (in chronological order) to new aids.

As regards existing aids, para 1 stipulates that the Commission shall keep under constant review all systems of aid existing in the MS. It shall propose to them 'any appropriate measures required by the progressive development or by the functioning of the internal market'.1786 If these measures are not taken, the Commission may initiate the procedure laid down in para 2. As regards new aids, para 3 provides that the Commission shall be informed of any MS plans to grant or alter aid. In a perspective exclusively based on primary law, the Commission – where it considers that any such plan is incompatible with the internal market having regard to Article 107 TFEU – shall initiate the procedure provided for in para 2.

<sup>1784</sup> It is true that in general a directive may also be addressed only to one MS. However, with regard to Article 106 TFEU (and its predecessors) the Court held that the Commission shall adopt a directive only 'where, without taking into consideration the particular situation existing in the various Member States, it defines in concrete terms the obligations imposed on them under the Treaty. In view of its very nature, such a power cannot be used to make a finding that a Member State has failed to fulfil a particular obligation under the Treaty'; case C-202/88 *France v Commission*, para 17; for the distinction between directives and decisions in this context see case C-163/99 *Portugal v Commission*, paras 25–28; see also Gil Ibáñez, Supervision 106 f. For the Commission's past administrative practice in this context see Jung, Art. 106 AEUV, paras 68 ff; Wernicke, Art. 106, paras 86 and 88.

<sup>1785</sup> See Gil Ibáñez, Exceptions 154 f.

<sup>1786</sup> Article 22 of Council Regulation 2015/1589 stipulates that these appropriate meas‐ ures shall take the form of a Commission recommendation, which appears as an adequate concretisation of primary law; see also 2.2.2.2.2. below; for Commission communications based on what is now Article 108 para 1 TFEU, and their excep‐ tionally binding effect, see H Adam, Mitteilungen 107–113.

We shall now explore the procedure laid down in para 2, which may apply both in the context of existing and in the context of new aids – as a follow-up to the procedure laid down in para 1 and para 3, respectively. Where the Commission finds that a certain State aid granted (or planned to be granted or altered) is not compatible with the internal market (within the meaning of Article 107 TFEU), or that such aid is being misused, 1787 it shall – having notified the parties concerned to submit their comments in the course of a formal investigation procedure<sup>1788</sup> – decide that the respective MS 'shall abolish or alter such aid within a period of time to be determined by the Commission'. This decision shall be published in the OJ.<sup>1789</sup> If the MS does not follow this order of the Commission within the period of time determined by the latter, the Commission – or any other interested MS – may, 'in derogation from the provisions of Articles 258 and 259', refer the matter to the CJEU 'direct[ly]'.<sup>1790</sup>

In order to find out what the Treaties prescribe, the legal quality of the output adopted by EU actors in the course of these procedures shall be assessed with an exclusive view to primary law. The concretisation of one of these procedures, namely the one on existing aid schemes, by means of secondary law shall be considered below (see 2.2.2.2.2.). Whether the procedure on existing aid is mixed or hard is unclear in terms of primary law, as the act proposing 'appropriate measures' according to Article 108

<sup>1787</sup> 'Misuse of aid' means aid used by the beneficiary in contravention of an authoris‐ ing Commission decision; see Article 1 lit g of Council Regulation 2015/1589.

<sup>1788</sup> For the formal investigation procedure see in particular Article 6 of Council Regulation 2015/1589 which shall apply in the cases of notified aid (Article 4 para 4), unlawful aid (Article 15 para 1 in conjunction with Article 4 para 4), misuse of aid (Article 20 in conjunction with Article 4 para 4) and existing aid (Article 23 para 2 in conjunction with Article 4 para 4). The 'parties concerned' in Article 108 para 2 TFEU include the respective MS. For the different types of decisions the Commission may take see Article 9 of Council Regulation 2015/1589.

<sup>1789</sup> Article 32 para 3 of Council Regulation 2015/1589. For the confidentiality require‐ ments in order to protect business secrets and other confidential information see Article 9 paras 9 f *leg cit*; for the non-disclosure of confidential or secret informa‐ tion in Commission soft law related to competition or State aid law see Ştefan, Soft Law 102.

<sup>1790</sup> That the Council may, exceptionally and unanimously, decide that aid which a MS is granting or intends to grant shall be 'considered to be compatible with the internal market' (subpara 3) and the devolution provided for in subpara 4 shall not be expanded on any further here. After all, this is not a mechanism aimed at ensuring compliance with EU law by a MS, but an exceptional possibility for the Council to legalise an aid which without this exception would not be in accordance with EU law.

para 1 TFEU could qualify as law, soft law or an action below the level of soft law.<sup>1791</sup> What may follow – a Commission decision and an CJEU judgement – are hard law acts. The procedure on new aid is a hard mecha‐ nism, as it entails a decision by the Commission which may be followed by a judgement of the CJEU. According to primary law, no soft law acts are involved in this procedure.<sup>1792</sup> The Commission's notification of the parties concerned and its invitation to submit their comments can be neglected in this context. The former is not normative,<sup>1793</sup> the latter merely *invites* the party addressed to make comments. Since none of the two options the addressee has – to submit or not to submit comments – seems to be preferred by the Treaty-maker or the Commission as originator of the invitation, and since therefore the MS addressed is not pushed to act in either way, this output arguably does not entail a (soft) normativity.

The action to the Court, which is implied in the words 'refer the matter to the [CJEU]', is, first, addressed to the Court itself and, second, does not constitute soft law. The procedure can be described as a special form of the Treaty infringement procedure,1794 as it provides the Commission with the possibility to claim before the Court that a MS has violated an 'obligation under the Treaties', without providing for the pre-litigation phase as laid down in Article(s) 258 (and 260) TFEU.

#### 2.2.1.1.3. Article 114 TFEU

The procedure laid down in Articles 114 paras 4 ff allows for the MS to deviate, within certain limits and following a certain procedure, from an EU harmonisation measure.1795 This leeway is granted to the MS in

1795 Differently: Article 27 TFEU on temporary deviations for economic reasons; for an early legal act providing for MS deviations from internal market law see the

<sup>1791</sup> In spite of its potential to be mixed in its application in a concrete case, this procedure shall be analysed here together with the procedure on new aid *sub titulo* 'hard mechanisms'.

<sup>1792</sup> Neither does Council Regulation 2015/1589 provide for the adoption of soft law acts in this context.

<sup>1793</sup> Compare the letter of formal notice adopted by the Commission under Articles 258 and 260 TFEU, respectively.

<sup>1794</sup> Case 301/87 *France v Commission*, para 23: 'This means of redress is in fact no more than a variant of the action for a declaration of failure to fulfil Treaty obligations, specifically adapted to the special problems which State aid poses for competition within the common market'.

order to ensure the protection of certain interests of high importance, eg the protection of health and life of humans or the protection of the environment. Para 4 concerns the case that a MS deems it necessary to maintain national provisions on grounds of major needs corresponding to the justificatory reasons for restrictions to the freedom of goods pursuant to Article 36 TFEU, or relating to the protection of the environment or the working environment.<sup>1796</sup> Para 5 concerns the case that a MS deems it necessary to introduce national provisions based on new scientific evidence relating to the protection of the environment or the working environment on grounds of a problem specific to that MS arising after the adoption of the harmonisation measure. In both cases – that of para 4 and that of para 5 – the MS concerned shall notify the Commission of these provisions and of the reasons why they should be maintained/introduced. In either case the Commission shall, within six months of the notification, approve or reject the national provisions involved, after having verified whether or not they are a means of arbitrary discrimination or a disguised restriction on trade between MS, and whether or not they constitute an obstacle to the functioning of the internal market.1797 Where the Commission does not adopt a decision within the six months period (and where this period is not extended in accordance with Article 114 para 6 subpara 3, either), the national provisions at issue shall be deemed to have been approved. A MS measure shall only be applicable once the Commission has approved it (standstill requirement).1798 Where a MS is authorised to maintain or in‐ troduce national provisions derogating from a harmonisation measure, the

Agreement of the representatives of the Governments of the Member States meet‐ ing within the Council of 28 May 1969 providing for standstill and notification to the Commission.

<sup>1796</sup> The incorporation of this regime was a compensation for the MS for dropping the unanimity requirement in internal market legislation by the SEA; see Eilmans‐ berger, Binnenmarktprinzipien 261f.

<sup>1797</sup> After that request the Commission is not required to hear the MS again before taking its decision; see case C-3/00 *Denmark v Commission*, para 50; joined cases C-439/05P and C-454/05P *Land Oberösterreich*, para 43.

<sup>1798</sup> See joined cases C-439/05P and C-454/05P *Land Oberösterreich*, para 42; case C-319/97 *Kortas*, para 20: 'request a derogation'; see also Maletić, Harmonisation 80; for a – procedurally speaking – similar mechanism even prior to the intro‐ duction of what is now Article 114 TFEU see Articles 8 f of Council Directive 83/189/EEC (now: Articles 5 f of Directive 2015/1535); see also Gil Ibáñez, Supervi‐ sion 118 ff.

Commission shall immediately examine whether to propose an adaptation to that measure.<sup>1799</sup>

Article 114 para 9 refers to the case of a MS making 'improper use' of its powers under paras 4 and 5 – a term which shall simply mean non-compliance with a MS's duties under these provisions, no qualification of the misbehaviour being required.<sup>1800</sup> Where the Commission (or<sup>1801</sup> a MS) considers such an improper use to be made, it may – a 'procedural "shortcut[]"'<sup>1802</sup> from Articles 258 f TFEU – bring the matter directly before the CJEU.<sup>1803</sup>

Maintaining or introducing national provisions may be approved or rejected by the Commission by means of a (legally binding) decision. Hence Article 114 paras 4 f TFEU provide for two hard mechanisms. A matter according to para 9 can be 'directly' brought before the Court which means that the pre-litigation phase known from the Treaty infringement procedure does not apply. Considering that the only legal act adopted in this context is the judgement of the Court, also this procedure constitutes a hard compliance mechanism. The Treaty-makers when inserting Article 114 para 9 TFEU drew inspiration from Article 348 para 2 TFEU (see below).1804 Both notification regimes – para 4 and para 5 – provide for an examination involving the Commission and they are applied before poten‐

1802 Andersen, Enforcement 129.

1804 See Classen, Art. 114 AEUV, para 259.

<sup>1799</sup> Article 114 para 7 TFEU; for an alternative road (limited to public health concerns) to having the Commission examine whether to propose new EU measures see para 8 *leg cit*.

<sup>1800</sup> With regard to other paragraphs of Article 114, the procedure envisaged in para 9 does not seem to be applicable; see Classen, Art. 114 AEUV, paras 259 f; Tietje, Art. 114 AEUV, paras 226 f. Also with regard to the safeguard clauses referred to in Article 114 para 10 TFEU, the special infringement procedure does not apply. The latter provision does not as such empower the MS, but obliges the legislator. Only the concrete safeguard clause laid down in an act of secondary law empowers the MS (see 2.2.1.2.6. below); *argumentum* 'subject to a Union control procedure' (to be laid down in secondary law); for the applicability of the shortened infringement procedure according to para 9 also in the context of para 10 see Khan, Art. 114 AEUV, para 21; Korte, Art. 114 AEUV, para 70, with further references; differently: Classen, Art. 114 AEUV, para 197.

<sup>1801</sup> The conjunction 'and' between 'the Commission' and 'any Member State' in Arti‐ cle 114 para 9 TFEU is to be understood as 'or'; see eg the German version of this provision; see also Article 348 para 2 TFEU, using the term 'or'.

<sup>1803</sup> See also Koops, Compliance 143, with further references.

#### *IV. MECHANISMS IN EU LAW TO ENSURE LEGAL COMPLIANCE*

tial Court proceedings.<sup>1805</sup> Arguably, this is the reason why the pre-litigation phase pursuant to Article 258 TFEU is dropped in these cases. Only where a MS has failed to notify the Commission of maintained/introduced nation‐ al provisions, the Commission may bring the matter before the Court, without a prior hearing being provided for by law.<sup>1806</sup> Since such non-noti‐ fication is a clear 'improper use of [the MS's] powers', a pre-litigation phase appears to be dispensable.<sup>1807</sup> What was said under 2.2.1.1.2. above with regard to the legal qualification of the action filed with the Court and with regard to the relationship of this procedure to Article 258 TFEU *mutatis mutandis* applies here, as well.

# 2.2.1.1.4. Article 348 TFEU

Like Article 114 para 9 TFEU, Article 348 para 2 TFEU stipulates that, by (explicit) derogation from the general Treaty infringement procedure, the Commission or any MS may bring the matter 'directly' before the CJEU 'if it considers that another [MS] is making improper use of the powers [at issue]'. The powers at issue here are the rights and competences of a MS related to its national security as laid down in Articles 346 and 347 TFEU, that is to say the right not to supply information the disclosure of which the MS concerned considers contrary to the essential interests of its security and the competence to take the measures considered necessary for the protection of the essential interests of its security which are connected with the production of or trade in arms, munitions and war material (Article 346 para 1 lit a and b).1808 It also includes the competence of MS to consult each other 'with a view to taking together the steps needed to prevent the

<sup>1805</sup> See 114 para 6, according to which the Commission may approve or reject the notified national provision.

<sup>1806</sup> By all means the MS can utter its view before the Court, once litigation has started; see Article 41 of the Statute of the CJEU and Article 124 of the Rules of Procedure of the CJEU.

<sup>1807</sup> See Classen, Art. 114 AEUV, para 260. Has the Commission approved the national provision, another MS may – for lack of 'improper use' of the MS's powers – only initiate an annulment procedure against the Commission decision (or initiate a regular Treaty infringement procedure pursuant to Article 259 TFEU); see case C-41/93 *France v Commission*, in which the Court annulled the Commission decision for lack of sufficient reasoning.

<sup>1808</sup> At the same time, these measures shall not adversely affect the conditions of competition in the internal market regarding products which are not intended for specifically military purposes (Article 346 para 1 lit b TFEU *in fine*). See case

functioning of the internal market being affected by measures which a [MS] may be called upon to take in the event of serious internal disturbances affecting the maintenance of law and order, in the event of war, serious international tension constituting a threat of war, or in order to carry out obligations it has accepted for the purpose of maintaining peace and international security' (Article 347). If measures taken under the regimes of Article 346 or Article 347 TFEU have the effect of distorting the conditions of competition in the internal market, the Commission shall, together with the MS concerned, examine how these measures can be adjusted to the rules laid down in the Treaties.

In case of Article 348 para 2 – ie when a MS has allegedly made improper use of the above powers – the Court shall, in order to protect security-re‐ lated interests of the MS, give its ruling *in camera*. That does not mean that the judgement is not read in open court pursuant to Article 37 of the Statute of the CJEU, but it means that confidential information contained in the judgement is edited out in the published version.1809 As regards a prior possibility for the MS concerned to make known its views, the discussion provided for in Article 348 para 1 TFEU is to be underlined. Given the fact that all cases of 'improper use' pursuant to Article 348 para 2 will lead to a distortion of the conditions of competition in the internal market, Article 348 para 1 seems to be an appropriate tool to give the MS concerned the possibility to utter its view *vis-à-vis* the Commission prior to (potential) Court proceedings.1810 As the normative output on the part of the EU in this mechanism only involves a Court ruling, it constitutes a hard mechanism.

C-372/05 *Commission v Germany*, para 70, according to which the provision now contained in Article 346 TFEU 'cannot however be read in such a way as to confer on Member States a power to depart from the provisions of the Treaty based on no more than reliance on those interests'; for the exclusion of so-called dual-use goods (which fall under the regime set under CCP) see case C-70/94 *Werner*, paras 8 ff.

<sup>1809</sup> See Dittert, Art. 348 AEUV, para 16.

<sup>1810</sup> Note that the consultations among MS according to Article 347 – as a discussion among peers, not involving the Commission – do not contribute to the MS' right to be heard being fulfilled; see Calliess, Art. 348 AEUV, para 3.

### 2.2.1.1.5. Article 144 TFEU

The regime of Articles 143 and 144 TFEU concerns only MS 'with a deroga‐ tion', that is to say non-euro MS.<sup>1811</sup> Where such a MS is 'in difficulties or is seriously threatened with difficulties as regards its balance of payments' 1812 and where these difficulties are 'liable in particular to jeopardise the func‐ tioning of the internal market or the implementation of the common com‐ mercial policy', the Commission shall examine the case and recommend measures for the MS to take. If these measures, together with the measures the MS may have taken of its own accord, do not prove sufficient to over‐ come the (threat of ) difficulties, the Council may, upon a recommendation of the Commission, grant mutual assistance. <sup>1813</sup> Where mutual assistance recommended by the Commission is not granted or where, together with the other measures taken, it turns out to be insufficient, the Commission shall authorise the MS1814 to take protective measures (in derogation from Union law1815), the conditions and details being determined by the Com‐ mission. This authorisation may be revoked and the conditions and details may be changed by the Council.<sup>1816</sup>

Where a sudden crisis1817 in the balance of payments occurs and mutual assistance according to Article 143 para 2 TFEU is not immediately granted, a MS with a derogation 'may, as a precaution, take the necessary protective measures'.1818 These 'must cause the least possible disturbance in the func‐ tioning of the internal market' and must not go beyond what is 'strictly necessary to remedy the sudden difficulties'.1819 The Commission and the other MS shall be informed of these measures when they enter into force

<sup>1811</sup> For the exact definition see Article 139 para 1 TFEU.

<sup>1812</sup> Article 143 para 1 TFEU; for the (required) causes of these difficulties see ibid.

<sup>1813</sup> Article 143 para 1 subpara 2 and para 2 TFEU; see Council Regulation 332/2002.

<sup>1814</sup> The MS must be 'in difficulties'. Apparently, a threat of difficulties is not sufficient at this stage.

<sup>1815</sup> See Häde, Art. 143 AEUV, para 13.

<sup>1816</sup> Article 143 para 3 subpara 2 TFEU.

<sup>1817</sup> That means a fast and unexpected emergence of a crisis situation; see Hölzer, Art. 144 AEUV, para 6.

<sup>1818</sup> Article 144 para 1 TFEU.

<sup>1819</sup> Article 144 para 1 TFEU.

at the latest.<sup>1820</sup> The Commission may (still) recommend to the Council the granting of mutual assistance as referred to above.<sup>1821</sup>

The Council, upon a Commission recommendation and after the Eco‐ nomic and Financial Committee has been consulted, may decide that the MS concerned shall amend, suspend or abolish the protective measures it has adopted. It may do so even if it has not granted mutual assistance pursuant to para 2.

This latter mechanism shall build the focus here.<sup>1822</sup> It is a mechanism to ensure that a MS, here: a non-euro MS, complies with EU law. Article 144 para 3 TFEU does not make explicit under which conditions the Council shall amend, suspend or abolish the protective measures, or when the Commission shall adopt its recommendation, respectively. A systematic interpretation, however, reveals that the standard of scrutiny is the lowest possible disturbance in the functioning of the internal market and the strict necessity of the protective measures.1823 In addition to that, the measures must be directed against the sudden crisis in the balance of payments, not against general difficulties regarding the balance of payments1824; for the latter, the regime laid down in Article 143 TFEU may be applicable.<sup>1825</sup>

<sup>1820</sup> Compliance with this duty to inform is a requirement for the legality of these measures; see joined cases 6/69 and 11/69 *Commission v France*, paras 28 ff.

<sup>1821</sup> Article 144 para 2 TFEU.

<sup>1822</sup> The entire system of remedying (threats of ) difficulties to the balance of payments of non-euro MS was outlined to put Article 144 para 3 TFEU in context.

<sup>1823</sup> Article 144 para 1 TFEU.

<sup>1824</sup> Whether or not the MS may uphold its protective measures causing the least possible disturbance in the functioning of the internal market and not going beyond what is strictly necessary, once the Council has granted mutual assistance pursuant to Article 144 para 2 in conjunction with Article 143 – and as long as it is not requested to amend, suspend or abolish them pursuant to para 3 – is not clear from the wording of Article 144. Even if the Council grants mutual assistance later, it has still 'not immediately taken' the respective decision and hence this requirement of para 1 is fulfilled. The term 'as a precaution' may suggest otherwise; see also Häde, Art. 144 AEUV, paras 1f, with a further reference; arguing that protective measures of a MS need to be abolished as soon as mutual assistance is granted: Bandilla, Art. 144 AEUV, para 11. Even where such coexistence of MS measures and mutual assistance would be unlawful as such, Article 144 para 3 TFEU would still not be redundant, as the Council may take such decision also without having granted mutual assistance.

<sup>1825</sup> See Häde, Art. 144 AEUV, para 3; see Kilpatrick, Bailouts 400, with regard to Hungary, Latvia and Romania, which have received assistance under Article 143 TFEU, and with further references.

### *IV. MECHANISMS IN EU LAW TO ENSURE LEGAL COMPLIANCE*

The fact that the Council only adopts a decision renders this compliance procedure a hard mechanism, as it entails only a legally binding act *vis-à-vis* the MS. The Commission recommendation, a soft law act which is required for the Council decision, cannot change this, because it is addressed only to the Council. For our classification of compliance mechanisms it is the acts addressed to the MS which matter. Neither is the (possible) grant of mutual assistance – whatever legal form it may take – to be considered here, as it is geared towards doing away with an undesirable economic situation in a MS, but not about ensuring compliance of acts of national law with EU law.

### 2.2.1.2. In secondary law

# 2.2.1.2.1. Article 13 para 1 of Directive 2001/95/EC

Directive 2001/95/EC of the EP and of the Council, based on what is now Article 114 TFEU (former Article 95 TEC) lays down rules on general product safety '[i]n order to ensure a high level of consumer protection'.<sup>1826</sup> In the given context, it is in particular its Article 13 para 1 which is of interest. It provides for the following procedure involving the Commission and 'various' MS: Where the Commission 'becomes aware of a serious risk from certain products to the health and safety of consumers in various Member States', it may adopt a decision requiring the respective MS to take certain counter-measures from among those laid down in Article 8 para 1 lit b to f (eg a warning of the risk to certain persons or a ban on marketing a product). It shall do so only after consulting the MS and, where 'scientific questions arise which fall within the competence of a Community Scientific Committee', after consulting the competent committee. The decision shall be adopted 'in the light of the result of those consultations' and, unless they concern specific, individually identified products or batches of products, shall be subject to (repeated) re-consideration after one year.<sup>1827</sup>

<sup>1826</sup> Recital 4 of Directive 2001/95/EC.

<sup>1827</sup> Article 13 para 2 of Directive 2001/95/EC. For further procedural requirements see paras 3–5 *leg cit*.

The procedure which is addressed here is the examination procedure, a comitology procedure according to Article 5 of Regulation 182/2011.<sup>1828</sup> Where the competent committee provides a negative opinion or where it does not provide an opinion on a certain draft, the Commission may not adopt the decision.<sup>1829</sup> The decision shall furthermore be adopted only where, at one and the same time, a) it emerges from prior consultations with the MS that they – the MS – differ significantly on the approach adopted or to be adopted to deal with the risk; b) the risk cannot be dealt with, in view of the nature of the safety issue posed by the product, in a manner compatible with the degree of urgency of the case, under other procedures laid down by the specific Union legislation applicable to the products concerned; and c) the risk can be eliminated effectively only by adopting appropriate measures applicable at Union level, in order to ensure a consistent and high level of protection of the health and safety of consum‐ ers and the proper functioning of the internal market (Article 13 para 1 lit a to c). For the duration of the validity of the Commission decision, the (prohibited) export to third countries of the product(s) concerned, the implementation period for the MS (20 days, unless the Commission specifies otherwise), and for the right to be heard of the parties concerned see Article 13 paras 2 to 5 of the Directive.<sup>1830</sup>

From a structural point of view, and in the MS' perspective focussed on here, this is a hard mechanism. The Commission tells the MS concerned in a form of law, namely a decision, which measures it shall take in order to ensure the health and safety of consumers pursuant to Union law. That the Commission in its decision-making depends on a positive opinion of a comitology committee is a different story.1831 These opinions are addressed to the Commission, not to the MS. The special rules regarding the duration of the validity of the decision and the period for its implementation by the MS as laid down in Article 13 paras 2 and 4 of the Directive are peculiarities which do not affect the decision's legal bindingness.

<sup>1828</sup> Article 15 para 2 of Directive 2001/95/EC, to which its Article 13 para 1 refers, in conjunction with Article 5 of Council Decision 1999/468/EC in conjunction with Article 13 para 1 lit c of Regulation 182/2011.

<sup>1829</sup> Article 5 para 3 and 4 subpara 2 lit b in conjunction with Article 13 para 1 lit c of Regulation 182/2011. For the rights of scrutiny of the EP and the Council regarding the Commission decision see Article 11 of Regulation 182/2011.

<sup>1830</sup> For a comparison of the Article 13 regime with its predecessor, Article 9 of Direc‐ tive 92/59/EEC, see Weatherill, Consumer Law 213–215.

<sup>1831</sup> The comitology opinions adopted in the course of an examination procedure are analysed under III.3.7.2.1. above.

### 2.2.1.2.2. Articles 70 f of Regulation 2018/1139

Regulation 2018/1139 of the EP and of the Council, based on Article 100 para 2 TFEU, lays down common rules in the field of civil aviation and establishes the EASA<sup>1832</sup>. Its Article 70 is entitled '[s]afeguard provisions'. Para 1 states that Regulation 2018/1139 and the tertiary law adopted on its basis 'shall not prevent a Member State from reacting immediately to a problem relating to civil aviation safety' where the following conditions are met (cumulatively): a) it involves a serious risk to aviation safety and imme‐ diate MS action is required to address it; b) it is not possible for the MS to adequately address the problem in compliance with Regulation 2018/1139 and the tertiary law based on it; c) the action taken is proportionate to the severity of the problem. In this case, the MS concerned shall immediately notify the Commission, the EASA and the other MS of the measures taken, their duration and the reasons for taking them. Subsequently, the EASA shall examine whether these conditions have actually been met.1833 If so, it shall assess whether it is able to address the problem identified by the MS by taking decisions according to Article 76 para 4 of Regulation 2018/1139, thereby obviating the need for MS action. In the affirmative, it shall take the appropriate decision to that effect and inform the MS thereof. If it deems that the problem cannot be addressed that way, it shall recommend to the Commission to amend any delegated or implementing acts adopted on the basis of Regulation 2018/1139 in order to address this issue.<sup>1834</sup>

If the EASA is of the opinion that the above conditions have not been met, it shall address a recommendation to the Commission as regards the outcome of this assessment. The Commission shall then assess itself wheth‐ er the conditions have been met.1835 Where it considers that the conditions have not been met or where it departs from the outcome of the EASA's assessment, it shall adopt implementing acts containing its decision and setting out its findings to that effect. If the implementing act confirms that

<sup>1832</sup> The original founding regulation – which Regulation 2018/1139 repeals – is Regula‐ tion 216/2008.

<sup>1833</sup> For the repository which serves the communication of viewpoints in this context see Article 74 of Regulation 2018/1139.

<sup>1834</sup> The Commission is not legally bound by the EASA's assessment. However, the au‐ thority of the EASA's output is considerable; see Riedel, Gemeinschaftszulassung 123; W Weiß, Agenturen 639; for the phenomenon of agencies predetermining Commission output more generally see Orator, Möglichkeiten 95–97.

<sup>1835</sup> Article 70 para 3 of Regulation 2018/1139.

the conditions have not been met, the MS concerned shall, upon notifica‐ tion of this act, immediately revoke the measure taken pursuant to Article 70 para 1 of Regulation 2018/1139.<sup>1836</sup> It appears that these implementing acts shall be adopted without the involvement of comitology committees pursuant to Regulation 182/2011.<sup>1837</sup> Where the EASA and the Commission agree that the conditions have been met, apparently no implementing act of the Commission is provided for to confirm the legality of the MS measure.

According to Article 71, MS may, 'in the event of urgent unforeseeable circumstances affecting [the persons subject to Regulation 2018/1139] or ur‐ gent operational needs of those persons', grant exemptions to requirements laid down in Chapter III of Regulation 2018/1139 (other than essential requirements) or in tertiary law based on that Chapter, if the following conditions are met (cumulatively): a) it is not possible to adequately address those circumstances or needs in compliance with the applicable requirements; b) safety, environmental protection and compliance with the

<sup>1836</sup> Article 70 para 4 of Regulation 2018/1139. Note, in this context, the deletion of a sentence provided for in Commission Proposal COM(2015) 613 final: 'The Member State concerned shall immediately terminate the measures taken pursuant to paragraph 1 upon the notification of that implementing decision' (Article 59 para 2 subpara 3). Where the Commission confirms that the conditions have been met (thereby departing from the EASA's assessment), it is unclear whether the Commission only adopts a general-abstract implementing act (addressing the underlying issue) or whether the respective MS, in addition to that, also receives a (positive) decision regarding the measure it has taken. At least under the pred‐ ecessor provision, Article 14 of Regulation 216/2008, the Commission seems to have adopted positive decisions; see Commission Decision of 6 February 2014 authorising Sweden and the United Kingdom to derogate from certain common aviation safety rules pursuant to Article 14(6) of Regulation (EC) No 216/2008 of the European Parliament and of the Council. These decisions apply until amend‐ ment of the relevant rule: see <https://www.easa.europa.eu/system/files/dfu/Arti cle%2014%206%20webpage%2020140520.pdf> pages 2 f, accessed 28 March 2023; for the decision-making procedures see ibid 1 and 5.

<sup>1837</sup> Article 70 does not make any reference to comitology procedures in accordance with Article 127 of Regulation 2018/1139. Note the difference between the text of Recital 57 of Commission Proposal COM(2015) 613 final which makes all Commission implementing acts subject to comitology and Recital 75 of Regulation 2018/1139 which merely stipulates that 'the majority of […] implementing powers […] should be exercised in accordance with Regulation (EU) No 182/2011'; for the procedure to be followed under the predecessor-Regulation see Article 8 of Regulation 182/2011 (Articles 14 and 65 para 7 of Regulation 216/2008 in conjunc‐ tion with Article 13 para 1 lit d of Regulation 182/2011); for the EASA's role in supporting the Commission in the adoption of implementing acts more generally see Riedel, Gemeinschaftszulassung 310–312.

applicable essential requirements are ensured; c) the MS has mitigated any possible distortion of market conditions as a consequence of the grant‐ ing of the exemption as far as possible; d) the exemption is limited in scope and duration to the extent strictly necessary and it is applied in a non-discriminatory manner. In this case the MS shall immediately notify the Commission, the EASA and the other MS of the measure, its duration and its reasoning.<sup>1838</sup>

Where the exemption was granted for a duration of more than eight consecutive months or where the same exemption was granted repetitively (by the same MS) and its total duration exceeds eight months, the EASA shall assess whether the above conditions have been met and shall, within three months, adopt a recommendation to the Commission as regards the outcome of this assessment.<sup>1839</sup> The Commission shall then, taking account of that recommendation, assess itself whether the conditions have been met. Where the conditions have not been met or where the Commission departs from the EASA's assessment, the Commission shall, within three months, adopt an implementing act containing its decision to that effect. If the implementing act confirms that the conditions have not been met, the MS concerned shall, upon notification of this act, immediately revoke the measure taken pursuant to Article 71 para 1 of Regulation 2018/1139. Where the EASA and the Commission agree that the conditions have been met, apparently no implementing act of the Commission is provided for to confirm the legality of the MS measure.<sup>1840</sup>

Articles 70 f of Regulation 2018/1139 contain two regimes under which an EU actor may tell a MS how to comply with EU law. According to Article 70, a MS may take measures reacting immediately to a problem relating to civil aviation safety. Upon a recommendation of the EASA, the Commission – where it deems that the conditions for MS action have not been met or where it departs from the EASA's assessment – may adopt an implementing act to address this problem. It may determine that the

<sup>1838</sup> See Article 71 para 1 of Regulation 2018/1139, also with regard to the possibility of applying mitigation measures.

<sup>1839</sup> Where the exemption was granted by the EASA itself, the procedure pursuant to Article 76 para 4 of Regulation 2018/1139 applies.

<sup>1840</sup> Article 71 para 2 of Regulation 2018/1139. Para 3 provides for the possibility of MS requesting the amendment of delegating and implementing acts regarding the demonstration of compliance with the essential requirements referred to above. Since this does not constitute a compliance mechanism within the understanding applied here, this provision shall not be discussed any further.

conditions have not been met – in which case the MS has to revoke its measure – or it may address the underlying issue otherwise: by a gener‐ al-abstract regulation of the issue or by simply confirming the legality of the MS measure. The second regime is laid down in Article 71: MS may grant exemptions from requirements laid down in the pertinent EU law under urgent unforeseeable circumstances affecting the persons subject to Regulation 2018/1139 or urgent operational needs of those persons. If the exemption(s) are granted for more than eight consecutive months, the following procedure applies: Following a recommendation from the EASA, the Commission either adopts an implementing act, or where it agrees with the EASA that the conditions are met does not adopt an act, thereby implicitly confirming the legality of the MS measure.

These regimes are, *vis-à-vis* the MS, hard mechanisms. The EASA's recommendation preceding the (possible) Commission acts in both proce‐ dures is addressed only to the Commission, with the MS being informed, accordingly. The respective MS only receives the Commission implement‐ ing act or, exceptionally under the regime of Article 70, the EASA decision – or, if no intervention is deemed necessary, no act at all.

#### 2.2.1.2.3. Article 29 para 2 of Regulation 806/2014

Regulation 806/2014 of the EP and of the Council, based on Article 114 TFEU, establishes uniform rules and a uniform procedure for the resolu‐ tion of credit institutions and certain investment firms in the framework of a Single Resolution Mechanism and a Single Resolution Fund. According to Article 29 para 1 of this Regulation, the national resolution authorities shall implement all decisions referred to in this Regulation.1841 These deci‐ sions emanate, for example, from the Council and the Commission, or, importantly, from the Single Resolution Board (SRB), a European agency established by this Regulation. Where a national resolution authority 'has not applied or has not complied with a decision' of the SRB according to Regulation 806/2014 or 'has applied it in a way which poses a threat to any of the resolution objectives under Article 14 or to the efficient implementa‐

<sup>1841</sup> For the room for manoeuvre left to the national resolution authorities when fol‐ lowing a decision of the SRB addressed to them see Article 6 para 7 of Regulation 806/2014.

tion of the resolution scheme', the SRB – in its executive session<sup>1842</sup> – may order an institution under resolution<sup>1843</sup> to take certain measures.<sup>1844</sup> Before doing so, the SRB shall notify – ideally 24 hours in advance – the national resolution authority or authorities concerned and the Commission of its intention to adopt a decision, thereby providing the details and reasons of the envisaged measures and details on when they are intended to take effect.<sup>1845</sup>

This is a special case: While it is the MS authorities which shall comply with SRB decisions (mostly regarding institutions under resolution), the SRB – in order to remedy non-compliance on the part of the MS – may address a decision directly to the institution concerned. This tool may be considered even more effective than a decision *vis-à-vis* the MS. Since a MS is obliged to comply with decisions of the SRB anyway, it is dubitable whether a second SRB decision *vis-à-vis* this MS would be more effective in ensuring MS' compliance than the first decision. By directly addressing those concerned, namely the institution under resolution, the SRB ousts the national resolution authority (evocation; see also 2.2.1.2.5. below). A decision of the SRB shall prevail over any previous decision adopted by the national resolution authority on that matter (Article 29 para 3 of Regulation 806/2014). Para 4 states that the national resolution authorities, when 'tak‐ ing action in relation to issues which are subject to [an SRB-decision] taken pursuant to paragraph 2', shall comply with that decision. This stipulation is to be understood against the background of the general rule, according to which '[a] decision which specifies those to whom it is addressed [here: the institution under resolution] shall be binding only on them'.<sup>1846</sup>

While this mechanism does not entail direct intervention of an EU actor *vis-à-vis* a MS, the SRB decision indirectly remedies a MS's non-compli‐ ance with pertinent EU law. It is a hard mechanism in that it does not allow a MS to deviate. Even though the SRB decision is not addressed to the MS, the latter is affected (and restricted in its actions) by the decision.

<sup>1842</sup> Article 54 para 1 lit b of Regulation 806/2014.

<sup>1843</sup> For this term see Article 3 para 1 subpara 14 in conjunction with Article 2 of Regulation 806/2014.

<sup>1844</sup> See Article 29 para 2 lit a-c of Regulation 806/2014.

<sup>1845</sup> Article 29 para 2 subparas 3 and 4 of Regulation 806/2014. For the publication of the resolution scheme or its main content (including decisions taken pursuant to Article 29 para 2) see Article 29 para 5 of Regulation 806/2014.

<sup>1846</sup> Article 288 para 4 TFEU; on this issue see also Geismann, Art. 288 AEUV, para 60, with further references.

### 2.2.1.2.4. Article 63 of Regulation 2019/943

The compliance mechanism at issue here – in a Regulation on the internal market for electricity, which is based on Article 194 para 2 TFEU<sup>1847</sup> – is placed within a regime under which new<sup>1848</sup> direct<sup>1849</sup> current interconnec‐ tors may, upon request and for a limited period, be exempted from certain requirements under this Regulation and under Directive 2019/944. We shall not dwell on these requirements here, for which an exemption may only be granted under certain conditions essentially intended to maintain competi‐ tion and the effective functioning of the energy market, <sup>1850</sup> but focus on the decision-making mechanism of this regime.

Having received the request, the national regulatory authority<sup>1851</sup> shall send a copy of it to the ACER and to the Commission for information.<sup>1852</sup> It is normally the national regulatory authorities of the MS concerned which decide on such a request, and exceptionally the ACER,1853 namely where the authorities concerned cannot reach an agreement within six months or upon their joint request. The ACER shall take its decision, having consulted the national regulatory authorities concerned and the applicants.1854 The decision – either of the national regulatory authorities concerned or of the ACER (the notifying bodies) – shall be notified to the Commission without delay with all the relevant information (reasoning, analysis of effect on competition etc).<sup>1855</sup>

Within an extendable period of 50 days from receipt of this notification, the Commission may adopt a decision requesting the notifying bodies to

<sup>1847</sup> Note that the predecessor regulation – Regulation 714/2009 – was still based on Article 114 TFEU; see Article 17 para 8 *leg cit* for the predecessor compliance mech‐ anism. For Article 194 TFEU as the new (ie Lisbon) legal basis for energy-related acts and further acts which have been based on it see Talus, Introduction 11–14; see also Ludwigs, Energierecht, para 142.

<sup>1848</sup> For the application of this regime to existing interconnectors in case of significant increases of capacity see Article 63 para 3 of Regulation 2019/943.

<sup>1849</sup> For the exceptional application of this regime to alternating current interconnec‐ tors see Article 63 para 2 of Regulation 2019/943.

<sup>1850</sup> See Article 63 para 1 of Regulation 2019/943.

<sup>1851</sup> See definition in Article 2 para 2 of Regulation 2019/943.

<sup>1852</sup> Article 63 para 7 of Regulation 2019/943.

<sup>1853</sup> Article 63 paras 4 f of Regulation 2019/943. For an overview of the tasks of this authority see Hauenschild, Agentur 108 f; Tišler, Agency 392 ff.

<sup>1854</sup> Article 63 para 5 of Regulation 2019/943.

<sup>1855</sup> See Article 63 para 7 of Regulation 2019/943.

amend or withdraw the decision to grant an exemption. The addressees of this decision shall comply with it within one month and shall inform the Commission accordingly.<sup>1856</sup> The Commission also takes a decision if it approves of an exemption decision.<sup>1857</sup> Apparently, a decision refusing the request to grant an exemption may not be objected to by the Commission under this regime.

The above procedure shall, *mutatis mutandis*, apply also where the national regulatory authorities decide to modify an (existing) exemption decision.<sup>1858</sup>

The Commission may, upon request or on its own initiative, decide to reopen the proceedings having lead to an exemption where there has been a material change in any of the facts on which the exemption decision was based, where the undertakings concerned act contrary to their com‐ mitments or where the decision was based on incomplete, incorrect or misleading information provided by the parties.<sup>1859</sup>

In terms of Commission output, this mechanism consists of one deci‐ sion. It is normally directed to national regulatory authorities (and excep‐ tionally to the ACER) and obliges them (it) to comply with it.1860 In the regular case that national regulatory authorities are concerned, it therefore qualifies as a hard compliance mechanism according to the terminology introduced under 1. above.

# 2.2.1.2.5. Articles 18 and 19 of Regulation 1093/2010

The procedures dealt with here are laid down in the Regulation establishing the EBA. Essentially the same procedures are laid down in the Regulations establishing the two sister authorities involved in the supervision and regu‐ lation of the financial market, the European Insurance and Occupational Pensions Authority (EIOPA) and the European Securities and Markets Au‐ thority (ESMA). These three founding Regulations are all based on Article

<sup>1856</sup> Article 63 para 8 subpara 3 of Regulation 2019/943, containing further details of the procedure on the notification of (requested) information to the Commission.

<sup>1857</sup> Article 63 para 8 subpara 5 of Regulation 2019/943.

<sup>1858</sup> Article 63 para 9 of Regulation 2019/943.

<sup>1859</sup> Article 63 para 10 of Regulation 2019/943.

<sup>1860</sup> Differently with regard to the earlier generation of compliance mechanisms in energy law: case T-317/09 *Concord*, paras 50–53; see also case T-381/09 *RWE*, paras 44–47.

114 TFEU and are largely drafted alike. In the following, the procedures laid down in Regulation 1093/2010 will be presented and analysed with a focus on the EU legal acts which may be adopted in the course of these procedures.<sup>1861</sup>

Article 18 is entitled 'Action in emergency situations'.<sup>1862</sup> Where the exis‐ tence of an emergency situation is established by the Council according to para 2 and in exceptional circumstances where the orderly functioning and integrity of financial markets or the stability of at least part of the financial system in the Union or customer and consumer protection is at risk,<sup>1863</sup> the EBA may adopt individual decisions requiring 'competent authorities'<sup>1864</sup> to take the necessary action in accordance with the acts referred to in Article 1 para 2 of Regulation 1093/2010 to address the emergency situation by ensuring that financial institutions and competent authorities satisfy the requirements laid down in this Union law.1865 Where a competent authority does not comply with such decision within the prescribed period of time, the EBA may, where the EU legislation at issue (and the regulatory and implementing technical standards based on it) is directly applicable to financial institutions, adopt an individual decision addressed to a financial institution requiring the necessary action to comply with its obligations under these rules, including the cessation of any practice. This shall apply only in situations in which a competent authority does not apply the above Union law or applies it in a way which appears to be a manifest breach of it,

<sup>1861</sup> For a more comprehensive analysis of these procedures see eg Michel, Gleichge‐ wicht 248–255; Weismann, Agencies 138–144, with further references.

<sup>1862</sup> Critically of the weak involvement of the Commission in this procedure: Michel, Gleichgewicht 254 f; note the insistence of the legislator – in view of the *Meroni* doctrine – to incorporate the Commission in another procedure involving a Euro‐ pean agency (the SRB): case T-628/17 *Aeris*, paras 127–130.

<sup>1863</sup> For the full description of these exceptional circumstances see Article 18 para 3 of Regulation 1093/2010.

<sup>1864</sup> Originally, the 'competent authorities' were only the national supervisory authori‐ ties (see Article 4 para 1 no 40 of Regulation 575/2013). With the SSM starting to operate in 2014, the ECB has – as far as the Regulation 1093/2010 is concerned – become a 'competent authority' as well (see Recital 45 of Council Regulation 1024/2013 and Article 4 para 2 (i) of Regulation 1093/2010).

<sup>1865</sup> This illustrates that meanwhile national authorities may be direct addressees of EU decisions; only pointing in that direction: case 310/85 *Deufil*, para 24; more strictly: *von Bogdandy/Bast/Arndt,* Handlungsformen 96 (fn 68); stressing that direct communication between the EU's administration (the Commission) and national authorities is the exception rather than the rule: Eekhoff, Verbundaufsicht 129.

and where urgent remedying is necessary to restore the orderly functioning and integrity of financial markets or the stability of at least part of the financial system in the Union.

Article 19 is about the settlement of disagreements between competent authorities in cross-border situations. The part of the procedure to be focused on here is laid down in paras 3 and 4. The background to this procedure is a disagreement of a competent authority about the procedure or content of an action, proposed action or inaction of another competent authority or cases where legislative acts referred to in Article 1 para 2 of Regulation 1093/2010 provide that the EBA may assist on its own initiative where there is a disagreement between competent authorities.<sup>1866</sup> Here the EBA may act as a mediator, assisting the authorities concerned in reaching an agreement. The EBA may set a time limit for conciliation between the competent authorities.1867 If the competent authorities concerned fail to reach an agreement in due time, the EBA may adopt a decision requiring them to take specific action or to refrain from action in order to settle the matter and to ensure compliance with Union law. Where a competent authority does not comply with the EBA decision, and thereby fails to ensure that a financial institution – or, in a case relating to the prevention and countering of money laundering or of terrorist financing, a financial sector operator1868 – complies with requirements directly applicable to the financial institution/financial sector operator, the EBA may adopt an indi‐ vidual decision addressed to a financial institution/financial sector operator which requires the necessary action to comply with its obligations under Union law, including the cessation of any practice.1869 In cases regarding the prevention of the use of the financial system for the purpose of money laundering or of terrorist financing, the EBA may also adopt such decision where the requirements are not directly applicable to financial sector opera‐

<sup>1866</sup> Article 19 para 1 of Regulation 1093/2010. Heed the mutual notification duties on the part of the competent authorities and the EBA, respectively, as laid down in paras 1a and 1b.

<sup>1867</sup> Article 19 para 2 of Regulation 1093/2010; for further details of this procedure as set out by the EBA itself see EBA Decision concerning rules of procedure for non-binding mediation between competent authorities, EBA/DC/2020/314.

<sup>1868</sup> For this term see Article 4 para 1a of Regulation 1093/2010 in conjunction with Article 2 of Directive 2015/849.

<sup>1869</sup> For the decision-making procedure in the EBA's Boards of Supervisors see Article 44 of Regulation 1093/2010; for a similar mechanism (regarding the authorisation of medicinal products) triggered by (qualified) disagreement between national authorities see Schütze, Rome 1412.

tors. To that effect, the EBA shall apply all relevant Union law and, where it is composed of Directives, the applicable national law transposing them. Where it is composed of Regulations explicitly granting options for MS, the EBA shall apply national law by which these options are exercised.

The regimes laid down in Articles 18 f of Regulation 1093/2010 are with‐ out prejudice to the Commission's powers under the Treaty infringement procedure.<sup>1870</sup> EBA decisions adopted under these regimes shall prevail over decisions adopted by the competent authorities on the same matter.<sup>1871</sup> In general, information on the identity of the competent authority or the financial institution/financial sector operator concerned and the main con‐ tent of the decision shall be published. Legitimate interests in confidentiali‐ ty shall be considered, though.<sup>1872</sup> For the prior notification of presumptive addressees of (future) decisions and their right to be heard see Article 39 para 1,1873 for the internal committee involved in the EBA's decision-making under Article 19 see Article 41 of Regulation 1093/2010.

The part of the procedure of Article 18 focussed on here is a hard mech‐ anism. It may entail two types of decisions – one addressed to competent authorities, the other one addressed to financial institutions. Both kinds of decisions are binding *vis-à-vis* their respective addressees. Where the competent authorities do not comply with the first kind, the EBA may make use of a competence which normally is for the competent authorities to exercise: to order financial institutions to take the necessary action to comply with their legal obligations (here: obligations under Union law). It is true that the competent authorities concerned are not the addressees of this (second) decision, but still it has the effect of ensuring their compli‐ ance. Not only is this indicated by the clarification that this power of the EBA is without prejudice to the powers of the Commission under Article 258 TFEU, but it is also made clear by the express determination that the decision prevails over any previously adopted decisions of the competent authorities on the same matter. Any (subsequent) action by the competent authorities in relation to facts which are subject to EBA-decisions under

<sup>1870</sup> For the Court's emphasis on the Commission's powers under the Treaty infringe‐ ment procedure see eg case C-359/93 *Commission v Netherlands*, paras 13 f.

<sup>1871</sup> Para 5 of Articles 18 f of Regulation 1093/2010.

<sup>1872</sup> Article 39 para 6 of Regulation 1093/2010.

<sup>1873</sup> See also Article 19 para 1b of Regulation 1093/2010; for further details of this procedure, eg the convention of an independent panel according to Article 41 para 3 of Regulation 1093/2010, see EBA Decision concerning rules of procedure for the settlement of disagreements between competent authorities, EBA/DC/2020/313.

this procedure 'shall be compatible with those decisions'.<sup>1874</sup> Due to this evocation '*à l'européenne*',<sup>1875</sup> the competent authority may not any more make effective use of its original competence to regulate the issue.<sup>1876</sup>

While the EBA in the Article 19-procedure is involved in the conciliation phase, it is not competent to adopt an act of (soft) law during this phase. Only when conciliation fails, the EBA may adopt a decision addressed to the competent authorities concerned. Where they do not comply with this decision, the EBA may – under certain circumstances – address a decision directly addressed to a financial institution/financial sector operator in order to give effect to its first decision (requiring compliance with EU law). What was said about the effects of this second decision in the context of Article 18, applies here as well. In addition to that, under Article 19 para 4 subpara 2 the EBA may even apply not directly applicable Union law and national law transposing Directives in order to ensure compliance with Union law by a financial sector operator. In conclusion, the regime laid down in Article 19 is a hard compliance mechanism.

### 2.2.1.2.6. Safeguard clauses

Article 114 para 10 TFEU allows – and, where the requirements are met: actually obliges the legislator to provide1877 – for the insertion of so-called safeguard clauses in harmonisation measures adopted on the basis of Arti‐ cle 114.1878 These safeguard clauses authorise MS to deviate – under strict conditions – from the harmonisation measure at issue, more precisely: 'to take, for one or more of the non-economic reasons referred to in Article 36 TFEU, provisional measures subject to a Union control procedure'. It is this Union control procedure, as shaped by secondary law, which constitutes a measure of ensuring compliance with Union law by the MS (normally undertaken by the Commission1879), not Article 114 para 10 TFEU itself. We

<sup>1874</sup> Article 18 para 5 and Article 19 para 5 of Regulation 1093/2010.

<sup>1875</sup> Kämmerer, Finanzaufsichtssystem 1285; see also 2.2.2.2.3. below.

<sup>1876</sup> For a similar mechanism see 2.2.1.2.3. above.

<sup>1877</sup> See Classen, Art. 114 AEUV, para 190; Korte, Art. 114 AEUV, para 57.

<sup>1878</sup> Even before the introduction of Article 100a TEEC (now: Article 114 TFEU) by the SEA the Council has inserted such safeguard clauses in secondary law; see Opinion of AG *Jacobs* in case C-359/92 *Germany v Council*, para 20. For safeguard clauses more generally see Eekhoff, Verbundaufsicht 158–164.

<sup>1879</sup> See Maletić, Harmonisation 91.

shall therefore take three exemplary acts of secondary law and analyse the relevant procedures more closely.<sup>1880</sup>

Article 23 of Directive 2001/18/EC on the deliberate release into the environment of genetically modified organisms stipulates the following: Where a MS has – 'as a result of new or additional information made available since the date of the consent and affecting the environmental risk assessment or reassessment of existing information on the basis of new or additional scientific knowledge' – 'detailed grounds' to consider that a genetically modified organism (GMO) approved under this Directive con‐ stitutes a risk to human health or the environment,<sup>1881</sup> it may provisionally restrict or prohibit the use and/or sale of that GMO on its territory.<sup>1882</sup> The MS shall immediately inform the Commission and the other MS of the measures it has taken, provide the reasons/information on which it is based, and indicate whether and, if so, how the conditions of the consent to the GMO (laid down in this Directive) should be amended or whether the consent should be terminated. Within 60 days of receipt of this infor‐ mation, the Commission shall take a positive or negative decision on the national measure in the course of an examination procedure.<sup>1883</sup>

*Vis-à-vis* the MS this procedure is a hard compliance mechanism, as the Commission output consists of a decision.

<sup>1880</sup> The Union control procedure has apparent similarities with the derogations pro‐ vided for in Article 114 paras 4 and 5 TFEU (see 2.2.1.1.3. above) and it is not entirely clear why the Treaty-makers have installed the two regimes in parallel; see Glaesner, Act 461–464; Maletić, Harmonisation 91f. For the possible co-existence of a safeguard clause and a free movement clause in one act of secondary law and their respective effects see de Sadeleer, Impact 344, with further references; see eg the free movement clause in Article 128 of Regulation 1907/2006.

<sup>1881</sup> Since environmental protection is not listed in Article 36 TFEU, it is unclear whether such a safeguard clause is in accordance with Article 114 para 10 TFEU; referring to the different viewpoints in the literature: Korte, Art. 114 AEUV, para 59 f.

<sup>1882</sup> Article 23 para 1 subpara 1 of Directive 2001/18/EC; for the case of a 'severe risk' see subpara 2.

<sup>1883</sup> Article 23 para 2 in conjunction with Article 30 para 2 of Directive 2001/18/EC in conjunction with Article 13 para 1 lit c of Regulation 182/2011; for the application of both safeguard clauses based on secondary law and of the derogation provided for in Article 114 paras 5 TFEU in the context of GMO see Rosso Grossman, Coexistence 148–150; Vos, Differentiation 167–169. For the balancing between the interest of the public to be informed and confidentiality claims see Articles 24 f of Directive 2001/18/EC.

Another safeguard clause is contained in Article 11 of Directive 2006/42/EC on machinery. It stipulates that a MS which 'ascertains' that machinery 'bearing the CE marking,[<sup>1884</sup>] accompanied by the EC declara‐ tion of conformity and used in accordance with its intended purpose or under reasonably foreseeable conditions, is liable to endanger the health or safety of persons or, where appropriate, domestic animals or property or, where applicable, the environment', it shall withdraw such machinery from the market, prohibit its placement on the market and/or its putting into service, or restrict its free movement. The MS shall immediately inform the Commission and the other MS of any such measure, thereby providing the reasons/information on which it is based.<sup>1885</sup> The Commission shall enter into consultation with the parties concerned, and shall then decide whether the measures are justified and communicate this decision to the respective MS, the other MS, and the manufacturer or 'his authorised representative'.1886 In principle, these decisions shall be published.<sup>1887</sup>

From the perspective of the MS which took the initiative (the examina‐ tion of compliance with EU law by the other addressees, the other MS and the manufacturer is not at issue in this procedure1888), the compliance mechanism at issue is hard. While for the consultation no legal output is provided for, the only act which the Commission subsequently adopts is a decision.

The third example of a safeguard clause to be discussed here is laid down in Article 129 of Regulation 1907/2006, the so-called REACH-Regu‐ lation, based on what is now Article 114 TFEU: Where a MS has 'justifiable grounds for believing that urgent action is essential to protect human health or the environment' of a substance/mixture/article, even if satisfy‐ ing the requirements of this Regulation, it may take provisional measures accordingly. In this case, the MS shall immediately inform the Commis‐ sion, the European Chemicals Agency (ECHA) and the other MS thereof,

<sup>1884</sup> See Article 5 para 4 of Directive 2006/42/EC; for the CE marking more generally see <https://ec.europa.eu/growth/single-market/ce-marking\_en> accessed 28 March 2023.

<sup>1885</sup> See in more detail Article 11 para 2 of Directive 2006/42/EC.

<sup>1886</sup> Article 11 para 3 of Directive 2006/42/EC; for the further procedure see paras 4–6 *leg cit* and Article 10 of the Directive.

<sup>1887</sup> Article 18 para 3 of Directive 2006/42/EC; see para 1 with regard to confidentiality issues.

<sup>1888</sup> The decision does have immediate effects on the machinery, it is true, but it is, if at all, only indirectly about the machinery's lawfulness.

thereby indicating the underlying reasons/information. Within 60 days the Commission shall adopt, in the course of an examination procedure,<sup>1889</sup> a decision either authorising the provisional measure for a defined period of time<sup>1890</sup> or requiring the MS to revoke it.<sup>1891</sup>

Also this procedure constitutes a hard compliance mechanism, as its only legal act addressed to the MS concerned is a (positive or negative) Commission decision.

# 2.2.1.3. Summary and résumé

While the compliance mechanisms presented here under the heading 'hard compliance mechanisms' differ from each other as regards in particular the subject matter or the EU bodies involved, there are some overarching characteristics which may explain why it is exclusively legally binding acts of EU law which are directed to non-compliant MS in these procedures.

First to the five compliance mechanisms laid down in primary law: Arti‐ cle 106 para 3 TFEU provides the Commission with the means considered necessary to ensure, in the context of public undertakings and undertakings to which MS grant special or exclusive rights, that MS comply with their duties under EU law, in particular under competition law in the wider sense.1892 With regard to the regimes set out in Article 108 para 2 TFEU on the one hand and Article 348 para 2 TFEU on the other hand, it is apparent that they provide for a shortened version of the Treaty infringement proce‐ dure. The main reason for this arguably is that possible impediments to the internal market shall be eliminated as soon as possible and that the MS concerned has already got a chance to utter its view (and the Commission

<sup>1889</sup> Article 129 para 2 in conjunction with Article 133 para 3 of Regulation 1907/2006 in conjunction with Article 13 para 1 lit c of Regulation 182/2011. For the comitology procedures under the REACH regime see Pawlik, Meroni-Doktrin 118–120.

<sup>1890</sup> See eg Commission Implementing Decision of 14 October 2013 authorising the provisional measure taken by the French Republic to restrict the use of ammonium salts in cellulose wadding insulation materials; <http://eur-lex.europa.eu/legal-co ntent/EN/TXT/PDF/?uri=CELEX:32013D0505&from=EN> accessed 28 March 2023.

<sup>1891</sup> For possible further steps in the procedure see Article 129 paras 3 f of Regulation 1907/2006; for the publication of information on evaluation results more generally see Articles 54 and 109 *leg cit*.

<sup>1892</sup> Even the compatibility with State aid law may be scrutinised on the basis of Article 106 para 2 TFEU; see eg case T-125/12 *Viasat*, para 51.

could examine this view) prior to the referral to the CJEU. This applies also with regard to Article 114 paras 4 f in conjunction with para 9 TFEU – procedures which are initiated only upon MS request.<sup>1893</sup> Therefore the pre-litigation phase as prescribed in Article 258 TFEU (see 2.1.2. and 2.1.3. above) may be considered superfluous.<sup>1894</sup> As regards Article 144 TFEU, it is arguably the *ultima ratio* competences a MS with a derogation is granted which justify (potentially) sharp action on the part of the EU. In addition to that, the MS action *a priori* is considered to be only provisional (*argumen‐ tum* 'as a precaution')<sup>1895</sup> and potentially detrimental to the functioning of the internal market.

With a view to the secondary law mechanisms listed above, most of which are based on what is now Article 114 TFEU, it appears that they either provide for exceptional MS competences, a situation of emergency or at least of urgency, and/or a prior possibility for the MS concerned to utter its view. Article 13 para 1 of Directive 2001/95/EC is about product safety (urgency), provides for a consultation of the MS concerned (point of view of MS concerned considered) and the hard law measure is based on a comitology procedure (points of view of all MS considered); Articles 70 f of Regulation 2018/1139 in the two regimes provide for an exceptional permission for MS to deviate from EU law (exceptional MS competence) and the decision-making by the Commission is coined by comitology (points of view of all MS considered).1896 Article 29 para 2 of Regulation 806/2014 deals with banking resolution, in general an area in which fast decision-making is required regularly (urgency). While the provision does not explicitly provide for a right to be heard of the MS concerned, it at least provides for its prior information. That this information is delivered on short notice (24 hours) underpins the urgency of final decision-making in this area; Article 63 of Regulation 2019/943 is about an exemption from

<sup>1893</sup> Article 108 para 2 TFEU: 'parties concerned'; Article 114 paras 4 and 5 TFEU; see Korte, Art. 114 AEUV, paras 111f, with further references; see case C-3/00 *Denmark v Commission*, para 49: 'It is therefore clear that the authors of the Treaty intended, in the interest of both the applicant Member State and the proper functioning of the internal market, that the procedure laid down in that article should be speedily concluded. That objective would be difficult to reconcile with a requirement for prolonged exchanges of information and observations'; with regard to Articles 346 f TFEU see eg Jaeckel, Art. 348 AEUV, para 7, with further references.

<sup>1894</sup> See Opinion of AG *Wahl* in case C-527/12 *Commission v Germany*, paras 25 f, with further references.

<sup>1895</sup> See Bandilla, Art. 144 AEUV, para 8.

<sup>1896</sup> Only exceptionally, it is the EASA which is competent to take the decision.

requirements laid down in EU law (exceptional MS competence); Article 18 of Regulation 1093/2010 is about an emergency situation (determined by the Council) which – according to the legislator – seems to justify cutting on the MS' right to be heard<sup>1897</sup>; according to Article 19 of Regulation 1093/2010, the EBA decision follows a conciliation procedure in the course of which the MS concerned appear to have sufficient possibilities to utter their respective view (points of view of MS concerned considered); the exemplary safeguard clauses laid down in secondary law on the basis of what is now Article 114 para 10 TFEU are by definition about safety issues (urgency) and entail an exceptional MS competence to (provisionally) deviate from Union law. The MS' viewpoints are considered either in the course of a comitology procedure or otherwise.

While it is not intended here to make generalisations beyond the hard compliance mechanisms addressed in this context, it must be reiterated that the exclusive use of hard law in our examples coincides with an exceptional MS competence to deviate from EU law (an exception to supremacy), time pressure and/or a possibility for the MS concerned to be heard prior to the adoption of the respective hard law act. After all, soft law is not the only means of giving a party to a procedure the possibility to utter its view. This conclusion is not to be understood as a confirmation of the lawfulness of the procedures laid down in secondary law (the primary law provisions, in an EU law perspective, being lawful by definition), but as a finding in the search for the rationale of setting up hard law mechanisms – both on the part of the MS when adopting primary law and on the part of the legislator when adopting secondary law.

<sup>1897</sup> See case T-510/17 *Del Valle Ruíz*, para 123, with regard to the applicability of this 'principle and fundamental right of the EU legal order' (also) to MS, and paras 146 f, with regard to the objective of ensuring financial market stability which may justify a limitation on this principle/right; case T-481/17 *SFL*, paras 250 f, with regard to a situation of urgency.

# 2.2.2. Mixed compliance mechanisms

2.2.2.1. In primary law

2.2.2.1.1. Articles 116 and 117 TFEU

Article 116 TFEU provides: Where the Commission considers 'a difference between the provisions laid down by law, regulation or administrative action' in MS to be 'distorting the conditions of competition in the internal market' and where it considers that this distortion is 'to be eliminated', it shall consult the respective MS. Article 116 addresses differences in the law of the MS which so far have not been subject to harmonisation, and which are – in principle – lawful.<sup>1898</sup> Only where they cause a distortion of the conditions of competition in the internal market which *needs to be eliminated* (that is to say: a very strong distortion1899) Article 116 may be applied. If the consultations do not result in an agreement on the elimination of the distortion, the EP and the Council shall adopt, in the ordinary legislative procedure, the 'necessary directives'. These directives should be addressed only to the MS concerned,1900 because it is not the purpose of Article 116 to harmonise, but to abolish a strong distortion of the conditions of competition in the internal market in one or more MS.<sup>1901</sup> Any other 'appropriate measures' laid down in the Treaties may be adopted. That is to say that consultations according to Article 116 para 1 do not entail a blocking effect (*Sperrwirkung*) for the application of other competence clauses, eg and in particular Article 114 TFEU.1902 The application of Article 114 TFEU may turn out to be more opportune, because it does not require

<sup>1898</sup> See Classen, Art. 116 AEUV, para 5; see Declaration No 26 which provides for a certain procedure to be followed in case a MS 'opts not to participate in a measure based on Title IV of Part Three of the [TFEU]'. In addition to that, the Declaration says, a MS may ask the Commission to examine the case on the basis of Article 116 TFEU.

<sup>1899</sup> See M Schröder, Art. 117 AEUV, para 9; Tietje, Art. 116 AEUV, para 15, both with further referenes.

<sup>1900</sup> See Classen, Art. 116 AEUV, para 31; M Schröder, Art. 116 AEUV, para 12; see also below; in favour of allowing also for EU-wide directives: Korte, Art. 116 AEUV, para 15.

<sup>1901</sup> Even though this regime is not governed by an administrative body, but by legis‐ lative actors, in spite of the limitation to administrative procedures announced under 1. above it is presented here. This is justified by the individual-concrete and thus *quasi*-administrative compliance thrust this mechanism displays.

<sup>1902</sup> See Classen, Art. 116 AEUV, paras 33 f; Tietje, Art. 114 AEUV, para 5.

proof of a distortion of the conditions of competition in the internal market which needs to be eliminated.

Article 117 TFEU is one step ahead of the situation described in Article 116. Here it is not a MS's law, regulation or administrative action currently in force which is at issue, but there is a mere 'reason to fear' that the adoption/amendment of a MS's law, regulation or administrative action may cause a distortion within the meaning of Article 116. While Article 116 forms the repressive prong of the regime, Article 117 has a preventive func‐ tion.<sup>1903</sup> A MS 'desiring to proceed' with the adoption/amendment shall consult the Commission. After that, the Commission shall recommend to the MS concerned measures 'as may be appropriate to avoid the distortion in question' (para 1). Para 2 stipulates that where a MS acts against this recommendation, other MS are not required, according to Article 116, to amend their own law in order to eliminate the distortion. This is to make clear that the active perpetrator, the MS causing the distortion, shall be obliged by Article 116, not the other (passive) MS which have not amended their *corpus* of law. This provision may also be understood as a clarification that directives according to Article 116 para 2 may only be directed against the MS actively distorting competition.1904 Furthermore, it says that where a MS ignores the Commission recommendation and thereby causes a distor‐ tion 'detrimental only to itself ', Article 116 shall not apply. While Articles 116 and 117 TFEU, leaving minor modifications apart, have been in force ever since the foundation of the EEC, their significance in practice has remained marginal.<sup>1905</sup>

A structural view suggests perceiving Article 116 and Article 117 TFEU together. When following this view, the regime is a mixed mechanism. Article 117 may result in a Commission recommendation addressed to the MS concerned, according to Article 116 – where the MS actually adopts the measure in question – a legislative act, namely a directive adopted by the EP and the Council addressed only to the MS concerned, may follow. Of course, both Articles may also be applied on their own, independently of the respective other Article. Then Article 117 is a soft mechanism, whereas Article 116 qualifies as a hard mechanism. Both provisions aim at abolishing an actual or impending non-compliance with Union law, because it can be assumed that a distortion of the conditions of competition in the internal

<sup>1903</sup> See M Schröder, Art. 117 AEUV, para 2.

<sup>1904</sup> See M Schröder, Art. 117 AEUV, para 8.

<sup>1905</sup> See Classen, Art. 116 AEUV, paras 1 and 35; M Schröder, Art. 117 AEUV, paras 1f.

market which 'needs to be eliminated' is contrary to EU law, in particular to the aims laid down in Article 3 para 3 TEU.<sup>1906</sup> More concretely, such a distortion may constitute an infringement eg of one of the fundamental freedoms or of State aid law. If this is the case, the Treaty infringement pro‐ cedure or the procedure laid down in Article 108 TFEU may be applied as alternative roads to ensure compliance with EU law. There is no derogatory relationship of speciality (*lex generalis* – *lex specialis*) between these proce‐ dures and the regime set out in Articles 116 f TFEU. Where the respective requirements are met, either regime may be applied.

# 2.2.2.1.2. Article 126 TFEU

Article 126 TFEU sets out one of the two main regimes of EU econom‐ ic policy, the excessive deficit procedure. The other regime is about the economic policy coordination and multilateral surveillance procedure as laid down in Article 121 TFEU (see 2.2.3.1.1. below). It is 'two forms of coordination', a 'softer' one and a 'harder' one.1907 While Article 121, the soft mechanism, is essentially about coordinating and monitoring the economic policy of the MS (and of the EU), Article 126 TFEU, the harder mecha‐ nism, is about remedying excessive deficits and possibly 'punishing'1908 the incriminated (Eurozone) MS. While a comprehensive analysis of the func‐ tioning in practice of these mechanisms must take into account also the pertinent secondary law,1909 in particular the Stability and Growth Pact (as amended), here it is the bare Treaty provisions which should be addressed. This is due to the aim of this chapter to clearly separate primary law from secondary law mechanisms, a distinction which is important when it comes

<sup>1906</sup> This provision of EU law is, admittedly, drafted in very broad terms. Note, how‐ ever, that the Court in the context of a Treaty infringement procedure has accepted the Commission's accusation that a MS has acted against the 'system, scheme, or spirit' of an EU measure; case C-202/99 *Commission v Italy*, para 23. The endeavour to combat such distortions is as old as the idea of a common market in Europe; see the Brussels Report on the general common market of 21 April 1956 ('Spaak Report'), 14 <http://aei.pitt.edu/995/1/Spaak\_report.pdf> accessed 28 March 2023.

<sup>1907</sup> Craig/de Búrca, EU Law 771f.

<sup>1908</sup> The inverted commas are due to the fact that not all of the measures laid down in Article 126 para 11 TFEU actually qualify as sanctions. Some are mere incentives – 'nudges' – to resolve the excessive deficit; see below; for the discussion on nudges see III.4.3.2.1. above.

<sup>1909</sup> See Article 121 para 6 TFEU, Article 126 para 14 TFEU; see also 2.2.2.2.4. below.

to the classification and legal assessment of compliance mechanisms which shall be undertaken in Part V of this work. What is more, Articles 121 and 126 TFEU as such are not mere fragments, but provide for a relatively detailed framework of the economic policy coordination and multilateral surveillance procedure and the excessive deficit procedure, respectively.

According to Article 126 para 1 TFEU, MS shall avoid excessive gov‐ ernment deficits.<sup>1910</sup> The procedure laid down in Article 126 TFEU is to 'encourage and, if necessary, compel the Member State concerned to reduce a deficit which might be identified'.<sup>1911</sup> The Commission is in charge of monitoring the budgetary situation and the government debt in the MS.<sup>1912</sup> The two criteria to be considered are laid down in para 2, the respective reference values are specified in Protocol No 12 to the Treaties; they shall not be expanded on here. Where a MS does not fulfil either or both of these criteria, the Commission *shall* or, where there is a mere risk of such an excessive deficit in a MS, *may* prepare a report,1913 upon which – in either case – the Economic and Financial Committee (EFC) shall formulate an opinion.1914 If the Commission deems that an excessive deficit in a MS exists or may occur (the latter case corresponding to the 'risk of an excessive defi‐ cit' referred to above), it shall address an opinion to the respective MS and inform the Council accordingly.1915 The Council shall then, on a proposal from the Commission and having considered observations which the MS at issue may have made, decide – on the basis of an 'overall assessment'<sup>1916</sup> – whether an excessive deficit exists.<sup>1917</sup>

Where the Council decides in the affirmative, it shall, 'without undue delay' and on a recommendation from the Commission, address recom‐ mendations to the MS concerned 'with a view to bringing that situation [ie the excessive deficit] to an end within a given period'.1918 Only where the MS does not take effective action in response to these recommendations

<sup>1910</sup> See Amtenbrink/de Haan, Governance 1088, describing this requirement as the most precisely drafted MS duty in Article 126 TFEU.

<sup>1911</sup> Case C-27/04 *Commission v Council*, para 70.

<sup>1912</sup> For the definitions of the term 'government' (and the terms 'deficit', 'investment' and 'debt') in the context of Article 126 TFEU see Article 2 of Protocol No 12.

<sup>1913</sup> Article 126 para 3 TFEU.

<sup>1914</sup> Article 126 para 4 TFEU.

<sup>1915</sup> Article 126 para 5 TFEU.

<sup>1916</sup> This assessment takes into account factors which go beyond the two criteria laid down in Article 126 para 2 TFEU; Häde, Art. 126 AEUV, para 34.

<sup>1917</sup> Article 126 para 6 TFEU.

<sup>1918</sup> Article 126 para 7 TFEU.

within the prescribed period, the Council may – on a recommendation from the Commission<sup>1919</sup> – make its recommendations public.<sup>1920</sup> Where a Eurozone MS persists in failing to comply with the recommendations, the Council may, on a recommendation from the Commission,<sup>1921</sup> give notice to the MS to take, again within a specified time limit, certain measures for the deficit reduction as proposed by the Council; this is done in the form of a Council decision.<sup>1922</sup> The Council may request the MS to submit the relevant reports, on the basis of a specific timetable.<sup>1923</sup> The applicability of the Treaty infringement procedure pursuant to Articles 258 f TFEU in the context of Article 126 paras 1 to 9 is explicitly excluded.<sup>1924</sup>

As long as a Eurozone MS<sup>1925</sup> fails to comply with the measures laid down in the Council decision, the Council may, on a recommendation from the Commission,<sup>1926</sup> apply or – subsequently – intensify one or more of the following measures:


According to para 12, the Council shall, on a recommendation from the Commission,1928 abrogate some or all of its decisions or recommendations referred to above to the extent that the excessive deficit in the respective


1928 Article 126 para 13 TFEU.

<sup>1919</sup> Article 126 para 13 TFEU.

<sup>1920</sup> Article 126 para 8 TFEU; for the requirements for such 'soft sanctions' to be effective see Hodson/Maher, Soft law 807.

<sup>1921</sup> Article 126 para 13 TFEU.

<sup>1922</sup> See Article 126 para 11 TFEU. When adopting measures relating to excessive defi‐ cits concerning Eurozone MS, the voting rights of non-euro MS in the Council shall be suspended; Article 139 para 4 TFEU.

<sup>1927</sup> Article 126 para 11 TFEU. Critically with regard to fines: Hahn/Häde, Währungs‐ recht 317.

MS has been corrected. Where recommendations have been made public, the Council – once the decision to publish the recommendations has been abrogated – shall make a public statement that an excessive deficit in the respective MS no longer exists.

From a structural point of view, and in the perspective of the MS, the following can be said about Article 126 TFEU. Contrary to the Treaty infringement procedure, at the EU level it is largely the Commission and the Council (instead of the CJEU) which act, with the Commission having monitoring tasks and important rights of initiative, but with the final 're‐ sponsibility for making the Member States observe budgetary discipline […] essentially [lying] with the Council'. <sup>1929</sup> The Commission monitors the situation in the MS and prepares a report where a MS does not comply or where there is a risk that it will not comply. This report is neither a legally binding nor a soft law act, as it merely sets out the Commission's observa‐ tions. It essentially contains information (and its analysis, respectively), not norms. The opinion which the EFC shall formulate on the report is mainly about whether or not an excessive deficit exists. It does not qualify as soft law *vis-à-vis* the MS concerned (which is regularly informed of the opinion).1930 Whether it qualifies as soft law *vis-à-vis* the Commission is du‐ bitable.1931 The subsequent opinion of the Commission which is addressed to the MS concerned and of which the Council is informed clearly qualifies as soft law. While it does not explicitly request a certain behaviour from the MS concerned, it *softly* determines the existence of an excessive deficit and hence of a situation entailing concrete legal consequences.

As regards the Council, it shall decide on a Commission proposal to state that an excessive deficit exists. Following this declarative decision, the Council may address recommendations to the MS concerned. These

<sup>1929</sup> Case C-27/04 *Commission v Council*, para 76.

<sup>1930</sup> See Hamer, Art. 126 AEUV, para 99.

<sup>1931</sup> The Committee's opinion constitutes a highly authoritative view on the question whether an excessive deficit exists or impends. It does not entail an explicit request *vis-à-vis* the Commission (not) to proceed according to Article 126. In view of the importance of its statement, however, it may be perceived as a suggestion (not) to do so. The question whether or not an excessive deficit exists is paramount for the question whether an Article 126-procedure is launched/continued. This output is not a classical expert opinion, but has a pronounced 'political' stance. It states the existence or non-existence of a legally decisive situation. Hence the author would qualify it as soft law; see Hamer, Art. 126 AEUV, para 99. For the Commission's duty to take 'fully into account' the Committee's opinion see Article 3 para 2 of Council Regulation 1467/97.

recommendations constitute soft law, that is to say they are norms which are legally non-binding. Where the MS addressed fails to comply with them, however, it may face negative effects. First, the recommendations may be published,<sup>1932</sup> second, they may be reinforced by a decision of the Coun‐ cil, setting a timetable for certain measures to remedy the deficit. In this decision, the Council may also request the MS to submit certain reports. This decision may again be reinforced by a set of other measures, most prominently fines. In a Eurozone MS's perspective, in a full procedure, ie a procedure encompassing all steps provided for in law, we therefore have the following steps of the Article 126 procedure: a soft law act (Com‐ mission opinion according to para 5) and a decision (Council decision according to para 6). This is the basis and the prerequisite for the following steps: recommendations, publication of these recommendations, decision, decision (and, in case of an intensification according to para 11: another decision). That Council decisions under Article 126 are regularly based on Commission recommendations – which the Council, within the frame of its broad discretion, is free to counteract by the measure subsequently adopted1933 – does not affect the (Eurozone) MS's perspective. Article 126 entails a mixed mechanism aimed at ensuring compliance of MS with EU law.<sup>1934</sup>

# 2.2.2.1.3. Article 271 lit a and d TFEU

Article 271 lit a and d lay down that the EIB (its Board of Directors) and the ECB (its Governing Council) shall, under certain conditions, enjoy the same powers as the Commission does under Article 258 TFEU. That is to say Article 271 lit a and d provide for two variants of the Treaty infringement procedure.

Let us start with Article 271 lit d TFEU which is to be read in conjunction with Article 35 para 6 of the Statute of the ESCB and the ECB. It states that where a NCB fails to fulfil an obligation under the Treaties or the Statute, the ECB shall deliver a reasoned opinion on the matter after having given

<sup>1932</sup> Up until this stage, *Ibáñez* claims the procedure to have a 'non-binding character'; Gil Ibáñez, Supervision 110.

<sup>1933</sup> See case C-27/04 *Commission v Council*, para 80; see also Hahn/Häde, Währungs‐ recht 316.

<sup>1934</sup> While for MS with a derogation the procedure is shorter, it still qualifies as a mixed mechanism.

the NCB concerned the opportunity to submit its observations. If the NCB concerned does not comply with the reasoned opinion within the period stated by the ECB, the latter may bring the matter before the CJEU.

This is a relatively simple procedure, inspired by Article 258 TFEU and entailing two formal steps.<sup>1935</sup> First, the ECB issues a reasoned opinion (in practice preceded by a letter of formal notice<sup>1936</sup>) to the respective NCB. This soft law act contains a clear command *vis-à-vis* the NCB. Where the latter does not follow this opinion, the ECB may file an action with the CJEU. The Court then renders a judgement – a legally binding act – to settle the matter. Article 271 lit d TFEU in conjunction with Article 35 para 6 of the Statute constitutes a mixed compliance mechanism.

Whereas the Commission pursuant to Article 258 TFEU is the general 'prosecutor' of MS violating EU law, it may not play this role where the alleged infringement by a MS is caused by the behaviour of its respective NCB.1937 This is due to the independence of the NCBs according to Article 130 TFEU. While they are independent *vis-à-vis* Union institutions, bodies, offices or agencies, and *vis-à-vis* any government of a MS or any other body, the NCBs form part of the ESCB. Pursuant to Article 129 para 1 TFEU, the ESCB shall be governed by the decision-making bodies of the ECB which is why the NCBs may receive, and shall comply with respectively, instruc‐ tions from the ECB.1938 They are not independent *vis-à-vis* the ECB.<sup>1939</sup> Therefore, it is consistent with the logic of this independence regime<sup>1940</sup> that violations of Union law committed by NCBs are not 'prosecuted' by the Commission (according to Article 258 TFEU), but by the ECB, which

<sup>1935</sup> See Gramlich, Wirtschafts- und Währungspolitik 625; Hahn/Häde, Währungs‐ recht 157. The comments made under 2.1. above in the context of Article 258 TFEU *mutatis mutandis* apply here as well.

<sup>1936</sup> See Hahn/Häde, Währungsrecht 160; see 2.1. above.

<sup>1937</sup> See Schima, Art. 271 AEUV, para 13; see also Karpenstein, Art. 271 AEUV, paras 26 f.

<sup>1938</sup> Article 14 para 3 of the Statute of the ESCB and of the ECB.

<sup>1939</sup> See Hahn/Häde, Währungsrecht 218.

<sup>1940</sup> But possibly not with the logic of the regular Treaty infringement procedure which lays down the liability of the MS also for the actions of independent bodies (eg national courts); see also Hahn/Häde, Währungsrecht 160, both with further references.

may bring the matter before the Court.1941 An (analogous) application of the sanctions regime according to Article 260 TFEU is excluded.<sup>1942</sup>

A similar regime applies with regard to the European Investment Bank – which also is a legal person of its own<sup>1943</sup> – according to Article 271 lit a TFEU.<sup>1944</sup> The main addressees of the acts adopted under this procedure are the MS. Whether Article 260 TFEU applies by analogy is contested.<sup>1945</sup>

### 2.2.2.2. In secondary law

### 2.2.2.2.1. Article 63 of Directive 2019/944

The mechanism addressed here is about compliance with Commission network codes and guidelines referred to in Directive 2019/944 – which is based on Article 194 para 2 TFEU – or in Chapter VII of Regulation

1942 See Ehricke, Art. 271 AEUV, para 21; Schwarze/Wunderlich, Art. 271 AEUV, para 11.


<sup>1941</sup> See Ehricke, Art. 271 AEUV, para 20, also with regard to the fact that Article 271 lit d TFEU explicitly addresses the NCBs, not 'the MS'. In case of the EIB the powers are granted in respect of MS (lit a *leg cit*), in case of the ECB in respect of NCBs (lit d *leg cit*); trying to explain the latter specificity: Potacs, Zentralbanken 38.

2019/943. <sup>1946</sup> These network codes and guidelines are legally binding (dele‐ gated or implementing acts).<sup>1947</sup> A national regulatory authority (herein‐ after: 'MS authority') or the Commission may request an opinion from the ACER on the compliance with these network codes and guidelines of a de‐ cision taken by a(nother) MS authority.<sup>1948</sup> Within three months the ACER shall provide its opinion to the requesting body (MS or Commission) and to the MS authority which has taken the respective decision. Where the MS authority concerned does not comply with the opinion within four months, the ACER shall inform the Commission.

Where the decision of the MS authority is relevant for cross-border trade, another MS authority may inform the Commission where it deems this decision not to be in compliance with the Commission network codes or guidelines. (In this case the Commission, not the ACER, is the first point of contact.)

Where the Commission, within two months of having been informed by the ACER or a MS authority, or – on its own initiative – within three months from the date of the decision, finds that the decision 'raises serious doubts as to its compatibility with the network codes and guidelines', it may decide to examine the case further.1949 It shall then, within four months of the decision to examine the case further, issue a final decision a) not to raise objections against the decision of the MS authority, or b) to require this authority to withdraw its decision for lack of compliance with the network codes or guidelines. In the latter case the MS authority shall withdraw its decision within two months and inform the Commission thereof. Where the Commission has not taken a decision to examine the case further or a

<sup>1946</sup> For similar mechanisms established with regard to other forms of energy see Gundel, Energieverwaltungsrecht, para 37.

<sup>1947</sup> See eg Article 58 para 1 of Regulation 2019/943. For the instrument of Commission guidelines more generally see W Weiß, Leitlinien(un)wesen; see also W Weiß, Verwaltungsverbund 149–151.

<sup>1948</sup> Article 63 para 1 of Directive 2019/944. The scope of scrutiny only covers these network codes and guidelines as tertiary law, not (also) the pertinent secondary law; see (for the predecessor regime) Gundel, Energieverwaltungsrecht, para 47 (fn 221); differently: the scope of scrutiny of the compliance mechanism laid down in Article 6 paras 5–7 of Regulation 2019/942; see 2.2.3.2.1. above.

<sup>1949</sup> For the possibility of the MS authority and of the parties to the proceedings to submit their observations see Article 63 para 5 of Directive 2019/944.

final decision within the respective periods, it shall be deemed not to have raised objections to the decision of the MS authority.<sup>1950</sup>

This regime involves a number of actors and different kinds of output. For our purposes, the essentials of the procedure are the following: Upon request by a MS authority or the Commission, the ACER shall issue an opinion on the compliance with the Commission network codes and guide‐ lines of the decision of a(nother) MS authority. This opinion is (also) addressed to the latter. Where it does not comply with the opinion, the Commission may, first, decide to examine the case further and, in this case, as a second step, may adopt a final decision directed to the MS authority at issue. That means that the MS authority first receives an EU soft law act. Where it does not comply, the Commission may possibly adopt a first decision (to examine the case further). If it has done so, it may adopt a second (final) decision addressed to the MS authority. Thereby the Commission can detect potential non-compliance with its network codes and guidelines by a MS authority and, where it turns out to be actual non-compliance, determine this failure on the part of the MS concerned. As this regime involves acts of both soft law and law which are addressed to an (allegedly) non-compliant MS (authority), it constitutes a mixed compliance mechanism. Only in the variant according to Article 63 para 4, according to which the Commission is addressed by a MS authority without the ACER rendering its opinion beforehand, the procedure is to be called a hard mechanism.<sup>1951</sup>

### 2.2.2.2.2. Articles 22 f and 28 of Council Regulation 2015/1589

This regime laid down in Council Regulation 2015/1589, based on Article 109 TFEU, is about the review of existing aid schemes pursuant to Article 108 para 1 TFEU.1952 Where the Commission considers that such an existing aid scheme is not or no longer compatible with the internal market, it

<sup>1950</sup> Article 63 paras 2–8 of Directive 2019/944; for the very similar predecessor mech‐ anism and for the German transposition see Koenig, Entflechtungszertifizierung 506 ff.

<sup>1951</sup> The exclusion of the ACER in this variant is mitigated by the fact that the ACER may address an opinion to the Commission on any matter, be it upon request or on its own initiative; Article 3 para 1 of Regulation 2019/942.

<sup>1952</sup> For this term see Article 1 lit b of Council Regulation 2015/1589; see also eg W Cremer, Art. 108 AEUV, para 3, with further references.

shall inform the MS concerned of its 'preliminary view' and shall give it op‐ portunity and time (one month; extendable) to submit its observations.<sup>1953</sup> Where these observations cannot dispel the concerns of the Commission, it shall address a recommendation to the MS, thereby proposing in par‐ ticular amendments, procedural requirements or the abolition of the aid scheme.<sup>1954</sup>

Where the MS accepts the recommendations, it shall inform the Com‐ mission thereof, and the Commission shall record that finding and inform the MS in turn. By this recorded acceptance the MS shall be bound to implement the recommended measures. Where the MS does not accept the recommendations, the Commission – if it still considers the recommen‐ ded measures to be necessary – shall initiate proceedings in accordance with Article 108 para 2 TFEU and Article 4 para 4 of Council Regulation 2015/1589. If the Commission finds that the aid scheme is not compatible with the internal market, or that it is being misused, it shall decide that the MS concerned shall abolish or alter such aid within a certain period of time (to be determined by the Commission1955).1956 Where the MS concerned does not comply with a conditional or negative decision, the Commission may refer the case to the CJEU directly, following which the CJEU shall render a judgement.<sup>1957</sup>

The sequence of recommendation and decision envisaged in this proce‐ dure is relatively common in EU compliance mechanisms (and, what is more, already sketched out in Articles 108 paras 1f TFEU). Where the MS does not follow (in this context that means: 'accept') the soft law act, it may eventually be forced to do so by law – in the form of a decision. A specificity of this procedure is that where the MS *accepts* the measures set out in the recommendation the Commission will record this acceptance, whereby the MS shall be legally bound. This does not change the soft law character of

<sup>1953</sup> Article 21 para 2 of Council Regulation 2015/1589.

<sup>1954</sup> Article 22 of Council Regulation 2015/1589.

<sup>1955</sup> See Article 108 para 2 TFEU.

<sup>1956</sup> For the publication of this decision in the OJ see Article 32 para 2 of Council Regulation 2015/1589; for other possible decisions see Article 9 *leg cit*.

<sup>1957</sup> Article 28 para 1 of Council Regulation 2015/1589 in conjunction with Article 108 para 2 TFEU. For the possibility to sanction a MS not complying with a Court judgement in this context (in accordance with Article 260 TFEU), see Article 28 para 2 of Council Regulation 2015/1589.

the recommendation.1958 What makes its content binding is the recorded acceptance of its addressee (agreed law).<sup>1959</sup> That the recommendation is also referred to as 'proposal'<sup>1960</sup> does not entail special effects. Since it is addressed to a MS (not: to the Council), the varying designation of the soft law act is insignificant.<sup>1961</sup> The possibility of the Commission to directly refer the case to the CJEU, thereby skipping the pre-litigation procedure as laid down in Article 258 TFEU, is in accordance with primary law, namely Article 108 para 2 TFEU (see 2.2.1.1.2. above). The Court may then add another hard law act in this – all in all: mixed – procedure.

# 2.2.2.2.3. Article 17 of Regulation 1093/2010

The regime to be discussed here is laid down in Article 17 of Regulation 1093/2010. In the following, it will be presented and analysed with a focus on the EU legal acts which may be adopted in the course of this proce‐ dure.<sup>1962</sup>

Article 17 of Regulation 1093/2010 provides for a possibility for the EBA to react to a breach of Union law by the competent national authorities.<sup>1963</sup>

<sup>1958</sup> For the bindingness of the determination of existing aid see case T-354/05 *Télévi‐ sion française*, paras 60–81; on the slightly different regime provided for in Article 9 of Regulation 1/2003 see case C-441/07P *Alrosa*, paras 47–50.

<sup>1959</sup> See Rusche, Art. 108 AEUV, para 11; for the legal bindingness of these agreements – normally dubbed 'guidelines', 'disciplines' or 'frameworks' – see H Hofmann, Rule-Making 165–169. For the consequences of non-compliance of a MS with a recommendation it has previously accepted see Rusche/Micheau/Piffaut/Van de Casteele, State Aid, para 17.515 (with regard to the predecessor provisions, Articles 18 f of Council Regulation 659/1999).

<sup>1960</sup> See the headings of Articles 22 f of Council Regulation 2015/1589. In the German version the German word for 'recommendation' is not used at all. Other language versions – *raccomandazione* (Italian), *recommandation* (French), *recomendación* (Spanish) – of the provision, however, suggest that it actually refers to a recom‐ mendation pursuant to Article 288 TFEU.

<sup>1961</sup> See case T-354/05 *Télévision française*, para 65.

<sup>1962</sup> For a more comprehensive analysis of this procedure see eg Michel, Gleichgewicht 243–248; Weismann, Agencies 133–138; for further procedural details see Article 39 of Regulation 1093/2010 and EBA Decision concerning rules of procedure for investigation of breach of Union law, EBA/DC/2020/312.

<sup>1963</sup> The Commission itself has framed this compliance mechanism in the context of ensuring an independent application of EU law; see Commission, 'EU law: better results through better application' (Communication), 2017/C 18/02, 3 f; for an EBA request to a competent national authority for investigation related to the

In this context, 'Union law' means the pertinent acts of secondary law as laid down in Article 1 para 2 of Regulation 1093/2010, including the regula‐ tory and implementing technical standards adopted by the Commission (with the EBA being strongly involved). The alleged breach<sup>1964</sup> (including the non-application) of Union law shall be investigated by the EBA, after having informed the competent authority concerned, on its own initiative (eg based on well-substantiated information from third parties) or upon a request from one or more of the following bodies: a competent author‐ ity, the EP, the Council, the Commission or the Banking Stakeholder Group.<sup>1965</sup> In deciding whether or not to open an investigation, the EBA disposes of a discretion comparable to that of the Commission under Article 258 TFEU.<sup>1966</sup> If the EBA decides in the affirmative, the competent authority concerned shall provide all the information the EBA considers necessary for its investigation. The EBA may also request information from other competent authorities. No later than two months from initiating its investigation, the EBA may address a recommendation to the competent authority concerned, setting out the action necessary to comply with Union law.1967 Within ten working days, the competent authority shall then inform

prevention and countering of money laundering and of terrorist financing (which may also result in a procedure pursuant to Article 17) see Article 9a of Regulation 1093/2010.

<sup>1964</sup> That in practice not any breach, but only a qualified breach of Union law is pursued by the EBA under Article 17 follows from Annex 2 to the pertinent Decision of the EBA's Board of Supervisors EBA DC 054. According to Recital 27 of Regulation 1093/2010, the 'mechanism should apply in areas where Union law defines clear and unconditional obligations'; for the wide interpretation of this term by the Court see case C-501/18 *Balgarska Narodna Banka*, para 88.

<sup>1965</sup> See Article 2 of EBA Decision concerning rules of procedure for investigation of breach of Union law, EBA/DC/2020/312; see also case T‑660/14 *SV Capital*, paras 69 f.

<sup>1966</sup> Case T‑660/14 *SV Capital*, paras 47 f; confirmed by C-577/15P *SV Capital*, para 40; Joint Board of Appeal, *C v EBA*, BoA-D-2022–01, paras 65–69; see also Simoncini, Regulation 161; for the EBA-internal division of powers regarding this question see Article 6 of EBA Decision concerning rules of procedure for investigation of breach of Union law, EBA/DC/2020/312.

<sup>1967</sup> For the decision-making procedures of the EBA – which apply, *mutatis mutandis*, also to the adoption of recommendations – see Article 39 of Regulation 1093/2010; for the contents of the (draft) recommendation in more detail see Article 5B para 6 of the Decision of the EBA's Board of Supervisors EBA DC 054. For the effect of this particular recommendation on national bodies see case C‑501/18 *Balgarska Narodna Banka*, paras 78–81; for the possible 'engagement' between the EBA and

the EBA of the steps it has taken or intends to take to ensure compliance with Union law.

Where the competent authority has not complied with Union law within one month, the Commission may, upon information by the EBA or on its own initiative, issue a formal opinion, thereby taking into account the EBA's recommendation, which requires the competent authority to take the action necessary to comply with Union law. The Commission shall do so no later than three months (possibly extended by one month) of the adoption of the EBA recommendation. The competent authority shall, within ten working days,<sup>1968</sup> inform the Commission and the EBA of the steps it has taken or intends to take to comply with the formal opinion.

Where the competent authority has not complied with the formal opin‐ ion in due time and where it is 'necessary to remedy, in a timely manner, such non-compliance in order to maintain or restore neutral conditions of competition in the market or ensure the orderly functioning and integrity of the financial system',1969 Regulation 1093/2010 provides for a further instrument. Explicitly without prejudice to the powers of the Commission under the Treaty infringement procedure, the EBA may, where the relevant requirements of the legislative acts at issue are directly applicable to finan‐ cial institutions/financial sector operators, adopt an individual decision addressed to a financial institution or, in cases regarding the prevention and countering of money laundering and of terrorist financing, to another financial sector operator which requires the necessary action to comply with its obligations under Union law, including the cessation of any prac‐ tice.1970 The EBA decision, if it is taken in the first place,1971 shall be in

the competent authority prior to the adoption of a recommendation see Article 17 para 2a of Regulation 1093/2010.

<sup>1968</sup> Article 17 para 6 of Regulation 1093/2010 ('period of time specified therein') ap‐ pears to suggest that the Commission may also allow for a longer period of time (arguably taking account of the complexity of the case).

<sup>1969</sup> See also Böttner, Mechanism 184 f.

<sup>1970</sup> For the publication requirements regarding EBA decisions taken under Article 17 see Article 39 para 6 of Regulation 1093/2010.

<sup>1971</sup> The EBA is not obliged to take action, even if the Commission's formal opinion was not complied with. For another procedure in which an EU institution provides a soft assessment (with high authority), upon which an EU agency may take a decision, see the ECB's 'failing or likely to fail' (FOLTF) assessment and the reso‐ lution decision the SRB may take after that (Article 18 of Regulation 806/2014); for the rationale of this sharing of tasks the Court held that 'the SRB, while not bound by the ECB's examination and view, did not err in law by taking the latter

conformity with the formal opinion of the Commission.1972 Article 17 para 6 subpara 2, introduced as part of a reform of Regulation 1093/2010 (by the end of December 2019), provides for a deviating procedure where the prevention of the use of the financial system for the purpose of money laundering or of terrorist financing is concerned and where the relevant requirements of the legislative acts at issue are not directly applicable to financial sector operators. Here the EBA may adopt a decision requiring the competent authority to comply with the Commission's formal opinion. If the competent authority does not comply with the EBA decision, the EBA may address a decision to the financial sector operator. To that effect, the EBA shall apply all relevant Union law and, where it is composed of Directives, the applicable national law transposing them. Where it is composed of Regulations granting options for MS, the EBA shall apply the national law by which these options are exercised.

In terms of output, this procedure is threefold,1973 exceptionally (in case of para 6 subpara 2) fourfold. An EBA recommendation to the competent authority may be followed – reinforced, as it were – by a formal opinion of the Commission. Where the competent authority does not comply with this opinion,1974 either, the EBA may – under certain conditions – address an individual decision to a financial institution/financial sector operator to enforce its legal view. Under para 6 subpara 2 the EBA may do so even where the relevant legislative acts are not directly applicable to financial

as its basis, since the ECB was the institution best placed to carry out the FOLTF assessment in respect of the applicant'; case T-280/18 *ABLV*, para 108. Applying these thoughts to the EBA regime at issue, we could say that the Commission – due to its competence and experience under the Treaty infringement procedure – is well placed as an actor involved in a mechanism which displays some similarities with the Treaty infringement procedure.

<sup>1972</sup> See Article 17 para 6 subpara 3 of Regulation 1093/2010. For the effects of the deci‐ sions and the public report on non-compliant competent authorities see Article 17 paras 7 f of Regulation 1093/2010.

<sup>1973</sup> See Recital 28 of Regulation 1093/2010: 'three-step mechanism'.

<sup>1974</sup> The (misleading) wording of Article 17 para 7 subpara 2 of Regulation 1093/2010 does not render the formal opinion binding upon its addressee. It merely clarifies that where a formal opinion is the last step in a concrete procedure, the compe‐ tent authority shall (or rather: should) comply with it. However, where an EBA decision was subsequently adopted, the competent authority (even if it is not the addressee) shall ensure compliance with this decision (not the formal opinion). According to Regulation 1093/2010, there should not be significant differences in the approaches taken in the two acts anyway (see below).

sector operators, but only after having addressed, in vain, the competent authority.

While the Commission's involvement arguably shall increase the political legitimacy of the procedure, the meaning of its formal opinion requires some more attention. Article 17 para 4 stipulates that the Commission shall 'take into account' the EBA recommendation when drafting its formal opinion. This means that it may deviate from the recommendation, eg in order to do justice to new arguments or evidence brought forward by the competent authority concerned.<sup>1975</sup> It may also decide not to adopt a formal opinion at all (*argumentum* 'may'), in which case the procedure comes to a halt. Where the EBA adopts a decision subsequent to a formal opinion, it shall, according to Article 17 para 6 subpara 3, be 'in conformity with' this opinion. This certainly suggests a larger degree of accordance than the phrase 'take into account'. To the extent that the formal opinion is legally binding (only) upon the EBA, it ensures the Commission a leading role in the procedure.1976 Conformity does not, however, mean identity. The EBA does have some room for manoeuvre, the scope of which has to be concretised case by case. Otherwise, the legislator could as well have empowered the Commission to adopt a (binding) decision instead of a formal opinion, the third step – the EBA decision – being abolished as superfluous.<sup>1977</sup>

This mixed procedure, if applied in full, entails two soft law acts – the EBA recommendation and the formal opinion of the Commission – and the EBA decision. While the recommendation and the formal opinion are both legally non-binding for the competent authority addressed, the formal opinion may have a higher *de facto* authority: First, because it stems from the Commission which is also competent to initiate Treaty infringement procedures (which it may do independently of an Article 17-procedure), and, second, because the formal opinion, unlike the EBA recommendation (with regard to the formal opinion), largely determines the content of its follow-up (the EBA decision). Except for the specific first case of Article 17 para 6 subpara 2, the individual decision is not addressed to the competent authority, it is true, but indirectly – via an evocation '*à l'européenne*' <sup>1978</sup> – it

<sup>1975</sup> See Michel, Gleichgewicht 245.

<sup>1976</sup> See Michel, Gleichgewicht 247, with further references.

<sup>1977</sup> Similarly: case T-317/09 *Concord*, para 52, with regard to preliminary output in a multiphased procedure.

<sup>1978</sup> Kämmerer, Finanzaufsichtssystem 1285; see also 2.2.1.2.5. above.

does away with the competent authority's breach of Union law and it limits its competences in this respect.<sup>1979</sup>

The existence of Article 17 para 6 subpara 2 proves the legislator's convic‐ tion that a decision directed to the competent authority is the more moder‐ ate interference with MS' decision-making power. Nevertheless, it did not take this route in the remaining cases. There, the formal opinion of the Commission, if not complied with by the competent authority addressed, may be directly followed by an EBA decision addressed to the financial institution/financial sector operator at issue.

# 2.2.2.2.4. The excessive imbalance procedure laid down in Regulations 1176/2011 and 1174/2011

Regulation 1176/2011 on the prevention and correction of macroeconomic imbalances is based on Article 121 para 6 TFEU, and hence is to lay down detailed rules for the multilateral surveillance procedure referred to in Article 121 paras 3 f TFEU. Among other things, this Regulation sets out the excessive imbalance procedure to correct 'severe imbalances, including im‐ balances that jeopardise or risks [sic] jeopardising the proper functioning of the economic and monetary union'. <sup>1980</sup> Upon a recommendation from the Commission which has previously carried out an in-depth review of a certain MS,1981 the Council shall adopt a recommendation (mentioned in Article 121 para 4 TFEU; see 2.2.3.1.1. below) establishing the existence of an excessive imbalance in that MS, and recommending the MS concerned to take corrective action.<sup>1982</sup>

Upon such a recommendation, a MS shall submit a corrective action plan to the Council and the Commission within a certain deadline. This

<sup>1979</sup> See Eekhoff, Verbundaufsicht 142, describing comparable evocation rights of the Commission as 'ein schärferes und zugleich wirksameres Intstrument als eine ver‐ bindliche Aufforderung' [a harsher and at the same time more effective instrument than a binding request].

<sup>1980</sup> Article 2 para 2 of Regulation 1176/2011.

<sup>1981</sup> For the in-depth review see Article 5 of Regulation 1176/2011; for the other EU actors to be informed of the Commission's assumption that a certain MS is affected by excessive imbalances see Article 7 para 1 *leg cit*.

<sup>1982</sup> In practice, the Commission has detected excessive imbalances in certain MS, but – in reaction thereto – has not applied the corrective arm. It merely intensified surveillance; Pierluigi/Sondermann, Macroeconomic imbalances 40.

plan shall set out the specific policy actions the MS has implemented or intends to implement and shall include a timetable for those actions.<sup>1983</sup> It shall be consistent with the BEPG which renders the latter legally binding for the purposes of this procedure.<sup>1984</sup> The Council and the Commission shall then, to put it short, assess the corrective action plan, and where the Council – upon a Commission recommendation – considers it sufficient, it shall endorse it by means of a recommendation setting out the details of the implementation of the plan. Where the Council, upon a Commission rec‐ ommendation, considers the plan insufficient, it shall adopt a recommenda‐ tion to the MS to submit, within two months as a rule, a new corrective action plan.<sup>1985</sup>

The Commission shall monitor implementation of the Council's approv‐ ing recommendation, for which purpose the MS shall submit progress reports (to be published by the Council).1986 The Commission shall then provide a – later to be published – report to the Council on whether or not the MS has taken corrective action in accordance with the Council recommendation. Where the MS has not done so, the Council – on a recommendation from the Commission – shall adopt a decision (applying reverse qualified majority voting1987) establishing non-compliance, together with a recommendation setting new deadlines for corrective action.<sup>1988</sup> Otherwise – ie where the MS has taken the corrective action recommended – the excessive imbalance procedure shall be considered to be on track and shall be held in abeyance, the Commission continuing to monitor. Where a MS is no longer affected by excessive imbalances, the Council, on a recommendation from the Commission, shall abrogate its recommenda‐

<sup>1983</sup> Article 8 para 1 of Regulation 1176/2011.

<sup>1984</sup> It ought to be emphasised that the BEPG, while setting out clear objectives, leave some discretion to the MS as to how to reach these objectives. The BEPG do not form part of the Stability and Growth Pact, although they are in places mentioned in its context. On the (wider) scope of the BEPG see Deroose/Hodson/Kuhlmann, Guidelines 828; for the consideration of country-specific recommendations in the excessive imbalances procedure see Bénassy-Quéré/Wolff, Imbalances 31.

<sup>1985</sup> For the publication requirements see Article 7 para 4 of Regulation 1176/2011.

<sup>1986</sup> For the details of the monitoring procedure, for the possility of the Council to amend its recommendations and for the possible revision of the corrective action plan by the MS see Article 9 para 4 of Regulation 1176/2011; for an enhanced surveillance mission the Commission may carry out see para 3 *leg cit*.

<sup>1987</sup> Article 10 para 4 subpara 2 of Regulation 1176/2011; for the application of reverse (qualified) majority voting in the Council see III.4.4. above.

<sup>1988</sup> For the information of the European Council and publication requirements see Article 10 para 4 subpara 1 of Regulation 1176/2011.

tions in accordance with Article 11 of Regulation 1176/2011 and publish this information (for the requirement of a *contrarius actus* see III.3.8. above).

In terms of output *vis-à-vis* the MS, the procedure looks as follows: Council recommendation regarding the existence of an excessive imbal‐ ance; following a corrective action plan submitted by the MS: Council recommendation on the details of implementation or Council recommen‐ dation to submit a new action plan; possibly Council decision establishing non-compliance and recommendation setting new deadlines; Council rec‐ ommendation abrogating its recommendations. All of this Council output can be adopted only upon an appropriate Commission recommendation. While the Council has discretion when acting on these Commission rec‐ ommendations,<sup>1989</sup> it is bound by the procedural route the Commission has taken.<sup>1990</sup> Therefore it appears that the Council may not, for example, adopt a recommendation that the corrective action plan is insufficient according to Article 8 para 3 where the Commission has recommended to consider it sufficient according to para 2. Where the required majority for a decision is not achieved, no decision is taken.<sup>1991</sup>

It is to be noted that all Council measures aimed at steering MS be‐ haviour are recommendations, ie legally non-binding. The only hard law measure – the decision according to Article 10 para 4 – merely establishes the MS's non-compliance, but does not require action. Action is required by a Council recommendation accompanying this decision.

The appearance of weakness of this regime is done away with by the sanctions regime to correct excessive macroeconomic imbalances as laid down in Regulation 1174/2011, based on Article 136 in conjunction with Article 121 para 6 TFEU, which applies only to Eurozone MS. According to Article 3 of this Regulation, the Council shall impose an interest-bearing deposit upon a recommendation of the Commission, where it has adopted a decision establishing non-compliance in accordance with Article 10 para 4 of Regulation 1176/2011.1992 The Council shall, again on a recommendation of the Commission, impose an annual fine where a) two successive Council

<sup>1989</sup> For the political expectation that the Council follows the Commission recommen‐ dations see (with regard to related regulations of the 'Six Pack') Schulte, Art. 121 AEUV, para 59.

<sup>1990</sup> See case C-27/04 *Commission v Council*, paras 80 f.

<sup>1991</sup> See case C-27/04 *Commission v Council*, para 31. As a way out of this predicament, the legislator has introduced reverse (qualified) majority voting in some proce‐ dures.

<sup>1992</sup> For the amount of this deposit see Article 3 paras 5 f of Regulation 1174/2011.

recommendations in the same imbalance procedure are adopted according to Article 8 para 3 of Regulation 1176/2011 and the Council considers that the MS has submitted an insufficient corrective action plan or b) two successive Council decisions in the same imbalance procedure are adopted establishing non-compliance in accordance with Article 10 para 4 of Regu‐ lation 1176/2011. In the latter case the already imposed deposit is converted into an annual fine.<sup>1993</sup> The sequence of two acts of the same legal quality content-wise addressing the same issue does not mean that the Council in the excessive imbalance procedure may adopt another recommendation or another decision to repeat its view. Rather, according to Article 8 para 3, the Council by means of a recommendation may request the submission of a new corrective action plan which is then again subject to scrutiny, and hence – if the Council is not satisfied with it – may be followed by a new recommendation to submit another corrective action plan.1994 In the case of Article 10 para 4 the Council may adopt a decision establishing non-compliance with the recommendation, and may set – by means of a recommendation – a new deadline. Where this deadline is not complied with either, the Council may establish this by means of a (second) decision. In that sense, the term 'successive' used in Article 3 para 2 of Regulation 1174/2011 does not exclude, in the second case, the adoption of a recommen‐ dation in between the two decisions.

The Council decisions on the imposition of sanctions are adopted by reverse qualified majority voting. <sup>1995</sup> Article 3 para 3, last sentence of Regu‐ lation 1174/2011 says: 'The Council may decide, by qualified majority, to amend the recommendation'. This means that the Council may, with a qualified majority, amend the Commission recommendation and thereby the content of its (future) decision.1996 This is to mitigate the shifting of

<sup>1993</sup> For the potential (partial) return of the paid amount see Article 3 para 7 of Regulation 1174/2011. For the exceptional reduction or cancellation of sanctions see para 6 *leg cit*.

<sup>1994</sup> See Obwexer, System 223.

<sup>1995</sup> In practice, no sanctions have been imposed so far; see Koll/Watt, Macroeconomic Imbalance 57.

<sup>1996</sup> See Commission Proposal COM(2010) 525 final, 6 f, with regard to the originally provided Commission *proposal* and the accordingly envisaged applicability of Article 293 para 1 TFEU.

power from the Council to the Commission which is brought about by this procedure.<sup>1997</sup>

Thus, the legal non-bindingness of the recommendation according to Article 8 para 3 and the lack of a command of the decision according to Article 10 para 4 of Regulation 1176/2011 are compensated for by the sanctions regime, at least with regard to the Eurozone MS. Legally speak‐ ing, this neither makes the recommendation binding nor does it make the declarative decision a command, but *de facto* it substantially increases their respective authority and the likelihood of compliance or at least of attempts to remedy the stated non-compliance by their respective addressees.<sup>1998</sup>

### 2.2.2.2.5. Article 7 para 4 of Regulation 806/2014

With regard to general information on Regulation 806/2014, see 2.2.1.2.3. above. According to Article 7 para 4, the SRB may – where necessary to ensure the consistent application of high resolution standards under Regu‐ lation 806/20141999 – address a warning to the relevant national resolution authority where it deems a national authority's draft decision to any entity or group (which the national authorities are principally competent to adopt in accordance with Article 7 para 2) violates Regulation 806/2014 or its – the SRB's – general instructions according to Article 31 para 1 lit a *leg cit*. <sup>2000</sup> The SRB shall be informed by the national resolution authorities of any measure according to para 3 (eg resolution plans or resolvability assess‐ ments) to that end.2001 The Board may also, 'in particular if its warning

<sup>1997</sup> See, with regard to a similar provision, Obwexer, System 224 f, with a further reference.

<sup>1998</sup> See Recital 11 of Regulation 1174/2011.

<sup>1999</sup> These standards are not a specific type of (soft law) act, but standards in the gener‐ al meaning of the term, that is: the relevant EU law in its correct interpretation; see Zavvos/Kaltsouni, Mechanism 127.

<sup>2000</sup> Note that, according to the wording of Article 7 para 4 lit a of Regulation 806/2014, only the (binding) general instructions are included, not: the (non-binding) guidelines which are also mentioned in Article 31 para 1. For the duty of national resolution authorities to submit their draft decisions to the SRB see Article 31 para 1 lit d.

<sup>2001</sup> Article 7 para 3 (penultimate subparagraph) of Regulation 806/2014. While this provision merely speaks of 'measures', it appears that what is actually meant are 'draft measures' (*argumentum* 'to be taken', 'closely coordinate with the Board when [ie: before] taking those measures').

[…] is not being appropriately addressed', *sua sponte* or upon request by the national authority concerned, exercise directly all of the relevant powers under Regulation 806/2014 also for entities or groups for which in principle the national authorities are competent under Article 7 para 3.<sup>2002</sup>

The character of the warning deserves further attention. While there is no general definition of this term in Union law, and no specific one in Regulation 806/2014, it is clear that in this case it qualifies as a soft law act. After all, the warnings are about non-compliance with Union law and they should be 'appropriately addressed' by its recipients, namely the national authorities. Both characteristics strongly convey that the warning suggests compliance in more or less detailed terms.

This is not a perfect mixed compliance mechanism, as the evocation (para 4 lit b) by the SRB may not only follow (non-compliance with) a warning according to lit a (*argumentum* 'in particular'). Since lit a and lit b are therefore potential alternatives, one could also perceive them separately as one soft and one hard compliance mechanism.2003 The term 'in particu‐ lar' and the common legal basis in one paragraph suggest, however, that lit a and lit b were rather conceptualised as *one* regime.

This regime involves a soft law act, a warning, which is sent to the national authority and which may be followed – where the warning has not been 'appropriately addressed' by the national authority – by a hard law act by means of which the SRB attracts competences of the national authorities, to ensure that they are exercised in compliance with Union law. As explained, the hard law act may also stand alone, without a preceding warning, but it shall 'in particular' be adopted where the warning has not been duly considered. Whether the SRB's decision suffices to ensure compliance with Union law depends on whether the national resolution authority has already adopted the measure at issue. If so, the SRB may take a decision *vis-à-vis* the entity or group according to Article 7 para 3 of Regulation 806/2014.2004 If the relevant national measure is still a draft, the national authority has – *qua* SRB decision – lost its competence to adopt it.

<sup>2002</sup> For the possibility of a participating MS to transfer these competences to the Board by a decision see Article 7 para 5 of Regulation 806/2014; see also J-H Binder, Resolution 137. In the context of the preparation of draft resolution plans and draft group resolution plans relating to specific entities or groups, the SRB may – in what is to be called a hard mechanism – address (binding) instructions to its national counterparts; Article 8 para 3 of Regulation 806/2014.

<sup>2003</sup> Note that the decision according to lit b is preceded by 'consulting' with the national authority.

### 2.2.2.2.6. Article 25 of Regulation 2016/796

Regulation 2016/796 on the European Union Agency for Railways (ERA) is based on Article 91 para 1 TFEU. The mechanism laid down in Article 25 which is at issue here builds on the MS' obligations under Article 8 para 4 of Directive 2016/798 and under Article 14 para 5 of Directive 2016/797, according to which MS shall submit the draft of new national rules on certain issues<sup>2005</sup> to the ERA and the Commission 'for consideration'.<sup>2006</sup> Upon receipt, the ERA shall examine the draft national rules within an extendable period of two months. Where the ERA deems the drafts to be in compliance with the relevant Union law,<sup>2007</sup> it shall inform the Commission and the MS concerned of its positive assessment. Where the ERA fails to inform the Commission and the MS concerned of its assessment within the (extended) period, the MS may proceed with the introduction of the rule.<sup>2008</sup>

Where the ERA's assessment is negative, the ERA shall inform the MS concerned and ask for its position on the assessment. If, following that exchange of views, the ERA maintains its negative assessment, it shall, within one month, address an opinion to the MS concerned,2009 stating the reasons why the draft national rules should not enter into force and/or be

<sup>2004</sup> This situation – an individualised act adopted by a national authority, which is contrary to EU law – may pose a challenge to supremacy; see eg Clausing/Kim‐ mel, § 121 VwGO, paras 116b-116h. Article 29 para 2 of Regulation 806/2014 does not apply here. Since the national authority has not violated an SRB decision (but, at most, a warning of the SRB), the requirements of this provision are not met. What is more, if Article 29 para 2 applied, there would be no need for the special regime of Article 7 para 4.

<sup>2005</sup> See the issues listed in Article 8 para 3 of Directive 2016/798 and in Article 14 para 4 of Directive 2016/797, respectively.

<sup>2006</sup> For the case of urgent preventive measures in case of Directive 2016/798 see its Article 8 para 5. Under the regime of Directive 2016/797 no special procedure applies: see its Article 14 para 4 lit b. Note that the ERA has true decision-making power in certain cases; see eg Article 10 or Article 17 of Directive 2016/798; see, in conjunction therewith, Article 14 of Regulation 2016/796. For the role of the ERA in ensuring compliance of MS with EU law see Versluis/Tarr, Compliance.

<sup>2007</sup> For the creation and qualification of rules of Union law referred to as 'common safety measures' (CSMs), 'common safety targets' (CSTs) and 'technical specifi‐ cations of interoperability' (TSI) which are highly relevant in this context see Granner, Verkehrsagenturen 229–232.

<sup>2008</sup> Article 25 para 2 of Regulation 2016/796.

<sup>2009</sup> On the ERA's opinions more generally, and in particular on its publication, see Article 10 of Regulation 2016/796.

applied, and inform the Commission accordingly.2010 The MS shall inform the Commission of its position on the ERA's opinion within two months, including its reasons in case of disagreement. Where the reasons provided are deemed insufficient or where the MS has failed to provide them, and where the MS adopts the respective national rule without 'paying sufficient heed' to the ERA's opinion, the Commission may adopt an implementing decision according to Article 291 TFEU to the MS concerned, requesting it to modify or repeal the rule.<sup>2011</sup>

This regime<sup>2012</sup> constitutes a mixed compliance mechanism. As the first act addressed to a MS, there is the opinion of the ERA, stating non-compli‐ ance of the draft national rule and implicitly (and in a legally non-binding way) commanding compliance with it. Following adoption of national rules which are not compliant with that opinion, the Commission may adopt a decision requiring the MS concerned to modify or repeal these rules, ie to comply with Union law.

### 2.2.2.3. Summary and résumé

Mixed compliance mechanisms provide for both soft and hard law acts adopted by EU bodies and addressed to MS. Compliance with Union law is first 'suggested' and, if the MS does not comply, eventually ordered by law. The mixed procedures presented here appear to be more generous towards the MS than hard ones. It should be borne in mind, though, that also hard compliance mechanisms regularly provide for a possibility for the MS concerned to utter their respective view. It is not so much different rights of MS which signify the increased generosity of mixed as compared to hard compliance mechanisms, but it is the extended time frame available for the MS and the (at least temporary) reduction of legal pressure exerted on it. Mixed compliance mechanisms – or at least those presented above, regardless of whether they are laid down in primary law or in secondary law – are not so much about matters considered very urgent, but about matters which allow for some time to be settled and/or in which the EU

<sup>2010</sup> Article 25 para 3 of Regulation 2016/796.

<sup>2011</sup> Article 25 para 4 of Regulation 2016/796.

<sup>2012</sup> A parallel regime is laid down in Article 26 of Regulation 2016/796 for the exami‐ nation of *existing* national rules, that is to say rules which are existing already (not: drafts) at the beginning of the procedure.

has a particular interest in (trying to reach) an amicable settlement of disagreements with MS.

The sequence of acts in mixed compliance procedures regularly is that one or more soft law acts preceed one or more hard law acts. It is a tiered procedure during which the pressure on the MS to comply is gradually increased. Article 271 lit a and d TFEU, laying down a parallel procedure to Article 258 TFEU, is an example for the increased amount of time available. Also the regimes laid down in Article 7 para 4 of Regulation 806/2014 and Article 25 of Regulation 2016/796, respectively, seem to address important, but not urgent issues. They are both about the examination of drafts of MS measures by EU bodies. Drafts are not yet in force, which is why the legislator catered for a more extended formalised exchange of views between the EU and the national level here. Similar in this respect is Article 117 TFEU which applies where there is a mere 'reason to fear' that (future) national measures may distort the conditions of competition in the internal market.<sup>2013</sup>

Another point is the question of competence. Where the policy field at issue is delicate because it addresses traditional prerogatives of sovereign states (such as fiscal policy or penal jurisdiction) or where the involvement of EU bodies the empowerment of which is subject to strict conditions is intended, a mixed compliance mechanism may appear to be more appro‐ priate than a hard compliance mechanism.

An example for the former are the multilateral surveillance procedure and the excessive deficit procedure. The MS as Masters of the Treaties have decided that the EU shall have a merely coordinating competence in the field of economic policy according to Article 5 TFEU. The compliance mechanisms contained in Articles 121 and 126 TFEU have to be understood in this light.2014 While economic policy coordination and the multilateral surveillance procedure laid down in Article 121 TFEU are entirely soft, the excessive deficit procedure of Article 126 TFEU also provides for legally binding Union acts which are, however, conceptualised as the *ultima ratio* in a long-winded procedure with many possibilities for the MS concerned

469

<sup>2013</sup> Article 116 TFEU, on the contrary, allows for a (hard) reaction where this risk has materialised. Taken together, as was set out above, these two provisions form a mixed compliance mechanism.

<sup>2014</sup> As provisions of primary law, Articles 121 and 126 TFEU are not to be examined with regard to their compliance with Article 5 TFEU, but they can be assumed to be set up against the background (and in the spirit) of the competence regime addressed therein.

to show its good will to tackle its fiscal problems (and thereby prevent or at least delay the adoption of hard law on the part of the EU). What is more, a decision determining the existence of an excessive deficit or even to impose a financial sanction is to be adopted by the Council – which makes the procedure more 'political' than if the Commission were in charge. From that we can deduce that the MS intended to provide certain competences for the EU in the field of economic policy, but that they wanted the intru‐ sion with this traditional MS prerogative to be mild. The considerable intensification of both regimes brought about by the so-called 'Six Pack' and 'Two Pack'<sup>2015</sup> in the course of the Euro crisis – the Regulations 1176 and 1174/2011 (transforming, in addition to a material extension, the soft multilateral surveillance procedure as laid down in Article 121 TFEU into a mixed procedure) have been addressed above – *qua* primacy of the TFEU could not do away with that approach *in principle*. 2016

Examples for conscious limitations to the empowerment of EU bodies are the regimes laid down in Article 39 of Directive 2009/72/EC, Article 17 of Regulation 1093/2010 and Article 25 of Regulation 2016/796. All these ca‐ ses involve specialised bodies mainly composed of MS representatives – Eu‐ ropean agencies which were established, among other things, to support the Commission in the implementation/enforcement of Union law.2017 Since the amount of powers such agencies may be vested with is limited in par‐ ticular by the so-called (and meanwhile reconsidered) *Meroni* doctrine,<sup>2018</sup> in our case specifically in order not to interfere with the Commission's central role as guardian of the Treaties (as one aspect of maintaining the EU's institutional balance), the legislator tried to do justice to the role of the Commission in the respective mechanisms. In the procedures involving the ACER and the ERA, respectively, these agencies adopt a soft law act *visà-vis* the MS, which may then be reinforced by the Commission in a legally binding way. In the regime laid down in Article 17 of Regulation 1093/2010 the role of the Commission is comparatively weaker, with the last act in the (possible) sequence of acts stemming from the EBA. However, with its formal opinion the Commission can largely predetermine the content of the ultimate – hard – EBA output.

<sup>2015</sup> See Craig, Administrative Law 207–209; see also 2.2.3.2.4. below.

<sup>2016</sup> See Antpöhler, Emergenz 382.

<sup>2017</sup> See Commission, European Governance – A White Paper, COM(2001) 428 final, 30.

<sup>2018</sup> See III.3.7.2.2. above and V.3.3.1. below.

In the case of Article 22 f of Council Regulation 2015/1589, the mixed character of the procedure was pre-determined in primary law – namely: Article 108 paras 1f TFEU. The cited provisions of the Council Regulation merely concretise primary law and clarify that the Commission's proposi‐ tion pursuant to Article 108 para 1 TFEU shall take the form of a recom‐ mendation.

#### 2.2.3. Soft compliance mechanisms

2.2.3.1. In primary law

2.2.3.1.1. Article 121 TFEU

The regime of Article 121 TFEU – in a primary law perspective – constitutes a soft compliance mechanism. For examples of its concretisation by means of secondary law see 2.2.2.2.4. above and 2.2.3.2.4. below.

Under Article 121 TFEU, the Council shall monitor, *inter alia*, <sup>2019</sup> the consistency of MS' economic policies with the broad economic policy guidelines (for the drafting and the adoption of these BEPG see III.3.5.2.1.2. above).2020 For that purpose, the MS shall forward information to the Commission 'about important measures taken by them in the field of their economic policy and such other information as they deem necessary'.<sup>2021</sup> Where it is established either that the economic policies of a MS are not consistent with the BEPG or that they risk jeopardising the proper func‐ tioning of the EMU, the Commission may address a warning to the MS concerned.2022 On a recommendation from the Commission, the Council may – in addition to a (potential) Commission warning – address the 'nec‐ essary recommendations' to the respective MS. It may, on a proposal from the Commission, make these recommendations public.2023 When adopting

<sup>2019</sup> For the broader scope of multilateral surveillance see Article 121 para 3 subpara 1 TFEU.

<sup>2020</sup> Article 121 para 3 subpara 1 TFEU.

<sup>2021</sup> Article 121 para 3 subpara 2 TFEU.

<sup>2022</sup> For this procedural step introduced by the Treaty of Lisbon see Louis, Economic Policy 288.

<sup>2023</sup> For the potential 'Prangerwirkung' [pillory effect] such publication may create see Häde, Art. 121 AEUV, para 14. While the recommendations according to Article 121 TFEU are legally non-binding, and a publication is the strongest means of 'enforcement', the initiation of a Treaty infringement procedure appears possible

recommendations addressed to Eurozone MS in the framework of multilat‐ eral surveillance, the voting rights of non-euro MS in the Council shall be suspended.<sup>2024</sup>

In terms of the output of EU institutions *vis-à-vis* the MS, Article 121 is a sequence of soft law acts: Starting with the BEPG, a recommendation adopted by the Council, over the Commission warning<sup>2025</sup> and eventually ending with the Council recommendation in case of inconsistency of MS economic policies with the BEPG (which may be published<sup>2026</sup>).<sup>2027</sup> Thus, Article 121 TFEU constitutes a soft compliance mechanism.

# 2.2.3.1.2. Article 148 para 4 TFEU

Article 148 TFEU provides for a regime of monitoring the employment situation in the Union. The European Council shall each year consider this situation and adopt conclusions accordingly, on the basis of a joint annual report by the Council and the Commission. Against the backdrop of these conclusions, the Council shall draw up guidelines annually (on a proposal from the Commission2028) which the MS shall take into account in their employment policies. These guidelines shall be consistent with the BEPG adopted pursuant to Article 121 para 2 TFEU.2029 Each MS shall provide

where a MS infringes its duty to participate in the cooperation laid down in Article 120 TFEU; see ibid, para 15, with further references.

<sup>2024</sup> Article 139 para 4 lit a TFEU. It is not clear how the reference to 'warnings' in this provision is to be understood. After all, it is the Commission – not the Council – which may adopt a warning under Article 121 para 4 TFEU. Notwithstanding this unclarity, it seems that the understanding underlying this provision is that a warn‐ ing (according to Article 121 para 4 TFEU) is a sub-category of recommendations.

<sup>2025</sup> See Verhelst, Reform 10 f, stating that warnings are legally non-binding, but silent as to their qualification as EU soft law; referring to the policy advice contained therein: Louis, Economic Policy 288 (fn 6); for the pillory effect of warnings see Feik, Verwaltungskommunikation 428.

<sup>2026</sup> The publication of recommendations is effected by a legally binding Council act, arguably a decision. But since this decision does not impose duties on the MS con‐ cerned and hence does not in principle alter the soft character of this mechanism, it shall be left aside here.

<sup>2027</sup> For the distinction of these measures in *ex ante* and *ex post* mechanisms see Amtenbrink/Repasi, Compliance 154 f.

<sup>2028</sup> And after consulting the EP, the ESC, the CoR and the Employment Committee referred to in Article 150 TFEU.

<sup>2029</sup> Article 148 paras 1 and 2 TFEU.

the Council and the Commission with an annual report on the principal measures taken to implement its employment policy in the light of these guidelines.<sup>2030</sup> The Council shall then – in short – examine compliance of the employment policies of the MS with the Council guidelines, and may – on a recommendation from the Commission – make recommendations to the MS accordingly.<sup>2031</sup>

The compliance mechanism focussed on here is embedded in the system of Article 148 TFEU which – for reasons of contextualisation – was presen‐ ted briefly. The Council recommendation according to para 4 seeks compli‐ ance of the MS concerned with EU law, namely with the (non-binding) Council guidelines for employment.<sup>2032</sup> The recommendation is addressed to single MS<sup>2033</sup> and is legally non-binding. Being adopted by the Council on the basis of the conclusions of the European Council, the latter convey high (political) authority; *Hemmann* considers them a politically 'macht‐ volles Instrument' [powerful instrument].<sup>2034</sup>

### 2.2.3.2. In secondary law

2.2.3.2.1. Article 6 paras 5–7 of Regulation 2019/942

The ACER disposes of a number of means to ensure – sometimes together with the Commission – compliance of the regulatory authorities in the MS.2035 Article 6 paras 5–7 of Regulation 2019/942, based on Article 194

<sup>2030</sup> Article 148 para 3 TFEU.

<sup>2031</sup> For the joint annual report to the European Council see Article 148 para 5 TFEU.

<sup>2032</sup> Such a recommendation may only be adopted upon a Commission recommenda‐ tion. The Council has a wide discretion on whether, and if so: in which way, to follow the Commission recommendation (*argumentum* 'may', 'if it considers it ap‐ propriate in light of that examination'); see Hemmann, Artikel 148 AEUV, para 11. The legal status of the guidelines is not entirely clear. Apparently they are adopted in the form of a Council decision (see eg Commission Proposal COM(2017) 677 final), arguably because Article 148 TFEU does not mention the 'recommendation' as the adequate legal form; see Hemmann above, para 14; for a discussion of whether these guidelines are legally binding see Braams, Koordinierung 39 f, with further references.

<sup>2033</sup> See Niedobitek, Art. 148 AEUV, para 18.

<sup>2034</sup> Hemmann, Artikel 148 AEUV, para 11; for the possibility of a publication of these recommendations see Steinle, Beschäftigungspolitik 371.

<sup>2035</sup> See eg Article 51 para 1 of Regulation 2019/943 or the mechanism addressed in 2.2.1.2.4. above.

para 2 TFEU, shall be taken as an example of a soft compliance mechanism here.<sup>2036</sup> At the request of one or more national regulatory authorities or the Commission, the ACER shall provide a 'factual opinion'<sup>2037</sup> on whether a decision of a regulatory authority complies with (binding<sup>2038</sup>) network codes and guidelines referred to in Regulation 2019/943,<sup>2039</sup> Regulation 715/2009,<sup>2040</sup> Directive 2019/944<sup>2041</sup> or Directive 2009/73/EC or with other relevant provisions of those directives or regulations. Thereby the ACER may also list which further information or other components the decision at issue should have contained.<sup>2042</sup> Where a regulatory authority does not comply with the opinion of the ACER within four months, the ACER shall inform the Commission and the MS concerned.<sup>2043</sup> Its opinion being a legally non-binding instrument,<sup>2044</sup> the ACER cannot force the regulatory authority to comply. This could be achieved by a Treaty infringement procedure subsequently initiated by the Commission,2045 or – at least with regard to some of the guidelines addressed here – in an extended, and mixed, procedure as set out eg in Article 43 of Directive 2009/73/EC. In the latter case, the Commission may – following a regulatory authority's non-compliance with an ACER opinion – take a (legally binding) decision requiring the regulatory authority concerned to withdraw its decision on the basis that the guidelines have not been complied with.<sup>2046</sup>

<sup>2036</sup> For another soft compliance mechanism involving the ACER see Article 63 para 8 of Regulation 2019/943.

<sup>2037</sup> Apparently and comprehensibly so, the ACER also utters its legal viewpoint in this opinion; see Tišler, Agency 397, with regard to the predecessor provision, Article 7 para 4 of Regulation 713/2009, which depicted the opinion as 'based on matters of fact'.

<sup>2038</sup> See eg Article 66 para 1 of Regulation 2019/943, with regard to guidelines and network codes; see also Recital 88 of Directive 2019/944. Framework guidelines – on the contrary – are explicitly qualified as non-binding (see eg Article 59 para 4 of Regulation 2019/943).

<sup>2039</sup> See Articles 58–61 of Regulation 2019/943.

<sup>2040</sup> See Articles 23 f of Regulation 715/2009.

<sup>2041</sup> This Directive does not only refer to guidelines and network codes of the Commis‐ sion but, misleadingly, also to guidelines of national regulatory authorities (eg in its Article 8 para 3). Since they do not constitute EU law, arguably the mechanism addressed here does not specifically aim at compliance with these acts.

<sup>2042</sup> See case T-671/15 *E-Control*, para 74, with regard to the predecessor mechanism.

<sup>2043</sup> Article 6 para 6 of Regulation 2019/942.

<sup>2044</sup> See case T-63/16 *E-Control*, paras 46 f, with regard to the predecessor mechanism.

<sup>2045</sup> See Tišler, Agency 397.

<sup>2046</sup> Article 43 para 6 lit b of Directive 2009/73/EC.

As was mentioned above, the procedure described here constitutes a soft compliance mechanism: Upon request by a national regulatory authority or the Commission, the ACER issues an opinion determining whether or not a decision of a national regulatory authority is in compliance with the relevant EU law. No further acts are provided for in the regime of Article 6 paras 5–7 of Regulation 2019/942.

### 2.2.3.2.2. Article 53 of Directive 2019/944

Directive 2019/944 is based on Article 194 para 2 TFEU and concerns common rules for the internal market in electricity. Its Article 53 which is at issue here sets out restrictions on electricity transmission operations by third-country actors. <sup>2047</sup> Where a certification is requested by transmis‐ sion system owners/operators controlled by third country nationals, the national regulatory authority shall notify the Commission.2048 The national authority shall then adopt a (positive or negative2049) draft decision on the certification within four months, which shall be notified to the Commission together with the relevant information with respect to that decision. MS shall provide (in their respective national law transposing the Directive) for the national authority concerned2050 to request, before the (final) decision is taken, an opinion from the Commission on whether a) the entity con‐ cerned complies with the requirements of Article 43 of the Directive and b) granting certification will not put at risk the security of energy supply to the EU.2051 The Commission shall examine the request and deliver an opinion within two months (which may be extended by two months).<sup>2052</sup>

<sup>2047</sup> For the parallel provision in the natural gas sector see Article 11 of Directive 2009/73/EC; see also Luca, Framework 132 f; Schweitzer, Funds 280.

<sup>2048</sup> Article 53 para 1 subpara 1 of Directive 2019/944; see also Article 53 para 1 subpara 2 and para 2.

<sup>2049</sup> For the legal reasons for a negative decision, a refusal, see Article 53 para 3 and para 8 of Directive 2019/944.

<sup>2050</sup> This can be the regulatory authority or the designated competent authority accord‐ ing to Article 11 para 3 lit b of Directive 2009/72/EC.

<sup>2051</sup> Article 53 para 5 of Directive 2019/944. The content of Article 43 need not be discussed any further in this context. Suffice it to say that it is the compliance with (relevant) Union law of the draft decision which is to be examined by the Commission.

<sup>2052</sup> For the details of the Commission's examination see Article 53 para 6 subparas 1f and para 7 of Directive 2019/944.

Where the Commission does not deliver an opinion within the prescribed period, the Commission shall be deemed 'not to raise objections to the de‐ cision' of the national authority.<sup>2053</sup> Upon receipt of the opinion (or expiry of the period), the national authority shall take its final decision on the certification, thereby taking 'utmost account' of the Commission's opinion (if any). The decision and the Commission opinion shall be published together. Where the final decision diverges from the Commission opinion, the MS concerned shall provide and publish, together with that decision, the reasoning underlying such decision.<sup>2054</sup>

Upon request of a national authority, the Commission shall send an opinion on the compliance of the authority's draft decision with specific EU law. The national authority shall take 'utmost account' of this opinion when adopting the final decision and the MS shall provide the reasons for any divergence. This emphasises the legal non-bindingness of the Commis‐ sion opinion, which is why the regime is to be called a soft compliance mechanism.

### 2.2.3.2.3. Article 33 of Directive 2018/1972

Directive 2018/1972 on a common regulatory framework for electronic communications networks and services is based on Article 114 TFEU. Its Article 33 which shall be focussed on here is entitled 'Procedure for the consistent application of remedies'.2055 For a certain category of (intended) measures to be taken by national regulatory authorities2056 the Commission may, within one month, notify the national regulatory authority concerned and the Body of European Regulators for Electronic Communications (BE‐ REC)2057 of its reasons for considering that the draft measure would create

<sup>2053</sup> Article 53 para 6 subpara 3 of Directive 2019/944.

<sup>2054</sup> Article 53 para 8 of Directive 2019/944; for further details of the regime see paras 9 f *leg cit*.

<sup>2055</sup> For a contextualisation of this procedure with regard to other procedures laid down in the very similar regime under the predecessor Directive 2002/21/EC: Kühling, Telekommunikationsrecht, para 69; for the practical application of this procedure by the Commission see ibid, para 71.

<sup>2056</sup> Namely those specified in Article 33 para 1 of Directive 2018/1972.

<sup>2057</sup> For an overview of organisation and tasks/powers of the BEREC see Van Cleynen‐ breugel, Supervision 66–68. The BEREC does not qualify as a European agency, only the Office does; <https://www.berec.europa.eu/en/berec-office/tasks-and -mission> accessed 28 March 2023. While the BEREC in the political negotiations

a barrier to the internal market or of its serious doubts as to its compatibili‐ ty with Union law. In this case, the draft measure shall not be adopted for a further three months following the Commission's notification. Otherwise – ie where the Commission has not made a notification – the national authority concerned may adopt the measure, taking 'utmost account' of any comments made by the Commission,<sup>2058</sup> the BEREC<sup>2059</sup> or any other national regulatory authority.

Within six weeks from the beginning of the three months period, the BE‐ REC shall issue a reasoned opinion on the Commission's notification.<sup>2060</sup> If the BEREC in its opinion (which is to be published) shares the serious doubts of the Commission, it shall cooperate closely with the national authority concerned – to which the opinion is (also) addressed – to iden‐ tify the most appropriate and effective measure. The national authority may, before the end of the three months, either amend/withdraw its draft measure, taking 'utmost account'2061 of the Commission's notification and of the BEREC opinion and advice, or maintain its draft measure.2062 If the national authority does not withdraw its draft measure anyway, the Commission may, within one month after the end of the three months period and taking 'utmost account' of the BEREC opinion (if any): a) issue a reasoned2063 recommendation requiring the national authority concerned to amend or withdraw the draft measure (including specific proposals to


on its establishment was originally envisaged as a European agency, these plans were later dropped by the Council and the European Parliament; see Schilchegger, Agenturen 123–125.

<sup>2058</sup> These comments still leave it 'for [the national] authority alone to decide whether to adopt that measure and to determine its content'; see, with regard to the predecessor regime, case T-109/06 *Vodafone España,* para 161; case T-295/06 *Base*, para 61. See also para 62 of the latter Order (stressing the cooperation required between the Commission and the national authorities) and its para 69 (qualifying the comments of the Commission as 'acte communautaire préparatoire').

<sup>2059</sup> For the authority of the BEREC's soft law output more generally see Article 4 para 4 of Regulation 2018/1971 (ie BEREC's founding regulation). For the BEREC's in-between position betwixt the Commission and the national authorities see Kühling, Telekommunikationsrecht, para 62.

<sup>2060</sup> For further details see Article 33 para 3 of Directive 2018/1972; for the preceding cooperation between the Commission, the BEREC and the national authority see para 2 *leg cit*.

<sup>2061</sup> See also the more general rule of Article 4 para 4 of Regulation 2018/1971.

that end), or b) take a decision to lift its reservations indicated in the course of its notification.<sup>2064</sup> For specific draft measures, the regime under para 5 lit c applies. Within an extendable period of one month of the Commission having acted as provided in either alternative, the national au‐ thority concerned shall communicate to the Commission and the BEREC the adopted final measure.<sup>2065</sup> Where the national authority decides not to amend or withdraw the draft measure on the basis of the Commission recommendation, it shall provide 'reasons'.<sup>2066</sup> The national authority may withdraw the draft measure at any time during the procedure laid down in Article 33.

Let us dwell on the structure of this procedure a bit more: Its first phase may be coined by a Commission notification. While this notification is likely to establish non-compliance in a legally non-binding way, the *act* of the notification does have a legally binding effect (laid down in the Directive),2067 namely that the national authority shall not adopt the draft measure for three months.<sup>2068</sup>

Where the Commission has not made a notification, the national author‐ ity shall take 'utmost account' of any comments the Commission (or other

<sup>2064</sup> Article 33 para 5 of Directive 2018/1972.

<sup>2065</sup> Article 33 para 6 of Directive 2018/1972.

<sup>2066</sup> The pleonastic wording of the predecessor provision, Article 7a para 7 of Directive 2002/21/EC, obliging the national authority to provide 'reasoned justification', has been substituted by a simpler expression.

<sup>2067</sup> Tobisch, Telekommunikationsregulierung 99–101 (with examples and with regard to the predecessor regime of Article 7a of Directive 2002/21/EC) qualifies it as 'opinion' pursuant to Article 288 TFEU. Whether the barrier to the internal market also (necessarily) constitutes a violation of EU law must be left open here.

<sup>2068</sup> Not respecting the notification requirement would arguably lead – for non-compli‐ ance with EU law – to the non-applicability of the national measure; see Commis‐ sion, Communication concerning the non-respect of certain provisions of Council Directive 83/189/EEC of 28 March 1983 laying down a procedure for the provision of information in the field of technical standards and regulations, 86/C 245/05, claiming that – with regard to a similar, but hard mechanism – 'without notifying the draft to the Commission and respecting the standstill obligation, the [national measure] thus adopted is unenforceable against third parties in the legal system of the Member State in question'; for the problem of non-notification or non-compli‐ ance with the standstill period in the context of another legal act see Commission, 'The Operation of Directive 98/34/EC from 1995 to 1998', COM(2000) 429 final, paras 74 f.

bodies involved) may have made. Where these comments contain norms, they (therefore) qualify as soft law.<sup>2069</sup>

A second phase may only follow where the Commission has adopted a notification in the first phase. It is initiated by the (potential) BEREC opinion on the Commission's notification which shall be made public. This opinion is also addressed to the national authority which shall, when amending or withdrawing its draft measure, take 'utmost account' of it. Otherwise, the national authority shall maintain the measure. The BEREC opinion clearly is an act of EU soft law, as it contains norms (it indicates whether 'the draft measure should be amended or withdrawn' and, if so, how<sup>2070</sup>) and is legally non-binding (*argumentum* draft may be main‐ tained).<sup>2071</sup> Where the BEREC does not share the Commission's doubts or where it does not issue an opinion, or where the national authority amends or maintains its draft, the Commission may, taking utmost account of the BEREC opinion, issue a recommendation to the national authority concerned. It thereby requires the latter to amend (and, if so, indicates in which way) or withdraw the draft measure. This is also clearly a soft law measure, as it contains norms and is legally non-binding.<sup>2072</sup>

The procedure is intended to cater for input from BEREC, the expert body in the field, whose main organ is composed of representatives of the national regulatory authorities,2073 while ensuring that the Commission – as the central administrative authority of the EU 'supervising'2074 the national authorities here – has the last (soft) word2075 in case the BEREC in its opinion deviates from the Commission's viewpoint, does not issue

<sup>2069</sup> For the content of comments adopted under the predecessor of Article 32 of Directive 2018/1972 (Article 7 para 3 of Directive 2002/21/EC) see Kühling, Tele‐ kommunikationsverwaltungsrecht, paras 51–53. For the guidelines adopted on the basis of the predecessor Directive 2002/21/EC see case C-410/09 *Polska Telefonia*.

<sup>2070</sup> Article 33 para 3 of Directive 2018/1972.

<sup>2071</sup> See also case C‑632/20P *Spain v Commission*, para 85.

<sup>2072</sup> Article 33 para 7 of Directive 2018/1972.

<sup>2073</sup> See Article 7 of Regulation 2018/1971. Also heed the statutory independence of the BEREC as laid down in Article 3 para 3 of Regulation 2018/1971; with regard to the BEREC's independence see also case C‑632/20P *Spain v Commission*, paras 119–121.

<sup>2074</sup> Opinion of AG *Cruz Villalón* in case C-518/11 *UPC Nederland*, para 52.

<sup>2075</sup> See Commission, EU Telecoms Reform, MEMO/09/513 (20 November 2009), para 9; critically with regard to the – in terms of the legal non-bindingness of the Commission recommendation – misleading German version of this document: Tobisch, Telekommunikationsregulierung 98.

an opinion, and/or in case the MS authority concerned amends2076 or maintains its draft measure. Otherwise it may 'take a decision to lift its reservations indicated in accordance with [Article 33] paragraph 1'.<sup>2077</sup> This 'decision' does not need to be legally binding. According to the *contrarius actus* doctrine, the repeal of a soft law act may also be effectuated by a soft law act of the same kind.<sup>2078</sup> The specific case of para 5 lit c, which allows the Commission to take a binding decision, shall not be addressed here.<sup>2079</sup>

### 2.2.3.2.4. Article 3 para 7 of Regulation 472/2013

Regulation 472/2013, based on Article 136 in conjunction with Article 121 para 6 TFEU, aims at strengthening the economic and budgetary surveil‐ lance of MS in the Eurozone experiencing or threatened with serious dif‐ ficulties with respect to their financial stability. Together with Regulation 473/2013, it forms the so-called 'Two Pack'.2080 While making a Eurozone MS subject to enhanced surveillance – a status on the prolongation of which the Commission shall decide every six months2081 – has a number of consequences,2082 here we shall focus on one specific measure, as laid down in Article 3 para 7. Where the Commission, on the basis of a review mission provided for in para 5 *leg cit*, deems further2083 measures to be required in order to address the sources or potential sources of difficul‐ ties,2084 and the financial and economic situation of the MS concerned has significant adverse effects on the financial stability of the Euro area or of its MS, it may propose to the Council the adoption of recommendations to that MS to adopt precautionary corrective measures or to prepare a draft

<sup>2076</sup> Where the national regulatory authority amends its draft measure *in accordance with* the BEREC opinion, the Commission in the majority of cases – ie if it does not object to the BEREC opinion in the first place – will lift its reservations pursuant to Article 33 para 5 lit b of Directive 2018/1972.

<sup>2077</sup> Article 33 para 5 lit b of Directive 2018/1972.

<sup>2078</sup> See III.3.8. above.

<sup>2079</sup> If lit c is applied in a concrete case, according to the terminology applied here this will transform the regime under Article 33 into a mixed mechanism.

<sup>2080</sup> For the 'Two Pack' more generally see Gerapetritis, Constitutionalism 54.

<sup>2081</sup> Article 2 para 1 subpara 3 of Regulation 472/2013.

<sup>2082</sup> See eg Borger, European Stability Mechanism 169 f.

<sup>2083</sup> That means: measures in addition to those referred to in the rest of Article 3 of Regulation 472/2013, in particular in its paras 3 f.

<sup>2084</sup> See Article 3 para 1 of Regulation 472/2013.

macroeconomic adjustment programme. Where the Council adopts such recommendations, it may decide to make them public.<sup>2085</sup>

This is a soft compliance mechanism, as it merely encompasses Council recommendations addressed to a Eurozone MS (which may be made public to increase the pressure on the MS concerned to comply). The final aim of these recommendations is to ensure that a Eurozone MS which is subject to enhanced surveillance<sup>2086</sup> again complies with its duties laid down in Article 120 TFEU.<sup>2087</sup>

### 2.2.3.2.5. Articles 16 and 17 of Regulation 1092/2010

Regulation 1092/2010 which is based on Article 114 TFEU sets up a regime for EU macro-prudential oversight of the financial system, in particular by creating a European Systemic Risk Board (ESRB).2088 This ESRB may issue recommendations2089 in accordance with Article 16 of Regulation 1092/2010 and address these general or specific recommendations, apart from the EU and specific EU bodies, to one or more MS or to one or more of the na‐ tional authorities in charge of (financial market) supervision, in charge of applying measures aimed at addressing systemic or macro-prudential risk or in charge of bank resolution.2090 These recommendations shall propose remedial action (possibly including legislative initiatives) where significant risks to the stability of the EU's financial system as circumscribed in Article 3 para 1 of Regulation 1092/2010 are identified, and shall contain a specified timeline for the policy response.<sup>2091</sup>

2090 Article 16 para 2 of Regulation 1092/2010.

<sup>2085</sup> For the further consequences this publication may entail see Article 3 para 8 of Regulation 472/2013.

<sup>2086</sup> See Article 2 para 1 of Regulation 472/2013.

<sup>2087</sup> See Wittelsberger, Art. 120 AEUV, para 3.

<sup>2088</sup> For organisation and tasks of the ESRB see eg Weismann, Agencies 106 ff; for the composition of its General Board see Article 6 of Regulation 1092/2010.

<sup>2089</sup> Under Article 16 of Regulation 1092/2010, the ESRB may adopt both warnings and recommendations. Whether the warnings qualify as soft law needs to be assessed case by case with a view to whether they actually contain a (soft) command. In general, warnings – unlike recommendations – may be uttered already at a stage where the risks at issue are not yet identified in full (see also Article 3 para 2 lit c and d; slightly different: Article 16 para 1). Here only the recommendations shall be addressed.

<sup>2091</sup> Article 16 paras 1f and Article 3 para 2 lit b of Regulation 1092/2010. For the trans‐ mission of these recommendations to the EP, the Council and the Commission

#### *IV. MECHANISMS IN EU LAW TO ENSURE LEGAL COMPLIANCE*

The addressees of the recommendation shall communicate to the EP, the Council, the Commission and the ESRB the actions they have under‐ taken 'in response' to the recommendation, and shall 'substantiate' any inaction,<sup>2092</sup> '[h]ence, recommendations issued by the ESRB cannot be simply ignored'.<sup>2093</sup> If the ESRB establishes – 'decides' – that its recommen‐ dation has not been complied with or that the addressees have failed to provide adequate justification for their respective inaction, it shall inform the addressees, the EP and the Council and, where relevant, the ESA concerned in accordance with Article 17 para 2 of Regulation 1092/2010. While a recommendation, according to Article 16, in principle is handled confidentially by the ESRB, it may make the recommendation public under the conditions laid down in Article 18 of Regulation 1092/2010. Where the ESRB makes a 'decision'<sup>2094</sup> (establishing non-compliance) pursuant to Article 17 para 2 with regard to a (published) recommendation, the EP may invite the Chair of the ESRB to present its 'decision', and the addressee may request to participate in an exchange of views.<sup>2095</sup>

While this mechanism does not necessarily aim at ensuring compliance with detailed provisions of EU law, it aims at ensuring compliance with an important objective of the EU, namely the stability of the financial system of the EU.2096 In this context, the ESRB shall 'contribute to the prevention or mitigation of systemic risks to financial stability in the Union that arise from developments within the financial system and taking into account macroeconomic developments, so as to avoid periods of widespread finan‐ cial distress' and to 'contribute to the smooth functioning of the internal

<sup>(</sup>and possibly to the ESAs) and for the criteria for the classification of risks in the economy see Article 16 paras 3 f *leg cit*.

<sup>2092</sup> Article 17 para 1 of Regulation 1092/2010.

<sup>2093</sup> Commission Proposal, COM(2009) 499 final, 5.

<sup>2094</sup> Even though the legislator in this context uses the terms 'decide' and 'decision' respectively, it is clear already by comparison with other language versions of Regulation 1092/2010 that this does not encompass a decision according to Article 288 TFEU. What is more, Article 3 para 2 of Regulation 1092/2010 does not mention a decision-making power of the ESRB.

<sup>2095</sup> Article 17 para 3 of Regulation 1092/2010. Also this effect rather indicates legal non-bindingness.

<sup>2096</sup> This aim can be subsumed under Article 3 TEU. For another example in which policy objectives, among others, constitute the threshold against which compliance with EU law is to be examined see Article 29 para 2 of Regulation 806/2014; see 2.2.1.2.3. above; also note, in this context, the wording of Article 4 para 3 subpara 3 TEU.

market and thereby ensure a sustainable contribution of the financial sector to economic growth'.<sup>2097</sup> While the concrete requirements to reach this objective may be, but are not necessarily explicitly laid down in EU law, the ESRB may explicate them in its recommendations.<sup>2098</sup>

This compliance mechanism is a soft mechanism, as it entails a recom‐ mendation addressed – among others – to one or more MS or to one or more of the relevant national authorities. This recommendation is legally non-binding,<sup>2099</sup> but non-compliance needs to be adequately justified by the MS/national authority concerned. The ESRB may increase the pressure to comply, or at least to justify non-compliance, by publishing the recom‐ mendations at issue.

# 2.2.3.2.6. Article 6 of Regulation 2019/452

The 'cooperation mechanism' laid down in Article 6 of Regulation 2019/452 establishing a framework for the screening of foreign direct investments into the Union, based on Article 207 para 2 TFEU, relates to foreign direct investements undergoing screening. <sup>2100</sup> Screening in this context means 'a procedure allowing to assess, investigate, authorise, condition, prohibit or unwind foreign direct investments'.2101 A screening is applied to foreign direct investments on the grounds of MS' security or public order. It is performed by the MS.2102 According to this provision, MS shall notify the Commission and the other MS of any foreign direct investment in their territory that is undergoing screening by providing certain information on it (eg the ownership structure of the foreign investor or the approximate value of the foreign direct investment2103). This notification may include a list of MS whose security or public order is deemed likely to be affected.<sup>2104</sup>

<sup>2097</sup> Article 3 para 1 of Regulation 1092/2010.

<sup>2098</sup> For the importance of the EU's objectives for the interpretation of Union law see eg Terhechte, Art. 3 EUV, para 12.

<sup>2099</sup> See Ruppel, Finanzdienstleistungsaufsicht 109.

<sup>2100</sup> For the cooperation mechanism regarding investments not undergoing screening see Article 7 of Regulation 2019/452.

<sup>2101</sup> Article 2 para 3 of Regulation 2019/452.

<sup>2102</sup> For the fact that these issues largely fall within the prerogatives of the MS see, even if in a different context, Articles 72–74 TFEU.

<sup>2103</sup> See Article 9 para 2 of Regulation 2019/452.

<sup>2104</sup> For this and further content of the notification see Article 6 para 1 of Regulation 2019/452.

Other MS, if they feel affected in that way or if they have information relevant for the screening, may comment *vis-à-vis* the MS undertaking the screening, normally within 35 days of being informed<sup>2105</sup> (also informing the Commission thereof, which shall again inform the remaining MS).<sup>2106</sup>

Where the Commission considers that a foreign direct investment is likely to affect more than one MS in the above way, or where it has relevant information on that investment, it may issue an opinion to the MS under‐ taking the screening. The Commission in principle *may* issue an opinion irrespective of whether there have been comments from the other MS, but *shall* issue an opinion ('where justified'), if at least one third of the MS consider that a foreign direct investment is likely to affect their security or public order. The Commission shall adopt its opinion normally within 35 days of being informed,<sup>2107</sup> and it shall inform the other MS that an opinion was issued.2108 Both the MS' comments and the Commission's opinion shall be reasoned ('duly justified')2109 and announced in advance.<sup>2110</sup>

Where a MS, as a result of its examination, duly considers that a foreign direct investment in its territory is likely to affect its security or public order, it may request the Commission to issue an opinion or other MS to provide comments.<sup>2111</sup>

Where the MS undertaking the screening exceptionally considers that its security or public order requires immediate action, it shall notify the other MS and the Commission that it intends to take a screening decision before the expiry of the deadlines for comments and opinions referred to above (normally 35 days). The other MS and the Commission shall then attempt 'to provide comments or to issue an opinion expeditiously'.<sup>2112</sup>

<sup>2105</sup> For the various deadlines set in this context see Article 6 para 7 of Regulation 2019/452.

<sup>2106</sup> Article 6 para 2 of Regulation 2019/452.

<sup>2107</sup> For the deadline regime see Article 6 para 7 of Regulation 2019/452.

<sup>2108</sup> Article 6 para 3 of Regulation 2019/452. For the effects of this opinion see also de Kok, Framework 45.

<sup>2109</sup> Article 6 para 5 of Regulation 2019/452.

<sup>2110</sup> For the *ex ante* notification procedure and for requests for further information see Article 6 para 6 of Regulation 2019/452.

<sup>2111</sup> Article 6 para 4 of Regulation 2019/452.

<sup>2112</sup> Article 6 para 8 of Regulation 2019/452.

The MS taking the final screening decision shall give 'due considera‐ tion'<sup>2113</sup> to the comments of the other MS and the Commission opinion.<sup>2114</sup>

The opinion of the Commission is a legally non-binding act, a soft law act.<sup>2115</sup> It aims at furthering an objective not only of the MS, but also of the EU, that is to protect security and public order. Therefore it is (also) about compliance with EU law. Since no (binding) follow-up action to a MS's non-compliance is provided for, this is a soft compliance mechanism.

### 2.2.3.3. Summary and résumé

Soft law acts are legally non-binding. The fact that some provisions require MS to take 'utmost account' may express enhanced (political) authority.<sup>2116</sup> EU soft law acts do not only 'preserve' MS competences, granting the pow‐ er to adopt them may2117 also 'preserve' the institutional balance of the EU. This is why EU bodies not established by primary law (in particular: Euro‐ pean agencies), in an attempt to stay within the frame set by *Meroni*, 2118 are often vested with the power to adopt soft law acts, less often with hard law powers.2119 The soft compliance mechanisms addressed here reflect this situation – as do the mixed mechanisms above (see the explanations under 2.2.2.3.).

Soft compliance mechanisms are the least intrusive compliance mecha‐ nisms in the categorisation applied here. As the hard and mixed compliance mechanisms, they can be found in various policy fields. As regards the two mechanisms laid down in primary law, it is apparent that they both are used in delicate policy fields – economic policy and employment policy –

<sup>2113</sup> The Commission's legislative proposal (leading to Regulation 2019/452) still re‐ quired the MS to take 'utmost account' of the Commission's opinion and to provide an explanation to the Commission in case it did not follow it; Article 9 para 5 of Commission Proposal COM(2017) 487 final.

<sup>2114</sup> Article 6 para 9 of Regulation 2019/452.

<sup>2115</sup> See also Commission Proposal COM(2017) 487 final, 3 (Explanatory Memoran‐ dum).

<sup>2116</sup> For (potentially) different degrees of authority see V.3.5. below.

<sup>2117</sup> Critically: Cannizzaro/Rebasti, Soft law 230.

<sup>2118</sup> For the *Meroni* doctrine see III.3.7.2.2. above; for its importance in the context of the EU's institutional balance see also V.3.3.2. below.

<sup>2119</sup> See eg Ştefan/Petri, Review 531f, with respect to the ACER.

in which the EU has only limited competences.2120 Against this background, it is understandable that the MS (as Masters of the Treaties) have chosen the type of compliance mechanism which puts the least strain on MS competences.

As regards our selection of soft compliance mechanisms laid down in secondary law, the following can be said. Under the regimes of Article 6 paras 5–7 of Regulation 2019/942, Article 53 of Directive 2019/944 and Article 33 of Directive 2018/1972, it is a (draft) decision of a national authority which is assessed with a view to its compliance with the relevant Union law. This assessment may be expressed by a soft law act of the Com‐ mission and/or a specialised EU body (European agency). Article 3 para 7 of Regulation 472/2013 is one more mechanism within the framework of the multilateral surveillance procedure. It applies only to Eurozone MS and provides for Council recommendations as a means of ensuring compliance of these MS with the relevant Union law. Its softness is sketched out in Article 121 TFEU. Articles 16 f of Regulation 1092/2010 empower the ESRB to issue recommendations, *inter alia* to the MS. What is special about this compliance mechanism is that it is about compliance with an objective of the EU. This objective – in broad terms – is laid down in Union law. Thus, it can be argued that also this mechanism is about MS' compliance with Union law. The broad objective – the stability of the EU's financial system – and its affecting national policy choices, but also the empowerment of a newly established body may have been the reasons for the legislator to content itself with the soft shape of this procedure. According to Article 6 of Regulation 2019/452, the Commission addresses an opinion to a MS in order to ensure that foreign direct investments do not go against MS' security or public order – again, this procedure seems to be intended to leave enough room for national policies.

<sup>2120</sup> For the similarity of the competence categories 'economic policy' and 'employment policy' see Article 5 TFEU; see also Krebber, Art. 145 AEUV, para 1.

# V. CLASSIFICATION AND LEGAL ASSESSMENT OF COMPLIANCE MECHANISMS

#### *1. Introduction*

Having presented a range of different compliance mechanisms in Part IV above, namely those laid down in primary law and a selection of those laid down in secondary law, we shall now address this material, as it were, with a view to its classification and with a view to its legal assessment, thereby also resorting to the more general findings on EU soft law which Part III above resulted in. It is important to bear in mind that our basis for discussion – to the extent it is composed of compliance mechanisms laid down in secondary law – is only exemplary. The conclusions made in this respect first and foremost relate to these examples. They will be of use also in the context of compliance mechanisms which are not addressed here, in which many of the properties of our sample recur, but they cannot claim universality in the sense that they could simply be 'extrapolated' to other compliance mechanisms.

The above categories of hard, mixed and soft mechanisms are one way to structure the large number of compliance procedures laid down in EU law. As most classifications, it entails a certain simplification in order to facilitate an image and an understanding of reality. It ought to be emphas‐ ised that this taxonomy does not intend to suggest too strict a separation of these three categories. Exceptionally, there is room for different interpre‐ tations which may lead to categorial overlaps within one mechanism.<sup>2121</sup> What is more: Where the Commission, as the most prominent actor in the mechanisms presented above, is empowered to perform a hard mechanism, in practice it may, *qua* Article 292 TFEU, address a recommendation to

<sup>2121</sup> See eg Article 108 TFEU. With regard to new aids it lays down, according to para 2 *leg cit*, a hard mechanism. With regard to existing aids it may be perceived – if the measures to be proposed pursuant to para 1 *leg cit* are qualified as soft law (in Articles 22 f of Council Regulation 2015/1589 the term 'recommendation' is used; see IV.2.2.2.2.2. above) – as a mixed mechanism. Where this qualification is refused, and thus the proposal of measures is located below the level of soft law, the mechanism remains to be hard.

the MS concerned prior to adopting a decision and thereby render the mechanism, in its application in the concrete case, a mixed mechanism.<sup>2122</sup>

The classification in terms of the legal quality of the output adopted on the part of the EU actors involved was chosen here because it is the use of soft law which builds the focus of this work. Notwithstanding, there are other factors according to which the mechanisms presented here could be structured, eg the (number of ) EU actors involved (mono-, bior even poly-institutional mechanisms), the respective policy-field, whether it is a general or a special mechanism, whether its application constitutes day-to-day administration or forms the reaction to an emergency situation), etc. These and other factors shall be addressed in the following chapter on classification (2.).

At first, the actors involved in the compliance mechanisms shall be sin‐ gled out with a view to getting an idea of where the mechanisms are to be localised in an institutional perspective. After all, compliance mechanisms do not only entail procedural questions such as the sequence of acts or sub‐ stantial questions such as the material law to be applied, but they also raise institutional questions (2.1.). Then the policy fields within which the com‐ pliance mechanisms presented here have been established shall be looked at with a view to answering the question whether there are certain (types of ) policy fields which are more likely to display compliance mechanisms than others (2.2.). This matter is strongly connected to the question of the legal basis of (secondary law-based) compliance mechanisms.2123 Eventually, the output-related structure of the mechanisms shall be referred to. While the categorisation in hard, mixed and soft mechanisms was applied in Part IV, here the sequence of EU output in the single procedures shall be pinpointed with a view to better understanding their respective structure (ie their 'log‐ ic'; 2.3.). Another point to be addressed is the various purposes of providing for the adoption of soft law acts in compliance mechanisms. Against the background of the purposes of soft law more generally, as addressed under III.5. above, here we shall try to reveal and after that discuss the purposes of the soft law acts provided for in (mixed and soft) compliance mecha‐ nisms (2.4.). A further point of interest is the deviation from the Treaty infringement procedure, the general compliance mechanism and hence the

<sup>2122</sup> See III.3.4.3.2. above.

<sup>2123</sup> The legal bases of compliance mechanisms shall be recapitulated and analysed in more depth in the chapter on the legal assessment of the compliance mechanisms (see 3.2. below).

genuine point of reference in this context: While it is apparent that the special compliance mechanisms are all different from the latter in one or the other respect, it shall subsequently be explored in which way exactly they deviate from it (2.5.). Finally, the question why a concrete compliance mechanism 'looks the way it looks' shall be addressed, and a number of the multifaceted (actual or potential) reasons for the various designs of compliance mechanisms (efficiency considerations, weak EU competence, genuinely 'political' reasons, etc) shall be disclosed (2.6.).

With respect to the legal assessment of the above compliance me‐ chanisms, the distinction between implementation and enforcement in the system of the Treaties and the characteristics of either of these categories shall be fleshed out. This distinction is central for various aspects of the legal assessment of the compliance mechanisms, which is why it stands at the beginning of Chapter 3 (3.1.). This leads us to the question of legality. Against the background of their categorisation in terms of implementation and enforcement, the secondary law-based compliance mechanisms pre‐ sented above shall be examined with a view to the primary law provisions on which they are based. In this context Article 114 TFEU, as the most frequently used legal basis, stays in the centre of the discussion. Other, more specific legal bases for the establishment of compliance mechanisms shall be approached thereafter (3.2.). In this context, we can also draw on the general discussion on the (primary law) competences of soft law contained in Part III (III.3.4. above). Subsequently, we will turn to the EU's institutional balance. It shall be scrutinised whether the secondary law-based compliance mechanisms distort this competence-related balance, in particular: whether they constitute – on their respective own or *in toto* – an unlawful deviation from the Treaty infringement procedure, or whether they are in accordance with this equilibrium (3.3.). As a next step, secondary law-based compliance mechanisms shall be perceived against the background of the principles of subsidiarity and proportionality, and the question to which extent these principles (should) influence the design of soft, mixed and hard mechanisms shall be addressed (3.4.). After that, the allegedly different effects of soft law with special attributes (eg requiring 'ut‐ most account' to be taken of them) as compared to 'regular' soft law shall be examined with a view to whether the former display a higher authority than the latter and whether therefore a hierarchy of soft law can be established in this context (3.5.). Eventually, the judicial review MS can avail themselves of against the EU output adopted in the course of compliance mechanisms shall be presented, thereby also drawing from the results of Chapter III.6. above (3.6.).

# *2. Classification*

# 2.1. The EU actors involved in compliance mechanisms

One way of looking at the above compliance mechanisms is to take an institutional perspective, that is to say to consider the EU bodies<sup>2124</sup> in‐ volved, and among them in particular the originators of output addressed to the MS concerned.<sup>2125</sup> Let us begin with the general compliance mecha‐ nism: The Treaty infringement procedure places the Commission at the core of the first, the administrative part of the procedure. Even in the rarely applied variant pursuant to Article 259 TFEU, which grants the MS the power to launch an infringement procedure before the Court, the Commis‐ sion may step in, in its genuine role as guardian of the Treaties. Only after the Commission has adopted its opinion or where it does not deliver a reasoned opinion within three months, the MS may bring the matter before the Court. While the Court takes the final decision on whether or not a MS has failed to fulfil an obligation under the Treaties, the Commission dominates the procedure, as it has a large measure of discretion in deciding whether or not the procedure is initiated in the first place;2126 as it has considerable leeway as to whether and, if so, when2127 it files an action with the Court; as it may – in negotiations with the MS concerned and hence 'diplomatically' – settle the dispute before or after the CJEU is addressed and therefore (or for other reasons, thereby again disposing of a certain

<sup>2124</sup> In compliance mechanisms *vis-à-vis* MS the non-EU actors are – as a matter of course – in particular the MS concerned (including its authorities).

<sup>2125</sup> For the increasing organisational 'Europeanisation' of the enforcement of EU law (also *vis-à-vis* individuals) see Scholten, Trend.

<sup>2126</sup> For an early example of the Court criticising the Commission's hesitation in this context: case 43/75 *Defrenne*, paras 72 f.

<sup>2127</sup> See case C-177/03 *Commission v France*, paras 17 f; see also Gil Ibáñez, Supervision 174, including references to further case law.

latitude2128) end the procedure until the hearing before the Court.<sup>2129</sup> The dominance of the Commission extends to the sanctions regime pursuant to Article 260 TFEU, which is generally shorter and has 'a much narrower ambit [than Article 258 TFEU]',<sup>2130</sup> but still leaves room for (also informal) communication between the Commission and the MS concerned.<sup>2131</sup> In spite of procedural differences, the core role of the Commission as 'promo‐ ter' of compliance with EU law *vis-à-vis* the MS is undisputed also in the procedures laid down in: Article 108 para 2, Article 114 para 9 (here, similar to Article 259 TFEU, also MS may act as promoters), Article 348 para 2 TFEU. According to Article 271 TFEU, exceptionally (ie under the condi‐ tions laid down in this provision) it is the EIB, and the ECB respectively, which enjoy the powers conferred upon the Commission by Article 258 TFEU.

Under Article 106 para 3 TFEU it is the Commission only which is empowered to ensure MS' compliance with EU law in the case of public un‐ dertakings and undertakings to which MS have granted special or exclusive rights.

According to other regimes laid down in the Treaties, the Commission is involved in ensuring MS' compliance with EU law, together with the Council (Articles 121, 126, 144 and 148 para 4 TFEU) or together with the European Parliament and the Council (Articles 116 f TFEU).

The dominance, or at least strong involvement, of the Commission in – leaving the exceptional variants to the Treaty infringement procedure according to Article 271 TFEU apart – all compliance mechanisms laid down in the Treaties, also beyond Articles 258 and 260 TFEU, fleshes out the role of the Commission as guardian of the Treaties as enshrined in Article 17 TEU.<sup>2132</sup>

<sup>2128</sup> For the example of the closing of infringement procedures (a so-called *classement*) and the treatment of complaints in the area of gambling in order 'to be more strategic in enforcing EU law' see Commission, Press release of 7 December 2017, IP/17/5109.

<sup>2129</sup> See Articles 147 f of the Rules of Procedure of the Court of Justice; see also eg Euro‐ pean Parliament, 'The relationship between the Commission acting as guardian of the EU Treaties and complainants: Selected topics' (Note, 2012) 6 f.

<sup>2130</sup> Joined cases C-514/07P, C-528/07P and C-532/07P *API*, para 119, with further references.

<sup>2131</sup> See Andersen, Enforcement 103.

<sup>2132</sup> See also Commission, 'A Europe of Results – Applying Community Law', COM(2007) 502 final, 3 f.

As regards (our selection of ) compliance mechanisms laid down in secondary law the scene looks decidedly different. In the hard mechanisms we can perceive an involvement of the Commission which ranges from it being the sole promoter (Article 13 para 1 of Directive 2001/95/EC) to the Commission being supported by the expertise of European agencies (Arti‐ cles 70 f of Regulation 2018/1139, EASA; Article 63 of Regulation 2019/943, ACER) or to (other) MS being involved (safeguard clauses), to the Com‐ mission being ousted by newly established European agencies (Article 29 para 2 of Regulation 806/2014, SRB; Articles 18 f of Regulation 1093/2010, EBA).

In the context of mixed compliance mechanisms, it is apparent that pow‐ ers are frequently shared between the Commission and European agencies, with either the Commission (Article 63 of Directive 2019/944, ACER; Article 25 of Regulation 2016/796, ERA) or the respective agency (Article 17 of Regulation 1093/2010, EBA) taking the upper hand.2133 In one instance the Commission alone2134 (Articles 22 f of Council Regulation 2015/1589), in another instance a European agency alone (Article 7 para 4 of Regula‐ tion 806/2014, SRB) conducts the respective compliance procedure. In the excessive imbalance procedure as laid down in Regulations 1176/2011 and 1174/2011, it is the Council together with the Commission.

When it comes to soft compliance mechanisms, there are again regimes providing for the Commission alone as promoter of compliance (Article 53 of Directive 2019/944; Article 6 of Regulation 2019/452), for the Com‐ mission together with another body (Article 6 paras 5–7 of Regulation 2019/942, ACER; Article 33 of Directive 2018/1972, BEREC) or for other institutions (Article 3 para 7 of Regulation 472/2013, Council) or bodies (Articles 16 f of Regulation 1092/2010, ESRB) on their respective own.

In case of some of the above – hard, mixed, or soft – regimes, the MS (eg in the form of national authorities) have the right to request the initiation

<sup>2133</sup> See Alberti, Actors 40 f; for a further category of compliance instruments involving European agencies see the example of the 'alert-warning system'; see Roadmap on the follow-up to the Common Approach on EU decentralised agencies 1, 2; see also eg Vos, Agencies 31 ff. For the 'political and not […] legal dimension' of this tool see Commission, 'Progress report on the implementation of the Common Approach on EU decentralised agencies', COM(2015) 179 final, 7.

<sup>2134</sup> Only later in the procedure the CJEU may take over the lead; see Article 28 of Council Regulation 2015/1589.

of the respective procedure, hence they are additionally involved.2135 This ought to be mentioned here, even though the focus of this sub-chapter is on EU actors participating in compliance mechanisms.

In conclusion, we can say that a comparison between compliance mech‐ anisms laid down in primary law and compliance mechanisms laid down in secondary law with regard to the actors involved (and in particular the originators of output addressed to the MS concerned) shows that the dominance of institutions in primary law is relativised in favour of newly established bodies, not only but in particular European agencies.<sup>2136</sup> Since the Commission takes the lead or is at least engaged in most of the respec‐ tive processes (often with a comitology committee being involved<sup>2137</sup>) it can be depicted as a constant. It is therefore fair to say that also with regard to the secondary law-based mechanisms presented here, the Commission – in a holistic perspective – is the main guardian of MS' compliance with EU law.

Another (potential) actor is the CJEU. While it is addressed in the context of the Treaty infringement procedure, its involvement, above all on the basis of Article 263 TFEU, hovers over all hard and mixed mechanisms. The Court is the highest-ranking interpreter of EU law and may, in the course of an annulment procedure, authoritatively decide on whether or not an act of an institution, body, office or agency of the Union which is intended to produce legal effects *vis-à-vis* third parties is in compliance with (higher-ranking) EU law. It may exercise the latter competence (only) when called upon (in particular by the MS concerned) in the course of or following the application of one of the above mechanisms, except for the soft ones. Since EU law acts not intended to produce legal effects *vis-à-vis* third parties pursuant to Article 263 para 1 TFEU are excluded from judicial

2137 For an early example of a compliance mechanism provided for in the field of air safety and for the questionable role the advisory committee of MS representatives has played in a rather 'political' case see Gil Ibáñez, Exceptions 164 f.

<sup>2135</sup> See in particular Article 114 para 9 and Article 348 para 2 TFEU, Articles 11 and 39 of Directive 2009/72/EC, Article 17 of Regulation 1093/2010, Article 7 para 4 of Regulation 806/2014, Article 6 paras 5–7 of Regulation 2019/942. In other mechanisms the competent EU actor may act upon a notification by one or more MS.

<sup>2136</sup> Describing the role of agencies in enhancing compliance as application of a 'co‐ ercive strategy' and a 'persuasive strategy' (among other measures: soft law): Ver‐ sluis, Catalysts 179; for the activity of European agencies or comitology committees leading to an institutionally mixed (EU) administration see Orator, Möglichkeiten 25, with further references; see also Britz, Verwaltungsverbund 51–53.

review, with soft mechanisms the route of an annulment procedure is not available to the MS (nor to any other of the potential claimants listed in Article 263; see also 3.6. below).

The competences the CJEU may exert with regard to hard and mixed compliance mechanisms should always be borne in mind as a potential add-on to the respective procedure, even if – unlike in the case of the Treaty infringement procedure and its variants – in a *locus legis* perspective they are laid down separately. The possibility of an application for judicial review with the CJEU in all hard and mixed mechanisms may be perceived as constituting a common factor with the Treaty infringement procedure and its variants, may the applicants in the former group of procedures be the MS concerned.<sup>2138</sup> While it is inherent in the Treaty infringement procedure and its variants, and not a separate procedure as in case of the other compliance mechanisms concerned, from a perspective of legal certainty the CJEU performs a similar role, namely that of a final authority which is only potentially addressed. On the other hand, there are significant differences between these procedures, in particular regarding their respec‐ tive *telos*: While the Treaty infringement procedure is aimed at establishing whether or not a MS has violated EU law, in the annulment procedure the Court primarily examines the legality of a Union act. Here the Court, if at all, considers the lawfulness of MS action only indirectly.

The possibility of an involvement of the Court always exists, but in practice it does not always materialise. By far not all legally binding acts adopted in the course of a compliance mechanism and addressed to a MS are made subject to judicial review (for various reasons2139), and, as was mentioned above (IV.2.1.2.), the vast majority of Treaty infringement procedures which are initiated end prior to the Court having rendered its judgement. Thus, in the given context, the Court as highest interpreter of EU law only sometimes comes into play directly, and in most cases the

<sup>2138</sup> With regard to the Treaty infringement procedure, in practice this is only excep‐ tionally the case.

<sup>2139</sup> Take the example of the fine imposed on Austria for the misrepresentation of gov‐ ernment debt by Council Implementing Decision 2018/818 (based on Regulation 1173/2011), against which Austria refused to file an action for annulment. This was because on the one hand it feared that the CJEU could increase the fine, and on the other hand the *Bund* as addressee of the fine could, according to national law, claim a refund from the *Land* Salzburg whose authorities actually took the incriminated action; see <https://derstandard.at/2000086136034/Oesterreich-bez ahlt-Millionenstrafe-der-EU-nach-Salzburger-Finanzskandal> accessed 28 March 2023.

allegedly correct interpretation of EU law is, for a concrete case, determined by other EU actors. This interpretation will usually be based on previous case law of the Court (if any), which allows the latter to indirectly influence these cases.

### 2.2. The policy fields and the primary legal bases concerned

All of the specific compliance mechanisms addressed above must root in a policy field in which the EU is competent to act. In addition to that, they must all be based on the Treaties, either directly or indirectly. If the compliance mechanism is laid down in primary law only, the legal basis indicates the policy field. Also in case of compliance mechanisms laid down (also) in secondary law, the policy field concerned and the primary legal basis applied regularly correspond to each other. If not, the legislator may have chosen the wrong legal basis, resulting in the unlawfulness of the act of secondary law.

While the Treaty infringement procedure due to its general scope can‐ not possibly be assigned to a specific policy field, the other primary law mechanisms – except for the variants of the Treaty infringement procedure, as laid down in Article 271 TFEU – are mainly laid down in the field of approximation of laws, competition law and economic policy. As regards the secondary law mechanisms, Article 114 TFEU – regardless of whether the mechanisms are hard, mixed or soft – is by far the most frequently applied legal basis, by a wide margin followed by other legal bases, such as those on economic, transport or energy policy. While the selection of secondary law compliance mechanisms taken here may not be representa‐ tive in all respects, the frequency with which the legislator made use of Article 114 TFEU is significant and it appears to indicate the importance of Article 114 TFEU as a legal basis for compliance mechanisms also beyond this selection. This marries well with the fact that Article 114 TFEU in general is considered one of the most important legal bases for secondary law.2140 As Article 114 TFEU relates to nothing less than the establishment and functioning of the internal market and hence has a very broad scope, it does not come as a surprise that also the legislative acts based on it provide for compliance mechanisms in a wide range of areas, such as

<sup>2140</sup> See eg Classen, Art. 114 AEUV, para 5.

product safety, electronic communications or the supervision of banks.<sup>2141</sup> What is more, Article 114 TFEU not only provides for the incorporation of safeguard clauses in secondary law, in its paras 4 ff it contains a compliance mechanism itself. The latter may have served as a source of inspiration for one or the other compliance mechanisms based on this provision.

In conclusion, we can say that compliance mechanisms are established in all kinds of policy areas, and that the prevalence of Article 114 TFEU as a legal basis for setting-up compliance mechanisms does not result in the possibility to assign them all to one concrete policy area. Due to the malleability of Article 114 TFEU and the encompassing nature of the EU's internal market concept, the contrary is the case. While this brief account of the primary legal bases used may therefore have helped little in localising our selection of compliance mechanisms, fleshing out the importance of Article 114 TFEU also in this context suggests a closer examination of what all (in particular: which kinds of compliance mechanisms) may be based on this provision (see 3.2.2. below).

# 2.3. The sequence and addressees of acts in compliance mechanisms

As regards the structure in terms of (individual-concrete) output (hard law, soft law), the mechanisms presented above can also be classified in terms of the peculiar sequence and the addressees of these acts.2142 The Treaty infringement procedure according to Article 258 TFEU provides for two acts: a soft law act, the reasoned opinion adopted by the Commission, and a hard law act, the judgement of the CJEU. The procedure follows the con‐ cept that first there should be an attempt to convince the MS concerned by means of soft law, and only when this attempt turns out to be unsuccessful, an intervention by means of law should be made (law as *ultima ratio*).<sup>2143</sup>

<sup>2141</sup> The scope of what is now Article 114 TFEU can be reduced by the creation of more specific legal bases, eg Article 194 TFEU on energy policy; for a compliance mechanism based on the latter Article see IV.2.2.1.2.4. above. Its predecessor was contained in Regulation 714/2009 which was still adopted on the basis of Article 95 TEC (the predecessor of Article 114 TFEU).

<sup>2142</sup> For a classification of similar 'models of enforcement', taking account, among other things, of the legal (non-)bindingness of the respective output see Schol‐ ten/Ottow, Design 85 ff.

<sup>2143</sup> For the role of the Commission opinion under Article 259 TFEU see 2.4.2. below. That the advantages of such a tiered approach may, according to the Court, be used also in other contexts is exemplified in case 245/81 *Edeka*, para 22; for

Most of the hard compliance mechanisms addressed here are composed of only one act. An exception forms Article 108 para 2 TFEU which pro‐ vides for a Commission decision (potentially) to be followed by a Court judgement. In derogation from Articles 258 f TFEU, the Commission (or any other interested MS) may directly refer the case to the CJEU if the MS addressed by the Commission decision does not comply with it within the prescribed time. Similarly, also the mechanism laid down in Article 114 TFEU may result in two subsequent acts. Another exception are Articles 18 f of Regulation 1093/2010, according to which the EBA addresses a decision to a national authority (for our purposes that means: a MS<sup>2144</sup>) which may – if not complied with – be followed by a decision addressed directly to the financial institution/financial sector operator concerned, with a (material) blocking effect preventing the competent authorities concerned from ruling on this matter in a different way and hence having legal effects also for the respective MS.<sup>2145</sup>

The mixed compliance mechanisms show a more complex structure. This is hardly surprising, as they – *qua* being mixed – need to provide for at least two acts: a soft one and a hard one. In all our examples of mixed pro‐ cedures, a hard law act is preceded by a soft law act. While EU law does not provide for any automatism in this respect, this sequence seems to confirm that the law-maker provides for an attempt to convince the MS by means of soft law, which is perceived as less dominating or even less aggressive and is actually less strongly interfering with (potential) MS prerogatives. Only where soft law acts fail in reaching this objective, that is to say where they are not complied with, a hard law act may be adopted to follow. This logic (which is known from the Treaty infringement procedure and its variants

another example in the context of information-gathering by the Commission in competition law see Article 18 paras 2 f of Council Regulation 1/2003. While the procedure was obligatorily tiered under Regulation 17/62 (case 136/79 *National Panasonic (UK)*, para 10), according to the letter of the law of Regulation 1/2003, the soft and the hard request for information may as well be used as alternatives; for the discretion the Commission has in this context see Hennig, Auskunftsver‐ langen, para 19.

<sup>2144</sup> See also 3.1.1.2.1.3. below.

<sup>2145</sup> The 'competent authority' may not only be national authorities but may – in the field of banking supervision – also be the ECB as empowered under the SSM; see Article 4 para 2 (i) of the Regulation 1093/2010 (as amended). In this case, it would be the ECB as an institution of the EU which is affected by the EBA decision; critical of the partial superiority of the EBA as compared to the ECB: Weismann, Agencies 195.

according to Article 271 TFEU) is followed in the mechanisms laid down in Articles 116 f TFEU, in Article 63 of Directive 2019/944, in Articles 22 f of Council Regulation 2015/1589, in Article 7 para 4 of Regulation 806/2014 and in Article 25 of Regulation 2016/796. All of these procedures provide for a soft law act adopted by the Commission (exceptionally: EIB/ECB) or a European agency (ACER, SRB, ERA) which is followed by a hard law act adopted by one (exceptionally: two) EU institutions (exceptionally: the SRB).

Slightly extended is the procedure laid down in Article 17 of Regulation 1093/2010, in which the recommendation of the EBA may be reinforced by the Commission's formal opinion. If also the latter is not complied with, the EBA may address a decision directly to a financial institution/financial sec‐ tor operator. In the specific case of the prevention of the use of the financial system for the purpose of money laundering or of terrorist financing, where the relevant legislative acts are not directly applicable to financial sector operators, the EBA decision addressed to the financial sector operator concerned is preceded by a decision addressed to the respective national authority.

An almost flamboyant compliance mechanism in the categories dis‐ cussed here is Article 126 TFEU. It starts with a Commission opinion which may be followed by a Council decision together with Council recommen‐ dations. If the recommendation is not complied with, the Council may give notice to the MS to take the respective measures by means of a decision. When also this does not help, the Council may impose sanctions (by means of a decision), and intensify them (by means of another decision), if need be. If we split this long-winded procedure, we can see that also here the logic described above is followed: The Commission opinion suggesting that there is an excessive deficit may be followed by a Council decision stating that – in a legal understanding – there actually is an excessive deficit. This decision is combined with Council recommendations initiating the next step of the procedure. Where these recommendations on how to remedy the excessive deficit in due time are not complied with, the Council may reinforce them by giving notice to the MS in the form of a decision. Where the MS does not react to this decision in a satisfying way, either, the Council may – as a third step – sanction the MS by means of a decision, and intensify the sanctions respectively (again by means of a decision).

Also the excessive imbalance procedure laid down in Regulations 1176/2011 and 1174/2011 needs to be split in order to understand the *telos* of the sequence of soft and hard law acts. At first, there is a Council rec‐ ommendation regarding the existence of an excessive imbalance, followed by a Council recommendation reacting to a corrective action plan submit‐ ted by the MS concerned. This recommendation lays down the details of implementation or requests the MS to submit a new action plan. If this rec‐ ommendation is not complied with, there may be a Council decision estab‐ lishing the MS' non-compliance with the last recommendation (this meets again the above described logic: soft exhortation, hard reinforcement), possibly (and only for euro MS) combined with a sanctioning decision (imposing an interest-bearing deposit), and a recommendation setting new deadlines. This recommendation initiates the second part of the procedure: The Council shall impose an annual fine where in one and the same procedure there are two successive Council recommendations requesting the submission of a new corrective action plan or two successive Council decisions establishing non-compliance.

Similarly to the hard mechanisms addressed here, also the selection of soft mechanisms is often composed of one act only. An explanation for this could be that within a soft mechanism a soft law act cannot be reinforced by a more compelling subsequent act. However, this argumentation does not consider the fact that also soft law acts may entail different degrees of authority, eg depending on the body adopting the act or, less often, on the category of act (see 3.5. below). This is exemplified by the procedure laid down in Article 33 of Directive 2018/1972, starting with mere comments of the Commission and potentially followed by a BEREC opinion, which may – in case of non-compliance by the MS addressed – again be intensified by a Commission recommendation. Thus, the regime of Article 33 provides for two (including the notification or comments: three) soft law acts aimed at ensuring compliance of a MS with EU law, the BEREC opinion and the Commission recommendation. The fact that both acts essentially serve the same purpose, to ensure compliance with EU law that is, but are named differently – opinion on the one hand, recommendation on the other hand – confirms the close proximity of these two acts. However, in accordance with the low-key conceptual distinction between recommendations and opinions fleshed out above, we can see that the (BEREC) opinion mainly reacts to the Commission notification, whereas with the (Commission) rec‐ ommendation its commanding character – a concrete (soft) command to a national authority to amend or withdraw the draft measure – is emphasised (see III.3.1.1. above). This procedure allows for both the Commission and an expert body to be involved and is, in that respect, similar to some of the mixed mechanisms. Also the multilateral surveillance procedure pursuant to Article 121 TFEU provides for a sequence of two soft law acts seeking to ensure compliance by the MS. The BEPG adopted by the Council in the form of a recommendation are the primary threshold against which compliance with EU economic policy is to be examined. In the course of this examination procedure, the Commission may address a soft law act (a warning) to a MS which may be reinforced by a Council recommendation.

The soft law mechanism enshrined in Article 148 TFEU (Council recom‐ mendation) allows for one soft law act to be adopted. Within this category also fall the procedures laid down in Article 6 paras 5–7 of Regulation 2019/942 (ACER opinion), Article 53 of Directive 2019/944 (Commission opinion), Article 3 para 7 of Regulation 472/2013 (Council recommenda‐ tion), Articles 16 f of Regulation 1092/2010 (ESRB recommendation) and Article 6 of Regulation 2019/452 (Commission opinion).

The output adopted by EU actors is regularly directed to one or more MS. Sometimes the addressee is specified as the national authority com‐ petent in the respective field,2146 sometimes other actors are additional addressees.2147 Exceptionally, the output is directed to private actors. This is the case with the mechanisms laid down in Articles 17–19 of Regulation 1093/2010 and Article 29 para 2 of Regulation 806/2014 respectively, pur‐ suant to which a decision – not a soft law act – is addressed to the financial institution/financial sector operator, and the institution under resolution respectively. Where the breach of EU law is caused by a private actor (not primarily by a MS2148), this is the most direct way of redressing it. Both under Articles 17–19 of Regulation 1093/2010 and under Article 29 para 2 of Regulation 806/2014 this way is provided for the case that the – principally competent – national authorities do not follow the application or interpre‐ tation of EU law as provided for by the respective EU actor involved.

<sup>2146</sup> See eg Article 63 para 8 of Regulation 2019/943, Articles 17–19 of Regulation 1093/2010, Articles 11 and 39 of Directive 2009/72/EC, Article 7 para 4 of Regula‐ tion 806/2014.

<sup>2147</sup> See eg Article 114 para 9 and Article 348 para 2 TFEU: Commission; Article 108 TFEU: (legal) persons concerned.

<sup>2148</sup> A MS (authority) not abolishing a breach of Union law by a private actor may be breaching Union law itself, but only due to the private actor's behaviour; see eg case C-265/95 *Commission v France*.

### 2.4. The purposes of soft law acts in compliance mechanisms

### 2.4.1. The question of command

As was discussed at some length in Part II above, soft law constitutes an *aliud* as compared to law. Public authority traditionally is regulated by and itself regulates by law. The use of soft law, while being on the rise, is still 'atypical'.<sup>2149</sup> This leads us to the purposes which the use of soft law serves. In the context of soft law more generally (but with a focus on collective rules), the French *Conseil d'État* has listed four functions (see III.5.1. above), each of which is related to law (as the regular case of rule-making): substitution (*substitut*), preparation (*préparation*), company (*accompagnement*) and permanent alternative (*alternative pérenne*). In the given context, we are dealing only with one segment of soft law, namely individual-concrete soft law. With regard to the mixed and the soft mech‐ anisms presented above – hard mechanisms do not entail any soft law act – we shall now examine whether and, if so, which of these and other functions or purposes soft law may serve in each case.

A possible point from where to start in this context is the command contained in soft law. What do the soft law acts adopted in the course of the compliance mechanisms presented here actually tell? Most of the time, these acts tell their respective addressee(s) to take, or to refrain from respectively, certain action. Sometimes, however, they appear to limit themselves to determining a certain situation, eg in the reasoned opinion according to the Treaty infringement procedure and its variants it is stated that an infringement exists. Whether or not it also contains a concrete (soft) command, ie a guideline on how to remedy this infringement, may vary from case to case.2150 However, even where there is no such explicit command, the statement of an infringement itself certainly effectuates an implicit command; it can be linked to a command.<sup>2151</sup> After all, it is com‐ mon sense that in a system based on law unlawful situations are undesirable and therefore to be rectified. That in a concrete case the establishment of an unlawful situation by means of soft law – and hence in a legally

<sup>2149</sup> See Arndt, Sinn 25, who describes the character of soft law as 'atypisch' [atypical].

<sup>2150</sup> In its reasoned opinions adopted in the course of a Treaty infringement proce‐ dure the Commission a number of times has stated the measures required for compliance, and hence has included an explicit command – without the CJEU's objection; see Gil Ibáñez, Supervision 94, with references to the CJEU's case law.

<sup>2151</sup> For this requirement see also II.2.1.1.1. above.

non-binding way – is acknowledged, and that the unlawful situation is subsequently rectified by its addressee may have different reasons, eg the authority of the concrete soft law act and its originator (in case of the Treaty infringement procedure: the reasoned opinion of the Commission). It may also be perceived by the addressee as the more 'peaceful' approach, especially where the infringement was done by mistake rather than on pur‐ pose.<sup>2152</sup> In our context, another potential reason ought to be mentioned: the further course of the procedure. Soft law may be followed by a hard law act (in case of the Treaty infringement procedure: a Court judgement) or may have (negative) effects for the addressee in a different procedure.<sup>2153</sup> The fact that soft law is thereby adopted in the shadow of hard law, as it were, certainly increases the 'persuasiveness' of its (implicit) command.<sup>2154</sup> Compliance with the soft law act at issue may create an expectation that thereby the unlawfulness is remedied, and that no follow-up action – either an act in the procedure (mechanism) at issue or the initiation/continuation of a Treaty infringement procedure – will be taken, at least not by the body which was involved in the procedure so far.2155 This applies to the Commission opinion softly determining the existence of an excessive deficit according to Article 126 TFEU; to the ACER opinion according to Article 63 of Directive 2019/944 which may consider the MS decision compliant or non-compliant with pertinent Union law, in the latter case containing an implicit command to remedy this non-compliance; to the Council rec‐ ommendation determining the existence of an excessive imbalance in the course of the excessive imbalance procedure; to the ERA opinion adopted

<sup>2152</sup> Note the words of Chayes/Chayes, Sovereignty 22: 'If we are correct that the principal source of noncompliance is not willful disobedience but the lack of capability or clarity or priority, then coercive enforcement is as misguided as it is costly'.

<sup>2153</sup> For example: The adoption of a reasoned opinion in the course of a Treaty infringement procedure may, if that is prescribed by secondary law, lead to the suspension of financial support granted within the framework of cohesion policy: see European University Institute, Research 22 and 47.

<sup>2154</sup> See Peters, Typology 426 f, using the expression 'shadow of the law' (emphasis in original); see also Aldestam, Soft Law 26; de Búrca/Scott, Introduction 6–10; Rošic Feguš, Soft law 54 ('shadow of hierarchy'); Ştefan, Enforcement 209; U Stelkens, Rechtsetzungen 407; discussing this phenomenon on a larger scale: Hér‐ itier/Lehmkuhl, Introduction; in the context of public international law: Franzius, Paris-Abkommen 524. For this metaphor in the context of private rule-making see Mnookin/Kornhauser, Bargaining, in particular 968 f.

<sup>2155</sup> See von Bogdandy/Arndt/Bast, Instruments 115, with a further reference.

pursuant to Article 25 of Regulation 2016/796, softly establishing that and why the draft national rules should not enter into force.

Also among the soft mechanisms both soft law acts establishing an in‐ fringement and soft law acts containing an explicit command are provided for in law. The Commission warning which may be adopted under Article 121 TFEU, for example, according to this provision either states that the economic policies of the MS addressed are not consistent with the BEPG or that they risk jeopardising the proper functioning of economic and mon‐ etary union. The subsequent Council recommendations, on the contrary, already according to the letter of the law contain soft commands on how to remedy the infringement (*argumentum* 'necessary recommendations'). The situation is similar under Article 33 of Directive 2018/1972, in which – following the Commission's notification – the BEREC may adopt an opinion on whether MS action complies with Union law, which may be followed by a Commission recommendation requesting concrete action. In our sample, it appears that the legislator (by tendency) has provided an opinion where a statement is to be made on whether or not a MS complies with relevant Union law (see eg Article 6 of Regulation 2019/452), and that it has envisaged a recommendation where concrete action is requested by its addressee (see eg Articles 16 f of Regulation 1092/2010). This is in accordance with a semantic distinction which can be drawn between these two categories of acts and which is sometimes, but not always reflected upon in general practice (see III.3.1.1. and III.3.9. above). However, and as was mentioned above, the line between statement and command is blurry, and where the law – in a systematic view – provides for concrete (negative) effects in case a MS does not adequately react to a statement made in an opinion (see eg Article 6 paras 5–7 of Regulation 2019/942), it is difficult to deny the (implicit) command effectuated by such an opinion.

#### 2.4.2. The Treaty infringement procedure

Before addressing the different purposes of soft law with a view to specific mixed and soft compliance mechanisms presented in Part IV above, we shall take a look at the purposes soft law has in the Treaty infringement procedure as the general compliance mechanism laid down in the Treaties. In the Treaty infringement procedure according to Article 258 TFEU (and in the respective variants) the soft law act – the reasoned opinion – serves a number of purposes detailing the broader objective, that is to convince its addressee to remedy the situation accordingly.2156 To a limited extent, the content of the reasoned opinion prepares a hard law act, namely the CJEU's judgement. Obviously, the CJEU is not bound in deciding the case, but the case itself is defined by the opinion (to which definition the subsequent action must stick<sup>2157</sup>) and in accordance with the principle *non ultra petita* the Court is bound by that definition or delimitation. For reasons of clarity, it ought to be stressed that the content of the reasoned opinion is coined already by the letter of formal notice, and that the Court is addressed only by an action. Thus, the reasoned opinion is only an in-between on the way to define the scope of the matter *vis-à-vis* the Court. Where the reasoned opinion suffices to convince the MS addressed and the alleged infringement is remedied accordingly, it also works as *alternative pérenne*, because it then settles this (individual) matter for good.<sup>2158</sup>

Under Article 259 TFEU the purpose of the Commission opinion is slightly different, as it is directed to two MS which regularly have opposing views on the matter at issue. In this procedure the Commission may not only support the allegations of the accusing MS, but it may as well express its view that the accused MS has not violated EU law. The Commission here exercises the function of a soft arbitrator, in its scope comparable to that of the Court.2159 If no recourse to the Court is made, the reasoned opinion also in this case may work as *alternative pérenne*.

### 2.4.3. The purpose of preparation – not always a matter of course

The purpose of soft law to prepare subsequent (binding) output is apparent in many compliance mechanisms. Sometimes, however, it is ousted by more dominant purposes of the soft law involved. In the following, we shall have a look at examples for both cases.

<sup>2156</sup> See case C-371/04 *Commission v Italy*, para 9, with further references.

<sup>2157</sup> For the strong link between the reasoned opinion and the action filed with the CJEU; see Gil Ibáñez, Supervision 178; Prete, Infringement 154–159, with referen‐ ces to the differentiated case law.

<sup>2158</sup> This does not imply an authoritative statement on the underlying questions of Union law. When considering that between 1978 and 2017 around 9,000 reasoned opinions were launched in Treaty infringement procedures which eventually re‐ sulted in around 2,000 Court judgements (in ca 90 % *against* the MS at issue), reasoned opinions seem to display a remarkable degree of effectiveness; for the data see Börzel, Noncompliance 28 f.

<sup>2159</sup> See Wunderlich, Art. 259 AEUV, paras 8 f, with further references.

In the regime laid down in Articles 116 f TFEU the Commission recom‐ mendation may – if it is adopted (prior to the directive) in the first place – serve as a preparation for a hard law act, a directive of the EP and the Council. While this is a legislative act, under Article 116 TFEU it is directed only to one or a small number of MS and hence has an individualised character.<sup>2160</sup>

Under Article 126 TFEU, the soft law acts each immediately preceding a hard law act are the Commission opinion and the Council recommenda‐ tion. The former allows for the involvement of the Commission (thereby objectifying the initiation of the procedure), the latter may allow for in‐ creased flexibility for both the Council and the MS addressed, meaning that the duration of the procedure is extended and thereby leaves the MS time to react. In a more political procedure such as the excessive deficit procedure it can be opportune for the decision-maker not to be forced to adopt a legally binding act, which may be appealed against only within a certain deadline and hence increases pressure on its addressee, but to resort to a soft law act which structures the procedure – a general (collateral) effect of soft law, in particular in mixed compliance mechanisms – and increases the political rather than the 'legal' pressure (ie: pressure to apply for judicial review in case of discontent with the command at issue). The preparation aspect of these acts is sidelined here by the Commission recommendations which are required for the Council to adopt its output. These (preparato‐ ry) Commission recommendations are addressed only to the Council and hence are not considered here *in extenso* (see also IV.1. above).

A purpose of the soft law acts discussed here which appears to be a sub-category of preparation is to involve the output of other (specialised or expert) bodies. This is apparent in the regimes laid down in Article 63 of Directive 2019/944 and Article 25 of Regulation 2016/796 in which a European agency (the ACER and the ERA, respectively) may adopt an opinion preceding a (possible) Commission decision. In Article 17 of Regu‐ lation 1093/2010 it is the way round. Following an EBA recommendation, the Commission may adopt a formal opinion, after which the EBA may adopt a decision. Here it seems that the Commission was involved in order to cater for democratic legitimacy and in particular in order not to shake the EU's institutional balance. <sup>2161</sup> The preparatory function is inherent in

<sup>2160</sup> See Eekhoff, Verbundaufsicht 127.

<sup>2161</sup> The content-wise continuity of the sequence of EBA/Commission acts (see IV.2.2.2.2.3. above) may be compared to the Treaty infringement procedure and

the involvement of a second body which adopts soft law, because the main reason for that certainly is to improve – in whichever way (expertise, political legitimacy, involvement of the different perspectives) – the quality of the subsequent hard law act provided for.<sup>2162</sup>

The mechanism laid down in Articles 22 f of Council Regulation 2015/1589 is mono-institutional, as is the procedure provided for in Article 7 para 4 of Regulation 806/2014. In the former, the Commission adopts recommendation and decision, in the latter the SRB addresses a warning to a MS, after which an SRB decision may follow. Also in these procedures the preparation aspect is apparent.

The excessive imbalance procedure laid down in Regulations 1176/2011 and 1174/2011 – from the perspective of the MS concerned – is mono-insti‐ tutional, as well. The preparatory function is fulfilled by the Commission recommendations (addressed to the Council), only upon which the Coun‐ cil may adopt its output *vis-à-vis* the MS concerned.

In the context of soft mechanisms, the purpose of preparation as pro‐ posed by the *Conseil d'État* – that is to say: preparation of law – does not apply for obvious reasons. However, in an adapted understanding, soft law may also serve the preparation of soft law.2163 The majority of soft mechanisms presented here consist of one act only, which means that no later (soft law) act can possibly be prepared in the course of these procedures. The mechanisms involving more than one act addressed to the MS concerned are laid down in Article 121 TFEU and in Article 33 of Directive 2018/1972, respectively. As regards Article 121 TFEU, the Com‐ mission warning addressed to a MS – a possibility introduced only by the Treaty of Lisbon – ought to increase the pressure on the MS.2164 It is not, however, to be seen as a preparation of the Council recommendations (potentially) subsequently adopted. They are rather prepared by the Com‐

the continuity required in its administrative phase; see W Cremer, Art. 258 AEUV, para 16.

<sup>2162</sup> For the potential of an improved reasoning of acts in this context see case T-576/18 *Crédit agricole*, para 138, with further references.

<sup>2163</sup> For the preparation of soft law by another act see the Commission's 'preliminary view' preceding its recommendation according to Articles 21f of Council Regula‐ tion 2015/1589. The 'preliminary view' arguably does not qualify as soft law, as it merely invites the MS concerned to submit its comments, but does not request compliance with the 'view'.

<sup>2164</sup> See Schulte, Art. 121 AEUV, para 14; see also Hattenberger, Art. 121 AEUV, para 34; Part, Art. 121 AEUV, para 26.

mission recommendation addressed to the Council. As regards the named procedure laid down in Directive 2018/1972, the Commission notification has the effect of preventing the adoption of the national draft measure at issue for three months. The BEREC may then issue an opinion on this notification, that is to say on its content. The Commission recommendation which may be released subsequently to the adoption (or for lack) of a BEREC opinion shall take "utmost account" of the latter (if any). Thus, if a BEREC opinion has been adopted, content-wise it has a strong influence on the Commission recommendation.<sup>2165</sup> The same is true, one procedural step ahead, for the Commission notification with regard to the BEREC opinion. The tiered output – Commission notification, BEREC opinion, Commission recommendation – ensures that both the Commission and the BEREC have a say, at the same time guaranteeing that the Commission has, if the procedure is applied in full, the first and the final word. There‐ fore, the Commission notification and the BEREC opinion meet a certain preparatory or preliminary function. At the same time, they are complete even without a 'follow-up' and have the capacity – where, in a concrete procedure, no subsequent act is adopted – to serve as the final EU output in a procedure.

# 2.4.4. The purpose of company ('accompagnement') and the right to be heard

As regards the function *accompagnement*, we have to bear in mind that in the given context, where ensuring the 'individual' compliance of a MS with Union law is at issue, it is mostly the application of a general rule in a concrete case which is at issue. This application often entails concretisa‐ tion,2166 even though concretisation is normally understood as a collective explanation, not only an explanation given with regard to an individual case.2167 The conception of the Conseil d'État as 'le droit souple comme

<sup>2165</sup> For the effects of the requirement to take 'utmost account' of soft law and similar epithets see 3.5. below.

<sup>2166</sup> With regard to concretisation as the main purpose of implementation see 3.1.1.2.1.2. below.

<sup>2167</sup> For the concretisation by means of EU soft law see Storr, Wirtschaftslenkung 41f, who stresses, taking the example of Commission guidelines in state aid law, that concretisation by means of soft law *may* also lead to undue complexity; on the other hand, the MS are sometimes very keen on the Commission providing

instrument d'accompagnement de la mise en œuvre du droit dur' seems to be broad enough to cover both aspects.<sup>2168</sup> However, since the application of rules to a concrete case is a characteristic of the executive more generally, not only of the one that is expressed by soft law, this cannot be referred to as a unique characteristic of the soft law acts provided for in the compliance mechanisms at issue here.

In the context of these mechanisms, another (potential) function of soft law is to be examined: that is to fully inform the MS addressed of the respective allegations and to thereby ensure a comprehensive right to be heard before a (hard law) decision is taken on the matter.<sup>2169</sup> In this regard, it is to be noted that the initiation of compliance mechanisms is regularly – if the matter is not urgent – preceded by a (comparatively) informal contact on the part of the competent EU institution, body, office or agency (most often: the Commission). Thereby the MS is informed about the concerns of the EU actor at issue and is given the possibility to utter its point of view, before the mechanism as laid down in EU law is initiated. This is relatively well documented in the context of the Treaty infringement procedure, but arguably also applies to most other compliance mechanisms. For this reason, the meaning of soft law acts to allow for the MS addressed to react and thereby to make use of its right to be heard at an early stage of the procedure may be limited.<sup>2170</sup> Soft law has a strong informative function, nevertheless.

# 2.4.5. Substitution or permanent alternative – two purposes which often overlap

In its study on soft law, the French *Conseil d'Etat* has argued that soft law substitutes law where the adoption – for legal or political reasons – is not feasible. It serves as a permanent alternative (to law), however, where it is deemed to be the *better* regulatory approach (see III.5.1. above). Sometimes these purposes overlap. When compliance mechanisms are created on the basis of secondary law, the legislator has to consider the Treaties. They

<sup>(</sup>non-binding) guidance on the interpretation of EU law, in particular in the field of agricultural or state aid law: U Stelkens, Rechtsetzungen 408.

<sup>2168</sup> Conseil d'État, Droit souple 25.

<sup>2169</sup> See also 2.5.4. below. The function of soft law to inform more generally is ad‐ dressed also under III.5.2.1. above.

<sup>2170</sup> See case C-371/04 *Commission v Italy*, para 9.

determine the scope of action of the legislator. Since this scope often is not very clear, also the motivation of the legislator may be blurred: Has it opted for a soft mechanism because it deemed a mixed one to go beyond its competences or because it considered the former as the better approach – or have both motives played a role in the decision-making?<sup>2171</sup> In the following analysis this difficulty is always to be borne in mind. With mixed mechanisms the situation is less unclear: Mixed mechanisms provide for the possibility to adopt hard law in the end, so here soft law normally is not applied because hard law is not (legally or politically) feasible. This may only be the case where different EU actors are involved and one of them, eg a European agency, under primary law may not be granted the power to adopt a legally binding act in the given circumstances, but the legislator deemed its involvement important and hence provided for a soft law competence. Apparently, in these cases also the preparation aspect plays a significant role (see 2.4.3. above).

An example would be the mechanism laid down in Article 63 of Direc‐ tive 2019/944 which involves the ACER. The ACER may adopt an opinion, upon which the Commission can launch a decision requiring the national authority concerned 'to withdraw its decision for lack of compliance with the Guidelines'. The reason why the ACER here was not granted also the decision-making power – in other contexts the ACER does have individual decision-making power2172 – probably was the fact that a decision to with‐ draw a national measure means a legally binding review of national acts, a category of competence which is politically highly sensitive, and which is – if at all – for the Commission to exercise (note Article 17 para 1 TEU).<sup>2173</sup> By granting merely the power to adopt an opinion, the legislator stayed

<sup>2171</sup> See Senden, Soft Law 168, emphasising that the Commission hardly ever discloses its motives for adopting soft law instead of (initiating the adoption of ) law; for both substitution and preparatory effects of EU soft law see Snyder, Effectiveness 19. For the purpose of 'gaining experience' with a certain body of regulation: Brohm, Mitteilungen 76.

<sup>2172</sup> See Hauenschild, Agentur 108 f.

<sup>2173</sup> See also the (later withdrawn) draft Interinstitutional Agreement on the operat‐ ing framework for the European regulatory agencies, COM(2005) 59 final, 12, according to which European agencies shall not 'have responsibilities entrusted to them with respect to which the EC Treaty has conferred direct decision-making powers on the Commission' – a guideline which would not have been infringed by granting to the ACER a decision-making power in this case. Nevertheless, the sen‐ sitivity of the matter must have been apparent to the Commission/the legislator; for another example in which the ACER's power to adopt legally binding acts (as

on the safe side, as this power is certainly in accordance with the ACER's '[g]eneral advisory role'.<sup>2174</sup>

Another example is the procedure laid down in Article 17 of Regulation 1093/2010. Here the EBA may – where the national authority does not comply with Union law – eventually adopt a decision directly addressed to a financial institution/financial sector operator, thereby overruling any decision the national authority may have taken on this matter. This mecha‐ nism is different from the one involving the ACER. The EBA does not order the national authority to withdraw a decision, but it may – under certain circumstances (and in specific cases preceded by a decision addressed to the national authority) – adopt a decision of its own, thereby ousting the national authority. The effects may be similar to that of the above procedure involving the ACER, but its *modus operandi* is different. What is more: Non-compliance with EU law need not necessarily find its expression in an act, but may as well be effectuated by the national authority's omission. Thus, the EBA decision may also step in where the national authority has failed to act. This does not alter the fact that the EBA may determine situations which are normally for the national authorities to decide and that this – where a competence of the EU can be confirmed in the first place – is primarily a task of the Commission.2175 The legislator seems to have taken into account this argument. After all, the decision the EBA may take under Article 17 of Regulation 1093/2010 needs to be 'in conformity' with the preceding formal opinion of the Commission and may only be adopted exceptionally, namely 'where it is necessary to remedy, in a timely manner, such non-compliance in order to maintain or restore neutral conditions of competition in the market or ensure the orderly functioning and integrity of the financial system'.2176 This makes the EBA's decision-making power subject to only exceptional application and strongly dependent on the Commission's prior assessment of the matter by means of soft law. While the latter does not substitute the hard law act, but strongly 'complements' it, in a concrete case the formal opinion of the Commission may very well end

proposed by the EP) was rejected by the Commission see Orator, Möglichkeiten 382, with further references.

<sup>2174</sup> Commission Proposal COM(2007) 528 final, 12.

<sup>2175</sup> In the Commission proposal for what later has become Regulation 1093/2010, a true decision-making power of the Commission was provided for: see Article 9 of Commission Proposal COM(2009) 501 final.

<sup>2176</sup> See also Böttner, Mechanism 184 f.

the procedure for good, namely where it succeeds (ie: is accepted/complied with).

The ERA's power to adopt an opinion on national draft rules under Article 25 of Regulation 2016/796 could not have been lawfully transformed into a hard law power. First of all, and as was mentioned above, this is be‐ cause the empowerment of EU actors to review national acts is considered very delicate. This is not normally a task for European agencies (without Commission involvement) in general, and in particular the ERA overall may only exceptionally adopt decisions.<sup>2177</sup>

The reverse qualified majority voting applied in the excessive imbalance procedure actually diminishes the power of the Council and at the same time increases the power of the Commission, only upon the recommenda‐ tion of which the Council may act, thereby possibly distorting the concrete balance struck between these institutions in Article 121 TFEU, para 6 of which – in case of Regulation 1174/2011: together with Article 136 TFEU – forms the legal basis of the excessive imbalance procedure.2178 Apart from the enforcement measures pursuant to Regulation 1174/2011, the soft and the hard law powers are assigned in the spirit of Article 121 TFEU. The voting mode, however, increases the importance of the soft law act the Commission conveys to the Council. In other words: The conferral of soft power (to the Commission) and hard power (to the Council), as laid down in primary law (Article 121 TFEU), is formally complied with, but reverse qualified majority voting in the Council assigns an importance to the Commission soft law act addressed to the Council which it does not have under primary law. In conclusion, soft law here does not actually substitute hard law, but the likelihood of the (normative) content of the soft law act to be taken over in the form of a hard law act is increased.<sup>2179</sup>

With respect to soft mechanisms, the question of substitution/permanent alternative is even more prevalent because with them, unlike with mixed compliance mechanisms, also the final act – if more than one act is provi‐ ded for at all – is a soft law act. Thus, if hard regulation at the end of a procedure is seen as the rule, the idea that in soft mechanisms soft law substitutes or constitutes a permanent alternative to law springs to mind.

<sup>2177</sup> Article 4 lit e of Regulation 2016/796. Under the old founding regulation of the ERA, Regulation 881/2004/EC, the ERA was not competent to adopt any deci‐ sions; see Article 2 lit a and b *leg cit*.

<sup>2178</sup> Critically with regard to reverse (qualified) majority voting more generally eg Ruffert, Crisis 1800 ff; Palmstorfer, Majority; Weismann, Central Bank.

<sup>2179</sup> See also III.4.4. above.

As indicated above, the reason for the legislator taking this route may be a lack of legal or political feasibility of a mixed/hard mechanism and/or the conviction that a soft approach is considered more adequate.

As regards the procedure laid down in Article 6 paras 5–7 of Regulation 2019/942, the remarks made above on the scope of the ACER's powers apply. In its legislative proposal for what has become Regulation 713/2009, the predecessor of Regulation 2019/942, the Commission has emphasised that the ACER's decision-making power shall be limited to certain cases with a cross-border dimension (*argumentum* 'concerning the infrastructure in the territory of more than one Member State').<sup>2180</sup> Whether this is what the Commission has originally wanted or whether this is an example of an adaptation of political wishes to legal feasibility must be left open here.

The mechanism laid down in Article 53 of Directive 2019/944 is about the (soft) examination of a national (draft) measure. As a form of on-going supervision of day-to-day MS administration it is considered a sensitive issue.2181 Even the Commission in the underlying legislative procedure em‐ phasised the importance of 'keeping national regulators' centre role in ener‐ gy regulation'.2182 It appears that here the Commission and the legislator did not intend to grant the ACER (further) hard law powers. Therefore in this case the purpose of permanent alternative arguably dominates.

Also the mechanism laid down in Article 33 of Directive 2018/1972 is about the examination of a draft measure of a national authority by the Commission. Here the Commission's lead – with the BEREC being in‐ volved as an expert body – was proposed by policy advice the Commission as initiator of the pertinent legislation had received.2183 While the Commis‐ sion proposed the involvement of comprehensive hard decision-making power on its part (where the BEREC shared the Commission's concerns), the legislator refused this.2184 From the legislator's perspective, this points in the direction of the permanent alternative purpose.

<sup>2180</sup> Commission Proposal COM(2007) 530 final, 11; with regard to the current Reg‐ ulation, the Commission held that 'the powers of ACER for those cross-border issues which require a coordinated regional decision' ought to be strengthened; Commission Proposal COM(2016) 863 final, 7.

<sup>2181</sup> For the notion of on-going supervision (control) see Busuioc, Accountability 607.

<sup>2182</sup> Commission Proposal COM(2016) 864 final, 12.

<sup>2183</sup> Commission Proposal COM(2016) 590 final/2, 12 f.

<sup>2184</sup> The room for decision-making power now laid down in Article 33 para 5c of Di‐ rective 2018/1972 – which was left aside in the above description of the mechanism as an exceptional variant – was significantly reduced in its scope (as compared

The mechanism laid down in Article 3 para 7 of Regulation 472/2013, concluding from its legal basis, shall concretise Article 121 paras 3 f TFEU (Article 121 para 6 TFEU) with regard to the Euro-MS (Article 136 TFEU). Since Article 121 paras 3 f TFEU only provide for soft law acts of the Council, it is consistent that the concretisation in secondary law does as well. The purpose of soft law here is not to interfere with what is considered a prerogative of the MS, but to flesh out EU powers upon which the MS as Masters of the Treaties have already agreed in the form of Article 121 TFEU. Whether the legislator actually would have preferred to go further is unclear, even though other measures to reinforce the multilateral surveil‐ lance procedure seem to suggest so.<sup>2185</sup>

The power of the ESRB to adopt a recommendation as laid down in Arti‐ cles 16 f of Regulation 1092/2010 was intended to be soft already when the Commission made the respective legislative proposal. Being conceptualised as a mere '"reputational" body', no hard law powers have been envisaged for the ESRB at all.2186 Therefore this seems to be an example of permanent alternative.

According to Article 6 of Regulation 2019/452, based on Article 207 para 2 TFEU,2187 the Commission may adopt an opinion on a planned foreign direct investment where it is 'likely to affect security or public order in more than one Member State, or has relevant information in relation to that foreign direct investment'.2188 Whereas in general the screening of foreign direct investments into the EU is up to the MS, the Commission may excep‐ tionally step in. Since it may only render an opinion, 'Member States keep the last word in any investment screening'.2189 From a legal perspective, it does not appear that under Article 207 para 2 TFEU the Commission could not have been granted more far-reaching powers, in particular a proper decision-making power.2190 The relatively weak involvement of the

to the Commission proposal); see Commission Proposal COM(2016) 590 final/2, 12 f.

<sup>2185</sup> See eg Regulation 1173/2011 as addressed under 3.6. below.

<sup>2186</sup> Commission Proposal COM(2009) 499 final, 5.

<sup>2187</sup> For different opinions on the correct legal basis and on the correct legislative procedure see Hindelang/Hagemeyer, Enemy 882 (fn 5); see also Klamert, Loyalty 216.

<sup>2188</sup> Article 6 para 3 of Regulation 2019/452.

<sup>2189 &</sup>lt;https://policy.trade.ec.europa.eu/enforcement-and-protection/investment-screen ing\_en> accessed 28 March 2023.

<sup>2190</sup> Sceptically: de Kok, Framework 45.

Commission seems to have been a political wish (to have law as the regular case substituted by soft law) rather than legal necessity.<sup>2191</sup>

### 2.4.6. Institutional transformation as a purpose of soft law?

Thus far, we have analysed the immediate purposes of soft law, thereby focusing on the specific acts (legal sources perspective). Now we shall take a more institutional perspective and examine whether soft law may be used as a tool to facilitate institutional transformation in the EU. The main question is the following: Do soft law powers play a core role in the process of secondary law-based specific compliance mechanisms 'competing with' and hence relativising the practical importance of the Treaty infringement procedure, whereby – at least *de facto* – the powers of the Commission and the Court under primary law are shifted in particular towards the Commission (under secondary law) and to European agencies?

It is true that a number of specific compliance mechanisms are laid down in primary law and hence already from a primary law perspective the Treaty infringement procedure is not envisaged as the exclusive compliance mechanism. It is also true that many compliance mechanisms – whether provided for in primary law or in secondary law – do not only allow for the adoption of soft law, but also of hard law acts. However, we have to acknowledge that the large majority of compliance mechanisms – of which only a small sample could be presented in Part IV above – is provided for in secondary law, with only few of them concretising a procedure which is already sketched out in the Treaties. Furthermore, it is apparent that compliance mechanisms frequently provide for the adoption of soft law on the part of the EU bodies in charge, sometimes exclusively, sometimes combined with the possibility to adopt hard law acts.

Against this background, it does not appear far-fetched to assume that soft law and the power to adopt it – whose inconspicuousness and result‐

<sup>2191</sup> See eg Deutscher Bundesrat, Empfehlungen der Ausschüsse, Drucksache 655/1/17, 1f <https://www.umwelt-online.de/PDFBR/2017/0655\_2D1\_2D17.pdf> accessed 28 March 2023; see also Commission Communication 'Welcoming Foreign Direct Investment while Protecting Essential Interests', COM(2017) 494 final, 9: 'The Commission fully acknowledges the need to maintain the necessary flexibility for Member States to screen foreign direct investments'. For the EU regime on foreign subsidies distorting the internal market which provides for a much stronger em‐ powerment of the Commission see Regulation 2022/2560.

ing capability of facilitating competence creep have been pointed out in this work<sup>2192</sup> – are important tools in bringing about the institutional shift described above comparatively smoothly, with little attention being drawn to it and thereby raising comparatively little (public) controversy. Chronologically speaking, soft law stands only at the end of this process. At the beginning, there is the legislator (and, respectively, the Commission) vesting existing or newly established EU bodies with soft law powers (and, respectively, proposing legislation to this effect). Therefore, at the outset, there is a legislative procedure. On this basis, the EU bodies in charge make use of their soft law powers, which – due to the legal non-bindingness of soft law and the limited possibilities of judicial review – largely exist and function beyond the attention of a broader public. Hence what appears as 'the calibration of different instruments and actors to deliver effective and legitimate forms of governance'2193 by the legislator may turn out to be strategic action, partly with good intentions and the aim to increase the effectiveness of Union law in the MS, aimed at ousting the Treaty infringement procedure.<sup>2194</sup>

While the role of soft law in transforming the institutional setting under‐ lying the Union regime of ensuring MS' compliance with EU law will be mapped out in particular in Chapter 3 below, it is intended here to raise awareness that this alleged institutional change may not be an altogether incidental development, but that it may actually be promoted by the Com‐ mission and the legislator, thereby using soft law and the powers to adopt it purposively as a tool to achieve this objective.

# 2.5. The deviation from the Treaty infringement procedure

### 2.5.1. The ubiquity of the Treaty infringement procedure

The specialty of the Treaty infringement procedure (including its variants) as the general compliance mechanism is its involvement of the CJEU, the institution with the highest authority in matters of EU law, more precisely in matters concerning the interpretation of EU law and the validity of

<sup>2192</sup> See for example III.5.2.2. above.

<sup>2193</sup> Armstrong, Character 214.

<sup>2194</sup> For the 'disdain for courts' as a characteristic of new governance more generally see Dawson, Waves 211.

secondary law. While its application is exceptionally excluded – eg pursuant to Article 126 TFEU – and while MS' compliance with EU law by national acts which have not yet been taken (draft acts) or with legally non-binding acts of EU law cannot be enforced via the Treaty infringement procedure, it may be initiated as a follow-up to most of the procedures addressed above – after all, nearly all of our selection of compliance mechanisms, also the soft ones, are about compliance with (hard) law. Understood that way, these compliance mechanisms are (potentially) upstream to a Treaty infringement procedure. They take place against the background of Article 258 TFEU (or its variants just mentioned), in particular, but not only where they are conducted by the Commission.<sup>2195</sup> Where a special compliance mechanism does not meet its respective primary *telos* (that is to settle the matter pursuant to Union law), the Commission regularly may initiate a Treaty infringement procedure. With a complete Treaty infringement pro‐ cedure – that is to say: including a CJEU judgement (and possibly a second one pursuant to Article 260 para 2 TFEU) – the concrete case is settled for good. Under EU law, there is no further legal instance to turn to. Not only may the Treaty infringement procedure be used to reinforce or back-up alternative compliance mechanisms – either as a Sword of Damocles whilst the alternative compliance mechanism is applied, or as a follow-up to the (unsuccessful) application of an alternative compliance mechanism.2196 In turn, also alternative compliance mechanisms may be used (ie menaced to be applied) in order to reinforce a Treaty infringement procedure already initiated – in particular with a view to settling the latter still in its adminis‐ trative phase. In a case related to fisheries policy, the Commission adopted a letter of formal notice (thereby initiating the Article 258 TFEU-procedure), suggesting that in case of non-compliance therewith it would adopt – in the course of a special compliance mechanism – preventive measures2197 to protect the threatened fish stock.<sup>2198</sup>

<sup>2195</sup> See, with regard to the procedure laid down in Article 7a of Directive 2002/21/EC, the predecessor provision of Article 33 of Directive 2018/1972, Alberti, Actors 39 f; Tobisch, Telekommunikationsregulierung 99.

<sup>2196</sup> One example for this are the ECB's opinions on national draft laws; see eg ECB Opinion CON/2019/20 on judicial relief granted to former holders of qualified bank credit, para 2.1.3., in which the ECB explicitly refers to its powers under Article 271 TFEU; see also, for a different policy field, Ştefan/Petri, Review 544, with regard to a soft compliance mechanism involving the ACER.

<sup>2197</sup> See Article 26 para 3 of Council Regulation 2371/2002 (now repealed by Regula‐ tion 1380/2013).

### 2.5.2. Time of intervention, discretion, and confidentiality

A comparison of the Treaty infringement procedure and its variants and other compliance mechanisms in terms of the time of intervention shows that the majority of them envisage a *post factum* monitoring of compli‐ ance.<sup>2199</sup> Only sometimes are MS obliged to submit draft measures for *ex ante* scrutiny by EU bodies (eg Article 6 paras 5–7 of Regulation 2019/942 or Article 53 of Directive 2019/944).<sup>2200</sup> The Commission has considered such scrutiny in advance an important tool to prevent future Treaty in‐ fringement procedures.<sup>2201</sup> Apart from that, the Commission – with regard to a similar compliance mechanism – mentioned in particular the leeway granted to MS, the possibility for the MS to check whether or not their respective drafts are in accordance with EU law, the dialogue between the Commission and the MS which is thereby facilitated,2202 and the respect for the subsidiarity principle.<sup>2203</sup>

As regards the amount of power, it is apparent that the Treaty infringe‐ ment procedure – as an example for 'police-patrol' supervision: moni‐ toring of compliance, remedying of violations, and discouraging of further breaches2204 – allows for a much wider discretion for the EU body in charge. This is particularly well documented with regard to the Commis‐ sion's power under Article 258 TFEU. The wide margin of its discretion,

<sup>2198</sup> See Commission, Press Release IP/03/1534; see also Andersen, Enforcement 189 f; for the use of the Treaty infringement procedure as a means to steer MS' behaviour in a thematically unrelated field see III.4.3.2.1. above.

<sup>2199</sup> This question – *ex ante* or *ex post* intervention – will be revived below in the con‐ text of distinguishing between implementation and enforcement (see 3.1.1.2.1.1.).

<sup>2200</sup> For a similar procedure see Article 11 para 4 of Council Regulation 1/2003, accord‐ ing to which national authorities have to present the 'envisaged decision or, in the absence thereof, any other document indicating the proposed course of action'; see also Commission Notice on cooperation within the Network of Competition Authorities, 2004/C 101/03, para 46; for the *ex ante* scrutiny of national provisions under the Services Directive (2006/123/EC) see eg its Article 15 para 7.

<sup>2201</sup> See already Commission, 'Better monitoring of the application of Community law', COM(2002) 725 final/4, 4 ff.

<sup>2202</sup> For the advantages of such dialogue for the EU see Commission, European Gover‐ nance – A White Paper, COM(2001) 428 final, 25; stressing the dialogical function of multiphase procedures more generally: case T-317/09 *Concord*, para 50.

<sup>2203</sup> See eg Commission, 'The Operation of Directive 98/34 in 2009 and 2010', COM/ 2011/853 final, 10.

<sup>2204</sup> Tallberg, Paths 615.

which has been confirmed by the Court on numerous occasions,<sup>2205</sup> has emerged by default rather than by design, in other words: It have originally been other reasons – initially concerns about the prestige of the MS,<sup>2206</sup> later on the Commission's limited (personnel) capacities<sup>2207</sup> – leading to the Commission's *selection* of the cases it pursues, not an explicit indica‐ tion in law. On the other hand, the Commission's discretion also reflects its role as a 'semi-political institution',<sup>2208</sup> a role which does follow from the law, that is from the Treaties in a systematic interpretation.<sup>2209</sup> The Commission's discretion includes the power to seek friendly solutions not only in the pre-litigation phase of the Treaty infringement procedure (when it is a matter of course<sup>2210</sup>), but also before that, in what was called a '"pre-pre-litigation" phase'<sup>2211</sup> (see IV.2.1.2. above). In the meantime, the Commission has disclosed in more detail how it selects its cases, how it 'prioritises' that is to say.2212 In the other compliance mechanisms, the discretion of the EU body in charge is, on a whole, more limited, especially where it is not an institution of the EU and hence, most of the time at least, its powers are not made explicit in primary law.2213 This more limited room for manoeuvre is particularly visible in cases where the competent EU

<sup>2205</sup> See eg joined cases T-479/93 and T-559/93 *Bernardi*, para 31. This discretion has also been acknowledged by the European Ombudsman in its own-initiative inqui‐ ry into the Commission's administrative procedures for dealing with complaints concerning Member States' infringement of Community law (1997); see, with further references to the case law, Prete, Infringement 39–41.

<sup>2206</sup> See Opinion of AG *Roemer* in case 7/71 *Commission v France*, 1026.

<sup>2207</sup> See Börzel/Hofmann/Panke/Sprungk, Member States 1374.

<sup>2208</sup> Andersen, Enforcement 69.

<sup>2209</sup> For early criticism see Audretsch, Supervision 199: 'It is difficult, and in principle even not quite proper, to combine a political function with that of an independent supervisor. Conflicting interests are then at stake'.

<sup>2210</sup> See already the early case 74/82 *Commission v Ireland*, para 13.

<sup>2211</sup> Andersen, Enforcement 47.

<sup>2212</sup> See Commission, 'Better Monitoring of the Application of Community Law' (Communication), COM/2002/725, 11f; more recently: Commission, 'EU law: Better results through better application' (Communication), 2017/C 18/02, with reference to further acts which are relevant in the given context.

<sup>2213</sup> The freedom to initiate or not to initiate a compliance mechanism amounts to a wide latitude which is not only procedural in nature. After all, it allows for the actor concerned – to some extent at least – to decide whether or not to 'permit' a (suspected) infringement of EU law.

actor cannot initiate the respective proceedings *sua sponte*, but may act only upon request, recommendation, etc by another actor.<sup>2214</sup>

With regard to confidentiality, it is to be noted that the output adopted in the course of a Treaty infringement procedure is regularly confidential up until the Court has launched its judgement (see IV.2.1.2. above),<sup>2215</sup> in accordance with Article 4 para 2 (third indent) of Regulation 1049/2001, stipulating that institutions 'shall refuse access to a document where dis‐ closure would undermine the protection of the purpose of inspections, investigations and audits, unless there is an overriding public interest in disclosure'.<sup>2216</sup> The reason for this is that disclosure of case material would thwart one main objective of the Treaty infringement procedure, that is to find a friendly settlement of the case. This applies in particular to the pre-litigation procedure. But since an amicable solution may be reached be‐ tween the Commission and the MS even after the Commission has filed an action with the CJEU, confidentiality extends to the litigation procedure.<sup>2217</sup> As soon as the Court has handed down a judgement under Article 258 TFEU, there cannot be confidentiality on the case anymore – not even with regard to a procedure according to Article 260 TFEU (should it be initiated later). This is because the purpose of the latter is not to find a friendly settlement.2218 Also the output adopted in the course of other compliance mechanisms is regularly confidential in accordance with Article 4 para 2

<sup>2214</sup> See eg Article 144 TFEU (Council acts upon Commission recommendation); Article 53 (Commission acts upon request by a national regulatory authority) and Article 63 of Directive 2019/944 (ACER acts upon request from the Commission or a national authority).

<sup>2215</sup> The mere fact that a Treaty infringement procedure is on-going, however, is not confidential. The Commission regularly informs the public of the steps taken during a Treaty infringement procedure via press releases; see eg <http://europa.e u/rapid/press-release\_IP-96-1239\_en.htm> accessed 28 March 2023.

<sup>2216</sup> See also case T-36/04 *API*, para 132 f, with further references. A different view – ap‐ plying a historical interpretation of this provision – is proposed by Krämer, Access 201f; for a complaint on how the Commission dealt with three requests for public access to documents concerning EU pilot and infringement procedures submitted to the European Ombudsman and for her appraisal see case 383/2022/NK.

<sup>2217</sup> See case T-191/99 *Petrie*, para 68. In case of a friendly settlement, the parties have to inform the Court of the abandonment of their claims; see Article 147 para 1 of the Rules of Procedure of the Court of Justice.

<sup>2218</sup> See, with regard to the CJEU's case law and with arguments in favour of maintain‐ ing this confidentiality policy, European Parliament, 'The relationship between the Commission acting as guardian of the EU Treaties and complainants: Selected topics' (Note, 2012) 10.

(third indent) of Regulation 1049/2001. This is appropriate, as also these compliance mechanisms principally aim at finding a friendly solution of the matter at issue. That confidentiality is the rule we can also – namely *e contrario* – conclude from the exceptional power of EU actors to publish certain acts (see eg Article 53 of Directive 2019/944 or Articles 16 f of Regu‐ lation 1092/2010). As a negative consequence of this confidentiality, it was brought forward that the respective cases cannot become 'comprehensive learning opportunities' for other MS.<sup>2219</sup>

# 2.5.3. Efficiency concerns and EU Pilot

Another point of difference between the Treaty infringement procedure and the other (hard and mixed) compliance mechanisms of our sample is the fact that in the former only the Court can state the unlawfulness of a MS action/inaction with legally binding effect, not the Commission (or the EIB or the ECB).2220 In the hard and mixed compliance procedures addressed above, the legally binding acts are taken by functionally administrative EU actors. In light of the amount of time the CJEU normally requires for rendering a judgement (in addition to the regularly extended pre-litigation phase of the procedure),2221 hard and mixed compliance procedures may be faster – and to that extent: more efficient – in providing for a legally binding account of the matter.2222 Only where the matter is brought before the Court, either via a Treaty infringement procedure or following an action for annulment submitted in due time, the latter normally being filed

<sup>2219</sup> Andersen, Enforcement 204 f, pointing at expert groups dealing with the imple‐ mentation of directives as another example of a compliance tool which may very well create these learning opportunities.

<sup>2220</sup> A fact that has given cause for criticism by the Commission; see Andersen, En‐ forcement 124 ff, with further references. The High Authority (and after 1967: the Commission) was granted such power pursuant to Article 88 TECSC.

<sup>2221</sup> See Bobek, Court 9 f.

<sup>2222</sup> The time it takes until a final decision is made in a procedure depends on the time granted to the EU actors involved to render their respective output, and the time to react allowed for the addressees (mostly: the MS). These time frames, if they are made explicit in the Treaties or in secondary legislation at all, vary significantly: see eg Article 126 TFEU (Council shall act 'without undue delay'; MS shall react 'within a given period'), Article 63 of Directive 2019/944 (ACER shall act within three months upon request; MS shall comply within four months), Article 17 of Regulation 1093/2010 (EBA shall act within two months after initiating the investigation; MS shall react within ten working days).

by the MS addressed, the Court has the possibility to provide the ultimate legal solution of the legal question(s) underlying the case.

The long duration of the pre-litigation phase and of CJEU decision-mak‐ ing is one aspect of said inefficiency of the Treaty infringement procedure, which may have lead to the multiple provision of upstream compliance mechanisms as a means to provide alternative (concurring) routes towards compliance.<sup>2223</sup> The comparative lack of efficiency of the Treaty infringe‐ ment procedure does not only root in its procedure, though, but also lies in the fact that the Commission neither has the power nor the capacity to investigate compliance within the MS' territories on its own.<sup>2224</sup> Rather, the Commission hinges on MS' cooperation and thus the procedure 'depends to a large degree on deliberation and knowledge-creation'.<sup>2225</sup> Already more than 20 years ago, *Gil Ibáñez* held that 'the Article 226 [now: Article 258 TFEU] procedure is no longer appropriate for a variety of infringements of EC law committed by Member States'.<sup>2226</sup>

Compliance mechanisms laid down in specific policy areas, on the con‐ trary, provide for a more intense monitoring of the MS in the respective field brought about also by closer contact with the relevant national author‐ ities,2227 whose representatives may – as is normally the case with European agencies2228 – even take part in the decision-making of the EU body in

<sup>2223</sup> See case C-359/92 *Germany v Council*, paras 46–50, appearing to acknowledge the efficiency argument in favour of secondary law compliance mechanisms; see also Gil Ibáñez, Tools 3.2.3.; Andersen, Enforcement 171, with regard to the fact that the Treaty infringement procedure and many other compliance mechanisms only address individual cases, but not 'clusters of compliance failures that apply to a large number of member states'; for the discontent this drawback of the Treaty infringement procedure has resulted in with regard to infringements of the right of establishment and the free movement of services see Commission Proposal for a Directive of the European Parliament and of the Council on services in the internal market, COM(2004) 2 final/3, 18: 'ineffective' and 'unmanageable' to address these infringements individually via the Treaty infringement procedure.

<sup>2224</sup> See European Parliament, 'The relationship between the Commission acting as Guardian of the EU Treaties and complainants: Selected topics' (Note, 2012) 10.

<sup>2225</sup> See Commission, 'Single Market Scoreboard. Performance per governance tool: Infringements (Reporting period: 12/2015–12/2016)' 13.

<sup>2226</sup> Gil Ibáñez, Exceptions 169.

<sup>2227</sup> See S Augsberg, Verwaltungsorganisationsrecht, paras 55–59.

<sup>2228</sup> For the Boards of European agencies in which exceptionally not all MS (or: not all the relevant national bodies) are (or historically were) directly represented see Chamon, Agencies 66 f.

charge.2229 While the Treaty infringement procedure pursuant to Article 258 TFEU addresses a MS as such, but is not preoccupied with the question which body within the State is responsible for the wrongdoing at issue (a federal, a provincial or a municipal body, a legislative, administrative or judicial body, a body bound by instructions or an independent body, etc), special compliance mechanisms often provide for direct interaction between the EU body in charge and the national authority responsible for the wrongdoing. This makes these 'alternative means of problem solving […] often more effective, quicker, and less expensive' and hence facilitates their introduction and application in order to render redundant 'systematic recourse to infringement procedures'.<sup>2230</sup> These efficiency gains may, to some extent at least, compensate for decreased (democratic) input-legitima‐ cy, eg where compliance mechanisms are governed by independent agen‐ cies instead of the Commission.<sup>2231</sup>

Over the past 20 years, the number of reasoned opinions pursuant to Article 258 para 1 TFEU launched by the Commission has considerably decreased from 533 in the year 2003 to 104 in the year 2022.2232 Also the number of complete performances of the procedure laid down in Article 258 TFEU – from formal notice to CJEU judgement – has become lower and lower, and still the average duration of a case with 112 weeks in 2022 is remarkably long.2233 What is more, even after the Court has launched its judgement in a case, the time until a MS complies can be considerable (and has been increasing significantly in the past few years), as the Commission's statistics indicate.<sup>2234</sup>

<sup>2229</sup> See Commission, 'A Europa of Results – Applying Community Law' (Communica‐ tion), COM(2007) 502 final, 3; see also Commission Proposal COM(2016) 863 final, 16 f, in which the Commission considers a deviation from this route with regard to the ACER, but eventually discards it; for the positive effects on MS' compliance with soft law this composition may have see III.4.4. above.

<sup>2230</sup> Commission, Press Release 'Internal Market: Commission presents ten-point plan for making Europe better off' (2003), IP/03/645.

<sup>2231</sup> See Orator, Möglichkeiten 350.

<sup>2232</sup> Commission, Report on monitoring the application of Community law 2003, COM(2004) 839 final, 4; number for the year 2022 taken from the Commission's online-database <https://ec.europa.eu/atwork/applying-eu-law/infringements-p roceedings/infringement\_decisions/> accessed 28 March 2023; for the earlier development of related figures since 1995 see Börzel/Knoll, Non-compliance 10.

<sup>2233 &</sup>lt;https://commission.europa.eu/law/application-eu-law/implementing-eu-law/i nfringement-procedure/2022-annual-report-monitoring-application-eu-law\_en> accessed 28 July 2023.

In order to tackle these efficiency concerns, the Commission – for the time prior to the initiation of a Treaty infringement procedure – has first introduced its EU Pilot in 2008. It is a confidential online database for communication between Commission services and MS authorities. Via this channel, MS authorities are asked to answer questions with regard to compliance issues, ie a potential (future) Treaty infringement case. The MS have ten weeks to provide answers to the questions, upon submission of which the Commission renders its comments – again within a period of ten weeks. Matters may be solved at this early stage without a Treaty infringement procedure being initiated in the first place.

The Commission deems the introduction of EU Pilot responsible for the decrease in the number of infringement proceedings.<sup>2235</sup> In 2016, 790 pilot cases were opened, 875 processed, and by the end of 2016, 1,175 pilot cases (including the backlog from preceding years) were still open. 72 percent of the closed pilot cases were closed due to a satisfactory answer on the part of the MS concerned.<sup>2236</sup> 233 Treaty infringement procedures were opened in 2016, following closure of EU Pilot cases.2237 In the years since 2017 the number of handled EU Pilot processes has gone down significantly (with again an increase starting in 2021).2238 In view of the depoliticised, more technical approach of EU Pilot, it could – even though in the discussion it is often linked to the procedure laid down in Article 258 TFEU – functionally be perceived as a special compliance mechanism. However, it also differs from the special compliance mechanisms presented and discussed above, as it is not about a suspected infringement, but it is about less: unclarities which, if not satisfactorily resolved, may give rise to infringement proceedings.

The legal quality of the Commission's assessment of the response of the MS appears to range below that of soft law. This is also due to the fact

<sup>2234</sup> See Commission, 'Single Market Scoreboard. Performance per governance tool: Infringements (Reporting period: 12/2017–12/2018)' 8 f.

<sup>2235</sup> See Commission, '28th Annual Report on Monitoring the Application of EU Law', COM(2011) 588 final, 5; for further explanations of the low number of Treaty in‐ fringement procedures, in particular in the rule of law context, see Scheppele/Ko‐ chenov/Grabowska-Moroz, Values 59-63.

<sup>2236</sup> See Commission, 'Monitoring the application of European Union law. 2016 Annual Report', COM(2017) 370 final, 21.

<sup>2237</sup> See Commission, 'Monitoring the application of European Union law. 2016 Annual Report', COM(2017) 370 final, 21.

<sup>2238</sup> See Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 18 f.

that EU Pilot preceeds a procedure – the Treaty infringement procedure – which itself begins with an only informal exchange of views which moves below the level of soft law. If at all, and leaving apart its attachment to the Treaty infringement procedure, EU Pilot can be classified as below-soft law mechanism. Due to the extended period of time EU Pilot requires (potentially in addition to the lengthy Treaty infringement procedure), the Commission late in 2016 has decided to stop 'systematically relying on the EU Pilot problem-solving mechanism' and to apply it only when it 'is seen as useful in a given case'.<sup>2239</sup> At least in the short term, the Commission ar‐ gues, this has resulted in an increase of Treaty infringement procedures.<sup>2240</sup> The *von der Leyen* Commission seems to be more convinced of the merits of EU Pilot (it 'has proven its value', the Commission now stresses), but the numbers are still comparatively low.<sup>2241</sup>

As concerns the special compliance mechanisms, no comprehensive data on their respective efficiency exists. However, against the backdrop of the efficiency concerns related to the Treaty infringement procedure, from the mere fact that special compliance mechanisms have been set in place we can at least deduce that the Commission and the legislator have assumed that their operation will lead to increased compliance rates in the respective policy fields (see also 2.6. below).

# 2.5.4. The MS' right to be heard

As regards the right to be heard (as one important component of the rights of defence) of the MS concerned, with regard to the Treaty infringe‐ ment procedure the Court held: '[T]he opportunity for the Member State concerned to submit its observations constitutes an essential guarantee required by the Treaty and, even if the Member State does not consider it necessary to avail itself thereof, observance of that guarantee is an essential formal requirement'.2242 The Court qualifies 'respect for the rights of the

<sup>2239</sup> Commission, 'Single Market Scoreboard. Performance per governance tool: In‐ fringements (Reporting period: 12/2016–12/2017)' 10 f.

<sup>2240</sup> See Commission, 'Single Market Scoreboard. Performance per governance tool: Infringements (Reporting period: 12/2017–12/2018)' 10.

<sup>2241</sup> Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 18 f.

<sup>2242</sup> Case 211/81 *Commission v Denmark*, para 9; see also case C-525/12 *Commission v Germany*, para 21: 'the purpose of the pre-litigation procedure is to give the Member State concerned an opportunity, on the one hand, to comply with its

defence' as a general principle of Union law which must be observed 'even in the absence of express provisions'.<sup>2243</sup> Against the background of proceedings pursuant to what is now Article 106 TFEU, the Court held that a MS 'must receive, before the decision which will be notified to it is adopted […], an exact and complete statement of the objections which the Commission intends to raise against it'.<sup>2244</sup> More recently, it held – again (also) with a view to the EU-MS relationship and partly repeating its earlier case law – that 'observance of the rights of the defence […] requires that the person against whom […] proceedings have been initiated should be placed in a position in which he may effectively *make known his views on the facts and the infringement of EU law that are raised against him before a decision appreciably affecting his interests is adopted*' (emphasis added).<sup>2245</sup>

In most of the mechanisms examined here the right to be heard is explic‐ itly provided for. In the Treaty infringement procedure, its MS-friendly prelitigation phase as laid down in Article 258 TFEU may, in the words of *Gil Ibáñez*, 'provoke an excessive and unjustified slowness in the procedure'.<sup>2246</sup> In the other mechanisms the respective provisions are, on a whole, less elaborate. Sometimes even no mention is made at all of possibilities for MS to utter their view on a case.2247 That does not mean, however, that prior to the adoption of the act at issue no communication takes place between the EU actor and the MS. In general, and in accordance with the above case law, a preliminary exchange of view in a relatively informal communication between the administrator and the *administré* is the rule rather than the exception.<sup>2248</sup> One of the exceptions may constitute the procedures initiated

obligations under EU law and, on the other, to avail itself of its right to defend itself properly against the objections formulated by the Commission'.

<sup>2243</sup> Joined cases C-48/90 and C-66/90 *Netherlands v Commission*, para 37, with a further reference; case T-510/17 *Del Valle Ruíz*, para 121, with further references. Where the rights of defence are granted, the EU actor in charge is allowed to gather information about a MS in a certain case even before the official decision to open an investigation has been taken; case C‑521/15 *Spain v Council*, para 62.

<sup>2244</sup> Joined cases C-48/90 and C-66/90 *Netherlands v Commission*, para 45; see also Hofmann/Rowe/Türk, Administrative Law 210 f.

<sup>2245</sup> Case C‑521/15 *Spain v Council*, para 61.

<sup>2246</sup> Gil Ibáñez, Supervision 97.

<sup>2247</sup> See eg Article 106 para 3, Articles 116 f TFEU, Articles 16 f of Regulation 1092/2010. For the problems compliance mechanisms may raise in terms of the right to be heard of the individuals concerned – a topic which shall not be elaborated on in this context – see Gundel, Energieverwaltungsrecht, para 37, with further references.

<sup>2248</sup> See Harlow/Rawlings, Process 60.

upon request of the MS concerned. Here the MS can utter its view in the request. Hence with regard to the mechanisms now laid down in Article 114 paras 4 f TFEU, for example, the Court has held that beyond that the Commission 'is not required to observe the right to be heard before taking a decision'.<sup>2249</sup>

In mixed mechanisms, the soft law act preceding the hard law act may serve as a tool to inform the MS concerned of the allegations made against it and to invite it to utter its view on them. After that, and only if necessary, the EU actor may adopt the hard law act provided for in the procedure. Whether in the course of soft mechanisms – or in mixed mechanisms in which, due to compliance on the part of the MS concerned in a concrete case, only a soft but no hard law act is adopted – MS have a right to be heard before the adoption of the soft law act is unclear.<sup>2250</sup> After all, it is uncertain whether an EU soft law act may 'appreciably affect[] [a MS's] interests'.2251 This threshold is certainly lower than the 'legal effects *vis-à-vis* third parties' required for an act to be subject to judicial review pursuant to Article 263 TFEU, and hence principally also soft law acts may meet it. In view of the wide range of soft law acts – with varying contents, political authority and consequences (of non-compliance, eg publication of the soft law act or of the fact of non-compliance) – the examination whether or not MS' interests are thereby appreciably affected, in my opinion, unlike with individual-concrete legally binding acts, is to be made case by case.

2.6. Why the compliance mechanisms 'look the way they look'

While one of the main reasons for deviation from the Treaty infringement procedure as the general compliance mechanism – the lack of efficiency of the latter – has been addressed above (2.5.3.), still the question remains unanswered why the alternative compliance mechanisms show such a great variety, particularly in terms of the output provided for, of the procedure to be followed, of the actors involved. For the various compliance regimes

<sup>2249</sup> Joined cases C-439/05P and C-454/05P *Land Oberösterreich*, para 43; see also para 38 of this case and case C-3/00 *Denmark v Commission*, para 50.

<sup>2250</sup> In the context of mixed mechanisms it is to be noted that it can never be predicted whether or not it will extend beyond the adoption of a soft law act in a concrete case, which is why the right to be heard would *always* have to be granted in advance.

<sup>2251</sup> See again case C-521/15 *Spain v Council*, para 61.

laid down in primary law some explanations can be found. A number of these mechanisms are just variants of the Treaty infringement procedure laid down in Article 258 TFEU. The adaptations (to the procedure as laid down in Article 258 TFEU) are minor and can be explained by the requirements of the respective procedures. In the context of the procedure which is now laid down in Article 108 TFEU, the Court held that State aid 'raises problems which presuppose the examination and appraisal of economic facts and conditions which may be both complex and liable to change rapidly',<sup>2252</sup> suggesting that in these cases constant monitoring and, if need be, a fast reaction may be more important than in cases of other infringements of EU law.

The excessive deficit procedure, to take an example strongly deviating from the Treaty infringement procedure, looks the way it looks because the MS – in view of the highly sensitive policy field, public spending being an epitome of national sovereignty2253 – consciously opted for a 'political' procedure with 'little automaticity',2254 a strong Council next to the Commission and no role for the CJEU that goes beyond Articles 263 and 265 TFEU.2255 The powerful tools the procedure provides (eg sanctions) are mitigated by its strong intergovernmental set-up. The soft procedure laid down in Article 121 TFEU better reflects the weak, only coordinating competence of the EU in the field of economic policy.2256 Also in the field of employment policy the soft mechanism laid down in Article 148 para 4 TFEU reflects the very limited EU competence.2257 Article 106 para 3 TFEU serves to complement in particular EU competition law and to avoid its circumvention by the MS with regard to public or privileged undertakings. In this procedure the Commission does not only have a con‐ trol function but also a regulatory function in the politically very sensitive area of public/privileged undertakings.2258 Its acts may concern only one

2256 See Article 5 para 1 TFEU.

2258 See Wernicke, Art. 106, paras 79 and 81.

<sup>2252</sup> Case C-301/87 *France v Commission*, paras 15 f; sceptically: Gil Ibáñez, Supervi‐ sion 102.

<sup>2253</sup> See references in Hamer, Art. 126 AEUV, para 5.

<sup>2254</sup> Fratianni/von Hagen/Waller, Maastricht 45. For the historical development of this 'gentle' approach in economic policy see Braams, Koordinierung 17 f.

<sup>2255</sup> For the exclusion of the application of Articles 258 f TFEU see Article 126 para 10 TFEU. This exclusion also encompasses the imposition of sanctions according to Article 260 para 2 TFEU; see Amtenbrink/Repasi, Compliance 162, with a further reference.

<sup>2257</sup> See Article 5 para 2 TFEU; see also Hemmann, Artikel 148 AEUV, para 2.

MS, but may also have a broader scope. Where a MS fails to comply, a Treaty infringement procedure may follow.<sup>2259</sup> Article 144 TFEU addresses an emergency situation in which the deviation from EU law on the part of the MS is exceptionally allowed. It lays down a pronounced 'political' mechanism with the Council as the central EU actor. Articles 116 f TFEU were intended to provide for 'Krisenmanagement' [crisis management] in a case in which a principally lawful MS action causes a serious distortion of the conditions of competitition in the internal market.<sup>2260</sup> It is primarily about finding an economically viable solution (as required by the law), not about establishing the non-compliance of a MS with primary law. Therefore the Treaty infringement procedure may have appeared to be inapt in this case.

In summary, we can say that – while they all serve the purpose of achiev‐ ing MS compliance – the above mechanisms are nuanced in one or the oth‐ er way. This shading, as it were, is to be understood against the concurrence of the inter-dependent factors of the category of EU competence at issue, the more detailed purpose of the concrete mechanism (implementation, enforcement,2261 exchange of views, monitoring, etc, for which 'ensuring compliance' is only an umbrella term) and the actual course of political negotiations in the respective Treaty-making Convention (which is very difficult to trace2262).<sup>2263</sup>

Also with regard to the design of the mechanisms laid down in secondary law, the majority of which is about day-to-day, rather 'technical' administra‐ tion, there is hardly any uniformity; in *Andersen's* words: 'When viewed as a whole, the measures do not form a coherent picture in the sense of a standardised or formalised procedure established with the purpose of supplementing the general EU infringement procedure'.2264 The question may be raised why they were necessary in the first place and why they vary so strongly from each other, why there is no common model of compliance mechanisms in secondary law.

<sup>2259</sup> See also Koops, Compliance 140, describing Article 106 TFEU against this back‐ ground as 'an extra phase in ensuring compliance'.

<sup>2260</sup> Classen, Art. 116 AEUV, paras 2 and 5.

<sup>2261</sup> These two purposes are addressed in more depth under 3.1. below.

<sup>2262</sup> See eg Thomas Müller, Wettbewerb 23 f.

<sup>2263</sup> Similarly with regard to compliance mechanisms laid down in public international law: Shelton, Compliance 120 ff.

<sup>2264</sup> Andersen, Enforcement 201; similarly: Gil Ibáñez, Exceptions 174.

Also for the compliance mechanisms laid down in secondary law, the need for fast-track<sup>2265</sup> procedures – or at least: procedures faster than the Treaty infringement procedure – has been a decisive factor for their crea‐ tion.<sup>2266</sup> What is more, the activation of these mechanisms rarely attracts the attention of a wider public, even if exceptionally the output may be published. They are regularly down-to-earth procedures not entailing dip‐ lomatic consternation, but – ideally at least – an exchange of (legal) views based on facts. Since the compliance mechanisms are primarily an expres‐ sion of 'prozedurale Kooperation' [procedural cooperation],<sup>2267</sup> a certain degree of mutual trust is necessary to ensure their smooth functioning. Especially with regard to more technical questions, public clamour and national shame are rarely helpful to facilitate cooperation.<sup>2268</sup>

Another reason for special procedures has been the need to involve additional expertise. While the Treaty infringement procedure does not, as a rule,2269 provide for the involvement of expert bodies, the secondary law mechanisms often make provisions for output from European agencies or other EU expert bodies – not only to (softly) demand compliance, but also to 'stimulate mutual learning processes among national regulatory

<sup>2265</sup> *Storr* uses the term 'fast track procedure' in the context of Article 17 of Regulations 1093–1095/2010 and in comparison to the Treaty infringement procedure; Storr, Agenturen 80.

<sup>2266</sup> Arguably this holds true for all secondary law compliance mechanisms presented here, but in particular it does for urgency or even emergency procedures such as the ones laid down in Article 13 para 1 of Directive 2001/95/EC (health safety), Articles 70 f of Regulation 2018/1139 (safety), Article 29 para 2 of Regulation 806/2014 (resolution of banks), Article 18 of Regulation 1093/2010 (supervision of banks); with regard to the allegedly limited possibilities of the Commission in cases of urgency: Commission, 'Communication on the handling of urgent situations in the context of implementation of Community rules – Follow-up to the Sutherland report', COM(93) 430 final, 40; see also the Report to the EEC Commission by the High Level Group on the Operation of Internal Market, presi‐ ded over by Peter Sutherland (1992; so-called 'Sutherland-Report'), in particular Recommendations 20, 27 and 31.

<sup>2267</sup> For this term as a component of the European *Verwaltungsverbund* see Schmidt-Aßmann, Einleitung 6.

<sup>2268</sup> For the balance between 'management' (in particular coordination, cooperation) and 'enforcement' (in particular sanctions) which ought to be struck in order to ensure compliance see Tallberg, Paths 632 f.

<sup>2269</sup> For the possibility of allowing for expert opinions in Court proceedings see eg Article 25 of the Statute of the CJEU.

authorities'.2270 Part of their expertise is also to be better informed in their respective field, that is to say to know better what goes wrong in which MS. This increases the likelihood of detection of an infringement. In hard and in particular in mixed procedures, these expert bodies often act together with the Commission,<sup>2271</sup> in soft procedures more frequently alone.<sup>2272</sup> That a European agency – that is to say: a body not established by primary law<sup>2273</sup> – is the only actor in a mixed or hard procedure is the case only ex‐ eptionally.<sup>2274</sup> This is in particular due to (actual or politically assumed) le‐ gal or political limits to vesting bodies not established by primary law with executive power, above all the requirement, as most prominently expressed in the so-called *Meroni* doctrine, to maintain the EU's institutional balance provided for in the Treaties (see 3.3.1. below).<sup>2275</sup> Some of the secondary law mechanisms presented here are mere concretisations of procedures laid down in the Treaties.2276 Therefore their respective structure – above all in

<sup>2270</sup> Groenleer/Kaeding/Versluis, Governance 1227; for mutual learning effects of peer reviews see Dawson, Soft Law 15; critically: Harlow/Rawlings, Accountability 7.

<sup>2271</sup> See the mechanisms laid down in Articles 70 f of Regulation 2018/1139, Article 63 of Directive 2019/944, Article 17 of Regulation 1093/2010, Article 25 of Regulation 2016/796.

<sup>2272</sup> See the mechanisms laid down in Article 6 paras 5–7 of Regulation 2019/942, Article 33 of Directive 2018/1972, Articles 16 and 17 of Regulation 1092/2010.

<sup>2273</sup> Some EU expert bodies, such as the European Public Prosecutor's Office or the EDA, are explicitly foreseen in primary law, but actually established they are – like all the other European agencies – by means of secondary law (in case of the European Public Prosecutor's Office: in the framework of enhanced cooperation); see also Chamon, Agencies 136 f.

<sup>2274</sup> See the mechanisms laid down in Article 7 para 4 and Article 29 para 2 of Regula‐ tion 806/2014, Articles 18 and 19 of Regulation 1093/2010; for the ambivalent role the Commission plays in the increasing empowerment of European agencies see Chamon, Agencies 123–126.

<sup>2275</sup> See, *ex multis*, Craig, Administrative Law 168–172; for the revisiting of the *Meroni* doctrine by the so-called *ESMA* case see eg Bergström, System. For the continu‐ ing relevance of *Meroni* also in the political discussion see the example of the Commission proposal on the establishment of the European Monetary Fund, COM(2017) 827 final, 13 f; in the context of the SRB's powers see case T-510/17 *Del Valle Ruíz*, paras 204 ff, dwelling on the legislative history of Regulation 806/2014; for the (ir)relevance of *Meroni* in the context of the EBA and the other financial market supervisory authorities see Annunziata, Remains.

<sup>2276</sup> See in particular Articles 22 f of Council Regulation 2015/1589 (Article 108 TFEU), the excessive imbalance procedure laid down in Regulations 1176/2011 and 1174/2011 (Article 121 TFEU; the latter in conjunction with Article 136 TFEU), Article 3 para 7 of Regulation 472/2013 (Article 121 TFEU, in conjunction with Article 136 TFEU).

terms of the actors involved and the categories of output – is, to some extent at least, preordained.<sup>2277</sup>

As mentioned above (2.5.2.), the Treaty infringement procedure – like most compliance mechanisms of our sample – provides for an *ex post* scrutiny of MS action, whereas some procedures laid down in secondary law (also) provide for the scrutiny of draft decisions of MS authorities, that is to say *ex ante* scrutiny.<sup>2278</sup> Thereby the violation of EU law shall be barred preventively.<sup>2279</sup> This has an influence on the interaction between the two parties (MS and EU), as at this early stage of decision-/rule-making they are necessarily more flexible in searching for a lawful solution in accordance with the interests/views of both of them.

All mechanisms presented here aim at reaching compliance with EU law on the part of the MS, but the approaches taken (or: the more detailed purposes) may differ from each other. In this context, all of the above 'objective' considerations may contribute to the actual shape of a mecha‐ nism. But even in view of such 'objective' factors there is never only one way the compliance procedure could possibly look like. Whether MS may request the initiation of the procedure, whether they have ten working days or three weeks to react to the Commission opinion, but also more fundamental questions, for example whether a mechanism should be mixed or only soft, are eventually – to some degree at least – the result of pure political bargaining (the weighing of the legislator's subjective motivations that is).2280 Or, as *Gil Ibáñez* has put it in this context: '[I]t seems clear that the creation of new procedures does not seem to obey a general strategy of fulfilling new needs demanded by all the areas characterised by certain features. In reality, nor can the financial consequences for the EC budget serve to justify all the far-reaching enforcing and supervising tools granted

<sup>2277</sup> Beyond this predetermination, the actors involved should be selected by the legis‐ lator with a view to the 'political, economic, and social characteristics of the sector at stake'; Scholten/Ottow, Design 91.

<sup>2278</sup> See the mechanisms laid down in Articles 70 f of Regulation 2018/1139, Article 7 para 4 of Regulation 806/2014, Article 25 of Regulation 2016/796, Article 53 of Directive 2019/944 and Article 33 of Directive 2018/1972.

<sup>2279</sup> For such preventive Commission measures more generally see Schmidt/Schmitt von Sydow, Art. 17 EUV, paras 37 ff.

<sup>2280</sup> The Commission Proposal COM(2013) 27 final (Article 21) for what has become Article 25 of Regulation 2016/796, for example, has provided for an ERA *recom‐ mendation* to be adopted. Article 25 of Regulation 2016/796, however, due to the Council's position at first reading in the ordinary legislative procedure, provides for the ERA to adopt an *opinion*.

to the Commission […]. Instead, those tools appear to be more the result of political bargaining within the Council [today we would have to add: 'and of the EP'], and between the latter and the Commission, on a case by case basis'.<sup>2281</sup> This is not necessarily harmful for the outcome and it is certainly not an EU-specific characteristic. It is simply a concomitant of collective decision-making of human beings and of the fact that for most policy problems there is more than just one (reasonable) solution.

# *3. Legal assessment*

3.1. Compliance mechanisms: implementation or enforcement?

3.1.1. The characteristics of implementation and enforcement

3.1.1.1. Introduction

As we have seen, the compliance mechanisms presented and discussed above are strongly inhomogeneous – procedurally, institutionally, and not least substantially. What unites them is in particular their shared overall purpose,2282 namely to ensure compliance with EU law by the MS.<sup>2283</sup> Taking a closer look at this apparently shared purpose, we notice that with regard to the broad objective to ensure compliance with EU law by the MS, the TFEU draws a basic line between the tasks of the Commission and the Council in their function as administrative bodies (let us call this implementation) and the traditional tasks of the CJEU (let us call this en‐ forcement), thereby reflecting upon a material separation of powers within the EU. In substantive terms, it was said, implementation 'concerns putting law into effect',2284 while enforcement (by the EU) is about compelling the

<sup>2281</sup> Gil Ibáñez, Tools 5.3., also with regard to MS' interests in establishing such mecha‐ nisms by means of secondary law.

<sup>2282</sup> For the more specific purposes of the soft law acts used in (some) of these mechan‐ isms see 2.4. above.

<sup>2283</sup> See Article 70 TFEU which provides, without prejudice to Articles 258–260 TFEU, for the possibility to perform an evaluation of the implementation of certain Union policies by MS' authorities.

<sup>2284</sup> Andersen, Enforcement 163; for the term 'implementation' see also Christian‐ sen/Dobbels, Rule Making 44 f.

relevant actors to comply with the law.2285 Enforcement 'only becomes rele‐ vant in the phase succeeding implementation, if the question arises whether EU law has been implemented, applied, and enforced effectively'.<sup>2286</sup> Both fields – implementation and enforcement – are aimed at ensuring compli‐ ance with EU law on the part of the MS.<sup>2287</sup> They approach this aim in a different manner, though.

Ideally, the compliance mechanisms presented and discussed above can be allocated to either of these categories – implementation or enforcement. Concluding from the Treaties, no third category is apparent in this context (*tertium non datur*). The allocation to either category is important for numerous aspects of the legal assessment of compliance mechanisms, such as the correct legal basis or the EU's institutional balance, and will thus reoccur throughout this chapter. It shall therefore be addressed right at its beginning. In spite of the lack of a third category, the above-mentioned material separation of powers underlying the distinction between imple‐ mentation and enforcement does not entail hermetic segregation. Thus, the differentiation between the two categories in practical terms also allows for cooperation between them.

In the specific context at issue here, we talk about procedures performed by EU institutions, bodies, offices and agencies *vis-à-vis* specific MS and their respective authorities. Thus, we have an interest in defining the two terms – implementation and enforcement – only in this individual-concrete relationship. Conversely, we are not concerned with the general-abstract implementation of EU law by EU actors. Neither are we concerned with

<sup>2285</sup> See Nollkaemper, Role 161, with further references. The distinction between coer‐ cion and persuasion does not represent the distinction between enforcement and implementation. It rather addresses the dichotomy of hard and soft approaches. However, as has been shown throughout this work, also soft regulation – especially if combined with a duty to explain non-compliance – may be rather a tool of coercion than a tool of persuasion; for the antipodes coercion and persuasion see also van Waarden, Harmonization 102; for different conceptions of 'enforcement' in EU law see Scholten, Enforcement 9 f.

<sup>2286</sup> Andersen, Enforcement 163.

<sup>2287</sup> For the relationship between the terms 'implementation' and 'compliance' on the one hand, and 'enforcement' and 'compliance' on the other hand see Nollkaemper, Role 160 f; see also IV.1. above. The Commission broadly utters: 'the best way to enforce EU law is to prevent breaches from happening in the first place'; Commis‐ sion, 'Enforcing EU law for a Europe that delivers' (Communication), COM(2022) 518 final, 1.

the implementation of EU law by the MS or the (in practice increasing2288) enforcement of EU law by EU bodies *vis-à-vis* individuals (in a broader sense), eg the transposition of an EU directive by a MS or the enforcement of EU competition law by the Commission *vis-à-vis* an undertaking.

In an attempt to define these two categories more closely in the specific context just recapitulated, we will flesh out the main characteristics of implementation and enforcement as laid down in primary law. On their own, most of these characteristics are not necessary, and none of them is sufficient for the allocation of a mechanism to either of the two categories. Therefore, they will only serve as indicators of the *ratio legis* of the act providing for the mechanism under scrutiny. The multitude of such indica‐ tors will create a flexible system which shall allow a concrete compliance mechanism to be assigned, at least by tendency, to either category.

# 3.1.1.2. Implementation and enforcement under the Treaties


Having provided above for a preliminary, rather general definition of the terms implementation and enforcement, we shall now flesh out the characteristics of these concepts – in our specific context, namely in the individual-concrete relationship between EU actors and MS actors – as laid down in primary law. Even though neither of them is defined in the Treaties, we can deduce some indicators from certain core provisions, in particular Articles 16 f TEU as well as Article 291 TFEU in the context of implementation and Article 19 TEU and Articles 258–260 TFEU in the context of enforcement. In a systematic perspective, however, also other provisions2289 are to be considered, and so is the dynamic case law of the Court. This reveals that Article 291 TFEU is not the only primary law provision which allows for individual-concrete implementing acts to be addressed to the MS (see also 3.1.1.2.1.3. below). Nevertheless, being

<sup>2288</sup> See Scholten, Trend 1349.

<sup>2289</sup> See the 'autonomous executive powers' referred to by Chamon, Member States 1506, eg Articles 42, 43 para 3, 78 para 3 or 103 para 1 TFEU. These powers directly serve the implementation of certain Treaty provisions.

the main general provision on the implementation of EU law, Article 291 TFEU certainly is to be given special weight. As regards Articles 258–260 TFEU, it is to be noted that these provisions establish the Court as the body in charge of determining infringements of EU law by the MS and to ensure compliance by the imposition of financial sanctions, if necessary. Thereby the Treaties have, primary law exceptions (eg the excessive deficit procedure) apart, monopolised the enforcement of EU law *vis-à-vis* the MS with the Court (which is, in the Treaty infringement procedure pursuant to Article 258 TFEU, requested to act by the Commission).

While under the Treaties the primary aim of implementation on the part of EU actors (according to Article 291 para 2 TFEU<sup>2290</sup>) is to create uniform conditions for the implementation (by the MS) of legally binding Union acts, more generally that is to say: to concretise EU law,<sup>2291</sup> the primary aim of enforcement is to determine and to end non-compliance with law, in our case: with EU law on the part of the MS (Treaty infringement procedure). Article 291 para 2 TFEU allows for an EU actor to implement (to concretise) EU law so that it is uniformly implemented (applied) by the MS. In the case of enforcement, compliance shall be reached by the deter‐ mination of an infringement.2292 Only when this infringement is established authoritatively may a change of action – compliance – be requested.

This dichotomy of ensuring compliance with EU law *vis-à-vis* the MS under the Treaties also has a chronological dimension, in that it is reflected upon by the point of time of intervention. Since concretisation (as a means of ensuring compliance) is their main purpose, implementing measures (of EU actors) will regularly be taken before, exceptionally also in the course of the relevant MS action (constituting the application of EU law).<sup>2293</sup> This can be described as the *ex ante* occurrence and the on-going occur‐ rence of implementing acts, the latter meaning that the EU implementing body is involved eg in a national procedure leading to the adoption of national implementing acts. An example is the procedure laid down in

<sup>2290</sup> For the broader understanding of this term prior to the Treaty of Lisbon see Craig, Lisbon 50; for the relevant case law prior to the entry into force of the Treaty of Lisbon see Mendes, Rule Making 31.

<sup>2291</sup> See also case C-427/12 *Commission v European Parliament/Council*, para 39: 'pro‐ viding further detail in relation to the content of a legislative act'.

<sup>2292</sup> See also joined cases C-514/07P, C-528/07P and C-532/07P *API*, para 119: 'Article 226 EC is designed to obtain a declaration that the conduct of a Member State is in breach of EU law and to terminate that conduct'.

<sup>2293</sup> See also case C-359/92 *Germany v Council*, para 47.

Article 121 TFEU (see IV.2.2.3.1.1. above), pursuant to which the national decision-making procedure as a whole is monitored. The former – *ex ante* occurrence – is exemplified by the adoption of general-abstract implement‐ ing acts – the regular case under Article 291 para 2 TFEU – on the basis of which the administrations of the MS take their respective decisions. Also the procedure laid down in Article 114 paras 4 ff TFEU (on the maintenance or introduction of national provisions; see IV.2.2.1.1.3. above) provides for *ex ante* intervention, since it is normally performed prior to the entry into force of the harmonisation measure.<sup>2294</sup>

In practice, it is sometimes difficult to draw a clear line between *ex ante* and on-going occurrence. For our purposes this is not necessary anyway, because both modes of intervention indicate implementation. For the over‐ lap between on-going and *ex post* occurrence see 3.1.1.2.1.2. below.

Enforcement action, on the contrary, is taken *post factum*, that is to say after the relevant MS action (*ex post* occurrence).2295 Before the incrimi‐ nated MS action has been taken, no infringement can be established.<sup>2296</sup> Whereas also *ex post* action, after an assessment of the other criteria, may turn out to qualify as implementation, the reactive character of a measure *in general* indicates enforcement.<sup>2297</sup>

Eventually, one word on the discretion of the legislator in granting imple‐ menting powers: As opposed to the regular case of MS implementation according to Article 291 para 1 TFEU2298 (which normally leaves a certain leeway to the MS and hence may lead to slightly different results in different MS), para 2 makes provision for the case that uniform conditions for the implementation are required by means of implementing acts of the

<sup>2294</sup> See Classen, Art. 114 AEUV, para 224, with further references. Article 114 para 9 TFEU again is addressing enforcement (a variant of the Treaty infringement procedure). This exemplifies the sometimes close entanglement between imple‐ mentation and enforcement.

<sup>2295</sup> See also Gil Ibáñez, Supervision 16, who describes enforcement in EU law as 'reactively related to compliance'.

<sup>2296</sup> Note the words on the Treaty infringement procedure of Scholten, Trend 1353: 'In any case, it is an *ex post* mechanism, a tool of last resort, a stick rather than a carrot' (emphasis in original).

<sup>2297</sup> Also the distinction between on-going and *ex post* occurrence may sometimes turn out to be difficult; see 3.1.1.2.1.2. below.

<sup>2298</sup> See already joined cases 89 and 91/86 *CNTA*, para 11, with a further reference; see also European Convention, 'Delimitation of competence between the European Union and the Member States – Existing system, problems and avenues to be explored', CONV 47/02, in particular 9 f.

Commission (or of the Council). A supplementation or amendment, if only of certain non-essential elements of a legislative act, is excluded from such implementation *qua* Article 290 para 1 TFEU. Even if we applied a wide understanding of the term implementation, Article 290 TFEU would not be relevant in our context, as it does not allow for the adoption of individ‐ ual-concrete acts directed to a MS.<sup>2299</sup> In the Court's words, the legislator has discretion when it decides to confer an implementing power pursuant to Article 291 para 2 TFEU (as opposed to delegated power under Article 290 TFEU), which is why here judicial review is limited to 'manifest errors of assessment as to whether the EU legislature could reasonably have taken the view […] that […] only the addition of further detail, without its nonessential elements having to be amended or supplemented [is required] and, secondly, that [the basic act at issue] require[s] uniform conditions for implementation'.2300 Its wide wording – interpreted also in light of Article 17 TEU which very generally obliges the Commission 'to ensure the application of the Treaties'2301 – in combination with this concessive case law makes Article 291 para 2 TFEU a versatile tool in the hands of the legislator.

As mentioned above, implementing acts adopted by the Commis‐ sion/Council pursuant to Article 291 para 2 – unlike delegated acts pur‐ suant to Article 290 TFEU – may also take the form of individual-concrete decisions.2302 This does not in principle appear to be contrary to the con‐ cept of implementation: While the purpose to create 'uniform conditions for implementing […] Union acts' primarily addresses general-abstract measures (which also in practice are the regular case of implementing

<sup>2299</sup> See Ilgner, Durchführung 227 and 254; Craig, Comitology 176.

<sup>2300</sup> Case C-427/12 *Commission v European Parliament/Council*, para 40; confirmed by case C-88/14 *Commission v European Parliament/Council*, paras 28–32; see also the more nuanced approach of AG *Mengozzi* in this case, in particular paras 30–38; for the scope of (non-)essential elements see Ritleng, Domain.

<sup>2301</sup> Note that the Commission in its Communication 'EU law: Better results through better application', 2017/C 18/02, states that under Article 17 para 1 TEU it is the Commission's responsibility to ensure not only the effective application of EU law, but also its implementation and enforcement (page 1); see also Senden, Soft Law 318.

<sup>2302</sup> See Article 2 para 2 of Regulation 182/2011: 'other implementing acts [than imple‐ menting acts of general scope]'; see case C-146/91 *KYDEP*, para 30; see Ilgner, Durchführung 219; Schütze, Rome 1418; see also Nettesheim, Art. 288 AEUV, para 27, who argues that implementing acts may be adopted in all the forms laid down in Article 288 TFEU, also in the form of recommendations and opinions.

measures), it can be argued that in some cases also the adoption of individ‐ ual-concrete measures addressed to MS serves the uniformity of application of EU law.<sup>2303</sup>

# 3.1.1.2.1.2. Different approaches towards ensuring compliance: concretisation and determination

It has already been stated that both implementation and enforcement are aimed at compliance. It should be noted, however, that whereas enforce‐ ment is aimed at the determination of an infringement, a different approach is inherent in implementation within the meaning of Article 291 para 2 TFEU. Implementing power on the part of the Commission or the Council is primarily directed towards the concretisation of EU law and thereby does away with or at least reduces the MS' leeway in applying a legally binding Union act. By establishing 'uniform conditions', any deviating application of the basic act becomes unlawful, also an application which would under normal circumstances – that is to say without an implementing act adopted by the Commission or the Council – be well within the MS' room for manoeuvre.<sup>2304</sup>

The distinction between implementation and enforcement becomes par‐ ticularly difficult where implementation takes the form of an individualconcrete measure directed to a MS *ex post* or in the course of an on-going procedure2305 – two forms which, in places, overlap. In these cases, both form and time of intervention at least resemble those of enforcement measures. That by means of implementing measures not only the slight deviations (which, for lack of the concrete implementing act, would be lawful) but also – *a fortiori* – the severe breaches of law may be tackled, is a collateral effect – an effect which increases the difficulty to distinguish between implementation and enforcement in this respect. If the procedural characteristics (reflecting upon the substantive differences between imple‐

<sup>2303</sup> Differently: Schlacke, Komitologie 319; U Stelkens, Unionsverwaltungsrecht 513 ff.

<sup>2304</sup> See also Tallberg, Paths 613, with a view to what he describes as the 'Management Approach': 'Non-compliance, when it occurs, is not the result of deliberate deci‐ sions to violate treaties, but an effect of capacity limitations and rule ambiguity. By consequence, non-compliance is best addressed through a problem-solving strategy of capacity building, rule interpretation, and transparency, rather than through coercive enforcement'.

<sup>2305</sup> See 3.1.1.2.1.1. above.

mentation and enforcement) are not indicative, we have to find out about the *telos* of the respective mechanism otherwise.<sup>2306</sup> Thereby, we are thrown back to our abstract definition: The focus of implementation is to concre‐ tise EU law, thereby reducing the likelihood of infringements, whereas the focus of enforcement lies on the determination of an infringement (or, if that is not possible: a lawful situation).

In order to illustrate the above problems of distinction, *Andersen* takes the example of the procedure laid down in Article 25 of (outdated) Council Regulation 2847/93. This procedure requires MS to carry out certain tech‐ nical controls to ensure compliance with specific objectives related to the EU's fisheries policy, and empowers the Commission to make proposals to the Council for the adoption of appropriate general measures where it has established that a MS has not complied with the aforementioned duty. *An‐ dersen* claims that 'the establishment of an infringement is not tantamount to an authoritative interpretation' and hence this regime does not encroach upon the powers of the CJEU.2307 With regard to the procedure at issue, the author would agree on the latter conclusion because here the Commission 'establishes' the infringement *in order to* make proposals to the Council for the adoption of appropriate general measures. It does not establish the infringement in the form of a decision *vis-à-vis* the MS concerned, but this determination is a mere prerequisite for the Commission to adopt propos‐ als, accordingly. As regards the (alleged) difference between the establish‐ ment of an infringement and the authoritative interpretation of EU law, the conceptual separation is acknowledged, but in practice the establishment of an infringement by means of a (binding) decision of an administrative EU body certainly entails an interpretation of EU law which bears a significant authority.2308 It is true that also in this case it is the Court – if it is called upon eg in the course of an annulment procedure – which has the final say. However, if this possibility were the only requirement for rendering lawful administrative output in this context, all mixed and hard compliance

<sup>2306</sup> For the difficulty to find out about the purpose of a law and its subjective implica‐ tions more generally see Schober, Zweck 3–5; see also 3.1.1.2.1.3. below.

<sup>2307</sup> Andersen, Enforcement 143.

<sup>2308</sup> That the binding interpretation of EU law *vis-à-vis* national authorities is a strong power which not any EU body may be granted is illustrated in case 19/67 *Bestuur der Sociale Verzekeringsbank*, 355, and in case 98/80 *Romano*, para 20, in which the Court declared – in a systematic interpretation of the EEC Treaty – a 'decision' of the Administrative Commission (an EU body) to be non-binding.

mechanisms would have to be accepted, no further examination of them being required (for this question see also 3.3.3. below).

The legislator's awareness of the fine line between the authoritative de‐ termination of an infringement – which is a prerogative of the Court<sup>2309</sup> – and the mere investigation of, and conclusion on, the correct application of EU law which may be necessary for an EU administrative body to perform its implementing powers is illustrated by the following example: While the Commission proposal for what later became (meanwhile outdated) Coun‐ cil Regulation 2371/2002 provided that '[a]ny loss to the common living aquatic resources resulting from a violation of the rules of the common fisheries policy attributable to any activity or omission by the Member State [to be established by the Commission] shall be made good by the Member State',<sup>2310</sup> the wording of the respective provision, in the course of the legislative procedure, was changed to: 'When the Commission has established that a Member State has exceeded the fishing opportunities which have been allocated to it, the Commission shall operate deductions from future fishing opportunities of that Member State'.2311 Thereby the Commission's power to establish a *violation* of EU law on the part of a MS was replaced by a power to compensate an excess of the allocated fishing opportunities, or in other words: make sure that the regime is uniformly applied. These two versions of the provision exemplify well the difference between (but also the proximity of ) implementing measures (aiming at the uniform application of EU law by the MS) on the one hand, and enforcement measures (focussing on its violation) on the other hand.

## 3.1.1.2.1.3. The indicative value of the material scope of and institutional questions relating to compliance mechanisms

In distinguishing the two realms of implementation and enforcement, the material scope of the mechanism at issue may have indicative value, as well. While the Treaty infringement procedure as the main enforcement mecha‐ nism of the EU has a general scope, the Treaties also provide for specific enforcement mechanisms, eg Article 108 TFEU. Implementing mechanisms may also have a broader scope (eg Article 114 paras 4 ff TFEU), but when

<sup>2309</sup> Andersen, Enforcement 144.

<sup>2310</sup> Commission Proposal COM(2002) 185 final, Article 23 para 4 (first sentence).

<sup>2311</sup> Article 23 para 4 subpara 1 of Council Regulation 2371/2002.

they are laid down in secondary law by tendency their respective scope is much narrower. This is not least due to the fact that Article 291 para 2 TFEU, the main provision on the implementation of Union law by EU actors, prescribes that the Commission (or exceptionally the Council) shall only be empowered where 'uniform conditions' for the implementation of EU law are required. Regularly, it is only specific rules (falling within a specific material scope) which require uniform conditions for implementa‐ tion – which is why the Commission is empowered specifically in many different acts of secondary law. While the Court over time seems to have loosened the level of determination required for the basic rules, thereby allowing for a broader qualitative scope of implementing powers (that means: a broader measure of discretion for the implementing EU actor),<sup>2312</sup> too broad a quantitative scope (granted eg for a Regulation as a whole) still rather speaks against implementation.

When comparing the various compliance mechanisms, it attracts atten‐ tion that some of them are directed to 'the Member State' concerned,<sup>2313</sup> others are more specifically directed to the national authority in charge.<sup>2314</sup> This distinction does not appear to be a peculiarity either of implementa‐ tion or of enforcement. Rather, the more specific compliance mechanisms tend to relate to the national bodies in charge, whereas more general com‐ pliance mechanisms – also the Treaty infringement procedure – relate to 'the Member State' concerned. In the former case, the national body reques‐ ted to comply with EU law is determined in the request for compliance, in the latter case the determination of the specific body or bodies in charge is up to the national sphere. This latter mode appears to consider national sovereignty, *in concreto*: the MS' procedural autonomy, to a larger extent.<sup>2315</sup> In complex (in particular: federally organised) national administrations it may be difficult for EU actors to find out which national bodies are in charge in a concrete case. Therefore EU law allows EU actors to address the MS as a whole, leaving the question of internal (national) competence up to this MS.<sup>2316</sup>

<sup>2312</sup> See case C-240/90 *Germany v Commission*, para 41; still more strictly: case 291/86 *Central-Import Münster*, para 13.

<sup>2313</sup> For example Article 13 para 1 of Directive 2001/95/EC; see IV.2.2.1.2.1. above.

<sup>2314</sup> For example Article 7 para 4 of Regulation 806/2014; see IV.2.2.2.2.5. above.

<sup>2315</sup> See already Constantinesco, Recht 299; see also Gil Ibáñez, Supervision 212–215.

<sup>2316</sup> For this question see also Schütze, Rome 1418 f, with further references; see case C-359/92 *Germany v Council*, para 38, according to which 'the measures taken for the implementation of Article 100a of the Treaty [now: Article 114 TFEU]

At the EU level, implementation is performed by administrative EU bodies. According to Article 291 TFEU these are the Commission and exceptionally the Council, but also other bodies, eg European agencies may be vested – on different legal bases – with implementing powers.<sup>2317</sup> Thus, where a legislative act providing for a compliance mechanism does not refer to Article 291 para 2 TFEU (including comitology as laid down in Regulation 182/2011) when regulating (the creation of ) the EU output to be adopted in the framework of this mechanism and/or where it empowers an EU body other than the Commission or the Council, this legislative act may still lawfully provide for implementation.<sup>2318</sup> The measure of inde‐ pendence these bodies dispose of varies.<sup>2319</sup> In particular the Commission and European agencies are accountable to at least one of the institutions comprising the (ordinary) EU legislator (that is the Council and the EP). They can normally act upon their own motion, sometimes they can act *also* upon request, rarely *only* upon request, mostly by other EU actors or MS (authorities).

Enforcement, on the other hand, idealtypically is performed by the CJEU, an independent body2320 which acts only upon request – in case of the Treaty infringement procedure this is a request (action) of the Com‐ mission (or a MS), which again can act on its (their) own motion.

The lawfulness of implementing acts may be scrutinised by the CJEU, if they are 'intended to produce legal effects *vis-à-vis* third parties' pursuant to Article 263 para 1 TFEU (see also 3.6. below). With enforcement according to Articles 258–260 TFEU the situation is different: Judgements of the Court are final.

are addressed to Member States and not to their constituent entities'. The Court referred to the division of competences between the *Bund* and the *Länder* in Germany, but its words could also be applied with regard to national authorities.

<sup>2317</sup> See also 3.3.4. below.

<sup>2318</sup> See case C-270/12 *United Kingdom v European Parliament/Council*, paras 78 ff: with regard to European agencies; case C-521/15 *Spain v Council*, para 43: with regard to the Council.

<sup>2319</sup> Note the CJEU which – in the context of national supervisory authorities, but still – stressed the importance of independence in the context of ensuring compliance by uttering that the independence of these bodies 'is intended to ensure the effec‐ tiveness and reliability of the monitoring of compliance with the [relevant law]'; case C-362/14 *Schrems*, para 41.

<sup>2320</sup> Article 19 para 2 TEU.

3.1.1.2.2. The preliminary reference procedure, the procedure pursuant to Article 218 para 11 TFEU, and the excessive deficit procedure – special cases

When trying to distinguish implementation and enforcement undertaken by EU bodies *vis-à-vis* the MS, the preliminary reference procedure inevi‐ tably attracts attention. Article 267 TFEU provides for a procedure in the course of which the CJEU shall answer questions of a court or tribunal of a MS on the interpretation of EU law, and the validity of EU law (other than primary law) respectively. The answer given by the CJEU is binding for the national court or tribunal which has referred the questions to the former.<sup>2321</sup> When it comes to the classification of this procedure, it is clear that this is not an enforcement procedure within the meaning fleshed out above.2322 It does not entail the review of MS action in terms of compliance with EU law. Rather, it is a procedure which may only be initiated upon request by a national court or tribunal and which is embedded in an on-going national court procedure. These latter characteristics may speak in favour of implementation. The fact that the Court takes action only upon a mostly2323 voluntary request by the national court or tribunal feebly speaks against this view. While content-wise it could be called implementa‐ tion, in institutional terms such qualification would clearly go against the implementation regime set up by the Treaties.2324 Hence the preliminary reference procedure should be qualified as a procedure *sui generis*. It is to be accepted as a special case of supporting national courts or tribunals in their handling of matters of EU law which – *qua* being provided for in pri‐ mary law – in legal terms cannot possibly conflict with the implementation and enforcement regimes set up by the Treaties.

<sup>2321</sup> Whether and, if so, to which extent the Court's answer is binding also beyond the case at issue is contested; see *Ehricke,* Bindungswirkung 44 ff, with further referen‐ ces; affirming an *erga omnes* effect in practice: Broberg, Preliminary References 107.

<sup>2322</sup> For a discussion of the increasing number of preliminary references in the context of enforcement see Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 7.

<sup>2323</sup> For obligatory requests see Article 267 para 3 TFEU, the *CILFIT* doctrine and the *Foto-Frost* doctrine of the Court; see case 283/81 *CILFIT* and, more recently, case C-561/19 *Consorzio Italian Management*; case 314/85 *Foto-Frost*.

<sup>2324</sup> In the literature, the preliminary reference procedure is sometimes referred to as an enforcement measure, but then this term, unlike here, is not used in conceptual opposition to implementation: see Broberg, Preliminary References 99, with a further reference.

The procedure according to Article 218 para 11 TFEU may have a charac‐ ter which is similar to that of the preliminary reference procedure, namely where it is initiated by a MS.<sup>2325</sup> The Court's opinion is then addressed to the respective MS. While this procedure is designed to clarify whether or not an agreement envisaged pursuant to Article 218 TFEU is in accordance with primary law, the Court in its past case law has also reflected upon the competences of the MS. While being politically useful, from a legal point of view this has been criticised as falling neither within the scope of Article 218 para 11 TFEU nor within the tasks and powers of the Court more generally.<sup>2326</sup> In the given context, this procedure is to be mentioned only as far as the Court addresses its opinion to a MS, thereby ensuring that it complies with Union law. However, since its opinion determines the legal situation of all the MS (intending to conclude the respective agreement) – and also the EU – its individual-concrete character is strongly diluted. Thus, the procedure pursuant to Article 218 para 11 TFEU is to be mentioned here, but in developing a Treaty-based distinction between implementation or enforcement it is far less important than the procedures laid down in Article 267 TFEU and in Article 126 TFEU, respectively.

The excessive deficit procedure as laid down in Article 126 TFEU (see IV.2.2.2.1.2. above) is another special case. It provides for the determination of an infringement of EU law – namely of the EU's deficit criteria – by a MS and, possibly, the imposition of sanctions, including financial sanctions.<sup>2327</sup> While these indicators clearly point in the direction of enforcement,2328 it is the Council, together with the Commission, which takes the lead in this procedure. The application of the Treaty infringement procedure is explicit‐ ly excluded. This clearly is an exception to the principle of the Treaties that the enforcement of EU law *vis-à-vis* the MS is ultimately performed by the Court. But since it is laid down in primary law, the classification along the lines of the above discussion – like in the case of the preliminary reference procedure – is of secondary importance anyway. Again, this procedure is to

<sup>2325</sup> For this procedure see also III.3.5.2.5. above.

<sup>2326</sup> See Lorenzmeier, Art. 218 AEUV, para 76; Schmalenbach, Art. 218 AEUV, paras 39–41, both with further references to the Court's case law.

<sup>2327</sup> For the application or rather non-application of the sanctions regime in the past see Bieber/Maiani, Enforcement 1066.

<sup>2328</sup> At the same time, it ought to be mentioned that the Council recommendations which may be adopted in the course of an Article 126-procedure shall give the MS concerned a guideline on how to remedy the violation of EU law (see 3.1.1.2.1.2. above). This indicator of implementation, however, stands back before the stronger enforcement thrust of Article 126.

be accepted as a special case which is protected from legal challenge *qua* its belonging to the topmost layer of EU law.

In view of these examples, it may be argued that the conclusion *tertium non datur* uttered above (3.1.1.1.) is challenged. However, these procedures do not form a specific third category of their own. They differ strongly from each other. While the preliminary reference procedure – and, to the limited extent measured above, the procedure laid down in Article 218 para 11 TFEU – could be dubbed 'implementation by the Court' (which in terms of the Treaties would be non-system), the excessive deficit procedure constitutes enforcement performed by administrative bodies – again an oddball under the Treaties. Insisting on an allocation to either of the two categories would be misleading in both cases. Perhaps these procedures in conceptual terms are best grasped as the famous exceptions confirming the rule.

### 3.1.1.2.3. Two further aspects: soft law and sanctions

We have not yet specifically addressed the issue of soft law and in particular soft mechanisms here. In the given context, they certainly play an ambiva‐ lent role. On the one hand, they seem to be rather submissive when it comes to an overlap with other institutions', bodies', offices' or agencies' (hard law) powers. On the other hand, the output created in the course of soft compliance mechanisms is largely excluded from review by the Court. The former means that a possible encroachment upon the competences of other bodies (in particular: the Court) *prima facie* appears to be less serious,2329 the latter means that the Court with regard to soft law cannot display its genuine role – that is to 'determine[] the scope of the provisions of the Treaties whose observance it is its duty to ensure'.<sup>2330</sup>

Enforcement by the Court shall not take place in the form of soft law. This is why individual-concrete soft law is only adopted by administrative bodies such as the Commission or European agencies, under the heading 'implementation'. This does not, however, as such exclude the possibility

<sup>2329</sup> See already Opinion of AG *Roemer* in cases 9–10/56 *Meroni*, according to whom the delegation of powers relating to 'supporting preparatory work and the purely technical implementing measures' to other bodies than the body taking the deci‐ sion in a certain procedure is unproblematic.

<sup>2330</sup> CJEU, Report of the Court of Justice on certain aspects of the application of the Treaty on European Union (Luxembourg, May 1995) 2.

of a material interference with the Court's enforcement powers by means of soft law. Not least in view of this risk, the institutional balance is to be considered also where only soft law powers are at issue (see 3.3. below).

With sanctions the situation seems to be clear only at first sight. Sanc‐ tions are a classical means of enforcement.<sup>2331</sup> Sanctions *vis-à-vis* the MS are an exception (see Article 260 TFEU, Article 126 para 11 TFEU), and they do not only aim at preventing future violations of EU law (deterrent effect), but in particular they serve to punish an actual violation of EU law (penal‐ ising effect).<sup>2332</sup> Since the EU cannot normally replace MS action by its own action,<sup>2333</sup> sanctions are the strongest means of ensuring compliance in this context. The reason why sanctions imposed on private actors – eg by the Commission for violation of EU competition law – do not bring about the same legal intricacy and sensitivity is that with regard to individuals/un‐ dertakings no general procedure comparable to the Treaty infringement procedure is provided for. Therefore, with EU bodies being empowered to impose sanctions on individuals/undertakings, the Court's role does not seem at risk of being challenged.2334 But also in the relationship EU-MS – which builds the focus of this discussion – the Court has accepted, to some extent at least, the use of sanctions by the EU administration. In the case *Spain v Council*, the Court has addressed this question, unsurprisingly against the background of a compliance mechanism, namely Article 8 of Regulation 1173/2011. This Regulation, as part of the so-called 'Six Pack',<sup>2335</sup> is intended to cater for 'the effective enforcement of budgetary surveillance in the euro area'. Its Article 8 empowers the Council to impose sanctions on a MS 'that intentionally or by serious negligence misrepresents deficit and debt data relevant for the application of Articles 121 or 126 TFEU,

2332 See Posch/Riedl, Art. 260 AEUV, para 49; Wunderlich, Art. 260 AEUV, paras 21f.

<sup>2331</sup> For the difference between financial correction (administrative measures adopted eg in the context of payments from the European Structural and Investement Funds) and enforcement through eg penalties see Andersen, Enforcement 182 f; on the application of financial corrections see, as an example, case C-332/01 *Greece v Commission* or case C-8/88 *Germany v Commission*, the latter being commented on by: Comijs, Priorities 305.

<sup>2333</sup> Note the words of Bieber/Maiani, Enforcement 1061: '[C]entralized enforcement, even at its strongest, must elicit (and cannot replace) "sincere cooperation" in the sense of Article 4(3) TEU'.

<sup>2334</sup> See Montaldo, Power 131–136.

<sup>2335</sup> For the 'Six Pack' more generally see Gerapetritis, Constitutionalism 53 f; for another compliance mechanism under the Six Pack allowing for the imposition of fines see IV.2.2.2.2.4. above.

or for the application of the Protocol on the excessive deficit procedure'. These sanctions are of an 'administrative nature'.<sup>2336</sup> The Court may not only annul the sanctioning decision, it may also reduce or increase the fine. Against this background, the Council – in the case at issue – has adopted Implementing Decision 2015/1289 addressed to Spain, thereby imposing on this country a fine of about 19 million euro.

The Court seems to accept the qualification of the Council decision as implementing act, but utters its doubts as to whether the according power could be based on Article 291 para 2 TFEU, as this provision 'relates solely to legally binding acts of the European Union which lend themselves in principle to implementation by the Member States […] but which, in contrast to the latter acts, must, for a particular reason, be implemented by means of measures adopted not by each Member State concerned, but by the Commission or the Council, for the purpose of ensuring that they are applied uniformly within the European Union'2337; and further: 'That is clearly not so in the case of an act which establishes a power consisting in the imposition of a fine on a Member State. Such an act does not lend itself in the slightest to implementation by the Member States themselves, as implementation of that kind involves the adoption of an enforcement measure in respect of one of them'.<sup>2338</sup>

Apparently, the Court does not see implementation and enforcement as opposites here, but rather deems enforcement a sub-category of implemen‐ tation.2339 It concludes that the Council decision at issue constitutes an implementing act in a more general sense, but not within the meaning of Article 291 para 2 TFEU. The Court thereby seems to exclude the possibility of granting the power to adopt financial sanctions on the basis of Article 291 para 2 TFEU (in conjunction with a material competence). The AG in this case, *Juliane Kokott*, expresses: 'The concept of "implementation" comprises both the drawing-up of implementing rules and the application of rules (of secondary legislation) to specific cases by means of acts of individual application. Imposing a fine thus appears to be an implementing

<sup>2336</sup> Article 9 of Regulation 1173/2011. For the possible lack of a 'moral reproach' under‐ lying these sanctions see Zuleeg, Enforcement 351.

<sup>2337</sup> Case C‑521/15 *Spain v Council*, para 48.

<sup>2338</sup> Case C‑521/15 *Spain v Council*, para 49.

<sup>2339</sup> In this case the Court was also concerned with the interpretation of Art 51 of the Court's Statute, eventually confirming its jurisdiction (and not that of the General Court) in spite of dealing with an implementing act; case C‑521/15 *Spain v Council*, paras 39–51.

measure [a *sui generis* implementing measure<sup>2340</sup>] and the power to adopt such a measure appears to be an implementing power'.<sup>2341</sup>

In view of the *ex post* occurrence of the fine, the determination of a violation of Union law (by a MS) it entails, and its focus on reacting to a violation of (not: concretising) the underlying material rule,<sup>2342</sup> the better arguments seem to speak in favour of qualifying this mechanism as an enforcement tool according to the regime set up above, irrespective of its legal basis.<sup>2343</sup> Also the fact that the sanction depends on intent or negligence of the actor having infringed the law, which means that the procedure is subjective in nature, points in the direction of enforcement.<sup>2344</sup> The same is true for the CJEU's power to reduce or increase the fine in the course of judicial review. The regular procedure as provided for in Articles 258–260 TFEU – the Commission confronts the MS and the matter may then be decided by the Court – is reversed in this case: the Commission, subject to formal adoption by the Council by reverse qualified majority, sets the fine and it is then the MS which may go to Court, bearing the general risk of litigation, and – above all – bearing the risk that the Court may even increase the fine.

Also from a different perspective the procedure is remarkable. Article 136 and Article 121 para 6 TFEU (in conjunction) – the 'effective enforcement' of which Regulation 1173/2011 is intended to serve – do not (explicitly) provide for any binding EU output, let alone sanctions. What is more, the power to reduce or increase administrative sanctions goes beyond a mere power of annulment (as laid down in Article 263 TFEU).<sup>2345</sup>

<sup>2340</sup> Opinion of AG *Kokott* in case C-521/15 *Spain v Council*, para 52.

<sup>2341</sup> Opinion of AG *Kokott* in case C-521/15 *Spain v Council*, para 47; see also para 48.

<sup>2342</sup> See also case C‑521/15 *Spain v Council*, para 53: 'deterring the Member States'; referring to the similarity of this procedure and the Treaty infringement procedure see Opinion of AG *Kokott* in case C-521/15 *Spain v Council*, para 53.

<sup>2343</sup> The question whether the enforcement power was lawfully granted by the legisla‐ tor – and hence whether the used legal basis is correct and sufficient – shall be adressed in a next step; see 3.2. below.

<sup>2344</sup> Even though the Treaty infringement procedure is said to be objective, if the procedure reaches the state of sanctions it can be assumed that the MS concerned has intentiously infringed EU law; see also Koops, Compliance 149.

<sup>2345</sup> See Article 261 TFEU, though, which allows the EP and the Council, or the Council on its own, to grant the Court 'unlimited jurisdiction with regard to the penalties provided for in [their respective] regulations'; see also 3.6. below; also note the discretion of the Court pursuant to Article 260 para 2 TFEU to deviate – in terms of financial sanctions – from the suggestions of the Commission.

### 3.1.1.2.4. Conclusion

The material distinction between implementation and enforcement on the basis of EU primary law is difficult, but worthwhile. It is difficult essentially because both spheres share one important aim – MS' compliance with EU law – and because the Treaties do not provide a comprehensive, let alone an explicit circumscription of these spheres. It is worthwhile because it will increase our understanding of the scope (including, in particular, the limits) of the – in this context – main prerogative of the EU's admin‐ istration (implementation *vis-à-vis* the MS) on the one hand, and of the main prerogative of the EU's judiciary (enforcement *vis-à-vis* the MS), on the other hand. Eventually, this differentiation will allow us to detect interferences – be they of a singular or of a structural kind – of one sphere with the respective other.

We can summarise the characteristics fleshed out in the course of the above comparison between the implementation and the enforcement of EU law, both undertaken by EU bodies *vis-à-vis* the MS, by using the following table:

*V. CLASSIFICATION AND LEGAL ASSESSMENT OF COMPLIANCE MECHANISMS*

*Table: Characteristics of individual-concrete implementation/enforcement by EU bodies vis-à-vis the MS*


As was underlined above (3.1.1.1.), none of these characteristics is, on its own, sufficient. Necessary appears, in our specific context of compliance mechanisms, only the individual-concrete character of the output – a char‐ acteristic which both categories share and which thus cannot serve as a differentiator, and, in case of enforcement, the legally binding nature of the output.

The remaining criteria are to be understood as idealtypical characteris‐ tics. If one of them is not met in a specific case, this shortcoming may be balanced by a strong stance on (most of ) the other criteria. For example: Where a compliance mechanism is clearly focused on the concretisation of EU law, its output occurs *ex ante*, is addressed to a specific MS and, due to its bindingness, subject to judicial review, then the fact that the EU body at issue cannot act *proprio motu* in this procedure does not stand in the way of qualifying this mechanism as an instance of implementation. While the primary aim and the character of the command, if they are clear, appear to be the most significant aspects – a legally non-binding act cannot be qualified as enforcement measure – it does not make sense to give each characteristic a certain weight so that the categorisation of a specific mechanism would be a matter of mere calculation. This could be feasible if we had a large number of characteristics, but in the given case it would only result in *pseudo*-accuracy and we would deprive ourselves of the leeway in analysing these highly heterogeneous mechanisms which is, in my view, necessary to do justice to their respective individuality.

Where safety or health concerns require a fast reaction – to take another example – this may 'compensate' the enforcement tendency of a compliance mechanism, and lead to the conclusion that – also in light of Article 114 paras 4 ff TFEU which protects these policy objectives – the (implement‐ ing) mechanism at issue was lawfully based on Article 114 TFEU.

In general, we need to bear in mind that while a clear allocation to either category may be desirable, there may be stalemate cases, or cases where only a tendency in either direction can be established.

What shall follow in the next sub-chapter is an investigation of the compliance mechanisms presented above against the background of these characteristics, with a view to classifying them as either belonging to the realm of implementation of EU law or to the realm of its enforcement. As was stated above, the qualification as enforcement mechanism of a secondary law mechanism performed by an EU administrative body as such raises concerns as to the legality of this measure – in particular against the background of the Treaty infringement procedure as the general regime for the enforcement of EU law by EU actors *vis-à-vis* the MS. However, the allocation of a mechanism to either implementation or enforcement can only be a first step in providing a comprehensive legal account of this mechanism. A more thorough analysis would require also to take into account, for example, the primary law provision on which this mechanism is based. Thus, the latter aspect and further legal aspects will be considered in the chapters below.

# 3.1.2. The categorisation of the compliance mechanisms

# 3.1.2.1. Introduction

Against the background of the (idealtypical) characteristics elaborated above, we shall now examine the compliance mechanisms presented in Part IV above with a view to allocating them to either category. The Treaty infringement procedure being the standard enforcement procedure of the EU, it does not come as a surprise that it meets all criteria of an enforce‐ ment mechanism. The possibility of sanctions is to be confirmed (Article 260 TFEU). A legal remedy is not available because the Treaty infringement procedure is reserved for the Court of Justice.2346 The criteria are also met by the variants of the Treaty infringement procedure laid down in primary law (see IV.2.2.1.1.2., IV.2.2.1.1.3., IV.2.2.1.1.4., IV.2.2.2.1.3. above), which, for lack of doubt, need not be analysed in more detail in this context.

From among the remaining compliance mechanisms, an analysis of the hard mechanisms shall be followed by an analysis of the mixed mech‐ anisms. Eventually, the soft mechanisms shall be addressed. While soft mechanisms, for lack of legally binding output produced in their respective course, cannot fall within the realm of enforcement, they shall be consid‐ ered here nevertheless with a view to the other criteria. Even though they cannot be called enforcement mechanisms, they may still display certain similarities with them, thereby possibly interfering with the EU's enforce‐ ment regime under the Treaties.

<sup>2346</sup> Article 256 para 1 TFEU *e contrario*.

### 3.1.2.2. Hard mechanisms

We shall start with the mechanisms laid down in primary law. The pro‐ cedure laid down in Article 106 para 3 TFEU (see IV.2.2.1.1.1. above) is ambiguous, but eventually an allocation to one category can be made. It may serve either the primary aim of concretisation or the primary aim to end non-compliance which is also why it may be initiated at any time – *ex ante*, in the course of a MS action which may lead to non-compliance, or *ex post*. The implementation list is indicated by the competence of the Commission to adopt, apart from a decision, a directive which *may* have a general-abstract scope.<sup>2347</sup> The legal bindingness of the output and its potential individual-concrete scope, on the contrary, are neutral properties. The other characteristics (Commission in charge, action upon its own motion, no sanctions, judicial review available) speaking in favour of imple‐ mentation, the mechanism appears to result in implementation rather than enforcement.

The mechanism laid down in Article 114 TFEU (see IV.2.2.1.1.3. above) rather provides for the implementation of EU law. It is mainly about the concretisation of the exceptional possibility of MS to deviate from a harmo‐ nisation measure under Article 114 TFEU. It allows for *ex ante*/on-going intervention, because the MS turn to the Commission asking whether they could maintain or introduce deviating national rules (which they present to the Commission). In both cases the national rules may enter into force only after the Commission's authorisation. The Commission's decision may be made subject to judicial review before the CJEU. The fact that the Commission acts only upon notification does not change the mechanism's implementation character. The possibility for the Commission to turn to the CJEU in case a MS makes 'improper use' of its powers in this context is a variant of the Treaty infringement procedure and hence to be qualified as enforcement procedure. This enforcement procedure is related to the implementing mechanism laid down in Article 114 TFEU, but conceptually nevertheless can clearly be distinguished from it.

Also with regard to the protective measures which a MS with a dero‐ gation may take in case of a sudden crisis in the balance of payments under Article 144 TFEU (IV.2.2.1.1.5. above) the categorisation is clear. The

<sup>2347</sup> Since Article 106 TFEU provides for the power to adopt a decision as an individu‐ al-concrete act, it appears that under this regime the directive *normally* is expected to have a general-abstract scope; see IV.2.2.1.1.1. above.

Council may take a decision upon a Commission recommendation, after the MS at issue has informed the Commission and the other MS of its action. The fact that the Council may not only require the MS to abolish the measures taken, but also to amend or suspend them suggests that the mechanism aims at creating a situation which is in accordance with EU law by concretising the latter (an exceptional deviation competence of the non-euro MS which is nevertheless drafted in comparatively general terms), not at determining an infringement on the part of the MS (even though the Council is acting *ex post*, that is to say after the measures of the MS have taken effect). That the Council acts on a recommendation from the Commission and hence not on its own motion does not challenge the implementation character of this procedure. The decision of the Council can be reviewed under Article 263 TFEU.

Article 13 para 1 of Directive 2001/95/EC (see IV.2.2.1.2.1. above) clearly provides for an implementing mechanism. While it envisages *ex post* inter‐ vention, it allows for a reaction to safety concerns in specific MS. The implementing act according to Article 291 TFEU (which is challengeable before the Court) may only be adopted if, among other things, the risk at issue cannot be eliminated other than by adopting this act. The Commis‐ sion acts on its own motion, thereby concretising the requirements under EU law. The temporary nature of (some of ) the Commission output2348 un‐ derlines its implementing (rather than enforcement) character. The Court itself – with regard to the predecessor mechanism of Article 13 of Directive 2001/95/EC – has acknowledged that the Treaty infringement procedure 'does not permit the results set out in Article 9 of the directive [92/59/ EEC] to be achieved',2349 thereby stressing the difference between the two regimes.

In order to determine whether they display an implementation or rather an enforcement thrust, the rules laid down in Articles 70 f of Regulation 2018/1139 (see IV.2.2.1.2.2. above) require some further analysis. The two regimes laid down in these provisions are about safety concerns or urgent unforeseeable operational circumstances or needs in the application of the relevant EU law. The purpose is to ensure a harmonised legal situation in all MS which at the same time does justice to the aforementioned concerns, circumstances or needs. Under the regime laid down in Article

<sup>2348</sup> See Article 13 para 2 of Directive 2001/95/EC.

<sup>2349</sup> Case C-359/92 *Germany v Council*, para 46.

70, the MS are allowed to deviate, under certain circumstances, from the relevant EU law to address safety concerns. Under the regime laid down in Article 71, the MS are allowed to deviate (temporarily or under certain circumstances) from the relevant EU law in case of urgent unforeseeable operational circumstances or needs. These characteristics speak in favour of concretisation, not of a focus on determining infringements of EU law by MS. While the mechanisms apparently allow for *ex post* intervention on the part of the EU and while the Commission acts upon a recommendation of the EASA, the EU output is aimed at concretising EU law by ensuring that the possibilities for a lawful derogation from EU law are correctly applied. These exceptional deviations for specific reasons seem to be an expression of implementation in the spirit of the Treaties.<sup>2350</sup> The fact that the Com‐ mission's (or exceptionally under Article 70: the EASA's) output can be made subject to judicial review pursuant to Article 263 TFEU underpins this view.

The mechanism laid down in Article 29 para 2 of Regulation 806/2014 (see IV.2.2.1.2.3. above) provides for a possibility of the SRB to directly address an institution under resolution in case a national resolution author‐ ity has not complied with an SRB decision. This measure is taken *ex post* upon the SRB's own motion and can be challenged – also by the MS – under Article 263 TFEU. Because of the short route to compliance by ousting the national authority this mechanism raises concerns as regards the principle of MS implementation of EU law pursuant to Article 291 para 1 TFEU.2351 However, it ought to be taken into account that the national authority has been addressed already by the SRB decision which it alleg‐ edly does not comply with. This suggests a certain consideration of the national prerogative of implementation,2352 but the shortcut also illustrates the focus on reaching compliance rather than concretising the law. After all, if a MS does not comply with EU law (here: an SRB decision), the regular route under the Treaties would be to initiate a Treaty infringement procedure. Apparently, the necessity of a Treaty infringement procedure shall be avoided by the shortcut. This ousting of the regular enforcement

<sup>2350</sup> See Article 114 para 10 or Article 192 para 5 TFEU.

<sup>2351</sup> See also Scholten, Trend 1350.

<sup>2352</sup> Also under the similar regime of Article 17 of Regulation 1093/2010 the EBA deci‐ sion directly *vis-à-vis* the financial institution/financial sector operator concerned is provided for only as an *ultima ratio*, that is to say when the national authority does not comply even upon request; see 3.2.3. below.

procedure *vis-à-vis* the MS – a purpose which is also reflected upon in its broad material scope – confirms the strong enforcement tendency of this mechanism. In conclusion, even though the SRB decision may be made subject to an action for annulment also by the MS,<sup>2353</sup> Article 29 para 2 of Regulation 806/2014 tends towards enforcement rather than towards implementation.

Article 63 of Regulation 2019/943 (see IV.2.2.1.2.4. above) provides for a mechanism in the course of which the Commission can ensure compliance with EU law by a MS (national regulatory authority) or exceptionally the ACER. The MS or exceptionally the ACER decide on the exemption from certain requirements under the relevant EU law for new direct current interconnectors upon their respective request. The Commission may scruti‐ nise positive decisions (that is to say: decisions granting the exemption) *ex post* and may approve of them or order the notifying bodies to amend or withdraw them. It is to be acknowledged that here exceptional deviations from EU law (exemptions) are at issue. We have come across the possibility of lawful deviations from EU law for MS above. Here these deviations are requested by undertakings. The effects for our purposes are the same. The handling of exceptional (lawful) deviations from EU law seems to fall within the realm of implementation, as it is strongly connected to the concretisation of these deviations – in the interest of as harmonised an application of Regulation 2019/943 as possible. Under this regime, the Commission normally takes its decisions upon its own motion.2354 The decisions may be made subject to judicial review (Article 263 TFEU) by the MS. All in all, this regime is to be classified as an implementing mechanism.

The considerations above on the mechanism laid down in Article 29 para 2 of Regulation 806/2014 *mutatis mutandis* also apply with regard to the, in some important respects, similar procedures laid down in Articles 18 and 19 of Regulation 1093/2010 (see IV.2.2.1.2.5. above). The fact that here an emergency situation, and a continuous competence conflict between two or more national authorities respectively, are at issue – that is to say: excep‐ tional and highly undesirable situations – lets the arguable interference with the Court's (enforcement) prerogatives appear in a more mellow light. The question whether European agencies (here: the EBA) – instead of the

<sup>2353</sup> Whether this is possible also for the 'skipped' national authority (on the basis of Article 263 para 4 TFEU) is unclear; Article 86 para 2 of Regulation 806/2014 does not seem to support this possibility.

<sup>2354</sup> The reopening of the procedures may also take place upon request; see Article 63 para 10 of Regulation 2019/943.

Commission – may be granted this *pouvoir* under the Treaties in the first place is addressed below (3.3.4.).

With the provisional measures subject to a Union control procedure as allowed for in Article 114 para 10 TFEU, we had a look at three examples laid down in secondary law (see IV.2.2.1.2.6. above). They all provide for the exceptional adoption of provisional national measures deviating from Union law. Immediately after their adoption (that is to say: *ex post*) they shall be scrutinised by the Commission, which may approve of them or re‐ quest for them to be withdrawn. These exceptions are a possibility to adapt the requirements of EU law to specific concerns, and hence in a way they lead to the concretisation of EU law. The Commission intervention occurs directly after the adoption of the act. It may be made subject to judicial review pursuant to Article 263 TFEU. Not least due to their framing by an explicit Treaty provision concerning the implementation of EU law, these examples of provisional measures qualify as implementing mechanisms.

#### 3.1.2.3. Mixed mechanisms

Articles 116 and 117 TFEU provide for a regime to address MS measures which distort the conditions of competition to a certain extent (Article 116) or where there is at least 'reason to fear' that they will do (Article 117; see IV.2.2.2.1.1. above). In both cases the national measures may principally be lawful, but – due to the distortion the different measures in different MS may create – the EU feels required to intervene. Article 117 provides for *ex ante*/on-going monitoring of rule-making at the national level which may result in a recommendation addressed to the respective MS. This part of the procedure therefore clearly has an implementing thrust. Under Article 116, the EP and the Council in the ordinary legislative procedure may adopt the 'necessary directives' which should be addressed to the MS concerned. This procedure appears to be neither implementation nor enforcement. It is not primarily about concretising or about determining a violation of EU law. Rather, it is about tackling a situation where the divergence of national regulation (the 'difference' pursuant to Article 116 para 1 TFEU) leads to an undesirable result – the distortion of the conditions of competition. Here this is done by legislative intervention aimed at doing away with (or rather: superseding) national measures of a single (or a small number of ) MS which cause the distortion. A legislative act does not constitute implemen‐ tation within the meaning of the Treaties, nor can it qualify as enforcement. Therefore, Article 116 TFEU stands outside the above categorisation. As a provision of primary law, it falls outside the scope of legal scrutiny.

The excessive deficit procedure as laid down in Article 126 TFEU has been addressed and qualified as a *sui generis* mechanism above (3.1.1.2.2.).

Article 63 of Directive 2019/944 (see IV.2.2.2.2.1. above) provides for a procedure to scrutinise the compliance of MS acts with network codes and guidelines (binding delegated or implementing acts) of the Commission. Upon request by another national authority or the Commission, the ACER may address an opinion to the national authority concerned. Where the authority addressed does not comply, the Commission – upon request or on its own motion – may investigate the case further (thereby informing the authority at issue) and, if it does so, shall eventually decide either not to raise objections against the national measure or to require the authority to withdraw its decision. It appears that ending non-compliance stands in the foreground here, not the further concretisation of (already concretising) implementing law (the Commission's network codes and guidelines). The measures are taken *ex post*, which further underpins the enforcement char‐ acter of this mechanism. While the ACER opinion – *qua* non-bindingness – cannot constitute an enforcement measure, the (possibly) ensuing Com‐ mission decision can. The fact that a binding decision is preceded by a soft law act does not exclude the enforcement character of the succeeding decision. On the other hand, the Commission may not only act upon re‐ quest, but also on its own initiative, and its decision can be reviewed by the Court following an action for annulment. These single indicators pointing in the direction of implementation do not, in my view, challenge the strong enforcement thrust of this mechanism. Its *telos* is similar to that of the Treaty infringement procedure. The fact that the Directive sets certain deci‐ sion-making deadlines both for the ACER and for the Commission suggests that fast decision-making is desired. Comparatively short decision-making periods are a general *desideratum* of legal procedures in a system based on the rule of law. It is not apparent, however, that in this case requirements of urgency or importance of the matter (politically) suggest, when it comes to enforcement, not to rely exclusively on the Treaty infringement procedure.

Articles 22 f and 28 of Council Regulation 2015/1589 flesh out the variant of the Treaty infringement procedure laid down in Article 108 TFEU (see IV.2.2.2.2.2. above). These provisions extend the administrative phase of the infringement procedure, essentially by providing for the competence of the Commission to send a recommendation to the MS concerned, proposing in particular amendments, procedural requirements or the abolition of the aid scheme. Where the MS addressed accepts this recommendation, it becomes binding upon it. If it does not accept it or if it fails to comply with the accepted recommendation, the Commission may render a decision that the MS concerned shall abolish or alter the aid at issue within a certain period of time. Where the MS does not comply with the Commission decision, the Commission may refer the case to the CJEU. This is an enforcement procedure, as provided for in Article 108 TFEU, which is complemented by some details by means of Council Regulation 2015/1589.

The mechanism laid down in Article 17 of Regulation 1093/2010 (see IV.2.2.2.2.3. above) procedurally involves the Commission and the EBA. It involves *ex post* intervention for 'breach of Union law' which in this context shall be limited to the legal acts listed in Article 1 para 2 of Regulation 1093/2010. Nevertheless, for an alleged implementing mechanism this is a remarkably broad scope. It is this broad scope and the fact that the determination of an infringement appears to stay in the foreground here which emphasise the enforcement character of this mechanism. On the other hand, it involves two soft law acts (recommendation, formal opinion), and only if these acts are not complied with may the EBA adopt a decision directly addressed to a financial institution/financial sector operator (in certain cases to be preceded by a decision addressed to the competent au‐ thority concerned), given this is 'necessary to remedy, in a timely manner, such non-compliance in order to maintain or restore neutral conditions of competition in the market or ensure the orderly functioning and integrity of the financial system'. These factors again limit the scope of the mecha‐ nism, even though the latter criterion allows a broad measure of discretion for the EBA. Like under Article 29 para 2 of Regulation 806/2014, the direct intervention *vis-à-vis* the individual actors appears to oust the Treaty infringement procedure which is the general mode of tackling non-compli‐ ance with EU law on the part of a MS (see 3.1.2.2. above). In conclusion, this mechanism displays a strong tendency towards enforcement.

The excessive imbalance procedure laid down in Regulations 1176/2011 and 1174/2011 (see IV.2.2.2.2.4. above) is related to Article 121 TFEU which provides for a soft compliance mechanism. The excessive imbalance proce‐ dure is mixed in the sense that it may also involve binding acts of the EU. The specific negotiation element underlying Article 121 TFEU is also reflec‐ ted upon in the excessive imbalance procedure. Compliance does play an important role, but also the development of an appropriate solution. It runs *ex ante* and *ex post*, but also monitors on-going decision-making in the MS. It leaves the putting into effect of EU law up to the MS. They have to submit their plans which are then subject to scrutiny by the Commission and the Council. The regime is about concretising EU law acts which are mainly about economic development and hence subject to different views – hence the negotiations – to a larger extent than regular legal norms. So far, it can be assumed that the excessive imbalance procedure is not an enforcement mechanism, but rather entails a very special kind of implementation which is strongly shaped by primary law, namely Article 121 TFEU. But, eventually, the sanctions which are provided for give the procedure an enforcement spin. Whether Article 121 TFEU actually allows for the provision of legally binding acts, including sanctions, by means of secondary law (based on its para 6) will be assessed below (3.2.4.).

The mechanism laid down in Article 7 para 4 of Regulation 806/2014 (see IV.2.2.2.2.5. above) provides for a warning and/or a decision of the SRB being addressed to a national resolution authority. The decision leads to an attraction of principally national tasks by the SRB. The *ex post* inter‐ vention and the primary aim to end non-compliance (thereby avoiding the initiation of a Treaty infringement procedure), namely by the SRB exercis‐ ing certain originally national tasks itself create an enforcement character which is mitigated only to a limited extent by the fact that the SRB acts on its own motion and by the fact that the decision can be made subject to judicial review under Article 263 TFEU. All in all, the mechanism has a clear propensity towards enforcement.

Article 25 of Regulation 2016/796 (see IV.2.2.2.2.6. above) prescribes that the MS submit certain (national) draft rules on certain issues to the ERA and the Commission, which then examine these draft rules with a view to their compliance with the relevant Union law. Where the ERA establishes non-compliance, it addresses an opinion to the MS concerned, at least implicitly requesting it to amend its drafts. If the MS does not react in an appropriate way, the Commission may adopt a decision requesting the MS to modify or repeal the draft rules. This mechanism entails *ex ante* intervention. It is about the concretisation of specific EU law which shall be performed, if need is, in a dialectic process involving the MS and the EU. Only where the MS fails to adequately react to the suggestions coming from the EU (ERA), the Commission may end the discourse by adopting an implementing act according to Article 291 TFEU, which may be challenged before the Court. Having examined the relevant criteria, Article 25 of Regu‐ lation 2016/796 clearly qualifies as an implementing mechanism.

### 3.1.2.4. Soft mechanisms

While soft mechanisms, due to their lack of legally binding output, cannot possibly qualify as enforcement according to the above scheme (see in particular 3.1.1.2.4.), in order to apply its remaining distinguishing features to further practical examples – thereby enriching the test sample, as it were – they will nevertheless be addressed here. What is more, also soft mechanisms, even though they cannot qualify as enforcement measures, may interfere with the latter, and hence they deserve attention also in this context.

The compliance mechanism laid down in Article 121 TFEU (see IV.2.2.3.1.1. above) is focussed on the concretisation of the economic policies of the MS to the extent they shall be coordinated under Articles 120 ff TFEU (that is to say: to the extent they fall within the competence of the EU). It entails *ex ante*, on-going and partly also *ex post* intervention. In particular, the Council may address recommendations to a certain MS. The Commission initiates the procedure on its own motion. The procedure as laid down in Article 121 TFEU – unlike some of the secondary law based upon it – does not provide for sanctions to be imposed on MS. The mechanism displays a clear implementation tendency.

Article 148 para 4 TFEU provides for a mechanism on MS' compliance with the Council's guidelines for employment according to its para 2 (see IV.2.2.3.1.2. above). Each year the Council shall – *ex post* – examine to which extent the MS have complied with these guidelines, and make rec‐ ommendations to the MS, if appropriate. The focus of this regime is to monitor compliance and, if need be, to concretise the guidelines in the form of recommendations. Thus, the procedure is similar to Article 121 TFEU – to which it also has a material link.2355 It appears to facilitate the implementation of EU (soft) law.

Article 6 paras 5–7 of Regulation 2019/942 (see IV.2.2.3.2.1. above) allows for the ACER to scrutinise – *ex post* and only upon request – decisions of national regulatory authorities with a view to their compliance with network codes and guidelines referred to in Regulations 2019/943 and 715/2009 and in Directives 2019/944 and 2009/73/EC, or with any other relevant provision of these legal acts. In the resulting opinion, the ACER may also refer to further information or other components the decision at

<sup>2355</sup> The guidelines according to Article 148 para 2 TFEU shall be consistent with the BEPG under Article 121 TFEU.

issue should have contained. This particularly broad scope of the ACER's power (not only in terms of the EU law threshold of assessment, but in particular in view of all decisions of the national regulatory authorities concerned potentially being under scrutiny) clearly points towards enforce‐ ment. This qualification seems to be reflected in the fact that the ACER addresses the national authority concerned by a 'factual opinion', arguably determining instances of non-compliance rather than recommending alter‐ native action.<sup>2356</sup> Whereas, for lack of legal bindingness, its output cannot qualify as enforcement pursuant to the above scheme (see in particular 3.1.1.2.4.), it is also questionable whether this regime could qualify as imple‐ menting mechanism.

Under Article 53 of Directive 2019/944 (see IV.2.2.3.2.2. above), the Commission shall examine draft decisions of national authorities upon their respective request. This examination shall result in an opinion ad‐ dressed to the national authority concerned, of which the latter shall take 'utmost account' when adopting its (final) decision. The *ex ante* interven‐ tion suggests concretisation, and together with the relatively narrow scope of the Commission's examination this clearly points in the direction of implementation.

Article 33 of Directive 2018/1972 (see IV.2.2.3.2.3. above) provides for a regime in the course of which a certain category of draft measures to be taken by national regulatory authorities is scrutinised by the BEREC and the Commission in two phases. The scrutiny essentially is an *ex ante*/ on-going assessment, with only the Commission recommendation possibly being adopted after the adoption of the national decision. The Commission may initiate the procedure on its own motion. The purpose of the latter is to ensure compliance with EU law on the part of the national regulatory authority, namely by concretising it. All in all, the essential properties of the regime speak in favour of its implementing character.

Article 3 para 7 of Regulation 472/2013 (see IV.2.2.3.2.4. above) empow‐ ers the Commission to propose to the Council the adoption of recommen‐ dations to a Eurozone-MS (under enhanced surveillance) to take certain measures to do away with significant adverse effects on the financial sta‐ bility of the Euro area or of its MS which emanate from that MS. This mechanism clearly is about concretisation. The broad concept of measures

<sup>2356</sup> See also 2.4.1. above. This specific argument should not be overrated, however, because – as was shown above – the difference between recommendations and opinions in practice sometimes approaches zero; see III.3.1.1. and III.3.9. above.

aimed at tackling significant adverse effects on the financial stability of the Euro area or of its MS is defined more closely in the form of more concrete instructions. The broadness of the concept at issue here is to be distinguished from a broad quantitative scope of a compliance mechanism as addressed above (see 3.1.1.2.1.3.). The Council here does not ensure compliance with a large number of EU rules, but rather with one (vague) concept. In fact, the vagueness of the latter seems to render it inadequate for enforcement. After all, for a MS a variety of (apparently) adequate steps are available to tackle the significant adverse effects, which may, depending on the economists which are consulted, very well be heterogeneous. Thus, the Council recommends what it deems to be the most suitable from among these steps. While it is an *ex post* measure in the sense that it entails a reac‐ tion to the omission of a Eurozone-MS (to take the appropriate steps), the eminent focus on concretisation renders this procedure an implementation regime (in accordance with Article 121 TFEU; see above).

The mechanism laid down in Articles 16 f of Regulation 1092/2010 (see IV.2.2.3.2.5. above) to some extent is similar to the mechanism under Regulation 472/2013 just addressed. This is because the EU law concepts which are to be concretised in its course are relatively broad: significant risks to the stability of the EU's financial system and the adequate policy response thereto. In the given procedure, it is the ESRB which may adopt a recommendation to one or more MS or to one or more of the national supervisory authorities. Like the above procedure in Regulation 472/2013, this mechanism applies *ex post*. That the MS addressed needs to justify non-compliance emphasises the dialogic nature of this process. It suggests that it is not about establishing a wrongdoing on the part of a MS, but rather about supporting a MS in taking appropriate action. All in all, it has an implementing nature.

Eventually, the 'cooperation mechanism' pursuant to Article 6 of Regula‐ tion 2019/452 is to be addressed (see IV.2.2.3.2.6. above). It intervenes in on-going national proceedings (screening of one or more foreign direct investments). The interest to be protected by the mechanism is the security or public order of one or more MS which is – not least due to Regulation 2019/452 – also an objective of the EU. Again the concept to be defined more closely by the Commission, thereby taking account of the views of the MS (if uttered), is comparatively broad. The procedure is about concre‐ tisation (note the self-description as 'cooperation mechanism'), not about establishing a wrongdoing of a MS. Thus, it qualifies as implementing mechanism.

3.2. The primary legal bases of compliance mechanisms established through secondary law

# 3.2.1. Introduction

In Part III above, the explicit legal bases for the adoption of soft law as laid down in primary law were discussed (III.3.4., III.3.5., III.3.6.). In addition to that, the possibility of implicit competences to adopt soft law was addressed (III.3.3.). Bearing these findings in mind, we shall now shift the focus to the primary legal bases of compliance mechanisms, as provided for in secondary law, thereby complementing the classification of the above compliance mechanisms in terms of 'implementation' and 'enforcement' (3.1.). This means that we are not looking at the legal possibilities to adopt soft law in general, but at the legal bases on which the legislator has provi‐ ded for compliance mechanisms. In case of mixed and soft mechanisms, the underlying procedures (also) allow for the creation of soft law.

The mechanisms established by the Treaties themselves are, for that very reason, in compliance with Union law, rendering redundant a further examination in this respect. The question raised in the given context is whether the establishment of secondary law-based mechanisms – not only, as more generally addressed in Chapter III above, of the soft law parts of it (if any) – is covered by the respective Treaty base. This examination is to be performed not only at the level of the EU's competence (in German litera‐ ture referred to as *Verbandskompetenz*), but also at the level of different EU bodies (*Organkompetenz*; see III.3.2. above). The latter will also play a role when the EU's institutional balance is addressed (see 3.3. below).

# 3.2.2. A frequently used legal basis: Article 114 TFEU

# 3.2.2.1. Overview

The legal basis most frequently used for the establishment of secondary law-based compliance mechanisms (in our sample) is Article 114 TFEU. All kinds of compliance mechanisms have been based on this norm – hard, mixed, and soft ones, those presenting the Commission as the main actor and those providing for governance by European agencies or other bodies, such as the BEREC or the ESRB. For the adoption of Directive 2009/72/EC, the predecessor of Article 114 was used in conjunction with other competence clauses, namely with what are now Article 53 para 1 and Article 62 TFEU. The frequent use which is made of it and the vagueness of its wording, two main characteristics of Article 114 TFEU, are certainly interlinked in that the former is due to the latter. It should be emphasised once more that Article 114 TFEU itself, in its paras 4–6, makes provision for a compliance mechanism allowing for the Commission to address a deci‐ sion to a single MS. It is to be acknowledged, though, that these provisions deal with the special case of an exceptional deviation from a harmonisation measure. Therefore, the existence of this mechanism cannot be used as a general argument in favour of basing compliance mechanisms on Article 114 (para 1) TFEU.

Article 114 TFEU was described as a 'finale Querschnittskompetenz' [final cross-sectional competence],<sup>2357</sup> that is to say its scope is to be concretised by the legislator (and eventually by the CJEU) in each case of application, with a view to the objectives of the internal market. This 'concretisation' is required also in the context of other Treaty provisions, but – due to its malleable wording – it is certainly pronounced in the case of Article 114 TFEU. Acts established on the basis of Article 114 TFEU shall lead to the approximation of the provisions laid down by law, regulation or administrative action in MS which have as their object the establishment and functioning of the internal market.2358 And while Article 114 by no means allows for a general power to regulate the internal market,<sup>2359</sup> the legislator, according to the Court, has 'a discretion, depending on the gen‐ eral context and the specific circumstances of the matter to be harmonised, as regards the harmonisation technique most appropriate for achieving the desired result, in particular in fields which are characterised by complex technical features'.<sup>2360</sup>

<sup>2357</sup> Tietje, Art. 114 AEUV, para 126.

<sup>2358</sup> Case C-270/12 *United Kingdom v European Parliament/Council*, para 100. This is in accordance with the general doctrine of the CJEU, according to which 'the choice of the legal basis for a legal act of the Union must rest on objective factors amenable to judicial review, which include *the aim and content of the measure*' (emphasis added); case C‑589/15P *Anagnostakis*, para 67, with further references.

<sup>2359</sup> See Moloney, Rule-Making 70. The introduction of a general power to regulate the internal market was discussed during the preparation of the SEA; see Streinz, Europarecht (9th edn) para 976.

<sup>2360</sup> Case C-66/04 *United Kingdom v European Parliament/Council*, para 45; see also case C‑358/14 *Poland v European Parliament/Council*, para 33, with many further references. In terms of the proportionality principle, the Court – in the context of Article 114 TFEU – applies its general standard of review, according to which a measure is unlawful 'only if [it] is manifestly inappropriate having regard to

A legal act has as its object the establishment and functioning of the internal market where 'it is actually and objectively apparent from th[is] legal act that its purpose is to improve the conditions of the establishment and functioning of the internal market'.<sup>2361</sup> A legal act based on Article 114 TFEU may, however, also contain measures 'for contributing to the im‐ plementation of a process of harmonisation' where they are 'closely linked to the subject matter of the acts approximating the laws, regulations and administrative provisions of the Member States. Such is the case in particu‐ lar where [a Union body] provides services to national authorities and/or operators which affect the homogenous implementation of harmonising instruments and which are likely to facilitate their application'.<sup>2362</sup>

This is particularly relevant for compliance mechanisms because their provision does not approximate the laws itself. However, they may be aimed at MS' compliance with approximating rules, thereby contributing to the implementation of a process of harmonisation. Due to their compliance-en‐ hancing function the existence of a close link to the subject matter of the respective legal act (as required by the Court) is at least probable; for our sample of compliance mechanisms based on Article 114 TFEU the existence of a close link will be tested under 3.2.3. below.

The Court stresses that even where certain measures are not 'aimed directly at improving the conditions for the functioning of the internal market', but whose 'purpose [it] is to ensure that certain prohibitions con‐ cerning the internal market and imposed in pursuit of that object are not circumvented', they may be adopted on the basis of Article 114.2363 In other words, the Court allows for the establishment of a regime ensuring compli‐ ance with the EU rules at issue, thereby 'completing' their approximating effect. Scholars have argued that in this context the above criteria must be interpreted as requiring an urgent necessity for a uniform application throughout the EU.2364 This necessity shall be examined below case by case with a view to the concrete policy field.

the objective which the competent institution is seeking to pursue'; case C-491/01 *British American Tobacco*, para 123. For a condensed and critical account of the manifest error case law of the Court see Craig, Administrative Law 472–474.

<sup>2361</sup> Case C-66/04 *United Kingdom v European Parliament/Council*, para 44; also note the wording in case C-398/13P *Inuit*, para 26: 'genuinely [aim] to improve the conditions for the establishment and functioning of the internal market'.

<sup>2362</sup> Case C-217/04 *United Kingdom v European Parliament/Council*, paras 44 f.

<sup>2363</sup> Case C-491/01 *British American Tobacco*, para 82.

<sup>2364</sup> See references in Michel, Gleichgewicht 120.

# 3.2.2.2. An appropriate legal basis for compliance mechanisms?

More specifically with regard to compliance mechanisms, the Court held – in the context of (meanwhile outdated) Regulation 460/2004, establish‐ ing, among other things, the European Network and Information Security Agency (ENISA<sup>2365</sup>) – that '[t]he legislature may deem it necessary to provide for the establishment of a Community body responsible for contri‐ buting to the implementation of a process of harmonisation in situations where, *in order to facilitate the uniform implementation and application of acts based on that provision*, the adoption of *non-binding supporting and framework measures* seems appropriate' (emphases added).<sup>2366</sup> The meas‐ ures which the Court referred to here were laid down in the ENISA's found‐ ing act, Regulation 460/2004.<sup>2367</sup> According to Article 10 of this Regulation, the ENISA shall provide (individual) advice and assistance falling within its scope, objectives and tasks, among others to national authorities.<sup>2368</sup> A soft compliance mechanism, on the part of the EU actor involved, even if not explicitly foreseen in this provision, would not go beyond these tasks and powers. It is therefore justified to assume that the Court here has in principle (and only implicitly) approved of the possibility to establish soft compliance mechanisms on the basis of Article 114 TFEU.2369 The Court has also confirmed the close link to the basic act (as mentioned above), as ENISA 'provides services to national authorities and/or operators which af‐ fect the *homogenous implementation of harmonising instruments and which are likely to facilitate their application*' (emphasis added).<sup>2370</sup>

<sup>2365</sup> Now referred to – under the same abbreviation – as European Union Agency for Cybersecurity; see <https://www.enisa.europa.eu/about-enisa> accessed 28 March 2023.

<sup>2366</sup> Case C-217/04 *United Kingdom v European Parliament/Council*, para 44; critical‐ ly: Ottow/van Meerten, Proposals 24. That material competences (eg Article 114 TFEU) also include certain organisational competences has been accepted even before the *ENISA* case, but the extent of the organisational competences was unclear; see Berger, Einrichtungen 62 f.

<sup>2367</sup> This Regulation meanwhile has been replaced first by Regulation 526/2013, then by Regulation 2019/881.

<sup>2368</sup> See case C-217/04 *United Kingdom v European Parliament/Council*, para 64; see now Recital 55 and Article 4 paras 1f of the successor Regulation 2019/881.

<sup>2369</sup> Critically as regards the conferral of advisory powers on the basis of this provision: Adamski, ESMA 816.

<sup>2370</sup> Case C-217/04 *United Kingdom v European Parliament/Council*, para 45; critical‐ ly: Opinion of AG *Kokott* in case C-217/04 *United Kingdom v European Parlia‐ ment/Council*, para 46; see also references made in Görisch, Verwaltung 240 f; for

While, in general, the extensive interpretation of an already vague provi‐ sion (Article 114 TFEU) beyond its obvious purpose – here the approxima‐ tion of national law, regulation or administrative action by EU law – may be criticised, it ought to be stressed that also a focus on verbal interpretation may provide an argument in favour of the Court's far-reaching case law. After all, Article 114 para 1 TFEU does not simply empower the legislator to approximate national rules relating to the internal market, but to adopt 'the measures for the approximation' of them.<sup>2371</sup> This may be understood as including legislative measures which do not themselves approximate national rules, but which provide for mechanisms the application of which (by Union bodies) leads to the approximation of national rules – or the approximation of the MS' application of Regulations based on that provi‐ sion<sup>2372</sup> – by providing a uniform concretisation of the pertinent (superior) Union law.2373 In the words of AG *Kokott*: The measure 'can provide for procedures which do not bring about approximation directly but only in a multi-stage model with intermediate steps'.<sup>2374</sup>

In conclusion, the Court has confirmed the feasibility of establishing soft mechanisms on the basis of Article 114 TFEU. <sup>2375</sup> Since empowering a newly established agency in this context is lawful, *a fortiori* this applies where the Commission is in charge. The widely drafted reference of the Court to 'supporting and framework measures' aimed at the 'uniform im‐ plementation and application of acts based on Article 114 TFEU' appears to indicate that both general-abstract and individual-concrete (soft law) measures are addressed.<sup>2376</sup>

2372 The adoption of Regulations on the basis of Article 114 TFEU has been considered problematic for their general direct applicability. However, the wording of this provision clearly (and in particular as compared to Article 115 TFEU) does not exclude the adoption of Regulations and by now a number of them has been adopted on Article 114 TFEU, partly even with the explicit approval of the Court; see Classen, Art. 114 AEUV, paras 125 f.

the role implied (annexed) powers play in the context of Article 114 TFEU see Orator, Möglichkeiten 214.

<sup>2371</sup> See also case C-359/92 *Germany v Council*, para 37.

<sup>2373</sup> See case C-66/04 *United Kingdom v European Parliament/Council*, para 59.

<sup>2374</sup> Opinion of AG *Kokott* in case C-66/04 *United Kingdom v European Parlia‐ ment/Council*, para 33.

<sup>2375</sup> The Court's judgement in the *ENISA* case does not constitute a legitimation of mixed/hard compliance mechanisms, though; see also Moloney, Rules in Action 219.

<sup>2376</sup> Case C-270/12 *United Kingdom v European Parliament/Council*, para 44.

Whether also mixed or even hard compliance mechanisms may lawfully be adopted on the basis of Article 114 TFEU remains to be considered. With reference to the above case law, the Court, in its judgement in the *ESMA* case, added that 'EU legislature, in its choice of method of harmonisation and, taking account of the discretion it enjoys with regard to the measures provided for under Article 114 TFEU, may delegate to a *Union body, office or agency powers for the implementation of the harmonisation sought*. That is the case in particular where the measures to be adopted are dependent on *specific professional and technical expertise* and the ability of such a body to *respond swiftly and appropriately*' (emphases added).<sup>2377</sup> While the Court specifically referred to the ESMA's powers to address legally binding acts to individuals/undertakings (Article 28 of Regulation 236/2012), it is to be noted that this Regulation also empowers the ESMA to adopt an opinion to the MS to ensure compliance.2378 Only where this opinion is not complied with, the ESMA may make use of its powers under Article 28 *leg cit* (*ultima ratio*).<sup>2379</sup>

The competence to adopt legally binding measures *vis-à-vis* MS – which is a necessary condition of most mixed and hard compliance mechanisms addressed here – appears, in this context at least, to be less problematic than the power to directly address individuals/undertakings.<sup>2380</sup> After all, it is the rule rather than the exception that it is MS who are addressed by

<sup>2377</sup> Case C-270/12 *United Kingdom v European Parliament/Council*, para 105. In terms of the level of detail, the Court required that the 'essential elements of the harmo‐ nising measure' are defined in the legislative act at issue and that the 'mechanism for implementing those elements [is] designed in such a way that it leads to a harmonisation within the meaning of Article [114 TFEU]'; case C-66/04 *United Kingdom v European Parliament/Council*, paras 48 f. The Court said so in the context of the Commission. These requirements must also (or even: *a fortiori*) apply where mere bodies, offices or agencies of the EU are empowered; overall sceptically: W Weiß, Future 345.

<sup>2378</sup> Article 27 of Regulation 236/2012; for the implementing powers of European agencies see already case T-187/06 *Schräder* (implicit confirmation of the CPVO's individual decision-making power); confirmed in case C-38/09P *Schräder*.

<sup>2379</sup> See case C-270/12 *United Kingdom v European Parliament/Council*, paras 108 ('where necessary') and 115 ('not taken sufficient measures'). The *ultima ratio* character of the measures directed to a market participant already follows from the pertinent Regulation 236/2012 (in particular its Recital 33 and its Article 28 para 3, last subparagraph *e contrario*). The ESMA opinion which is regularly addressed to the competent national authority in advance pursuant to Article 27 para 2 *leg cit* underpins the hierarchy of this sequence of acts.

<sup>2380</sup> This conviction is exemplified by the sequenced procedure laid down in Article 17 of Regulation 1093/2010 (see IV.2.2.2.2.3. above).

measures adopted on the basis of Article 114 TFEU or by measures adopted (by an EU body) on the basis of EU secondary law (which is again based on Article 114 TFEU).<sup>2381</sup> Such constellations lead to an influence on MS action, but they grant the latter some leeway in executing EU law *vis-à-vis* individuals/undertakings.<sup>2382</sup> The power of EU bodies to address legally binding measures to individuals/undertakings, on the contrary, regularly *replaces* MS competence/action.<sup>2383</sup> On the other hand, the possibility of EU bodies giving instructions to national authorities, thereby creating a hi‐ erarchy of these two levels of administration is not reflected in the Treaties, whereas pre-emptive action by EU bodies – at the cost of MS action – has a certain tradition in EU law.<sup>2384</sup>

Thus, from the perspective of EU law, each of the two *modi* has its respective *pro* and *contra*. Specifically with regard to Article 114 TFEU as interpreted by the Court, however, we can conclude: If the power to adopt individual-concrete measures binding upon individuals or undertakings may, under certain circumstances, be feasible, due to the fact that measures based on Article 114 TFEU are normally directed to the MS this must also be true for the power to adopt legally binding individual-concrete measures addressed to one or a number of MS.

In the context of another compliance mechanism involving the adoption of hard law acts,2385 the Court has expressly confirmed compliance of the mechanism with Article 100a TEEC (today Article 114 TFEU): 'In certain fields, and particularly in that of product safety, the approximation of general laws alone may not be sufficient to ensure the unity of the

<sup>2381</sup> See case C-217/04 *United Kingdom v European Parliament/Council*, para 44.

<sup>2382</sup> Still critical with regard to such constellations: Gundel, Energieverwaltungsrecht, para 47.

<sup>2383</sup> This is also the logic behind tiered compliance mechanisms like the one underly‐ ing the *ESMA* case. Another example, taken from our sample, is Article 19 of Regulation 1093/2010; more generally on the problem of the power of EU bodies to 'lift implementation powers' of the MS: Opinion of AG *Jääskinen* in case C-270/12 *United Kingdom v European Parliament/Council*, para 50; critically in the context of Article 291 TFEU: Nettesheim, Art. 291 AEUV, para 31.

<sup>2384</sup> See Schütze, Rome 1404 f, with further references. It has been argued by many scholars – under previous Treaty versions – that a power of the Commission or the Council to address legally binding instructions to national authorities would be *ultra vires*: Constantinesco, Recht 299; Eekhoff, Verbundaufsicht 130–139; Gil Ibáñez, Supervision 205; Scheuing, Impulse 334–336; Schöndorf-Haubold, Verwaltung 46; von Bogdandy/Arndt/Bast, Instruments 96; see also Kahl, Verwal‐ tungsverbund 366**.**

<sup>2385</sup> See Article 9 of Directive 92/59/EEC, the predecessor act of Directive 2001/95/EC.

market. Consequently, the concept of "measures for the approximation" of legislation must be interpreted as encompassing the Council's power to lay down measures relating to a specific product or class of products and, if necessary, individual measures concerning those products'.<sup>2386</sup>

It should not be omitted here to remark that the Court's approach is not uncontested. One of the main issues in this context is whether the power to adopt individual-concrete measures, be they addressed to individ‐ uals/undertakings or to MS, can be qualified as *approximation of laws*. In the literature and in legal practice a strong opposition to the individualised application of EU law by an EU body on the basis of Article 114 TFEU can be found.<sup>2387</sup> Suffice it to quote AG *Jacobs* here, who uttered: 'It is one thing to lay down rules which must be uniformly applied in all Member States, another to take the decisions which apply the rules to individual cases. It is clear that, under certain provisions of the Treaty, the Council may delegate to the Commission both the power to lay down rules and the power to take individual decisions.[] Article 100a, in contrast, is concerned exclusively with the harmonisation of national provisions. It follows that Article 100a may be used only to adopt measures which lay down uniform rules; the application of those rules to individual cases is then a matter for the national authorities'.<sup>2388</sup>


<sup>2386</sup> Case C-359/92 *Germany v Council*, para 37; see also para 39; repeated in case C-217/04 *United Kingdom v European Parliament/Council*, para 106; see also over‐ view of relevant case law by Schütze, Rome 1393 ff. Still cautiously: Commission Proposal COM(97) 619 final, para 11: 'Conferring on the Commission the power to take a Decision requesting a Member State to take rapid and appropriate measures to remove an obstacle to trade is necessary if one of the objectives of the Community is to be attained […] and therefore the proper functioning of the internal market […]. Besides, conferral of this power is not, directly or indirectly, associated with harmonization, within the meaning of Article 100a of the Treaty. The purpose of the Regulation is action by the Commission which does not call into question the laws, regulations and administrative provisions of the Member States as such'. That is why the Commission proposed the flexibility clause – not Article 100a TEC – as the correct legal basis for the respective act.

### 3.2.3. The compliance mechanisms based on Article 114 TFEU

#### 3.2.3.1. Introduction

Against the background of the above analysis in particular of the Court's case law, in the following we shall examine those compliance mechanisms of our sample which were based on Article 114 TFEU with a view to their compliance with this provision, in particular with a view to whether there is a close link to the subject matter of the respective legal act, as required by the Court, and with a view to their respective proportionality.<sup>2389</sup>

The stated objectives of the acts of secondary law based on Article 114 TFEU and containing the compliance mechanisms presented above are the safety of different kinds of products,<sup>2390</sup> to set up an efficient and effective single European framework for the resolution of entities and ensuring the consistent application of resolution rules,2391 to provide a harmonised framework for cross-border exchanges of electricity, <sup>2392</sup> to provide a system that is in line with the objective of a stable and single Union financial mar‐ ket for financial services,2393 the creation of a fully operational internal elec‐ tricity market,2394 the participation and cooperation of national regulatory authorities in order to facilitate the uniform application of the legislation on the internal markets for electricity and natural gas throughout the Un‐ ion,2395 to achieve a harmonised framework for the regulation of electronic communications services, electronic communications networks, associated facilities and associated services,2396 and an effective macro-prudential oversight of the Union financial system. <sup>2397</sup> The examination of these acts

<sup>2389</sup> The safeguard clauses according to what is now Article 114 para 10 TFEU (see 2.2.1.2.6. above) shall not be addressed here, because their legal basis (para 10) is much more specific than para 1 *leg cit*, and hence potential incompatibilities would be more apparent. The three examples provided above seem to be in accordance with Article 114 para 10 TFEU (and the respective predecessor provisions).

<sup>2390</sup> Directive 2001/95/EC; Directive 2001/18/EC; Directive 2006/42/EC; Regulation 1907/2006; for product safety as an early example of a policy field with an EU-MS network to supervise compliance with and enforce EU law see Gil Ibáñez, Supervi‐ sion 298 f.

<sup>2391</sup> Regulation 806/2014 (Recital 122).

<sup>2392</sup> Regulation 2019/943 (Recital 74).

<sup>2393</sup> Regulation 1093/2010 (Recital 8).

<sup>2394</sup> Directive 2009/72/EC (Recital 62).

<sup>2395</sup> Regulation 2019/942 (Recital 16).

<sup>2396</sup> Directive 2002/21/EC (Recital 41).

<sup>2397</sup> Regulation 1092/2010 (Recital 33).

shall not be comprehensive, but shall be limited to the compliance mechan‐ isms referred to above, based on the assumption that these acts all comprise measures for the approximation of MS' law, regulation or administrative action (as interpreted broadly by the Court; see 3.2.2.2. above), and based on the assumption that the objectives *stated* in these acts – which all serve the achievement of the internal market as defined in Article 26 TFEU, in particular: the facilitation of at least one of the fundamental freedoms – are their respective *actual* objectives.<sup>2398</sup> The follow-up question is whether this holds true also for the single compliance mechanisms enshrined in these legislative acts, that is to say whether they are at least 'closely linked to the subject matter of the acts approximating the laws', as required in the case law referred to above.

### 3.2.3.2. Hard and mixed mechanisms

The mechanism laid down in Article 13 para 1 of Directive 2001/95/EC is restricted in different ways. First, it may only be applied in case of a serious risk from certain products to the health and safety of consumers. In addition to that, the following criteria must be met: The MS' approaches to deal with this risk differ significantly from each other, the risk cannot be dealt with otherwise due to the urgency of the matter, and the risk can be eliminated effectively only by adopting measures applicable at Union level in order to ensure a consistent and high level of protection of the health and safety of consumers and the proper functioning of the internal market (*ulti‐ ma ratio*).2399 Second, if scientific questions falling within the competence of an EU scientific (comitology) committee arise, the Commission must consult this committee. Third, MS' competences are not skipped by the Commission decision, but – and that, in comparison, appears as a weaker form of interference with MS' prerogatives – the Commission is requiring a MS to exercise these competences in a certain way. In view of that, there seems to be a close link to the subject matter of Directive 2001/95/EC,<sup>2400</sup> fleshed out in a proportionate way<sup>2401</sup> . Note in particular that – with regard

<sup>2398</sup> For the importance of these stated objectives see eg case C-270/12 *United Kingdom v European Parliament/Council*, para 53.

<sup>2399</sup> Article 13 para 1 lit a-c of Directive 2001/95/EC.

<sup>2400</sup> See case C-359/92 *Germany v Council*, para 35.

<sup>2401</sup> See case C-359/92 *Germany v Council*, paras 44 ff.

to the predecessor mechanism of Article 13 of Directive 2001/95/EC – the Court has confirmed the necessity of the procedure, in particular in view of the urgency required.<sup>2402</sup>

Under the mechanism laid down in Article 29 para 2 of Regulation 806/2014, the SRB may order an institution under resolution to take cer‐ tain action to comply with an earlier SRB decision which the national resolution authority in charge has not respected. Decisions of EU bodies directly addressed to market participants are, especially where there is an explicit competence of the MS in this respect, considered to be exceptions, sometimes legally problematic exceptions. In this case, the interference with the MS competence by the SRB follows from the breach of an earlier SRB decision on the part of the competent national authority. The decision is obligatory also for the national authority concerned and shall only be taken if it 'significantly addresses the threat' at issue.2403 Before adopting a decision addressed to an institution under resolution, the SRB shall notify the national resolution authority concerned and the Commission no later than 24 hours in advance.2404 This very short notice has to be seen in the context of the urgency of the matter.2405 It reduces to a minimum the possibility of an exchange of views between the EU and the national level. In terms of judicial protection, the MS concerned is not worse off than if the SRB had addressed its decision to the national authority. In both cases, it may file an action for annulment with the Court.2406 The mechanism is strongly attached to the rest of Regulation 806/2014 and clearly serves the 'efficiency and uniformity of resolution action'.2407 However, the strong enforcement tendency of this mechanism (already addressed under 3.1.2.2. above) renders it doubtful whether Article 114 TFEU is a sufficient legal basis for it.

Another compliance mechanism of Regulation 806/2014 is laid down in its Article 7 para 4. It allows the SRB 'to exercise directly all of the rele‐

<sup>2402</sup> Case C-359/92 *Germany v Council*, para 50.

<sup>2403</sup> Article 29 paras 2 and 4 of Regulation 806/2014.

<sup>2404</sup> Note that a request pursuant to Article 29 para 2 lit c of Regulation 806/2014 is subject to further conditions.

<sup>2405</sup> In the course of the negotiations on what became Regulation 806/2014, the ECB and the Commission have stressed the importance of fast-track decision-making in bank crisis management; see Howarth/Quaglia, Road 133. Considering the necessity of urgent decision-making: case T-510/17 *Del Valle Ruíz*, para 414.

<sup>2406</sup> Article 263 para 2 TFEU.

<sup>2407</sup> Recital 87 of Regulation 806/2014; see also Alexander, Banking Union 178, with further references.

vant powers under this Regulation', either after a warning to the national authority concerned has not been reacted to accordingly or even without a predating warning. But also in the latter case consultations or a request of the national authority concerned must precede. Where neither a warning has been adopted nor a request filed by the national authority concerned, the proportionality of the direct intervention is highly questionable and may only be justified by danger ahead in the concrete case. The Regulation does not make provision for such a restriction of short-term intervention. For the enforcement thrust of this mechanism – which reflects the propor‐ tionality concerns uttered here – see 3.1.2.3. above.

The mechanisms laid down in Regulation 1093/2010,<sup>2408</sup> more precisely the '"low-level" enforcement powers'<sup>2409</sup> provided for in its Articles 17–19, shall be addressed together. Article 17 is a general (mixed) mechanism aimed at ensuring compliance with the relevant EU law. It allows for an *ex post* scrutiny of the actions of the competent authorities. It encompasses all kinds of action and the threshold against which this action is to be exam‐ ined covers the relevant financial market law.2410 It is carefully drafted, apart from the action taken by the EBA also allowing for Commission interven‐ tion. It provides for the possibility of the EBA decision addressed directly to a financial institution/financial sector operator which shall only be adopted 'where it is necessary to remedy, in a timely manner, such non-compliance in order to maintain or restore neutral conditions of competition in the market or ensure the orderly functioning and integrity of the financial system'2411 (as an *ultima ratio* measure2412). While there are concerns as to the powers the EBA is vested with all in all,2413 also the mechanism laid down in Article 17 seems to be problematic, in particular with a view to its proportionality. Unsurprisingly, this assessment is related to the strong enforcement tendency of this mechanism (as established above, see 3.1.2.3.),


<sup>2408</sup> While the discussion here shall focus on the compliance mechanisms, not on the entire legal act by means of which it is established, it ought to be noted that Regulation 1093/2010 in its entirety much more caters for the harmonisation of law (by the EBA) than it harmonises the law itself; for an examination of the Regulation 1093/2010 against its legal basis see Fahey, Emperor.

<sup>2409</sup> Fahey, Emperor 589.

<sup>2410</sup> See Article 1 para 2 of Regulation 1093/2010.

<sup>2413</sup> Critically with regard to the appropriateness of Article 114 TFEU for the ESMA's (the EBA's sister authority) regulatory powers: Moloney, Rule-Making 71.

for which it is doubtful whether Article 114 TFEU provides an appropriate basis.<sup>2414</sup>

As regards the procedure laid down in Article 18 of Regulation 1093/2010, it is clear that it is an emergency mechanism which requires – for it to be applied – a Council decision<sup>2415</sup> establishing the existence of an emergency situation and 'exceptional circumstances, where coordinated action by com‐ petent authorities is necessary to respond to adverse developments which may seriously jeopardise the orderly functioning and integrity of financial markets or the stability of the whole or part of the financial system in the Union'.<sup>2416</sup> Also here the EBA decision directly addressed to financial institutions is a measure of last resort. It applies only where a competent authority has failed to comply with a decision from the EBA addressed to it. Again the route via influencing the competent authority (the national authority and potentially also the ECB2417) is preferred, and only where it does not lead to the aspired result – compliance with EU law on the part of the competent authority that is – the direct way to the respective market actor is available. Here the exceptionality of the application of this mecha‐ nism (reflected also in the requirement of a Council decision establishing an emergency situation) mitigates the concerns on the primary legal basis uttered above with regard to the Article 17-procedure.

In Article 19 of Regulation 1093/2010 the sequence of (possible) acts corresponds to Article 18. The Article 19-mechanism does not require a Council decision to be activated, though. It is about settling disagreements between competent authorities and hence means that also here the EBA intervenes only where the uniform application of the relevant Union law is at risk. While normally the EBA may only get involved upon request by at least one of the relevant competent authorities, provision is made for action by the EBA *proprio motu*, namely in cases specified in the relevant legislation and where, on the basis of 'objective reasons', disagreement

<sup>2414</sup> See also Gil Ibáñez, Exceptions 170 and 173, who mentions as legal bases for new enforcement mechanisms either Art 352 TFEU or a new provision introduced in the course of a Treaty revision.

<sup>2415</sup> Note that pursuant to the pertinent Commission proposal the Commission should have been in charge here (instead of the Council); see Article 10 of Commission Proposal COM(2009) 501 final.

<sup>2416</sup> Article 18 paras 2 f of Regulation 1093/2010; with regard to emergency measures requiring MS to take certain action see case C-359/92 *Germany v Council*, para 33.

<sup>2417</sup> Article 2 para 2 lit f of Regulation 1093/2010.

between competent authorities from different MS can be determined.<sup>2418</sup> Since disagreements between national authorities are likely to constitute obstacles to the functioning of the internal market,<sup>2419</sup> and since such disagreements are (from *an ex* ante perspective) not unlikely to occur,<sup>2420</sup> there seems to exist a sufficiently strong link to the material scope of Article 114 TFEU.

# 3.2.3.3. Soft mechanisms

The mechanism set out in Article 33 of Directive 2018/1972 provides for the scrutiny of planned MS measures *vis-à-vis* operators – that means, in this context, undertakings providing or authorised to provide a public electronic communications network or an associated facility.2421 The Com‐ mission may notify the national authority concerned (and the BEREC) when it deems the draft measure to 'create a barrier to the single market or [when it has] serious doubts as to its compatibility with Union law'.2422 The BEREC may issue an opinion where it agrees with the Commission. If not, the Commission may adopt a recommendation to the national authority concerned to amend or withdraw its decision, or it may decide to lift its reservations against the draft. Article 33 of Directive 2018/1972 is entitled 'Procedure for the consistent application of remedies' – a procedure which was considered necessary (and suggested by a study2423) to ensure Commis‐ sion influence on the adoption of remedies by national authorities.2424 In the relevant Commission proposal, a hard law power for the Commission was envisaged which – in the course of the legislative procedure – was

<sup>2418</sup> Article 19 para 1 subpara 2 of Regulation 1093/2010.

<sup>2419</sup> See Recital 32 of Regulation 1093/2010: 'to ensure the correct and consistent application of Union law'.

<sup>2420</sup> See Report of the High-Level Group on Financial Supervision in the EU (February 2009; so-called *de Larosière* Report) 77.

<sup>2421</sup> Article 2 para 29 of Directive 2018/1972.

<sup>2422</sup> Article 33 para 1 of Directive 2018/1972. For the comments the Commission may make even if it does not object to the draft measure see subpara 2 *leg cit*.

<sup>2423</sup> See Commission Proposal COM(2007) 697 final, 5.

<sup>2424</sup> For the remedies to be applied see references in Article 33 para 1 of Directive 2018/1972; for a hierarchisation of these remedies see Article 14 para 3 *leg cit*; for the 'schleichende[n] Machtzuwachs' [creeping increase of power] of the Commis‐ sion in EU network regulation more generally see Ludwigs, Netzregulierungsrecht 608, with further references.

converted into a mere soft law power.<sup>2425</sup> In the fast-moving area of tele‐ communications the Treaty infringement procedure has proved particularly fragile<sup>2426</sup> – an argument which may count in favour of the necessity (as part of the proportionality test) of the compliance procedure(s) laid down in Directive 2018/1972. With a view to Article 114 TFEU, the mechanism – due to its softness, but also due to its focus on the approximation of the administrative action of MS (close link to the general purpose of Directive 2018/1972) and the *ex ante* intervention it provides – seems feasible.

According to the procedure laid down in Articles 16 f of Regulation 1092/2010, the ESRB may issue recommendations – *inter alia* – to MS or their respective supervisory authorities, asking for a policy response on the part of the addressees. It may do so where significant risks to the achievement of the objective of, to put it briefly, maintaining the stability of the financial system of the EU as stipulated in Article 3 para 1 of Regulation 1092/2010 arise. Non-compliance may lead to further inter‐ action and may, as a measure of last resort, result in the publication of the recommendations.2427 In spite of its limitation to cases of risk (as set out above), the mechanism still leaves a margin of appreciation and a broad scope of action for the ESRB. Practice suggests that the adoption of these recommendations is hardly exceptional – to date (and since its establishment by 1 January 2011) the ESRB has adopted dozens of these recommendations, some of which are addressed to specific MS.2428 The adoption of recommendations in principle is in accordance with the main objective of Regulation 1092/2010, namely to establish EU macro-prudential oversight of the financial system. However, where the recommendations have a strong enforcement thrust (eg requesting the abolishment of certain administrative decisions *ex post* due to their alleged infringement upon EU law) or where they request a concrete legislative initiative of a certain MS this – not least in view of the potentially broad scope of these recommenda‐ tions – appears to be problematic; in the former case because enforcement other than by the Treaty infringement procedure and its variants is regu‐

<sup>2425</sup> See Recital 11 of Commission Proposal COM(2007) 697 final, which lead to the adoption of Directive 2009/140/EC amending – *inter alia* – Directive 2002/21/EC.

<sup>2426</sup> See Kühling, Telekommunikationsrecht, para 73, with an example.

<sup>2427</sup> The ESRB Secretariat has published a 'Handbook on the assessment of compliance with ESRB recommendations' (2016) <https://www.esrb.europa.eu/pub/pdf/reco mmendations/160502\_handbook.en.pdf> accessed 28 March 2023.

<sup>2428</sup> See <https://www.esrb.europa.eu/mppa/recommendations/html/index.en.html> accessed 28 March 2023.

larly unlawful, in the latter case because of the interference with national sovereignty.<sup>2429</sup>

With a view to Article 114 TFEU, the EU legislator expressly refers to the case C-217/04 *United Kingdom v European Parliament/Council*, stressing the close link of the ESRB's tasks 'to the objectives of the Union legislation concerning the internal market for financial services'.<sup>2430</sup>

# 3.2.4. Other primary legal bases

3.2.4.1. Hard mechanisms

Article 100 para 2 TFEU empowers the EP and the Council to lay down, pursuant to the ordinary legislative procedure and thereby consulting the Economic and Social Committee and the Committee of the Regions, 'ap‐ propriate provisions for sea and air transport'. While the rest of the TFEU's title 'Transport' does not apply to sea and air transport,2431 Article 100 para 2 provides for regulatory action on the part of the EU in these fields. In spite of this peculiarity, systematically speaking Article 100 para 2 belongs to the title 'Transport', which is why Articles 56 f TFEU do not apply.<sup>2432</sup> The other provisions of primary law do apply to sea and air transport. This includes in particular the internal market objective pursuant to Article 26 TFEU, the other fundamental freedoms and the competition rules. On the level of secondary law, also the freedom of services has been realised in the field of transport. In this context, also the 'additional aim' of Regulation 2018/1139 – the founding act of the EASA which is based on Article 100 para 2 TFEU – is to be mentioned, namely to 'facilitate the free movement of goods, persons, services and capital'.<sup>2433</sup>

The question now is whether Article 100 para 2 TFEU may serve as a legal basis for the compliance mechanism laid down in Articles 70 f of

2430 Recital 31 of Regulation 1092/2010.

<sup>2429</sup> Article 16 para 1 of Regulation 1092/2010: 'recommendations […] for legislative initiatives'. For these concerns see also 3.3.3. below.

<sup>2431</sup> Article 100 para 1 TFEU.

<sup>2432</sup> Article 58 para 1 TFEU: *argumentum* 'transport'; see eg Fehling, Art. 100 AEUV, para 10; for the nevertheless applicable 'principle of the freedom to provide services' see case C-92/01 *Stylianakis*, paras 23 f, with further references.

<sup>2433</sup> Article 1 para 2 lit d of Regulation 2018/1139; with regard to the importance of one of the predecessor regulations for the internal market programme of the then Community see Riedel, Gemeinschaftszulassung 4.

Regulation 2018/1139. The title on transport contained in the TFEU refers to executive functions of the Commission (Article 95 para 4 and Article 96 para 2). Even though these provisions do not apply in the context of sea and air transport, the author would agree with *Riedel* who argues that – in view of the systematic belonging of what is now Article 100 para 2 TFEU to EU transport policy – executive functions on the part of the EU cannot be excluded.<sup>2434</sup> The EASA's powers under its founding regulation have been described as considerable and – in the context of European agencies – as unprecedented.<sup>2435</sup> However, 'the contorted way in which the regulatory powers have been granted, and the multiple controls to which those powers are subjected, show that it was not intended to grant them a clear hierarch‐ ical authority over their national counterparts'.<sup>2436</sup> Here it is not the full *pouvoir* of the EASA which is at issue, but the tasks and powers related to ensuring MS compliance which the EASA and the Commission are vested with pursuant to Articles 70 f of Regulation 2018/1139. Article 100 para 2 TFEU entitles the legislator to lay down 'appropriate provisions' for sea and air transport.2437 As regards the modalities – in particular: the output the Commission or newly established EU bodies may be empowered to adopt – which may be covered by these provisions, Article 100 para 2 TFEU appears to be (even) more encompassing than Article 114 para 1 TFEU. In view of that, it seems adequate to analogously consider the Court's case law adopted in the context of Article 114 TFEU (see 3.2.2.2. above),2438 both as regards the empowerment of the Commission and the establishment and empowerment of the EASA. While here only the powers specified above are at issue, the legality of the establishment of the EASA is an important preliminary question.2439 One of the tasks of the EASA is to contribute to

<sup>2434</sup> See Riedel, Gemeinschaftszulassung 245.

<sup>2435</sup> See Dehousse, Delegation 790, with regard to the EASA's *de facto* regulatory autonomy.

<sup>2436</sup> Curtin/Dehousse, Agencies 195, with regard to the predecessor regulation; in agreement with them on this point: Opinion of AG *Jääskinen* in case C-270/12 *United Kingdom v European Parliament/Council*, para 24.

<sup>2437</sup> For this broadly drafted Treaty provision and the competences it may possibly confer see already Priebe, Entscheidungsbefugnisse 81–83; see also Riedel, Ge‐ meinschaftszulassung 248–250.

<sup>2438</sup> See case T-317/09 *Concord*, paras 46 f for an exemption mechanism laid down in Article 22 (in particular its para 4) of Directive 2003/55/EC, based *inter alia* on what is now Article 114 TFEU, which is similar to that laid down in Articles 70 f of Regulation 2018/1139.

<sup>2439</sup> See eg Orator, Möglichkeiten 468–470.

the uniform application of Regulation 2018/1139.<sup>2440</sup> While this objective corresponds to the (wide) wording of Article 100 para 2 TFEU, the content of the regimes laid down in Articles 70 f of this Regulation – on the moni‐ toring of national safeguard measures and the handling of legal exemptions, respectively, each time providing for a Commission decision adopted upon an EASA recommendation – appears to serve this aim in a proportionate manner.

Regulation 2019/943 is based on Article 194 para 2 TFEU which allows for the EP and the Council, acting in accordance with the ordinary le‐ gislative procedure, to establish the measures necessary to achieve the following objectives: to ensure the functioning of the energy market; to ensure security of energy supply in the Union; to promote energy efficiency and energy saving and the development of new and renewable forms of energy; and to promote the interconnection of energy networks. Article 63 para 8 of this Regulation provides that the Commission may, after a market participant (a new interconnector) has requested the competent national authority to take a certain decision (to grant an exemption from certain provisions of this Regulation and of Directive 2019/9442441), request the national authority concerned (or exceptionally: the ACER) to amend or withdraw the decision to grant an exemption. The Commission shall do so within an (extendable) period of 50 days (*ex post* scrutiny). The predecessor Regulation – Regulation 714/2009 – in its Article 17 para 8 provided for nearly the same mechanism. It was still based on the more general Article 95 TEC (now: Article 114 TFEU). This reflects the proximity of Article 114 TFEU (as the more general provision) and the legal basis of Regulation 2019/943, Article 194 para 2 TFEU, as the more specific rule. It is apparent that Article 114 paras 4 ff TFEU is similar as regards the *ex post* scrutiny and in that it aims at balancing internal market concerns on the one hand and ensuring MS prerogatives, on the other hand. Since it is contained in Article 114 TFEU, it may – as regards its procedural specificities – be understood as a role model for other compliance (imple‐ menting) mechanisms.2442 This applies first and foremost to mechanisms based on Article 114 TFEU, but also to implementing mechanisms more

<sup>2440</sup> See Recital 57 of Regulation 2018/1139.

<sup>2441</sup> Article 63 para 1 of Regulation 2019/943.

<sup>2442</sup> Paras 4 and 5 of Article 114 TFEU cannot serve as a legal basis for compliance measures laid down in secondary law, though; see case C-359/92 *Germany v Council*, para 18.

generally, especially if they are based on a somewhat related provision. The compliance mechanism at issue here limits the risk of a heterogeneous application of EU law provisions throughout the EU which the exemption regime handled by national authorities bears. Furthermore, being aimed at ensuring 'the smooth functioning of the internal market for electricity',<sup>2443</sup> it is strongly linked to the rest of Regulation 2019/943. In conclusion, the mechanism seems to be correctly and proportionately established under Article 194 para 2 TFEU.

# 3.2.4.2. Mixed mechanisms

The mechanism laid down in Article 63 of Directive 2019/944, based on Article 194 para 2 TFEU as well, provides for the intervention of the ACER and the Commission in case a decision of a national regulatory authority does not comply with the Commission network codes and guide‐ lines referred to in this Directive or in Regulation 2019/943. It is a tiered procedure starting with an ACER opinion. Since these acts shall 'provid[e] the minimum degree of harmonisation required to achieve the aim of this Directive',2444 the link to the subject matter of the Directive appears sufficiently close. *Gundel* held that since no serious conflicts between the Commission and the national regulatory authorities have become known in the energy sector, the Commission's power raises subsidiarity concerns required to be addressed in a justification.2445 The scarcity of conflicts seems to emphasise that the Treaty infringement procedure not only in a legal but also in a practical perspective would be the adequate procedure to tackle these conflicts.2446 In that light, coverage by Article 194 para 2

<sup>2443</sup> Recital 63 of Regulation 2019/943.

<sup>2444</sup> Recital 92 of Directive 2019/944.

<sup>2445</sup> Gundel, Energieverwaltungsrecht, para 47 with regard to the predecessor Directive 2009/72/EC. In my view, the lack of such conflicts is also an argument against the likelihood that obstacles to the functioning of the internal market will emerge in the future. This may be one of the reasons why Directive 2019/944, unlike its predecessor, was not based on Article 114 TFEU. For the likelihood of future impediments to the internal market as a requirement for making use of Article 114 TFEU (preventive harmonisation); see also Classen, Art. 114 AEUV, para 71, with references to the Court's case law.

<sup>2446</sup> See Orator, Möglichkeiten 472 f, with regard to a similar mechanism (empowering the ACER).

TFEU may seem questionable, also with regard to proportionality (more concretely: the necessity criterion enshrined in it).

Article 109 TFEU provides for the Council, on a proposal from the Com‐ mission and after consulting the EP, to make any 'appropriate regulations' for the application of Articles 107 f TFEU. While this term resembles the wording of Article 100 para 2 TFEU,<sup>2447</sup> it is to be noted that in particular Article 108 TFEU predetermines to a large extent the procedure to be ap‐ plied in the context of State aid. The main novelty of the compliance mech‐ anism laid down in Articles 22 f of Council Regulation 2015/1589 (based on Article 109 TFEU) is that the proposal of the Commission, if accepted by the MS concerned, becomes binding upon the latter. This innovation seems to be located well within the leeway granted to the Council in Article 109 TFEU, as it allows the Council to flesh out Article 108 TFEU (in accordance with the latter).2448 In view of the strong coinage of this mechanism by primary law, its enforcement character – which is also reflected upon in the Council Regulation – does not appear to be a problem.

Article 121 para 6 TFEU allows for the EP and the Council to adopt 'detailed rules for the multilateral surveillance procedure referred to in paragraphs 3 and 4' in the form of regulations adopted pursuant to the ordi‐ nary legislative procedure. Pursuant to Article 136 para 1 TFEU the Council shall, in accordance with the relevant procedure from among those referred to in Articles 121 and 126 (except for Article 126 para 14), adopt measures specific to the Euro-MS: a) to strengthen the coordination and surveillance of their budgetary discipline; b) to set out economic policy guidelines for them, while ensuring that they are compatible with the guidelines adopted for the whole of the Union and that they are kept under surveillance. Regulation 1176/2011 is based on Article 121 para 6 TFEU and Regulation 1174/2011 on the same provision in conjunction with Article 136 TFEU.<sup>2449</sup>

While the corrective action plans to be submitted by the MS – which are not explicitly foreseen in the Treaties and hence can be considered an invention of Regulation 1176/2011 – do not appear to go beyond 'detailed

<sup>2447</sup> The word 'regulations' is to be understood strictly, though. It limits the instruments the Council may make use of under this provision to regulations according to Article 288 TFEU.

<sup>2448</sup> See Erlbacher, Art. 109 AEUV, para 4. Accordingly, and unsurprisingly so, the express objective of Council Regulation 2015/1589 is to lay down detailed rules for the application of Article 108 TFEU.

<sup>2449</sup> On the importance of Article 136 TFEU for a variety of crisis measures, among them Regulation 1174/2011, see Hinarejos, Crisis 32–34.

rules',2450 the sanctions regime of Article 1174/2011 is to be seen critically. After all, an interest-bearing deposit or an annual fine is not even hinted at in primary law.<sup>2451</sup> On the contrary, Article 121 TFEU is clearly established as a tender mechanism based on a dialogue between the EU and the MS which does not seem to envisage any kind of financial sanctions.<sup>2452</sup> Whether Article 136 TFEU which empowers the Council 'to strengthen the coordination and surveillance of the[] budgetary discipline' of Euro-MS allows for such reinforcement (with regard to Euro-MS) is contested.<sup>2453</sup>

It is true that the Court has confirmed, in the context of environmental protection through criminal law, that the EU legislator may provide for 'the application of effective, proportionate and dissuasive […] penalties' in order to ensure that the rules laid down by it are 'fully effective'.<sup>2454</sup> It did so with regard to a Council framework decision which obliged MS to incorporate criminal penalties in their respective national law, thereby leaving to the MS the choice of the criminal penalties to be applied (as long as they met the three criteria above).2455 However, here a comparatively elaborate procedure to be applied is laid down in primary law, and the legislator shall only complement the 'detailed rules' of it.2456 What is more, sanctions to be imposed on MS are politically more sensitive and hence less common than sanctions – to be more closely defined by the MS themselves – to be imposed, again by the MS themselves, on individuals or undertakings.<sup>2457</sup>


<sup>2450</sup> See Häde, Art. 121 AEUV, para 23, also criticising (with further references) the re‐ verse majority voting which applies to the adoption of Council recommendations in this context.

<sup>2451</sup> See also Obwexer, System 227.

Sanctions to be imposed on MS for non-compliance with EU law, ie as a measure of enforcement, in primary law are primarily provided for in Arti‐ cle 260 TFEU. By the introduction of fines, it was said, 'the EU competence for the coordination of economic policy is surreptitiously transformed into a competence for the adoption of *binding substantive policy decisions*' (em‐ phasis in original).<sup>2458</sup> The objective of Regulation 1174/2011 in its entirety – to create enforcement measures to correct excessive macroeconomic imbal‐ ances in the Euro area that is – sits uneasy with the thrust both of Article 136 para 1 and Article 121 para 6 TFEU which is focussed on coordination and plain surveillance.<sup>2459</sup> Apparently, the Court thinks otherwise. In a case on Regulation 1173/2011 on the effective enforcement of budgetary surveillance in the Euro area, again based on Article 136 and Article 121 para 6 TFEU, the Court made no objections to the sanctions provided for therein (see 3.1.1.2.3. above).<sup>2460</sup>

Article 91 para 1 TFEU stipulates that for the purpose of implementing Article 90 on the common transport policy, the EP and the Council shall, in

to take urgent action against Member States which fail in these obligations, using sanctions where necessary' (page 3). In the ensuing legislative proposal of the Commission, COM(97) 619 final, the Commission sanctions were dropped and the Commission contented itself with the mere power to adopt a decision estab‐ lishing the infringement, combined with the reference to the fact that individuals could have this decision 'rapidly enforced before the national courts and could, under the ways and means of national redress, obtain provisional measures, com‐ bined with penalty payments or fines, to prevent extension or aggravation of the obstacle, to end the alleged infringement and, if appropriate, achieve compensa‐ tion for the loss suffered' (page 3). In other words, the original plan of Commission sanctions *vis-à-vis* MS gave way to reliance on national means of redress (against the respective MS) which may be open to the individuals concerned; for an earlier contribution to the discussion on the introduction of sanctions in internal market law see Commission, 'Completing the Internal Market', COM(85) 310 final, para 154, according to which the Commission will 'continue its general action […] in order to correct violations rapidly and effectively. It will closely combine its actions of prevention and cure, and it will consider the possible introduction of sanctions'.

<sup>2458</sup> Bieber/Maiani, Enforcement 1071. For the Commission's search for alternative 'incentives' to ensure compliance in economic policy more generally see its Com‐ munication 'Enhancing economic policy coordination for stability, growth and jobs – Tools for stronger EU economic governance', COM(2010) 367 final, 8–11. For the conditionality-based incentives, in the context of the so-called 'umbrellas', see III.2.2.4.2.2. above; for traditional and new forms of conditionality in cohesion policy see Bachtler/Mendez, Cohesion 126-130.

<sup>2459</sup> See de Sadeleer, Architecture 366 f; Palmstorfer, Krise 170 f, both with further references.

<sup>2460</sup> See case C‑521/15 *Spain v Council*, eg para 44.

accordance with the ordinary legislative procedure, lay down: a) common rules applicable to international transport to or from the territory of a MS or passing across the territory of one or more MS; b) the conditions under which non-resident carriers may operate transport services within a MS; c) measures to improve transport safety; d) any other appropriate provisions. In view of the wide wording of this provision, in particular the legislator's competence to adopt 'any […] appropriate provisions',<sup>2461</sup> it appears that the regime laid down in Article 25 of Regulation 2016/796 – on the scrutiny of (draft) decisions of national authorities, which was qualified above as a clear case of implementation – is in accordance with primary law. While the ERA may only adopt an opinion, the Commission may adopt a binding decision requesting the national authority to modify or repeal the adopted decision, thereby being advised by a comitology committee. Having said that, it is surprising that among the objectives of this Regulation – namely to establish the ERA to formulate common solutions on matters concerning railway safety and interoperability2462 – no mention is made of its aiming for a uniform application of these common solutions and of the relevant Union law more generally.<sup>2463</sup>

### 3.2.4.3. Soft mechanisms

According to Article 6 paras 5–7 of Regulation 2019/942 (for its legal basis in primary law, Article 194 para 2 TFEU, see 3.2.4.2. above), the ACER may adopt opinions on the decisions of national regulatory authorities. It examines compliance of these decisions with binding network codes and guidelines referred to in Regulation 2019/943, Regulation 715/2009, Direc‐ tive 2019/944 or Directive 2009/73/EC, or with other relevant provisions of these directives or regulations. The scope of norms compliance with which is scrutinised overlaps with that of the mechanism laid down in Article 63 of Directive 2019/944, but also goes far beyond it. However, whereas under the latter mechanism the Commission may step in with a binding decision, under Article 6 paras 5–7 of Regulation 2019/942 the ACER opinion is

<sup>2461</sup> For the width of the words 'appropriate provisions' see the remarks made above in the context of Article 100 para 2 TFEU (3.2.4.1.).

<sup>2462</sup> See Recital 45 of Regulation 2016/796.

<sup>2463</sup> Note that the submission of draft national rules to the ERA and the Commission for consideration is also provided for in Article 8 para 4 of Directive 2016/798 and Article 14 para 5 of Directive 2016/797.

the only output on the part of an EU actor. In view of that there are no concerns with respect to Article 194 para 2 TFEU as the primary legal basis.

Article 53 of Directive 2019/944 lays down a procedure for the case that certification is requested by a transmission system owner or a transmission system operator which is controlled by a person from a third country. Such third country investments are to be treated with care, as transmission system owners or operators controlled by persons from a third country could pose a risk to the security of the energy supply to the Union and hence to the functioning of the internal market. Therefore the Commission shall 'guarantee that companies from third countries respect the same rules that apply to EU based undertakings in both letter and spirit',<sup>2464</sup> a task which is closely linked to the subject matter of Directive 2019/944, namely to provide common rules for the internal market in electricity. The national regulatory authorities have to inform the Commission of their respective draft decisions on such a request. National law shall provide for the nation‐ al authorities to ask for a Commission opinion in this case, of which they shall take 'utmost account'.2465 The intervention on the part of the EU constitutes classical implementation in that it envisages an *ex ante* scrutiny of MS action (draft decisions). It is a soft mechanism, and MS may – under the conditions laid down in Article 53 para 8 of the Directive – lawfully deviate from it. This and the delicate scenario which is addressed (risk to the security of the energy supply to the EU) in my view render the mechanism compliant with Article 194 para 2 TFEU.

The mechanism providing for a Council recommendation to adopt precautionary corrective measures or to prepare a draft macroeconomic adjustment programme as laid down in Article 3 para 7 of Regulation 472/2013, the weakest regime based on Article 136 and Article 121 para 6 TFEU which is referred to here (see also 3.2.4.2. above), seems to be in accordance with the coordination framework laid down in primary law.

Article 207 TFEU stipulates that the EP and the Council shall adopt the measures defining the framework for implementing CCP by means of regulations adopted in accordance with the ordinary legislative procedure (para 2). Since the Treaty of Lisbon, this core Treaty provision on CCP in its para 1 explicitly states that this policy field shall be based on uniform principles also with regard to foreign direct investment. On the basis of

<sup>2464</sup> Commission Proposal COM(2007) 528 final, 7, leading to the predecessor Direc‐ tive 2009/72/EC.

<sup>2465</sup> Article 53 para 8 of Directive 2019/944.

this competence clause (Article 207 para 2 TFEU) Regulation 2019/452 was adopted. Its Article 6 allows for the Commission to adopt an opinion where MS' security or public order requires immediate action. In view of the legal non-bindingness of the Commission output and the generous wording of Article 207 para 2, according to which 'the measures defining the frame‐ work for implementing the [CCP]' may be regulated, and in light of the fact that this mechanism pertains the main objective of the Regulation, namely to provide 'a policy response to protect legitimate interests with regard to foreign direct investments that raise concerns for security or public order of the Union or its Member States',<sup>2466</sup> it appears that the compliance mechanism laid down in Article 6 of Regulation 2019/452 has a sufficient legal basis in primary law.

# 3.3. The institutional balance of the EU

### 3.3.1. Introduction

Having attempted a categorisation of the compliance mechanisms presen‐ ted above in terms of whether they constitute implementation or enforce‐ ment (3.1.) and having assessed whether they are in accordance with their respective primary legal bases (3.2.), we shall now address them with a view to their influence on the EU's institutional balance.

The institutional balance of the EU is an important principle regarding the intra-EU distribution of powers. It was first pronounced, albeit still under a different name, by the Court in its *Meroni* judgements of 1958, in which it referred to the 'balance of powers which is characteristic of the institutional structure of the Community'.2467 As a principle 'characteristic

<sup>2466</sup> Commission Proposal COM(2017) 487, 2.

<sup>2467</sup> Cases 9–10/56 *Meroni*, 151; see also case 25/70 *Einfuhr- und Vorratsstelle*, para 4: 'institutional balance', and – more recently – case C-409/13 *Council v Commission*, para 64; case C-928/19P *EPSU*, para 48. For the changing terminology see Michel, Gleichgewicht 74. For the distinction between 'separation of powers' and 'institu‐ tional balance' see Opinion of AG *Trstenjak* in case C-101/08 *Audiolux*, paras 104 f; for the nevertheless existing relationship between the two principles see Conway, Separation 319–321, and references in Orator, Möglichkeiten 219 (fn 187). The principle of institutional balance is not expressly mentioned in the Treaties, with the exception of Protocol No 7 annexed to the Treaty of Amsterdam on the appli‐ cation of the principles of subsidiarity and proportionality: '[t]he application of the principles of subsidiarity and proportionality shall respect the general provisions

of the institutional structure of the European Union', it 'requires that each of the institutions must exercise its powers with due regard for the powers of the other institutions',<sup>2468</sup> that is to say paying tribute to its respective *Organkompetenz*.

The institutional balance, as alluded to in Article 4 para 3 in conjunction with Article 13 para 2 TEU,<sup>2469</sup> is actually struck by the distribution of powers laid down in the Treaties.<sup>2470</sup> Hence it is the concrete balance laid down in law, not a State-theorist, 'ideal' balance which is at issue.<sup>2471</sup> What is more, the balance is not struck between the three separate powers according to *Montesquieu*, <sup>2472</sup> but primarily between all EU institutions, bodies, offices and agencies. This entails a separation in the sense that the powers of the respective other actors (institutions, bodies, offices and agencies) are acknowledged, but also a cooperation aimed at a meshing of the powers of the various actors. Any act of secondary law fleshing out the powers of an EU body may potentially lead to a distortion of the EU's institutional balance.2473 Thus, also the compliance mechanisms presented above, as far as they are laid down in secondary law, may distort the EU's institutional balance. In this context, the question whether they allow for the adoption of law (mixed and hard mechanisms) or whether they only provide for the adoption of legally non-binding output (soft mechanisms) plays an important role. While soft law powers have the capacity to entail a distortion of the institutional balance, as well,2474 the interference with other actors' competences it may cause by tendency is smaller. Where the legal norm providing for a compliance mechanism, eg a Regulation, leaves it unclear whether a soft or a hard law power is delegated, it must be perceived as a soft law power. This is because 'a delegation of powers

and objectives of the Treaty, particularly as regards [...] the institutional balance'; see also Opinion of AG *Wathelet* in case C-425/13 *Commission v Council*, para 68.

<sup>2468</sup> Case C-24/20 *Commission v Council*, para 83, with further references.

<sup>2469</sup> See case C-413/11 *Germanwings*, para 16.

<sup>2470</sup> See Joerges/Vos, Structures 84 f, also with regard to how the concept of institution‐ al balance may as well reflect upon the separation of tasks between the EU and the MS.

<sup>2471</sup> See Priebe, Entscheidungsbefugnisse 78 (fn 39); see also case 138/79 *Roquette Frères*, para 33: 'institutional balance intended by the Treaty', and case C-11/00 *Commission v European Central Bank*, para 174.

<sup>2472</sup> See, however, Lenaerts/Verhoeven, Balance40 f.

<sup>2473</sup> See eg Everson, Governance 147 f, with further references. This applies also to public international law: see eg Rossolillo, Compact.

<sup>2474</sup> See Senden, Soft Law 334–336, with further references; Gentile, Review 487.

cannot be presumed', as the Court held in *Meroni*. <sup>2475</sup> Therefore, in case of doubt as to the power which is delegated, it is the weaker power – in our case: a soft law power – which is to be assumed.<sup>2476</sup>

Exceptionally, the Treaties explicitly allow for the shaping of the relation‐ ship between EU institutions, bodies, offices and agencies by means of secondary law. Article 103 para 1 in conjunction with para 2 lit d TFEU allows for the Council to define, by means of regulations or directives, the respective functions of the Commission and of the Court in applying Articles 101 and 102 TFEU. It could be concluded *e contrario* that where the Treaties do not make express provision for this possibility, the relationship between EU bodies may not be re-shaped.<sup>2477</sup> This would also concern our compliance mechanisms. And while this exclusionary deduction appears to be too strict, the EU's institutional balance certainly requires closer examination in the context of compliance mechanisms.

Therefore, in the following sub-chapters, the questions arising from com‐ pliance mechanisms in relation to the EU's institutional balance shall be ap‐ proached in the following order. After addressing – in this specific context – the meaning of the Treaty infringement procedure and the prerogatives of the CJEU it entails (3.3.2.), it shall be explored whether the role the Court may play in other (mixed and hard) compliance mechanisms – in particu‐ lar via the annulment procedure – can remedy the institutional changes brought about by these alternative compliance mechanisms (3.3.3.). Even‐ tually, the delegation of implementing (and enforcement) powers beyond Article 291 para 2 TFEU shall be assessed. Such forms of delegation are in‐ creasingly common, in particular – but not only – where European agencies act as delegates (3.3.4.).

# 3.3.2. The Treaty infringement procedure

As a general statement with regard to alternatives to the Treaty infringe‐ ment procedure, the Court has held that 'it must be recalled that special procedures in a directive can neither derogate from nor replace the powers

<sup>2475</sup> Cases 9–10/56 *Meroni*, 151.

<sup>2476</sup> Similarly in case 98/80 *Romano*, para 20.

<sup>2477</sup> Articles 101 and 102 TFEU in practice are primarily applied to private undertak‐ ings, but they may be – and actually have been – applied also to public undertak‐ ings and to MS more generally; for public undertakings see also Article 106 TFEU; for the indirect obligations of MS in general see case 231/83 *Cullet*, para 16.

of the Commission under Article 226 EC [now: Article 258 TFEU]'.<sup>2478</sup> It appears that this finding also applies with regard to other acts of secondary law than a directive (which was at issue in the given case).<sup>2479</sup> In other words, it is not allowed – legally infeasible – by means of secondary law to derogate from or to replace the powers of the Commission under Article 258 TFEU. As a consequence, the application also of the administrative phase of the Treaty infringement procedure shall remain untouched by secondary law-based compliance mechanisms – be they governed by the Commission itself or by other EU bodies, eg European agencies. <sup>2480</sup> The latter may be applied next to the Treaty infringement procedure or, if that is what the Commission deems more opportune in a concrete case, either procedure may be applied exclusively.<sup>2481</sup> *In practice* the former possibility – two procedures being applied in parallel – seems to be a rare scenario, which is why the variety of compliance mechanisms laid down in secon‐ dary law may have an impact on the role of the Commission under the Treaty infringement procedure. Against this background, it is remarkable that the Commission in its recent Communication 'Enforcing EU law for

<sup>2478</sup> Case C-424/07 *Commission v Germany*, para 36, with a further reference. See also case C-156/21 *Hungary v European Parliament/Council*, paras 166-168, with regard to the (questionable) circumvention of the procedure laid down in Article 7 TEU (no 'parallel' procedure).

<sup>2479</sup> This view was acknowledged by the Commission (and the legislator): see Com‐ mission, 'The European Agencies – The Way Forward', COM(2008) 135 final, 5, in which it emphasises that 'agencies cannot be entrusted with powers which may affect the responsibilities which the Treaty has explicitly conferred on the Commission (for example, acting as the guardian of Community law)'.

<sup>2480</sup> See case C-146/91 *KYDEP*, para 30, addressing a 'reminder' by the Commission of the obligations under EU law of a certain MS. A general legal basis for *ad hoc* enforcement procedures was discussed in the negotiations on the Nice Treaty, but was eventually not introduced; see Andersen, Enforcement 144–148. For the 'exclusivity' of the institutions' core competences see Griller/Orator, Everything 25; exemplifying the EP's pushing for an empowerment of agencies in this respect and the Commission's opposition: Alberti, Actors 46.

<sup>2481</sup> Case T-461/93 *An Taisce*, paras 35 f; see also order in appeal case C-325/94P *An Taisce*, paras 24–26. The Commission itself has declared, however, that even if no inter-dependence is required by law, in its administrative practice of handling the two procedures – the Treaty infringement procedure on the one hand, and the procedure laid down in Article 24 of Regulation 4253/88 on the other hand – there should be 'a degree of consistency'; Commission, 'Sixteenth Annual Report on monitoring the application of community law – 1998', COM(1999) 301 final, 35. In some cases this is explicitly stipulated in (secondary) law; see eg Article 1 para 4 of Regulations 1093–1095/2010 in general, and Article 17 para 6, Article 18 para 4 and Article 19 para 4 *leg cit* in particular.

a Europe that delivers', while acknowledging the importance of other 'key bodies' (in particular national courts and authorities) when it come to enforcing EU law, does not expressly acknowledge the role of European agencies in this context.<sup>2482</sup>

However, it is not only the Commission's power as 'guardian of the Treaties' which may be affected by special compliance mechanisms. Also the power of the CJEU as the ultimate interpreter of EU law according to the Treaties may be concerned. On a number of occasions the Court has stressed – thereby referring to what is now Article 344 TFEU – that it does not condone any rivals in this respect.<sup>2483</sup> It is important to note that these cases concerned international treaties to which MS were parties and which provided for the jurisdiction of an international court. In this context it is to be understood when the Court said that a certain agreement 'is likely adversely to affect the allocation of responsibilities defined in the Treaties and, hence, the autonomy of the Community legal order, respect for which must be assured by the Court of Justice'.2484 This would result in a distor‐ tion of the institutional balance and, furthermore, would be problematic in a rule of law perspective. International agreements concluded by the MS must accept 'the binding nature of the Court's case-law or the autonomy of the Community legal order'2485 for them to be in compliance with Union law. The mere 'risk that a judicial forum other than the Court will rule on the scope of obligations imposed on the Member States pursuant to Com‐ munity law'2486 challenges the Court's competences laid down in primary law.2487 The Court stuck to this line when asked about the possibility of an

<sup>2482</sup> See Commission, 'Enforcing EU law for a Europe that delivers' (Communication), COM(2022) 518 final, 8.

<sup>2483</sup> What is more, the Court has excluded certain international courts from the catego‐ ry of 'courts or tribunals of the MS' pursuant to Article 267 TFEU: case C-284/16 *Slowakische Republik v Achmea*, in which the Court stressed that the arbitral tribu‐ nal established by means of a bilateral investment treaty between two MS – for lack of 'links with the judicial systems of the Member States' – cannot be qualified as a 'court or tribunal of a Member State' within the meaning of Article 267 TFEU (in particular paras 47–49); see also Peers, Form, *passim*, but in particular 71.

<sup>2484</sup> Opinion 1/91 *EEA I*, para 35; for the discussion on an unchangeable core of primary law see Calliess, Art. 13 EUV, para 29 f; Curtin, Structure 63–66; see also Bieber*,* Limites*.*

<sup>2485</sup> Opinion 1/92 *EEA II*, para 29.

<sup>2486</sup> Case C-459/03 *Commission v Ireland*, para 177.

<sup>2487</sup> See also the Court's critical stance on the 'Fund Tribunal' in Opinion 1/76 *Water‐ way vessels*, in particular para 22.

accession of the EC/EU to the (European) Convention for the Protection of Human Rights and Fundamental Freedoms (ECHR): 'The interpretation of a provision of EU law, including of secondary law, requires, in principle, a decision of the Court of Justice where that provision is open to more than one plausible interpretation. If the Court of Justice were not allowed to provide the definitive interpretation of secondary law, and if the European Court of Human Rights, in considering whether that law is consistent with the ECHR, had itself to provide a particular interpretation from among the plausible options, there would most certainly be a breach of the principle that the Court of Justice has exclusive jurisdiction over the definitive inter‐ pretation of EU law'.<sup>2488</sup>

While the mechanisms at issue in these cases were clearly established outside the EU legal order, namely in public international law,<sup>2489</sup> also with regard to mechanisms established on the basis of Union law the Court em‐ phasised the singularity of its role. In the context of the Treaty infringement procedure, the Court held that 'the rights and duties of Member States may be determined and their conduct appraised only by a judgment of the Court'.2490 In spite of its important role in the administrative phase of this procedure in law and practice (today, more than 90 % of the infringement cases are settled prior to being referred to the Court2491), the Commission here does not act as a rival of the Court (in conformity with primary law), as its decision to discontinue proceedings does not constitute a declaration of lawfulness of the MS behaviour at issue.<sup>2492</sup>

<sup>2488</sup> Opinion 2/13 *ECHR II*, paras 245 f; see also case 2/94 *ECHR*, in which the Court made clear (paras 30 and 35) that an accession of the then EC to the ECHR could only be brought about via a Treaty amendment. Due to 'constitutional significance' it could not be based on the flexibility clause (now: Article 352 TFEU).

<sup>2489</sup> In the case of the Unified Patent Court the CJEU denied an interference with Union law: see case C-146/13 *Spain v European Parliament/Council*, and case C-147/13 *Spain v Council*. In principle, the Court seems to apply the same criteria, irrespective of whether powers are delegated to a private body (*Meroni* cases) or to a body of public international law (*Laying-up fund* case); see also Priebe, Entscheidungsbefugnisse 115.

<sup>2490</sup> Joined cases 142/80 and 143/80 *Amministrazione delle Finanze dello Stato*, para 16; see also case C-191/95 *Commission v Germany*, para 45; case T‑258/06 *Germany v Commission*, para 153.

<sup>2491</sup> See Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 21. Already nearly 25 years ago, *Ibáñez* conceived that a 'new social complex‐ ity requires more active intervention on the part of administrations, so that the maximum number of cases may be resolved in the pre-judicial or pre-litigious phase'; Gil Ibáñez, Supervision 1.

Also with respect to financial sanctions according to what is now Article 260 TFEU, the Court highlights its sole jurisdiction to impose sanctions on a MS for failing to comply with its judgement.<sup>2493</sup> AG *Geelhoed* mentions a 'functional reason[]' for that: It is 'the Court which is best placed to assess to what extent the situation pertaining in the Member State concerned does or does not comply with its first judgment under Article 226 EC and to assess the seriousness of a continued infringement having regard to all interests involved'.<sup>2494</sup> This is, the Court complements, because the imposition of financial sanctions under Article 260 TFEU 'is not intended to compensate for damage caused by the Member State concerned, but to place it under economic pressure which induces it to put an end to the breach established'.<sup>2495</sup> Article 260 para 2 TFEU must therefore be regarded as a 'method of enforcement'.<sup>2496</sup>

In the above references, the Court has dealt with its role under the Treaties, notably under the Treaty infringement procedure. It has not at the same time addressed, let alone sanctified special compliance mechanisms. The Court does not only clarify the distribution of powers between the Commission (or the MS pursuant to Article 259 TFEU) and the Court within the Treaty infringement procedure, it also seems to imply that in terms of *enforcement* there can be no alternative to the Treaty infringement procedure.

The wide category of compliance mechanisms does not in the first place encompass enforcement measures, but – as we have seen – in particular implementing measures (see 3.1. above). While implementing powers of the EU are explicitly provided for in the Treaties (generally in Article 291 para 2 TFEU) and consequently – on their respective own – the underlying mech‐ anisms 'do not upset the horizontal division of powers and tasks established in the Treaty',2497 enforcement mechanisms bear a high risk of negatively affecting the Court's powers under the Treaty infringement procedure and of thereby tilting the EU's institutional balance.2498 This is possible for each

<sup>2492</sup> See case 74/69 *Hauptzollamt Bremen-Freihafen*, para 9; joined cases 15–16/76 *France v Commission*, para 27.

<sup>2493</sup> See case C-304/02 *Commission v France*, para 90.

<sup>2494</sup> Opinion of AG *Geelhoed* in case C-304/02 *Commission v France*, para 24.

<sup>2495</sup> Case C-304/02 *Commission v France*, para 91.

<sup>2496</sup> Case C-304/02 *Commission v France*, para 92.

<sup>2497</sup> Andersen, Enforcement 143; for implementing mechanisms more generally see ibid 163 ff.

<sup>2498</sup> See also Gil Ibáñez, Supervision 127.

single enforcement mechanism established by secondary law which is not grounded on an adequate (in this context that means: explicit) legal basis in the Treaties. Having said that, not only the single (unlawful) mechanisms may threat the EU's institutional balance. Also the entirety of (on their respective own: lawful) compliance mechanisms (as just mentioned: in the majority of cases they are implementing mechanisms), taken collectively, may pose a risk to the Court's prerogatives under Articles 258–260 TFEU. This is because also implementing mechanisms share the broad objective of the Treaty infringement procedure, to ensure MS' compliance with EU law that is (see 3.1.1.2.1.2. above). With the individual-concrete implementation of Union law – in the spirit of Article 291 para 2 TFEU this is assumed to be an exception – becoming more and more frequently provided for in secondary law, the Court's role under the Treaty infringment procedure seems to be under pressure.

With regard to this latter, summative effect, it could be argued that the establishment of compliance mechanisms step by step – even if each single mechanism as such may be lawful – has transformed the institutional balance of the EU without the text of the Treaties having been changed in that respect. The increasing number of such mechanisms and the longterm decrease in the number of Treaty infringement procedures2499 at least suggest that there is a risk that the Court's role in ensuring compliance is slowly being reduced to cases which cannot be tackled by an alternative compliance mechanism, for example the late transposition of directives.<sup>2500</sup> There may be further explanatory factors, eg fewer late transposition cases due to the significant increase of the regulations-directives ratio in favour of the former, in particular over the past 15 years,2501 or the effectiveness of EU Pilot. Mention should also be made of compliance-related trends other than the compliance mechanisms at issue here, in particular the

<sup>2499</sup> See <https://ec.europa.eu/info/publications/annual-reports-monitoring-applicat ion-eu-law\_en> accessed 28 March 2023; see also Koops, Compliance 98; Börzel, Noncompliance 97, providing for a timeline stretching from 1978 to 2017.

<sup>2500</sup> For the comparatively high proportion of new infringement procedures relating to late transposition see European Commission, 'Monitoring the Application of Union Law. 2018 Annual Report, Part I: general statistical overview' (2019) 18; for a similar trend with regard to EU enforcement *vis-à-vis* individuals see Scholten, Trend 1357 f.

<sup>2501</sup> See Börzel, Noncompliance 105.

increasing popularity of incentive-based measures.2502 These measures may be applied as a means to ensure compliance with EU law, as well, and may corroborate the just described development away from the Treaty in‐ fringement procedure. In addition to that, it is to be acknowledged that the (decreasing) number of Treaty infringement procedures – while still 'the most comprehensive, valid, and reliable measurement of noncompliance in the EU'<sup>2503</sup> – only gives a hint at the development of non-compliance. Most instances of non-compliance are not reported, let alone made subject to a Treaty infringement procedure.<sup>2504</sup>

In the context of the German Federal Constitution this phenomenon of a gradual transformation has been described as 'schleichender Verfassungs‐ wandel' [creeping constitutional change].<sup>2505</sup> In EU-related scholarship the broader term 'integration by stealth'<sup>2506</sup> is more common, mostly pertaining to the EU-MS relationship. With reference to the risk emanating from international agreements concluded between the MS, *Pescatore* warned of 'une révision froide' [a cold revision] of the then EEC Treaty.2507 If at all, we talk about a different kind of transformation here, namely an intra-EU development, but also in this context the above terms are telling metaphors.

The institutional balance of the EU is not a narrowly tailored regime, but one that allows for some flexibility.2508 However, this flexibility does not mean that the principle of institutional balance is entirely inapt to serve as a limit to integration without Treaty change. It is, as was said before, not only exceptional enforcement mechanisms but also the sheer multitude of (alternative) compliance mechanisms (enforcement and implementing mechanisms based on secondary law) which bear the risk of a re-weighting by stealth of the EU's institutional balance, first and foremost to the detri‐ ment of the Court's role under the Treaties.

<sup>2502</sup> In the context of the rule of law: Halmai, Possibility 1; for the COVID-19 Recovery Plan see de Witte, COVID-19.

<sup>2503</sup> Börzel, Noncompliance 34.

<sup>2504</sup> See Börzel, Noncompliance 21 and *passim*.

<sup>2505</sup> The term is widely used in German literature; see, as one example, Vorländer, Verfassung 235 f, with further references.

<sup>2506</sup> See in particular Majone, Dilemmas.

<sup>2507</sup> Pescatore, Ordre 144.

<sup>2508</sup> See eg Schmidt-Aßmann, Verwaltungsrecht 396 f.

### 3.3.3. Judicial review

A question which arises in the given context is the following: Does the possibility of judicial review following a mixed or hard special compliance mechanism – and hence the possibility of a Court judgement on the case – remedy the ousting of the Court in the procedure up to then? In my opinion, the fact that decisions adopted by the Commission or other ad‐ ministrative bodies of the EU in the context of special (mixed and hard) compliance procedures may be made subject to judicial review on the basis of Article 263 TFEU does not remedy – in terms of the EU's institutional balance – the deviation from the Treaty infringement procedure. While, legally speaking, the applicability of the Treaty infringement procedure (if the respective requirements are met) remains untouched, the existence and application of compliance mechanisms *de facto* relativises the Court's role as 'a single judicial body […] which can give definitive rulings on the law for the whole of the Community',2509 but in particular it relativises its role under the Treaty infringement procedure by reducing the likelihood of such a procedure being launched in a concrete case. The fact that the Court may be called upon in the course of annulment procedures can hardly serve as a compensation. An increase in the number of annulment procedures cannot remedy a reduced role under the Treaty infringement procedure – at least not qualitatively. After all, the Court's role under Article 263 TFEU is remarkably different from that under Articles 258 and 260 TFEU. Under the Treaty infringement procedure, the Court is asked to examine MS' behaviour and to rule on whether or not it is in accordance with EU law. Under the conditions of Article 260 TFEU, it may even impose financial sanctions on non-complying MS. In the course of an annulment procedure, on the contrary, the Court may – by deciding on the lawfulness of EU output adopted in the course of a specific compliance mechanism – only indirectly utter its view on the lawfulness of MS behaviour. It is a mere review of the assessment of a MS' action2510 by an EU body which takes

<sup>2509</sup> CJEU, 'Report of the Court of Justice on certain aspects of the application of the Treaty on European Union' (Luxembourg, May 1995) 3.

<sup>2510</sup> In the compliance mechanisms addressed here it is MS' action which is at issue. As we have seen, it is only exceptionally the case that decisions are addressed to individuals in the framework of these mechanisms.

place in this procedure, as opposed to the original assessment of MS action by the Court in the course of a Treaty infringement procedure.<sup>2511</sup>

In addition to that, there is a factual aspect which can hardly be over-em‐ phasised: The legally binding acts adopted in the course of (mixed and hard) compliance mechanisms may be brought before the CJEU, but they may as well not. In the former case, it is the MS concerned which has to take action (unlike in the case of Treaty infringement procedures pursuant to Article 258 TFEU, as initiated by the Commission). In the latter case, the EU administrative body has the final say on the assessment of the facts underlying the concrete case. It may be argued that this is also the case with the Commission's decision to end a Treaty infringement procedure prior to a Court decision. However, such a termination does not necessarily settle the matter for good, and certainly it does not settle the matter with a legally binding decision. Also, it does not harm the MS concerned, but is regularly perceived as a relief. Furthermore, this particular power of the Commission is explicitly provided for in primary law.

Where the final output on the part of the EU body in charge is soft, that is to say in soft compliance mechanisms, the path of Article 263 TFEU may not be embarked upon in the first place. Instead, the MS addressed may simply not abide by this output. If it does, however, adapt its behaviour to that prescribed in the respective soft law act (for whichever reason; see III.4. above), the prescription laid down in this act in the given case effectively is final.2512 Thereby, also mixed mechanisms which *in concreto* end with a soft law act or soft mechanisms may affect the EU's institutional balance or the division of powers between the EU and its MS.2513 The ESRB's soft powers under Articles 16 f of Regulation 1092/2010 shall serve as an example for the latter. Where the ESRB uses this power to proactively suggest legislative action,2514 either to the EU or to the MS, this raises concerns. After all, the right to initiate legislative action is – where the EU has the respective competence – up to the Commission (or exceptionally other actors, as laid down in the Treaties) or – where a MS competence is at issue – up to the respective national actors, in particular the MS' govern‐

<sup>2511</sup> For the purposes of the procedures laid down in Articles 258 and 260 TFEU in the view of the Court see joined cases C-514/07P, C-528/07P and C-532/07P *API*, para 119.

<sup>2512</sup> Note the facts referred to in case T-116/89 *Prodifarma*, para 83.

<sup>2513</sup> See also Dawson, Soft Law 4.

<sup>2514</sup> See Article 16 para 1 of Regulation 1092/2010.

ments.2515 The freedom in principle to propose or not to propose legislative action is one of the core guarantees in democratic constitutions. Secondary law does not seem to be an adequate means to allow for interference in this context. The non-binding character of the recommendations may render them comparatively weak, but in the above case they still seem to interfere with other EU and/or national actors' competences.<sup>2516</sup> In the former case, this challenges the EU's institutional balance, in the latter case the principle of conferral.

It is possible – for the final output of hard, mixed and soft mechanisms alike – that the Court gets the chance to examine one of these acts in a separate procedure (eg following an action for the EU's non-contractual liability) or that it is directly made subject to a preliminary reference,<sup>2517</sup> but it is not too probable that such procedures are initiated with direct or indirect regard to an act addressed *to a MS or a national authority*. As men‐ tioned above, the likelihood of judicial review by the Court is not a legal, but a factual issue. Nevertheless, it ought to be taken into account when considering the question whether or not the Court's role envisaged in the Treaties is being ousted by secondary law-based compliance mechanisms.

In conclusion, from an institutional balance perspective the openness of the output adopted in the course of secondary law-based compliance mechanisms to judicial review by the CJEU under different procedures is hardly sufficient to remedy the Court's being skipped in the mechanisms themselves – and mostly also in the Treaty infringement procedure (where the requirements for its application are met in the first place). After all, in practice the performance of a specific compliance mechanism most of the time – factually, not legally – excludes the initiation of a Treaty infringe‐ ment procedure in the case at issue.

<sup>2515</sup> With regard to binding requests to that effect see Andersen, Enforcement 164.

<sup>2516</sup> Describing the ESRB's powers, in particular to adopt recommendations, as 'fairly strong tools': Saarenheimo, Policy 31; similarly, but attesting the ESRB to be 'just on the right side of the constitutional line': Ferran/Alexander, Oversight Bodies 23 f.

<sup>2517</sup> The option of an *incidenter* ruling pursuant to Article 277 TFEU is only available for acts of general application, hence irrelevant in the given context.

3.3.4. The neglection of Article 291 para 2 TFEU and the empowerment of European agencies

### 3.3.4.1. Introduction

The output of the compliance mechanisms at issue here normally ranges somewhere in between EU (legislative) rule-making and the application of these rules in everyday administration (*vis-à-vis* individuals/undertakings) which, in principle, is a prerogative of the MS (pursuant to Article 291 para 1 TFEU). Most of the compliance mechanisms are not located at the poles of this scale, that is to say most of them neither result in rule-mak‐ ing nor in decisions *vis-à-vis* individuals/undertakings. Only exceptionally legislative powers are involved and, again only exceptionally and given that a MS does not comply even upon request by the EU body in charge, the compliance mechanisms provide for the application of EU law *vis-à-vis* an individual/undertaking.

Above all, the compliance mechanisms qualifying as 'implementation' according to the distinction made above (3.1.) are expressions of a strong administrative (in particular: procedural) cooperation between the EU and the national level.2518 The individual/undertaking concerned by the action of the national authority at issue (eg the grant of a concession) may not take notice of the EU action involved (eg the scrutiny of the draft concession), and may therefore perceive of it as purely national administration.2519 In a macro-perspective, however, these mechanisms qualify as mixed admi‐ nistration (involving the EU and the national level).2520 Even if we look at the above mechanisms only from an EU (not from a MS) perspective, many of them display strong links to the national administrations, as expressed, for example, in the activity of European agencies or comitology committees. This development – the increasing administrative entanglement between the EU and the national level – has been going on for decades now,<sup>2521</sup>

<sup>2518</sup> Note the remark of *Bilder* who – in the context of (soft) international arrange‐ ments – stressed that it is not only compliance that matters in this context, but the facilitation of cooperation between nations; Bilder, Compliance 65.

<sup>2519</sup> For this problem see J Hofmann, Protection 464.

<sup>2520</sup> See Schütze, Rome 1419 ff; for an alternative (procedural) approach towards this phenomenon see Hofmann/Rowe/Türk, Administrative Law 11–18.

<sup>2521</sup> For this development see eg Hofmann/Rowe/Türk, Administrative Law 259–264. For the concept of composite administration (*Verwaltungsverbund*) see Schmidt-Aßmann, Introduction 6–8; for the Commission's vocabulary of 'networks' (in

and the traditional depiction of the implementation of EU law through the distinction between direct execution of EU law on the one hand and indirect execution on the other hand more and more appears to be insufficient.<sup>2522</sup> Also the general principle of the implementation of EU law by the MS (Article 291 para 1 TFEU) has become strongly relativised. It remains a mere 'Grundlinie mit erheblichen Abweichungstoleranzen' [base line with a significant deviation tolerance], as *Schmidt-Aßmann* has put it.<sup>2523</sup> While the Treaties only rarely request this procedural/institutional mixity, a number of provisions appear to presuppose or at least to facilitate such cooperative approach.<sup>2524</sup>

Here we shall analyse in more depth the fact that European agencies are vested with hard or soft law powers *vis-à-vis* the MS in a number of the above compliance mechanisms, an empowerment which has raised concerns in particular regarding the Commission's constitutional role as guardian of the Treaties.<sup>2525</sup>

2523 Schmidt-Aßmann, Verwaltungsrecht 382.

A large body of literature on the role of European agencies has emerged in recent years. Suffice it to refer to: Chamon, Agencies, in particular 29–44, with regard to

German: *Verbundverwaltung*) see Communication 'Management of Community programmes by networks of national agencies', COM(2001) 648 final, 4.

<sup>2522</sup> See J Hofmann, Protection 441f. For the general trend away from direct or indirect execution of EU law towards a cooperative implementation, involving both EU and national administrators see Britz, Verwaltungsverbund 46; Schaller, Intensi‐ vierung 415 f, with further references: 'vollzugsprogrammierende und -steuernde Wirkung' [programming and steering effect of national enforcement by EU actors].

<sup>2524</sup> See, for example, the loyalty principle as laid down in Article 4 para 3 TEU; the reference to the mechanisms for control by MS in Article 291 para 3 TFEU; Article 197 TFEU on administrative cooperation between the EU and the national level; see S Augsberg, Verwaltungsorganisationsrecht, para 12.

<sup>2525</sup> See eg the EP's concern (with regard to European agencies) about a 'consequent risk of the Commission's executive role being dismantled and fragmented into a plethora of bodies that work largely in an intergovernmental manner'; European Parliament, Resolution on the draft interinstitutional agreement presented by the Commission on the operating framework for the European regulatory agencies, P6\_TA(2005)0460, C.5.

See also the considerations on the role of the Commission as guardian of the Treaties in Commission Proposal COM(2007) 528 final (page 13) for what has become Directive 2009/72/EC; for the Commission's wariness not to challenge its role as guardian of the Treaties see Commission, 'European agencies – The way forward' (Communication), COM(2008) 135 final, 5: 'agencies cannot be entrusted with powers which may affect the responsibilities which the Treaty has explicitly conferred on the Commission (for example, acting as the guardian of Community law)'; see also Groenleer/Kaeding/Versluis, Governance 1227.

# 3.3.4.2. European agencies and Article 291 para 2 TFEU

Primary law does not explicitly provide for the empowerment of European agencies. Against this background, the fact that agencies – in many compli‐ ance mechanisms – are concerned with the weighing of different (public) interests (brought forward by the national authorities) may be considered as affecting the EU's institutional balance. <sup>2526</sup> In a systematic view, however, the Treaties do reveal some hints at the possibility of an empowerment of agencies (see also III.3.7.2.2. above), to which the Court – thereby arguing in favour of the principal possibility to vest agencies with (implementing) decision-making powers – has already referred: 'Under Article 263 TFEU, the Union bodies whose acts may be subject to judicial review by the Court include the "bodies, offices" and "agencies" of the Union. The rules governing actions for failure to act are applicable to those bodies pursuant to Article 265 TFEU. Article 267 TFEU provides that the courts and tribu‐ nals of the Member States may refer questions concerning the validity and interpretation of the acts of such bodies to the Court. Such acts may also be the subject of a plea of illegality pursuant to Article 277 TFEU'.<sup>2527</sup> Later it added that since the empowerment of agencies is not mentioned in Article 291 TFEU (nor in Article 290 TFEU), '[i]t is therefore the case-law, and in particular the [*Meroni* judgement], which laid down the principles governing the delegation of powers. Next, the [*ESMA* judgement] applied

agencies' powers/output and many further references; for the soft law powers of agencies see W Weiß, Agenturen 634–637; for the use of agencies as a new mode of governance more generally see Dawson, Waves 213.

<sup>2526</sup> See the proposal of the *de Larosière* Group to vest the ESAs with the power to adopt 'binding supervisory standards'; Report of the High-Level Group on Financial Supervision in the EU (February 2009; so-called *de Larosière* Report) 55 f. This proposal was dropped by the Commission 'as it would conflict with the Treaty based responsibilities of the Commission and give the Authorities dis‐ cretionary powers, requiring a revision of the Treaty'; Commission Staff Working Document, SEC(2009) 1234, 13; see Orator, Möglichkeiten 472 f, with regard to a mechanism empowering the ACER.

<sup>2527</sup> Case C-270/12 *United Kingdom v European Parliament/Council*, para 80; see also Nettesheim, Art. 291 AEUV, paras 38 f; Chamon, Line 1633. Arguing against those who claim decision-making powers of agencies *vis-à-vis* MS to constitute a dis‐ tortion of the EU's institutional balance: Fischer-Appelt, Agenturen 121f; confirm‐ ing the legal possibility of vesting European agencies with implementing powers (which principally belong to the Commission's *pouvoir* according to Article 291 TFEU): Orator, Möglichkeiten 402; for the meaning of Article 277 TFEU in the given context see W Weiß, Agenturen 647 f.

those principles to cases where autonomous powers have been conferred on an agency by the EU legislature'.<sup>2528</sup>

Let us have a closer look at Article 291 TFEU now. This provision allows for the Commission (exceptionally: the Council) – on the basis of a ma‐ terial EU competence<sup>2529</sup> – to be vested with implementing powers in cer‐ tain cases.<sup>2530</sup> There appears to be a general assumption that the Commis‐ sion may adopt both individual-concrete and general-abstract measures, addressed to one or more MS or individuals, as the case may be.<sup>2531</sup> While European agencies are not even implicitly mentioned in Article 291 TFEU, against the background of the Court's case law quoted above it cannot be read as excluding European agencies to be vested with like powers.<sup>2532</sup>

<sup>2528</sup> Case T-510/17 *Del Valle Ruíz*, para 208; apparently in favour of applying only the criteria laid down in the *ESMA* judgement when assessing European agencies: case T-755/17 *Germany v ECHA*, paras 138 f.

<sup>2529</sup> See Schütze, Rome 1398 f, with regard to the question whether Article 291 para 2 TFEU could be seen as an independent legal basis for implementing powers of the Commission (or the Council).

<sup>2530</sup> For the wide discretion the creator of the legally binding Union act has with regard to the implementing powers it delegates see case C-427/12 *Commission v European Parliament/Council*, para 40; for the subsidiarity of the empowerment of the Commission (or the Council) under Article 291 para 2 TFEU see Ruffert, Art. 291 AEUV, para 2.

<sup>2531</sup> See case 16/88 *Commission v Council*, paras 11 and 16; case C-440/14P *Iranian Oil*, para 36; see also Ilgner, Durchführung 178; U Stelkens, Rechtsetzungen 385, with many further references; for arguments against the lawfulness of basing the Commission's power to adopt individual-concrete acts on Article 291 para 2 TFEU: ibid 285 f; W Weiß, Agenturen 645.

<sup>2532</sup> See also Michel, Gleichgewicht 115. Heed the example of regulatory technical standards (RTS) and implementing technical standards (ITS), to be adopted by the Commission with the support of the ESAs (pursuant to Articles 10 ff and 15 of Regulations 1093–1095/2010 which are again referring to Article 290 and Article 291 TFEU); see also Simoncini, Regulation 79 ff; forward-looking: Euro‐ pean Parliament, Resolution on the implementation of the legal provisions and the Joint Statement ensuring parliamentary scrutiny over decentralised agencies, P8\_TA(2019)0134, recital 21. For the involvement of comitology also in the case of ITS see Moloney, Rule-Making 74 f; for the relativisation of Commission power see Harlow/Rawlings, Process 281–283; critically: statement of the Commission in relation to Articles 290 f TFEU in: Council, Addendum to 'I/A' item note (10 November 2010), 15649/10 ADD 1, 1; see also case C-146/13 Spain v European Parliament/Council, paras 77 f, in which the Court could be understood to suggest that actors other than the Commission (or the Council) may be vested with implementing powers where 'uniform conditions for implementing legally binding Union acts' are not needed.

#### *V. CLASSIFICATION AND LEGAL ASSESSMENT OF COMPLIANCE MECHANISMS*

What follows from the empowerment of European agencies in this con‐ text is that the guarantees laid down in Article 291 para 2 TFEU and, on that basis, in Regulation 182/2011 are being thwarted. In legislative practice, that is to say in the acts providing for agencies' powers, no provision is made for MS control, eg by analogy to Article 291 TFEU. However, it is hard to imagine that where the Commission's powers have to be made subject to specific MS control, agencies may exercise the same kind of powers without such control.<sup>2533</sup> An assumption to that effect could lead to a legislative flight into 'implementation through agencies', the legislator thereby side-lining the MS control provided for in Article 291 TFEU. It was argued that the 'openness' of this provision – listing implementation by the MS, the Commission and the Council as possibilities – suggests that also other actors may be involved in one or the other way.<sup>2534</sup> The wording of Article 291 TFEU does not exclude an involvement of other bodies, such as European agencies, it is true, but only where it is limited to an advisory role.2535 As soon as MS may be addressed by these bodies, if only by soft law, in my view this would go beyond a merely advisory, assisting capacity. It would constitute the actual *exercise* of implementing powers.

It could be argued that the composition of agencies' main decision-mak‐ ing bodies allows for MS representation pursuant to Article 291 para 3 TFEU anyway,2536 rendering superfluous the application of comitology.<sup>2537</sup>

<sup>2533</sup> Sceptically: Volpato, Delegation 120 (also with regard to the ESMA judgement); for alternative (and diversified) accountability regimes for agencies see ibid 99f, with further references.

<sup>2534</sup> See W Weiß, Agenturen 641.

<sup>2535</sup> That the Commission may take advice from various actors when exercising im‐ plementing powers does not follow from the wording of Article 291 TFEU, but it is clear from the Commission's traditionally applied practice (also under the predecessors of this provision), that is: to regularly take advice from a variety of different sources.

<sup>2536</sup> See also W Weiß, Agenturen 656 f. For the representation of each MS as the regular case see Busuioc, Agencies 724; for other voices in the literature in favour of the lawfulness (with a view to Article 291 TFEU) of vesting agencies with implementing powers see references in Volpato, Delegation 97.

<sup>2537</sup> See Schütze, Rome 1423 f, with further references; W Weiß, Agenturen 657 f. For the similarity of comitology committees and agencies in that 'they juxtapose technical and scientific information with political discourse': Everson/Joerges, Europeanisation 530. In case of the technical implementing standards according to Article 15 of Regulations 1093–1095/2010, for example, it appears that (also) the legislator was convinced of the adequacy of replacing – some argue: in the form of a *lex specialis* to Regulation 182/2011 – control through comitology committees by

However, first, there have been decision-making boards of agencies in which not all MS were represented<sup>2538</sup> and, second, it is not always MS representatives in the strict sense of the word of which these boards are composed.<sup>2539</sup> In more recent agencies, such as the SRB, the ESAs or the ACER, it is representatives of independent national authorities taking the most important decisions, not ministerial bureaucrats which are bound by (political) instructions from their respective ministers.<sup>2540</sup> A board com‐ posed of such persons is not 'a form of Council of Ministers writ small'<sup>2541</sup> (comparable to comitology committees) anymore, but a body of independ‐ ent experts, one per each MS.<sup>2542</sup> The latter characteristic also applies to the Commission, by the way, whose general absolution from MS control under Article 291 TFEU would be simply unconceivable. Thus, it is not conclusive that European agencies should be generally exempted from the

2542 See also Chiti, Governance 52 f; Ruffert, Unabhängigkeit 407, with a further refer‐ ence. It ought to be mentioned that also comitology committees may potentially involve other experts than officials; see Schmidt-Aßmann, Verwaltung 1393.

control through European agencies; see Weismann, Agencies 130–132, with further references.

<sup>2538</sup> See Chamon, Agencies 66 f.

<sup>2539</sup> Due to the broad scope of the respective provisions, arguably MS would be free to appoint an independent person as their respective representative. However, overall the secondment of officials from the national ministries seems to be the rule; see III.3.7.2.1. above; see also Egeberg/Trondal, Agencies 871, with a further reference; Simoncini, Regulation 137 f; with regard to the MS' practice, taking the example of Europol: Groenleer, Autonomy 283. Only exceptionally the founding act requires the appointment (by each MS) of an independent person; see Article 12 para 1 lit a of Council Regulation 168/2007, with regard to the European Union Agency for Fundamental Rights.

<sup>2540</sup> With the ESAs it is to be noted that they also dispose of a Management Board composed of seven members of the Board of Supervisors (including its Chairper‐ son). The SRB is composed, in addition to the representatives of the national resolution authorities (and non-voting members of the Commission and the ECB), of five full-time members (including its Chair) who – on their own – also form the executive session of the Board which is vested in particular with management and preparatory tasks. The main decision-making body of ACER, its Administra‐ tive Board, is composed of nine members appointed by the Council (five), the Commission and the EP (two each). Its Board of Regulators, on the contrary, is composed of one senior representative of the MS' regulatory authorities; see Articles 18 and 21 of Regulation 2019/942.

<sup>2541</sup> Hertig/Lee, Predictions 8, with regard to the ESC. For the professional qualities of the Board members as required in EU law (in particular: the agencies' founding acts) see Chamon, Agencies 71.

requirements laid down in Article 291 TFEU and in Regulation 182/2011 when exercising implementing powers.<sup>2543</sup>

The above questions related to the EU's institutional balance are pressing where the *final* output *vis-à-vis* a MS in a (secondary law-based) compli‐ ance mechanism is adopted by an agency, but also where an agency's soft law output (*vis-à-vis* a MS) may be followed by Commission output they are by no means irrelevant. Also in the latter case the agency contributes to settling the matter, not least because the procedure in practice may end at the soft law level (due to compliance on the part of the MS concerned), without the MS as a whole having a possibility to exert control in accord‐ ance with Regulation 182/2011.

In practice, also the Commission and the Council have been granted implementing powers or at least *prima facie* implementing powers under different secondary law-based regimes.2544 In particular, it has occurred that the Commission was, by a delegating/implementing act, vested with the power to adopt implementing acts, with no provision being made for MS control according to the Comitology Regulation.2545 In the context of soft law, the involvement of comitology committees anyway appears to be the exception rather than the rule.2546 Such bypassing of comitology under Article 291 para 2 raises serious concerns. There is a risk inherent in comitology (namely that the MS participate in controlling their own com‐ pliance, and hence may 'protect' each other from administrative action2547), it is true, but this cannot lead to ignoring a clear principle of primary law, as enshrined in Article 291 TFEU,2548 the creators of which certainly took this risk into account.

<sup>2543</sup> See also Wörner, Verhaltenssteuerungsformen 371, with regard to the ESAs; for the motives of the legislator to empower European agencies instead of taking the route of comitology-based implementation by the Commission: Armstrong, Character 193 f, with further references.

<sup>2544</sup> See case C‑521/15 *Spain v Council*, and its analysis above (3.1.1.2.3.).

<sup>2545</sup> See F Schmidt, Art. 291 AEUV, para 13. For the example of an even more far-reach‐ ing implementing power assumed by the Commission see Andersen, Enforcement 133 f.

<sup>2546</sup> That comitology may also be applied to soft law is argued by Senden, Rulemaking 70; for the 'under-proceduralisation' of the adoption of soft law more generally see Chamon, Regulators 334; for the procedures for the adoption of EU soft law (by European agencies) see also Rocca/Eliantonio, Soft Law 15 ff.

<sup>2547</sup> See also Gil Ibáñez, Exceptions 170 f.

<sup>2548</sup> The bypassing of comitology was considered to be unlawful in joined cases T-261/13 and T-86/14 *Netherlands v Commission*, paras 49 f; with regard to insti‐

### 3.4. The principles of subsidiarity and proportionality

### 3.4.1. Introduction

The principles of subsidiarity and proportionality are important precepts of EU law which apply to all EU bodies, regardless of whether they per‐ form legislative, executive or judicative functions. While the principle of subsidiarity applies only in areas which do not fall within the exclusive competence of the EU, the principle of proportionality applies throughout. According to the former, 'the Union shall act only if and in so far as the objectives of the proposed action cannot be sufficiently achieved by the Member States, either at central level or at regional and local level, but can rather, by reason of the scale or effects of the proposed action, be better achieved at Union level'.2549 The latter – a traditional legal principle in many legal orders2550 – under EU law requires that 'the content and form of Union action shall not exceed what is necessary to achieve the objectives of the Treaty'.2551 The question of the necessity of compliance mechanisms in addition to the Treaty infringement procedure (and other primary law tools aimed at increasing compliance, such as the preliminary reference procedure or the claim for State liability) in particular is to be understood against the background of what has been dubbed the EU's 'commitment-compliance gap'.<sup>2552</sup>

In the following, the question whether the compliance mechanisms laid down in secondary law (as presented above) are in accordance with the principle of subsidiarity and the principle of proportionality shall be ad‐ dressed.

tutional balance concerns of this practice more generally: Bast, Categories 923; Senden/van den Brink, Checks 38.

<sup>2549</sup> Article 5 para 3 TEU. Note that the word 'therefore' known from the Maastricht wording of this principle has now been abolished, arguably because it unduly links the examination of the two factors. It was/is a common misunderstanding that if the MS cannot sufficiently achieve a certain objective, the Union is *therefore* in a better position to do so. There may be cases in which there is a correlation with that result, but there is certainly no such causality; critically: Schima, Subsidiari‐ tätsprinzip 107.

<sup>2550</sup> For an early reference to proportionality in international case law see case *Portugal contre Allemagne*, Report of International Arbitral Awards II, 1028 (D.2.).

<sup>2551</sup> Article 5 para 4 TEU.

<sup>2552</sup> See eg Börzel, Governance 12.

### 3.4.2. The principle of subsidiarity

#### 3.4.2.1. Compliance mechanisms in general

The outdated Protocol on the Application of the Principles of Subsidiarity and Proportionality (as annexed to the Treaty of Amsterdam) – which was more extensive on the character of subsidiarity than the current Protocol No 2 – described subsidiarity as 'a guide as to how those powers are to be exercised at the Community level'; and further: '[s]ubsidiarity is a dynamic concept and should be applied in the light of the objectives set out in the Treaty. It allows Community action within the limits of its powers to be expanded where circumstances so require, and conversely, to be restricted or discontinued where it is no longer justified'. While this Protocol is not in force any more, and while the new Protocol has not taken over the quoted passages, they still appear to be a valid account of the subsidiarity principle.

It has been said that the principle of subsidiarity limits the exercise of the EU's competences.2553 In order to do justice to this principle, EU actors shall always consider – in the words of Protocol No 2: 'ensure constant respect'2554 to – subsidiarity when acting on the basis of an EU competence which is not exclusive.2555 In this vein, the Commission has opted for proposing to the legislator the adoption of a directive instead of a regulation in a number of cases.2556 In the given context, it is not only the legislative, but also the executive competence which stays in the foreground. The EU has exclusive legislative competence in the policy fields listed in Article 3 para 1 TFEU, but the administrative execution ('implementation') of the legislation adopted in these fields is, pursuant to Article 291 para 1 TFEU, normally up to the MS.2557 Provision for executive tasks of EU bodies

2556 See Craig/de Búrca, EU Law 126.

<sup>2553</sup> See eg Craig/de Búrca, EU Law 126; for a different understanding: Goos, Gedank‐ en 39 f.

<sup>2554</sup> Article 1 of Protocol No 2 on the Application of the Principles of Subsidiarity and Proportionality.

<sup>2555</sup> For the limits of this duty to consider see case C-233/94 *Germany v European Parliament/Council*, paras 26–29, according to which an explanation of the legisla‐ tor as to why it considers that its action is in conformity with the principle of subsidiarity is sufficient, an explicit reference to the principle of subsidiarity not required. For the requirement of legislative drafts to be 'justified with regard to the principles of subsidiarity and proportionality' see Article 5 of Protocol No 2 on the Application of the Principles of Subsidiarity and Proportionality.

<sup>2557</sup> See eg Obwexer, Art. 2 AEUV, para 10.

enshrined in a legislative act (eg the compliance mechanisms at issue here) is only lawful where the Union also has an according executive competence (eg Article 108 TFEU as a specific competence or, as a general competence, Article 291 para 2 TFEU, or Article 352 TFEU).<sup>2558</sup> This applies also to the right to give individual-concrete instructions to national authorities as a form of indirect execution by EU bodies.<sup>2559</sup> On a whole, the Court has tak‐ en a generous approach in that respect, implying – with regard to what is now Article 114 TFEU<sup>2560</sup> – such competence to an EU (legislative) compe‐ tence (see also 3.2.2.2. above).<sup>2561</sup> The outdated Protocol on the Application of the Principles of Subsidiarity and Proportionality states more broadly that '[t]he application of the principles of subsidiarity and proportionality […] should take into account Article F(4) of the Treaty on European Union, according to which "the Union shall provide itself with the means necessary to attain its objectives and carry through its policies"' (para 2). While 'well established national arrangements and the organisation and working of Member States' legal systems' should be respected, '[w]here appropriate and subject to the need for proper enforcement, community measures should provide Member States with alternative ways to achieve the objectives of the measures' (para 7).

With regard to the compliance mechanisms presented here, it is apparent that many of them are established in a field in which the MS are princi‐ pally in charge of implementation. The question whether a compliance mechanism was established on an appropriate legal basis was assessed above (3.2.). What matters in the context of subsidiarity is whether, on the assumption that the legal basis chosen by the legislator is appropriate, the action performed by EU actors in the course of a given procedure can actually be better achieved at Union level – in other words: can be achieved only worse at MS level. As regards supervisory mechanisms in which the Commission acts as a supervisor, the Court has once declared that the application of certain EU law provisions (on export refunds) 'is a matter for the national bodies appointed for this purpose and that the Commission

<sup>2558</sup> Leaving this question open: case T‑31/07 *Du Pont de Nemours*, paras 203–205; for this topic see also Schütze, Rome.

<sup>2559</sup> Arguing against a general right of EU bodies to give individual-concrete instruc‐ tions: Biaggini, Theorie 115; Constantinesco, Recht 299.

<sup>2560</sup> For the wide wording of Article 114 TFEU ('measures') which supports such generosity see Schütze, European Union Law 336.

<sup>2561</sup> See case C-359/92 *Germany v Council*, paras 37 ff; case C-66/04 *United Kingdom v European Parliament/Council*, paras 47–50.

has no power to take decisions on their interpretation but may only express its opinion which is not binding upon the national authorities'.<sup>2562</sup> In a different context, the Court held that it is not for the EU and its bodies to 'act in place of the Member States and to prescribe for them the measures which they must adopt'.<sup>2563</sup> Only to the extent that these measures have the effect of distorting the conditions of competition in the internal market – that is to say: the functioning of the internal market (uniform application of the pertinent EU law) – the EU does have a competence to act, thereby taking due account of the MS' discretion.<sup>2564</sup>

Compliance mechanisms are not primarily about replacing MS' execu‐ tive powers. Normally they apply where national execution in a concrete field or category of situations has turned out to be dysfunctional. Against this background, *Eekhoff* refers to the typical conflicts of interest of the MS which need to be enclosed by empowering an independent EU body, and hence – with regard to 'klassische Aufsichtsverfahren' [classic supervisory processes] – argues in favour of accordance with the principle of subsidiari‐ ty.2565 Also where the MS may make use of legal exceptions under EU law, the risk of MS pursuing their individual interests at the cost of the EU's objectives is apparent. Thus, with a view to subsidiarity considerations, a control regime performed by EU bodies is feasible. Another pertinent case is the adoption of safeguard and/or emergency measures. Here the EU in places may be better equipped to act than the MS.2566 While maintaining the public order and safeguarding the internal security falls within the

<sup>2562</sup> Case 133/79 *Sucrimex*, para 16.

<sup>2563</sup> Case C-265/95 *Commission v France*, para 34.

<sup>2564</sup> Case C-265/95 *Commission v France*, para 35.

<sup>2565</sup> See Eekhoff, Verbundaufsicht 207 f, referring to the Final Act of the Conference of the Representatives of the Governments of the MS (Amsterdam, 2 October 1997), 43th Declaration, according to which 'the administrative implementation of Community law shall in principle be the responsibility of the Member States in accordance with their constitutional arrangements. This shall not affect the supervisory, monitoring and implementing powers of the Community Institutions as provided under Articles 145 and 155 of the Treaty establishing the European Community'. In other cases, most prominently in the field of competition law, subsidiarity considerations (among others) have lead to a de-centralisation of ad‐ ministrative execution by Regulation 1/2003; see Commission, 'Better Lawmaking 2003' (Report), COM(2003) 770 final, 20 f. For the 'tension between the collective Community interest and that of individual [MS]' in the context of subsidiarity see Craig, Subsidiarity 76 f.

<sup>2566</sup> See Gil Ibáñez, Tools 5.3., who says: 'In fact, it can be said that a specific area should not serve to justify a specific procedure, but the specificity of the situation

MS' competences,2567 the Commission was vested with special powers of control in order to prevent protectionist measures adopted by the MS uni‐ laterally *sub titulo* 'safeguard measures' from disturbing the functioning of the internal market.<sup>2568</sup> Even at the level of primary law such constellations can be found. When it comes to restrict the fundamental freedoms for certain justificatory reasons, such as public morality, public policy, public security, protection of health and life of humans, animals or plants, the MS – to the extent the details of these justificatory reasons have not already been harmonised by legislation – enjoy a measure of discretion.<sup>2569</sup> Where harmonisation measures have been adopted on the basis of Article 114 TFEU, the regime of its paras 4 ff may apply (see IV.2.2.1.1.3. above). Articles 346–348 TFEU follow a similar logic with regard to MS' national security.

When assessing a planned initiative with a view to its compliance with the subsidiarity principle, the most important considerations for the 'Union relevance' are 'the geographical scope, the number of players affected, the number of Member States concerned and the key economic, environmental and social impacts. In addition, the analysis determines in qualitative – and as far as possible in quantitative – terms, whether there is a significant cross-border problem'.2570 These manifold considerations2571 – in combina‐ tion with other factors, such as the Commission's focus with regard to the question whether MS action *is*2572 (not: *can be*, as the wording of Article 5 para 3 TEU stipulates) sufficient and the lenient case law of the Court<sup>2573</sup> –

<sup>(</sup>i.e., the urgency or the possibility of irreparable damage), which usually will apply to more than one area'.

<sup>2567</sup> See Article 72 TFEU; see also case C-265/95 *Commission v France*, para 33.

<sup>2568</sup> See Andersen, Enforcement 179.

<sup>2569</sup> See eg Article 36 TFEU for the written justificatory reasons. For the unwritten ones see the pertinent case law starting with *Cassis de Dijon*.

<sup>2570</sup> Commission, Annual Report 2016 on Subsidiarity and Proportionality, COM(2017) 600 final, 3.

<sup>2571</sup> The risk of diversity in the application of EU law may be said to be inherent in the sheer fact that it is – at the national level – enforced by 27 different administrations; see Craig, Subsidiarity 76, with examples.

<sup>2572</sup> Commission, Annual Report 2016 on Subsidiarity and Proportionality, COM(2017) 600 final, 3: 'assess whether action at national, regional or local level is sufficient to achieve the objective pursued'.

<sup>2573</sup> See case T-339/04 *France Télécom*, para 73 and the critical discussion by Schütze, Rome 1414 f; see also more generally Kadelbach, Art. 5 EUV, para 47, with referen‐ ces to the pertinent case law. The Court seems to be mainly concerned about compliance with the minimum requirements of the duty to give reasons on the part of the EU actors. Note that the explicit duty of a justification with regard to

suggest that the subsidiarity test is performed with a certain bias in favour of Union action.<sup>2574</sup> The recent multiple crises – in places disclosing the dysfunctioning of traditional (weaker) forms of cooperation among MS authorities – may have reinforced this tendency.<sup>2575</sup>

# 3.4.2.2. Implementing and enforcement mechanisms in particular

Also with regard to implementing mechanisms, the core question in this context is: Can implementation be sufficiently achieved by the MS or can it rather be better achieved at Union level? Where Article 291 para 2 TFEU is applied, this question essentially boils down to: Are uniform conditions required for implementation? If that is the case, then there will regularly be no subsidiarity concerns because uniform conditions can regularly be better achieved by one actor (the EU) than by 27 actors.<sup>2576</sup>

With regard to mechanisms displaying an enforcement tendency the matter is less clear. The enforcement of EU law *vis-à-vis* the MS is nothing which the MS could possibly do themselves. They can comply with EU law, in which case enforcement procedures will not have to be applied, but that is a different issue. The introduction of a new enforcement measure *vis-à-vis* the MS does not even impose a new burden on the MS, since even before its introduction EU law could be enforced via the Treaty infringe‐ ment procedure. The question whether an enforcement mechanism is nec‐ essary in view of the existence of the Treaty infringement procedure may

the principles of subsidiarity and proportionality in Article 5 of Protocol No 2 only relates to legislative acts. For lack of a specific rule, with regard to other acts (also soft law acts) the regular reasoning requirements apply also with regard to subsidiarity/proportionality; for these requirements see case C-62/14 *Gauweiler*, para 70: 'not required to go into every relevant point of fact and law'; critically as regards the sometimes substantially weak reasoning of the Commission: Wittinger, Satelliten 616 f.

<sup>2574</sup> See Commission, Report 'Better Lawmaking 2003', COM(2003) 770 final, 23, in which – as a justification for establishing the REACH regime – the large number of MS (by then: soon 25) and the technical complexity of the matter (ie the registration, evaluation, authorisation and restriction of chemicals) are pointed at.

<sup>2575</sup> See Moloney, Rules in Action 219, with further references; Senden/van den Brink, Checks 21. With regard to an increased generosity *vis-à-vis* grand anti-crisis measures more generally see Fahey, Emperor 582 and 594 f.

<sup>2576</sup> For heterogeneous legal frameworks in the MS as a justification for Union action under the principle of subsidiarity more generally see case C-491/01 *British Ameri‐ can Tobacco*, paras 182 f.

reasonably be posed, but not in the context of 'subsidiarity'.2577 However, where – in a macro-perspective – the enforcement of EU law *vis-à-vis* indi‐ viduals/undertakings is at issue (which is the case with most compliance mechanisms addressed above), the question *de lege ferenda* may very well be: Either enforcement by the national authorities on their respective own (potentially combined with some institutional or procedural requirements under EU law), or combined with a monitoring regime governed by an EU body *vis-à-vis* the national authorities in charge,<sup>2578</sup> or enforcement *visà-vis* the individuals/undertakings directly by an EU body? Option one is least, option three most likely to interfere with the principle of subsidiarity. With regard to the monitoring by the EU body, again it can be argued that this is nothing which the MS on their own could possibly take care of. The aforementioned old Protocol is more telling in this respect than the word‐ ing of Article 5 para 3 TFEU. It states, *inter alia*, the following: 'Regarding the nature and the extent of Community action, Community measures should leave as much scope for national decision as possible, consistent with securing the aim of the measure and observing the requirements of the Treaty' (para 7). Applied to our context, these considerations may result, in a concrete case, in the legislator's decision in favour of introducing a mixed rather than a hard mechanism or a soft rather than a mixed mechanism.

In view of the, materially speaking, questionable justiciability of the sub‐ sidiarity principle before the Court, it appears that the implementing/en‐ forcement mechanisms presented above, where they are not predetermined to a large degree in primary law or falling within the exclusive competence of the EU anyway, are in accordance with the subsidiarity principle.2579 It is to be noted that all of them either relate to safety issues, are 'technical' to the extent that expertise beyond legal expertise is required and/or cater for exceptional interventions for the sake of the uniform application of law,<sup>2580</sup> in particular in cross-border cases. In general, these characteristics seem to

<sup>2577</sup> For the necessity criterion as part of the proportionality test, which is applicable to both implementing and enforcement mechanisms and which is examined (also) against the backdrop of the Treaty infringement procedure, see 3.4.3.1. below.

<sup>2578</sup> These are then either implementing or enforcement mechanism, depending on the concrete procedure envisaged.

<sup>2579</sup> For the (contested) issue of the justiciability of the principle of subsidiarity see Panara, Enforceability; Portuese, Subsidiarity.

<sup>2580</sup> Sceptically with regard to such direct intervention in the context of the subsidiarity principle: Gundel, Energierecht, para 58.

be a sound basis for justifying the limitation of what the old Protocol calls 'scope for national decision'.

With a view to the legislative proceedings leading to the adoption of our sample of compliance mechanisms, it ought to be mentioned that the Commission in the older legislative proposals did not find it necessary to provide for an in-depth consideration of the question of subsidiarity.<sup>2581</sup> The most recent proposals are more elaborate in this respect.<sup>2582</sup> Thus, there seems to be an increasing willingness on the part of the Commission to explain its considerations on the subsidiarity principle.<sup>2583</sup>

3.4.2.3. On the issue of soft law

The fact that for safeguard and/or emergency measures – due to the risks at stake – often fast action is required may be used as an argument against extensive exchanges of views between the EU and the MS level within the framework of the respective compliance procedure. By tendency, it also speaks against the use of individual-concrete EU soft law, as it prolongs the time until a legally binding decision is made – by the EU body in charge or by the MS authority, as the case may be. However, the latter reservation can be met by setting in place adequate (tight) deadlines for reaction (to soft law).

Recalling the above quotation from the old Protocol, according to which 'the nature and the extent of Community action […] should leave as much scope for national decision as possible', we can infer that also (the power

<sup>2581</sup> It did provide for a brief consideration of subsidiarity in the following propos‐ als: COM(2009) 501 final, 3 f, leading to Regulation 1093/2010; COM(2010) 527 final, Recital 16, leading to Regulation 1176/2011; COM(2010) 525 final, Recital 15, leading to Regulation 1174/2011; COM(2013) 27 final, Recital 36 and page 6, presenting different policy options that where weighed in the light of varying considerations, among them the subsidiarity principle, and leading to Regulation 2016/796; COM(2009) 499 final, Recital 20, leading to Regulation 1092/2010.

<sup>2582</sup> See Proposals COM(2015) 613 final, 4 f, leading to Regulation 2018/1139; COM(2013) 520 final, 6 f, leading to Regulation 806/2014; COM(2016) 864 fi‐ nal, 10 f, leading to Directive 2019/944; COM(2016) 863 final, 10 f, leading to Regulation 2019/942; COM(2016) 861 final, 9 f, leading to Regulation 2019/943; COM(2016) 590 final/2, 4 f, leading to Directive 2018/1972.

<sup>2583</sup> This does not contradict the arguable bias in favour of Union action referred to above (3.4.2.1.).

to adopt) soft law acts must meet the subsidiarity threshold.<sup>2584</sup> At the same time, it ought to be stressed that soft law – by leaving a leeway for MS in their respective decision-making – in principle may serve the facilitation of the subsidiarity principle.<sup>2585</sup> This is why the use of soft law in mixed and soft compliance mechanisms *prima vista* is to be welcomed with a view to the subsidiarity principle.<sup>2586</sup> That is, in principle, also acknowledged by EU actors, eg the Commission.<sup>2587</sup> From this it does not follow, though, that soft law powers on the part of EU bodies meet the subsidiarity principle by default.<sup>2588</sup> Nor does it mean that where the subsidiarity principle applies EU soft law must always be considered as a first choice. It appears that regulation by law regularly is the more promising approach with a view to what the old Protocol calls 'securing the aim of [a] measure and observing the requirements of the Treaty', one of the latter being legal certainty.

3.4.3. The principle of proportionality

3.4.3.1. The Treaty infringement procedure as the elephant in the room

The principle of proportionality is traditionally examined in the course of a test, in EU law most prominently when it comes to the infringement of human rights or fundamental freedoms.2589 This test regularly encompasses the following criteria: objective of general interest, suitability, necessity, and proportionality in the narrow sense, the latter assessing the meansends relationship of the measure. The proportionality test may also be

<sup>2584</sup> See Majone, Agencies 267–269; Snyder, Institutional Practice in the European Community 202 f. This also seems to be the understanding of the legislator; see Commission, Report 'Better Lawmaking 2003', COM(2003) 770 final, 8. For the (non-)applicability of the old protocol with regard to soft law acts and for the risk for subsidiarity the use of soft law may pose see Knauff, Regelungsverbund 413– 415; Wörner, Verhaltenssteuerungsformen 288–290 and 354.

<sup>2585</sup> See Senden, Soft Law 23 and 206 f: 'preference for recommendations and similar instruments over legislation whenever possible'; Rawlings, Soft law 230.

<sup>2586</sup> See also Peters, Soft law 33. Describing international soft law as a measure suitable to protect national sovereignty: Kanehara, Considerations 85.

<sup>2587</sup> See examples provided by Andone/Greco, Burden 90 f.

<sup>2588</sup> See Braams, Koordinierung 204, with further references.

<sup>2589</sup> For an example from the field of Common Agricultural Policy see case 265/87 *Schräder*, para 21.

applied in the given context,2590 when the legislator creates compliance mechanisms, irrespective of whether they qualify as implementing or as enforcement mechanisms.<sup>2591</sup> The implementation and the enforcement of EU law generally qualify as objectives of the Treaties and hence meet the criterion 'objective of general interest'. With implementation and enforce‐ ment mechanisms deserving this name, suitability can be taken for granted. When it comes to the necessity criterion, with enforcement mechanisms inevitably and in each case the question must be raised whether in partic‐ ular the Treaty infringement procedure – the regular and universal EU enforcement mechanism applied *vis-à-vis* the MS – is insufficient. But also other (existing) mechanisms may be taken into account (eg Article 114 paras 4 ff TFEU).<sup>2592</sup> These sufficiency arguments must be included in the examination of necessity, and only if the mentioned procedures turn out to be insufficiently effective in the policy field at issue, the creation of a new mechanism can be deemed necessary. Proportionality in the narrow sense means that the means applied (ie the compliance mechanism at issue) may not be disproportionate to the ends achieved (ie – at best – enhanced compliance rates).

The argument that the Treaty infringement procedure is too burdensome in principle and hence should be 'complemented' by various compliance mechanisms laid down in secondary law cannot be accepted without fur‐ ther specifications. The Commission, for example, has uttered its discontent with the Treaty infringement procedure on many occasions, claiming that in certain policy fields it is inappropriate to ensure compliance in due

<sup>2590</sup> For the different thrust of the proportionality test, depending on the context in which it is applied, see Schima, Art. 5 EUV, para 71.

<sup>2591</sup> For the (alternative) primary aims of compliance mechanisms see 3.1.1.2.1.2. above.

<sup>2592</sup> Since implementing and enforcement mechanisms both aim at ensuring compli‐ ance, it is in particular the existence of the Treaty infringement procedure (and its variants) and implementing mechanisms laid down in primary law, but, apart from that, also compliance mechanisms established under secondary law which need to be taken into account. In a broader perspective, also other mechanisms aimed at ensuring compliance with Union law, such as the preliminary reference procedure, ought to be considered.

time,2593 and that complementary mechanism may be useful.<sup>2594</sup> During the legislative procedure for what later became Council Regulation 2679/98 on the functioning of the internal market in relation to the free movement of goods among the MS, which allows the Commission to demand informa‐ tion from a MS in which 'an obstacle to the free movement of goods among Member States which is attributable to [that] Member State'<sup>2595</sup> occurs and, if need be, to request the MS concerned to take certain measures, the Commission held: '[T]he procedures provided under Articles 169 [now: Article 258 TFEU] and 186 of the Treaty [now: Article 279 TFEU] are not suitable for removing this obstacle in due time' in view of the objectives of the Regulation which are 'to ensure rapid restoration of the free movement of goods when it is impeded in such a way as to seriously disrupt the proper functioning of the internal market'.<sup>2596</sup> It concluded that '[t]hese are consequently special situations to which the appropriate response is specific means of action. The proportionality of the proposed mechanism is therefore based essentially on the speed and the binding force of the Com‐ mission's intervention in response to the situations described above'.<sup>2597</sup>

In the field of feed and food safety, to take another example, the Com‐ mission complained that '[a]lthough this procedure [ie the Treaty infringe‐ ment procedure] is a powerful instrument, the time constraints imposed

<sup>2593</sup> Note that alternative compliance mechanisms are not only a result of the insuffi‐ ciency of the Treaty infringement procedure, but may also be seen as instruments to prevent the Treaty infringement procedure from becoming less efficient, be‐ cause they provide for an interpretation of EU law by an EU body and thus arguably reduce the likelihood of Court proceedings (thereby contributing to faster decision-making by the Court of Justice in the cases submitted to it). As *Hofmann*, *Rowe* and *Türk* said (in a different context, but still): '[I]t would be highly undesirable for reasons of efficiency, and in respect of the workload of the Court, were every issue of interpretation and application of European law to be resolved purely through litigation'; Hofmann/Rowe/Türk, Administrative Law 569; for exemplary evidence of high compliance rates of such preventive regimes see Schmidt/Schmitt von Sydow, Art. 17 EUV, para 42; see also Pelkmans, Recognition.

<sup>2594</sup> See Commission, 'Better Monitoring of the Application of Community Law' (Communication), COM(2002) 725 final, 15 f.

<sup>2595</sup> Article 1 para 1 of Council Regulation 2679/98.

<sup>2596</sup> Commission Proposal COM(97) 619 final, paras 11f.

<sup>2597</sup> Commission Proposal COM(97) 619 final, para 14. For the early warning system laid down in Council Regulation 2679/98 see Leible/Streinz, Art. 34 AEUV, para 138; for 'speedy enforcement' as a requirement potentially justifying the introduc‐ tion of an alternative compliance mechanism see Gil Ibáñez, Exceptions 166.

on it render it impractical where a failure to implement Community law requires prompt action to safeguard feed and food safety'.<sup>2598</sup>

Also the Court, in a case on the Commission's power to require a MS, under certain circumstances, to take certain temporary measures in order to tackle a serious and immediate risk from a product to the health and safety of consumers in various MS,<sup>2599</sup> conceded that 'even if [Treaty infringement procedures] were initiated and held by the Court to be well founded, it is not certain that a declaration by the Court to that effect would enable the objectives set out in the directive to be achieved as effec‐ tively as would be the case by a Community harmonisation measure'.<sup>2600</sup> The Court thereby addresses one of the main characteristics of the Treaty infringement procedure: its declaratory nature. While MS are required to take the necessary means to comply with a Court judgement,<sup>2601</sup> specific actions can be requested from MS only by means of special procedures (mainly to be found in secondary law).

### 3.4.3.2. The specific compliance mechanisms

As part of the examination of the necessity criterion, it is required to positively scrutinise whether the proposed mechanism could actually do away with the insufficiency of the existing regime. This can be affirmed eg where the procedure at issue is faster and experience (eg with a similar procedure in a comparable policy field) has shown high compliance rates. Thus, not only the drawbacks of the Treaty infringement procedure but also the (expected2602) effectiveness of the new mechanism needs to be scrutinised under the heading 'necessity'. For example: According to Direc‐ tive 2015/1535 MS have to notify the Commission *ex ante* of certain techni‐

<sup>2598</sup> Commission Proposal COM(2003) 52 final, Recital 56.

<sup>2599</sup> See the mechanism laid down in Article 9 of Council Directive 92/59/EEC.

<sup>2600</sup> Case C-359/92 *Germany v Council*, para 49. There is no doubt that the Treaty infringement procedure does not only aim at an authoritative declaration of law‐ fulness or unlawfulness of MS behaviour, but – in the latter case – also of bringing that behaviour to an end; see joined cases 15–16/76 *France v Commission*, para 27.

<sup>2601</sup> Article 260 para 1 TFEU.

<sup>2602</sup> While it is true that the actual effectiveness of a measure can only be measured *ex post*, empirical data such as experience with similar measures may reasonably predict a high degree of effectiveness also in the case at issue; for the numerous difficulties in measuring a norm's effectiveness see Shelton, Compliance 131 ff, focussing on international human rights norms.

cal rules they intend to adopt. In the course of a soft procedure (Articles 5 f ), the Commission can suggest amendments to these drafts where it does not deem them to be compliant with internal market law, more concretely: where it thinks that 'the measure envisaged may create obstacles to the free movement of [goods or services] within the internal market' (Article 7 para 2 *leg cit*). In order to emphasise the necessity of this mechanism, the legislator uttered that 'it is essential for the Commission to have the neces‐ sary information at its disposal before the adoption of technical provisions [by the MS]' in order to 'promote the smooth functioning of the internal market' (Recitals 5 and 3). What is necessary follows from the needs (ie the concrete problems) of the specific policy field in which a compliance mechanism is intended to be established.<sup>2603</sup> As a consequence, alternative compliance mechanisms may rather be introduced with a focus on specific, problematic situations. With a new general compliance mechanism it would – apart from other serious legal obstacles – be very difficult to prove its necessity.<sup>2604</sup>

With a view to proportionality in the narrow sense, it can be said that the more a compliance mechanism interferes with the MS' sphere, the more specific and problematic the situation thereby tackled must be. The legislator seems to have been guided by this thought when limiting the right of intervention on the part of the EU body in charge to 'clear and manifest' infringements by the MS.<sup>2605</sup>

In procedural terms, it is in particular the form the intervention takes which matters in this context. The legislator may prescribe measures on a scale ranging from relatively weak measures, eg a duty of the MS to inform the EU body in charge *ex post*, to relatively strong measures, eg the possibil‐ ity for the EU body in charge to arrogate a MS competence, in order to

<sup>2603</sup> See Gil Ibáñez, Exceptions 161.

<sup>2604</sup> Instead, the Commission has tried to improve the functioning of the existing regime; for efficiency-driven steps taken by the Commission in the context of the Treaty infringement procedure see Commission, 'EU law: Better results through better application' (Communication), 2017/C 18/02, 14–16; for other 'conservative' reforms (that is to say reforms not requiring a Treaty revision) see Gormley, Infringement 75, with further references.

<sup>2605</sup> See eg Article 8 of Directive 92/13/EEC or Article 3 para 1 of Directive 89/665/ EEC. For the difference between a clear and manifest infringement and a 'regular' infringement according to Article 258 TFEU see case C-359/93 *Commission v Netherlands*, para 14. For one of the compliance mechanisms addressed above, see Article 18 para 4 of Regulation 1093/2010: non-application or 'manifest breach' (and necessity of urgent remedying).

adjust the intensity of a compliance mechanism. Also the legal (non-)bind‐ ingess of the EU output provided for is to be borne in mind. The mere fact that the output adopted within the framework of a certain mechanism is legally binding cannot lead to the disproportionality of a mechanism, though (see III.3.3.2.1. above).<sup>2606</sup> Conversely, also the introduction of a soft mechanism is not *per se* a proportionate means of ensuring compliance. It can be not suitable (to achieve the aim of enhanced compliance) or it can constitute an overreaction (in proportionality terms: 'not necessary'). While the institutions' awareness of the need to justify the use of soft law also against the principle of proportionality is comparatively low,<sup>2607</sup> the considerate use both of soft and hard measures is an important procedural requirement to render a measure proportionate.

# 3.4.2.3. Inter-institutional proportionality considerations in the legislative process and the final decision-making power of the Court

The example of Commission Proposal COM(2000) 162 final shows that, with a view to the co-decision procedure, the trilogue between Commis‐ sion, Council and EP – institutions which often have different agendas in a given policy field – serves well the purpose of 'sanding off' the original Commission proposal. This may result in the suppression of ambitious innovations, but it may also do away with disproportionalities in the pro‐ posal. The proposal referred to here suggested providing the Commission *inter alia* with the power to suspend the putting into circulation within the Community and exports to third countries of a product to be used in animal nutrition which was likely to pose a serious risk to human or animal health or to the environment. In case of emergency, it should be possible to take this measure without prior consultation of the MS. The Council pointed out that – were the Directive adopted as proposed by

<sup>2606</sup> See Knauff, Regelungsverbund 412 f (fn 88). However, the adoption of hard law may very well go beyond the EU's competences. The line between these two questions – Is a certain measure proportionate? Does the adoption of a certain measure fall within the competence of the EU (and the EU body at issue)? – may sometimes be particularly fine; against the background of the proportionality principle – and given that the relevant criteria are met – in favour of a general preferability of soft law: R Geiger/Kirchmair, Art. 5 EUV, para 18; Schima, Art. 5 EUV, para 73, with further references.

<sup>2607</sup> See Andone/Greco, Burden 91.

the Commission – there would be a 'risk of abuse' on the part of the Commission. The result of the legislative procedure, Directive 2001/46/EC, relied more strongly on MS powers and, for the case that the Commission interferes, for stronger MS involvement.<sup>2608</sup> While here the main issue was dubbed 'risk of abuse', this can be translated into 'possibility of an excess' which reveals that it is essentially a question of proportionality.

Eventually, the proportionality of a measure is decided upon by the Court, applying a relatively liberal threshold ('manifestly incorrect'<sup>2609</sup>). In assessing the relevant factors, the Court has emphasised the 'legislature's broad discretion, which implies limited judicial review of its exercise'.<sup>2610</sup> This limited judicial review, however, requires 'that the Community institu‐ tions which have adopted the act in question must be able to show before the Court that in adopting the act they actually exercised their discretion, which presupposes the taking into consideration of all the relevant factors and circumstances of the situation the act was intended to regulate. It follows that the institutions must at the very least be able to produce and set out clearly and unequivocally the basic facts which had to be taken into account as the basis of the contested measures of the act and on which the exercise of their discretion depended'.2611 The Court has stressed that especially in a policy field 'which entails political, economic and social choices on [the part of the legislator], and in which it is called upon to undertake complex assessments […] the legality of a measure […] can be affected only if the measure is manifestly inappropriate having regard to the objective which the competent institution is seeking to pursue'.2612 There‐ fore, the fault tolerance of the Court towards the legislator's proportionality considerations in these constellations is relatively high.

In conclusion, it is to be noted that there are many ways in which a pro‐ portionate (special) compliance mechanism may be built. The core issues of whether a certain measure is 'necessary' and whether it is proportionate in the narrow sense depends in particular on the policy field, on the nature of the competence on the part of the EU, and on the situations to be tackled.

<sup>2608</sup> For the Commission proposal and the ensuing legislative negotiations see Ander‐ sen, Enforcement 191–195.

<sup>2609</sup> See eg case C-233/94 *Germany v European Parliament/Council*, para 56.

<sup>2610</sup> Case C-310/04 *Spain v Council*, para 121.

<sup>2611</sup> Case C-310/04 *Spain v Council*, paras 122 f.

<sup>2612</sup> Case C-508/13 *Estonia v European Parliament/Council*, para 29; case T-510/17 *Del Valle Ruíz*, paras 107 f, each with further references; see also case C-210/03 *Swedish Match*, para 48 and the references made therein.

Commission Proposal COM(2017) 487 final for introducing the compli‐ ance mechanism for the screening of foreign direct investments into the EU (the legislative result of which is addressed above; see IV.2.2.3.2.6.), for example, displays an elaborate consideration of proportionality, addressing a number of different points (which the author has italicised), thereby suggesting that the necessity of the Regulation in general, but also of the compliance mechanism at issue here, was pondered:

'Moreover, the proposal introduces the possibility for the Commission to screen foreign direct investments which are likely to affect projects or programmes of *Union interest on security and public order grounds*. Projects or programmes of union [sic] interest include in particular those involving a substantial EU funding, or established by Union legislation regarding critical infrastructure, critical technology or critical inputs. In order to ensure transparency and legal certainty, an *indicative list* of projects or programmes of Union interest is included in the annex to the Regulation. The scope of the screening remains *limited to likely threats to security and public order*. The Commission will be able to provide an *opinion to the Member States* in which the investment is planned or completed, *while entrusting the final decision on the appropriate response to those Member States*' (emphasis added).<sup>2613</sup>

3.5. The effects of soft law in compliance mechanisms: varying degrees of authority?

# 3.5.1. Introduction

The legal effects of soft law have been addressed in Part III above (4.2.). The findings elaborated in this chapter apply also in the given context. While the Court's rather casuistic case law draws limits to any attempt to make generalisations, there seems to be a principal duty of MS to consider EU soft law which is addressed to them (see III.4.2.2.1.2. above).2614 We may repeat *Senden's* words that in this context there is an 'obligation of effort, as opposed to an obligation of result'.<sup>2615</sup> This excludes a *general* duty to

<sup>2613</sup> Commission Proposal COM(2017) 487 final, 9.

<sup>2614</sup> See, as an example from the more recent case law, case C-28/15 *Koninklijke*, para 41.

<sup>2615</sup> Senden, Soft Law 350.

give the reasons for non-compliance.<sup>2616</sup> In some cases, however, a duty to give reasons is explicitly laid down in law.<sup>2617</sup> This is an obligation (of result) the MS has *vis-à-vis* the EU (body concerned), but potentially also *vis-à-vis* individuals/undertakings. Where the follow-up action of the MS is addressed to an individual/undertaking, this individual/undertaking has a right to a sufficiently reasoned act – under the respective national legal standards. In addition to that, where the national authority is implementing Union law<sup>2618</sup> the individual/undertaking has a right based on EU law to be given the reasons for the decision,<sup>2619</sup> and hence (if applicable) also the reasons for non-compliance with the EU soft law act at issue (addressed to the decision-maker).<sup>2620</sup> In this context, the then Court of First Instance held, with a view to the requirements of an adequate reasoning under EU law, that the decision-making body, to the extent it disregards the opinion at issue (here: a scientific assessment provided by an expert committee), 'must provide specific reasons for its findings by comparison with those made in the opinion and its statement of reasons must explain why it is disregarding the latter'. The expressed reasons (for disregarding the opinion) 'must be of a scientific level at least commensurate with that of the opinion in question'.2621 While the legally non-binding acts dealt with here are rarely scientific in nature, it can be abstracted from this case law that the more

<sup>2616</sup> Differently: H Adam, Mitteilungen 83.

<sup>2617</sup> Article 17 para 1 of Regulation 1092/2010, for example, stipulates that a national authority shall provide 'adequate justification' for inaction with regard to ESRB recommendations.

<sup>2618</sup> Article 51 para 1 CFR.

<sup>2619</sup> For the enhanced reasoning requirements in case of decisions (here: of the Com‐ mission) entailing a power of appraisal see case C-269/90 *Technische Universität München*, paras 14 and 27.

<sup>2620</sup> Article 41 para 2 (third indent) CFR. Only where the meaning for the eventual output of the soft law act is entirely subordinate, it (ie the arguments made therein) may legitimately be skipped in the reasoning. This is in accordance with the Court's case law, pursuant to which '[i]t is not necessary for the reasoning to go into all the relevant facts and points of law'; case C-89/08P *Commission v Ireland*, para 77, with a further reference; for further specifications – the reasoning also depends on the content of the measure at issue, the nature of the reasons and the interest of the addressees (or other concerned parties) in obtaining explanations – see case T-63/16 *E-Control*, para 68, with further references. This case law on what is now Article 296 TFEU may be applied also with regard to Article 41 CFR; see also the Explanations relating to the Charter of Fundamental Rights with regard to the latter provision.

<sup>2621</sup> Case T-13/99 *Pfizer*, para 199.

specific an act is, the more specific the reasons for non-compliance must be (where a duty to give reasons in the case at issue has been affirmed).

Against the background of these general remarks, in the following we will consider the varying obligations to react to soft law, as laid down in the compliance mechanisms addressed above, and at their respective effects.

### 3.5.2. Degrees of authority or mere legislative wordiness?

Against the background set out above, we shall – in the context of our selection of compliance mechanisms – address the question whether the soft law acts provided for therein display different degrees of authority. While it is remarkable that nearly all soft or mixed secondary law-based compliance mechanisms at issue provide for the adoption of opinions and/or recommendations, that is to say those two categories of legally nonbinding acts explicitly provided for in Article 288 TFEU, an examination of the individual degrees of authority is suggested by the varied wording of explicit duties to consider soft law in the legal acts setting out these mechanisms. These duties may oblige EU actors – eg the Commission has to 'take into account' the EBA recommendation under the regime of Article 17 of Regulation 1093/2010 or the ESAs shall take 'utmost account' of ESRB warnings/recommendations pursuant to Article 36 para 6 of Regulations 1093–1095/2010 – but, more importantly in the given context, may also be imposed on MS (and their authorities). The latter shall eg 'appropriately address[]' the SRB warning adopted under Article 7 para 4 of Regulation 806/2014, give 'due consideration' to a Commission opinion (Article 6 of Regulation 2019/452) or 'pay[] sufficient heed' to the ERA opinion launched on the basis of Article 25 of Regulation 2016/796; they are obliged to take 'utmost account' of EU soft law (Article 53 of Directive 2019/944, Article 10 para 2 of Directive 2018/1972), to provide a 'reasoned opinion' for non-compliance (Article 33 of Directive 2018/1972) or to 'adequately justify any inaction' (Articles 16 f of Regulation 1092/2010). Do these seeming qualifications actually increase the authority of the soft law concerned, do they elevate the duty to consider of the addressee to a duty to provide the reasons for non-compliance?

As a first – apparent – finding we can state that all these *epitheta*, as it were, give proof of the non-bindingness of the acts at issue.<sup>2622</sup> On a second

<sup>2622</sup> See also case T-295/06 *Base*, paras 63 f; case C‑689/19P *VodafoneZiggo*, para 55.

level, we need to reiterate the distinction between obligations of effort and obligations of result. The latter are established with the requirement to provide a 'reasoned opinion' for non-compliance or to 'adequately justify any inaction'. These are clear commands that go beyond a duty to consider and hence non-compliance with these soft law acts – *qua* secondary law – entails an obligation to provide the reasons for that. It is important to note that the obligatory effect of these soft law acts does not emanate from the soft law acts themselves, but from the relevant secondary law.<sup>2623</sup> Let us take an example: The national competent authorities' duty to inform the ESAs in case of non-compliance with their guidelines and recommendations according to Article 3 para 2 subpara 2 of Regulations 1093–1095/2010 is laid down in secondary law, not in the guidelines or recommendations themselves.<sup>2624</sup> As regards the duty to report according to subpara 4 *leg cit*, which applies '[i]f required by that guideline or recommendation', it ought to be highlighted that this does not lead to the partial bindingness of these guidelines/recommendations.2625 Such an effect, if it were confirmed, from a legal perspective would be highly questionable, like the adoption by European agencies of general-abstract rules binding upon MS more generally.2626 In order to ensure conformity with primary law, this passage must be interpreted as a general duty to report cases of non-compliance (which is based on secondary law) regarding which the ESAs may state – with regard to certain guidelines/recommendations – that they do (not) insist on compliance with this duty to report. The declared non-insistence

2625 See also Russ/Bollenberger, Leitlinien 811.

<sup>2623</sup> Stressing, in the context of public international law, this separation of the soft law act on the one hand and the act stipulating certain duties with regard to this soft law act, on the other hand: Griller, Übertragung 156, with many further references. Also the exceptional case of soft law being transformed into hard law is effected by secondary law, in our example by Articles 22 f of Council Regulation 2015/1589 (see 3.2.4.2. above); mitigating the meaning of this distinction: Tridimas, Indeterminacy 61.

<sup>2624</sup> Further examples of the authority of certain soft law acts being increased by secon‐ dary law are Article 17 para 2 lit b of Regulation 2018/1139, according to which the Commission may not change the content of the EASA's draft implementing meas‐ ures 'without prior coordination with the Agency', or Article 10 para 1 subparas 6–8 of Regulations 1093–1095/2010 which also require coordination (defined in more detail in the *lex citata*) of the Commission with the ESAs when the former intends to change the content of a draft regulatory technical standard as prepared by one of the ESAs.

<sup>2626</sup> For the exceptional adoption of general-abstract legally binding rules by European agencies see Bergström, System 219 and *passim*.

(by not mentioning the duty to report in the guideline or recommendation) binds the ESAs, but entitles the national authorities.<sup>2627</sup>

The obligations of effort are expressed by a duty to 'take into account', or even to take 'utmost account' of the soft law act at issue, or to 'pay[] suffi‐ cient heed' to it.<sup>2628</sup> Assuming that there is a general duty to consider soft law (see also 3.5.1. above), it is unclear to which extent these special require‐ ments add anything thereto.<sup>2629</sup> They could as well be considered merely declaratory in nature. But even if the duties of MS were thereby increased, practically compliance with them could be scrutinised only where there was a related duty to give reasons or even to justify non-compliance.<sup>2630</sup> In case of the compliance mechanisms presented here, such duty may emanate from the duty to provide the reasons for a decision addressed to an individual/undertaking – however, if not explicitly regulated otherwise, this duty exists only *vis-à-vis* the (individual) addressee (see above), not *vis-à-vis* the creator of the respective soft law act.<sup>2631</sup>

In case C-424/07 *Commission v Germany*, the Court of Justice quotes the duty of national authorities to take 'utmost account' of certain Commission output, at the same time stressing their 'broad discretion in order to be able to determine the need to regulate a market according to each situation on a case-by-case basis'.2632 In another case, the Court deduced from the duty to 'take utmost account' of Commission recommendations in combi‐

2630 See Láncos, Core 763. Such a duty seems to be assumed by Tobisch, Telekommuni‐ kationsregulierung 105.

<sup>2627</sup> For the self-obliging effect of soft law see III.4.2.2.2.4. above.

<sup>2628</sup> It is unclear in general which of these terms ought to be most compelling. Only sometimes the legislator in one and the same act uses two different terms, thereby explicating different duties: see, for example, Regulation 2019/452, stipulating that MS have to take (only) 'due consideration' of a Commission opinion in one procedure (Article 7 para 7), but 'utmost account' of it (including the provision of an explanation in case of non-compliance) in the other procedure (Article 8 para 2 lit c).

<sup>2629</sup> Arguing that in these cases non-compliance is only justified where otherwise a breach of EU law or of national law (interpreted in accordance with EU law) would occur: von Danwitz, Verwaltungsrecht 251, with further references; see also Koe‐ nig/Neumann/Senger, Ausgestaltung 367, who tautologically say that 'besonderes Gewicht' [special emphasis] should be granted to acts of which, according to the law, 'utmost account' is to be taken.

<sup>2631</sup> In case C-69/13 *Mediaset*, para 31, the Court – invoking Article 4 para 3 TEU – seems to address the EU-MS relationship (not: the relationship of the national body *vis-à-vis* the individual), and mentions a duty to state reasons. The Court does not elaborate on this duty, though (eg in the following paragraph).

<sup>2632</sup> Case C-424/07 *Commission v Germany*, para 61.

nation with a duty to provide the reasons for non-compliance that the national authorities shall 'follow, as a rule, the guidance contained in [the] Recommendation', and shall only depart where the recommendation 'is not appropriate to the circumstances'.<sup>2633</sup> This seems to suggest that the 'utmost account' requirement, materially speaking, ought not to be overestimated. The effects of other *epitheta* – such as: 'appropriately address[]' – are not explicit on whether or not they include a duty to justify inaction, either. While it does not seem to imply a duty to justify non-compliance *vis-à-vis* the creator of the soft law act at issue, it may be interpreted as emphasising a duty to provide – in case of non-compliance – an appropriate reasoning in the final act (addressed to an individual/undertaking). From a compe‐ tence perspective this would be acceptable, as the EU legislator may certain‐ ly increase the threshold for the reasoning required in the output *vis-à-vis* the individual/undertaking – after all, in such a case Union law would be implemented within the meaning of Article 51 para 1 CFR, which renders applicable the CFR. The effect would be limited and (more importantly) substantially questionable, though: According to Article 41 CFR, there is a general duty to state the essential reasons for administrative output. According to this understanding, the above epithet would – in addition to that – require consideration in the form of a statement of reasons also where the soft law act at issue did not contain essential arguments for the final output. In light of this odd result, it can be doubted that this is what the legislator intended to prescribe. Rather, the additions in question may be understood as a merely declarative hint to the addressee to have a close look at essential arguments which may (possibly) be contained in the soft law act at issue.

## 3.5.3. Considerations in the legislative process and conclusions

While the above considerations were focussed on the literal interpretation of these additions, their respective effect needs to be examined also by tak‐ ing into account the context and other specificities of the concrete case.<sup>2634</sup> For example: The Commission proposal for what later became Directive 2009/72/EC – the predecessor of Directive 2019/944 – concerning common

<sup>2633</sup> Case C-277/16 *Polkomtel*, para 37.

<sup>2634</sup> See Brohm, Mitteilungen 163 f; Thomas Müller, Soft Law 116.

rules for the internal market in electricity<sup>2635</sup> in its Article 8b paras 8 and 10 – in the context of certifications relating to third countries – provided for a Commission decision by means of which a national regulatory authority could be obliged to amend or withdraw its decision on certification. In the course of the political negotiations on the legislative proposal, the decisionmaking power of the Commission was replaced by a mere power to adopt an opinion. In order to 'ensure the consistent application of those rules across the Community', the Directive states that the regulatory authorities 'should take utmost account of the Commission's opinion when the former take decisions on certification'.<sup>2636</sup> It appears that here setting the require‐ ment for the national authorities to take utmost account corresponded to an attempt to mitigate the effects of the replacement of the originally envisaged decision by an opinion. It is not a legislative command, but – bearing in mind the objective to ensure the consistent application of the relevant Union law – a strong legislative prompt to comply with the opinion, and – we could add, with a view to the genesis of the Directive – to deviate from it only when important reasons so suggest. This goes beyond the effects of 'regular' EU soft law.2637 Since there is no duty to provide the reasons for non-compliance in this case, it is difficult for EU actors (in particular: the Commission or the Court) to examine whether or not utmost account was taken of the Commission opinion. It is first and foremost the follow-up output of the addressee, in our example the decision addressed to the individual applicant for a certification, in partic‐ ular the reasoning contained therein, which will indicate the arguments for non-compliance the national authority has employed (see also 3.5.2. above). Also the national authorities themselves will have an interest in making clear their respective reasons for non-compliance. Otherwise, they may be held responsible for non-compliance without good reasons.

<sup>2635</sup> Commission Proposal COM(2007) 528 final.

<sup>2636</sup> Recital 24 of Directive 2009/72/EC.

<sup>2637</sup> See Tobisch, Telekommunikationsregulierung 104, with many further references.

All in all, the 'qualification' of soft law may lead – in a concrete case<sup>2638</sup> – to a hierarchy of different soft law acts. <sup>2639</sup> A hierarchy of acts – as long as it is provided for by the law-maker, which regularly is the case – may exist also among the different layers of legally binding acts, eg primary or consti‐ tutional law and, relatively lower-ranked, secondary or sub-constitutional law.<sup>2640</sup> The norms of the different layers are equally binding, but they differ in rank, and the respective lower layers of norms need to be in compliance with the respective higher layers. True soft law remains to be legally nonbinding (or, in the understanding of Article 263 para 1 TFEU: not intended to produce legal effects *vis-à-vis* third parties), but non-compliance with soft law acts with a qualification (set by law) may trigger legal duties, such as the duty to give reasons or the duty not to deviate for reasons of minor importance.<sup>2641</sup> The effect in practice is comparable to that of a binding rule allowing for certain exceptions (possibilities of lawful deviation, eg in case of 'important reasons'; see II.2.1.3.1. above). What is more, in a systematic view this approach may bear a slight risk of trivialising the meaning of soft law without such attributes. However, and as we have seen above (III.4.2.2. and III.4.2.3.), also this *regular* EU soft law is to be 'considered', ie to be dealt with – at least for the time being. In the long run, and with the use of qualifications increasing, the comparison with these qualified acts may downplay the authority of regular soft law, though. Once again, it ought to be emphasised that this hierarchy is created not by the respective soft law itself, but by the legal framework established in particular by secondary law.

The effectiveness of this qualified soft law does not only depend on the legal duties just addressed. There is also a factual side to the effectiveness of (both qualified and 'unqualified') soft law, viz the factual authority of its

2641 See also Arndt, Sinn 166.

<sup>2638</sup> There does not seem to be a general hierarchy of soft law acts in EU law, though. If at all, it is due to a strong connex with law; see Knauff, Regelungsverbund 501f; AG *Kokott* in her Opinion in case C‑398/13P *Inuit*, para 92, appears to suggest such a hierarchy when claiming that a certain (legally non-binding) act 'has the status of *at least* a recommendation' (emphasis added).

<sup>2639</sup> See case 815/79 *Cremonini*, para 8, in which reference is made to a hierarchy of national/international standards set up by EEC law; see, as another example, Article 17 of Regulation 1093/2010, according to which the EBA's recommendation only needs to be 'taken into account' when the Commission takes a follow-up action on the case at issue (in the form of a formal opinion; para 4 *leg cit*), whereas the EBA, when taking a decision following the formal opinion, has to make sure that it is 'in conformity with' the latter (para 6 *leg cit*); see 2.2.2.2.3. above.

<sup>2640</sup> For the only implicit hierarchy between *legally binding* acts in the then Communi‐ ty legal order see Senden, Soft Law 54 f.

creator and of the 'procedural surroundings'. If a soft law act contains an irresistibly convincing line of argumentation, or where it may be followed by a legally binding decision, or where a soft procedure is paralleled by a mixed or hard procedure on the same or a related matter which may end if the addressee complies with the soft law act, these factors, on their respective own or in combination, may as well increase the likelihood of compliance with it (see 2.4.1. above).

# 3.6. Legal protection for the Member States

# 3.6.1. Introduction

In complementation of what was said above on the possibilities of applying for judicial review of soft law (see III.6. above), we shall now turn to the legal, in particular judicial protection available for MS affected by acts adopted by EU bodies in the course of a compliance mechanism.

Where these (binding) decisions are adopted by European agencies, they may be appealed against before the respective agency's appeal body (nor‐ mally named 'Board of Appeal'),2642 given its competence to review the act at issue; non-binding agency output such as opinions normally may not. 2643 While not all agencies dispose of an appeal body, the founding regulations of the more recent decision-making agencies regularly provide for one.<sup>2644</sup> They are composed of a couple of experts in the policy field concerned (three or more). A MS (authority) addressed by an agency decision may turn to the Court only after it has turned to the appeal body (if set up and if an appeal is admissible).2645 The action for annulment then has to

<sup>2642</sup> Whether the procedures before these appeal bodies are administrative or judicial in nature is contested, but need not be elaborated on further in this context; see Chamon, Agencies 338 f; for the treatment of soft law by agencies' Boards of Appeal more generally see Alberti, Position 271.

<sup>2643</sup> See eg case T‑63/16 *E-Control*, para 49.

<sup>2644</sup> See eg Article 85 of Regulation 806/2014; Articles 58 f of Regulations 1093– 1095/2010; Articles 25 f of Regulation 2019/942; Article 165 of Regulation 2017/1001, referring – in the context of the EUIPO – to several Boards of Appeal, among them a Grand Board.

<sup>2645</sup> Heed, in this context, the reform of Article 58a of the Court's Statute, as reques‐ ted by the Court late in 2022; <https://curia.europa.eu/jcms/upload/docs/appli‐ cation/pdf/2022-12/demande\_transfert\_ddp\_tribunal\_en.pdf> accessed 28 March 2023.

be brought against the appeal decision.2646 Only where an agency does not have an appeal body or where its decision is not appealable may the MS (authority) file an action for annulment against the (original) agency decision.<sup>2647</sup>

3.6.2. The action for annulment and the action for failure to act

Beyond the fundamental statement that legally binding acts may be subject to an annulment procedure, whereas legally non-binding acts may (princi‐ pally) not, we shall take a look at one of the procedural specificities of most compliance mechanisms addressed above, that is their tiered structure. The Court has determined that '[i]t is […] settled case-law that, in the case of acts adopted by a procedure involving several stages, and particularly where they are the culmination of an internal procedure, it is in principle only those measures which definitively determine the position of the Commis‐ sion or the Council upon the conclusion of that procedure which are open to challenge and not intermediate measures whose purpose is to prepare for the final decision'.2648 Acts of an entirely preparatory kind cannot as such be complained against, but they may be considered in the course of the (annulment) procedure against the final act (the preparation of which they served).2649 Thereby an indirect, incidental judicial control of preparatory, potentially soft law acts may be achieved – not only where the final decision explicitly refers to the preparatory output and/or where the creator of the final decision is bound to consider or even to conform to the preparatory output (addressed also to the MS concerned),2650 but also where the linkage is only implicit.

<sup>2646</sup> See case T-102/13 *Heli-Flight*, paras 27 f.

<sup>2647</sup> See eg Article 61 of Regulations 1093–1095/2010 (Board of Appeal) or Article 86 para 1 of Regulation 806/2014 (Appeal Panel).

<sup>2648</sup> Case C-147/96 *Netherlands v Commission*, para 26, with a further reference; with regard to procedures involving both EU and national actors see case C-219/17 *Berlusconi*, paras 43 ff.

<sup>2649</sup> See case 60/81 *IBM*, para 10; case T-55/01R *Asahi*, para 62; case T-317/09 *Concord*, para 44, each with further references; see also more recently: case T-671/15 *E-Con‐ trol*, paras 26–28, and the discussion by Ştefan/Petri, Review 543 f; joined cases C‑551/19P and C‑552/19P *ABLV*, paras 40 ff.

<sup>2650</sup> Take Article 17 para 6 of Regulation 1093/2010 as an example, according to which the EBA decision needs to be 'in conformity' with the preceding formal opinion of the Commission, as addressed to the national authority concerned.

Thus, where the MS addressed in the course of a mixed or hard mecha‐ nism does not share the legal view of the EU body/bodies in charge – and where it does not manage to convince the latter of its own view – it is well advised to have the respective procedure performed in its entirety. Thereby it may achieve that a (final) legally binding act is adopted, against which it may then turn to the Court by filing an action for annulment pursuant to Article 263 TFEU. In the exceptional cases of mechanisms providing for binding acts preceding the respective final (binding) act, MS can also file an action against these acts. Legally binding acts with a specific addressee cannot be supposed to have a merely preparatory character.

Where the mechanism is soft, the MS can lawfully deviate from the EU body's output anyway, without having to challenge it before the Court. However, in case of doubt, it is advisable to address the Court, if only to have it dismiss the action for lack of (intended) legal effects on third parties of the challenged act. Otherwise, there remains a risk that the act turns out to be binding – eg in the course of a Treaty infringement procedure which the Commission or another MS may launch – and the addressee ends up having violated EU law, although it merely intended to ignore a soft law act. Having said this, a conviction under a Treaty infringement procedure would be a rare exception, as the Court has to take account of the addressee's trust in the appearance of the act – a trust which is, if justified (ie if the act is akin to a soft law act), legally protected.<sup>2651</sup>

When it comes to the active legitimation under Article 263 TFEU, the question arises whether national authorities fall within the privileged cate‐ gory of 'Member State[s]' (para 2) or whether they qualify as non-privileged actors pursuant to para 4. A qualification of any national authority as 'Member State' within the meaning of para 2 would result in the power of national authorities to challenge any legally binding output of EU insti‐ tutions, bodies, offices and agencies – a power which would be entirely non-system.2652 Thus, it is not surprising that the Court has confirmed that territorial authorities such as municipalities or *Länder* are non-privileged

<sup>2651</sup> Also in other procedures, eg with regard to State liability, this trust has to be taken into account (by the competent national court or, upon a preliminary request, by the Court of Justice), most suitably in the context of the 'sufficiently serious breach'; see the settled case law beginning with joined cases C-46/93 and C-48/93 *Brasserie du Pêcheur*, para 51.

<sup>2652</sup> Note Article 267 TFEU which limits a related power – namely to ask the Court about the validity of secondary law – to national courts and tribunals (where this question is relevant in a concrete case before them).

claimants.<sup>2653</sup> Other national authorities disposing of legal personality or at least the power to act as a party in procedures according to national law may as well file an action under Article 263 para 4 TFEU.<sup>2654</sup> Public bodies which do not meet these criteria regularly act for a legal person, in federally organised States in particular for a territorial authority. Then it is, legally speaking, the latter authorities which are addressed or otherwise di‐ rectly and (as regularly required) individually concerned by the EU output rendered in the course of compliance mechanisms, and which may hence file an action for annulment. As a 'Member State' within the meaning of Article 263 para 2 TFEU only the central (federal) government of a MS may file an annulment action with the Court.<sup>2655</sup>

Mixed and hard compliance mechanisms provide for national authorities to be addressed by binding EU output. Thus, in these cases it will not be difficult to prove that the requirements for an action under Article 263 para 4 TFEU ('addressed') are met. Where a national authority intends to challenge, under Article 263 TFEU, EU output addressed to another national authority (or to any other addressee for that matter), it has to prove its direct and – unless in case of regulatory acts pursuant to para 4 – individual concern.<sup>2656</sup>

Alternative (and less practicable) routes to a (possibly only incidental) review of acts adopted in the course of a compliance mechanism are a subsequent Treaty infringement procedure, a subsequent preliminary refer‐ ence procedure (following, for example, an action for State liability) or a subsequent procedure pursuant to Article 340 para 2 TFEU (see III.6.3. and III.6.4. above). The final decision whether or not such procedures are

<sup>2653</sup> See, each with further references, case 222/83 *Differdange*, para 8; case C-15/06P *Regione Siciliana*, in particular para 29; case C-444/08P *Açores*, para 31; case C‑872/19P *Venezuela v Council*, paras 44–46; joined cases T-132/96 and T-143/96 *Sachsen/Volkswagen*, para 81 (see also references in para 72). This appears to be consistent also against the background that State liability applies to the MS in their entirety. Thus, the alternative step – not to be held responsible for a breach of EU law, but to challenge the respective act of EU law before the Court – should be up to the MS as a whole (not to its single bodies). For the conceptual linkage between a MS' liability (here: under the State aid regime) and Article 263 para 1 TFEU see the Commission's argumentation in joined cases T-132/96 and T-143/96 *Sachsen/Volkswagen*, para 68.

<sup>2654</sup> See eg joined cases C‑177/19P to C‑179/19P *Germany v Commission*, paras 69 f, with further references; see also Stotz, Aktivlegitimation, paras 69 f.

<sup>2655</sup> See Dörr, Art. 263 AEUV, para 11.

<sup>2656</sup> See joined cases T-269/99, T-271/99 and T-272/99 *Guipúzcoa*, para 41.

initiated may be influenced by the MS authority concerned, but eventually it is taken by different actors.

The possibility to launch an action for failure to act under Article 265 TFEU – which has turned out to be practically insignificant<sup>2657</sup> – is to be mentioned here for the sake of completeness. While MS may not only invoke the failure to adopt a binding act, but also the failure to adopt a non-binding act 'in infringement of the Treaties', non-privileged claimants (including MS authorities) are limited to complain about the failure to adopt binding acts.<sup>2658</sup> In the compliance mechanisms presented above the decision whether or not to adopt a soft law act regularly falls within the discretion of the respective EU body (*argumentum* 'may'). Therefore – in addition to the limited admissibility of related actions under Article 265 TFEU – the non-adoption of such acts regularly will not constitute an 'in‐ fringement of the Treaties' (ie of EU law). Also with regard to the adoption of binding acts under the compliance mechanisms presented above, the EU bodies regularly dispose of a certain room for manoeuvre. But even if they do not, MS (or its authorities) regularly have no interest in receiving a decision urging them to comply with EU law, which makes actions for failure to act highly improbable also in this context.

3.6.3. The MS' motivation to seek judicial protection

Whether or not the MS seek judicial protection against acts of the EU (or – exceptionally – an EU body's failure to act) normally depends on a variety of factors, among which is what could be referred to as the 'legal context'. A special case to be mentioned here is the Court's power to scrutinise fines imposed on a MS by an EU body. In the example addressed above (see 3.1.1.2.3.), namely Article 8 of Regulation 1173/2011, the Council was granted the power to impose a fine on a MS who has intentionally or by serious negligence misrepresented its deficit and debt data to be transferred to the Commission. Being in charge of reviewing this fine, if requested, the Court may confirm, reduce or – and this is remarkable – increase the respective amount. This power has a deterrent effect in that it may prevent the MS concerned from filing an action even if it deems the Council act (including

<sup>2657</sup> See eg Dörr, Art. 265 AEUV, para 3.

<sup>2658</sup> *Argumentum* 'other than a recommendation or an opinion'; see also W Cremer, Art. 265 AEUV, para 5, with further references.

the fine) to be unlawful, especially where the fine is comparatively low.<sup>2659</sup> The risk of the Court refusing the action and eventually even increasing the fine may be considered too high – if justifiably so or not is irrelevant. The possibility of a *reformatio in peius* increases the power of the Court, it could be said, and *de iure* this is correct.<sup>2660</sup> In view of the chilling effect just mentioned, however, *de facto* it is the power of the Council which is enhanced (because the likelihood of its decision being judicially reviewed is reduced). In light of that, the given example – while *prima facie* adequately reflecting upon the Court's systematic importance in the EU – can be interpreted as another example of the Court's being ousted from its dominant role in the enforcement of EU law as epitomised by the Treaty infringement procedure.

On the other hand, what was referred to above as 'legal context' may also increase the likelihood of a MS (authority) seeking judicial protection. For example: Where a MS violates certain rules of the EU's excessive deficit regime (as laid down in primary and secondary law), in addition to the sanctions provided for therein EU law (in the context of 'macroeconomic conditionality') allows for other negative effects, such as the partial or complete suspension of financial support coming from the EU's Structur‐ al and Investment Funds.2661 Similar conditionality dynamics have been known in EU law eg in connection to Treaty infringement procedures, which may be applied once the Commission has addressed a reasoned opinion to a specific MS.2662 Regularly, these suspensions are based on the contract concluded between the EU and the respective MS ('Partnership Agreements')2663 which renders the possibilities of judicial review directed

<sup>2659</sup> For a practical example, note the case of Austria which was fined because of a misrepresentation of deficit and debt data concerning the *Land* Salzburg and which eventually refused to challenge the respective Council act, even though it deemed the act to be unlawful; see <https://www.consilium.europa.eu/en/press/pr ess-releases/2018/05/28/land-salzburg-austria-fined-for-misreporting-government -debt-data/> accessed 28 March 2023.

<sup>2660</sup> For the principal possibility of vesting the Court with such power see Article 261 TFEU; with specific regard to the possibility of a *reformatio in peius* see Booß, Art. 261 AEUV, para 15; W Cremer, Art. 261 AEUV, para 6; Ehricke, Art. 261 AEUV, para 8, each with further references.

<sup>2661</sup> See Article 23 para 9 of Regulation 1303/2013; for the broader framework see Klamert, Durchsetzung 164 f.

<sup>2662</sup> See European University Institute, Research 22 and 47.

<sup>2663</sup> The limited number of suspensions in practice cannot do away with the legal *problematique* these arrangements entail; see European Court of Auditors, 'Ex

specifically to these suspensions very limited. If applied, these arrangements exert a strong pressure on the MS concerned.

These circumstances may lead to the MS pushing for judicial review in the related compliance mechanism, hoping for the Court to confirm the respective MS's own legal view. However, it may as well cause the MS to give in and to comply without further ado in order to do away with the financial disadvantages it is confronted with. These concerns are not directly related to the compliance mechanisms at issue here, but in the larger 'legal context' they certainly ought to be taken into account.

ante conditionalities and performance reserve in Cohesion: innovative but not yet effective instruments' (Special Report, 2017) 47; see also European University Institute, Research 34 f.

# VI. CONCLUSION

#### *1. Summary*

Having set out in Part I the fundamental research questions to be addressed in this work, the methodology applied in it and the research gap it is intended to fill, in Part II terminological and conceptual questions relating to the idea of soft law were approached. After a discussion of the historical origins of soft law, which are prominent in particular in the field of public international law, a number of different approaches towards this phenom‐ enon, as proffered in the literature, have been presented. The subsequent comparison of these different schools of thought not only displayed the manifold ways in which the term soft law may be used, but also allowed the author to contrast some of them with his own position.

This position is based on a positivist perception of law and, consequently, also of soft law, which does not leave room for degrees of legal bindingness, as suggested by some scholars, but discerns legal non-bindingness (as op‐ posed to legal bindingness) as the core criterion in defining soft law (as opposed to law). Soft law was defined – for the purposes of this work – as norms, enacted by entities thereby exercising public authority and thereby aiming at steering human behaviour, which are legally non-binding according to the interpretatively established will of its creators (or, as an expression of self-obligation, legally binding only upon the creators them‐ selves).

Once defined, in a next step the characteristics of soft law were analysed in more depth, delimitating it from law, from further sets of norms – namely custom and, as a related form of law, customary law, morals, and regulation by private actors – and from other output of public bodies. Thereby also the cases of doubt have been inspected more closely, in which soft law intersects other categories of norms or other output of public bodies. As two examples of a conceptual overlap between law and soft law – arguably the most common case of doubt – first the Court's *Kadi* saga on the relationship between UN law and the EU's human rights regime and, second, its case law on the position of WTO law in the EU legal order were discussed more thoroughly. They illustrated that rules which are legally binding in one legal order – here: public international law – may be relativised, in one or the other way, by the highest legal authority in another legal order, to the effect that their position – in this other legal order – reaches or at least comes close to that of soft law.

On the foundation of these conceptual observations which have been, practical examples apart, largely abstracted from specific legal orders, in Part III the focus was shifted to the EU legal order and thus was dedicated to fundamental questions of EU soft law. Some of the most important ques‐ tions EU soft law raises relate to its potential originators and addressees, its legal bases (in other words: the competence regime applicable to soft law), its effects, in particular its legal effects, its purposes, and the available possibilities of judicial review. These sets of questions were addressed in Chapters 2–6 of Part III, whereas Chapter 1 provided an introduction and an overview of the historical and current use of soft law in the EU legal order.

The account of originators of EU soft law proffered in Chapter 2 dis‐ played the variety of potential creators, ranging from 'expected candidates', like the institutions and the bodies, offices and agencies of the EU, to in this context more exotic actors like the MS and non-EU bodies who may – exceptionally, but still – be empowered to adopt EU soft law. It was shown that *prima facie* EU soft law may have strings attached to different legal or‐ ders – most prominently: to EU law *and* to public international law – which renders difficult the allocation of the respective acts and the determination of the legal effects applying to them. Special attention was drawn to the case of Memoranda of Understanding as concluded with beneficiary MS in the context of the so-called umbrellas during the Eurozone crisis, both in terms of their relation to EU law and public international law and in terms of their legal (non-)bindingness. Eventually, the (potential) addressees of EU soft law were referred to and, in a final sub-chapter, a conceptual line was drawn – in the context of the EU legal order – between soft law and legally non-binding acts other than soft law, thereby adding on to the more general remarks on 'other output of public bodies' in the final sub-chapter of Part II.

In Chapter 3 the EU's competence regime applicable to the creation of soft law was discussed. Beginning with an analysis of the meaning of Article 288 TFEU (mentioning two kinds of non-binding legal acts: 'recommenda‐ tions' and 'opinions'), the applicability of the core rule in the context of the EU's competences, the principle of conferred powers, was examined. A closer analysis of this principle and the relevant case law of the CJEU did not provide for a clear answer to this question. The author deems there to be sound reasons to argue in favour of the applicability of the principle of conferred powers also in the context of soft law, though. On the assumption that the principle of conferral applies in this context, special features of the EU's general competence regime were discussed. The implied powers doctrine and – as a subset of it – the *argumentum a maiore ad minus* contribute to making it much more flexible and concessive. With regard to internal soft law, the relevant competence mostly follows from the EU bodies' right to self-organisation, in respect of which they dispose of a large measure of discretion. Subsequently, the explicit competence clauses allow‐ ing for the adoption of recommendations and/or opinions as laid down in the Treaties were presented, thereby distinguishing between general and special competence clauses. In addition to that, selected primary law com‐ petences to adopt EU soft law other than recommendations and opinions, and selected special competence clauses enshrined in EU secondary law and in public international law were discussed. Eventually, after the effects for EU soft law of a lack of legal basis had been shed light on, selected questions approached at the beginning of Chapter 3 – in particular the question of whether or not the principle of conferral is applicable – were revisited with a view to the findings which the preceding analysis of the general and special competence clauses allowed us to make. While these findings could not dispel all doubts in this context, they seemed to confirm the applicability of the principle of conferral also with regard to soft law rather than to dismiss it.

Chapter 4 was concerned with the effects of EU soft law, in particular with its legal effects, that is to say the effects resulting from law. These legal effects were addressed with regard to two groups of addressees, namely the MS on the one hand and the institutions, bodies, offices and agencies of the EU, on the other hand. In both cases the effects following from the pertinent case law of the CJEU were presented, after which the potential legal bases for these effects – which the Court in its judgements rarely makes explicit – were listed and analysed in some depth. As a complement, the factual – that is to say: the non-legal – effects of soft law were addressed, both generally and with regard to the EU context. Eventually, the possibility of effects displaying legal as well as factual aspects – referred to here as 'mixed effects' – were expounded and illustrated with examples from the EU context.

In Chapter 5, two approaches for the categorisation of the purposes of soft law more generally were outlined, which were then applied in the EU context, thereby also referring to illegitimate purposes EU soft law may serve according to the will of its creators. It was shown that these categorisations fit and are worthwhile with a view to imposing a conceptual order on the manifold purposes of EU soft law.

In the final chapter of Part III, Chapter 6, the possibilities of judicial review of EU soft law were fathomed, in particular with a view to the annul‐ ment procedure and the preliminary reference procedure. While under the former procedure true EU soft law – for lack of (intended) legal effects *visà-vis* third parties – cannot be annulled, the latter procedure leaves much room for national courts and tribunals to have the CJEU consider EU soft law. From among the other Court proceedings, the procedure following an action for damages according to Article 340 para 2 TFEU may reasonably lead to a consideration of EU soft law by the Court. Also the applicability of the *incidenter* review, as laid down in Article 277 TFEU, to soft law acts of general application was taken into consideration.

Following the account of the main legal questions on EU soft law provi‐ ded for in Part III, the focus was narrowed to compliance mechanisms in which EU soft law *may* – and in many cases actually *does* – play a pivotal role in ensuring that MS abide by EU law. Part IV was dedicated to a presentation of the compliance mechanisms – as defined, for the purposes of this work, more closely in Chapter 1 – laid down in primary law and a selection of those provided for by the legislator in secondary law. Con‐ sequently, Chapter 2 presented, first, the Treaty infringement procedure as the general compliance procedure established in primary law. Second, compliance mechanisms other than the Treaty infringement procedure – that is to say: special compliance mechanisms, whose material scope is regularly limited to one policy field – were presented. They were divided in three different categories: 'hard mechanisms', 'mixed mechanisms', and 'soft mechanisms'. Within the category of hard mechanisms fall compliance mechanisms which only provide for hard law measures being addressed by the EU body/bodies in charge to the respective MS. These mechanisms do not (explicitly) provide for the adoption of soft law *vis-à-vis* the MS concerned. Mixed mechanisms allow the EU body/bodies in charge – in the specific sequence envisaged in each procedure – to address both soft and hard law acts to the MS concerned. Soft mechanisms only envisage the adoption of soft law acts. In their respective course, the behaviour of the MS concerned may exclusively be steered in a legally non-binding way. The special compliance mechanisms laid down in primary law have been allocated to either of these categories. In addition to that, for each category six mechanisms provided for in secondary law were presented. The main characteristics of the presented mechanisms were then condensed in a summary.

Part V was preoccupied with the analysis of the compliance mechanisms presented in Part IV. This analysis was split in two parts, one on the classi‐ fication (Chapter 2) and one on the legal assessment of the mechanisms (Chapter 3).

In terms of classification, the creation of a taxonomy of the mechanisms under a variety of different aspects was attempted. These aspects encom‐ passed the actors involved and the policy fields affected (as indicated in the primary law basis of these mechanisms). Also the output-related struc‐ ture of the mechanisms was addressed more thoroughly, going beyond the broad separation in hard, mixed and soft compliance mechanisms which underlay Part IV. Here, for example, the concrete sequence of (hard, soft and hard, or only soft law) acts was examined. Thereafter, the focus was shifted to soft law and its purposes in the context of compliance mecha‐ nisms. On the basis of the findings of Chapter III.5., in which the purposes of EU soft law more generally had been addressed, the special purposes it is intended to meet in compliance mechanisms were fleshed out. Broadly speaking, it turned out that the general purposes of EU soft law are also reflected upon in mixed and soft compliance mechanisms. In addition to that, soft law in compliance mechanisms was recognised as a tool silently bringing about institutional transformation. Eventually, the deviation of compliance mechanisms from the Treaty infringement procedure as the general compliance mechanism laid down in the Treaties was investigated, allowing the legislator to compensate some of the drawbacks of the latter procedure. Finally, the objective and subjective reasons why compliance mechanisms are designed the way they are designed were explored, thereby taking account of the concrete legal history of the compliance mechanisms addressed here.

The chapter on legal assessment first introduced, in the context of indi‐ vidual-concrete EU measures addressed to a MS in order to ensure its compliance with EU law, the fundamental distinction between implementa‐ tion and enforcement. These two concepts root in the Treaties, in particular in Article 291 TFEU on the one hand, and Articles 258–260 TFEU on the other hand. Having described in more depth the similarities of and the differences between these two regimes, a number of characteristics – a list of indicators – was established. In most cases these indicators allowed for the allocation of (our selection of ) compliance mechanisms to either implementation or enforcement, or at least to a tendency towards either of these categories. This allocation was necessary as a first step in addressing the question to which extent the Treaties allow for the establishment of compliance mechanisms by means of secondary law.

In a next step, the primary legal bases of the compliance mechanisms laid down in secondary law were inspected with a view to their respective ade‐ quacy. Special attention was paid to Article 114 TFEU – a frequently used basis for EU secondary law in general, and for compliance mechanisms in particular. But also other Treaty provisions used as legal bases for setting up the sample of compliance mechanisms at issue here were interpreted with a view to their adequacy in the given context.

Strongly interwoven with both the distinction implementation/enforce‐ ment and the question of primary legal bases is the issue of the EU's institutional balance. This balance is struck by the Treaties and may not be distorted by means of secondary law. Against this background, it was examined whether and, if so, to which extent the secondary law-based compliance mechanisms at issue here – either on their respective own or in their entirety – challenge the EU's institutional balance, in particular by limiting the role of the Commission and the Court under the Treaty infringement procedure. In this context, also the principles of subsidiarity and proportionality are to be mentioned. They were analysed with a view to whether they suggest the use of soft law rather than hard law, in the context of compliance mechanisms this means: whether they serve as a guideline for the legislator, pointing in the direction of granting soft law powers rather than hard law powers.

The next sub-chapter was dedicated to the *epitheta* to be found in some acts of secondary law, requiring the respective addressees of soft law, for example, to take 'utmost account' of or to pay 'sufficient heed' to it. It was explored whether this actually strengthens the legal effects of soft law or whether it is nothing more than verbal ornamentation.

Finally, the legal protection which a MS may avail itself of in the context of compliance mechanisms was addressed in more detail. Apart from agen‐ cies' Boards of Appeal (if any) which may serve as an instance of legal protection, it is in particular two Court proceedings which were looked into here: the annulment procedure pursuant to Article 263 TFEU and the procedure following an action for failure to act pursuant to Article 265 TFEU.

### *2. Closing remarks and outlook*

### 2.1. Soft law

Soft law has become an important complement and sometimes competitor of law, the latter standing at the core of the development of the EU as a *Rechtsgemeinschaft*. <sup>2664</sup> Whereas law rather relies on the general authority of the regime it belongs to, the effects of soft law much more strongly root in the authority of its specific creator, not least because here compliance is not justiciable. Thus, while both law and soft law are power-based regimes, it appears that with soft law this power basis is individualised to a larger extent, which is why in practice its effectiveness seems to show a greater variance. Nevertheless, soft law must also be perceived as a whole, that is to say as a *general* phenomenon which is to be addressed by generalised questions. Some of these questions – with regard to the EU legal order – have been addressed in this work.

As a conclusion with regard to EU soft law, let me reiterate three (related) issues: 1. Soft law should be clearly recognisable as legally non-binding and leave room for deviating behaviour – both *de iure* and *de facto*. 2. The competence regime applicable in the context of soft law in general needs to be clear (eg principle of conferral or *in dubio* approach). While obscurity in single cases can hardly be avoided, a generally obscure competence situa‐ tion facilitates a decision-/policy-making culture in which soft law is – as a matter of course – adopted whenever hard rules are legally or otherwise not feasible, or where their adoption is at least cumbersome. This could easily lead to over-regulation and to a relativisation of (hard law) competences, neither of which seems to conform to the spirit of fundamental principles of EU law, such as the rule of law2665 or the principle of subsidiarity. 3. In the EU, the lack of a clear competence regime is combined with limited possibilities of judicial review. But even in view of the restrictiveness of Article 263 TFEU, the Court would have the possibility to scrutinise soft

<sup>2664</sup> Note the famous *dictum* of the Court in case 294/83 *Les Verts*, para 23: 'It must first be emphasized in this regard that the European Economic Community is a Community based on the rule of law, inasmuch as neither its Member States nor its institutions can avoid a review of the question whether the measures adopted by them are in conformity with the basic constitutional charter, the Treaty.' See also Trubek/Trubek, Governance 539, distinguishing rivalry, complementarity and transformation as three possible relationships between law and soft law.

<sup>2665</sup> Attempting, in view of EU soft law, a 'dynamic' conceptualisation of this principle: Dawson, Soft Law 14–16.

law more intensely, thereby – it is true – modifying its case law. Maybe this is even what it ought to do in order to counter-balance the lax application in practice of a competence regime which is fuzzy already in theory (ie according to the letter of the law).

Regularly, policy-making actors conceive of soft law as a convenient way to make rules where the adoption of legal rules would be unduly complicated, and if its addressees feel obliged because for them it is not clear whether the rules are legally binding or not – all the better. From a legal point of view, this attitude cannot go uncriticised. A more considerate, transparent, and traceable use of EU soft law – and here in particular the above issues 1 and 2 are addressed – would contribute greatly to doing away with its somehow dubious reputation. A clear commitment to and indication of its respective legal basis in (primary or secondary) EU law would be an important part of this approach – requirements which, argua‐ bly in the vast majority of cases, would not be difficult to meet. Another part would be an improved consideration of the 'truth of form' principle, that is to say a clear indication of the intended legal effects (in particular: legal bindingness or legal non-bindingness). These measures would affect eg the communications/recommendations which the Commission has, in places, adopted where an according Commission proposal has failed during a legislative procedure, or highly authoritative interpretative acts of the Commission relating to provisions of primary law, in particular in the field of competition and State aid law.

As regards the communications/recommendations replacing (failed) leg‐ islation, the potential of abuse is palpable. Where a legislative procedure has resulted in failure of the initiating proposal, also this should be accepted as a decision of the legislator – admittedly less authoritative than a positive decision. The clandestine conversion of a proposal to generally applicable (soft law) rules certainly does not display a high degree of deference to the legislator. In places, the Court has countered this attempted usurpation of rule-making power, namely were the act at issue presented the current state of law in an incorrect manner. It is important that the Court resolutely opposes attempts of abusing soft law in that way. Otherwise the indignation about such practice may fade out, as experience with respect to another legal source, the directive, has shown. In this case the Court has accepted the unorthodox way in which the legislator made use of it, namely as a tool to bring about highly detailed harmonisation or even unification of laws, with hardly any leeway for the MS transposing these acts. Originally, this has been considered abusive, as well,2666 but meanwhile – due to the Court's continuous approval – it has become a widely accepted purpose of directives.<sup>2667</sup>

As regards the interpretative acts, they are – under names such as 'Com‐ munication' or 'Guidelines' – declared legally non-binding, but *de facto* non-compliance on the part of the States or the undertakings will most probably be interpreted by the Commission not only as non-compliance with the interpretative soft law act, but also as a violation of the interpreted act, that is to say the underlying legal provision. This will again lead to adverse effects, eg a declaration of incompatibility with the internal market by the Commission in case of State aid or a fine imposed on an undertaking for violation of Article 101 or Article 102 TFEU. In terms of the actual effects – and presumably also in terms of the effects intended by the Commission when adopting these acts – there hardly seems to be any difference as compared to a legally binding act. Would not the truth of form principle dictate to adopt these acts as legally binding acts? The legislator – on the basis of Article 103 TFEU and Article 109 TFEU, respectively – could extend the power of the Commission in this respect. The Court – and here issue 3 above is addressed – has largely accepted such acts as legally non-binding and has, consistently, held in particular actions for annulment filed against them to be inadmissible. Only exceptionally, namely where the interpretation suggested in one of these acts went against EU law, did the Court confirm the admissibility of an action pursuant to Article 263 TFEU and subsequently annull these acts. That the Court annuls these acts where they violate EU law is to be embraced. However, while the Court in examining the admissibility of actions against such soft law acts considers the case in depth, it still allows for much room for manoeuvre for the EU actors concerned, in particular the Commission. By making use of this discretion, the latter may take undue regulatory action – in particular: a far-reaching and unprecedented interpretation of a certain rule – even without violating EU law. Thus, only the blatant cases are taken up by the Court, whereas more modest, but still practically important soft law will slip through, resulting in the inadmissibility of the respective action.

<sup>2666</sup> See Constantinesco, Recht 622–624, referring to the 'herrschende Meinung' [the prevailing view].

<sup>2667</sup> See Geismann, Art. 288 AEUV, para 41; Nettesheim, Art. 288 AEUV, para 113; Ruffert, Art. 288 AEUV, para 25; Schroeder, Art. 288 AEUV, para 54, each with further references.

Recent Opinions of AG have suggested a more progressive, a more liberal approach in confirming the admissibility of *prima facie* soft law.<sup>2668</sup> So far the Court has stuck to its case law, though.<sup>2669</sup> By regularly refusing the admissibility of actions in these cases, on the one hand the Court limits the legal importance of these acts and does justice to the wording of the Treaties and its case law so far. On the other hand, it may risk ignoring the factual meaning of these acts and it may miss an opportunity to render its authoritative view on this phenomenon more generally. A counter-argu‐ ment to the latter point of criticism could be that there already is a Court procedure which – in terms of its admissibility requirements – makes the Court easily accessible, namely the preliminary reference procedure.<sup>2670</sup>

Soft law has been and will remain to be a concept with many faces. Thus, whoever enters the scholarly discussion about it initially has to clarify his/her understanding of the term. This is not a shortcoming of the term, but actually indicates its sufficient flexibility to describe a multi-faceted phenomenon. But also on a more general level – in public discourse, eg in the media – the term soft law would probably stand the test as an intuitive and overall appropriate description of the subject matter: rules which are something less than, but still closely related to – or: 'in the penumbra of '<sup>2671</sup> – law. Unfortunately, as of now, the lively scholarly discussion does not seem to have evoked a continuous public debate – or at least this debate has not lead to a change in the use which is made of soft law. The merits and the risks of soft law have been elaborated in the literature in some depth, but – turning to the EU – the originators of EU soft law do not seem to make use of it too considerately.

In the EU soft law (addressed to MS) has been a success story, not least because its apparent ambition is low: It does not order compliance, it only suggests it. An overall picture of MS' compliance with EU soft law can hardly be drawn due to the multiplicity of different acts adopted in a variety of different situations. However, there is non-representative, but still

<sup>2668</sup> See eg the Opinion of AG *Bobek* in case C-16/16P *Belgium Commission*, paras 123 ff.

<sup>2669</sup> See Korkea-aho, Soft Law 290, who has limited hope in the Court adapting its approach towards soft law. For a reversal of the judicature of the French *Conseil d'État* in this context see Gundel, Rechtsschutz 600 f.

<sup>2670</sup> See Ştefan, Soft Law 20 ff, referring to arguments against and in favour of an extensive consideration of EU soft law before the CJEU.

<sup>2671</sup> Peters/Pagotto, Perspective 28.

evidence of remarkably strong compliance rates of EU soft law.<sup>2672</sup> Future research will be preoccupied with examining more closely the effects of soft law in selected policy fields, also and in particular from an empirical angle.<sup>2673</sup>

While it was argued that EU soft law may facilitate further integration, we should not forget about the core characteristic of this category of rules – its legal non-bindingness. Where it is addressed to MS, in principle each of them can decide for itself whether or not to comply. This constitutes a parallel to intergovernmental decision-making where each State – here: each MS – has to consent to (or, in case of consensus: not to veto) a certain measure. Where a MS does not 'consent' to a measure, it will not apply it. With soft law, the MS can decide anew in each case in which the soft law act would be applicable. Perceived from that angle, EU soft law – due to its (partly) 'intergovernmental' character – rather seems to work against Union method style integration. While this explanation may not be entirely satisfactory, either, at least it reminds us not to uncritically follow the beaten track, here: the dogma according to which soft law facilitates further integration of the EU. The EU is a *Rechtsgemeinschaft*, after all.

### 2.2. Compliance mechanisms

When it comes to the application of EU law in day-to-day administration, the MS and their respective authorities are the key actors. In this role, they cooperate with the Commission, eg in the field of competition law, in order to implement EU law *vis-à-vis* individuals/undertakings. At the same time, they are both addressees and creators of compliance mechanisms – creators either as Masters of the Treaties or, as participants in the Council, as (co-)legislator. *Schmidt-Aßmann* has described this as the 'eigentümliche[] triadische[] Rollenstruktur' [peculiar triadic role structure]2674 of the MS. This role structure in my view can hardly be overestimated when dealing with compliance mechanisms, as it discloses that the MS approve – not in each individual case, but in principle – of the implementation/enforcement of EU law *vis-à-vis* themselves. This is remarkable not least due to the fact that one of the findings of this work points to instances of a materialising

<sup>2672</sup> See eg Hartlapp, Soft law.

<sup>2673</sup> See already Eliantonio/Korkea‐aho/Stefan, Soft Law.

<sup>2674</sup> Schmidt-Aßmann, Verwaltung 1382.

risk of a creeping expansion of the EU's enforcement powers *vis-à-vis* the MS – an expansion MS in general seem to be wary of. In case of secondary law-based compliance mechanisms, however, they have – with their appro‐ val in the Council – significantly contributed to this expansion.

With regard to the relationship between the Treaty infringement proce‐ dure and special compliance mechanisms laid down in secondary law, essentially three statements can be made:

1. Compliance mechanisms – enforcement mechanisms as such, but also, due to their large quantity, the implementing mechanisms in place – lead to a restriction of the competences of the Commission and the CJEU under the Treaty infringement procedure.

2. The legislator is providing for special compliance mechanisms more and more frequently, leading to an increase in the total number of compli‐ ance mechanisms throughout the policy areas which are shaped by the EU.

3. The number of pending Treaty infringement cases has decreased dras‐ tically since the mid-2000s – in spite of an ever increasing amount of EU rules (secondary law) and in spite of the fact that the number of MS has nearly doubled since 2004, two factors which one would intuitively assume to boost the number of infringement procedures.

These three statements justify the assumption – which, as a matter of course, requires further research to be proven – that a causal relationship exists between the increasing number of compliance mechanisms in secon‐ dary law and the decrease in the number of Treaty infringement procedures performed in practice. Additional reasons for this decrease may have been the Commission's selective approach in pursuing violations of EU law, the introduction of EU Pilot, conditionality-based regimes and other EU tools created to improve compliance.<sup>2675</sup>

A solid account of this question would also require a quantitative analy‐ sis, reviewing the decreasing number of Treaty infringement procedures over the years, and examining possible correlations or even causalities on the part of the alternative compliance mechanisms in that respect. On a basic level, this would involve offsetting the decrease in the number of Treaty infringement procedures with a (potential) increase in the number

<sup>2675</sup> See eg Commission, 'Enforcing EU law for a Europe that delivers', COM(2022) 518 final, 15 f. Also private enforcement is to be taken into account in this context. However, private enforcement can only play a complementary role, not least because in some policy fields it is not available (eg EMU, Schengen agreement); see Bieber/Maiani, Enforcement 1058 f.

of applications of alternative compliance mechanisms and of the number of annulment procedures launched against the hard law output adopted in the course of (mixed or hard) compliance mechanisms. Arguably, this research cannot be done in a comprehensive manner, thereby taking account of all compliance mechanisms laid down in secondary law. Rather, such research is manageable only with regard to selected policy fields, and only step by step a more encompassing picture of potential causalities may thereby be drawn. This research may also be required to take account of different compliance cultures in the different MS, which may lead to geographically heterogeneous results.<sup>2676</sup>

Since the early days of the European Communities, the Treaty infringe‐ ment procedure has borne a strong political dimension. Objective and comprehensive legal enforcement has been impeded by:


And still the Commission has – for a long time – attempted to further decrease the number of Treaty infringement procedures.2677 The Treaty infringement procedure is perceived as suitable to settle 'big cases', that is cases of principle, or obvious violations, but not to be applied on an everyday basis to solve comparatively minor legal issues.

In view of the wide-spread discontent with the Treaty infringement pro‐ cedure, for the latter purpose a large number of compliance mechanisms has been set up in particular in secondary law. Their functioning is regular‐ ly supported by an improved information flow between the national and the EU level, and they allow for fast(er) decision-making by the Commis‐ sion or other, even more technocratic bodies. However, it is to be noted that most of them have an implementing thrust, aiming at the concretisation of EU law rather than – like the Treaty infringement procedure – at the determination and subsequent removal of its violation. On a meta-level,

<sup>2676</sup> See Tomkin, Enforcement 292. For MS which have traditionally been weak in complying with EU law see Ioannidis, Members 476.

<sup>2677</sup> See references by Koops, Compliance 119.

these different aims can be unified under the larger objective of ensuring MS' compliance with EU law.

In its State of the Union 2012 Address, then President of the Commission *Barroso* uttered his concern about 'threats to the legal and democratic fabric in some of our European States', pronouncing a claim for a 'better developed set of instruments' to monitor observance of this fabric in the(se) MS – 'not just the alternative between the "soft power" of political persua‐ sion and the "nuclear option" of Article 7 of the Treaty [on European Union]'.<sup>2678</sup> The Treaty infringement procedure could be added to the instruments already available, but also this tool – while having confirmed the Commission's criticism in some cases<sup>2679</sup> – does not seem perfectly suited to address the underlying, rather structural problems referred to by *Barroso*. <sup>2680</sup> Neither could the specific compliance mechanisms at issue in this work prevent or at least contain these developments.

The latter compliance mechanisms are intended to deal with politically less loaded, but still relevant violations of EU law, thereby partially compen‐ sating for the staidness of the Treaty infringement procedure. Facilitating MS' observance of the fundamental principles on which the EU is built is a relatively new focus of the EU's broad objective of achieving 'compliance with Europe'2681 – a focus which, in addition to the mechanisms just men‐ tioned, requires (and in part has already led to the creation of ) new tools.

<sup>2678</sup> Barroso, State 10; for this speech and the ensuing 'Rule of Law Initiative' of the Commission see Besselink, Bite 134–136.

<sup>2679 &</sup>lt;https://www.dw.com/en/top-eu-court-rules-against-polish-judicial-reform/a-5111 4974> accessed 28 March 2023.

<sup>2680</sup> See also Gormley, Infringement 75 f, with further references.

<sup>2681</sup> This term is inspired by Falkner/Treib/Hartlapp/Leiber, Europe.





697


Streinz, Rudolf, *Europarecht* (10th edn, CF Müller 2016)

Streinz, Rudolf, *Europarecht* (9th edn, CF Müller 2012)


# Index

#### **A**

abuse of soft law 383 accountability 76 ACER 368, 453, 498, 502, 505, 509, 512, 556, 558, 561, 581, 582, 586 *acta iure imperii* 103 action for annulment 631 – national authorities 632, 633 – privileged and non-privileged claim‐ ants 632 action for failure to act 634 addressees of EU soft law 184–187 alternative steering instruments 89, 90 approximation of laws – bank resolution 431, 572 – concretisation 565 – cross-sectional competence 565 – electricity market 572 – electronic communications 572 – ENISA 567 – improper use of exceptions 421, 422 – individual-concrete measures 571 – macro-prudential oversight of the Union financial system 572 – market for financial services 572 – national measures 420 – natural gas market 572 – product safety 572 – soft law 567

#### **B**

*Balgarska Narodna Banka* case 310 BEPG 262, 274, 471 BEREC 476, 499, 503, 507, 512, 577 – reasoned opinion 477 *Bernadotte* case 213 binary concept of law 84, 91 breach of Union law 457

*argumentum a maiore ad minus* 220–229

#### **C**

CEBS 170 CEIOPS 170 CEN 141 CENELEC 141 CESR 170

*CIRFS* case 311 CJEU – Opinion 544 Code of Conduct 100 collision rules 105 comitology 586, 600, 606 – soft law 299–302 Commission – formal opinion 458, 460 – guardian of the Treaties 491 – opinion 484 – reasoned opinion 477 committees – opinions 247–251 communication 195 competition and effective functioning of the energy market 433 compliance – concepts 397 compliance mechanisms – addressees 522 – approximation of laws 495, 567–579 – categorial overlaps 487 – comparison with Treaty infringement procedure 494 – competence conflict 556 – confidentiality 519 – delimitation 399 – deviation from Treaty infringement pro‐ cedure 526 – dialogic nature 563 – discretion 518 – dominance of the Commission 493 – efficiency 520 – emergency situation 556 – EU bodies involved 490–495 – *ex ante* intervention 517–519, 531, 557, 559–562, 578, 587 – explanations for variation 528 – *ex post* intervention 531, 555–561, 563, 575, 581 – judicial review 493–495, 574, 597–599 – legal basis (in primary law) 495, 496 – measures *vis-à-vis* individuals/undertak‐ ings 569


#### *Index*

– proportionality 573 – publicity 529 – reinforcement of Treaty infringement pro‐ cedure 516 – sanctions 634 – sequence of acts 497–500 – structure 496–500 – subsidiarity 609 comply or explain 120, 121, 366 concepts of soft law 84–90 conditionality 635 *Conseil d'État* 158, 374, 378, 380, 501, 506– 508 consensus 86, 138, 283, 287, 647 *consuetudo* 137 *contrarius actus* 311, 480 CoR 250 co-regulation 171 COREPER 249 Council – recommendations 480 – reverse qualified majority voting 371, 462, 464, 511, 548 creation of EU soft law – EIB 286 – High Representative 285 – MS 291, 292 – national parliaments 290, 291, 371 custom 136–138 customary law 136–138

#### **D**

declaratory acts 118, 237, 251, 314, 618, 626 definition of law 108 definition of soft law 107 determination of a norm 120, 163, 187, 238, 240, 541 difficulties as regards MS' balance of pay‐ ments – mutual assistance 424, 426 – necessary protective measures 424 direct effect of WTO law 128–132 direct execution of EU law 601 disagreement between national authori‐ ties 577 – disagreements between competent authorities in cross-border situa‐ tions 436 distortion of the conditions of competition in the internal market 444 duty to consider soft law 622–630 duty to give reasons 357–359, 623, 626, 628

#### **E**

EASA 166, 233, 555, 579 EBA 372, 456, 498, 510, 556, 559 – application of national law 437, 459 – evocation 438, 460 ECB – independence 206 – opinions 245, 246 – recommendations 244, 245 EDA 240 EFC 248, 449 effectiveness of (soft) law 112, 113 *effet utile* 207, 212, 214, 348, 362 efficient breach theory 135 EFSA 292, 369 EFSF 159, 173 EFSM 173 EIB 261, 291, 301, 371 EMA 166 enforcement – CJEU 542 – determination of an infringement 538 – *ex post* occurrence 536 – general definition 532–534 – indication of legal bindingness 96 – material scope 540 – national authorities 541 – primary aim 535 – private law 111 – public international law 110, 111 – sanctions 546–548 – soft law 109, 545 – subsidiarity 612 enhanced cooperation 285 equality 347, 360 ERA 467, 498, 502, 505, 511, 560 ESAs 625 ESC 250, 269 ESM 173 ESMA 150, 253 *ESMA* case 569, 602 ESOs 141 ESRB 481, 513, 563, 578, 598 – recommendations 481 ETSI 142 European agencies 368, 591, 600–606 – appeal bodies 630 European Court of Human Rights 593 EU soft law other than recommendations and opinions 293–297 examination procedure 427 excessive deficit procedure 274, 306, 309, 318, 446, 544 – sequence of acts 498

– structure 527

excessive imbalance procedure 559 – sequence of acts 498 *Expedia* case 326 extra-legal steering effects 88

#### **F**

factual effects of soft law – human nature 363–365 – nudging 365–367 facultative consultation 250 FIFA 98 financial assistance 172–178, 180 flexibility clause 215, 303 formal requirements of law 114 *forum internum* 139 FRA – soft law 303–306

#### **G**

general principles of law 87, 350 genetically modified organism 439 *Grimaldi* case 208, 221, 322, 324, 325, 327 guidelines 195 guidelines on employment policy 472

#### **H**

hard mechanisms – concretisation 553 – enforcement 553–557 – *ex ante* intervention 553 – implementation 553–557 – on-going intervention 553 hierarchy of soft law acts 629

#### **I**

IEC 141 IMF 174 *imperium* 102 implementation – concretisation 538 – European agencies 542 – *ex ante* occurrence 535 – general definition 532–534 – judicial review 537 – material scope 540 – national authorities 541 – on-going occurrence 535 – primary aim 535 – sanctions 546–548 – soft law 545 – uniform conditions 541

implementing mechanisms – subsidiarity 612 implied powers 212–219 – competence-accessory 214 – necessity 217 – objective-accessory 214 indirect execution of EU law 601 informative acts 146–148, 187–191 initiation of decision-making – Council 271, 272 initiation of rule-making – ECB 282–284 initiation of (soft) decision-making/rulemaking – Commission 262–267 instant custom 138 institutional balance 226, 485, 505, 511, 514, 515, 530, 546, 588–599, 602 integration by stealth 206, 596 internal soft law – characteristics 229–233 – competence 233–235 international agreements – ECB 283 international *courtoisie* 92 International Law Commission 103 international public authority 89 interpretation of national law in line with Union law 342, 343 interpretation rules 105, 107 IOC 98 ISO 141 *ius cogens* 85, 125 *ius dispositivum* 121

### **J**

judicial review – preparatory acts 631 – *reformatio in peius* 635 judicial review of soft law – action for damages 392 – annulment procedure 385 – incidenter review 395 – internal legal effects 386 – legal effects *vis-à-vis* third parties 385, 387

– *prima facie* soft law 387

#### **K**

*Kadi* cases 122–128 *Kergall* case 234 *Koninklijke* case 328 *Kotnik* case 341

*Index*

#### **L**

lack of legal basis 308–312 legal certainty 344, 346, 360 legal positivism 87, 88, 114, 637 legal referencing 144–146 legislative procedure 628 – trilogue 620 legitimate expectations 344, 346, 360 *lex imperfecta* 118 *lex sportiva* 98 *Lotus* case 83 *Louwage* case 350 loyal cooperation – fisheries policy 339 – State aid law 340

#### **M**

*Macevicius* case 234 macroeconomic imbalances 461, 506 – sanctions 463 *McBride* case 210 *Meroni* doctrine 303–306, 485, 530, 588– 590, 602 mixed compliance mechanisms 468 mixed effects of soft law 370–373 mixed mechanisms – competence 469 – enforcement 557–560 – implementation 557–560 – lack of urgency 469 morals 139, 140 MoU 160 – conditionality 179 – legal bindingness 179–184 – legal order 172–178 MS competence to deviate from EU law 443 MS' right to be heard 443 multilateral surveillance procedure 583

#### **N**

naming and shaming 120, 366 *Nefarma* case 210 network codes and guidelines 453 NGOs 100, 103, 142 non-normative output 118, 187 non-self-executing norm 86 norms adopted by private actors 96–104 *Nuclear Weapons* case 80, 93 *numerus clausus* of soft law acts 115, 150, 195–199, 297, 312, 313, 316

#### **O**

*Obliegenheit* 96, 97 obligation – degrees of bindingness 85, 87 – dimensions of obligation 85 – obligation of effort 338, 622, 625, 626 – obligation of result 338, 622, 625 *obligatio naturalis* 83 obligatory consultation 250 occurrence of EU soft law 152–157 occurrence of soft law 79–81 Official Journal of the European Union – C-series 198 – L-series 198 *opinio iuris* 81, 93, 136–138 opinion 192–194 – *reactio* 312 Opinions of AG 279–282 ordinary legislative procedure 189–191, 223, 225, 239–243, 255, 261, 291 ordinary Treaty revision procedure 287 *Organkompetenz* 201, 215, 216, 233, 237, 564, 589 *Organleihe* 307

#### **P**

*pacta de contrahendo* 86 policy requirements – action in emergency situations 435, 576 – civil aviation safety 428 – difficulties as regards MS' balance of pay‐ ments 424 – electricity transmission operations by third-country actors 475 – essential interests of security 423 – orderly functioning and integrity of finan‐ cial markets 435 – stability of the financial system 435 – urgency 443, 558, 573, 574 – urgent unforeseeable circumstances in civil aviation 429 preliminary reference procedure – creators of eligible acts 389 – effects of eligible acts 389 – enforcement 543 – implementation 543 – questions of interpretation 390 – questions of validity 390 – standard of review 392 presumption of legal force 79, 95, 181, 182 principle of conferral 238


principle of private autonomy 83 private authority 106 procedural rules 114 programmatic acts 114, 118, 155, 333 proportionality 218, 221, 607–622 – complex assessments 621 – definition 607 – judicial review 621 – necessity 618, 619 – proportionality in the narrow sense 615, 619 – test 615 proposal – legal qualification 294–296 PSC 247 public international law – comparison with private law 83, 84 – idiosyncrasies 82 publicity 104, 114, 188, 232, 578 purposes of soft law 80, 501–515 – avoidance of law 381 – circumvention 80, 255, 256 – company 376, 378, 507, 508 – concretisation 80, 374 – facilitating right to be heard 524–526 – factual purposes 376 – harmonisation 374 – information 526 – legal purposes 376 – permanent alternative 376, 378, 382, 508–514 – preliminary commitment 80


#### **Q**

questions and answers 195

#### **R**

reasons for adoption of soft law – inclusivity 382 – lack of competence to adopt law 382 reciprocity 130 recommendation 192–194 – *actio* 312 references – dynamic 144 – static 144 relative normativity 87, 88, 91 religious norms 105, 112

resolution 195

retroactivity 347

risk


#### **S**

safeguard clauses 438 sanctions 584, 594 – judicial review 634 screening of foreign direct investments 483 self-bindingness of soft law 108, 110, 232, 301, 351, 354, 378 self-regulation 171 separation of powers 76, 532 serious difficulties with respect to their finan‐ cial stability 480 services of general economic interest 416 sincere cooperation between the EU institu‐ tions 253, 354 – comitology 356 Six Pack 546 soft decision-making/rule-making – Commission 267–270 – Council 272–275 – ECB 284, 285 – European Council 275, 276 solidarity clause 285 SPC 249 special legislative procedure 239, 242, 260 SRB 498, 555, 560, 574 – evocation 466 – warning 465, 466 Stability and Growth Pact 446 standards 140–146 State aid – binding proposal 583 – existing aid 417, 418, 454 – infringement procedure 419 – misuse 418 – new aid 417, 419 steering of human behaviour 67, 114 subsidiarity 219, 220, 256, 607–622 – definition 607

#### *Index*


#### **T**

*Tillack* case 335 traditional distinction between law and nonlaw 76 transport policy 579 Treaty infringement procedure – burden of proof 410 – Commission 490, 491 – Commission's discretion 518


#### **U**

Union control procedure 438 UN Security Council 122–128 *usus* 137, 138

#### **V**

validity 92, 119, 128, 255, 310, 389, 390, 427, 515, 543, 602 *Verbandskompetenz* 201, 215, 216, 233, 236, 244, 248, 249, 564 *Verwaltungsverbund* 368 voidness 308–310

#### **W**

will of the norm-creator 78, 95, 96, 108, 110, 112, 115, 118, 129, 637 WTO Dispute Settlement Body 131