Oliver Kossack Pariahs or Partners?

**Political Science** Volume 153

**Oliver Kossack**, born in 1985, studied social and cultural sciences (BA) and European studies (MA) in Germany and Denmark. From 2015 to 2020, he was a project manager at the NGO Cultures Interactive in Berlin, Germany, focusing on the prevention of right-wing extremism and the promotion of democratic values and human rights. Since 2021, he has been working as personal adviser to the president of the Europa-Universität Viadrina Frankfurt (Oder). His research interests include radical right parties and movements, the post-communist transformation and strategies for preventing illiberal trends and right-wing extremism in Germany and (Eastern) Europe.

Oliver Kossack

## **Pariahs or Partners?**

Patterns of Government Formation with Radical Right Parties in Central and Eastern Europe, 1990-2020

This book is based on a PhD dissertation defended by the author at the Faculty of Social and Cultural Sciences of the European University Viadrina in Frankfurt (Oder) on 6 May 2022. The dissertation was supervised by Prof. Dr. Michael Minkenberg.

This publication was supported by funds from the Publication Fund for Open Access Monographs of the Federal State of Brandenburg, Germany.

#### **Bibliographic information published by the Deutsche Nationalbibliothek**

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

This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 (BY-SA) which means that the text may be remixed, build upon and be distributed, provided credit is given to the author and that copies or adaptations of the work are released under the same or similar license.

Creative Commons license terms for re-use do not apply to any content (such as graphs, figures, photos, excerpts, etc.) not original to the Open Access publication and further permission may be required from the rights holder. The obligation to research and clear permission lies solely with the party re-using the material.

#### **First published in 2023 by transcript Verlag, Bielefeld © Oliver Kossack**

Cover layout: Maria Arndt, Bielefeld Copy-editing: Aaron Martin, Munich Printed by: Majuskel Medienproduktion GmbH, Wetzlar https://doi.org/10.14361/9783839467152 Print-ISBN 978-3-8376-6715-8 PDF-ISBN 978-3-8394-6715-2 ISSN of series: 2702-9050 eISSN of series: 2702-9069

Printed on permanent acid-free text paper.

## **Contents**




## **List of tables**





```
Table A3.1 Calibration of the higher-order condition of socio-cultural similarity
(SOCCUL) 386
```
## **List of figures**


## **List of parties and electoral alliances**

#### *Bulgaria*


#### *Czech Republic*


#### *Estonia*


#### *Hungary*


#### *Latvia*



#### *Romania*


#### *Slovakia*


#### 20 Pariahs or Partners?


## **Acknowledgements**

My interest in right-wing radicalism emerged outside of academia. Growing up in rural Eastern Germany in the 1990s and 2000s, the radical right pervaded much of my teenage life. I am very grateful tomy family and friends,mostimportantly Fliege, Jana, Marian, Anika, Nippel, Thessa, Clemens, Dana, and Elvi who raised and socialised me in a way that I never felt attracted by these groups. When I had to take a side, I joined the hippies, punks, metalheads, indies, and ravers who all gathered in the Nevermind, the only place for alternative subcultures in the region. A whole generation is eternally grateful to Ilka, Ivo, Lutz and all of the others who ran this place, even while working regular jobs.

My constant personal struggle with right-wing radicalism also sparked my academic interest in this phenomenon. I wanted to know what was going on inside the radical right and why so many people my age joined them. The Europa-Universität Viadrina in Frankfurt (Oder) has been the base for my academic research during the last 15 years. I owe special thanks to Michael Minkenberg for his patient and encouraging supervision of my dissertation. His full support continued even after I started a new job outside academia at Cultures Interactive, which took my focus off this project. I would also like to thank Timm Beichelt for his constructive feedback on my academic projects. My gratitude extends to Malisa Zobel, Anja Hennig, Artur Kopka, Tobias Peter, Zsuzsanna Végh, Anca Florian, Andrea Pirro, and Cas Mudde, who provided valuable feedback on bits and pieces of my research in various colloquia and workshops at the Viadrina, and to Michaela Grün, who always motivated me to continue my academic work. I owe special thanks to Aaron Martin for his thorough, skilful, and patient editorial work on the manuscript.

On several research trips and, in particular, during numerous seminars, workshops, and conferences with Cultures Interactive, I had the opportunity to meet experts, such as Oľga Gyárfášová, Grigorij Mesežnikov, Radu Cinpoeş, Péter Krekó, András Bíró-Nagy, Kiril Avramov, Louis Wierenga, Petra Mlejnková, Vladislava Vondrová, György Tatár, Tina Gažovičová, Radka Vicenová, Lukáš Zorád, Maros Chmelík, Kita Boncheva, Dorota Bregin, Magdalena Chustecka, Júlia Vida, and János Wágner, who shared with me their insights on the radical right, politics, culture, history, and night life in different parts of Central and Eastern Europe. I would like to thank the entire team at Cultures Interactive for the professional exchange and the opportunity to work in and on Central and Eastern Europe. All these encounters helped to improve this book and, more importantly, my personality.

During my time at the Viadrina I also met Maria, Fo, Pınar, Elaine, Bartek and Léa, some of the best friends you can imagine. Utopia e.V. and in particular our music venue, the Garage, came to replace the Nevermind, which had to close after so many influential years. Thank you to Anika, Kermit, Janek, Rhena, Simon, Sarah, Franzi, Robert, Laura, Willi, Basti, Frank, Jochi, and everyone else who spent their time and energy to run this place that so often created the necessary diversion from my academic work. Among all my friends, Anika holds a special place, for we have been partners in crime ever since we met in school and after school in the Nevermind. Thank you for your friendship and the great company over all these years.

Finally, I would like to thank some people without whom this book would have never seen the light of day. My parents and grandparents always encouraged me to pursue my ideas and dreams, even if they seemed utterly strange to them.Thank you for the freedom, your love and trust! I am also indebted to Bartek who supported me with this project from the very beginning. His feedback on the final manuscript and many earlier drafts has been invaluable, as has his friendship outside of academia. I owe my deepest gratitude to Ulli. Thank you for your love, friendship, and patience, for keeping me sane and for having my back, particularly during the final stages of the project.

This book is dedicated to our cats Paula, Szymek, and Marie.They could not care less but they have been a constant source of happiness and distraction during a long and demanding research process.

## **1. Introduction**

#### **1.1 Research interest and research question**

The presence of radical right parties has become the norm in European democracies, even though large parts of the public reject their ultra-nationalist and illiberal ideology. In 2000, when the conservative Austrian People's Party (ÖVP) invited the radical right Freedom Party of Austria (FPÖ) to become a junior partner in the Austrian government, all member states of the European Union (EU) sanctioned Austria.This drastic measure indicates that political elites inWestern Europe viewed radical right parties as pariahs, even after two decades of substantial electoral success.The external intervention, however, remained largely inconsequential for Austria's position in the EU, and it did not prevent the ÖVP from further cooperation with the radical right. Since this pivotal event, the participation of radical right parties in government, either in Austria or in other Western European EU member states, has not triggered such a diplomatic outcry.

The story of radical right parties participating in government coalitions is somewhat different in the Central and Eastern European EU member states. Despite limited electoral success and organisational instability (Minkenberg 2002, 336, 2017; Mudde 2005a), radical right parties have entered government from the onset of the post-Communist transformation. The first governments that included radical right parties were formed in Estonia, Romania, and Slovakia in the early 1990s, but their government participation has remained neither limited to these countries nor to this time period. Contrary to the events described in Austria, the government participation of radical right parties in this region never caused similar protestations from European political elites.

In the past three decades, radical right parties have thus had the opportunity to directly influence political developments from the highest public office in many Central and Eastern European countries. Moreover, when radical right parties participate in government, this indirectly impacts national party systems by shifting the policy positions and narratives of their competitors to the right—even after they have left office (Pytlas and Kossack 2015; Pytlas 2016; Minkenberg et al. 2021). In light of these facts, the present study seeks to answer the following research question: What explains the government participation of radical right parties in Central and Eastern Europe in the first three decades after the fall of Communism?

## **1.2 Research overview and argumentation**

## Theoretical framework

The formation of governments with radical right parties in Central and Eastern Europe has received very limited scholarly attention (Fagerholm 2021). Nevertheless, by combining the research on radical right parties and the formation of government coalitions, this study draws on the literature from two established branches of comparative politics. Regarding the research on radical right parties, Cas Mudde (2007, 2) notes that works on this party family "might already outnumber the combined total of books on all party families together". He is quick to add, however, that there is still much to discover. The central areas of research on radical right parties include topics such as their ideology (Ignazi 1992; Minkenberg 1998; Mudde 2000b; Carter 2005; Pirro 2016), organisational structure (Art 2011), and the societal demand for these parties, often in view of explaining their success or failure at the polls (Scheuch and Klingemann 1967; Betz 1994; Kitschelt and McGann 1995; Norris 2005; Meguid 2005,2008;Mudde 2010;Minkenberg 2013; Pytlas 2016).More recently, scholars have devoted increasing attention to the impact of radical right parties on politics and society (Williams 2006; W. M. Downs 2012; Minkenberg 2015a, 2015b; Pytlas and Kossack 2015; Minkenberg et al. 2021), including their participation in government in Western European democracies (Minkenberg 2001; Bale 2003; Heinisch 2003; Akkerman 2012; Akkerman and Lange 2012; Zaslove 2012). It took about a decade before researchers also started paying attention to radical right parties in the new democracies of Central and Eastern Europe (Ramet 1999; Beichelt and Minkenberg 2002; Mudde 2005b) and yet another decade for more comprehensive comparative studies to emerge (Minkenberg 2009, 2015b; Pirro 2016, 2017; Pytlas 2016; Minkenberg et al. 2021).

The most comprehensive study on the participation of radical right parties in coalition governments focuses only on Western Europe (de Lange 2008). De Lange finds that the same criteria which explainmainstream party participationin government—namely parties' pursuit of public office and the goal of implementing their preferred policies—also apply to the radical right. More precisely, the ideological distance of radical right parties to the formateur<sup>1</sup> , particularly on the issue of im-

<sup>1</sup> The formateur of a coalition is the party that leads coalition bargaining and usually also appoints the prime minister. In some countries, constitutional provisions comprise the formal appointment of a formateur, mostly by the head of state, whereas in others the formateur is selected on the basis of a code of conduct in the country's political arena.

migration, and their seat share in parliament, determine whether or not they enter government or remain in opposition (de Lange 2008, 119; see also de Lange 2012).

Two political developments also influence the participation of radical right parties in government in Western Europe (Bale 2003; de Lange 2008). The rise of green and radical right parties caused party systems to become increasingly polarised, pitting left-wing and right-wing camps against one another. In this environment, conservative and Christian democratic parties were often unable to form centre-right majorities without involving radical right parties. Furthermore,in order to win back voters centre-right parties shifted their policy positions towards those of the radical right, particularly with regard to immigration.Thus, the electoral success of the radical right, coupled with their impact on mainstream parties' policy positions, paved the way for radical right parties to enter government in Western Europe.

The only comparative study of government formation which includes a discussion of radical right parties in Central and Eastern Europe is an analysis of 22 European countries which focuses on all radical—left and right—parties (Fagerholm 2021). This study provides some support for the importance of ideological and electoral factors for explaining radical right parties' inclusion in, and exclusion from, government. Yet, due to the broad scope of the study and because the results are based on only a few cases from Central and Eastern Europe which are skewed by the Latvian case, the conclusions remain rather tentative.Therefore, the author himself emphasises the need for further research (Fagerholm 2021, 16).

Existing research on party competition with radical right parties in Central and Eastern Europe offers alternative explanations for why these parties' gain entrance into government coalitions so frequently. Minkenberg and several co-authors, for instance, develop a model (see Fig. 1) for assessing the impact of radical right parties on other parties' policy positions, the political culture, and the quality of democracy, focusing in particular on processes of interaction between radical right parties and their competitors (Minkenberg 2015a, 2017; Minkenberg et al. 2021; see also Meguid 2005, 2008). The model does not distinguish between electoral and post-electoral party competition, such as government formation (Benoit and Laver 2006, chap. 2), but it does highlight the general importance of mainstream parties' strategic reactions to the radical right.The model also illustrates that whether mainstream parties cooperate with, or distance themselves from, the radical right depends on a number of factors, including the perceived (electoral) threat of radical right parties, their policy positions, the configuration of party systems, and the cultural context. Thus, more general research on party competition with the radical right in Central and Eastern Europe points to similar explanatory factors as the research on government formation with radical right parties in the western part of the continent.

Research into coalition politics provides another point of reference for this study (for an overview, see Laver and Schofield 1998; Kropp, Schüttemeyer, and Sturm 2002b; Müller, Bergman, and Strøm 2008). Within this literature, the formation of coalition governments has been one of the earliest and most prominent puzzles that scholars have tried to solve. Initially, researchers sought to predict the composition of coalitions after a given election, while other works tried to explain the participation of individual parties in government (de Winter and Dumont 2006).

*Figure 1.1: Model of party competition with radical right parties in Central and Eastern Europe*

Source: Minkenberg et al. 2021, 651.

The mainstream of coalition research follows a context-sensitive rational choice approach. Hence, scholars believe that coalition formation results from parties' strategic choices in the pursuit of office and policy, which are constrained by contextual factors, such as the institutional environment, the configuration of party systems or historical trajectories (de Swaan 1973; Strøm 1990a; Strøm, Budge, and Laver 1994; Müller and Strøm 1999, 2000b; Kropp, Schüttemeyer, and Sturm 2002a; Mitchell and Nyblade 2008).

Research on coalition formation focuses mainly on Western European democracies, and studies on Central and Eastern Europe are still rather rare (Grotz and Weber 2011; Döring and Hellström 2013; Savage 2014, 2016; Bergman, Ilonszki, and Müller 2019a).However,one of themore generalinsights foundin these worksis that the rich theoretical body of coalition research can be applied in Central and Eastern Europe as well. Furthermore, there is substantial agreement that office-oriented and contextual factors play an important role in explaining government formation in Central and Eastern Europe (Grotz and Weber 2011; Döring and Hellström 2013; Savage 2016).The influence of policy-related factors, by contrast, remains disputed. While most works subscribe to a predominantly office-oriented notion of coalition formation in Central and Eastern Europe (Döring and Hellström 2013; Savage 2016; see also Bergman, Ilonszki, and Müller 2019a), one study finds empirical evidence showing a significant relationship between parties' policy positions and participation in government (Savage 2014). Savage (2014, 558) argues convincingly that the frequently observed irrelevance of programmatic competition in Central and Eastern Europe results from a mis-conceptualisation of the region's policy space, which cannot be adequately captured by the classic (Western European) left-right dimension.

#### The radical right and party competition in Central and Eastern Europe

This brief overview of research on radical right parties and coalition formation suggests that the characteristics and preferences of political parties play an important role in government formation, as does the context of coalition bargaining. Therefore, specific features of Central and Eastern European radical right parties, as well as the social and political context, must be taken into account when attempting to explain their path(s) to power. Though the present study assumes that radical right parties, and the political processes underlying party competition in the democratic Central and Eastern European countries, function equivalently to those in Western Europe, it acknowledges that the region's post-Communist, context-specific characteristics must not be ignored (Pytlas 2018; see also Minkenberg 2002, 2015a, 2017; Pirro 2016; Mudde 2017).

The first important difference betweenWestern and Central and Eastern Europe concerns the conditions under which radical right parties emerged in both parts of the continent. In Western Europe, the rise of radical right parties is often seen as a counter-reaction to the post-materialist value changes that beganin the 1960s (Ignazi 1992, 2003; Minkenberg 1998; see also Inglehart 1977). In Central and Eastern Europe, they emerge after 1989 in the context of a "triple transition", which included the mammoth tasks of building new economic and political systems in new—or at least newly independent—nation states (Offe 1991; see also von Beyme 1996; Elster et al. 2000). In both parts of the continent, massive modernisation surges accompanied the emergence of the radical right, and they were even more severe in Central and Eastern Europe than in the West (Beichelt and Minkenberg 2002; Pytlas 2016; Minkenberg 2017). Due to the importance of state- and nation-building during the post-Communist transformation, the issue of national identity is also highly salient in Central and Eastern Europe (Offe 1991; von Beyme 1996; Bunce 2005). In such an environment, nativist ultranationalism, the ideological core of the radical right, has resonated with political and societal actors since the onset of the transformation process.

Second, the party systems of Central and Eastern Europe are peculiar in terms of their structural and content dimensions, and this distinguishes them from their Western European counterparts.On the structural level, post-Communist party systems consisted of wholly new parties, with the exception of Communist successors and few revived historical parties (von Beyme 1996, 127–29; Cabada, Hloušek, and Jurek 2014, 53; see also Ágh 1998; Elster et al. 2000). However, parties did not compete in a tabula rasa situation, even in the early phase of the transformation (Elster et al. 2000, 131). Detailed party programmes were often unavailable during this period and individual personalities enjoyed considerable influence over the political process, but political parties formed around certain core values that provided voters and other parties with some guidance from the outset (Hloušek and Kopeček 2010, 9–10).The familiarity with parties and their positions had increased markedly after a decade's worth of party competition, elections, and the government participation of different political forces (Ágh 1998; Toole 2000; Pop-Eleches 2010). Hence, by the turn of the millennium, it became possible to speak of a basic level of party system institutionalisation in the region, even though the level of stability was considerably lower than in the established Western European democracies (Enyedi and Casal Bértoa 2018; Emanuele, Chiaramonte, and Soare 2020).

Third, the conceptualisation of the policy space is particularly important when it comes to the ideological configuration of Central and Eastern European party systems. The crucial issue here is the nature and alignment of cleavages, or divides (Deegan-Krause 2007), which structure party competition. There is a broader debate about whether or not it is possible to apply the classic left-right dimension to Central and Eastern Europe and whether the socio-economic and socio-cultural divides that constitute this dimension are predominantly reinforcing or rather crosscutting (Kitschelt et al. 1999; Beichelt 2001; Marks et al. 2006; Deegan-Krause 2007; Rohrschneider and Whitefield 2009; Casal Bértoa 2014; Cabada, Hloušek, and Jurek 2014). The view here is that socio-economic and socio-cultural divides are reinforcing in some countries and cross-cutting in others. Therefore, this study opts for a two-dimensional conceptualisation of policy space, using both a socio-economic and socio-cultural dimension (see also Chapter 2).Thus, the issue of their alignment is not predefined by the researcher but becomes an empirical question. Moreover, the regime divide, which represents the contestation between the successor parties of the Communist regime and the oppositional forces, has constituted an overarching conflict dimension in Central and Eastern Europe. Particularly during the early years of the post-Communist transformation, this divide had the potential to overshadow other socio-cultural and socio-economic conflicts (Beichelt 2001; Grzymała-Busse 2001).

#### Basic argument

The basic argument of this study is grounded on the premise that, despite the specific features of post-Communist democracies, radical right parties and party competition in Western and Central and Eastern Europe are fundamentally, and functionally, equivalent. Thus, the study argues that the government participation of Central and Eastern European radical right parties depends on their strategic choices in pursuit of policy and office as well as the context in which they operate. More precisely, the seat share of these parties in parliament, their ideological distance from the formateur, particularly with regard to socio-cultural issues, and the configuration of the party systems, explain why they enter government or remain in opposition.

In light of the aforementioned specifics of the Central and Eastern European context, however, this basic argument needs to be qualified. The main reason for these qualifications is the development of post-Communist politics and societies during the course of the transformation process. The regime divide and the mammoth task of overseeing the triple transition severely impacted the region particularly in the first post-Communist decade. Against this background, three specifications shall be made to the basic argument concerning the early phase of the transformation:


<sup>2</sup> The opposition between competing political camps entails affective and/or ideological elements. Affective polarisation refers to "the extent to which groups dislike each other", whereas ideological or issue-based polarisation concerns "the extent to which they disagree with each other" on particular policies (Nugent 2020, 2–3; see also Iyengar et al. 2019; Wagner 2021).

such an essential part of the transformation process that governments could hardly afford major disagreement on their approach to economic policy. Hence, similar positions on socio-cultural issues alone are not sufficient for parties to form a government together during this transformational period.

These specifications of the basic argument indicate that government formation with radical right parties in Central and Eastern Europe is a complex process. Hence, this study argues that explaining this outcome requires paying close attention to the interplay of party characteristics, ideological preferences, and party system configurations.

## **1.3 Research design**

The present study is located in the sphere of medium-level concepts and middlerange theories, the classic domain of the comparative method in political science (Sartori 1970, 1040–46; Lane and Ersson 1996, 5–6; Lauth, G. Pickel, and S. Pickel 2009, 69). It involves theory-testing and theory-generating elements (Gerring 2017, 263–70). It draws on existing theoretical knowledge in the field of government formation, in part even with radical right parties, and puts these theories to a test in the context of Central and Eastern European democracies. The main goal of this study, however, is to identify different patterns—or configurations of factors—that explain why Central and Eastern European radical right parties enter government or remain in opposition. Hence, it applies a configurational approach that investigates the interaction of different explanatory factors and their impact on a specific outcome rather than testing the probabilistic effects of independent variables on dependent variables. Since fairly little is known about the interactions between the individual factors that explain the outcome of government formation with radical right parties, particularly in Central and Eastern Europe, this study also pursues a theory-generating approach when examining these complex causal relations.

Given the aim to shed light on the variety of patterns that explain why Central and Eastern European radical right parties enter government or remain in opposition, a diverse case selection will be applied. More precisely, cases were selected in order to show variation with regard to each potential explanatory factor,or even configurations of factors, as well as the outcome (Seawright and Gerring 2008; see also Gerring 2017, 89–91). For this reason, the study examines government formation in countries from various regions of Central and Eastern Europe: the Visegrad Four from Central Europe, the Baltic states of Estonia and Latvia, and Bulgaria and Romania from Southeastern Europe.<sup>3</sup> Analysing government formation in eight countries from three different parts of Central and Eastern Europe should prevent the study from generating an explanation that applies only to a specific sub-region.The analysis covers the first three decades after the fall of Communism, more precisely the period from the first free elections in each respective country in the early 1990s until the end of 2020.These criteria produce a total of 48 cases,in which radical right parties were present in parliament and thus had the chance to enter government.

The research design faces the challenge of integrating a relatively large number of cases with a configurational approach that is usually the domain of case studies (Müller, Bergman and Strøm (Müller, Bergman, and Strøm 2008, 33–35; see also Ragin 1989; King, Keohane, and Verba 1994; Brady and Collier 2004; Gerring 2017). It attempts to ease the trade-offs between small-N and large-N research (Gerring 2017, chap. 11; see also Chapter 4) with the help of Qualitative Comparative Analysis (QCA). Charles Ragin developed this method particularly for such medium-N settings with the aim of combining "the best features of the case-oriented approach with the best features of the variable-oriented approach"(Ragin 1989, 84; see also Ragin 2000, 24–26; Rihoux 2009; Schneider and Wagemann 2012). However, the reason for choosing this method is not only the number of cases, but also the fact that QCA, as a configurational case-oriented method, is particularly well suited for investigating causal complexity beyond individual case studies. Here, QCA has an advantage over statistical methods because it preserves the specific configuration of the cases throughout the analysis. Thus, the cases do not disappear behind individual variables (Ragin 1989, x; Schneider and Wagemann 2012, 76–78; Marx, Rihoux, and Ragin 2014, 120).

Integrating case-specific configurations into a structured, cross-national analysis of a medium-to-large number of cases still comes at a cost. It is almost impossible for a researcher to investigate this many cases with the analytical depth that is characteristic of comparative case studies. Therefore, in order to obtain the level of familiarity with the cases required to interpret the configurations of explanatory factors in a meaningful way, it is necessary to limit the number of conditions. The selection of these conditions will be based on existing theoretical knowledge about government formation in Central and Eastern Europe and with radical right parties across the continent. Moreover, even with regard to the limited number of explanatory factors, the present study cannot dig as deep into causal mechanisms as

<sup>3</sup> The former Yugoslav republics of Croatia and Slovenia are not included, because Tito's Yugoslavia began challenging the hegemony of the Soviet Union during the Stalinist era and developed much more independently from Moscow than either the Central and Eastern European satellite states or the Baltic Soviet republics (Rothschild 1993, chap. 3). Additionally, Croatia only gained independence in 1995 after four years of war and it entered the EU in 2013, several years later than the other member states in the region.

is possible in single, or small-N comparative, case studies which use process tracing or similar methods. The primary goal of this study is to investigate the causal relationships at play in government formation with radical right parties in Central and Eastern Europe. However, it will also address the causal mechanisms that connect the individual explanatory factors in the various explanatory patterns, though only to a limited extent.Thus, in terms of Gerring's (2017, 244) typology of trade-offs between small-N and large-N research, the research design charts a middle path between causal depth and breadth as well as between the study of causal effects and causal mechanisms.

Furthermore, the research design reflects the assumption that the patterns of government formation with radical right parties may vary over time. It divides the period under investigation into two phases, namely the time before and after the so-called "third-generation elections" (Pop-Eleches 2010). Third-generation elections are those elections that take place after parties from the two main political camps in a given country have been in power at least once. This allows both voters and political competitors to better assess their policies as well as their strategic behaviour, which is a key prerequisite for structured party competition (Savage 2016; see also Sartori 1976).The first third-generation elections in all eight countries took place around the year 2000, so this threshold, more or less, distinguishes between the first post-Communist decade and the two following decades, when Central and Eastern European party systems reached a basic level of consolidation. Moreover, democracy and market economy had also been firmly established by this time, at least on a procedural level, and all countries had begun formal accession negotiations with the EU (Beichelt 2004; Vachudova 2005).

## **1.4 Contributions and limitations**

#### Contributions

The present study contributes to the existing research on party competition in several respects. First, it provides original empirical insights into the process of government formation with radical right parties in Central and Eastern Europe. Here, the study confirms the tentative conclusions from the existing research (Fagerholm 2021) by showing that electoral results and ideological preferences of radical right parties play an important role in explaining why they enter government or remain in opposition. Yet, additional findings also explain how these factors interact with each other and with the configuration of the party system.

More precisely, the study reveals that the patterns of government formation with radical right parties in the early years of the post-Communist transformation differ significantly from those in the consolidating decades. It finds a clear transformational pattern that results from the triple transition's impact on the nascent party systems of Central and Eastern Europe. In this period, radical right parties' proximity to the formateur on both the socio-economic and socio-cultural dimensions is a necessary condition for government participation. The electoral weakness of most radical right parties at this stage is certainly no advantage, but it does not prevent them from entering government in the fragmented party systems of the region either.Moreover, the regime divide affects government formation in the period before the first third-generation elections, even if it does not divide party systems into two oppositional camps that are unable to form coalitions together.

After the turn of the millennium, the patterns of government formation with radical right parties in Central and Eastern Europe increasingly resemble those in the western part of the continent (Bale 2003; de Lange 2008, 2012). The existence of a deeply polarised opposition in the party system, originating mostly from conflicts other than the regime divide, and the ideological distance from the formateur on the socio-cultural dimension, become the most important factors for explaining government formation. The importance of the radical right's seat share in parliament also increases in this period. However, the high degree of fragmentation that still exists in many Central and Eastern European party systems continues to help electorally weak radical right parties to gain access to power.

Although this study primarily focuses on radical right parties,it also contributes to the research on government formation more broadly. In addition to generating new empirical insights, it advances the discussion about concepts and methodological approaches within this discipline.Mostimportantly,it confirms that a time- and context-sensitive approach contributes analytical value to the study of party competition in Central and Eastern Europe (Ekiert and Hanson 2003b). The results show that there are indeed substantial differences in the explanatory patterns of government formation with radical right parties before and after the first third-generation elections. This aspect should be relevant for other areas of comparative politics and social science research in Central and Eastern Europe and beyond. Western European democracies, for instance, have also witnessed dramatic contextual changes during the past decades, such as the post-materialist value change since the late 1960s, the fall of the Iron Curtain in 1989/90 or the economic crisis in the late noughties, which could also prove to be pivotal turning points upon closer examination (Inglehart 1977; Hernández and Kriesi 2016).

Moreover, the findings present insights into the connection between the fragmentation of the party system and the seat share of radical right parties in parliament as well as the interaction between their socio-economic and socio-cultural positions. These findings improve our knowledge about the strategic decisions of radical right parties and their competitors during party competition unrelated to government formation and, thus, about the impact of the radical right on politics and society in Central and Eastern Europe. The study also emphasises the importance of causal complexity and a configurational approach. In most cases, the outcome of government formation can only be explained when focusing on the interaction of different explanatory factors. QCA has proven to be a useful tool for this type of investigation.

### Limitations

Like all comparative empirical research, this project has certain limitations. First, as a study with a medium number of cases, it sacrifices some analytical depth in order to identify cross-national explanatory patterns. At the same time, however, the casebased research design and the regional focus set limits on the generalisability of the results.Therefore, the present study aims only to attaining internal validity (Gerring 2017, 232, chaps. 9–10).However,in light of the presumed functional equivalence betweenWestern and Central and Eastern Europe, the results of this study should offer some theoretical insights to scholars interested in studying government formation with radical right parties in Western Europe as well.

Second, the study works with a theory-based analytical model and aims to explain the participation of radical right parties in government with the help of the conditions specified therein. Even though country-specific case studies make up a sizeable part of the study, they remain purely descriptive.They serve primarily to introduce the cases and to gather the data required for the comparative analysis.Thus, the country case studies do not attempt to inductively identify further explanatory factors or to comprehensively illuminate the causal mechanisms behind the formation of governments with radical right parties in individual countries.

Third, the study shares a common feature with the vast majority of academic research on the formation of government coalitions in that it essentially views coalition negotiations as a black box. It does not attempt to trace processes that take place largely behind closed doors. This task can only be completed using individual case studies (e.g. Müller 1999) or large-scale international research projects that have the necessary resources for a large number of qualitative interviews with highlevel politicians in numerous countries. This project simply lacks the resources for such an undertaking.

Fourth and finally, the study deals exclusively with the formation of governments. It neither examines the cooperation between coalition parties while they are in office nor does it investigate the direct and indirect effects that radical right participation in government has on politics, society, or the radical right parties themselves. However, by explaining the formation of governments with radical right parties, it contributes to a better understanding of the processes that lead to these effects.

#### **1.5 Outline of the study**

As a first step, the next two chapters present the theoretical and conceptual frameworks used in this study. Chapter 2 provides a definition of radical right parties and discusses the characteristics of this party family in Central and Eastern Europe. It goes on to address the development of the post-Communist party systems in order to introduce the context in which government formation with radical right parties takes place. The chapter also discusses similarities and differences between radical right parties and party systems in Central and Eastern and Western Europe, thus identifying where context-specific modifications are necessary. Chapter 3 is dedicated to theories of coalition formation. It provides an overview of the most important theories found in the discipline and assesses whether these have received empirical support in Western and in Central and Eastern Europe.

The research design and methodology are the subject of Chapter 4. This chapter starts by outlining the general features of the research designs foundin the literature on government formation and introducing QCA as the principal research method of this study. It then identifies the most promising explanatory factors based on the discussions found in Chapters 2 and 3 and combines them into an analytical model. As a last step, the chapter discusses the operationalisation of the individual components of the analytical model.

The empirical section begins with the descriptive country case studies in Chapters 5 and 6, which follow a uniform structure based on the factors specified in the analytical model. Chapter 5 covers Central Europe, and Chapter 6 discusses the Baltic and Southeastern European countries. Chapter 7 summarises the data and carries out the calibration of set membership. This procedure generates a uniform dataset from the empirical data presented in the two previous chapters, which is necessary to prepare the data for analysis with QCA.

On the basis of this dataset, Chapters 8 and 9 provide a comparative analysis of government formation with radical right parties. Chapter 8 covers the period before the first third-generation elections, and Chapter 9 the two consolidating decades. As is usual in QCA, this analysis aims to identify necessary and sufficient conditions for the participation of radical right parties in government. In line with good practice in QCA (Schneider and Wagemann 2010), the negative outcome—the exclusion of radical right parties from government—is examined separately. In conclusion, Chapter 10 summarises the results and compares the patterns found in the two periods under investigation. It then discusses the implications of these findings for studying radical right parties and party competition.

## **2. Radical right parties in the post-Communist party systems of Central and Eastern Europe**

After the fall of Communism, it took about a decade before comparative research on right-wing radicalism started to turn its attention to Central and Eastern Europe(Ramet 1999;Mudde 2000a,2005b; Beichelt andMinkenberg 2002;Minkenberg 2002). Since then, scholars have discussed whether or not radical right parties in post-Communist Central and Eastern Europe constitute a phenomenon *sui generis*. This discussion also touches upon the issue of whether or not these parties are comparable to their Western European brethren, and further, whether or not they can be studied using the theoretical and methodological toolkit developed by scholars of radical right parties in Western Europe (Minkenberg 2002, 35, 2017; Mudde 2007, 3–4; Pirro 2016; Pytlas 2018). Since these issues are also relevant in the context of this study, this chapter examines the characteristics of radical right parties in Central and Eastern Europe and the region's party systems.

### **2.1 The Central and Eastern European radical right**

Defining radical right parties remains one of the most contested issues in research on the radical right. Without indulging in extensive taxonomic debates, this section presents a working definition that is suitable for comparative research on radical right parties and discusses the main features of this party family in Central and Eastern Europe.

The terminology in the existing literature can be separated into rather specific concepts, such as neo-Nazism or ethno-nationalism, and into broader ones, such as the far, radical, or extreme right. Despite ongoing debates about labels, most research on radical right parties ends up studying a very similar set of usual suspects (for an overview, see Mudde 2007, 11–13; Minkenberg 1998, 29–31). Nevertheless, the choice of a particular terminology and definition leads to specific theoretical and conceptual ramifications. Given the centrality of the parties' fierce opposition to immigration, some research on Western Europe, for instance, has defined the party family as "anti-immigrant" (van der Brug, Fennema, and Tillie 2005). Such a definition, however, implies that these are single-issue parties which makes it difficult to transfer findings to Central and Eastern Europe, where immigration was hardly on the radical right's agenda before the so-called migration crisis in 2015 (Minkenberg 2002, 346, 2017, 48–49; Mudde 2007, 19). Others have done quite the opposite and defined the radical right using a bundle of different issues (Mudde 2000b; Heitmeyer 2002; O. Decker, Brähler, and Geißler 2006).While offering a very precise description, such "shopping list" definitions can also be over-specific and thus limit the scope to particular temporal or spatial contexts (Minkenberg 1998; Pytlas 2016, 24). It is quite striking, for instance, that gender issues have rarely been discussed in connection with the ideological platform of the radical right until recently. In the last decade, however, authors have concluded that issues of gender identity, reproductive rights, homo- and transphobia, or even women's rights with a distinct anti-Muslim twist, have become key issues in radical right mobilisation (Kováts and Põim 2015; Akkerman 2015).The above-mentioned example of anti-immigrant parties also alludes to the spatial limitations, since this concept would not have addressed the ideological core of Central and Eastern European radical right parties until recently. Several studies, particularly in the German context, also highlight the glorification of National Socialism as a key feature of the radical right ideology (O. Decker, Kiess, and Brähler 2012, 18), but members of this party family in other countries would credibly reject this claim.

A comparative analysis of radical right parties in several countries and over a period of 30 years warrants a definition that captures the broader ideological core of the party family. Therefore, the present study follows Pytlas (2016, 25) and defines right-wing radicalism as an ideology based on "mythicized nativist ultra-nationalism". The focal point of this ideology is the mythicized image of a homogenous nation, a naturalistic *Volksgemeinschaft*, which is constructed by combining different criteria of inclusion and exclusion, such as race, ethnicity, or religion that can vary over time and between "nations" (see also Minkenberg 1998, 33; Mudde 2007, 19). Nativism adds the notion that "states should be inhabited exclusively by members of the native group ('the nation') and that non-native elements, (persons and ideas) are fundamentally threatening to the homogenous nation-state", which underlines the exclusionary nature of right-wing radicalism and sets it apart from liberalising nationalisms of minorities (Mudde 2007, 19).

The ideology of the radical right bears a strong exclusionary thrust and is therefore always directed against the values and principles of liberal democracies (Minkenberg 1998, 34; Pytlas 2016, 25; see also Mudde 2007). Political parties that adhere to this ideology, however, cannot be automatically conceived as anti-system parties in the sense that they aim to overthrow the liberal democratic order as such. There are nonetheless members of the radical right party family that pursue precisely this goal. Such parties constitute a distinct sub-group of the radical right and will be termed extreme right. In contrast to other authors who perceive opposition to the

democratic system as the ideological core of extreme right parties (Backes and Jesse 1996; Ignazi 2003; Carter 2005), the definition applied in this study considers their anti-systemness to be a secondary ideological feature. In this vein, the present study understands right-wing radicalism—and extremism—not as a fringe phenomenon, disconnected from a supposed democratic centre. Rather, "key aspects of the populist radical right ideology are shared by the mainstream, both at the elite and mass level, albeit often in a more moderate form" (Mudde 2010, 1178; see also Minkenberg 1998, 34–35; O. Decker, Brähler, and Geißler 2006, 12; Pytlas 2016, 7).

The ideological core of nativist ultra-nationalism is shared by radical right parties across Europe. However, there are two developments in Central and Eastern Europe which created a favourable environment for radical right mobilisation and set the region apart from Western Europe. First, there is the legacy of unfinished stateand nation-building which, in turn, led to the high salience of nationalism in post-Communist Europe (Beichelt and Minkenberg 2002; Pirro 2016; Pytlas 2016; Minkenberg 2017, chap. 3.2). The idea of the nation took root in Central and Eastern Europe when the region was ruled by multinational empires. Hence, nationalism started out as anti-imperialist independence movement which emphasised ethnic, cultural, linguistic, or (mythical) historical events to define the common identity of the nation (Schöpflin 1996; Bunce 2005). State- and nation-building always includes elements of ethnic and political nationalisms, both of which have been present in Central and Eastern as well as Western Europe (Shulman 2002; Blokker 2005; Pytlas 2016). However, as a result of the region's imperial history, ethnic and (religio-) cultural elements outweigh political ones when defining the boundaries of Central and Eastern European nations (Bunce 2005, 422–24; Grzymała-Busse 2015; Pytlas 2016, 50–55; Minkenberg 2017, 45).

Brubaker (1996, 4–6, chap. 3) models the specific nationalisms that emerged in the new Central and Eastern European states after World War I as a triadic nexus of mutually reinforcing interactions between nationalising states, national minorities, and external homelands. National minorities in many of the new states sought to fulfil the promise of statehood, or at least greater autonomy. As a reaction to such minority nationalist stances, the "core nation" continued its nationalist project by reinforcing the unity between national identity and state borders (see also Bunce 2005). In addition, most national minorities, such as Germans or Hungarians in Czechoslovakia, possessed external homelandsin theimmediate vicinity which supported their claims on the basis of shared national identity. In turn, the potential threat from these kin states also served as an integrative element in the national identity of the core nation. In this vein, nationalism continued to play a major role in the mostly short-lived attempts of democratic statehood in Central and Eastern Europe. In fact, these nationalist dynamics are often deemed largely responsible for these states' return to right-wing authoritarian rule in the 1920s and 1930s (Hobsbawm 1995, chap. 5; Minkenberg 2017, 45–47).

The idea of the nation continued to play a role under Communist rule, despite the strong internationalist thrust in the Communist ideology (Brubaker 1996, chap. 2; von Beyme 1996, chap. 3; Bunce 2005). The Soviet Union's recipe to deal with its multinational population was the creation of an ethno-federation,in which the subunits were structured along ethno-cultural lines and formally enjoyed a high degree of autonomy. Effectively, however, Moscow sought tight control over the republics, for instance by sponsoring and co-opting national elites, supporting the development of nationally defined institutions, and catering to the socio-economic needs of the population. By doing so, the regime successfully penetrated these territories, hoping that citizens' identification with the Communist regime and ideology would eventually substitute national identity. Moscow also extended the ideas of ethnofederalism, such as securing hegemony through sponsoring and control of national elites, to the independent states of Central and Eastern Europe within the Communist bloc (Bunce 2005, 426–27). While this strategy aimed ultimately at eliminating nationalism from the Communist sphere, the regimes in some satellite states, such as Romania or Bulgaria, actively invoked nationalism as a tool to secure power in light of growing discontent, thus developing a specific ideology of national Communism (Ishiyama 1998).

In the post-Communist era, the issue of nation-building and national identity resurfaced with full force (von Beyme 1996, chap. 3; Elster et al. 2000; Bunce 2005, 441–43) and rendered the ideological core of radical right ideology highly salient. Depending on the specific national context, different facets of nationalism gained prominence in the Central and Eastern European countries.They comprised, for instance, hostility towards ethnic minorities, including Roma; irredentist claims that were previously silenced by Moscow's hegemonic power; clericalism and ethno-religious nationalism; a strong anti-Communist or, in countries with a national Communist history, even a pro-Communist thrust (Minkenberg 2002, 2015a, 2017; Buštíková 2015, 2018; Pirro 2016; Pytlas 2018). The immigration issue, in contrast, has not played a role in radical right mobilisation until the so-called migration crisis in the mid-2010s.

Another difference between Western and Central and Eastern Europe concerns processes of modernisation in society. The (new) radical right that has emerged in Western Europe since the 1980s can be characterised as a counter-movement to rapid modernisation, and more precisely to the post-industrial transformation and the post-materialist value change, which occurred during the 1960s and 1970s. This "silent revolution" resulted in the rise of progressive left-libertarian actors, often labelled as green parties (Inglehart 1977). The nativist ultra-nationalist ideology of the radical right, focusing on the issues of immigration and law and order, appealed to those voters who struggled with the growing liberalisation and individualisation of the economic, political, and cultural spheres resulting from these modernisation processes. Therefore, Ignazi (1992) refers to the emergence of the radical right

in Western Europe as a "silent counter-revolution", which created a new conflict dimension in Western European politics and societies which initially cut across existing cleavage structures (see also Betz 1994; Minkenberg 1998; Ignazi 2003; Pirro 2016).

Central and Eastern European radical right parties came to life under quite different circumstances. After the fall of the Iron Curtain, post-Communist societies were confronted with the mammoth task of (re-) building a new economic and political order in new—or at least newly independent—nation states, a challenge often referred to as the "dilemma of simultaneity" (Offe 1991, 872). Compared to other post-Communist regions, the countries of Central and Eastern Europe managed these complex and far-reaching tasks much better than expected and established functioning democracies and capitalist economies rather quickly (von Beyme 1996; Elster et al. 2000). Yet, at the same time, the region experienced a dramatic economic decline in the first half of the 1990s, which was even worse than the Great Depression of 1929 (Merkel 2010, 329–39), and resulted in massive economic hardships as well as losses to social status and economic security among large parts of the population (Minkenberg 2017, 13–14). Thus, the emergence of Central and Eastern European radical right parties took place in a context of rapid modernisation as well, even though the causes of modernisation were different from those in Western Europe. Moreover, the modernisation shifts after the fall of Communism have been even "more far-reaching, deeper and complex than in the West" a generation before (Beichelt and Minkenberg 2002, 5–6; see also Pytlas 2016; Minkenberg 2017).

In such an environment, radical right parties clearly appealed to those who experienced economic hardships and status insecurity by providing an alternative course to political and economic liberalisation. It would be short-sighted to credit the support for radical right parties to socio-economic grievances alone, however. The supporters of the radical right also embrace the particular concept of national identity that these parties convey (Pytlas 2016, 5–7). In other words, they can be better characterised as "axiological modernization losers" who "perceive the post-communist state- and nation-building path as a threat to the integrity, values, and interests of the radicalized interpretation of 'the nation'" (Pytlas 2016, 7). Under these circumstances, an opposition to post-material values, and the parties that represent them, hardly affected the emergence of radical right parties in Central and Eastern Europe (Pirro 2016, 36).

In light of the economic hardships during the post-Communist transformation, it is hardly surprising that socio-economic issues feature quite prominently in the ideological platforms of many Central and Eastern European radical right parties—mostly in the form of left-leaning positions combined with a strong nationalist element. This policy of "social-national economics" (Pirro 2016, 41) generally accepts the framework of free-market economy but advocates for national protectionism and social security provided only to those who are considered members of the nation. In this vein, the socio-economic positions of the radical right are strongly linked to the socio-cultural core of their ideology (Pirro 2016; Buštíková 2018; see also Łapiński 2004; Mudde 2007; Minkenberg and Pytlas 2013; Pirro 2017).

The discussion has shown that an exclusionary, nativist ultra-nationalism constitutes the overarching ideological core of the radical right. However, the manifestations of right-wing radicalism, and the people who radical right parties consider members of the in-group, or the out-group, can differ between countries and regions as well as over time. The immediate adoption of the immigration issue and anti-Muslim racism in Central and Eastern Europe in the mid-2010s, or the increasing importance of anti-LGBTIQ+ mobilisation for the radical right across Europe, underlines that these parties are able to adapt to changing contextual conditions (Pytlas 2018).

Particularly in the context of Central and Eastern Europe,it is important to highlight the distinction between radical right parties and radical right politics (Mudde 2018, 261; see also Pytlas 2018). Mainstream parties have adopted various elements of radical right politics, not least due to the salience of nationalism in the region. Some scholars even speak of a radicalisation of the mainstream in Central and Eastern Europe (Minkenberg 2013; see also Pytlas and Kossack 2015; Pytlas 2016).

In some cases, the boundary separating mainstream parties that use radical right politics from radical right parties is becoming increasingly blurred. Since the 2010s, for instance, the Hungarian Alliance of Young Democrats (Fidesz) and the Polish Law and Justice (PiS) have embraced radical right politics to such an extent that some scholars include them in the radical right party family. In the mid-2010s, Minkenberg (2017, 2, 24) mentions that both parties have drifted toward the radical right, but still labels Fidesz as a "right-wing populist party" and PiS as a "nationalconservative" one. Pytlas (2016) conceives of PiS and Fidesz as "nearby competitors" of the radical right, but he also shows that both parties increasingly apply radical right frames and thus gravitate towards the ideology of *völkisch* nationalism (see also Sata and Karolewski 2020, 12–14; Markowski 2020, 1516). A few years later, Mudde argues that "[a]fter regaining power in 2010, Orbán quickly transformed Hungary into an illiberal democracy (or even a competitive authoritarian regime) and Fidesz into a far-right party" (Mudde 2020, 302). In another recent article, he includes PiS and Fidesz into the radical right party family and describes them as "transformed conservative parties" (Mudde 2019, 32). Hence, in both cases, the notion of conservative parties that underwent a gradual transformation into radical right ones seems plausible.

Because of the gradual nature of this transformation, however, it is difficult to pinpoint exactly when PiS and Fidesz ultimately joined their new party family, if this is possible at all. In the context of the present study, however, it is necessary to determine whether or not these parties belong to the radical right party family during each instance of government formation. Given that this project lacks the resources to carry out a detailed analysis of the transformation of these parties, and because Fidesz and PiS did not belong to the radical right party family for most of the period under study, the government participation of these—transformed radical right—parties in the 2010s is not included in the analysis. Moreover, this study analyses government formation in minority situations in parliament. It argues that, under these circumstances, the dynamics of coalition bargaining fundamentally differ from situations where one party controls an absolute majority (see Chapter 4.4). Since Fidesz and PiS have constantly won more than half of the seats in parliament during the time scholars consider them to be transformed radical right parties, the majority status also justifies their exclusion from the analysis. However, in the late 1990s and mid-2000s, before their transformation, Fidesz and PiS appear in the analysis as formateurs of coalitions while radical right parties were present in parliament.

The Latvian parties Latvian National Independence Movement (LNNK) and For Fatherland and Freedom (TB), and the Estonian National Independence Party (ERSP) experienced a reverse transformation. In the former Soviet republics, nation-building took place in the presence of a large Russian-speaking minority. During the first years of independence, political elites debated how to treat this minority, particularly with regard to citizenship. Although nativism was an essential part of LNNK's, TB's and ERSP's agenda, the family affiliation of these parties remains controversial (Pettai and Kreuzer 1998; Muižnieks 2005; Poleschuk 2005; Mudde 2007; Bennich-Björkman and Johansson 2012). Empirical research on the ideology of Central and Eastern European parties in the early 1990s is scarce (Mudde 2007, 53), and the Baltic states received even less attention than other regions in Central and Eastern Europe (Auers and Kasekamp 2009, 242; Mudde 2018, 260), so determining the party family affiliation for these parties is difficult.

In Latvia, the LNNK and TB began to moderate their nativist positions in the mid-1990s (Dehmel and Reetz 2011, 217; Bennich-Björkman and Johansson 2012). Both parties and their alliance, For Fatherland and Freedom/Latvian National Independence Movement (TB/LNNK), are sometimes included in the radical right party family, particularly in the first half of the 1990s (Auers and Kasekamp 2015; Minkenberg 2017, 72). Other research, however, characterises them as nationalist or (national) conservative (Muižnieks 2005, 120; Bennich-Björkman and Johansson 2012), and sometimes the classification even varies within a single study (Dehmel and Reetz 2011). Due to the lack of unambiguous classification and the shortage of empirical research on the ideology of the LNNK and TB, both parties and their alliance, TB/LNNK, are excluded from this study.<sup>1</sup> The literature agrees, however,

<sup>1</sup> The People's Movement for Latvia (TKL), a flash party founded by former LNNK member Joachim Siegerist and elected to parliament in 1995, definitely qualifies as a radical right party (Muižnieks 2005, 103–4; Dehmel and Reetz 2011, 117; Mudde 2018, 256). However, empirical

that the Estonian ERSP can be considered a radical right party until its merger with the more moderate Pro Patria in 1995. Pro Patria and, in particular, the party which resulted from the merger of these two parties, the Pro Patria Union, were never radical right (Kasekamp 2003, 404; Poleschuk 2005, 60; Mudde 2007, 143; Reetz and Thieme 2011, 103; Bennich-Björkman and Johansson 2012). The present study follows this assessment and conceives of the ERSP as a radical right party until 1995.

Table 2.1 provides an overview of radical right parties in Central and Eastern Europe. It includes only those parties that passed the threshold of parliamentary representation at least once, because representation in parliament is a vital precondition for participation in coalition formation. The table illustrates that some parties, like the Hungarian Justice and Life Party (MIÉP) in 1998, were only present in parliament for a single term, while others, such as the Slovak National Party (SNS), celebrated repeated success at the polls. Moreover, several countries witnessed more than one radical right party in their national parliament and some legislatures even included two radical right parties at a time, for instance the Romanian parliament between 1992 and 1996.


*Table 2.1: Radical right parties in Central and Eastern Europe and their presence in parliament since the first free elections*

information on this short-lived party, in particular its ideological positions, is so scarce that it cannot be included in this study.


Source: Own compilation; parties in parentheses are not included in this study.

\* Electoral alliance of VMRO, the NFSB and several small parties and organisations.

\*\* Electoral alliance of the PF and Ataka.

## **2.2 Central and Eastern European party systems: The context of coalition politics**

Having discussed Central and Eastern European radical right parties, this section now turns to the party systems in which they compete and interact with other parties. Sartori (1976, 44) defines a party system as "the system of interactions resulting from inter-party competition". With regard to Central and Eastern Europe, Savage underlines the importance of party systems as a context for government formation, while also highlighting how their fluidity causes problems in the region: "Party systems provide the essential structure of the coalition-bargaining environment, as they contain information on the parties' relative bargaining weights and preferences. Each party in the system uses this information when making decisions on potential coalition partners. What distinguishes party systems in new democracies from those of established democracies is the lack of routinized interactions between parties, which brings a higher level of uncertainty" (Savage 2016, 503–4).

The literature assesses the institutionalisation and stabilisation of post-Communist party systems quite differently. In a brief summary of the academic debate, Thorlakson (2018) shows that the arguments of proponents and critics of party system stabilisation do not necessarily contradict each other, but that the disparities often result from emphasising different elements of the party systems. Those who stress party system fluidity often highlight structural features, such as volatility and the emergence of new parties, while advocates of stabilisation focus rather on the content of party competition.The following discussion outlines the main arguments in this debate and relates them to the present study.The discussion follows Sartori's (1976) classic distinction between the structural-numerical and the ideological configuration of party systems.

## 2.2.1 The structural stabilisation of party systems in Central and Eastern Europe

The political developments in Central and Eastern Europe since the inter-war era had an impact on the structure of the emerging party systems in the region after 1989. The region's inter-war democracies were quickly toppled by either domestic authoritarian forces, or external political powers, such as the Nazi regime that occupied Czechoslovakia in 1938/9 (Elster et al. 2000, 37–38; Minkenberg 2017, 45–46). Thus, there was little time for democratic parties to take root in society. In the early days of Communist rule after World War II, the independent, Central and Eastern European countries adopted constitutions that closely mirrored the Soviet model, including one-party rule and tight control over every sector of society. Even though the individual regimes certainly developed their own specific traits after Stalin's death and several uprisings in the 1950s and 1960s, for instance with regard to domestic, foreign, or economic politics, none of the Communist parties in the Central and Eastern European satellite states risked their hegemony by tolerating party competition (Rothschild 1993). Even in countries where bloc parties existed, they were never truly independent from the Communists and posed no challenge to the ruling elite (Cabada, Hloušek, and Jurek 2014, 48).

With two notable exceptions, most parties that emerged in Central and Eastern Europe after 1989 were newly created political entities without any roots in historical party politics (von Beyme 1996, 127–29; Cabada, Hloušek, and Jurek 2014, 53). First, some leaders attempted to revive historical parties from the inter-war era, though most were unsuccessful (Cabada, Hloušek, and Jurek 2014, 44–47, 182). Second, and more importantly, the Communist parties themselves continued as more or less reformed political forces in the new party systems. The organisational continuity that equipped the Communist successor parties with substantial personal and financial resources, gave these parties an advantage over their newly founded competitors (von Beyme 1996, 133–35). Moreover, even in countries that witnessed a strong and well-organised political opposition, for example the Polish Solidarność, these movements did not necessarily transform into equally strong and well-organised political parties after 1989 (Ekiert and Kubik 1999). Hence, new parties played a major role in party system formation after 1989, but party competition did not entirely take place in a "tabula rasa" situation (Elster et al. 2000, 131).

Due to the large number of new parties, Central and Eastern European party systems were weakly structured and characterised by a high degree of uncertainty, particularly in the early phase of the post-Communist transformation. The opposition to the Communist regime mostly established broad umbrella organisations, so-called forum or movement parties, which claimed to act in the national interest rather than as advocates for particular strata or groups of society. In organisational terms, they were characterised by horizontal structures and blurry borders between the party and civil society.The opposition's distaste for vertical power structures and their (over-) emphasis on civil society is hardly surprising given their experience under four decades of authoritarian one-party rule. Despite the anti-elitist appeal of many of these parties, elites and individual personalities played an important role in their development from the very beginning (Ágh 1998, 102–4; see also Geddes 1995; von Beyme 1996).

Moreover, the early design of post-Communist democracies favoured the influence of political parties over other interest groups, providing a strong incentive for political entrepreneurs to form parties, which often resembled small elite organisations without formal organisational structures.Though for different reasons, power was concentrated among individual personalities and elites in the forum and the entrepreneurial parties as well. The new members of this political class consolidated and extended their power after the first elections, when they entered parliament and government (Ágh 1998, 104–8).The dominance of political elites in weakly organised political parties resulted in an "overparticization" and "parliamentarization" (Ágh 1998, 105) of Central and Eastern European party systemsin the first half of the 1990s. However, Ágh (1998, 109–12) identifies two stabilising trends which followed the dissolution of the forum parties and the institutional learning gained from the West. First, these processes contributed to horizontal differentiation in the party system, because more distinctive parties emerged from the catch-all forum parties, leaving room for a broader spectrum of political elites. Second, a growing vertical differentiation could be observed, since interest groups and civil society became more independent from political parties (see also Enyedi and Casal Bértoa 2018, 434–35) Thus, by the end of the 1990s, a professional political class had emerged and the functional differentiation improved parties' ability to represent societal interests.

Other observers were less optimistic about the prospects for party system institutionalisation at that time (Mair 1997; Elster et al. 2000). In one of the most recent and comprehensive empirical accounts of party system institutionalisation<sup>2</sup> and stability in Central and Eastern Europe, Enyedi and Casal Bértoa (2018) illustrate that the initial scepticism was not completely unwarranted. Low levels of party membership, for instance, remain a characteristic feature in post-Communist parties. In fact, the average share of party members in the electorate is even decreasing in Central and Eastern Europe. Such downward trend is also visible in Western Eu-

<sup>2</sup> The concept of party system institutionalisation was mainly developed and advanced by Mainwaring to explain democratic transformations in Latin America (e.g. Mainwaring and Scully 1995; Mainwaring 2018).

rope, but these states started from significantly higher levels of party membership (see also Cabada, Hloušek, and Jurek 2014, 121–23; Minkenberg 2017, 57–58).

The patterns of government formation have become more stable in Central and Eastern Europe since the 1990s, even though the level of party system closure still remains below those in the established democracies of Western Europe.<sup>3</sup> Furthermore, the decreasing intensity of fragmentation also points to a stabilisation of Central and Eastern European party systems. In fact, fragmentation, reflected in the effective number of parties that compete in elections or enter parliament, is the only indicator of party system stability examined by Enyedi and Casal Bértoa (2018, 440) which shows no statistically significant difference between East and West. The authors even find the average effective number of parliamentary parties to be exactly the same in both parts of the continent in the period since 2010. The structural feature which sets post-Communist party systems apart from their West European counterparts most clearly, however, is their high degree of volatility, reflecting the gains and losses of parties in an election compared to the previous one. The average volatility in post-Communist democracies is twice as high as in Western Europe between 1990 and 2016 (24 per cent versus 12 per cent) and four out of five elections result in a change of more than 15 per cent of the votes between competing parties (Enyedi and Casal Bértoa 2018, 435–37).

Much of the volatility in post-Communist party systems results from the frequent emergence of new parties, which often make remarkable electoral gains. Powell and Tucker (2014, 131) find that more than 70 per cent of the electoral volatility between 1989 and 2009 originates from new parties (see also Tavits 2008a).<sup>4</sup> The continuous rise and fall of new parties has prompted scholars to speak of a new party sub-system, in which "multiple parties shar[e] a common and distinct pool of ideas, voters, and elites" (Haughton and Deegan-Krause 2015, 69). Therefore, the individual new parties that emerge in these party systems should not be regarded as completely isolated phenomena. Moreover, Sikk (2005) points out that these parties are not always true newcomers to the political scene. He rather argues that many of the region's allegedly new parties are the offspring of existing political circles or parties and only few are "genuinely new". Genuinely new parties are those that are not "successors to any previous parliamentary parties, have a novel name and structure, and

<sup>3</sup> The concept of party system closure reflects the stability of the patterns of government formation, based on the alternation in government, the familiarity of the format of governments, and the parties' access to power (Enyedi and Casal Bértoa 2018, 426; see also Mair 1997).

<sup>4</sup> Tavits (2008a) makes an important theoretical contribution when investigating the causal relationship between electoral volatility and the emergence of new parties. While the literature suggests causal effects between these two factors in both directions, she illustrates that the emergence of new parties, including splits and mergers, is a cause of electoral volatility in Central and Eastern Europe and not vice versa.

do not have any important figures from past democratic politics among their major members" (Sikk 2005, 399). Such legitimate newcomers, however, are less successful than the high levels of volatility suggest, since their gains account for only about 20 per cent of the overall volatility in the first decade of democratic rule in Central and Eastern Europe. Many genuinely new parties did not even manage to enter parliament, and those that did often disappeared as quickly as they emerged (Sikk 2005, 402–6). Later studies which follow Sikk's (2005) coding approach for defining new parties corroborate his conclusion. Emanuele, Chiaramonte, and Soare's (2020) results show that less than a third of the total volatility in Central and Eastern Europe between 1990 and 2016 can be credited to genuinely new parties.<sup>5</sup> Moreover, they find that electoral changes caused by new parties are somewhat lower in the 2010s, when compared to the previous two decades, while volatility resulting from shifts between existing parties has increased during this period. Based on these findings, they conclude that some "'core' parties of the system have finally succeeded in creating (more) stable and enduring loyalties with their voters" while new parties remain a relatively frequent phenomenon in the region (Emanuele, Chiaramonte, and Soare 2020, 317).

Overall, Central and Eastern European party systems have undergone a process of consolidation over the last three decades, even though many indicators of their institutionalisation do not match the scores of established party systems in Western Europe. This is hardly surprising, however, given that they are much younger. Existing patterns of convergence between both regions, for instance regarding fragmentation and volatility, are not only a result of Central and Eastern Europe catching up with theWest. In particular, the converging levels of volatility also result from the steep increase of volatility in Western Europe after the economic crisis of 2008/9 (Emanuele,Chiaramonte,and Soare 2020).Hence, even scholars who aremore sceptical about the structural stabilisation of Central and Eastern European party systems agree that they provide a sufficiently stable context for the application of "Western" concepts and theories (Cabada, Hloušek, and Jurek 2014, 185–186, 189). Where coalition politics are concerned, these scholars also claim that political parties have been "key playersin government formation"in Central and Eastern Europe ever since "the very first months and years" of the transformation when only the "torso of the party-political structure" existed (Cabada, Hloušek, and Jurek 2014, 151).

<sup>5</sup> The results also diverge from previous studies since these authors try to avoid biased case selection. First, a temporal bias emerges from the frequent comparison of Central and Eastern European elections since 1990 with the whole post-war period in Western Europe, which can obscure similar trends in both regions in the same period. Second, Western European patterns are often compared to a broader sample of post-Communist countries (Powell and Tucker 2014; Enyedi and Casal Bértoa 2018), which includes post-Soviet or Balkan countries that differ significantly from the post-Communist EU member states in terms of democratic consolidation (Emanuele, Chiaramonte, and Soare 2020, 312–13).

## 2.2.2 Political divides in Central and Eastern European party systems

The ideological configuration of Central and Eastern European party systems shows a higher degree of stability, even though it differs significantly from the established Western European party systems (Bakke and Sitter 2005; Enyedi 2008; Rohrschneider and Whitefield 2009). In their seminal work on cleavage structures in Western European party systems, Lipset and Rokkan (1967) show how party competition resulted from cleavages between societal groups and their collective interests, which political parties then articulated. They identified the conflicts between labour and capital, urban and rural interests, centre and periphery, and state and church as the primary, structuring elements of Western European politics (Lipset and Rokkan 1967, 14).

While cleavage theory has become one of the most prominent approaches to studying party systems in democratic countries, some scholars have cast serious doubts about its applicability to the post-Communist party systems of Central and Eastern Europe.They argue that the region's political parties lack programmatic coherence and a solid grounding in societal conflicts (Elster et al. 2000; Innes 2002). Other research, however, provides evidence for the emergence of predictable issuebased party competition along relevant societal conflicts soon after the fall of the Iron Curtain (Kitschelt 1995; von Beyme 1996; Kitschelt et al. 1999; Beichelt 2001; Marks et al. 2006; Rohrschneider and Whitefield 2009). Returning to the concept of cleavages allows for some reconciliation between these conflicting positions. Bartolini and Mair (1990, 214–16) characterise a cleavage as having an "empirical", a "normative", and an "organisational" element. Hence, in order to speak of a cleavage there must be a socio-structural division in society, which is reflected in the selfconsciousness of these groups, and which results in political organisation and mobilisation.The authors also state that other terms are needed for situations in which all three elements of a cleavage are not present. Deegan-Krause (2007, 539–40) suggests the term "difference" when only one of the elements is present. A "divide" describes the simultaneous presence of two elements but not a "full cleavage", where all three elements are present. In this vein, he proposes the term "position divides" for situations that combine structural and attitudinal differences, "census divides" when structural and organisational elements are present at the same time, and "issue divides" when a conflict merges an attitudinal basis with organisational representation but lacks demographic roots.

Based on this analytical distinction, Deegan-Krause (2007) illustrates that full cleavages are indeed rare in Central and Eastern Europe, because political competition has shallow roots in the demographic structure of society.<sup>6</sup> When settling for

<sup>6</sup> Deegan Krause (2007, 543) argues that Western European democracies have evolved in a similar direction. The erosion of class or religious identity, and the development of new con-

the notion of issue divides, however, there is ample evidence for a linkage between the conflicts in society and programmatic positions of political parties in Central and Eastern Europe (Kitschelt 1995; von Beyme 1996; Kitschelt et al. 1999; Beichelt 2001; Marks et al. 2006; Rohrschneider and Whitefield 2009). Yet, the question remains, which issue divides structure party competition in the region and how different divides relate to each other.

Klaus von Beyme (1996, chap. 7) attempts to capture the conflict structure of post-Communist party systems by supplementing Lipset and Rokkan's traditional cleavages with four new conflict dimensions.<sup>7</sup> He adds, however, that some of them overlap and not all are relevant across the entirety of Central and Eastern Europe.The cleavage between labour and capital, for instance, did not fully materialise in post-Communist Europe, because the egalitarian politics of the Communist regimes prevented a capitalist, bourgeois elite from emerging and entering party politics. On the opposite end and as a consequence of a wide-spread, anti-socialist bias in the region, social democratic parties remained underdeveloped.<sup>8</sup> Beichelt (2001, 182–90) finds that five salient divides structure party competition in Central and Eastern Europe, and he groups them into socio-economic and socio-cultural ones. The socioeconomic dimension contains the conflict between labour and capital as well as urban versus rural interests. He adds that sectoral differentiation may emerge over time, which would render the socio-economic dimension somewhat similar to the economic cleavage in Western Europe, but not in the strict sense of class-based voting (see also von Beyme 1996; Deegan-Krause 2007). In the socio-cultural sphere, Beichelt's classification includes the conflict between centre and periphery as well as ethno-linguistic and religious divides. Here, he expects different issues to align and merge into integrated party policies. Finally, Beichelt (2001, 190–94) argues that these divides would only fully unfold after the so-called regime divide between Communist successor parties and parties that have their roots in the opposition to the former regime had vanished (see also Grzymała-Busse 2001).

flicts, such as the one between materialism and post-materialism, narrowed the societal basis of political parties, casting doubt on the existence of full cleavages in Western Europe as well.

<sup>7</sup> Von Beyme (1996, 129) lists a total of eight cleavages: labour versus capital, city versus countryside, secular versus religious, Westernisers versus nationalists, centre versus periphery, materialism versus post-materialism, centralism versus decentralisation, and libertarian versus bureaucratic.

<sup>8</sup> In some countries, reformed Communist successor parties took the place of social democratic parties, but their position in the party system was strongly affected by other cleavages, most importantly the regime divide. The Czech Republic is an outlier here, as a social democratic party without any roots in the former regime emerged as one of the most stable and successful political forces in the country.

Regardless of the exact divides, there are considerable within-region differences between Central and Eastern European party systems. Rohrschneider and Whitefield (2009) acknowledge these differences, but conclude that party competition still follows a similar pattern across the region which can be described as a "structured diversity". It is structured in a way that party positions "coalesce around a proreform versus antireform dimension"—liberal socio-economic positions correspond with liberal socio-cultural ones in the pro-reform pole, and the anti-reform pole unites economic protectionism and cultural conservatism (Rohrschneider and Whitefield 2009, 299–300). The diversity then arises from differences in issue salience. The authors argue that a country-specific set of salient issues, resulting from socio-historical trajectories and the agency of political parties, constitutes the specific content of national party competition. Regarding the relation and hierarchy between different issue dimensions, they conclude that "economic issues constitute the common basis for party competition in the region and other conflicts add a country-specific flavour" (Rohrschneider and Whitefield 2009, 298).

While the literature widely agrees that individual issue divides within the socioeconomic and the socio-cultural dimensions reinforce each other, the unidimensional concept of party competition along a pro-reform versus anti-reform dimension remains contested. Some authors provide evidence in support of this notion (Marks et al. 2006; see also Kitschelt 1995), whereas others find that socio-economic and socio-cultural divides align in some countries and cut across each other in others (Kitschelt et al. 1999; Deegan-Krause 2007; Casal Bértoa 2014). The findings regarding the hierarchy between these two cumulative issue dimensions are similarly inconclusive. Kitschelt and co-authors (Kitschelt et al. 1999; Buštíková and Kitschelt 2009), for instance, tend to agree with the predominance of distributional conflicts in Central and Eastern Europe. Yet, there is ample evidence that socio-cultural conflicts about citizenship, minority rights, religion, language, gender, or the interpretation of history are of equal, or even higher, importance (von Beyme 1996; Bunce 2005; Enyedi 2008; Pirro 2016; Pytlas 2016; Minkenberg 2017). In a comparison of Western European and post-Communist democracies, Deegan-Krause summarises the situation regarding cleavage structures as follows: "Economic issues […] are not necessarily the best way to compare the strength of issue divides in East and West. Although economy-related divides emerged throughout post-communist Europe, non-economic issue [sic] also aligned closely with party preference"(Deegan-Krause 2007, 543–44). Pytlas (2016, 6) points out that different socio-cultural conflicts have a tendency to reinforce each other and result in highly polarised "value wars" (Ágh 2001, in Pytlas 2016, 6) between deeply divided political camps.

These observations make it difficult to reduce a content analysis of party competition in Central and Eastern Europe to one single dimension. Given the salience of socio-economic and socio-cultural issue divides, and reinforcing divides within each dimension, the present study applies a two-dimensional concept of the policy space, using comprehensive socio-economic and socio-cultural dimensions. Another reason for focusing on broader socio-economic and socio-cultural dimensions is that the issue positions of Central and Eastern European parties are less stable than those of Western European parties; however, the vast majority of them developed an ideological, or value, core that remains relatively constant and offers orientation to voters and other parties (Hloušek and Kopeček 2010, 9–10). In such a situation, where parties have rather shallow roots in the society's structural differences, the use of the broader ideological dimensions can be beneficial.

#### 2.2.3 Party competition with radical right parties in Central and Eastern Europe

How do these developments and characteristics of post-Communist parties and party systems affect radical right parties and their participation in government? Central and Eastern European radical right parties are electorally less successful and consistent than their Western European counterparts (Minkenberg 2002, 336, 2017, 101; Mudde 2005a). Yet, the discussion of the structural development of parties and party systems in the region has revealed that these features are not limited to radical right parties. High levels of volatility and the constant appearance of new parties indicate that fluctuating electoral fortunes affect other parties just as much as the radical right. Several new parties even managed to enter government immediately after their electoral breakthrough. In such an environment, radical right parties did not need one or two decades of organisational consolidation and electoral growth to gain executive power (Minkenberg 2017, 129), as was the case in Western Europe (Bale 2003; de Lange 2008, 2012).

Despite the perpetual advent of new parties, an electorally and organisationally stable core of established parties populatemany Central and Eastern European party systems, reflecting the fact that it is possible for parties to survive in the long term. Stabilising patterns of governmentindicate that these established partiesmight also have advantages when it comes to participating in government. Hence, while organisational instability and limited, or short-lived, electoral successes are not necessarily an obstacle to their participation in government, parties which display electoral consistency and organisational consolidation should still have advantages over new weakly institutionalised parties.

The ideological configuration of the Central and Eastern European party systems has implications for government formation with radical right parties as well. Socio-cultural divides feature prominently in the region's party systems. Given the politicisation and salience of the core issues of radical right parties and the presence of corresponding attitudes in significant parts of society (Zick, Küpper, and Hövermann 2011), the limited electoral success of these parties may come as a surprise. However, the salience of nationalism also helps explain the relatively poor performance of radical parties at the polls. Radical right parties are not the only suppliers of radical right politics. In Central and Eastern Europe, they face fierce competition from mainstream parties that offer similar policies, though often in a more moderate fashion or with a different framing (Minkenberg and Kossack 2015; Pirro 2016; Pytlas and Kossack 2015; Pytlas 2016; Minkenberg 2017, chap. 6). However, while mainstream parties' openness to radical right politics has a negative impact on the radical right in the electoral arena, it may facilitate cooperation in government.

The ideological configuration of Central and Eastern European party systems also has a bearing on the potential partners of radical right parties in government. In Western Europe, centre-right parties, which combine conservative socio-cultural positions with liberal socio-economic views, have been the natural allies of the radical right. While not all Western European radical right parties adopted Kitschelt's winning formula of combining ultra-nationalism with economic liberalism (Kitschelt and McGann 1995), conservatives and Christian democrats are still situated closest to the radical right in a two-dimensional space (see Figure 2.1). Green and socialist parties are located in the opposite quadrant and represent their fiercest competitors.<sup>9</sup>

The picture in Central and Eastern Europe is quite different and less clear. Here, radical right parties tend to combine their ultra-nationalist agenda with social-national economics, which places them in the lower left quadrant of Figure 2.1. The party families present in this quadrant represent a diverse group of potential allies and coalition partners for radical right parties, including social democrats and conservatives. Both party families, however, can be found in the opposite quadrant as well. This diversity results from the specific Communist and transformational legacies of different Central and Eastern European countries (Kitschelt et al. 1999). Most social democratic parties in the region are reformed Communist successor parties. Particularly in countries with a patrimonial Communist regime, such as Bulgaria or Romania, these parties had embraced elements of nationalism in the Communist era, and they underwent a slow and partial process of structural and ideological reform after 1989. Therefore, they maintained a nationalist and protectionist profile, which situated them in close ideological proximity to the radical right (Ishiyama 1998). In other countries, the social democrats developed as part of the opposition or, more frequently, as credibly reformed Communist successors with relatively liberal socio-cultural and even socio-economic positions (Hloušek and Kopeček 2010, chap. 2).

<sup>9</sup> Even though the general alignment that Marks et al. (2006) suggest is viewed rather critically in the case of Central and Eastern Europe and this study's definition of radical right parties differs from the equivalent of "Radical Tan" parties used by those authors, the figures still illustrate the general patterns of party positions discussed here.

*Figure 2.1: Party positions in Western and Central and Eastern Europe in the mid-2000s*

Source: Marks et al. 2006, 158–59.

The conservative party family is also located in different quadrants. According to Hloušek and Kopeček (2010, chap. 8), Central and Eastern European conservative parties can be divided into liberal and nationalist branches. National-conservative parties, such as the Polish PiS and the Hungarian Fidesz before they transformed into full-fledged radical right parties, attribute great importance to nationalism and (Christian) religious values. They often advance national-protectionist economic policies, which renders them potential allies of the radical right as well. Liberal conservative parties are rather opposed to the radical right, because they do not share their national-protectionist positions, and they often reject the exclusionary nationalism of their national-conservative and radical right counterparts.

The emptiness of the upper left quadrant in Central and Eastern Europe can be understood as a legacy of the Communist and transformational periods. The oppositional "pro-reform" forces, including those Communist successor parties that underwent credible changes pursued a rather centrist, or even liberal, economic agenda due to the widespread anti-Communist bias (von Beyme 1996, 125), placing them in the upper right. Green parties, which traditionally occupy the left-libertarian end of the new politics dimension in the West, are missing in Central and Eastern European party systems because a post-materialist silent revolution has not occurred in the region.

## **2.3 Different but similar: Parties and party systems in Central and Eastern Europe compared to the West**

The discussion of radical right parties and the configuration and development of party systems points to similarities and differences between Western and Central and Eastern Europe. Central and Eastern European radical right parties emphasise, at least in part, different issues than their Western European counterparts, and the political mainstream in Central and Eastern Europe has been more open to radical right politics from the beginning. Moreover, party systems are less stable and institutionalised than in the West, particularly in the first post-Communist decade. They are becoming more consolidated in the following decades but this development is proceeding rather slowly, it is not always linear, and it exhibits significant intra-regional variation. Value conflicts related to state- and nation-building play a more important role in Central and Eastern Europe than in most Western European countries. However, socio-economic issues have never been absent from party competition in Central and Eastern Europe, particularly during the early phase of the post-Communist transformation when rebuilding the whole economic system was one of the top priorities on the political agenda. Many of these specific features of Central and Eastern European politics are related to the region's historical legacies from the Communist and transformational period, and, in part, even from the pre-Communist era (Jowitt 1991; Crawford and Lijphart 1995; Kitschelt et al. 1999; Ekiert and Hanson 2003a; Pop-Eleches 2007; LaPorte and Lussier 2011; Ekiert 2015). These legacies serve as text and context for radical right parties: They are "revived […] and reinterpreted" in the parties' ideology and they affect their "cultural and structural opportunities" (Minkenberg 2009, 454; see also Pirro 2016).

Nevertheless, there are also substantial similarities between radical right parties and party systems in Western and Central and Eastern Europe.The party systems of both regions are converging in terms of stabilisation and institutionalisation. Their convergence does not result from a one-sided movement of Central and Eastern Europe catching up with the established party systems in the West, but from opposite trendsin the party systems of both regions.Thus, they aremeeting somewherein the middle.Moreover, societal roots of cleavages—or divides—are eroding in both parts of the continent, and increasing polarisation of socio-cultural value conflicts is not a unique feature of Central and Eastern European party systems either. These similarities suggest that the fundamental patterns of party competition are comparable in Western and Central and Eastern Europe. Moreover, radical right parties assume a similar role in party competition in East and West. They participate in democratic politics in order to advocate for the idea, and supremacy, of a homogenous national community in their own nation-state. By doing so, they oppose the basic values of liberal democracy—and sometimes the democratic system itself.

These similarities observed in the discussion support Mudde's (2007, 3–5) plea for pan-European research on radical right parties,whereas the differences resonate with Minkenberg (2002), who considers the Central and Eastern European radical right a phenomenon *sui generis* (see also Pirro 2016). However, Minkenberg neither rejects the use of established concepts and theories, nor does Mudde neglect contextual differences between both parts of the continent. The present study follows a middle path and integrates both perspectives. This approach conceives of radical right parties and party competition in Western and Central and Eastern Europe as functionally equivalent (Pytlas 2018). At the same time, it emphasises the need to adapt and modify "Western" concepts and theories in order to account for specific features of radical right parties and party competition in Central and Eastern Europe. Thus, the present study does not start from scratch in its quest to explain the government participation of radical right parties in Central and Eastern Europe, but it draws on the rich body of literature on the radical right and government formation in Western European democracies. Moreover, evaluating contextualised versions of these theories in light of new empirical data has the additional benefit of providing "broader lessons relevant to the study of radical right politics across Europe and in 'the West'" (Pytlas 2018, 11).

## **3. Theories of government formation**

This chapter provides an overview of the research on government formation. It starts with a discussion of the rational choice approach, then presents different theories of coalition formation, and finally examines the empirical evidence related to these theories in Western and Central and Eastern Europe. This literature review will distinguish between coalition-centred and party-centred theories as well as the impact of contextual constraints on coalition formation. To conclude, the chapter will outline the findings associated with government formation and radical right parties.

#### **3.1 The rational choice approach in research on government formation**

Most theories of coalition formation follow the rational choice paradigm. De Swaan (1973, 12–25) summarises the main features of the rational choice approach in the context of coalition formation (see also Strøm 1990a). The rational choice approach is based on the assumption that fully informed actors make decisions based on their preferences in order to maximise the utility of a given outcome. In order to be applicable to real-world politics, however, this assumption must be relaxed. Time, contextual constraints, or a lack of resources may limit an actor's ability to gather the information needed to make a decision. Hence, they will never have—and in most situations, they do not even strive for—complete information. Instead, they collect only accessible information which is most relevant to their decisions. Similarly, actors may not always seek to maximise their utility, but rather settle for "a solution that might satisfy their aspirations" (de Swaan 1973, 14). The picture is complicated even further when decisions are based on competing preferences. Here, decisionmaking involves multiple trade-offs between different goals which can hardly be maximised simultaneously.

Strøm and Müller (1999) have convincingly demonstrated that political parties decide their course of action based on considerations related to gaining representation in public office, most importantly the national government; to implementing their preferred policies; and to winning popular support during elections. In the authors' words, political parties seek policy, office, and votes (see also Strøm 1990a). When forming coalitions, parties must confront the trade-offs between these goals (Strøm and Müller 1999, 9–13). Among these goals, however, only policy and office have an intrinsic value for the parties, while vote maximisation is rather instrumental—it is mainly a tool for gaining the capacity to implement policies and/or to enjoy the spoils of public offices, such as key leadership positions and financial revenue for the party (Strøm and Müller 1999, 6–8; see also Sartori 1976, 327). The distinction between intrinsic and instrumental goals is important for reconciling the theoretical assumptions that underlie the majority of works on electoral, and post-electoral, party competition. As Benoit and Laver (2006, 41–42) point out, research on party competition in the electoral arena mostly follows the Downsian (1957) tradition, which views parties first and foremost as actors seeking to maximise votes. Coalition theories, however, are concerned with post-electoral party competition and conceive of parties primarily as office- and policy-seekers. Hence, the notion that vote maximisation is first and foremost an instrumental goal serving the purpose of getting into public office and/or implementing certain policies is paramount for maintaining consistent assumptions about parties' strategic behaviour in both spheres of party competition.

Scholars who emphasise the importance of the national context for coalition formation (von Beyme 1984, 389) frequently criticise the rational choice paradigm for relying on unrealistic and simplifying assumptions. Strøm (1990a) points out, however, that rational choice-based coalition theories account for a good deal of the context in which coalition formation takes place. He illustrates, for instance, that specific contextual configurations can affect the incentives for political parties to prefer a policy-, office-, or vote-seeking strategy. Though coalition formation always takes place in a specific social, political, cultural, and even temporal context, and these elements definitely affect the bargaining process, political parties' pursuit of policy, office, and votes also impacts the outcome. Therefore, it is also misguided to over-emphasise the problems associated with the rational choice paradigm and completely forsake any efforts to draw broad conclusions related to coalition formation. In order not to abandon the goal of cross-national generalisations, the present study follows Strøm's (1990a, 566) strategy,maintaining the basic assumptions of the rational choice approach, but at the same time incorporating the influence of contextual factors.

Unlike *homo economicus* in the original rational choice theory, political parties are collective actors, not individuals. Even though coalition negotiations are usually conducted by a few representatives of the parties' leadership, these elites need to consider the positions of competing factions within their parties during the bargaining process. However, the vast majority of research on coalition formation conceives of parties as unitary actors (Laver and Shepsle 1996, chap. 12; Laver and Schofield 1998, chap. 2; see also Benoit and Laver 2006, 41; Müller, Bergman, and Ilonszki 2019, 26). The presence of intra-party competition between rank-andfile members and elites or different factions and organisational units cannot be dismissed. With regards to coalition formation, however, Laver and Schofield (1998, chap. 2) conclude that the treatment of parties as unitary actors is generally justified, if they share common goals and do not resemble mere "coalitions of factions" (Irving 1979, in Laver and Schofield 1998, 20; see also Benoit and Laver 2006). Therefore, the present study follows the majority of research on coalition formation and treats parties as unitary actors.

## **3.2 Theories of coalition formation and their empirical results: Central and Eastern Europe and Western Europe compared**

## 3.2.1 Coalition-centred theories

The coalition-centred branch of research on government formation has produced a wide range of theoretical propositions that aim at explaining or predicting the composition of coalitions.<sup>1</sup> The first formal theories in the game theoretic tradition conceived of political parties as pure office-seekers. These policy-blind theories assume that parties seek to translate their parliamentary seat share into maximum control over as many cabinet posts as possible by reducing the costs of negotiating with other parties in the process of coalition formation (Laver and Schofield 1998, 92–94; Dumont, de Winter, and Andeweg 2011, 7; Müller, Bergman, and Ilonszki 2019, 15–16). In this vein, the theory of the minimal winning coalition (von Neumann and Morgenstern 1953, in Laver and Schofield 1998, 92; see also Riker [1962] 1984) holds that parties seek to form coalitions based on the barest majority possible, such that a minimal winning coalition would lose its majority if one party left the coalition.This approach minimises the number of coalition members competing for political influence while also guaranteeing the backing of the parliament in a vote of (no) confidence. Other office-oriented theories propose that parties aim at reducing bargaining costs by forming only those coalitions that include as few parties as possible (Leierson 1968, in de Swaan 1973, 65) or the smallest number of seats sufficient for reaching a majority in parliament.This last type is referred to as minimum winning coalition (Riker [1962] 1984, 32–33; Laver and Schofield 1998, 94–95).

Several empirical studies show that minimal winning coalitions are indeed the most frequent coalition type in Western Europe. Depending on the exact sample of countries and time period covered, between 30 and 40 per cent of governments in post-war Western Europe followed the logic of minimal winning coalitions (Laver and Schofield 1998, 95; Martin and Stevenson 2001; Mitchell and Nyblade 2008, 207; Bergman, Ilonszki, and Müller 2019b, 538).The minimum winning proposition,

1 For an overview, see e.g. Laver and Schofield (1998) or Müller, Bergman and Strøm (2008).

however, has proven to be too rigid. Parties seem to prefer more stable majorities that forgive the occasional defection (Laver and Schofield 1998, 96; Dumont, de Winter, and Andeweg 2011, 8). The minimum parties proposition also finds less support inWestern Europe than the minimal winning proposition (Laver and Schofield 1998, 95).

However, these purely office-oriented theories were criticized for several reasons. First, they fail to predict the correct outcome of coalition formation in (at least) half the cases. Second, the theory of the minimal winning coalition usually produces several equiprobable coalitions of that format,making it difficult to discern how well it actually performs.Third, critics note the questionable assumption that parties are purely office-seeking (Laver and Schofield 1998; Martin and Stevenson 2001; Benoit and Laver 2006; de Winter and Dumont 2006).

Scholars have sought to remedy these shortcomings by incorporating parties' policy preferences into theories of coalition formation, based on the assumption that ideological proximity reduces bargaining costs while also facilitating cooperation and policymaking among government members (Laver and Schofield 1998, 96–98; Dumont, de Winter, and Andeweg 2011, 8–9). The minimal connected winning theory (Axelrod 1970, 166–75; see also Laver and Schofield 1998, 97–102), for instance, argues that parties should form only such majority coalitions that are ideologically connected or, more precisely, situated next to each other in a unidimensional policy space.

In a similar vein, the policy distance, or minimal range, theory (de Swaan 1973, chap. 5), posits that parties seek to minimise the policy range of a coalition on the left-right dimension. This theory exists in closed and open versions. In the closed version, it requires all coalition parties to be connected. Since de Swaan's (1973, 88) theory also includes the majority element, the closed minimal range theory is very similar to the minimal connected winning theory. The open minimal range theory, however, is primarily concerned with the policy range of the coalition and less with the position of the individual parties in relation to each other. Hence, the open version allows for opposition parties to be situated between the coalition partners.

Another policy-oriented explanation of coalition formation is the median party proposition. It holds that the party of the median legislator, or the member of parliament with an equal number of representatives to the left and right, will be part of the government coalition. Assuming that no member of parliament votes against their party's policy preferences, no policy-consistent majority can be formed without the median party (Laver and Schofield 1998, 111).

Empirical studies on coalition formation inWestern Europe show that including ideological proximity significantly improves the explanatory power of formal coalition theories, such as the minimal connected winning and the minimal range theory (Martin and Stevenson 2001; Mitchell and Nyblade 2008). Around 80 per cent of all coalitions in Western European democracies included the median party (Laver and

Schofield 1998, 113; see also Bergman, Ilonszki, and Müller 2019b, 540). Yet, similar to the theory of the minimal winning coalition, the median party proposition usually yields multiple equiprobable outcomes (Dumont, de Winter, and Andeweg 2011, 9).

Before turning to the predictive capacity of these theories in Central and Eastern Europe, it should be noted that coalition governments occur more frequently in this region than in Western Europe. Institutional design plays a significant role in explaining the low frequency of single-party governments in Central and Eastern Europe. Most importantly, none of the post-Communist countries opted for a majoritarian electoral system, a key institution for providing individual parties with an absolute majority in parliament (Grotz andWeber 2011, 100–101; Bergman, Ilonszki, and Müller 2019b, 538).

When compared to Western Europe, the formal office-oriented theories correctly predict a similar share of coalitions in Central and Eastern Europe. Minimal winning coalitions, for instance, are also the most frequent type of coalitions in post-Communist democracies (Grotz and Weber 2011, 101–2; Bergman, Ilonszki, and Müller 2019b, 538; see also Savage 2016). Some empirical evidence even supports the minimum parties proposition in Central and Eastern Europe (Savage 2016, 519). The share of minority governments and surplus coalitions differ between these regions, but the difference is not dramatic. Bergman and his collaborators find that 24 per cent of all governments in Central and Eastern Europe between 1990 and 2014 are oversized coalitions (compared to 23 per cent in Western Europe) and 38 per cent are minority governments (compared to one-third in Western Europe). They also highlight significant intra-regional differences in both parts of the continent (Bergman, Ilonszki, and Müller 2019b, 538–39; Müller-Rommel et al. 2008, 813).

The policy-oriented theories are rarely tested in Central and Eastern Europe, which is probably related to difficulties in measuring the policy space in the post-Communist democracies (see Chapter 2.2). Grzymała-Busse (2001) and Savage (2016) include policy distance in their analyses of government formation in Central and Eastern Europe. However, Grzymała-Busse (2001, 91) finds that only ten and 24 per cent of the coalitions in her study minimised the ideological distance on a socio-cultural and socio-economic dimension, respectively. In Savage's (2016, 519) model, ideological distance has no significant effect on the composition of coalitions. Grotz and Weber (2011, 204–5) acknowledge the problems associated with measuring policy distances in Central and Eastern European party systems. Therefore, they use the equally imperfect concept of party families to operationalise ideological proximity. They find little support for the minimal connected winning proposition, since only 24 per cent of the minimal winning coalitions consisted of parties from similar party families. Similar to Western Europe, the median party is also included in three out of four governments in Central and Eastern Europe (Savage 2016, 540; Bergman, Ilonszki, and Müller 2019b, 540). Due to the limited

impact of ideological proximity, the majority of researchers conclude that officeseeking is more influential than policy considerations in the coalition negotiations of Central and Eastern Europe (Bergman, Ilonszki, and Müller 2019b, 566; see also Döring and Hellström 2013; Savage 2016).

In addition to these formal theories, scholars also advance empirically oriented, non-formal theories of coalition formation. Strøm, Budge and Laver (1994, 311), for instance, argue that incumbent coalitions have an advantage in coalition bargaining if the institutional setting renders them the "reversion point" when parties cannot agree on an alternative government. Other authors have proposed that the continuation of incumbent governments results from parties' attempts to reduce transaction costs by working with familiar partners. Thus, they can build on established routines and trusting relations instead of starting over with new coalition partners which may entail a higher degree of uncertainty (Bäck and Dumont 2007, 474–75; Martin and Stevenson 2010, 504).

Martin and Stevenson (2010) find empirical support for a positive incumbency effect in Western Europe based on both institutional settings and parties' preferences for familiar partners. Moreover, the authors draw attention to the often-overlooked issue of government termination.They show that the incumbency effect also depends on themode of termination of theincumbent coalition andits electoral performance (Martin and Stevenson 2010, 515–16). In Central and Eastern Europe, incumbency had a negative electoral effect. 84 per cent of the incumbent governments suffered electoral losses averaging 37 per cent fewer seats in the following parliament (Bergman, Ilonszki, and Müller 2019b, 564–65; see also Roberts 2008). Roberts (2008) refers to this as "hyperaccountability" of governments in Central and Eastern Europe. Savage confirms this incumbency disadvantage, but he adds that it only applies to government formation immediately after elections. When governments are formed mid-term, incumbent governments even have an advantage (Savage 2016, 524–28; see also Döring and Hellström 2013, 684).

## 3.2.2 Party-centred coalition theories

While the classic policy- and office-oriented theories of coalition formation explain the composition of coalitions as a whole, party-centred theories focus on the coalition membership of individual parties. By shifting the focus to political parties as the "building blocks" of coalitions (Müller and Strøm 2000a, 6), they provide a partial remedy to the problem of equiprobabilty. A few studies apply a broad scope (Warwick 1996; Döring and Hellström 2013; Savage 2014), but most party-centred research focuses either on particular party types and families (Dumont and Bäck 2006; Druckman and Roberts 2007; Dunphy and Bale 2011; Zaslove 2012; de Lange 2008; Grotz and Weber 2013; Gherghina and Jiglău 2016) or on the effect of particular factors, such as experience in previous government (Tavits 2008b; Martin and

Stevenson 2010) or electoral success (Mattila and Raunio 2002, 2004), on a party's coalition membership.

Several party-centred theories use the characteristics of individual parties to explain their participation in government. Similar to the office-oriented, coalitioncentred theories, the seat share of parliamentary parties features prominently among these characteristics. Döring and Hellström (2013, 693–94) find strong evidence supporting the assumption that election winners enter governing coalitions. This finding holds true, if the largest party also becomes the formateur of the coalition (Warwick 1996, 488; Martin and Stevenson 2001, 43; Savage 2014, 556). The electoral result also matters for smaller parties. Some studies show a linear correlation between the size of a party's parliamentary group and their chance to enter government in Western and Central and Eastern Europe (Mattila and Raunio 2004; Döring and Hellström 2013). Another study on coalition formation in Western Europe, however, finds a negative correlation between the size of potential junior partners and their chances to become coalition members, which suggests that formateurs seek to maximise their own influence in government by choosing partners that are just big enough to secure a working majority (Warwick 1996, 499).

Some scholars argue that not only absolute electoral results but also gains and losses of a party compared to the previous elections can affect their coalition membership. Electoral gains can be understood as the voters' intention to bestow a party with more responsibility, whereas losses signal their negative assessment of a party's previous performance and/or what it offered for the future (Warwick 1996; Döring and Hellström 2013; see also Dumont and Bäck 2006; de Lange 2008). Mattila and Raunio (2004, 280) study coalition formation in 15 Western European democracies, and they find that vote gains, but not losses, have a significant effect on a party's coalition membership. Döring and Hellström (2013, 693–95), however, find that losses also reduce parties' chances to enter government in both parts of the continent.

Another structural feature of political parties that can influence their chances of entering a coalition government is political experience. In Western Europe, previous experience in government matters most when a party is a member of the incumbent government (Martin and Stevenson 2001, 2010; Bäck and Dumont 2007). Incumbency increases a party's chance to become the formateur of a coalition regardless of whether it was previously the prime minister party or a junior partner. However, an incumbent prime minister party is less likely to become a junior partner in the next government (Mattila and Raunio 2004, 279–81; Martin and Stevenson 2010).Moreover, the mitigating effects of (non-) conflictive government termination and electoral success also apply to the party level (Martin and Stevenson 2010). Due to hyperaccountability, incumbent parties have no significant advantage in Central and Eastern Europe (Döring and Hellström 2013, 694; Bergman, Ilonszki, and Müller 2019b, 564). Similar to coalitions as a whole, however, they are more likely to remain in office when governments re-form mid-term.

With regard to Central and Eastern Europe, Grotz and Weber (2011) introduce the concept of seniority, which covers experience in parliament as well as in government. Empirically, however, the authors determine that governing parties are not necessarily more experienced than those in opposition (Grotz and Weber 2011, 205–6; see also Savage 2016). Similarly, new parties were not significantly more likely to enter government than parties with parliamentary experience (Savage 2016, 524–25). Research on Western Europe also concludes that experience in parliament or in any government prior to the incumbent one does not create an advantage for parties in coalition bargaining (Martin and Stevenson 2001; Dumont and Bäck 2006).

Another structural feature to be addressed here is party organisation. Maor (1998), for instance, finds that effective channels of intra-party conflict resolution, usually found in well-organised and decentralised parties, contribute to their bargaining power. Similarly, Druckman (1996) points out that the reduction of factionalism has a positive effect on government stability in Western democracies. In Central and Eastern Europe, the effects of party organisation on coalition formation have not yet been subjected to a comparative analysis. However, Tavits (2013) finds that an effective party organisation has a positive impact on the "success", "survival", and "unity" of Central and Eastern European parties. By showing that effectively organised parties are better able to "successfully overcome any crises" and "keep their representatives unified and the party cohesive in office" (Tavits 2013, 195), this study provides at least indirect support for a similar effect of party organisation in Central and Eastern Europe.

Whether parties enter government and remain in opposition can also depend on the ideological preferences of these parties and their competitors. Research on government formation in Western Europe shows that the ideological distance of a potential junior partner from the formateur, or a party's ideological distance from the median party, affectsits chances to enter government (Warwick 1996;Martin and Stevenson 2001; Mattila and Raunio 2004; Döring and Hellström 2013). Median parties themselves are in a favourable position to become formateurs of a coalition government (Warwick 1996; Martin and Stevenson 2001, 43; Mattila and Raunio 2004).

In Central and Eastern Europe, Savage (2016, 519) confirms that the median party ismore likely than other parties to beincludedin government.Döring and Hellström (2013) show that the ideological distance between a party and the median of a universal left-right dimension is not significantly related to entering government in Central and Eastern Europe, concluding that ideology has no effect on a party's coalition membership in the region. Savage (2014), however, provides strong evidence that the ideological preferences of Central and Eastern European parties have an impact on their participation in government. Similar to the findings in Western Europe, he shows that parties closer to the median are more likely to enter government and that a party's ideological proximity to the formateur is a key determinant of junior coalition membership.

#### 3.2.3 The impact of context factors on government formation

A wide range of context factors can constrain parties' options and decisions in coalition politics, for example the institutional frameworks, party systems, or historical trajectories.

#### Institutional constraints

With regards to the legal-constitutional framework, Strøm, Budge, and Laver (1994) mention, for instance, cabinet formation rules, cabinet operation rules, and legislative rules. Cabinet formation rules comprise provisions that regulate the process of government formation. In some countries, the constitution contains a specific procedure for selecting a formateur, or stipulates whether a newly formed government requires an investiture vote in parliament. Cabinet operation rules, such as the modes of cabinet decision making and the distribution of power among cabinet members, can indirectly influence the negotiations preceding coalition formation. Similarly, the rules for cabinet termination, such as the existence of constructive, or destructive, votes of no confidence, might be taken into account when deciding upon the format of a coalition.

Legislative rules can affect the parliamentary majority that a government needs to survive and pass legislation. The electoral system, for instance, impacts how parties are represented in parliament. Majoritarian systems more often create large parliamentary groups and single-party governments than (semi-) proportional ones (Strøm, Budge, and Laver 1994, 314–16; Buzogány and Kropp 2013, 279; Nikolenyi 2014, 10–11). But even within the group of semi-proportional and proportional systems, electoral thresholds or the number and size of districts affect the distribution of seats in parliament (Nikolenyi 2014, 25–27). Moreover, federalism and bicameralism can impact coalition formation, particularly when they occur together in the form of a second chamber of parliament composed of federal state representatives that have the power to block legislation (Kropp, Schüttemeyer, and Sturm 2002b, 20).

Yet, generalisations about the individual effects of specific institutional factors on coalition formation are quite difficult. Whether a single institutional factor constrains or facilitates government formation can change profoundly depending on the overall institutional framework (Strøm, Budge, and Laver 1994, 325–26).<sup>2</sup> On a very general level, the literature agrees that electoral systems affect government formation. Across Europe, proportional systems generate minority situations in parliament more frequently than majoritarian systems, which often empower a single party with a clear electoral majority. Consequently, the dominance of proportional systems in Central and Eastern Europe is one of the key explanations for the low level of single-party majority governments in the region (Laver and Schofield 1998, 204; Bergman, Ilonszki, and Müller 2019b, 536).

The vast majority of research on Western Europe further agrees that the absence of investiture votes favours the formation of minority governments (Müller and Strøm 2000c, 567–69; Mitchell and Nyblade 2008, 229; Bergman, Ersson, and Hellström 2015, 360–61).<sup>3</sup> The institutional frameworks in all Central and Eastern European countries include some version of an investiture vote, which is why this factor cannot explain the variance that exists across the region. Nikolenyi (2014) finds, however, that minority governments are more likely to form in Central and Eastern European polities where the parliament's involvement in the process of coalition formation is not limited to the "reactive role to confirm, or reject, the prime ministerial appointment made by the head of state" (Nikolenyi 2014, 32).

## Party systems

Following Sartori (1976), the characteristics of party systems can be distinguished into numerical-structural features, such as fragmentation, and ideological ones, such as polarisation. Both feature prominently in research on coalition formation in Western and Central and Eastern Europe. Similar to the institutional factors, however, these party system characteristics should not be examined in isolation from one another (Dodd 1976, 139).

The formation of (majority) governments becomes more complex when the fragmentation of party systems increases (Dodd 1976; Kropp, Schüttemeyer, and Sturm 2002b; Müller, Bergman, and Strøm 2008). Depending on the unit of analysis, empirical studies find various effects of fragmentation on government formation. In their study on the government participation of individual parties, Döring and Hellström (2013) demonstrate that fragmentation has no effect in either part of Europe (see also Savage 2014). Warwick (1996, 495), however, shows that parties in more fragmented party systems in Western Europe aim at reducing the number of (small) coalition members in order to minimise bargaining costs. Grotz and Weber's

<sup>2</sup> Strøm, Budge, and Laver (1994, 309) also mention that some institutional choices result in hard constraints that fully eliminate certain coalitions while others create soft constraints that make them less likely.

<sup>3</sup> There has been some doubt as to whether this condition alone is sufficient for the emergence of minority governments (Müller, Bergman, and Ilonszki 2019, 32).

(2011, 202–3) coalition-centred research on Central and Eastern Europe indicates that minimal winning coalitions occur more regularly in compact, non-fragmented party systems, whereas oversized coalitions are formed more frequently when fragmentation is high. Somewhat contrary to Warwick, they argue that including more parties than necessary provides the coalition with a safety net in the fluid environments of Central and Eastern European democracies. Both arguments are plausible and not necessarily contradictory, if stable coalitions can be formed by a small number of large parties. In highly fragmented party systems with many small parties, however, coalition formation usually entails a trade-off between minimising the number of coalition members and maximising stability.

Research on government formation investigates few structural-numerical characteristics of party systems other than fragmentation. Herman and Pope (1973, in Keudel-Kaiser 2014, 60) demonstrate that minority governments are more likely to be formed in Western European party systems when one large party comes close to controlling a majority in parliament. Keudel-Kaiser (2014, 242–43) finds no similar effect in Central and Eastern Europe because parties were rarely that strong. She shows, however, that the dominance of two electorally strong parties, neither of which attain an individual majority, can contribute to the formation of a minority coalition in the region.

The findings regarding the impact of the polarisation of party systems in Western Europe are rather mixed. Mitchell and Nyblade (2008, 228–31) test various indicators of polarisation, including the share of extremist parties, the policy range of parliamentary parties, and the policy range weighted by the parties bargaining power, but they find only limited effects of these variables on the format of government. Other studies, however, show that the polarisation of party systems can facilitate the formation of single-party governments (Bergman, Ersson, and Hellström 2015, 359), minimal winning coalitions (Indridason 2011, in Bergman, Ilonszki, and Müller 2019b, 537), or minority governments (Martin and Stevenson 2001, 46; see also Dodd 1976, chap. 7).

The findings are similarly inconclusive in Central and Eastern Europe.The study by Bergman, Ersson, and Hellström (2015, 360) shows no significant effect of polarisation on the format of governments. Grotz and Weber (2011, 203), however, find that minimal winning coalitions are rare in deeply polarised party systems. Keudel-Kaiser's (2014) study on the formation of minority governments in Central and Eastern Europe shows that both the structural-numerical and the ideological configuration of party systems have a strong impact on this particular outcome. In addition to the electoral dominance of two parties, she finds that the presence of noncoalitionable parties, "a lack of coalition partners sharing the main policy positions with the formateur" (Keudel-Kaiser 2014, 257) and, in particular, strong ideological divides between two opposing camps, facilitate the formation of minority governments.These results underline that the impact of the structural-numerical and ideological characteristics of a party system on government formation is quite complex and that it often depends on the specific configurations in which they occur (Dodd 1976, 139; Mitchell and Nyblade; see also Laver and Shepsle 1996).

### Historical context factors

Historical trajectories of a country or region can also influence government formation, as observed in the Western European context, for instance, by the limited sovereignty of some countries after World War II, the scepticism towards Communist parties during the Cold War, or the process of EU integration and the leverage the EU exerts on national politics (Kropp, Schüttemeyer, and Sturm 2002b, 32–37). In Central and Eastern Europe, the focus turns immediately to Communist and transitional legacies. Communist successor parties, for instance, have played an important role in party systems across the region, but these parties have dealt with their past quite differently and, consequently, taken different paths in the post-Communist era. Some of them, such as the Polish Democratic Left Alliance (SLD) or the Hungarian Socialist Party (MSzP), undertook credible reforms and transformed into socialist or social democratic parties, whereas the Communist Party of Bohemia and Moravia (KSČM) in the Czech Republic maintained their orthodox Communist ideology after 1989 (Ishiyama 1997; Grzymała-Busse 2002). In relation to coalition politics, Druckman and Roberts (2007, 24) find that these parties are disadvantaged in several ways, but primarily because "other parties, and particularly their electorates, will view Communist successor parties not only in terms of their legislative strength and ideology, but also in terms of their identity as representatives of the old regime". As a result, their chances to participate in government are significantly lower than those of other parties. If they manage to enter government, Communist successor parties often participate in surplus coalitions and receive a smaller number of ministerial portfolios than they deserved according to their seat share (see also Savage 2016).

In addition, Grzymała-Busse (2001) shows that the regime divide between Communist successor parties and their oppositional competitors serves as a superstructure of coalition formationin the post-Communist democracies of Central and Eastern Europe.This divide overshadowed other determinants of government formation in the region and predicted the outcome of coalition bargaining better than the formal office- and policy-based theories (see also Savage 2016). However, this effect has decreased over time because new politicians gradually replaced old elites and Communist successor parties are evaluated by their policy rather than their Communist identity (Grzymała-Busse 2001, 89; Kropp 2008, 526; Savage 2016). Savage (2016, 526) finds, however, that the regime divide has not lost all its power even after the turn of the millennium.

Another prominent context factor that affects party competition and the transformation process in post-Communist Central and Eastern Europe more broadly is the role of the EU (Kropp, Schüttemeyer, and Sturm 2002b, 32–33; Vachudova 2005; Raunio 2009; Vachudova and Hooghe 2009; Haughton 2011; Börzel and Schimmelfennig 2017; Bochsler and Juon 2020). The shared goal of EU membership, for instance, provided the glue which held together the broad anti-Mečiar coalition in Slovakia in 1998, after the EU had threatened to put the accession negotiations on hold in case the prime minister remained in office (Pridham 2002; Vachudova and Hooghe 2009, 201). In Romania, too, the goal of ensuring the country's EU accession contributed to PDSR's decision not to renew the coalition with the radical right PRM in 2000 (Cinpoeş 2015, 288).<sup>4</sup> Research on the impact of the EU on democratic consolidation and party politics in Central and Eastern Europe suggest that the EU's leverage was strongest before the countries joined the EU (Vachudova 2005; Haughton 2011; Börzel and Schimmelfennig 2017; Bochsler and Juon 2020). In particular during the period of formal accession negotiations, when the majority of Central and Eastern European parties and the public supported EU membership (Beichelt 2004, 44–45; Vachudova 2005, 74, 237), Euroscepticism could reduce parties' coalitionability (Kropp, Schüttemeyer, and Sturm 2002b, 33). In these years, few radical right—and typically Eurosceptic—parties managed to enter parliament, which suggests that the EU had an electoral impact on party competition with radical right parties in Central and Eastern Europe before the government formation stage.

#### 3.2.4 Summary

The literature review demonstrates that research on coalition formation has generated a deep reservoir of cumulative knowledge (Laver and Schofield 1998; Kropp, Schüttemeyer, and Sturm 2002b; Müller, Bergman, and Strøm 2008). The formal coalition-centred theories, in particular the office-oriented minimal winning proposition and the policy-oriented minimal connected winning and minimal range propositions, continue to provide valuable insights into coalition formation across Europe. However, non-formal coalition-centred and party-centred theories as well as the inclusion of context factors have helped not only to better explain the format of government coalitions but also their exact partisan composition.

In a widely recognised study on coalition formation in Western Europe, Martin and Stevenson advanced a comprehensive statistical model that was able to explain the composition of about half the coalitions. They considered this a great success

<sup>4</sup> In Western Europe, the case of Norway tells a similar story. In the aftermath of the negative referendum on EU membership in 1972, different positions on EU membership prevented a centre-right government from forming (Jahn 2002, 232). A more recent study also provides some empirical evidence for the impact of EU membership on coalition politics in Denmark (Juul Christiansen and Brun Pedersen 2012).

"given thatinmost of the bargaining situations […] hundreds –and frequently, thousands – of coalitions could potentially form a government" (Martin and Stevenson 2001, 47). At the same time, however, they note that a general theory of government formation that connects all the individual elements of their model is still missing (Martin and Stevenson 2001, 48–49). Other authors have been more critical and argued that this model is "lumping together two dozens of variables drawn from three main schools […] and therefore lacks parsimony and internal consistency" (de Winter and Dumont 2006, 180). Regardless of the theoretical connection of the individual elements, these observations demonstrate that government formation is a complex process, the outcome of which depends on the interaction of a large variety of factors.

The discussion in this chapter has provided an overview of explanatory factors and assessed the empirical evidence to determine their impact on government formation in Western and Central and Eastern Europe. The results of this discussion are summarised in Table 3.1, which lists the individual theories and explanatory factors while also indicating whether the empirical support found in the literature is present, absent, or inconclusive. The overview indicates that many of the theories that emerged from research on Western European democracies provide insights into coalition formation in Central and Eastern Europe as well.

The most striking similarity between both parts of Europe concerns office-oriented explanations of the format of coalitions and the participation of individual parties in government. The minimal winning proposition and parties' electoral results are key determinants of coalition formation in Western and Central and Eastern Europe. The median party proposition that combines parties' pursuit of office and policy also finds empirical support across the continent. Moreover, incumbency has an effect on the format of governments and the chances of individual parties to enter coalitions across Europe. However, the direction of the incumbency effect is often negative in Central and Eastern Europe, where incumbent parties and coalitions tend to suffer severe losses at the polls.


*Table 3.1: Explanations for government formation and their empirical support in Western and Central and Eastern Europe*

Source: Own composition, based on de Lange 2008, 101–2.

+ and – indicate the presence or absence of empirical evidence for an impact of the respective factors on coalition formation; o denotes mixed or inconclusive findings; n.a. indicates that there was no information available.

The empirical findings in the literature differ between Western and Central and Eastern Europe most evidently regarding the impact of ideological preferences on coalition formation. The coalition-centred minimal winning and minimal range theories are tested less frequently in Central and Eastern Europe and if so, they explain a substantially lower share of coalitions than in Western Europe (Grotz and Weber 2011; Bergman, Ersson, and Hellström 2015; see also Savage 2016; Bergman, Ilonszki, and Müller 2019b). Döring and Hellström (2013) come to a similar conclusion regarding the impact of parties'ideological positions on coalition membership. All these studies use the traditional left-right dimension to account for party ideology. Only one study that applies a context-sensitive approach and constructs a leftright dimension based on the most salient issues in each country comes to a different conclusion (Savage 2014). Based on such conceptualisation of the ideological space, Savage finds strong evidence for an impact of parties' ideological positions on government formation in Central and Eastern Europe as well.

Thus, if the specific features of the ideological space in Central and Eastern Europe (see Chapter 2.2) are taken into account, party competition and coalition formation seem to follow fundamentally similar rules in Western and in Central and Eastern Europe. Whether parties enter government or remain in opposition depends on similar trade-offs between policy, office, and votes across the continent. These trade-offs might take a different shape depending on the regional context, but this context can differ between countries within Western or Central and Eastern Europe as much as between both regions (Mitchell and Nyblade 2008; Bergman, Ilonszki, and Müller 2019b). The discussion has shown, for instance, that the influence of the institutional setup or the configuration of the party system can be quite specific in every country. However, as has already been highlighted in Chapter 2, there are some context-specific features that exist in most Central and Eastern European party systems, such as the regime divide. Hence, when attempting to explain government formation with radical right parties in this region, the present study can draw on established theories of coalition formation, but must take the interaction of different explanatory factors as well as the specific features of the regional context into account.

The remainder of this chapter takes a closer look at the limited body of research on government formation with radical right parties in order to ascertain, whether there are additional explanations to be considered with particular regard to the government participation of this party family.

#### **3.3 Government formation with radical right parties**

Most research on government formation with radical right parties focuses on Western European democracies. Since the 1980s, many radical right parties have entered parliamentinWestern European democracies. It took several years,or even decades, however, before they lost their pariah status and were invited to become junior partners in government (de Lange 2008). In most Western European countries, radical right parties are no longer ostracised, but have gradually become "normal" political competitors. Therefore, de Lange concludes, "although some have interpreted the government participation of radical right-wing populist parties as revolutionary, in fact it is merely the logical consequence of the electoral growth of these parties" (de Lange 2008, 224).

Consequently, de Lange (2008) shows that coalition formation with radical right parties follows a similar logic as coalition formation in general, and similar theories explain why they enter government or remain in opposition. She demonstrates, for instance, that two-thirds of the coalitions with radical right parties in Western Europe are correctly predicted by either the minimal winning, the minimal range, or the minimal connected winning theory. Among these three formal theories, the minimal range proposition exhibited the greatest explanatory power, producing fewer equiprobable results and better explaining the non-membership of radical right parties in coalition governments than the other two (de Lange 2008, 154–55). She concludes that the "minimal range theory clearly outperforms the other formal coalition formation theories and therefore offers the best explanation for the formation of government coalitions in which radical right-wing populist parties have participated.The theory suggests that policy ranges of coalitions are the paramount factor when parties evaluate the coalition alternative" (de Lange 2008, 155).

De Lange's party-centred analysis confirms the important role of office and policy in coalition formation with radical right parties. The seat share of radical right parties and their ideological distance to the formateur are significant predictors of government participation (de Lange 2008, 118–19). Moreover, Zaslove (2012) shows that the organisational strength and stability of radical right parties helps them to enter government in Western Europe. More precisely, he argues that centralised leadership and the capacity to mobilise support in civil society, as well as maintaining an oppositional appeal while being in government, are crucial characteristics for radical right parties who wish to continue participating in government.

On a structural level, increasing polarisation within Western European party systems contributes to the government participation of the radical right. Bale (2003) and de Lange (2012) highlight two factors related to radical right parties, and the strategic reactions of their mainstream competitors, which facilitate their ability to gain executive power (see also Zaslove 2012): First, the increasing electoral support for radical right parties often places them in a pivotal position within the conservative camp because mainstream conservative parties depend on their votes if they want to form a centre-right majority coalition. Otherwise, these parties find themselves in the undesirable position of cooperating with centre-left parties, and this is only possible when polarisation is mild enough to make such a grand coalition

viable. Second, radical right parties managed to politicise their socio-cultural core issues, particularly immigration. Centre-right mainstream parties often applied an accommodative strategy and incorporated the policy positions of the radical right into their own platforms, which contributed to the polarisation of party systems and helped normalise radical right parties and politics (Meguid 2005, 2008).

Government formation with radical right parties in Central and Eastern Europe has received limited scholarly attention so far.Minkenberg (2017) offers a descriptive analysis of this issue in his volume on the radical right in Central and Eastern Europe. He points out that, in contrast to Western Europe, some radical right parties entered coalitions "only a few years after the part[ies] had been formed or shortly after the onset of democratization" (Minkenberg 2017, 129). Moreover, these parties became junior partners in coalition governments with both centre-right and centreleft parties in Central and Eastern Europe.

Fagerholm (2021) seeks to explain why radical parties are included in, or excluded from, government in Western and Central and Eastern Europe. His study on radical right and left parties highlights that government formation with radical (right) parties is a complex phenomenon that requires multicausal explanations. He identifies various combinations of different office- and policy-related factors which explain the government participation of radical parties. They participate in government, for instance, if they make electoral gains and face an electorally weak and ideologically compatible formateur. In some cases, radical parties also enter government if their policy positions are rather distant from those of the formateur, but only if both parties are located on the same side of the ideological spectrum (Fagerholm 2021, 270–71). If the formateur and the radical party are on opposite sides, the radical party remains in opposition (Fagerholm 2021, 273–74).<sup>5</sup>

Fagerholm (2021, 274) acknowledges, however, that the results of his study cover only a limited share of the instances of government formation with radical parties in Europe. The explanations for their inclusion in government is predominantly based on cases from a few countries, particularly Latvia. His explanation for the exclusion from government, in contrast, is better suited for radical left parties and Western Europe. Hence, the author concludes that "although the models provide intriguing explanations of single cases, it is unlikely that they are able to tell us much about general European trends" (Fagerholm 2021, 274).

Despite Fagerholm's(2021) pioneering work,much about government formation with radical right parties in Central and Eastern Europe remains to be discovered.

<sup>5</sup> However, this factor alone is not a sufficient condition for explaining the exclusion of radical parties from government. In the sufficient solution paths, it is combined with either a too small or too large seat share of the radical parties in parliament, a moderation of their ideology and losses at the polls, or large ideological distance to a strong formateur (Fagerholm 2021, 273–74).

The limited body of research on the topic suggests that the explanations for radical right parties' inclusion in, and exclusion from, government are similar to those of the formation of governments in general. Whether these parties get into power depends on their own agency and that of their competitors, which is based on the trade-off between policy, office, and votes, as well as on the constraining and facilitating effects of the context in which these parties operate.This chapter has pointed out which of the explanations of government formation received the most empirical support in Central and Eastern Europe, and with regard to radical right parties. It has thus laid the foundations for developing an analytical model in the following chapter.

## **4. A case-based, configurational, and time-sensitive research design for studying government formation with radical right parties**

This chapter outlines the research design of this study. It first discusses the research designs found in previous studies of government formation and then elaborates on this project's case-oriented, configurational approach using QCA. After introducing the method, it goes on to develop an analytical model and to advance the hypotheses that guide the comparative analysis. The final section of this chapter addresses the operationalisation of the explanatory conditions and the outcome.

## **4.1 Research designs in the study of government formation**

In addition to the theories of coalition formation, this research field has also developed certain traditions with regard to research design. Müller, Bergman, and Strøm (2008, 33–35) introduce a classification that builds on a distinction between complete and parsimonious theoretical approaches and intensive and extensive empirical designs. A very common research design for studying government formation combines a complete theoretical approach with empirical extensiveness or, in other words, uses a large variety of theories to explain the outcome of coalition formation in a large number of cases. These designs usually apply statistical methods (Warwick 1996; Martin and Stevenson 2001; Mitchell and Nyblade 2008; Döring and Hellström 2013; Bergman, Ersson, and Hellström 2015; Savage 2016).The combination of a parsimonious theoretical model with an extensive empirical design is somewhat less prominent. This strategy is particularly suited for investigating the influence of one, or a few, explanatory factors in a large number of cases. The classic gametheoretical works (Axelrod 1970; de Swaan 1973), and recent studies on the impact of electoral success, ideology, organisational structure, and incumbency on government participation (Mattila and Raunio 2004; Tavits 2008b; Martin and Stevenson 2010; Savage 2014) fall into this category, for instance. Such designs are usually also a domain of statistical methods.

The second common strategy for studying government formation uses (comparative) case studies to provide a comprehensive explanation of government formation in an individual case or a specific country over a certain period of time (Müller and Strøm 2000b; Kropp, Schüttemeyer, and Sturm 2002a; Bergman, Ilonszki, and Müller 2019a).These studies thus combine an intensive empirical design with a complete theoretical approach and use qualitative methods, such as process tracing.Quite often, such studies do not stand alone, but are published in edited volumes that apply a standard analytical framework across all cases in order to provide a cross-national summary of the results.These worksmake aninvaluable contribution to the research field as they generate and compile consistent cross-national data and provide rich explanations of government formation in individual countries. In their cross-national analysis, however, this literature remains predominantly descriptive, leaving the investigation of causal relations and mechanisms to future research. The fourth and final research design in this typology combines an intensive case-oriented empirical design with a narrow theoretical focus, but authors use this strategy relatively rarely (Juul Christiansen and Brun Pedersen 2012).

Thus, research on government formation reflects the gap between qualitative case-oriented and quantitative variable-oriented research that exists in comparative social science researchmore broadly(Ragin 1989; King,Keohane,and Verba 1994; Brady and Collier 2004). Choosing between a qualitative or quantitative approach entails a trade-off between analytical depth and generalisable results (Gerring 2017, chap. 11). With 48 cases where radical right parties could have possibly participated in government, the present study falls into the lower end of extensive empirical designs, which usually use statistical methods. However, the primary goal of this study is to identify different patterns, or configurations, of factors that explain the participation of radical right parties in government. It is thus more interested in the complex interaction of explanatory factors and less in the probabilistic effects of individual variables. For this reason, this study seeks to apply a configurational caseoriented research design despite a medium number of cases (Müller, Bergman, and Strøm 2008, 34).

## **4.2 Analysing government formation with QCA**

This project uses QCA as a method, because it was specifically developed for configurational case-oriented and medium-N research designs. In the subtitle of his seminal introductory volume on QCA, Charles C. Ragin describes his motivation for developing this method as "moving beyond qualitative and quantitative strategies" (Ragin 1989). In the context of ongoing methodological controversies between qualitative and quantitative research in comparative politics and social science (King, Keohane, and Verba 1994; Brady and Collier 2004; Gerring 2017), Ragin attempted to find a middle ground between the generalisability of quantitative research and the thick, case-specific explanations typical of qualitative studies by developing a method for medium-N research (Ragin 1989, vii–ix; see also Ragin 2000, chap. 1). He explicitly intended to integrate "the best features of the case-oriented approach with the best features of the variable-oriented approach" (Ragin 1989, 84).The extent to which he achieved this goal is still a matter of debate, however (Seawright 2005; Jacobs 2009; Rihoux and Lobe 2009; Rihoux 2013).

Despite these methodological discussions, it is true that QCA combines at least some of the advantages of both approaches. As in quantitative methods, it reduces complexity and increases the replicability of research findings by reducing cases to a combination of factors, which can then be subjected to a reproduceable, comparative analysis of causal relationships. As with qualitative methods, QCA examines specific configurations of factors instead of probabilistic effects of individual variables. Therefore, the method is better able to account for causal complexity and the interactions between different explanatory factors in the respective case (Rihoux 2009, 367).<sup>1</sup> This capacity makes QCA the method of choice in this study. Its casebased approach is particularly well suited for evaluating hypotheses which entail complex interactions of various explanatory factors.

When compared to small-N or large-N research, studies with an intermediate number of cases make up only a very small part of social science research (Ragin 2000, 24–26). The possibility of reducing complexity and retaining the configurational character of the individual cases enables QCA to conduct a comparative analysis of complex causal relationships in more than a handful of cases. However, the presence of a medium number of cases is not the only reason for selecting this method. In her study of government participation of radical right parties in Western Europe, which includes a similar number of cases as this study, de Lange (2008) demonstrates that this phenomenon can also be investigated using a primarily quantitative-statistical methodology. Hence, this project's case-oriented approach also favours the selection of QCA because it employs configurational settheoretic methods which are more appropriate than quantitative methods focused on probabilistic effects to test hypotheses in cases which exhibit high levels of causal complexity.

QCA has evolved significantly and become more diverse since its introduction in the late 1980s; therefore, choosing QCA also requires the selection of a specific methodological variant (Rohlfing 2019). Most importantly, it is necessary to determine whether the original crisp-set QCA (csQCA) (Ragin 1989), fuzzy-set QCA

<sup>1</sup> Statistical methods can also account for more complex causal relations, for instance by analysing interaction effects. It has been demonstrated, however, that they reach their limits when interactions of more than two variables are involved (Braumoeller 2003, in Schneider and Wagemann 2012, 297).

(fsQCA) (Ragin 2000, 2008) or multi-value QCA (mvQCA) (Cronqvist 2004; Cronqvist and Berg-Schlosser 2009) applies best to the project's research questions.<sup>2</sup> Crisp set QCA works with dichotomous concepts, meaning that conditions or outcomes can be either present or absent. Fuzzy-set QCA and mvQCA allow for a richer and more fine-grained representation of concepts. Multi-value QCA is not an option in this study because it was developed specifically for multinomial concepts, which do not apply to any of the relevant conditions here. The choice is thus between fuzzy and crisp sets. Schneider and Wagemann (2012, 277) argue that fuzzy sets are preferable to crisp sets whenever possible as they enable researchers to capture the complexity and gradual nature of social reality with a lower loss of information and because they place higher demands on the parameters of fit (see below for further discussion of parameters of fit). Rohlfing (2020) shows, however, that the latter is not always true, but depends on actual empirical observations. He also points out that the choice should be based on "the research interest in set relations between differences in kind (crisp) as opposed to differences in degree (fuzzy)" (Rohlfing 2020, 86). While dichotomous explanatory factors can be integrated into fsQCA without great difficulty, the inclusion of a dichotomous outcome is more problematic (Schneider and Wagemann 2012, 277). As this study examines the participation of radical right parties in government, which is a dichotomous rather than a gradual concept, it opts for csQCA despite the continuous nature of several explanatory factors.<sup>3</sup> Therefore, when this study refers to QCA in general, it refers to csQCA unless explicitly stated otherwise.

## 4.2.1 QCA as a set-theoretic method

QCA gains analytical traction from its ability to analyse causal relationships between necessary and sufficient conditions and the outcome in question. A condition is necessaryif the conditionis a superset of the outcome or,in other words,a set relation of necessity exists "if, whenever the outcome is present, the condition is also present" (Schneider andWagemann 2012, 329).A sufficient condition exists,in turn, "if whenever the condition is present, the outcome is also present"or the condition is a subset of the outcome (Schneider and Wagemann 2012, 333).

<sup>2</sup> The differences in the methodological foundations of the respective methods, such as the difference between crisp and fuzzy sets and the corresponding algebras, will not be discussed here (for a detailed methodological discussion, see Ragin 2000, 2008; Schneider and Wagemann 2012).

<sup>3</sup> It is not impossible to calibrate the outcome as a fuzzy set (Fagerholm 2021), so a supplementary fsQCA has been carried out as part of the robustness check (Schneider and Wagemann 2012, chap. 11.2). This analysis yields similar results to the csQCA, albeit with lower coverage scores (see Appendix I).

Set-theoretic methods are well-suited for analysing complex causal relations, since they are able to reveal, and account for, the fact that an outcome is not always caused by the same condition(s). More precisely, set-theoretic methods are able to capture three aspects of causal complexity (Schneider and Wagemann 2012, 76–82; see also Ragin 1989, chap. 2; Ragin 2000, chap. 4).The first one is precisely the notion that an outcome can have different causes, also referred to as equifinality. Second, they are particularly suitable for studying conjunctural causality, which is the assumption that an interaction between different factors can lead to a particular outcome. And third, they can account for asymmetric causality, which means that either the presence, or the absence, of a certain condition can cause the outcome, depending on the other factors that occur in combination with it.

As indicated in the discussion of the empirical model and the hypotheses, all three aspects of complex causality play an important role in the context of this study. There is, for instance, the assumption that different configurations of conditions explain why radical right parties are included in, or excluded from, government based on different temporal, or regional, contexts. Furthermore, the discussion so far suggests that it is rather unrealistic to believe that one individual factor can adequately explain the government participation of radical right parties. This analysis also expects to find examples of asymmetric causality, at least in some conditions. For instance, both small and large seat shares could lead to government participation of radical right parties depending on other pertinent factors.

INUS conditions are particularly important when investigating complex causality with QCA. The acronym INUS stands for an "insufficient but necessary part of a condition which is itself unnecessary but sufficient for the result" (Mackie 1974, in Schneider and Wagemann 2012, 79). An INUS condition is therefore by itself neither necessary nor sufficient for explaining an outcome, but it is an indispensable part of a combination of factors that together constitute a sufficient condition. Hypotheses 2a provides a good example of INUS conditions (see below). It posits that radical right parties should participate in government if they are on the same side as the formateur in a party system characterised by bipolar opposition and/or if their socioeconomic and socio-cultural positions are proximate to the formateur. Socio-economic and socio-cultural proximity are conceptualised as INUS conditions. The hypothesis proposes that only their joint occurrence should cause the outcome, which implies that individually, each condition is insufficient. Moreover, neither of them, nor their joint occurrence alone, is hypothesised to be a necessary condition for government participation, because, in party systems characterised by bipolar opposition, the outcome could also occur regardless of whether radical right parties' socioeconomic and socio-cultural positions are proximate to the formateur.Thus, according to this hypothesis, neither of the three conditionsmust be presentin all sufficient explanations for the government participation of radical right parties.

## 4.2.2 QCA in three steps

An empirical study with QCA can be divided into three main steps: 1) case selection and model building, 2) data analysis, often referred to as the analytical moment, and 3) interpretation of results (Rihoux and Lobe 2009). Good practice in QCA dictates that the researcher should carry out two separate analyses—one for the outcome and one for its negation (Schneider and Wagemann 2010). This two-staged process results from QCA's sensitivity to asymmetric causal relationships, according to which the absence of a condition does not automatically lead to the non-occurrence of the outcome, when the presence of that condition contributes to its occurrence (Schneider and Wagemann 2012, 81–83). In practice, QCA thus implements the three analytical steps, in particular step two and three, with regard to the outcome and its negation.<sup>4</sup>

## Step 1: Case selection and model building

The first step is to familiarise oneself with the cases, and the theoretical literature, relevant to the research question. A deep understanding is particularly important in exploratory research designs.Research designs based on a diverse case selection also require a high level of familiarity with the cases because the investigator must select the ones that represent as many configurations of conditions as possible (Seawright and Gerring 2008; Gerring 2017). QCA practitioners emphasise that the method is not only a technique for data analysis, but also a specific research approach that requires an intensive dialogue between theories, concepts, and empirical evidence. Hence, developing an analytical model may include rejecting the initial model specifications on the basis of empirical evidence or rethinking the operationalisation of individual explanatory factors (Rihoux and Lobe 2009, 230–33; Ragin 1989, chap. 9; Schneider and Wagemann 2012, 10–12).

Model building in QCA requires paying particular attention to the number of cases and conditions. Crisp-set QCA, in particular, has been criticised for generating explanatory models from random datasets (Marx 2010, 139–41). Marx (2010) shows, however, that this problem only exists if the proportion of cases to conditions is too high. His results indicate that a valid model in a QCA including 14 cases, such as the analysis of the period before the first third-generation elections in this study, should not contain more than five conditions. In an analysis of 34 cases, such as the one related to the consolidating decades, the model can include up to six conditions, which he also identifies as the maximum in small-to-medium-N research (Marx 2010, 149–52, esp. Table 5).Moreover, limiting the number of explanatory conditions ensures that the researcher knows the cases well enough to interpret the re-

<sup>4</sup> For an overview of the methodological foundations and the practical application of QCA, see e.g. Schneider and Wagemann (2012) or Mello (2021).

sults in a meaningful way—particularly the causal mechanisms behind the configuration of conditions in the solution paths (Mello 2021, 30). QCA practitioners have developed different approaches for reducing the number of conditions (Mello 2021, 31–34): For example, based on prior theoretical consideration, researchers can apply the same analytical model throughout the entirety of the analysis.They may also use different analytical models, either in order to test rival theories or to explore the explanatory power of several sets of conditions. If explanatory conditions can be distinguished according to their levels of causal proximity to the outcome, it is possible to conduct a two-step analysis of remote and proximate conditions (see also Schneider and Wagemann 2006; Schneider 2019).This study follows the first approach and develops a single model that will be applied to both periods under investigation.This model will include the most promising theories to explain why Central and Eastern European radical right parties enter government or remain in opposition based on the discussion in Chapter 3.

#### Step 2: Software-assisted analysis of necessary and sufficient conditions

The second step of a QCA is the computer-assisted data analysis. It begins with the creation of a data matrix that summarises the set membership of all cases in all conditions and the outcome. This process is also referred to as the calibration of cases (Schneider andWagemann 2012, 32).Here, the researcher defines the so-called "threshold of indifference" which marks the crossover point between set membership and non-membership (Rihoux and Lobe 2009, 233) and assigns the cases' set membership—or non-membership. This decision can be based on objective facts, theoretical concepts, and/or empirical evidence collected as part of the research process (Schneider and Wagemann 2012, 32). The dataset can then be presented in a truth table*,* where each row displays one of the theoretically possible combinations of the binary conditions. By allocating the cases to the corresponding rows, the truth table also reveals which of the possible configurations of explanatory factors have been empirically observed and which have not (Ragin 1989, 87–89; Rihoux and Lobe 2009, 233–34; Schneider and Wagemann 2012, chap. 4; for an illustration, see Table 9.2).

QCA proceeds with separate analyses of necessary and sufficient conditions with the help of appropriate software, such as fsQCA (Ragin and Davey 2016) or the QCA package for R (Duşa 2019). While the analysis of necessity focuses primarily on the individual conditions, the analysis of sufficiency usually generates several solution paths, each of which can include more than one explanatory condition. When analysing necessary and sufficient conditions, QCA researchers use three parameters of fit, namely the measures of consistency, coverage, and relevance of necessity (RoN) (Ragin 2006; Schneider and Wagemann 2012, 233–38). Consistency indicates the degree to which there is a perfect set relation between the outcome and a necessary or sufficient condition. It thus provides a certain relaxation of the

deterministic nature of csQCA, which enables it to better deal with "noisy social science data" that does not always contain perfect set relations (Schneider and Wagemann 2012, 117). The consistency value can range between 0 (no set relation) and 1 (perfect set relation). In order for a condition to qualify as necessary, its consistency should reach at least a value of 0.9 (Schneider and Wagemann 2012, 143).

The coverage and RoN of necessary conditions determine whether a condition with a sufficiently high consistency score is analytically or trivially necessary (Ragin 2006, 302–3; Schneider and Wagemann 2012, 144–47).The coverage of sufficient conditions measures the proportion of the outcome that is explained by the respective sufficient condition or solution path. In the analysis of sufficient conditions, the aim is always to achieve the highest values for both parameters, but low coverage is less problematic than low consistency. Low coverage only indicates that there are other explanations for the outcome, whereas low consistency points at cases that contradict the theoretical assumptions of the analytical model (Schneider and Wagemann 2012, chaps. 5.2–5.3).

One of the main challenges in applied QCA is the problem of limited diversity. In almost every empirical study, theoretically possible configurations of explanatory factors go unobserved in the actual dataset. In QCA, limited diversity manifests itself in so-called logical remainders, or truth table rows that do not correspond with an empirical case. Schneider andWagemann(2012, 155–75) describe different strategies for dealing with limited diversity. First, the analysis can be limited to only those configurations which represent empirically observed cases; therefore, the outcome of all logical remainders should be coded as 0. The disadvantage of this strategy is that it limits the possibilities for reducing the complexity in the solution. Alternatively, all logical remainders could be coded as 1 and thus be included in the minimisation. The disadvantage of this strategy is obviously that it generates a solution based on a large number of unobserved configurations. The third strategy takes a middle path. Here, the researcher includes a limited number of logical remainders in the analysis based on specific criteria, most importantly so-called directional expectations about the impact of certain conditions. If there are good theoretical reasons to believe that only the presence of a certain condition should relate to the outcome, and there is no initial empirical evidence to suggest otherwise, then only truth table rowsin which this conditionis present should beincludedin theminimisation.

The standard analysis procedure using the fsQCA software (Ragin and Davey 2016) performs three minimisations that generate different solution terms.The first one is the complex, or conservative, solution, which excludes all logical remainders. The second one is the parsimonious solution, which includes all logical remainders that reduce the complexity of the solution term. In a third step, an intermediate solution can be crafted based on the researcher's directional expectations (Ragin 2008, 173–75; Schneider and Wagemann 2012, 175–77; see also Ragin 2018). Schneider and Wagemann (2012, 198–219) point out, however, that the parsimonious solution generated by the software's standard analysis is often based on untenable assumptions about logical remainders. Therefore, they propose the (theory-guided) Enhanced Standard Analysis which modifies the procedure in several ways. They suggest that logical remainders should only be included in the minimisation if they do not contradict previous findings related to necessary conditions and if they do not contain combinations of factors that are impossible in the world as we know it.

Furthermore,it is important to note that all steps of the minimisation,including those leading to the conservative solution, are different ways of expressing the information contained in the truth table. Therefore, Schneider and Wagemann (2012, 107) highlight the researcher's discretion when choosing between different solution terms: "The principle that more than one solution term is an acceptable and logically correct representation of the data in the truth table is a general feature of QCA. The decision on which solution formula to choose as the basis for the substantive interpretation of the available information depends on many research-specific issues that have nothing to do with formal logic."

#### Step 3: Interpretation of results

The interpretation of results involves two main aspects: Relating the necessary and sufficient conditions to the empirical evidence in the individual cases and identifying cross-case patterns that allow for (limited) generalisations (Rihoux and Lobe 2009, 235–37; Schneider and Wagemann 2010, 2012, 280–81). Rihoux and Lobe (2009, 236) describe the case-based interpretation of results in QCA as follows: "each case is a 'black box', and the QCA minimal formula acts like a flashlight which indicates some precise spots to be looked at to better understand the outcome." The interpretation thus goes back to these illuminated spots and makes sense of the configuration in the solution formula.

Depending on the research design, it is possible either to illustrate the mechanisms behind the causal relations indicated by the solution term or to identify additional relevant conditions by looking at deviant cases (see also Schneider and Rohlfing 2013; Beach and Rohlfing 2018). The researcher can also interpret the results beyond individual cases, for instance by assessing a cluster of cases covered by a particular solution path in order to reveal what unites these cases and sets them apart from others. Solution paths can also be compared in terms of their relative weight, based on their individual coverage. Moreover, the interpretation can focus on the role of an individual condition, for instance, if it is present in multiple solution paths. However, the researcher must not ignore the configurational logic of QCA and should thus be careful not to discuss the impact of this condition independently from the configuration of the solution path(s) in which it occurs. Finally, of course, scholars may also make generalisations about the results based on the empirical evidence (Rihoux and Lobe 2009, 236; Schneider and Wagemann 2010).

#### **4.3 Towards an analytical model**

This section integrates the theoretical framework, the discussion of concepts and contexts, and the general remarks on the research design into an analytical model. The main objective is to identify those theories of coalition formation that help to explain the outcome of government formation with radical right parties in Central and Eastern Europe and to draw out the connections between them in a meaningful way. The starting point for this endeavour is the existing knowledge on government formation with radical right parties in Western and, to a more limited extent, also in Central and Eastern Europe.

Research on the government participation of radical right partiesinWestern Europe has shown that, after an initial period of exclusion,mainstream parties came to regard these parties as "normal" coalition partners. Once they lost their pariah status, their participation in government could be explained with the help of office- and policy-related factors.The literature identifies the parliamentary seat share and the ideological proximity of radical right parties to the formateur, particularly regarding socio-cultural issues such as immigration, as decisive factors for explaining how they came to power. Moreover, the increasing polarisation in many Western European party systems, in part a result of the radical right's ascension, has further contributed to these parties' participation in government. In an environment characterised by polarised oppositions between two competing camps, conservative parties were often no longer able to form right-of-centre majority governments without cooperating with electorally successful radical right parties. Further, conservative mainstream parties' rightward shifts reduced their ideological distance from radical right parties, which eased cooperation from a policy-seeking perspective (Bale 2003; de Lange 2008, 2012).

The only study that addresses government formation with radical right partiesin Central and Eastern Europe provides some evidence that radical right parties' electoral successes and ideological preferences impact their inclusion in, or exclusion from, government. At the same time, however, these results highlight the need for further enquiry because they cover only a limited number of instances of government formationin the region(Fagerholm 2021).Themodel of party competition with the radical right in Central and Eastern Europeintroducedin Chapter 1 also suggests that party-level electoral characteristics and policy preferences, as well as structurelevel cultural factors and party system configurations, can affect the participation of radical right parties in government (Minkenberg et al. 2021; see Figure 1.1). Thus, it corresponds with literature on government formation, which also points to the relevance of these factors in Central and Eastern Europe (see Chapter 3, esp. Table 3.1).

The discussion of radical right parties and party systems shows that party competition functions similarly in Western and Central and Eastern Europe. Hence, the basic argument of this study posits that, similar to Western Europe, government formation with radical right parties in Central and Eastern Europe can be explained with the help of office- and policy-oriented theories, as well as the configuration of the party system. Despite this functional equivalence, however, there are certain features of the regional context that need to be taken into account, most notably in the early phase of the post-Communist transformation. This phase, in particular, is affected by the regime change and the triple transition, which entailed large-scale political and economic transformations and the recurrence of state- and nationbuilding. Likewise, party systems were more fluid and less institutionalised while party competition was less programmatic, although most parties developed an ideological core almost immediately after the fall of Communism. For this reason, the basic argument needs to be qualified with regard to three particular features of the early phase of the transformation: First, due to the salience of state- and nationbuilding immediately after 1989, Central and Eastern European radical right parties do not need to undergo a period of normalisation before they are considered as coalition partners. Second, the regime divide provides a powerful source of polarised oppositionin post-Communist party systems.Third,in addition to the sociocultural dimension, the socio-economic one is crucial for issue-based party competition because of the paramount role that economic reforms played in the context of the regime change.

#### 4.3.1 Selecting the most promising explanatory conditions

Based on these preliminary considerations, this section presents the analytical model for studying government formation with radical right parties in Central and Eastern Europe.The model focuses on the characteristics and preferences of radical right parties as well as the configuration of the party system. One of the main challenges here is to limit the total number of conditions in the model because of the configurational approach used in this study and the methodological requirements of QCA.

### Characteristics and ideological preferences of radical right parties

Among the numerical and structural characteristics, the effect of electoral fortunes on radical right parties' participation in government has received substantial empirical support.The most decisive factor has been the size of a party's parliamentary group, referred to here as parliamentary strength. In the context of government formation, the number of legislative seats controlled by a party is more directly related to its participation in government than its vote share.Therefore, research on government formation generally uses the party's parliamentary, rather than its electoral, strength. Electoral gains and losses compared to the previous elections can also influence the outcome of government formation, but the parliamentary seat share has a greater impact on government participation (de Lange 2008, 118–19; Döring and Hellström 2013; Fagerholm 2021). Therefore, the parliamentary strength of radical right parties is the only numerical characteristic at the party level that is included in the analytical model.

Other characteristics of radical right parties, such as their previous political experience, incumbency, and organisational structure, will not be included in the model. While the literature provides no empirical evidence for an impact of a parties' political experience on their participation in government in Central and Eastern Europe (Grotz andWeber 2011, 205–6; see also Savage 2016; Fagerholm 2021: Appendix D3), scholars have found that former ruling parties are at a disadvantage when it comes to government formation opportunities immediately after elections, while incumbency creates an advantage when new governments are formed during the course of a legislative session (Savage 2016, 524–28; see also Roberts 2008; Döring and Hellström 2013; Bergman, Ilonszki, and Müller 2019b). However, since the negative post-electoral incumbency effect is also reflected in the seat share of the radical right, this factor will not be included.The organisational structure of the radical right is excluded for a different reason. Here, missing empirical evidence results from a lack of reliable and comparative data on the internal structures of radical right parties in Central and Eastern Europe, which cannot be obtained in this project either.

The literature discusses the impact of parties'ideological preferences on government formation in Central and Eastern Europe controversially. This study follows those who argue that ideology plays a crucial role in party competition and government formation in this region (Savage 2014; Fagerholm 2021). These works demonstrate that parties' ideological positions affect government formation in Central and Eastern Europe in general, and with radical right parties in particular,if the conceptualisation of the policy space pays attention to the regional context. Therefore, this study applies a two-dimensional conceptualisation of the policy space using separate socio-cultural and socio-economic dimensions instead of the classic, unidimensional left-right dimension.Here,it diverges from Savage (2014) who constructs a single, country-specific left-right dimension based on issue salience. This project prefers the two-dimensional approach for two reasons: First, the socio-economic and socio-cultural divides are aligned in some countries but cross-cutting in others (see Chapter 2.2). Hence, separating the dimensions can account for cross-national variation more effectively than a single dimension. Second, an analysis which uses both the socio-economic and socio-cultural dimensions generates more detailed insights into party competition with radical right parties than studies focused on only one dimension (Spies and Franzmann 2011).<sup>5</sup>

Now that these general issues have been addressed, the discussion turns to selecting concrete ideology-based explanatory factors. Since Fidesz and PiS are not coded as radical right parties (see Chapter 2.1), all of the radical right parties in this study are junior coalition partners. Several studies demonstrate that ideological distance to the formateur is the best predictor of government participation for small parties, including the radical right (Bäck and Dumont 2008; de Lange 2008; Mattila and Raunio 2004). Therefore, the analytical model will include the ideological preferences of radical right parties using their socio-economic and socio-cultural distances to the formateur.<sup>6</sup>

The distance to the formateurismore relevant than distance to the median party, because despite its important ideological position, as a junior coalition partner the median party could lack real bargaining power. In Western Europe, the ideological distance of the radical right to the largest conservative, or Christian democratic, party can also provide information about their chances of participating in government, as these party families are their usual coalition partners (Bale 2003; de Lange 2008, 2012). The potential allies of radical right parties in Central and Eastern Europe, however, can be found in a wide range of party families.

#### Context factors

The configuration of the party system, particularly the existence of bipolar opposition, can help radical right parties enter government (Bale 2003; de Lange 2008, 2012).Therefore, the model will include this factor to account for the ideological configuration of the party systems. Such a "bifurcation" of the party system is also related to the formation of minority governments in Central and Eastern Europe (Keudel-Kaiser 2014, 245–46). In the early transformation phase, bipolar opposition was usually rooted in the regime divide (Grzymała-Busse 2001; see also Beichelt 2001). By including bipolar opposition in the party system, the analytical model thus indirectly addresses the impact of the regime divide as well. Polarisation in the classic Sartiorian sense is not included here, since the empirical evidence does not show that it is an influential factor for predicting radical right participation in Central and Eastern European government coalitions.

To account for the structural-numerical configuration of the party system, the analytical model includes the classic indicator of party system fragmentation. In

<sup>5</sup> In his study on government formation with radical right and radical left parties, Fagerholm (2021, 263) also uses both dimensions, but he analyses only how socio-cultural issues relate to radical right parties and how socio-economic ones affect the radical left.

<sup>6</sup> Radical right parties rarely occupy the median in a one- or multi-dimensional ideological space, so the median party theory does not apply (Laver and Schofield 1998).

Central and Eastern Europe, Grotz and Weber (2011, 202–3) find that fragmentation is a relevant explanatory factor when combined with the seat share of parties in parliament. Moreover, including a factor related to the structural-numerical dimension of party systems addresses party system fluidity in Central and Eastern Europe. Two-party electoral dominance also affects government formation in Central and Eastern Europe (Keudel-Kaiser 2014, 242–43). However, this factor is only present in five of the 48 cases covered by this study, so due to this low diversity, it is not included in the model.

In sum, this section identifies five factors that are most relevant for explaining the government participation of radical right parties in Central and Eastern Europe. The first three, the parliamentary strength of radical right parties as well as their socio-cultural and socio-economic distance to the formateur, refer to the characteristics and ideological preferences of the radical right. The fragmentation and the existence of a bipolar opposition add two context factors at the level of the party system.

### Supplementary analysis of the composition of coalitions

This project seeks to explain why radical right parties enter government or remain in opposition, so the primary unit of analysis is the radical right party, not the government as a whole. Nevertheless, the composition of coalitions will be used as a heuristic tool that contributes to a better, more comprehensive, understanding of government formation with radical right parties in Central and Eastern Europe (de Lange 2012, 903). In order to do so, the analysis draws on established office- and policyoriented, coalition-centred theories of government formation. Among the officeoriented theories, the theory of the minimal winning coalition (Riker [1962] 1984) is particularly useful for explaining the coalition format in Central and Eastern Europe. The analysis will thus take into account whether the government that formed was a minimal winning coalition, a minority government, or an oversized coalition. The assumptions that parties form coalitions with as few parties as possible (Leierson 1968, in de Swaan 1973) or with the lowest possible number of seats needed to control a parliamentary majority (Riker [1962] 1984) have been widely rejected across the continent. Hence, the analysis omits these two theories.

The classic policy-oriented coalition-centred theories of government formation have received limited empirical support in Central and Eastern Europe. This study argues, however, that this is largely the result of misconceptualising the policy space in Central and Eastern Europe.Therefore, it will use the minimal range theory in its open version (de Swaan 1973) and the minimal connected winning coalition theory (Axelrod 1970) in order to assess the ideological range of coalitions with radical right parties. The socio-economic and socio-cultural dimension will be examined separately.

#### 4.3.2 Bringing the temporal dimension in

The last step in developing the analytical model concerns the temporal dimension. Time can affect political processes in many different ways, such as the duration and speed of processes or the timing of events (Grzymała-Busse 2011).The present study argues that the patterns of government formation change with growing temporal distance to the fall of Communism, because certain characteristics of the immediate post-Communist period, such as the fluidity of party systems or the impact of the regime divide, are in decline. Thus, temporality refers here to the duration of these processes of change. There are signs that Central and Eastern European party systems began stabilising around the turn of the millennium (Ágh 1998, 109–12; Toole 2000; Emanuele, Chiaramonte, and Soare 2020, 317). At that time, the influence of the regime divide on party politics was also declining in most countries.

Yet, instead of using the year 2000 as a fixed temporal threshold, the periodisation is based on the qualitative threshold of the first third-generation elections. Third-generation elections mark the point when two competing ideological camps, mostly parties from both sides of the regime divide, have governed a country (Pop-Eleches 2010, 236–38).<sup>7</sup> Pop-Eleches (2010) introduces the concept of election generations in an analysis of protest voting and the emergence of "unorthodox parties" in Central and Eastern Europe. He argues that participating in government results in voter dissatisfaction with ruling parties and, consequently, electoral losses that open a window of opportunity for new challengers. While this specific argument is of minor importance here, the underlying assumptions about structured party competition are relevant. Only after the first third-generation elections, the electorate is able to base their ballot decision on the actual track record of political parties and leaders from different camps and not merely on promises in manifestoes or electoral campaigns.The same applies to political parties, who also have a much better understanding about the behaviour and policy preferences of a competitor that has already participated in government. Hence, from the point of view of analysing structured party competition, the first third-generation elections mark a crucial stage of party system stabilisation.

In the countries covered by this study, the first third-generation elections were held around the year 2000 (see Table 4.1).Thus, the periodisation based on the qualitative threshold of the first third-generation elections reflects country-specific trajectories in the developments of party competition. At the same time,it ensures that the periodisation in all countries is similar enough to still enable a cross-national comparison. Moreover, the first third-generation elections also (roughly) coincide

<sup>7</sup> The discussion of the configuration of the party systems in the case studies in Chapters 5 and 6 provides further details on the divisions and competing camps in the respective countries.

with the establishment of democracy and a market economy, at least at the procedural level, as well as the introduction of formal accession negotiations with the EU (Beichelt 2001; Vachudova 2005).<sup>8</sup>


*Table 4.1: Election generations in Central and Eastern Europe*

Source: Pop-Eleches 2010, 234.

After critics initially stressed QCA's limited capacity to address temporality, QCA practitioners developed various procedures to include temporal sequencing into their analyses (Caren and Panofsky 2005; Ragin and Strand 2008; Mahoney, Kimball, and Koivu 2009; Rubinson 2019; see also Schneider and Wagemann 2012). This study, however, seeks to explain government formation with radical right parties in two different periods of time, rather than the temporal sequence of different explanatory factors. When incorporating this aspect of temporality into QCA, some scholars introduce an additional time period condition. As the present study covers two periods, this option would be compatible with the binary logic of csQCA. However, since the number of truth table rows increases exponentially with each additional condition, this strategy would also increase the number of logical remainders. In order to avoid this problem, the strategy employed here is to

<sup>8</sup> The actual date of EU accession could define a temporal threshold as well. While EU membership was a hallmark in the political development of Central and Eastern European countries, the criterion of the first third-generation election better reflects the changes in party competition most relevant to this study. The discussion suggests that crucial changes already occurred in the pre-accession period. The supplementary analysis and robustness checks, which use EU membership as temporal threshold (see Appendix I), support this interpretation.

conduct two separate analyses for the respective periods and to subsequently compare their results (Schneider and Wagemann 2012, 265–66). Dividing the dataset will also lead to an increase in logical remainders, particularly in the shorter period before the first third-generation elections, but this increase is much smaller than the exponential one caused by an additional condition.

#### 4.3.3 The analytical model

Figure 4.1 summarises these specifications in a graphical model. The upper section of the model contains the five conditions that shall explain why Central and Eastern European radical right parties enter government or remain in opposition. The two separate arrows illustrate the periodisation based on the criterion of the first thirdgeneration elections, which reflects the assumption that the patterns of government formation with radical right parties are expected to differ in the period before and after that threshold. The bottom section of the model adds the composition of the government as a whole. The dashed arrow that points towards the outcome denotes that the composition of coalitions is used as a heuristic and supplementary tool to better understand why radical right parties are included in, or excluded from government, rather than serving as an explanatory factor or an outcome to be explained itself.

The individual explanatory conditions in the model are connected by the overarching assumptions that government formation is a result of a) the strategic choices that parties make in their pursuit of policy and office, and b) the contextual constraints which limit those parties' decisions. Hence, the outcome of government formation hinges on the interaction of all the different conditions at the level of the individual parties and the party systems.

*Figure 4.1: Analytical model of government formation with radical right parties in Central and Eastern Europe*

Source: Own composition.

## 4.3.4 Hypotheses

Before advancing some hypotheses recall that this study combines theory-testing and theory-generating approaches (Gerring 2017, 263–70).On the one hand,it seeks to evaluate existing coalition theories in the specific context of government formation with radical right parties in Central and Eastern Europe. On the other hand, this study combines various theories in a context-specific analytical model. The exploratory nature of this approach aims at generating original theoretical insights, which is also reflected in the following hypotheses.

The first set of hypotheses concerns the numerical-structural factors—the parliamentary strength of radical right parties and the fragmentation of party systems. Radical right parties in Central and Eastern Europe were, on average, less successful in elections than in Western Europe (Minkenberg 2002, 336, 2017, 101; Mudde 2005a). Over the last three decades, however, they have improved their average electoral results, particularly in the 2010s.<sup>9</sup> With the exception of Romania, radical right parties entered parliament in all countries covered by this study in this decade. The party systems of Central and Eastern Europe are more fluid than in Western Europe, even though there is a trend towards convergence, particularly with regard to fragmentation (Enyedi and Casal Bértoa 2018; see also Emanuele, Chiaramonte, and Soare 2020).

The literature finds contradictory results and diverging theoretical arguments regarding how party system fragmentation and parliamentary strength affect radical right parties' participationin government. Forinstance, some scholars argue that parties benefit from gaining a large seat share becauseitincreases their contribution to the government's majority in parliament and thus their bargaining power (Mattila and Raunio 2004; Döring and Hellström 2013). However, other studies also point out that the seat share of small parties should not be too large, because when junior partners are strong, the formateur must yield power to them (Warwick 1996, 499; see also Fagerholm 2021). A high degree of party system fragmentation increases the complexity of the bargaining situation (Dodd 1976; Kropp, Schüttemeyer, and Sturm 2002b; Müller, Bergman, and Strøm 2008), because more parties are needed to form a majority. Thus, high levels of fragmentation improve the chances for small parties to enter government. Since parties aim to reduce bargaining complexity by limiting the number of coalition partners while still ensuring a stable majority, however, small parties with a relatively large seat share should have an advantage when party systems are fragmented (Warwick 1996, 495; Grotz and Weber 2011, 202–3).This discussion illustrates the interplay between the two structural-numerical factors in the model, and it suggests that several configurations of the two factors enable radical right parties to enter government in Central and Eastern Europe.There is one, however, that should create a clear disadvantage—a small seat share in a compact party system (see Table 4.2). Moreover, the development of party systems and election results of radical right parties in Central and Eastern Europe suggests differences in the two periods under investigation. Hence, the following hypotheses regarding the impact of the two structural-numerical factors shall be evaluated:

*H1a: Radical right parties that are large, and/or in fragmented party systems, enter government, but radical right parties with a small seat share in a compact party system remain in opposition.*

<sup>9</sup> The average vote share for radical right parties in the elections examined in this study is 7.9 per cent in the 1990s, 10.3 per cent in the 2000s and 10.8 per cent in the 2010s. These numbers include only those radical right parties that passed the threshold of representation in the respective country.

*H1b: Inthe period beforethe firstthird-generation elections, predominantly small radical right parties in fragmented party systems enter government.*

*H1c: In the period after the first third-generation elections, predominantly large radical right parties in compact party systems enter government.*

*Table 4.2: Theoretical expectations about the impact of seat share and fragmentation on government participation*


Source: Own compilation.

The second set of hypotheses concerns the socio-economic and socio-cultural distance between radical right parties and the formateur as well as the existence of bipolar opposition in the party system. The literature review demonstrates that parties prefer to form governments with partners that hold similar ideological positions (Axelrod 1970; Swaan 1973; see also Laver and Schofield 1998; Savage 2014). Western European radical right parties enter government when their socio-cultural distance to the formateur is small and bipolar opposition in the party system constrains coalition formation (Bale 2003; de Lange 2008, 2012). There are several features of the Central and Eastern European context, however, which demand adaptions of these findings and their theoretical underpinnings. First, the importance of the economic transformation, and the social hardships that it caused for large parts of the population in Central and Eastern Europe, make it impossible to disregard the socio-economic dimension from an analysis of government formation in the region, particularly in the period prior to the first third-generation elections. Second, this transformational period is strongly affected by the regime divide, which confined the choice of possible coalition partners to the parties within their respective camps (Grzymała-Busse 2001). Hence, the empirical analysis assesses the following hypothesis regarding the impact of ideological factors on government formation with radical right parties in the transformational period:

*H2a: Before the first third-generation elections, radical right parties enter government if their socio-cultural and socio-economic distance to the formateur is small and/or they are situated on the same side of a bipolar opposition as the formateur.*

In the consolidating decades, however, the salience of both the regime divide and the transformation of the economic system should decrease, and the patterns of government formation with radical right parties in Central and Eastern Europe should resemble those in the western part of the continent more closely. Hence, the following hypothesis will be evaluated in this period:

*H2b: After the first third-generation elections, radical right parties enter government if their socio-cultural distance to the formateur is small and/or they are situated on the same side of a bipolar opposition as the formateur.*

Further hypotheses can be formulated regarding the composition of coalitions with radical right parties. Because the composition of coalitions is not the primary subject of this study, these hypotheses remain descriptive.Althoughmainstream parties in Central and Eastern Europe frequently adopted elements of radical right politics, some parties—and parts of society—are still critical towards governing with radical right parties. Therefore, the starting point for these hypotheses is the assumption that mainstream parties prefer moderate coalition partners over the radical right. In order to prevent conflicts that may arise from cooperating with the radical right, mainstream parties should not govern with radical right parties unless their participation is required to secure a majority. Supporting a minority government can be particularly advantageous for radical right parties, since keeping "one foot in and one foot out of government" (Zaslove 2012, 435) enables them to influence government policies while simultaneously upholding their oppositional appeal (see also Albertazzi and McDonnell 2005; Dumont, de Winter, and Andeweg 2011, 9–10).This constellation also makes it possible for mainstream parties to distance themselves from the controversial positions of their radical right support parties. Hence, the first hypothesis regarding the numerical format of coalitions is as follows:

*H3a: Radical right parties are included in government as junior partners in minimal winning coalitions or as support parties for minority governments, but they do not participate in oversized coalitions.*

Mainstream parties should certainly not govern with the radical right if they disagree on socio-cultural policies because these are the radical right'smost salient core issues. Therefore, socio-cultural differences are particularly apt to cause conflicts within governments that include radical right parties. Hence, the following hypothesis guides the supplementary investigation of the ideological range of coalitions with radical right parties:

#### *H3b: Governments with radical right parties are socio-culturally homogeneous.*

Similar to the above argument regarding the ideological factors at the party level, governments with radical right parties should also be socio-economically homogenous in the period before the first third-generation elections. However, since there is no suitable comparative data on party positions available for the transformational period (see below), this hypothesis cannot be evaluated in this study.

## **4.4 Operationalisation and measurement**

After outlining the analytical model and presenting the hypotheses, this section turns to the operationalisation of the individual factors in the analytical model.

#### 4.4.1 The outcome: What counts as a coalition government?

Before defining the outcome, it needs to be re-stated that the present study is concerned primarily with the formation of multi-party governments. Sometimes, the electorate equips a single party with an absolute majority of seats in parliament. Such parties are in the position to form a single-party majority government. They can fill the cabinet exclusively with their own representatives and do not depend on other parties' support in a vote of (no) confidence in parliament. Hence, the dynamics of government formation in majority situations are fundamentally different from thosein which no single party controls amajorityin parliament andinter-party cooperation is required (Müller, Bergman, and Strøm 2008, 7; Mattila and Raunio 2004, 278).

Multi-party coalition governments include a cabinet comprised of several ministers which constitutes the country's top executive body as well as parliamentary groups of political parties that support this cabinet with a legislative majority. Alternative definitions emphasise particular aspects of government coalitions.Müller, Bergman, and Strøm (2008, 6), for instance, prioritise the government's executive branch and define the government or, more precisely the cabinet, as "the sharing of executive office by different political parties" and a "coalition party" as "a party that has at least one designated representative that enjoys voting rights in the country's top executive policymaking body". Like Dodd (1976), this project is more interested in the partisan composition of the government coalition than the cabinet.Therefore, it defines a coalition government as formalised cooperation between legislative parties for the purpose of sharing executive power.

In empirical research on government formation, it is also important to define what constitutes a new (coalition) government. Therefore, such counting rules receive more attention in the literature than the definition itself. In this study, a new government is formed a) after an election takes place, b) when the partisan composition of the government has changed, or c) when the party of the prime minister alternates (Müller and Strøm 2000a, 12–13). Given this study's primary interest in the partisan composition of the government, a change of the prime minister alone does not account for a change of government (Müller and Strøm 2000a, 12; Müller, Bergman, and Strøm 2008, 6).

Even though these rules seem quite clear and straightforward to implement, there are certain challenges that need to be addressed. First, parties often form electoral alliances in order to improve their chances of entering parliament (Müller, Bergman, and Ilonszki 2019, 17).<sup>10</sup> These alliances can entail commitments for future cooperation in a joint parliamentary group or (coalition) government. It is sometimes difficult to determine whether or not electoral alliances should be considered a single entity in post-electoral coalition formation. The Bulgarian Ataka and the Latvian National Alliance (NA), for instance, competed in their first national elections in 2005 and 2010, respectively, as electoral alliances. Both alliances, however, formed a joint parliamentary group and eventually merged into a full-fledged political party a few months after their election into parliament. Therefore, they will be treated as political parties from the outset. The situation is more ambiguous in the cases of the Patriotic Front (PF) and United Patriots (UP) in Bulgaria. In the run-up to the 2014 elections, the PF was created as an electoral alliance including the Internal Macedonian Revolutionary Organisation (VRMO), the National Front for the Salvation of Bulgaria (NFSB), and some peripheral parties and organisations. Prior to the 2017 parliamentary elections, Ataka joined this alliance, which was subsequently re-named UP. Even though these alliances show lower levels of internal cohesion than the previous examples, they formed common parliamentary groups and jointly engaged in coalition negotiations. For this reason, they will also be treated as single entities in coalition formation, even though this stretches the unitary actor assumption (see Chapter 3.1).

The second challenge concerns the treatment of minority governments. Minority governments are a relatively frequent phenomenon, and they have proven to be viable alternatives to majority governments in European democracies (Strøm 1990b; Keudel-Kaiser 2014). The party-centred definition of coalition governments introduced above can also include support parties of minority governments if cooperation with the governing party (or parties) is formalised. This study follows those scholars who argue that parties which consistently support a minority government

<sup>10</sup> Several countries have introduced higher electoral thresholds for electoral alliances which often increase with the number of parties included in the alliance.

may be considered part of the government, since they exert strategic influence on its activities and agenda well beyond individual policy proposals (de Lange 2008; see also Zaslove 2012; Fagerholm 2021). De Lange (2008, 41) introduces two criteria for formal support parties of a minority government. First, their support needs to be based on a mutual agreement and, second, the support relationship between the parties must be publicly known. Although formal support parties are considered part of the government coalition under these circumstances, there is still a substantial difference between support parties and actual coalition members—support parties do not receive cabinet posts.Therefore, a new government will be counted if a support party formally enters the ruling coalition or a coalition party resigns from the cabinet but continues as an official support party, even though the partisan composition of the government does not change in either situation.

The last issue to be addressed are so-called caretaker governments. Caretaker governments are provisional governing bodies which generally serve for a short period of time before (early) elections touched off by a government crisis. They often entail technocratic cabinets agreed upon by the majority of parties in parliament. Conventionally, caretaker governments do not make substantial policy decisions (Conrad and Golder 2010). These constraints illustrate that the power of caretaker governments is limited and that their formation does not involve the same trade-offs between parties' pursuit of policy and office, which drive regular government formation. Hence, caretaker governments will be excluded from the analysis of government formation in this study.

## 4.4.2 Party-centred conditions and party-system features

### Parliamentary strength

Compared to the outcome, the operationalisation of the parliamentary strength of radical right parties is rather straightforward. Parliamentary strength is measured by the percentage of seats in parliament. In countries with bicameral parliaments, the distribution of seats in the lower chamber will be used.The data on the distribution of seats is drawn from the database on Parties and Elections in Europe (Nordsieck 2021).

#### Ideological distance to the formateur

The analytical model includes the ideological distance between radical right parties and the formateur of a coalition on the socio-economic and socio-cultural dimensions. Given the problems associated with conceptualising and adequately measuring ideological positions in Central and Eastern Europe, operationalising this condition requires special attention. In modern comparative politics, two approaches have become predominant when measuring party positions: Expert surveys (Benoit and Laver 2006; Jolly et al. 2022) and the standardised analysis of party programmes by the Manifesto Project (Volkens et al. 2021).<sup>11</sup>

Marks et al. (2007) present a comprehensive discussion of the advantages and disadvantages of both approaches. The authors argue that expert surveys benefit from the ability to draw on multiple sources of information, such as programmatic documents, interviews with party members, and parties' behaviour, which scholars can use to assess party positions. Expert surveys also provide easily quantifiable and comparable data. They are disadvantaged, however, by the subjectivity of the experts' judgements, asymmetrical information regarding the individual parties, and the limited availability of data, particularly before the turn of the millennium. The Manifesto Project uses primary party documents and thus clearly distinguishes between what parties say and how they behave. This approach also enables the Manifesto Project to provide data on party positions retrospectively. However, the drawback of the Manifesto data is that programmes are strategic documents in which parties emphasise, or conceal, certain positions for tactical reasons (Marks et al. 2007, 26–27). In addition, the Manifesto Project sometimes captures the salience of a given issue dimension rather than the party's position on that issue (Kitschelt 2007, 1180).

The fact that party programmes are strategic documents is particularly important to this study. First, in order to attract a broader electorate and to avoid legal prosecution, radical right parties are known to downplay their ultra-nationalist, racist, and anti-democratic positions in official proclamations (Kitschelt 2007, 1180). It is therefore important to go beyond official programmatic documents when evaluating their ideological positions (Pytlas and Kossack 2015, 109). Second, political parties in Central and Eastern Europe often lack detailed programmatic documents, particularly in the early years of the post-Communist transformation. For these reasons, this study, like others in the field (Pirro 2016; Pytlas 2016; Minkenberg et al. 2021), draws on expert surveys to determine the positions of radical right parties and formateurs. Following Kitschelt's (2007, 1081) recommendation, the quantitative expert survey data will be supplemented by a qualitative assessment of party positions based on secondary literature.

The Chapel Hill Expert Survey (CHES) is the most comprehensive expert survey on party positions. In Central and Eastern Europe, the CHES dataset covers the period between 2002 and 2019 in five waves (Jolly et al. 2022). In addition to positions on various issue dimensions, it also contains indicators that measure the socio-economic (LRECON) and the socio-cultural positions of the parties (GALTAN).<sup>12</sup> These

<sup>11</sup> For a discussion of other approaches to measuring the ideology of radical right parties and their advantages and disadvantages, see Mudde (2007, 33–41) and Kitschelt (2007).

<sup>12</sup> In his study on government formation with radical left and right parties across Europe, Fagerholm (2021, 9, Appendix C) calculated indicators for the socio-economic and socio-cultural

indicators correspond with the two-dimensional conception of party competition applied in this study and will therefore be used here. A problem with the CHES data, and expert surveys more generally, however, is the lack of data on Central and Eastern European parties prior to the turn of the millennium. In this period, the socioeconomic and socio-cultural positions of radical right parties and formateurs will be assessed on the basis of secondary literature (Keudel-Kaiser 2014, 49–50). In order for these positions to be comparable, the study places parties from the pre-survey era on the same 11-point GALTAN and LRECON scales as the CHES.

## Party system fragmentation

When evaluating the fragmentation of party systems, Sartori (1976, 121–27) opted to simply count all relevant parties, but a party was only considered relevant if it had either coalition or blackmail potential. Building on Sartori, Laakso and Taagepera (1979) presented their calculation for the effective number of parties, which adds the relative strength of the parties according to their share of votes or parliamentary seats. The effective number of parties has become the standard measure of party system fragmentation in contemporary research. Therefore, this study uses the effective number of parliamentary parties (ENPP), to measure fragmentation. Unlike the effective number of electoral parties, which includes all parties running for an election, this measure is limited to parties that achieve parliamentary representation, which is a precondition for their access to government formation.The ENPP is calculated as follows:

"ENPP=1/Σsi²,

where s<sup>i</sup> is the proportion of seats of the ith party" (Casal Bértoa 2013, 401). For most of the cases covered by this study, the effective number of parties is taken from the database on Who Governs in Europe (Casal Bértoa 2021).

### Bipolar opposition

Bipolar opposition exists "when party competition as a whole is structured along a specific and deep dividing line" (Keudel-Kaiser 2014, 70), which separates the political parties of a country into two competing camps that are unable to cooperate with each other. Whether or not bipolar opposition constrains government formation to

dimensions, respectively, based on specific items in the Manifesto dataset. While this approach mitigates the problems associated with the Manifesto Project's over-emphasis of the socio-economic dimension in its general left-right indicator (Savage 2014, 550; Pytlas 2016, 74), it does not solve the problems of the limited availability of detailed party programmes in Central and Eastern Europe in the early 1990s and the tendency of radical right parties to downplay their radical ideology in programmatic documents.

coalitions within these camps can only be determined by a qualitative assessment of the main lines of conflict in each party system.

Research has demonstrated that bipolar oppositions have different causes. A distinction can be made between affective and ideological polarisation (Nugent 2020; see also Iyengar et al. 2019; Gidron, Adams, and Horne 2020). Ideological polarisation results from fundamentally different views on policies. It thus refers to "the extent to which they [parties] disagree with each other" (Nugent 2020, 3). The concept of affective polarisation has gained popularity in the context of the United States' polarised two-party system, but it provides a useful tool for the analysis of party competition in European multi-party systems as well (Gidron, Adams, and Horne 2020; Wagner 2021). In contrast to ideological polarisation, the affective dimension concerns partisan identities and expresses "the extent to which groups dislike each other" (Nugent 2020, 3). The literature on affective polarisation is primarily concerned with voters' dislike for other parties, but some authors have begun to show how this antipathy is also connected to the relations among political elites (Banda and Cluverius 2018; Gidron, Adams, and Horne 2020). Because the present study deals with inter-party competition rather than party-voter relations, it applies the general ideas of affective and ideological polarisation to the elite level. In order to determine the intensity and nature of bipolar opposition in the party system, this study draws on the party system literature, and information about national election campaigns (Keudel-Kaiser 2014).

#### 4.4.3 Format and ideological range of the government coalition

The format of coalitions with radical right parties is also determined by an analysis of secondary literature, such as the EJPR Political Data Yearbook. When comparing the format of governments with radical right parties to the format of governments without the radical right, the dataset on coalition politics compiled by Bergman, Ilonszki, and Müller (2019a) and their team will be used.This dataset includes a distinction between minimal winning coalitions,minority governments, and oversized coalitions.

Despite being very comprehensive, this dataset does not include information on the ideological range of Central and Eastern European governments. Hence, this study draws on the measures of parliamentary strength and ideological positions in order to assess whether the coalitions meet the criteria of a minimal connected winning or an open minimal range coalition. Because a qualitative assessment of the positions of all government parties is beyond the scope of this study, the evaluation of the ideological range of coalitions is limited to the period covered by the CHES data, which largely corresponds to the period after the first third-generation elections.

## **4.5 The structure of the empirical analysis**

The empirical analysis follows the three steps of QCA. Chapters 2 – 4 already cover much of the first step, which entails the dialogue between theoretical knowledge and empirical evidence typical of QCA. For the sake of a coherent presentation, the discussions in these chapters did not always reflect the iterative nature of this process. The descriptive case studies in Chapters 5 and 6 add the last piece of the first step. They provide a description of the cases, focusing on the explanatory factors in the analytical model. The second step of the analysis begins in Chapter 7, which discusses the calibration of cases and generates the binary dataset required for the analysis of necessary and sufficient conditions. The following two chapters analyse this data and interpret the results. Chapter 8 deals with the transformational period and Chapter 9 discusses the consolidating decades. The results from both periods are then compared as part of the conclusions in Chapter 10.

## **5. Government formation with radical right parties in Central Europe: The Visegrad Four**

This chapter introduces the cases from the four Central European countries covered by this study, the Czech Republic, Hungary, Poland, and Slovakia. All country case studies follow the same structure: First, they describe the formation and composition of governments when radical right parties were present in parliament. Second, the case studies turn to party system fragmentation and bipolar opposition. The third and final section in each country report covers the parliamentary strength of radical right parties and their ideological proximity to the formateur on the socioeconomic and socio-cultural dimensions.

## **5.1 Czech Republic**

### 5.1.1 Government formation with radical right parties in the Czech Republic

In the Czech Republic, the first radical right party entered parliament in 1992. However, the Rally for the Republic – Republican Party of Czechoslovakia (SPR-RSČ) was only able to survive for a few years. In the early elections of 1998, the party failed to pass the electoral threshold and quickly disappeared from the political scene thereafter. The SPR-RSČ was ostracised by all other parties in parliament and even regarded itself as a fundamental opposition to the system with no intention of participating in government (Čakl andWollmann 2005, 48; Minkenberg and Kossack 2015, 351; Minkenberg 2017, 106).

Government formation after the 1992 elections was strongly influenced by the negotiations over the dissolution of the Czechoslovak Federation. The clear winner of the elections, Václav Klaus' Civic Democratic Party (ODS) formed a coalition with three other conservative parties, the Civic Democratic Alliance (ODA), the Christian and Democratic Union – Czechoslovak People's Party (KDU-ČSL) and the Christian Democratic Party (KDS), with whom it had run in an electoral alliance (Grotz 2000, 349).The ODS also won the following parliamentary elections in 1996, albeit by a thin margin over the Czech Social Democratic Party (ČSSD). However, the incumbent coalition no longer controlled a majority in parliament, which is why the ČSSD was alsoincludedin the negotiations. In the end,ODS,KDU-ČSL,andODA formed aminority coalition that was tolerated by the ČSSD. In return for its support, the ČSSD received the post of parliamentary speaker and other offices (Grotz 2000, 367–68; Novák 2003, 154–55).<sup>1</sup>

After the early termination of the Klaus II cabinet in 1998 and the electoral demise of the SPR-RSČ, it took 15 years for another radical right party to enter the Czech parliament. In the 2013 parliamentary elections, Dawn of Direct Democracy (Úsvit), which had just been established prior to the elections by political newcomer Tamio Okamura, passed the electoral threshold.The victorious ČSSD, however, was not interested in cooperating with Úsvit and instead formed a coalition with the Christian democrats and the populist anti-establishment party ANO 2011 (ANO) (Hloušek and Kaniok 2014, 12). Úsvit's participation in government was out of the question for all other parliamentary parties, and even Okamura himself, whose anti-establishment stance had been instrumental to his party's success, showed no interest in entering coalition negotiations (Hloušek, Kopeček, and Vodová 2020, 132).

In 2017, ANO won the parliamentary elections and its founder, and leader, Andrej Babiš became the formateur of the next government. Neither Okamura nor Babiš ruled out cooperation between ANO and Okamura's new radical right party, Freedom and Direct Democracy (SPD).<sup>2</sup> They even held exploratory talks, thus cutting the cordon sanitaire against radical right parties that had existed in the Czech Republic ever since the fall of the Iron Curtain. These talks did not result in SPD's participation in government, though. Eventually, ANO and the ČSSD formed a minority coalition, which enjoyed the formal support of the Communist Party of Bohemia and Moravia (KSČM). Thus, the SPD remained in opposition, although there was occasional cooperation with the governing parties (Hloušek, Kopeček, and Vodová 2020, 158–61). The ANO-ČSSD minority government thus brought a major change to Czech politics, as it marked the end of the cordon sanitaire visà-vis both the radical right as well as the KSČM.

<sup>1</sup> The KDS formally merged into the ODS in the run-up to the 1996 parliamentary elections (Grotz 2000, 360).

<sup>2</sup> Due to massive internal conflicts between Okamura and other Úsvit parliamentarians shortly after the 2013 elections, a "coup" finally took place in the 9-member party assembly. Okamura's opponents decided to found a new party without him. After being virtually expelled from his own party, Okamura launched the SPD (Hloušek, Kopeček, and Vodová 2020).


*Table5.1:FormatandideologicalrangeofgovernmentsintheCzechRepublic*

Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in parenthesesindicatethesmallestpossibleideologicaldistanceofamajoritygovernment

Table 5.1 provides an overview of the governments that were formed while radical right parties held parliamentary seats in the Czech Republic. As regards their numerical format, two of the four governments were minimal winning coalitions and minority governments, respectively. Data on the ideological range is only available for the two most recent coalitions. Neither of them meets the criteria for the open minimal range or the minimal connected winning theory, since socio-economically and socio-culturally more homogeneous coalitions would have been possible.<sup>3</sup> In the case of the Babiš government, however, it should be noted that it is considerably more homogeneous when the support party KSČM is taken out of the equation. Overall, the instances of coalition formation observed here reflect a general tendency to form ideologically heterogeneous "rainbow coalitions" in the Czech Republic (Mansfeldová and Lacina 2019, 145–46).

## 5.1.2 The configuration of the Czech party system

## Fragmentation

The Czech national assembly elected in 1992 consisted of 4.8 effective and eight actual parties and was thus quite fragmented (see Table 5.2). Even an ideologically incompatible coalition of ODS and KSČM, the two largest parliamentary groups, would have controlled only a very slim majority. All realistic majority alliances required the cooperation of at least three parties. In 1996, the fragmentation had decreased considerably, but the formation of coalitions remained complex. Again, only a coalition of the two largest parties, ODS and ČSSD, would have permitted a twoparty majority government.


*Table 5.2: Fragmentation of the Czech party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

<sup>3</sup> The Babiš government does not fulfil the majority criterion either, which is required by both theories.

During the following decade, which was marked by the absence of radical right parties, the Czech party system developed into one of the most compact ones in Central and Eastern Europe. The average effective number of parties in the 2000s was only 3.4 (Enyedi and Casal Bértoa 2018, 442). From 2010 onwards, however, there was significant dealignment, characterised not only by a marked increase in fragmentation but also by one of the highest levels of volatility in Central and Eastern Europe during this time period (Enyedi and Casal Bértoa 2018). After the 2013 parliamentary election, the effective number of parliamentary parties peaked at 6.1. Here, no party managed to control more than a quarter of the seats in parliament. Hence, at least three parties were needed to form a majority government. Despite the complexity of this bargaining situation and the limitations from the cordon sanitaire vis-à-vis the radical right Úsvit and the Communist successor party, KSČM, government formation ultimately went rather smoothly, resulting in the above-mentioned three-party majority coalition.

After the 2017 parliamentary elections, nine parties entered the Czech parliament. Fragmentation remained relatively high, and government formation continued to be rather complex.The decrease in the effective number of parties, from 6.1 to 4.8, resulted mainly from ANO's strong position. The party controlled 78 of the 200 seats, more than three times as many as the ODS which came in second. Mathematically, only these parties were large enough to successfully form a two-party majority government. The other seven parliamentary groups were so small that ANO would have needed at least two of them to reach amajority.Amajority against ANO,in turn, would have required the cooperation of at least six of the other eight parliamentary parties.

#### Bipolar opposition

Soon after the founding elections in 1990, socio-economic divides were most salient in Czech politics. The main contenders were the liberal-conservative ODS on one side and the social democratic ČSSD on the other. The Christian democratic KDU-ČSL stood between the two parties and was coalitionable in both directions (Vodička 2005, 147; Cabada, Hloušek, and Jurek 2014, 93; Mansfeldová and Lacina 2019). In addition to the socio-economic divide, however, territorial issues, such as the status of the Czechoslovak federation and aspirations for Moravian autonomy, played an important role in 1992. Furthermore, the interpretation of the Communist past was still on the agenda. ODS leader Václav Klaus, for instance, believed that even the ČSSD, and some of his former companions who criticised his neoliberal policies, were too comfortable with the former regime (Grotz 2000, 327–29; Balík and Hloušek 2016, 105–6). Most Czech parties supported the existence of the Czechoslovak federation, and with the exception of the SPR-RSČ and the KSČM, there was also widespread agreement on the general path toward the country's Western integration (Keudel-Kaiser 2014, 92–93). Since parties' positions on these divides were

not aligned with those on the socio-economic one, there was no clear-cut bipolar opposition in the Czech party system in 1992 (Balík and Hloušek 2016, 110; Kitschelt et al. 1999, 226–30).

By the 1996 parliamentary elections, the socio-economic conflict had deepened, undergoing qualitative changes. The debate over the transformation of the economic system was increasingly sidelined by distributional conflicts (Keudel-Kaiser 2014, 92; Balík and Hloušek 2016, 110; Mansfeldová and Lacina 2019, 133). The balance of power between ODS and ČSSD became more equal due to the latter's increasing popularity. Already in 1996, the conflicting socio-economic positions made a joint government composed of these two parties hardly conceivable (Novák 2003, 154; Keudel-Kaiser 2014, 94). Some of the other divides that were relevant in 1992 still played their part in 1996. The regime divide, for instance, lost intensity but did not cease to exist. All parties had agreed on a cordon sanitaire vis-à-vis the KSČM, which constituted a serious constraint on coalition formation (Grzymała-Busse 2001, 97–98; Vodička 2005, 144; Mansfeldová and Lacina 2019, 132).The divide between the liberal democratic mainstream as well as the KSČM and the radical right SPR-RSČ also remained intact. Hence, the bipolar opposition between ODS and ČSSD was much more decisive in 1996 than in 1992, but there were still too many relevant divides to speak of a bipolar opposition between two camps. In the following decade, when no radical right party was present in the parliament, Czech politics remained dominated by socio-economic issues and the opposition between ODS and ČSSD. However, some secondary issues always remained salient and prevented a clear-cut bipolar opposition from emerging (Balík and Hloušek 2016, 108–9).

In the 2010s the ideological configuration of the Czech party system changed significantly. Various populist anti-establishment parties emerged and sparked debates over the corruption among elites after 1989 (Balík and Hloušek 2016, 109). Against this background, Mansfeldová and Lacina (2019, 134) speak of a tripolar party system, which consists of the "traditional, established centre-right, represented by the ODS, KDU-ČSL, and TOP 09 [...], the traditional left represented by the ČSSD and KSČM", and "the new populist, 'non-political politics' protest pole made up by ANO 2011 and Úsvit". They add that the left cannot be regarded as a coherent pole, since its constituents are quite divided as well. Whether the KSČM's support for the minority government of ANO and ČSSD in 2017 marks a permanent break in the cordon sanitaire vis-à-vis the Communists remains to be seen. In any case, there is no bipolar opposition in the Czech party system in the 2010s. Table 5.3 summarises the ideological configuration of the Czech party system in the periods when radical right parties were present in parliament.


*Table 5.3: Bipolar opposition in the Czech party system*

Source: Own compilation.

## 5.1.3 Characteristics and preferences of Czech radical right parties

#### Parliamentary strength

The SPR-RSČ entered parliament for the first timein 1992,whenit was still the Czech National Council in the Czechoslovak Federation. Even though the Czech Republic was not yet independent, the 1992 elections are included in the analysis because the Czech and Slovak National Councils showed significant similarities to national parliaments and functioned as such after the Velvet Divorce in January 1993. For instance, an almost entirely different set of parties competed in each part of the federation, the campaigns focused on specifically Czech or Slovak issues, respectively, and these parties took decidedly different approaches to the federal question (Grotz and Weber 2011, 200).

The radical right SPR-RSČ managed to enter parliament twice in 1992 and 1996, but it never achieved substantial electoral successes (see Table 5.4). In the fragmented parliament of 1992, the SPR-RSČ was one of six parliamentary groups, each of which controlled between 14 and 16 of the 200 available seats. From a purely numerical point of view, the bargaining position of these small parties, including the SPR-RSČ, was relatively weak. In 1996, the party fared better at the polls and won 18 seats, but it still remained one of the smallest parties in the Czech parliament with rather limited bargaining power.

In the 2000s, the soon-to-be-banned Workers' Party (DS) and its successor were the only radical right parties of any significance in the Czech Republic, even though they never threatened to pass the threshold of representation (Mareš 2015). In 2013, however, Tamio Okamura's first party, Úsvit, entered parliament with 6.9 per cent of the votes and 7.0 per cent of the seats. Yet, this party proved incapable of parliamentary work and dissolved almost as quickly as it had emerged. Okamura's new party, the SPD, was somewhat more successful at the polls in 2017, achieving the first double-digit result of a Czech radical right party at the national level. In the nine-party

parliament, this result put the SPD tied for third place with the Pirates in terms of parliamentary seats.


*Table 5.4: Election results and parliamentary strength of radical right parties in the Czech Republic*

Source: Own compilation based on data from Nordsieck 2021.

#### Ideological distance to the formateur

Because the SPR-RSČ entered parliament in the 1990s, there is no quantitative expert survey data available. Therefore, the socio-economic and socio-cultural positions of this radical right party and the formateur must be obtained through a qualitative assessment using secondary literature (see Chapter 4.4.2).

The ODS received the mandate to form the 1992 and 1996 Czech governments. More than most other parties, the ODS of the early 1990s, with its leading figure Václav Klaus, supported a big-bang approach to economic transformation.The party's economic agenda at that time favoured the privatisation of state-owned property and quickly establishing the institutions of a neoliberal free market economy. ODS' attempt to prioritise Czech investors in the privatisation process constitutes an outlier in the party's otherwise comprehensive privatisation policy. It was not until the mid-1990s that some members in the party began calling for certain elements of a social welfare state. To the SPR-RSČ, socio-economic issues were less salient compared to their socio-cultural concerns. However, the radical right party had a rather favourable position towards the privatisation of state property, because it viewed this as an expression of its distinctly anti-Communist stance. Otherwise, however, the party had a rather left-leaning agenda, including strong elements of welfare chauvinism, which differed from the neoliberal programme of the ODS (Vodička 1997, 114, 130, 2005, 162; Bugajski 2002, 237, 246). In light of these positions, the ODS is placed on the liberal end of the socio-economic dimension, with a score of 8.50 in 1992 and 7.50 in 1996, while the SPR-RSČ is placed slightly on the protectionist side, with a score of 4.00 in both years. Overall, therefore, there is a relatively large ideological distance between the radical right SPR-RSČ and the conservative ODS

on the socio-economic dimension in both years of government formation (see Table 5.5).

As far as the socio-cultural dimension is concerned, the ODS of the early 1990s focused on building a liberal democracy founded on individual rights and freedoms. The party's privatisation policies involved minor nationalist tendencies, but otherwise, the conservative elements in the party were clearly subordinate to the goals of democratisation, liberalisation, and integration with the West. Only after several liberals left the ODS in 1998, the party began developing into the conservative and Eurosceptic party it is today (Bugajski 2002, 246).The SPR-RSČ agreed with the ODS on the nationalist approach to privatisation and was even more vocal in this regard; however, the two parties had little else in common concerning socio-cultural issues. The SPR-RSČ favoured an authoritarian regime, including the reinstatement of the death penalty. In addition, the party held firm anti-minority sentiments, particularly against Roma, Jews, Germans, and the foreign workers who had come to the Czech Republic from former Communist allies. Party leader Miroslav Sládek advocated for a Czechoslovak state based on the 1918 borders, which entailed irredentist claims to reintegrate Carpathian Ruthenia (Bugge 1994, 161; Bugajski 2002, 257; Čakl and Wollmann 2005, 32). Sládek upheld this position even after the Velvet Divorce, when this part of Ukraine no longer shared a border with the Czech Republic, and established a symbolic branch of his party there (Mareš 2015, 212). Accordingly, the SPR-RSČ receives a score of 9.0, close to the TAN end of the socio-cultural dimension in 1992 and 1996. The ODS is placed at the GAL end of the dimension, albeit somewhat closer to the centre in 1996 (4.00) than in 1992 (3.50) due to the emergence of more conservative tendencies since the mid-1990s.

The ideological profile of the radical right Úsvit was less clear than that of the SPR-RSČ in the 1990s. At the beginning, Úsvit's 2013 election campaign focused on strengthening direct democracy as a core plank in the party's populist antiestablishment platform. Okamura increasingly adopted racist positions, however, most notably directed against Roma (Havlík 2014, 45; Hloušek and Kaniok 2014, 6; Hloušek, Kopeček, and Vodová 2020, 130–31). It is difficult to gauge the positions of this leader-centred flash party, but Okamura has been less radical than other figures from the radical right, such as his Czech "predecessor" Sládek, and this is reflected in Úsvit's GALTAN placement in the CHES (7.71). Úsvit's socio-economic positions were even less clear. Okamura's rare statements on these topics remained vague and contained both pro-business elements and promises of social safety (Hloušek, Kopeček, and Vodová 2020, 131; see also Stegmaier and Linek 2014), which led the CHES to assign Úsvit a centrist score of 5.33 on the socio-economic dimension.


*Table 5.5: Socio-economic and socio-cultural distance between radical right parties and formateurs in the Czech Republic*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions.

In 2013, the ČSSD acted as the formateur. With its roots in the opposition to the Communist regime, the ČSSD is one of the few successful members of the social democratic party family in Central and Eastern Europe that is not a Communist successor party. Its positions are rather typical of European social democratic parties. The ČSSD generally favours a market economy but criticises its neoliberal manifestation. It prefers a certain degree of state regulation of the economy and a strong welfare state (Vodička 2005, 157). The party's 2013 manifesto includes, for instance, demands for a higher minimum wage, progressive taxation, and tax increases for large enterprises (Havlík 2014, 46). On socio-cultural issues, the ČSSD is located at the liberal end, which sets it apart from some other (nominally) social democratic parties in the region, such as the Slovak Smer (see Chapter 5.4). The party's core programmatic documents advocate for the rights of ethnic and social minorities. There is, however, a certain gap between the rather progressive party elite and large parts of the party's membership and electorate (Koubek and Polášek 2017, 16).These positions are reflected in the party's CHES scores of 2.71 and 4.43 on the socio-economic and socio-cultural dimension, respectively, resulting in a moderate distance between ČSSD and Úsvit.

Okamura's second party, the SPD, emphasised nativism as the core component of its radical right ideology.The Roma minority remained one of the main targets of Okamura's agitation, but in the context of the "migration crisis", he also presented the SPD as a hard-line anti-immigrant and anti-Muslim party. Okamura's party

was instrumental in politicising the immigration issue in the Czech Republic and claimed ownership of this essential radical right issue. In addition, the SPD opposed the EU, whose policies it blamed for increasing immigration. Okamura even called for the Czech Republic to leave the EU (Hloušek, Kopeček, and Vodová 2020, 158–61). The party's placement at 9.37, near the TAN pole of the socio-cultural dimension, reflects these positions. Similar to his previous party, Okamura remained largely silent on, and indifferent to, socio-economic issues, as the party's CHES score of 4.67 on this dimension indicates.

ANO shares Okamura's populist anti-establishment appeal, even though its leader, Andrej Babiš, is one of the wealthiest entrepreneurs in the country. In one of his main campaign slogans, Babiš argued that the country must be run like a firm in order to be successful (Buštíková and Guasti 2019; see also Hanley and Vachudova 2018). When it comes to tangible policy positions, however, the party's profile is rather vague. In the socio-economic sphere, the CHES places ANO slightly left of centre (4.50), which adequately reflects the party's position. When Babiš founded ANO, he criticised the incumbent government for its neoliberal policies (Stegmaier and Linek 2014) and in the 2013 coalition negotiations, he opposed tax increases (Havlík 2014, 48). Other research, however, places the party right of centre on the socio-economic dimension. It characterises ANO as a party with a clear pro-market orientation and an "economically liberal vision of empowered citizen-consumers", but also acknowledges some rather left-leaning ideas, such as support for elements of a sharing economy (Hanley and Vachudova 2018, 281). ANO's socio-cultural positions, for instance on gender issues, remain vague and indifferent. At the same time, however, Babiš' anti-establishment appeal entails a somewhat authoritarian and anti-pluralist thrust (Hanley and Vachudova 2018, 281–82). Hence, the CHES places the party slightly to the TAN end of the socio-cultural dimension (5.73).

#### 5.1.4 Summary

The Czech Republic is the only Central and Eastern European country in this study where a radical right party has entered parliament but not government. The longstanding cordon sanitaire vis-à-vis the radical right prevented the SPR-RSČ and Úsvit from joining a government coalition. Together with poor internal organisation, intra-party conflicts, and scandals, the non-negotiable stance taken by the mainstream parties might have also contributed to the short lifespan of these two radical right parties (Čakl and Wollmann 2005, 32–33; Tavits 2013, 217; Hloušek, Kopeček, and Vodová 2020, 127–34). Whether Okamura's new party will be able to establish itself in the long term remains to be seen, but the eroding cordon sanitaire and Okamura's improved organisational skills put the SPD in a favourable position (Hloušek, Kopeček, and Vodová 2020, 157–66).

However, like its unsuccessful predecessors, the SPD was not included in government after the 2017 parliamentary elections. Instead, the pattern of ideologically broad coalitions typical of the post-Communist Czech Republic continued.This general pattern is, at least in part, a result of the non-coalitionability of the radical right parties and the KSČM until the second half of the 2010s. In this respect, too, it will be interesting to see whether the formats of future coalitions will change following the dealignment in the party system and the erosion of the cordon sanitaire surrounding the radical left and right. In light of these changes, it is not certain that the success of radical right parties in the Czech Republic will remain limited to a strong showing on the opposition bench in parliament.

## **5.2 Hungary**

## 5.2.1 Government formation with radical right parties in Hungary

The 1998 parliamentary elections marked a breakthrough for the first radical right party in post-Communist Hungary, the Hungarian Justice and Life Party (MIÉP). The Alliance of Young Democrats (Fidesz) won the election and its leader, Viktor Orbán, received the mandate to form the new government. He entered into coalition negotiations with two other conservative parties, the Hungarian Democratic Forum (MDF) and the agrarian Independent Smallholders' Party (FKgP), which quickly agreed on a joint coalition (Lomax 1999, 123). Even though the radical right MIÉP remained in opposition, the party supported the coalition during the investiture vote and on several important matters over the course of the legislature. When the FKgP became an uncomfortable coalition partner, the informal support from the radical right provided the government with an additional element of safety. For this reason, Ilonszki(2019,226) even lists theMIÉP as a support partyin theOrbán I government, but she also makes clear that this support was not wanted, or formally recognised, by the governing coalition. Thus, the MIÉP does not fulfil the criteria for a support party applied in this study (see Chapter 4).

As an oversized coalition, the Orban I government fails the criteria of the minimal range and minimal connected winning theories (see Table 5.6). It was nonetheless ideologically quite homogeneous, and the three parties were connected on the socio-economic and socio-cultural dimensions.Thus, the additional inclusion of the MDF did not increase the ideological range of the government too much.


*Table5.6:FormatandideologicalrangeofgovernmentsinHungary*

Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in paren- theses indicate the smallest possible ideological distance of a majority government.

## 5.2.2 The configuration of the Hungarian party system

## Fragmentation

The Hungarian party system of 1998 featured only 3.5 effective parliamentary parties and was thus quite compact (see Table 5.7). Hungary was a forerunner in terms of party system institutionalisation in Central and Eastern Europe due to the rapid concentration of political parties and the consistently low fragmentation of the party system (Enyedi and Casal Bértoa 2018). The transfer of power and the cooperation between the reformed Communist successor, the Hungarian Socialist Party (MSzP), and the liberal Alliance of Free Democrats (SzDSz), in 1994 also went smoothly. Hence, the party systems initially showed stable development towards moderate pluralism and centripetal party competition, to use the Sartorian terminology. Since 2010, however, the situation has changed, and Fidesz has become the dominant party, aided by the illiberal reforms Orbán has enacted to consolidate his own power.

From a purely numerical perspective, the bargaining situation after the 1998 elections was of moderate complexity. The effective number of 3.5 parliamentary parties resulted from two dominant parties, Fidesz and MSzP, which controlled 38 and 35 per cent of the seats in parliament, respectively. The third-strongest party, FKgP, held only 12.4 per cent of the seats.Overall, there were six mathematically possible minimal winning coalitions. Fidesz could have formed a two-party minimum winning coalition, either with the MSzP or the FKgP. The MSzP, in contrast, would have required the FKgP and either of the three small parties to form a minimal winning coalition.



Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

#### Bipolar opposition

The Hungarian party system of the early 1990s was characterised by multiple salient divides. In the context of the first free elections in 1990, the regime divide played a major role in the country, as did socio-economic and value conflicts, the latter pitting nationalists against cosmopolitans (Grotz 2000, 231; Cabada, Hloušek, and Jurek 2014, 96; Ilonszki 2019, 208). The MDF represented the conservative forces, while the SzDSz and Fidesz stood for the more liberal wing (Grotz 2000, 224–225, 239–234). The SzDSz' support for the conservative government's constitutional changes demonstrated, however, that the ideological differences did not prevent the parties on this side of the regime divide from cooperation (Grotz 2000, 240). In the mid-1990s, however, a bipolar opposition between conservative and liberal camps began to take shape. The coalition of MSzP and SzDSz after the 1994 parliamentary elections symbolised the erosion of the regime divide, at least in the sphere of coalition politics, and the two parties constituted the new liberal pole in Hungarian politics. Fidesz, in contrast, turned towards the conservative camp after the electoral defeat in 1994 (Grotz 2000, 265–66).

In the 1998 parliamentary elections, the bipolar opposition consisted of a national-conservative camp, dominated by Fidesz, and a left-liberal camp, led by the MSzP. The conservative parties even ran joint candidates against the MSzP in the constituencies (Grotz 2000, 267). This tactical decision was highly relevant because the Hungarian electoral system had strongmajoritarian elements.The bipolar opposition was based on congruent socio-economic and socio-cultural policy positions, but it also entailed an affective dimension that involved and perpetuated elements of the regime divide (Grotz 2000,275–76; Cabada,Hloušek, and Jurek 2014, 96; Ilonszki 2019, 208). Even though the opposition would become more intense in the times to come, coalitions between the camps were already impossible in 1998 (see Table 5.8).


Source: Own compilation.

#### 5.2.3 Characteristics and preferences of Hungarian radical right parties

#### Parliamentary strength

The radical right MIÉP managed to enter parliament after its second campaign in 1998, but the party gained only 5.5 per cent of the votes. Due to the complex Hungarian electoral system, the party received only 3.6 per cent of the 386 seats in the Hungarian parliament (see Table 5.9), making it the smallest parliamentary group. In 2002, the party narrowly missed clearing the five per cent threshold and never recovered from this electoral defeat.


*Table 5.9: Election results and parliamentary strength of radical right parties in Hungary*

Source: Nordsieck 2021.

## Ideological distance to the formateur

MIÉP's radical right ideology is strongly connected to the so-called trauma of Trianon. When signing the Treaty of Trianon after World War I, Hungary lost twothirds of its territory and large numbers of its population to neighbouring countries (Pytlas 2016, 156). To MIÉP, ethnic Hungarians living outside the country are an integral part of the Hungarian nation, and the party seeks to reclaim former Hungarian territories. The party also considers anyone who does not support their revisionist and irredentist views as a traitor to the nation (Karsai 1999, 136–39; Bernáth, Miklósi, and Mudde 2005, 82). Additionally, MIÉP's ideology includes an ethno-religious concept of the nation which excludes various minorities, such as the Roma, Jews, or LGBTIQ+. Party leader István Csurka was known to be a particularly notorious anti-Semite (Karsai 1999, 142–43; Bernáth, Miklósi, and Mudde 2005, 83; Krekó and Mayer 2015, 187). Consequently, the CHES places the party at 9.67 on the socio-cultural dimension (see Table 5.10).

In the socio-economic sphere, MIÉP advocated national-protectionist policies but it did not reject capitalism per se. The party demanded, for instance, that companies either be nationalised or, if private, be run only by Hungarians. International investors were often portrayed as enemies and part of an alleged international Jewish conspiracy.The party's ultimate goal was to establish a closed, national economic and social system that benefitted only those who belonged to the imagined, homogeneous Hungarian nation or, in Karsai's (1999, 140–41) words, "capitalism controlled by the 'Christian-national' elite" (see also Bock 2002, 285; Bernáth, Miklósi, and Mudde 2005, 83). These positions are also reflected in MIÉP's CHES score of 4.00 on the socio-economic dimension.

After the 1998 parliamentary elections, Fidesz acted as the formateur. Fidesz was established as a liberal opposition movement against the Communist regime, but underwent a double transformation, first into a national-conservative party in the second half of the 1990s and then into a radical right party in the mid-2010s (Bayer 2005; Pytlas 2016; Minkenberg 2017; Mudde 2020; see also Chapter 2.1). This study, however, is only interested in the party's positions during and after the first transformation. Fidesz' pro-market stance during the early 1990s had already vanished by 1998 in favour of a rather national-protectionist socio-economic agenda. In the context of the 1998 parliamentary elections, Fidesz called for limiting foreign investments and fortifying a chauvinist welfare state (Lomax 1999, 121; Bayer 2005, 178–79; Pytlas 2016, 40). Accordingly, the CHES places Fidesz slightly to the left of the centre on the socio-economic dimension (4.67).

In socio-cultural terms, the party also adopted the ethno-religious understanding of nationhood held by its conservative and radical right competitors. Shortly before the 1998 elections, Orbán arranged for the Holy Crown of St Steven, an important symbol of Hungary's religious and nationalist forces that is also closely linked to the idea of a Greater Hungary, to be displayed in parliament (Pytlas 2016, 40, 156). This act is only one example of Fidesz' mythical reinterpretation of Hungarian national history (Pytlas 2016, chap. 6; see also Bayer 2005, 184). In the 1998 election campaign, Orbán also accused the incumbent MSzP-SzDSz government of betraying ethnic Hungarians in neighbouring Romania and Slovakia because they had signed treaties with both countries (Bayer 2005, 178). Even the illiberal efforts to weaken the system of checks and balances and a democratic civil society, implemented by the Orbán governments since 2010, were visible as early as the late 1990s (Bayer 2005, 180–81). Fidesz' GALTAN score in the 2002 CHES wave (8.15) aligns with the party's nationalist and authoritarian policy preferences.

*Table 5.10: Socio-economic and socio-cultural distance between radical right parties and formateurs in Hungary*


Source: Chapel Hill Expert Survey (Jolly et al. 2022).

#### 5.2.4 Summary

When MIÉP entered parliament for the first and only time in 1998, the party system was quite polarised and the national-conservative camp that also included MIÉP emerged victorious. MIÉP was ideologically quite close to Fidesz, the formateur of the 1998 government, but Orbán still excluded the radical right from government. The party's behaviour in parliament suggests, however, that it would have been prepared to cooperate more closely with the conservative government if it had depended on the support of the radical right.

## **5.3 Poland**

## 5.3.1 Government formation with radical right parties in Poland

The radical right League of Polish Families (LPR) entered the Sejm for the first timein 2001. In this year, a social democratic electoral alliance between the reformed Communist successor party, the Democratic Left Alliance (SLD), and the Labour Union (UP) won the parliamentary elections. SLD's leading candidate, Leszek Miller, considered several options, but ultimately established a minimal winning coalition with the agrarian Polish People's Party (PSL) (see Table 5.11). LPR's participation in government was neither a viable option for the formateur nor for the radical right party itself (Jasiewicz and Jasiewicz-Betkiewicz 2002; Millard 2010, 114).<sup>4</sup>

During the term, the government lost public support after multiple corruption scandals, internal conflicts within the SLD, and intra-coalitional disputes between the SLD and the PSL. For these reasons, Miller announced the expulsion of the PSL in 2003. The remaining minority coalition continued in office but had to rely on issue-based support from other parliamentary parties (Jasiewicz and Jasiewicz-Betkiewicz 2003). One day after Poland's accession to the EU, on 2 May 2004, the Miller government resigned, and Marek Belka (SLD) was elected as the new prime minister. However, the partisan composition of the government remained stable (Jasiewicz and Jasiewicz-Betkiewicz 2004) and the Belka government served as a "de facto caretaker" until the next parliamentary election in 2005 (Jasiewicz and Jasiewicz-Betkiewicz 2006, 1232).

The 2005 parliamentary election heralded the demise of the crisis-ridden SLD. Moreover, it witnessed a duel between two parties from the post-Solidarność camp, the liberal-conservative Civic Platform (PO) and the national-conservative Law and Justice (PiS). Although many observers expected these two parties to form a governing coalition, fierce competition in the presidential election, scheduled shortly after the parliamentary election, ultimately prevented them from cooperating. Instead, PiS forged an alliance with the populist Self-Defence of the Republic of Poland (SO) and the radical right LPR, led by PiS backbencher Kazimierz Marcinkiewicz.<sup>5</sup> The LPR and the SO initially served as support parties for a PiS minority government

<sup>4</sup> This coalition is classified as an oversized coalition in Table 5.11 since the SLD and the UP did not form a joint parliamentary group despite their electoral alliance, and a coalition of the SLD and PSL would have controlled a majority in parliament (Jasiewicz and Jasiewicz-Betkiewicz 2002).

<sup>5</sup> PiS' party leader Jarosław Kaczyński spearheaded the 2005 campaign, but he gave way to Marcinkiewicz in order not to jeopardise the presidential candidacy of his twin brother, Lech Kaczyński (Millard 2010, 143).

(Jasiewicz and Jasiewicz-Betkiewicz 2006; Millard 2010, 136–38). Later in 2005, however, PiS entered into negotiations with the two support parties and the PSL in order to form a more stable government.These negotiations resulted in a formal coalition between PiS and its former support parties, LPR and SO. The Marcinkiewicz II majority government assumed office in May 2006 and two months later,Marcinkiewicz was replaced as prime minister by PiS party leader Jarosław Kaczyński (Jasiewicz and Jasiewicz-Betkiewicz 2006, 2007; Millard 2010, 143–44). After a little more than a year of continuous quarrels, scandals, and ministerial resignations, the Sejm finally removed the incumbent government and voted for early elections in September 2007 (Jasiewicz and Jasiewicz-Betkiewicz 2008; Millard 2010, 144–47).This transition marked the demise of the LPR, which fell well short of the five per cent threshold and subsequently disappeared from Polish politics.

Although LPR's participation in government was fraught with conflict between the ruling parties, the coalition was ideologically very homogeneous. When PiS, SO and LPR entered a formal coalition in 2006, it met the requirements for both the minimal range and the minimal connected winning coalition on the socio-economic and socio-cultural dimensions. Ideologically, these parties were already proximate in 2005, but it was not until a year later that they also met the majority criterion required by both formats. The two governments under prime minister Miller in the previous term were ideologically more heterogeneous. In socio-economic terms, SLD, UP, and PSL were quite close, but a more homogeneous majority coalition was still possible. On the socio-cultural dimension, however, the ideological range of the 2001 Miller I government was rather large. It narrowed considerably after the expulsion of the PSL in 2003. Yet, the coalition no longer controlled a majority and was thus neither a minimal range nor a minimal connected winning coalition by that point.

#### 5.3.2 The configuration of the Polish party system

#### Fragmentation

The Polish party system has become less fragmented and more institutionalised since the early 1990s (Enyedi and Casal Bértoa 2018, 443). After the first free elections in 1991, a total of 29 parties entered parliament and the effective number of parliamentary parties reached as high as 10.9 (Toole 2000; Casal Bértoa 2021). However, after the introduction of a parliamentary threshold rule, both figures dropped rapidly, so that the effective number of parties in Poland has been near the Central and Eastern European average since the 2000s (Enyedi and Casal Bértoa 2018).

The fragmentation scores for the 2001 and 2005 legislatures are slightly below and above four, respectively (see Table 5.12).The increase from 2001 to 2005 resulted from the sharp electoral decline of the SLD after Miller's term in office. While the SLD almost gained an absolute majority in 2001, the party struggled to re-enter parliament in 2005, thus leaving room for new parties. PiS and PO emerged as more or less equal competitors, which led to a relatively even distribution of power within the Polish party system, and a higher effective number of parties. After the 2001 elections, the majority situation in parliament resulted in a bargaining situation of limited complexity. The SLD was in such a strong position that it could have formed a minimal winning coalition with any of the other six parliamentary parties, except its ally UP. In 2005, the bargaining situation became more complex since both PiS and PO could have formed various minimal winning coalitions, even though victory left PiS with more options than PO.

### Bipolar opposition

The regime divide between the reformed Communist successor party, SLD, and the alliance of post-Solidarność parties structured the Polish party system of the 1990s (Grzymała-Busse 2001, 94–96; see also Millard 2010). The situation began to change around the turn of the millennium, however.


*Table 5.11: Format and ideological range of governments in Poland*

Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.Partiesinof

\*

=

 Surplus = oversized

 parentheses support party a minority government. \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government,coalition.  \*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in paren- theses indicate the smallest possible ideological distance of a majority government.


*Table 5.12: Fragmentation of the Polish party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

\* The government is classified as an oversized coalition since the SLD and the UP did not form a joint parliamentary group despite their electoral alliance, and a coalition of the SLD and PSL would have controlled a majority in parliament (Jasiewicz and Jasiewicz-Betkiewicz 2002).

The 2001 parliamentary elections saw the collapse of the two hitherto dominant forces from the post-Solidarność camp, the Solidarity Electoral Action (AWS) and the Freedom Union (UW). Some of their voters turned to PO and PiS (Millard 2010, 113). Moreover, two new parties, the radical right LPR and the populist SO also entered the Sejm.These developments altered the conflict structure in the Polish party system.The regime divide was still present but, for the first time, a coalition between the SLD and former opposition parties seemed possible(see Table 5.13).The differentiation within the post-Solidarność camp also highlighted the divide between liberal and conservative forces. Hence, the oppositional constellations in the Polish party systems had become somewhat more diverse in the context of the 2001 parliamentary elections (Millard 2010, 114).

The 2005 elections yielded a similar result for the SLD as the 2001 elections for AWS and UW, even though the SLD did not drop completely out of parliament.These shifts in the balance of power were accompanied by another change in the ideological configuration of the Polish party system. The regime divide became less salient, while the importance of the divide between liberal and conservative forces, in particular between PO and PiS, gained momentum. Both parties differed in their socioeconomic and socio-cultural policies, but these issue-based differences were still reconcilable.The affective polarisation between both parties grew rapidly during the 2005 presidential election, however, and rendered cooperation impossible (Szczerbiak 2007; Millard 2010, chap. 7; Keudel-Kaiser 2014, 183–87). Thus, the 2005 elections mark the beginning of the deep bipolar opposition between "'social-solidaristic' and 'liberal' visions of Poland" (Szczerbiak 2007, 204) that continues to shape the Polish party system even today.


*Table 5.13: Bipolar opposition in the Polish party system*

Source: Own compilation.

## 5.3.3 Characteristics and preferences of Polish radical right parties

#### Parliamentary strength

The LPR was founded shortly before the 2001 parliamentary elections and immediately won 7.9 per cent of the vote (see Table 5.14). The party benefitted from an electoral system reform which changed from the d'Hondt to a modified Sainte-Laguë formula, resulting in 8.3 per cent of the seats for the radical right. Nevertheless, the LPR was the smallest parliamentary group in the Sejm between 2001 and 2005 (Millard 2010, 112).

Because the new electoral formula prevented the SLD-UP coalition from winning a majority in parliament in 2001, the Miller government decided to return to the d'Hondt formula, which benefits large parties, in the 2005 parliamentary election (Millard 2010, 112). Therefore, the LPR obtained only 7.4 per cent of the Sejm seats in 2005, despite a slight improvement at the polls.The party thus fell well short of its result in Poland's first elections to the European Parliament in 2004, in which it gained almost 16 per cent of the vote, a result due in part to very low voter turnout (Millard 2010, 125). Thus, the LPR remained among the smallest parties in parliament in the 2005 legislature.



Source: Nordsieck 2021.

### Ideological distance to the formateur

During its two terms in parliament between 2001 and 2007, the LPR faced quite different formateurs, the reformed Communist successor party, SLD, and the nationalconservative PiS.The LPR itself was deeply rooted in Poland's national Catholic right wing.The origins of the party's ideology go back to Roman Dmowski's inter-war endecja movement.The LPR picked up on this tradition and carried an ethno-religious idea of the Polish nation, an ultra-conservative image of the family, and a virulent anti-Semitisminto parliament (Pankowski and Kornak 2005, 159; Grün and Stankiewicz 2006; see also Kasprowicz 2015; Pytlas 2016). It was also the only parliamentary party at that time to unequivocally oppose Poland's accession to the EU based on an alleged threat to national sovereignty (Pankowski and Kornak 2005, 159; Millard 2010, 134). Pytlas (2016, 92) describes LPR's Euroscepticism as being rooted in the idea that the Polish nation is a "bulwark of Christianity" defending against the EU, which is the embodiment of a morally corrupted, Western "civilization of death". In a similar vein, the party advocated for a comprehensive ban on abortion, presenting itself as a fierce opponent of gender diversity and the LGBTIQ+ community (Pytlas 2016, chap. 4; see also Hennig 2010).

The LPR was clearly positioned on the left end of the socio-economic spectrum. The party's socio-economic policies were based in Catholic social teaching and connected to the socio-cultural core issues of the party (Łapiński 2004). As for specific policies, the party campaigned for the re-nationalisation of key industries, against cutsin the welfare system,and for taxation and social systems which supported their traditional understandings of the family (Millard 2010, 131–33).

The SLD had clearly dissociated itself from its Communist past and developed a moderate social democratic profile. The party preferred a "sensitive privatisation" and sought to reduce unemployment, tomoderatelyincrease social benefits, but also to reduce taxes(Millard 2010, 106).As regards the socio-cultural sphere, the SLD took a secular, liberal stance, calling for gender equality and a liberal abortion policy. It was also staunchly pro-European (Millard 2010, 104–6).Thus, there is a certain proximity to the LPR on socio-economic issues, but the socio-cultural distance between the two parties could hardly be greater (see Table 5.15). Hence, LPR and SLD were both located on the left side of the socio-economic dimension, but the formateur was more moderate than its radical right competitor. This is also indicated by the parties' CHES scores of 2.00 (LPR) and 4.25 (SLD), respectively. In line with the parties' positions, however, the CHES places both parties on opposite ends of the GAL-TAN dimension. The LPR is close to the TAN pole (9.75), whereas the SLD occupies a position in the liberal spectrum (1.88).


*Table 5.15: Socio-economic and socio-cultural distance between radical right parties and formateurs in Poland*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions.

PiS, in contrast, was much closer to the LPR on the socio-economic and sociocultural dimensions. Initially, the party championed law and order, as suggested by its name. Quite quickly, however, PiS began emphasising the idea of national Catholicism, although in a slightly more moderate fashion than the LPR.The party's leading figures, the twin brothers Jarosław and Lech Kaczyński, announced that their ideological and historical roots do not lie in Dmowski's national movement but rather referred to the more liberal, inter-war national movement of Józef Piłsudski (Pankowski 2010, 155–57; Pytlas 2016, 30–31, chap. 4). Nevertheless, the Kaczyński brothers believed that it was impossible to "build a patriotic party without people of national-Catholic convictions" (Pankowski 2010, 156). This statement underlines PiS' ethno-religious concept of nationhood and thus signals a crucial similarity to the radical right LPR. Several of the parties' policies also reflect this resemblance. For example, PiS aimed to preserve the influence of the Catholic church in society, to establish the traditional family as the backbone of Polish society, and to curb women's reproductive rights as well as the rights of the LGBTIQ+ community. The only major issue where PiS' and LPR's positions differed was Poland's accession to the EU, which PiS did not oppose (Millard 2010, 134). In socio-economic terms, PiS positioned itself on the left. The party advocated progressive taxation, a large-scale social housing programme, and it promised tax and welfare benefits to married couples and (traditional) families. The party generally favoured a private economy, but it also wanted to keep key sectors under state control (Millard 2010, 131–33).

Based on their shared positions, PiS and LPR proposed a comprehensive reform package during their 2005 election campaigns.This proposal for a "Fourth Republic" was initially introduced by PiS but quickly embraced by the LPR. It aimed at transforming Poland into a national-Catholic society with a strong government and a law and order regime, which would be achieved by substantial policy and even constitutional changes, a massive lustration and anti-corruption campaign, and a new social contract that placed the traditional—ethnic Polish and Catholic—family at the centre of Polish national identity (Millard 2010, 127; Pytlas 2016, 30–31). The CHES placements of PiS and LPR in the 2006 wave reflect the parties' socio-economic and socio-cultural proximity.The LPR receives themaximum GALTAN score of 10.00 and a score of 1.17 on the socio-economic dimension. PiS' placement at 2.00 provides an adequate reflection of the party's socio-economic policies, but its GALTAN score of 9.50 seems somewhat exaggerated, probably resulting from inflated perceptions of the polarisation in the Polish party system at the time (Pytlas and Kossack 2015, 117–18). It is therefore adjusted to 9.00.

### 5.3.4 Summary

At a time of change in the Polish party system, when the regime divide was losing salience and its previous representatives suffered massive electoral losses, the radical right LPR managed to enter the Sejm twice. In 2005, when the opposition between PO and PiS took shape in the context of a heated presidential race, the rift between these two parties even propelled the LPR into power. However, it was not only the bipolar opposition in the party system, but also LPR's ideological proximity to the formateur, PiS, that made the radical right party a viable junior partner. The whole PiS-SO-LPR government is one of the most ideologically homogeneous governments with a radical right party in the entire study.

Nevertheless, the cooperation between PiS and LPR did not last long due to the numerous conflicts within and between the governing parties (Millard 2010, 143–47; see also Jasiewicz and Jasiewicz-Betkiewicz 2006, 2007, 2008). Moreover, after their joint government, PiS engaged in a strategy of co-optation, taking over positions and narratives from the LPR (Pytlas and Kossack 2015; Pytlas 2016). PiS succeeded with these tactics and eliminated its radical right competitor in the 2007 elections. However, the party maintained the radical positions and moved even further towards the right later on. Thus, the LPR contributed to PiS' transformation into a radical right party (see Chapter 2.1), even though it was present in parliament and government for only a short period of time.

#### **5.4 Slovakia**

#### 5.4.1 Government formation with radical right parties in Slovakia

Over the last three decades, the Slovak National Party (SNS) has been one of the most electorally consistent radical right parties in Central and Eastern Europe. The party entered the Slovak National Council in 1992, which became the first parliament of the independent Slovak state in 1993. Vladimír Mečiar's Movement for a Democratic Slovakia (HZDS) emerged as the undisputed winner of the 1992 elections, but it fell two seats short of a majority in parliament. Of the other four parties in parliament, Mečiar included the radical right SNS as a junior partner in the first coalition government to rule the newly independent Slovak state.

Personal and ideological tensions within the SNS and HZDS parliamentary groups led several members to defect causing the government's majority in parliament to shrink quickly. Mečiar attempted to win further support for his cabinet midway through 1993, but even though his attempts failed and the government was left without a majority in parliament, the opposition was not united enough to vote the government out of office (Malová 1994). It took until March 1994 for the opposition to close ranks and remove the incumbent government from power. The former HZDS parliamentarian, Jozef Moravčík, was elected to replace Mečiar as prime minister. However, the Moravčík government is regarded as a caretaker government, since the parties also agreed to call early elections, which gave the government little room to manoeuvre (Malová 1995).

The HZDS again won the parliamentary elections in 1994, but with fewer parliamentary seats than in 1992. Mečiar initially entered into coalition negotiations with the reformed Communist successor, the Party of the Democratic Left (SDĽ), and the Christian Democratic Movement (KDH). Unlike his former coalition partner, both parties controlled enough seats to secure a majority in parliament. Since Mečiar could not convince either one to govern with him, he turned to the smaller parliamentary parties, including the radical right SNS.The SNS agreed to renew cooperation with Mečiar who also managed to secure the support of the Union of the Workers of Slovakia (ZRS), a party that stood in ideological continuity with the Communist regime (Malová 1995). Despite the different ideological backgrounds of the constituent parties, this three-party majority coalition lasted the entire legislature and helped Mečiar to establish an autocratic, illiberal and clientelist regime which deeply divided the country.

Even though the HZDS once again emerged as the strongest party in the 1998 parliamentary elections, it was unable to muster the support it needed to form a majority. Mečiar's illiberal rule led the opposition parties to forge a broad alliance prior to the 1998 elections.This alliance, united in its opposition to a common threat, won enough votes to replace Mečiar. Mikuláš Dzurinda of the liberal-conservative Slovak Democratic Coalition (SDK) acted as the formateur of the oversized rainbow coalition with the Party of the Hungarian Coalition (SMK), the social-liberal Party of Civic Understanding (SOP), and the SDĽ that placed Slovakia back on track towards democratisation. The SNS, as a member of the former government, joined its coalition partner HZDS in opposition (Malová and Učeň 1999; Bugajski 2002, 296–97).

Due to a party split, the SNS failed to enter parliament for the first time in the 2002 elections (People Against Racism and Milo 2005, 214–15; Pirro 2016, 87). After settling this internal dispute, the party returned to parliament in 2006. In the meantime, the balance of power in the Slovak party system had shifted significantly. The liberal-conservative wing of the anti-Mečiar camp suffered substantial losses, the SDĽ had dropped out of parliament altogether, and Robert Fico's nominally social democratic Direction (Smer) had become the strongest party. In the run-up to the 2006 parliamentary elections, Fico declared that he was prepared to negotiate with parties from both sides of the regime divide. Among the two viable options, a coalition with either SMK and KDH or with HZDS and SNS, Fico opted for the latter and thus paved the way for the SNS to return to power. This coalition with the radical right and former autocratic ruler Mečiar caused international concern, particularly in the transnational Party of European Socialists, which Smer had joined in 2005 (Malová and Učeň 2007; Haughton and Rybář 2008, 248–49; Mesežnikov 2008, 10). These concerns, however, hardly affected Fico and his coalition, which remained in office for the entire term.

Domestically, the coalition did not harm Fico's popularity either. Smer even improved its electoral result in the 2010 parliamentary elections.The junior partners of the incumbent coalition, however, suffered heavy losses. The SNS barely managed to clear the five per cent threshold and the HZDS dropped out of parliament, never to return. Fico initially received the mandate to form the government, but because Smer and the SNS were not large enough by themselves and because the other parties were reluctant to cooperate with Fico and the radical right, he was unable to secure a parliamentary majority.Therefore, the mandate was passed to Iveta Radičová of the Slovak Democratic and Christian Union – Democratic Party (SDKÚ-DS), who had already negotiated with the other parties. Radičová eventually formed a majority coalition with the KDH and two new parties, the neoliberal Freedom and Solidarity (SaS) and Most-Híd. The latter had replaced the SMK as the main representative

of the interests of the Hungarian minority in Slovak politics (Deegan-Krause and Haughton 2012).This four-party coalition broke up after less than two years, following conflicts over the European financial crisis. The early termination of this coalition resulted in the 2012 snap election,in which the SNS failed to gain parliamentary representation for the second time (Malová and Učeň 2013).

Once again, however, the party returned to parliament in 2016. With Marian Kotleba's People's PartyOur Slovakia(ĽSNS),a second radical right partymanaged to enter the national parliament alongside the SNS in 2016. Smer once again emerged victorious and its leader, Robert Fico, was tasked with the formation of a new government. Fico had already mentioned his preference for a coalition with the SNS during the election campaign, but the two-party alliance was well short of a majority.Moreover, the composition of parliament had changed significantly compared to the last time these two parties had been in government together. KDH and SDKÚ-DS had disappeared, while three parties,including the radical right ĽSNS, were participating in either their first or second term. Since all parties had ruled out cooperation with the ĽSNS and some smaller parties were unwilling to govern with each other, the only potential partners remaining for Smer and the SNS were the new centre-right party Network (Sieť) and Most-Híd (Rybář and Spáč 2016). Because Slovakia would soon take over the Presidency of the Council of the European Union, some parties preferred not to have a caretaker government hold that prestigious position, but they also feared that calling new elections would further strengthen Kotleba. Therefore, the four parties—Smer, SNS, Sieť and Most-Híd—decided to form a coalition despite obvious ideological differences. Even before the investiture vote, some deputies from Sieť and Most-Híd left their parliamentary groups. Sieť lost so many deputies that it fell short of the minimum number required for a parliamentary group. Overall, however, these defections did not threaten the government'smajorityin parliament, which ultimately assumed officeinMarch 2016. Later that year, however, Sieť withdrew its support for the government completely, which was formally reduced to Smer, SNS and Most-Híd. Since some of the former Sieť deputies had joined Most-Híd, the coalition still had a majority (Baboš and Malová 2017). This three-party coalition remained in power until the next regular elections in 2020, although it was shaken by the massive public protests following the murder of Jan Kuciak and Martina Kušnirova in 2018, which forced several members of the cabinet to resign, including prime minister Fico himself.

The protest movement, and the political parties that emerged from it, achieved great success when their candidate, Zuzana Čaputová, was elected president in 2019. In the 2020 parliamentary elections, however, they narrowly missed the threshold. While support for the representatives of the protest movement dwindled, the populist anti-establishment party Ordinary People and Independent Personalities (OĽaNO) benefited from the discontent with the previous government. OĽaNO won the elections and its leader, Igor Matovič, became the formateur

of the new government. On the far right, the ĽSNS repeated its result from 2016, although other parties continued to ostracise Kotleba, whereas the SNS dropped out of parliament again. Matovič went on to form an oversized four-party coalition with the SaS, the nationalist anti-establishment party We Are Family (Sme Rodina), and the liberal-conservative party For the People (ZĽ) of former president Andrej Kiska. Matovič indicated that he sought a broad alliance in order to control a threefifths majority in parliament, which is required to amend the constitution.

The Matovič government stands in a certain continuity with the Dzurinda I government that was formed to oust Vladimír Mečiar in 1998. Except for these two oversized coalitions and two brief periods of minority rule, Slovakia was always ruled by minimal winning coalitions. Thus, unlike in the neighbouring Czech Republic, political parties in Slovakia seem to be rather reluctant to form minority governments, even when three or four parties with different ideological orientations are required to secure a majority in parliament (see Table 5.16). Only one of the governments that were formed when radical right parties were present in parliament meets the criteria for a minimal range coalition—the Radičová government of 2010 had the smallest possible ideological range on the socio-economic dimension. At the same time, however, the parties accepted great socio-cultural heterogeneity. When using the less restrictive minimal connected winning theory, the majority of the coalitions for which data are available were connected on the socio-economic dimension.Only the short-lived Fico III government was not.<sup>6</sup>

The governments that included radical right parties also show a relatively high degree of ideological homogeneity. The first government under the leadership of Robert Fico that assumed office in 2006 was even a minimal connected winning coalition on both the socio-economic and socio-cultural dimension.When considering the ideological positions of HZDS and SNS in the 1990s, the 1992 coalition of the two parties is ideologically close on both dimensions as well. After Sieť's defection, the 2016 Fico IV government is connected on the socio-economic dimension. Due to the participation of the SNS and Most-Híd, the Fico III and Fico IV governments are socio-culturally somewhat heterogeneous.Thus, there are some indications that governments with radical right parties in Slovakia might be ideologically more homogeneous than the average government in the country, particularly on the sociocultural dimension. As for the format, the governments that included radical parties were all minimal winning coalitions.

<sup>6</sup> The Matovič government was also socio-economically connected, but it does not fulfil the criteria of the minimal connected winning theory, because it was an oversized coalition.


*Table5.16:FormatandideologicalrangeofgovernmentsinSlovakia*


Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

 \*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in parentheses indicate the smallest possible ideological distance of a majority government.

### 5.4.2 The configuration of the Slovak party system

### Fragmentation

Table 5.17 shows that the fragmentation of the Slovak party system has been rather high.The only exceptions were the years 1992 and 2012.The effective number of parliamentary parties in 2012 was 2.9 (Casal Bértoa 2021), but it is not listed in the table since no radical right party was present in parliament. In both years, a single, dominant party, HZDS or Smer, respectively, came very close to, or even reached, a parliamentary majority. In 1992, the HZDS was only two seats short of a majority. Hence, the party was able to form a majority coalition with any of the other four parliamentary parties.The other elections covered here resultedinmore fragmented parliaments and thus more complex bargaining situations.

The increased effective number of parties in 1994 reflects the larger number of actual parties in parliament and the less dominant position of the HZDS.The HZDS lost seats and was unable to from a two-party majority coalition with any of the six remaining parliamentary parties. Moreover, a majority coalition no longer required all of the other parliamentary parties to stand united against the HZDS. Thus, the number of possible minimal winning coalitions was much higher in 1994 than in 1992.


*Table 5.17: Fragmentation of the Slovak party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

In 1998, fragmentation increased even though the number of actual parliamentary parties dropped from seven to six. The increase in the effective number of parties corresponds with HZDS' continuing electoral decline. The party controlled less than a third of the seats in parliament and was only one seat ahead of the SDK. Apart from a hypothetical coalition composed of the two strongest parties, at least three parties were needed to control a majority in parliament. From a purely numerical perspective, the bargaining situation was therefore quite complex. Both the HZDS and the SDK could have formed various minimal winning coalitions.

The 2006 parliamentary elections marked the beginning of Smer's electoral success. As in 1998, a total of six parties were present in parliament, but the balance of power between the two largest parties, Smer and SDKÚ-DS, was much more uneven.The option to form a minimal winning, grand coalition remained mathematically possible, but other than that Smer could form only minimal winning coalitions of at least three parties. Likewise, the SDKÚ-DS would have needed three or more junior partners to reach a majority.

In 2010, Smer controlled more than 40 per cent of the seats and was thus in a much stronger position than in 2006. Fico could have formed a two-party majority coalition with any parliamentary party except the SNS. Thus, the bargaining situation in 2010 was of moderate complexity and resembled that of 1994.

Smer's electoral success peaked in the 2012 parliamentary elections when it won 83 of the 150 seats in the Slovak parliament and formed a single-party government. In 2016, Smer remained the strongest party, but its size was reduced to less than 50 seats, while only one of remaining seven parties controlled more than 20. Hence, Smer still dominated the bargaining process. The extremely high fragmentation of 5.7 effective parliamentary parties, however, reflects the multitude of possible threeparty, minimal winning coalitions.

The number of parties that competed in 2020 was similar to previous parliamentary elections. Because several parties and alliances failed to reach the threshold of representation by a relatively narrow margin, the effective number of parliamentary parties fell significantly, from 5.7 to 4.4. In total, more than 20 per cent of valid votes were cast for parties that did not enter parliament (Havlík et al. 2020). Smer's loss of popularity continued after the massive public protests against the government. Nonetheless, the party came in second to OĽaNO, which controlled more than a third of the seats. Even though the electoral system kept the fragmentation of the 2020 legislature to a relatively moderate level, the election still resulted in a somewhat complex bargaining situation. Of the ten possible minimal winning coalitions, all but one involved more than two parties.

#### Bipolar opposition

While the relatively high levels of fragmentation allowed for many different majority coalitions to form in Slovakia, the number of realistic government coalitions was often constrained by the ideological configuration of the Slovak party system.The 1992 elections were still held in a united Czechoslovakia, but the Czech and Slovak party systems were already quite independent. In Slovakia, the future of the federation, and Slovakia's role in it, was the dominant issue in the 1992 campaign. The Slovak parties fundamentally agreed on the goal of greater autonomy, but their positions differed significantly regarding how much autonomy they preferred. Initially, only the SNS advocated for a fully independent Slovak state and the end of the federation. The HZDS, however, approached this position in the run-up to the 1992 elections as well. However, fundamentally different views on the federal question existed between the Slovak and the Czech parties rather than within the Slovak party system (Szomolányi and Mesežnikov 1997, 141; Grotz 2000, 338–42; Haughton and Rybář 2008, 235). Thus, there was no bipolar opposition to constrain coalition formation in Slovakia in 1992 (see Table 5.18).

During Mečiar's first term in office, the bipolar opposition between illiberal and democratic forces that characterised the Slovak party system of the late 1990s and early 2000s, began to emerge. In the first half of the 1990s, prime minister Mečiar challenged the authority of Slovakia's democratic institutions and, with the support of the SNS, attempted to establish an autocratic and clientelist regime that stood in opposition to the economic and, in particular, the democratic transformation of the country (Carpenter 1997, 212–13; Grotz 2000, 392–93). By the 1994 parliamentary elections, however, the fronts had still not hardened, and a clear-cut bipolar opposition did not influence government formation. Mečiar himself had referred to the ousting of his government by the broad opposition alliance, centred around interim prime minister Moravčík, as a coup, but he was still prepared to cooperate with individual parties from the opposition camp, as were some parties that had voted him out of office, such as the SDĽ (Malová 1995, 469; Szomolányi and Mesežnikov 1997, 139–40; Grotz 2000, 392).

During Mečiar's second term in office, the opposition between his autocratic government and the democratic opposition intensified considerably.This divide between democratic, pro-Western forces and the autocratic, nationalist camp was the dominant issue in the 1998 Slovak parliamentary elections. Several liberal and conservative parties in the anti-Mečiar alliance merged into one party, the SDK,in order to improve their chances of winning. The SMK and the SDĽ did not join this party, but they were also firmly in the oppositional camp.The aim of this broad anti-Mečiar coalition was to bring the country back on track towards democracy and the rule of law, while ensuring the country's integration into Western alliances, most importantly the EU (Pridham 2002; Vachudova and Hooghe 2009, 201; see also Hloušek and Kopeček 2008; Haughton and Rybář 2008). Hence, at that time, the Slovak party system was clearly divided into two oppositional camps that were unable to cooperate with each other. This opposition included ideological differences and intense affective polarisation between the competing parties.

The oppositional alliance removed the Mečiar government in 1998, which ensured Slovakia's return to Europe. At the end of prime minister Dzurinda's first term in office, the bipolar opposition between the democratic forces and the Mečiar camp still dominated the 2002 parliamentary elections. None of the incumbent parties was willing to form a coalition with the HZDS, although Mečiar had promised to break with his autocratic past (Hloušek and Kopeček 2008).

In the mid-2000s, the liberal democratic regime was firmly established in Slovakia and parties competed over socio-economic and socio-cultural issues. The socio-economic dimension was most salient, but socio-cultural conflicts,in particular regarding the rights of the Hungarian minority in the country, played a role as well. Socio-economic issues also dominated the 2006 election campaign. After the collapse of the SDĽ, Smer became the strongest force on the socio-economic left and faced the SDKÚ-DS as the main liberal-conservative contender. The opposition on the secondary, socio-cultural dimension ran primarily between the radical right SNS and the SMK, the political representative of the Hungarian minority in the country (Haughton and Rybář 2008). Both conflict dimensions reinforced each other, thus constituting a bipolar opposition between a national-protectionist camp, consisting of Smer, SNS, and HZDS, and a liberal-conservative one, comprising the SDKÚ-DS, the SMK, and the KDH. This constellation somewhat resembled the situation in the late 1990s, but the affective polarisation between both camps was still mild and coalitions across camps, for instance between Smer, KDH, and SMK, remained a realistic option.

This conflict structure had further intensified by 2010. Due to the economic crisis in Europe, socio-economic issues remained high on the agenda and shaped the conflict between the two dominant parties, Smer and SDKÚ-DS. But the ethnic divide also gained salience following Fidesz' triumph in the Hungarian parliamentary elections earlier that year. The bipolar opposition was further reinforced by Fico's style of government and his controversial personality.Thus,in 2010, the Slovak party system was again divided into two camps that were unable to cooperate with each other (Haughton, Novotná, and Deegan-Krause 2011; Deegan-Krause and Haughton 2012).

When the radical right returned to parliament in 2016, the Slovak party system had undergone another transformation. Various corruption scandals had weakened Smer and the liberal-conservative camp. At the same time, several new anti-corruption and anti-establishment parties, such as OĽaNO, Sme Rodina, and ĽSNS, entered party competition (Rybář and Spáč 2016). The emergence of corruption as an important issue in Slovak politics, as well as the introduction of various new parties with different ideological backgrounds, put an end to the bipolar opposition that had constrained coalition formation at the beginning of the decade.The Slovak party system of the late 2010s was rather characterised by multi-polar oppositions.

The 2020 elections were overshadowed by the political earthquake which followed the murder of Jan Kuciak and Martina Kušnirova in mid-2018. This incident sparked the largest protests in the country since the Velvet Revolution (Havlík et al. 2020, 221–22), boosting the salience of corruption and anti-establishment sentiments even further. Thus, Havlík and his co-authors also argue that populism had a decisive impact on the electoral success of parties in 2020: "In sum, the rise of populist parties and the fact that populism became the only viable alternative to the previous government were two of the most important results of the 2020 general election in Slovakia. Yet, a closer look at the ideological and electoral background of the populist challengers provides us with a more complicated picture" (Havlík et al. 2020, 230). The different ideologies that accompanied the populist anti-establishment appeal of the new parties indicate that the oppositions in the Slovak party system have become even more diverse in 2020 than they were in 2016, despite the unanimous rejection of Smer.


*Table 5.18: Bipolar opposition in the Slovak party system*

Source: Own compilation.

## 5.4.3 Characteristics and preferences of Slovak radical right parties

## Parliamentary strength

Although the SNS has been representedin parliament relatively consistently over the past three decades, the party's electoral results fluctuate regularly (see Table 5.19). In 1992, the SNS achieved a solid result by winning 7.9 per cent of the vote, which translated into ten per cent of the parliamentary seats and substantial support for the first Mečiar government. In 1994 the party barely managed to re-enter parliament. Its meagre seat share of only six per cent rendered it the smallest parliamentary group. Somewhat untypical ofincumbent partiesin Central and Eastern Europe (Roberts 2008; Bergman, Ilonszki, andMüller 2019b), the SNS made significant electoral gains in 1998, receiving 9.1 per cent of the vote and 9.3 per cent of the seats in parliament. In 2001, the party split and spent one term in extra-parliamentary opposition before rebounding in 2006 to record a result of 11.7 per cent of the vote, becoming the third-strongest party in parliament.

Since then, however, the SNS has struggled. By 2010, it had lost half of its vote share, and in 2012, it failed to enter parliament for the second time in its history. There have even been serious doubts as to whether the SNS would be able to recover from this defeat (Gyárfášová and Mesežnikov 2015). After a change in the party leadership in 2012 and the subsequent expulsion of the notorious long-time chairman, Ján Slota, a year later (Pirro 2016, 88), the party recovered once again and returned to parliament in 2016.The 2020 parliamentary elections, revealed, however, that the SNS' revival was temporary, as it fell well short of the five per cent threshold.


*Table 5.19: Election results and parliamentary strength of radical right parties in Slovakia*

Source: Nordsieck 2021.

The emergence of a second, more radical party on the far right, Marian Kotleba's ĽSNS, has contributed to weakening the electoral support for the SNS (Gyárfášová and Mesežnikov 2015; Mesežnikov and Gyárfášová 2017). In 2016, the ĽSNS entered the national parliament for the first time after unsuccessful attempts in 2010 and 2012.The party received a surprisingly strong eight per cent of the votes, due in part to the successful mobilisation of first-time voters (Mesežnikov and Gyárfášová 2017, 32). Kotleba and his party repeated this result in 2020. Due to high levels of disproportionality in the 2020 election, the ĽSNS received 11.3 per cent of the seats in parliament.

#### Ideological distance to the formateur

#### **The radical right parties: SNS and L'SNS**

The original SNS was the first political partyin Slovakia and existed from 1871 to 1938. When it was founded in 1989, the new SNS reclaimed continuity with this historical organisation (Pirro 2016, 86). The ideology of the SNS built on an ethno-cultural idea of a Slovak national identity that involved references to the threat of Hungarian domination as well as a religious component, most evident in the attempt to rehabilitate the inter-war clerico-fascist state and its central figures, Andrej Hlinka and Jozef Tiso (Gyárfášová and Mesežnikov 2015, 230–31; Pirro 2016, 89–91; Pytlas 2016, chap. 5).

Under this ideological roof, different issues have dominated the party's platform over the past three decades. In the run-up to the 1992 parliamentary elections, when Czechoslovakia still existed, Slovak national independence was the SNS' defining theme. Before the HZDS adopted this position during and after the 1992 campaign, the SNS was the main proponent of an independent Slovakia and voiced strong resentment towards the Czech population in the federation (Cibulka 1999, 116–17; Pirro 2016, 86). Once the country gained independence, the SNS turned towards the Hungarian minority in Slovakia as its main enemy. The SNS blamed the Hungarians for all the ills of the country and accused them of supporting the irredentist policies of their kin state, Hungary. Long-time party leader, Ján Slota, was infamous for his public anti-Hungarian outbursts. Moreover, the SNS' ideological platform included racist policies that targeted the Roma minority and aimed at their social exclusion (People Against Racism and Milo 2005, 113–14; Gyárfášová and Mesežnikov 2015, 234; Pirro 2016, 95–96; Pytlas 2016, chap. 5).

When Slota was replaced as party leader by Andrej Danko in 2012, the latter attempted to moderate the party's positions. Slota's expulsion from the party in 2013 was supposed to send a clear signal in this regard. Overall, however, Danko's strategy of moderation was rather hollow and did not bring about major programmatic changes (Rybář and Spáč 2016; see also Pirro 2016, 88–91). In the second half of the 2010s, and particularly in the 2016 election campaign, the SNS also focused on the omnipresent immigration issue in an attempt to profit from widespread xenophobia. Due to the competition from other parties, such as Smer and the ĽSNS, however, these efforts remained rather unsuccessful (Harris 2019, 551; Rybář 2020, 241). In light of these socio-cultural positions, the SNS receives a GALTAN score of 9.00 in 1992 and 1994. This position resembles the party's placements by the CHES experts,

which oscillate around 9.00 as well (8.77 in 2002, 8.64 in 2006, 9.21 in 2010, and 8.94 in 2019) (see Table 5.20).


*Table 5.20: Socio-economic and socio-cultural distance between radical right parties and formateurs in Slovakia*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions. \* The SNS' placement is based on the 2002 CHES wave, which did not include SDK.

On the socio-economic dimension, the SNS has consistently followed a nationalprotectionist course characterised by "strong elements of etatism, paternalism, and an inclination to redistributive social policy"(Gyárfášová and Mesežnikov 2015, 229). At the same time, the party held a more positive view on the free market economy than other radical right parties in Central and Eastern Europe (Pirro 2016, 93). The SNS clearly favoured privatisation, but only to Slovak investors. The SNS also criticised Prague's neoliberal approach to economic transformation and advocated for a stronger welfare state (Szomolányi and Mesežnikov 1997, 143; Gyárfášová and Mesežnikov 2015, 229–31).

In the 2000s, the party's socio-economic platform became more liberal. Its 2006 election programme, for instance, included the goal of "building an efficient, competitive and effective economy" (SNS 2006, in Pirro 2016, 93), which included the deregulation of the economy as well as the support for small and medium-sized domestic businesses and agriculture. Nevertheless, the fundamental chauvinistic and paternalistic approach to socio-economic policy remained (Pirro 2016, 93). In 2010, the SNS emphasised etatism and protectionism in response to the economic crisis, proposing a plan for the state to control key industries and infrastructure while also replacing the flat tax with a progressive taxation model. In its 2010 manifesto, the SNS explicitly criticised neoliberal economics, and in 2012, it published a memorandum with an even stronger national-protectionist thrust. This document was put aside, however, once Danko became the party's new chair later that year, because, while he still preferred the state to hold a majority position in key infrastructure businesses and favoured progressive taxation, he also acknowledged the need for foreign investment and compliance with European institutions (Gyárfášová and Mesežnikov 2015, 230; Pirro 2016, 93–94). CHES scores for the SNS on the socio-economic dimension reflect the party's positions quite well. Only the position in 2006 is adjusted from 5.00 to 4.50, indicating that the SNS leaned slightly to the left. In the period before 1998, where no quantitative data are available, the SNS is placed at 3.50 on the socio-economic dimension, based on the qualitative assessment of its positions in the secondary literature.

In contrast to the SNS, Marian Kotleba's ĽSNS openly opposes the democratic system itself (Harris 2019) and can therefore be regarded as an extreme right party (see Chapter 2.1). Nevertheless, the SNS and the ĽSNS share an ideological core.The ĽSNS also embraces an ethno-religious concept of the Slovak nation and glorifies the clerico-fascist inter-war state (Mesežnikov and Gyárfášová 2017, 27–30; see also Harris 2019). However, Kotleba's rise in the first half of the 2010s was largely a result of his outspoken racism. He capitalised on widespread resentment against Roma in the Slovak population and used this issue to stand out against the SNS, which primarily targeted the Hungarian minority (Kluknavská and Smolík 2016, 341; Mesežnikov and Gyárfášová 2017, 25; Řádek and Miroslav 2019, 47–48).

During the "migration crisis" in the mid-2010s, the ĽSNS shifted its focus and campaigned on protecting the ethnic Slovak and Catholic nation from Muslim immigrants. Similar to other radical right—and even some mainstream—parties across Europe, the ĽSNS portrayed immigrants as terrorists and the men as a threat to Slovak women (Mesežnikov and Gyárfášová 2017, 27–30). Here, the ĽSNS clearly outperformed the SNS and competed for ownership of the immigration issue with Smer, which held similarly xenophobic and racist positions (Androvičová 2017). In addition, the party targeted the LGBTIQ+ community, was clearly anti-European and anti-Semitic (Kluknavská and Smolík 2016; Gyárfášová and Mesežnikov 2015; Harris 2019). Hence, the ĽSNS' GALTAN score of 9.81 in the 2019 CHES wave provides an adequate reflection of its radical socio-cultural positions.

The ĽSNS' socio-economic positions are deeply intertwined with its nativist ultranationalism. The party rejects foreign investment and the influence of the EU in the country while also seeking to re-nationalise important industries and resources. In the early 2010s, the party even called for the reintroduction of the Slovak koruna as the national currency. In addition, the ĽSNS formulated the goal of full employment and aimed to expand the welfare state, for instance by supporting families or providing free health care.These measures were designed to exclude people who are not considered members of the Slovak nation, particularly Roma (Gyárfášová and Mesežnikov 2015, 238–39; Řádek and Miroslav 2019). In light of these positions, the party's CHES score on the socio-economic dimension (3.94) seems too close to the centre and is therefore adjusted to 2.50.

#### **The formateurs: HZDS, SDKÚ-DS, Smer, and OL'aNO**

In the past three decades, the two Slovak radical right parties faced different formateurs. The first one was Mečiar's HZDS in 1992. In the run-up to the 1992 elections, Mečiar approached the SNS' position on the independence issue, using similarly ethno-nationalist rhetoric (Mesežnikov 2008, 11–12, 2009, 41–43; Učeň 2009, 29; Stanley 2011, 258–59). The HZDS was also critical of the Hungarian minority in Slovakia. The party argued, for instance, that the Hungarian minority should not be responsible for their own affairs, and Mečiar even suggested that Hungarians should be resettled in their kin state (Bugajski 2002, 294). Mesežnikov (2008, 12) notes that the positions of the nationalist wing within the HZDS did not differ significantly from the SNS. Moreover, Mečiar disregarded the values and principles of liberal democracy and sought to establish an autocratic and illiberal regime. Based on these positions, the HZDS receives a GALTAN score of 7.50 in 1992 and 1994, placing the party clearly on the nationalist end of the socio-cultural dimension but somewhat closer to the centre than the SNS.

The HZDS held rather leftist socio-economic positions, preferring a gradual transformation of the economic system and rejecting neoliberal economic policies. Initially, Mečiar's opposition to the orthodox neoliberalism of Václav Klaus' ODS in Prague (see Chapter 5.1) was the main reason for his scepticism towards the continuation of the Czechoslovak federation in the early 1990s (Szomolányi and Mesežnikov 1997; Bugajski 2002, 312; Vachudova 2008, 870). At that time, the socioeconomic platform of the HZDS favoured state regulation of the economy and an incremental approach to the privatisation of state property, including preferential treatment for Slovak investors. Clientelism and cronyism were also an essential part of the party's policy in this context, though not in its official programme (Szomolányi and Mesežnikov 1997, 143; Fisher 2006, chap. 4). Therefore, the party is placed at 3.00 on the socio-economic dimension in 1992 and 1994, reflecting a moderate centre-left position.

In 1998, the SDK was tasked with government formation. Since the SDK united several parties, its positions were not always homogeneous. On socio-economic issues, however, the party clearly stood for market liberalism, but its platform also included some elements of a welfare state. In the socio-cultural sphere, the SDK shared the goal of reinstating a liberal democratic regime, and the rule of law, after Mečiar's illiberal rule. The party's agenda also entailed Slovakia's quick integration into NATO and the EU. In addition to the liberals, there were also conservative and Christian democratic factions within the party.They did not adhere to the ethno-religious nationalism found in the SNS and the HZDS, however, which was evident in their positive attitude towards the Hungarian minority (Bugajski 2002, 301; Fisher 2006, 162–64). Since SDK's liberal and pro-democratic positions outweigh the conservative tendencies in the party in 1998, it receives a GALTAN score of 4.50, which places the party on the liberal side of the socio-cultural dimension. As regards socioeconomic issues, the dominance of the liberal wing results in a score of 7.50.

In 2010, the re-organised successor of the SDK, the SDKÚ-DS, acted as the formateur. In the 2010 parliamentary election, the SDKÚ-DS positioned itself as a liberal-conservative party with a clear focus on socio-economic issues. Neoliberal ideas, such as support for privatisation, deregulation of the health sector, and the defence of the flat tax featured in the party's economic programme. Since the electoral decline of the HZDS had ruled out the return of Mečiar, the SDKÚ-DS could no longer rally behind opposition to his regime. Despite somewhat stronger conservative currents, when compared to the SDK in the late 1990s, the party remained fundamentally pro-Western, secular, and supportive of the Hungarian minority (Malová and Učeň 2007, 1105; Haughton and Rybář 2008, 237; Štefančík 2008), which is reflected in its GALTAN placement in the 2010 CHES wave (5.14).

In 2006 and 2016, Smer won the election and received the mandate to form the government. In the beginning, the party's ideological platform was rather vague and party leader Fico presented his nominally social democratic party as a third way between the two oppositional camps that had shaped Slovak politics during the turn of the millennium (see above). Over the years, however, Smer's centre-left socio-economic profile consolidated. In this regard, the party filled the void left by the SDĽ after its demise. In the 2006 election campaign Smer opposed SDKÚ-DS' plans for privatising the health and energy sectors. Fico also wanted to introduce progressive taxation and formulated the goal of establishing a comprehensive welfare state in Slovakia—often using populist anti-establishment rhetoric (Malová and Učeň 2007, 1106; Haughton and Rybář 2008, 244; Mesežnikov 2008, 10; Pytlas 2016, 34–35; Malová 2017, 10). In 2016, Smer was still a centre-left party regarding socio-economic issues, but after two terms in government,its positions had moved somewhat closer to the centre (Malová 2017, 10–11). The party's CHES scores of 2.36 and 3.50, respectively, mirror this positional shift.

Smer's socio-cultural positions were out of step with other parties in the social democratic party family, which are usually located at the GAL end of the spectrum (Hloušek and Kopeček 2010). Moreover, Pytlas (2016, 142) notes that in the mid-2000s, "the value profile of the party was an enigma" to many observers. However, Smer occasionally used nationalist rhetoric during the 2006 campaign, deliberately playing the anti-Hungarian card in order to appeal to culturally conservative voters (Mesežnikov 2008, 10; Rybář and Deegan-Krause 2008, 511; see also Pytlas 2016, chap. 5). During Fico's first term in office from 2006 to 2010, Smer's framing and policies revealed the party's nationalist leaning (Pytlas 2016, 47–50; see also Pytlas and Kossack 2015). By 2016, there was little doubt about Smer's sociocultural positions. Fico and his party were among the most prominent voices stirring hatred and mistrust against immigrants and refugees. He opposed a European quota system for the distribution of refugees, and he did not refrain from using blatant racism, repeatedly associating Islam and refugees with terrorism (Androvičová 2017; Rybář and Spáč 2016). In contrast to Smer's previous campaigns which focused on socio-economic issues (Rybář and Deegan-Krause 2008), the growing salience of socio-cultural issues for Smer is also illustrated by the fact that the party used "We will defend Slovakia!" as central campaign slogan in 2016 (Baboš and Malová 2017, 237). Smer's GALTAN score in the 2006 CHES wave (4.43) reflects the party's enigmatic socio-cultural position at the time. Overall, however, Smer's occasional use of nationalist rhetoric in the 2006 campaign and the policies it pursued during its subsequent term in office point to a moderate TAN instead of a moderate GAL position. Hence, the party's placement on the socio-cultural dimension in 2006 is adjusted to 6.50. In 2019, the CHES places Smer even closer to the TAN pole (7.88), matching the party's rightward shift.

The 2020 parliamentary elections changed the political landscape in Slovakia significantly and presented a relatively new party, OĽaNO, with the opportunity to form a government. OĽaNO has been established in 2011, but other than being anti-establishment, it had a vague ideological profile.<sup>7</sup> Consequently, observers characterised OĽaNO as "pro-conservative, but with eclectic and incoherent positions" (Bútora 2013, 20). By the 2020 elections, there was at least some evidence that OĽaNO held rather liberal socio-economic views (Rybář and Spáč 2016). Moreover,

<sup>7</sup> Following the anti-establishment appeal of its founding members, first and foremost Igor Matovič, OĽaNO was not registered as a political party but as a political movement, which made no difference in practice (Hloušek, Kopeček, and Vodová 2020, 104).

the party had expressed support for conservative Catholic values, such as a traditional image of the family and the opposition to a liberal abortion policy (Havlík et al. 2020, 218; Hloušek, Kopeček, and Vodová 2020, 107–8). OĽaNO's CHES scores of 6.00 on the socio-economic and 6.97 on socio-cultural dimension mirror the qualitative assessment.

#### 5.4.4 Summary

Slovakia has been the Central and Eastern European country where the radical right has been in government for the longest time during the three post-Communist decades.The SNS has participated in five coalitions and held public office for a total of 14 years. These coalitions were characterised by a relatively small socio-economic range, which highlights the important role of the radical right for the country's socio-economic left. The coalition of Smer, SNS, and HZDS in 2006 was also quite homogeneous on the socio-cultural dimension. Due to the lack of data on the positions of Slovak parties in the 1990s, it is not possible to draw clear conclusions as to whether the 1992 and 1994 governments were minimal connected winning or minimal range coalitions, but the ideological proximity of the HZDS and SNS suggests that their socio-economic and socio-cultural range was rather small.

Only the unorthodox Fico III and IV governments, which held office between 2016 and 2020, were ideologically rather heterogeneous. Several factors contributed to prime minister Fico's ability to successfully form a government coalition despite considerable socio-economic and socio-cultural differences, particularly between the SNS and Most-Híd. First, Slovak parties were rather sceptical towards forming minority governments in general. Secondly, Slovakia was scheduled to preside over the Council of the European Union, and leaders wanted to avoid forming either a caretaker or minority government during this period. Third, changes in the SNS' leadership and the political representation of the Hungarian minority enabled the formation of the four-party government that included both these antagonistic forces. Within the SNS, long-time chairman and most aggressive anti-Hungarian voice in the party, Ján Slota, had been replaced by the more moderate Andrej Danko after the party's electoral defeat in 2012. One year later, Slota was even expelled from the party. On the other side, the SMK had been replaced by Most-Híd as the representative of Hungarian minority interests in Slovak politics. The new party did not carry the baggage of the deep bipolar opposition between the pro- and anti-Mečiar camps of the late 1990s. Whether the SMK and the SNS under Slota's leadership would have been able to cooperate in a coalition government seems rather questionable.

More recently, the SNS has struggled at the polls and it is uncertain whether the party will be able to recover from its electoral defeat once again. The radical right did not disappear from the Slovak parliament with the SNS in 2020, though. Marian Kotleba's ĽSNS, which entered national parliament already in 2016, repeated this success in the 2020 parliamentary elections. The other parties, however, have shown no inclination to lift the cordon sanitaire against the ĽSNS that they had established right after Kotleba's breakthrough in 2013, when he surprisingly beat the Smer candidate in the second round of the regional elections and became governor of the Banská Bystrica region (Mesežnikov and Gyárfášová 2017, 21). Thus, Slovak governments might not include a radical right party anytime soon. Yet, even if this is the case, three decades of party competition with the radical right have primed the mainstream with radical right politics, most of all Robert Fico's Smer.

## **6. Government formation and the radical right in Northeastern and Southeastern Europe: The Baltics and the Balkans**

Afterintroducing radical right parties and government formationin the four Central European countries in the previous chapter, this one turns to the two Baltic states, Estonia and Latvia, and the two Southeastern European countries, Bulgaria and Romania. Again, each case study first outlines the formation and composition of governments when radical right parties were present in parliament, before discussing the configuration of party systems, as well as the parliamentary strength of the radical right and their ideological distance to the formateur.

## **6.1 Estonia**

#### 6.1.1 Government formation with radical right parties in Estonia

The radical right Estonian National Independence Party (ERSP) entered government after Estonia's first free elections in 1992. The ERSP became a junior partner in a coalition government under Mart Laar, the leader of the victorious national-conservative Pro Patria alliance.The third party in the 1992 coalition was the Moderates (M). In the run-up to the 1995 parliamentary elections, Pro Patria and the ERSP merged into the Pro Patria Union which is not considered a radical right party (see Chapter 2.1).

It was not until 2015 that another radical right party, the Conservative People's Party of Estonia (EKRE), entered parliament. However, EKRE faced a cordon sanitaire and was thus never considered as a candidate for government by the other parties (Braghiroli and Petsinis 2019, 439). The 2015 government was formed by the Reform Party (ER), a neoliberal party that had split from Pro Patria in the 1990s, the Moderates, now renamed Social Democratic Party (SDE), and the national-conservative Isamaa (I). However, this coalition lasted only until 2016, when the Centre Party (EK) replaced the ER as prime minister party of the coalition. At that time, there was a change of leadership within the EK, the largest opposition party, which had not governed for more than a decade due its controversial party leader, Edgar Savisaar (Lagerspetz and Vogt 2013). When Savisaar was replaced by Jüri Ratas in 2016, the two junior partners left the incumbent government almost immediately to form a new government under Ratas and the EK (Mölder 2017).

The 2019 Estonian parliamentary elections saw a close race between ER and EK. Both parties distanced themselves from EKRE during their campaigns, but postelectoral government formation revealed the fragility of the cordon sanitaire surrounding the radical right. After the ER won the elections, everything seemed set for the parties to form either a grand coalition of ER and EK or a three-party union of ER, SDE and Isamaa (Hartleb 2019). However, Jüri Ratas, EK's top candidate, was unwilling to settle for the role of junior partner in a government led by his biggest rival. Despite public protests against EKRE's inclusion in the Estonian government, both domestically and internationally, he broke the cordon sanitaire and formed a majority coalition with EKRE and Isamaa (Rankin 2019; Mölder 2020).This coalition did not survive the full term, however, and was replaced by the grand coalition of ER and EK in January 2021, although without former prime minister Ratas.

Most coalitions formed in Estonia since 2000 have been minimal winning coalitions, and this includes those with radical right parties (see Table 6.1). In the 1990s, however, this format was rare. Of the six governments formed in Estonia before 2000, only two were minimal winning coalitions, including the 1992 Pro Patria-ERSP-M government. The other four governments were either minority governments or oversized coalitions (Pettai 2019, 186).

Data on the ideological range of the coalitions is only available from the early 2000s onwards and thus only for the governments formed while EKRE was present in parliament. All three coalitions show a relatively broad ideological range on both the socio-economic and socio-cultural dimension. The 2015 and 2016 governments fulfil neither the criteria for the minimal range nor the minimal connected winning theory. The "conservative coalition" (Walker 2019; Mölder 2020, 119) of EK, Isamaa and EKRE, however, was socio-economically and socio-culturally connected. Despite an ideological distance of almost three points, this three-party coalition also had the smallest possible socio-cultural range in parliament, thus rendering it a minimal range coalition on this dimension.

*Table 6.1: Format and ideological range of governments in Estonia*


Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in paren- theses indicate the smallest possible ideological distance of a majority government

## 6.1.2 The configuration of the Estonian party system

## Fragmentation

As in many Central and Eastern European countries, the Estonian party system of the early 1990s was highly fragmented. Almost 40 parties competed during the first free electionsin 1992 and seven parties or electoral alliances entered parliament (Pettai and Kreuzer 1998). The effective number of parties in the 1992 Estonian parliament was 5.9 (see Table 6.2). The bargaining situation was quite complex, since any majority coalition required at least three parties. Even Pro Patria, which won the elections quite comfortably, controlled less than 30 per cent of the seats.

Since the mid-2000s, however, the Estonian party system has undergone structural consolidation resulting in, among other things, a low number of new parties and moderate levels of fragmentation by Central and Eastern European standards (Lagerspetz and Vogt 2013; Enyedi and Casal Bértoa 2018, 430–31). The bargaining situation in 2015 was still relatively complex, despite a significantly lower number of effective parliamentary parties. The two strongest of the six parties in the Riigikogu, however, controlled a similar number of seats (30 and 27 seats), as did the two medium-sized (15 and 14 seats) and small parties (eight and seven seats). As a result, apart from the grand coalition, it would have taken at least three parties to reach the 51-seat majority, and neither of the two large parties had a decisive advantage. The fragmentation of the Estonian party system decreased further after the 2019 election to 4.2 effective, and five actual, parliamentary parties, reducing the number of possible minimal winning coalitions to five. Due to the relatively equal distribution of seats between the parties, however, the bargaining situation was not straightforward, since all parties had more than one option to enter government.


*Table 6.2: Fragmentation of the Estonian party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

#### Bipolar opposition

In the early 1990s, the ethnic divide between the Estonian majority and the large Russian-speaking minority was at the centre of Estonian politics and deeply intertwined with the regime divide, which also entailed issues related to the speed of economic reform and the urban-rural cleavage (Duvold, Berglund, and Ekman 2020, 42–45). The proportion of ethnic Estonians in the country had fallen from 95 to 61 per cent between 1945 and 1989 as a result of Moscow's Russification policy (Saarts 2011, 96–97; Pettai 2019, 174). After independence, the Estonian majority introduced an "ethnic democracy" (Braghiroli and Petsinis 2019, 438), which meant that the "Estonian political system, its polity, continuously obtains its basic energy from a discourse portraying it as a national *Gemeinschaft*, a community of ethnic Estonians" (Lagerspetz and Vogt 2013, 66, italics in original).Thus, nativism was a vital element of the Estonian political mainstream at that time (Mudde 2007, 53–54).

Under this ethnic democracy policy,members of the Russian-speaking minority could only obtain citizenship, and thus the right to political participation, through the process of naturalisation. By the end of 1991, more than a quarter of Estonia's 1.5 million residents were stateless and did not have the right to vote in the 1992 parliamentary election. Neither did they have the right to stand for election, which prevented the emergence of Russian-speaking minority parties in the country. A majority of Estonian parties favoured leaving the 1991 citizenship regulation untouched after the 1992 elections, so these restrictions remained unchanged for several more years. Even after large parts of the Russian-speaking minority had completed the naturalisation process and received the right to vote, no relevant party emerged to cater exclusively to minority interests. The majority of the Russian-speaking electorate rather supported existing parties, in particular the EK (Pettai 2019, 175).Thus, there was no bipolar opposition in the Estonian party system during the early 1990s (see Table 6.3).

In light of the broad consensus among the parties regarding the minority issue, socio-economic divides gained importance in Estonia from the mid-1990s to the mid-2000s, although the ethnic divide never disappeared completely. As the Russian-speaking minority was more supportive of redistributive socio-economic policies than the Estonian majority in the country, their underrepresentation in the political arena resulted in a party system lacking a representative on the socio-economic left. In the 2000s, the main opposition in the Estonian party system ran between a "national-neoliberal camp"(Lagerspetz and Vogt 2013,66), consistingmainly of ER and Isamaa, and the more centrist EK, which received most of its support from the Russian-speaking minority and leaned slightly to the left on socio-economic issues.The SDE sided with the EK, but its socio-economic positions were more liberal than the party's label suggests (Saarts 2011, 96–97; Pettai 2019, 174). Lagerspetz and Vogt (2013, 55) even argue that the EK and SDE could pass for centre-right parties in other European countries (see also Reetz 2011), as is also indicated by their membership in the liberal ALDE Party in the European parliament.

In the second half of the 2000s, the relocation of a Soviet soldier's memorial from the centre to the outskirts of Estonia's capital Tallinn caused the ethnic divide to resurface and escalate. The ER took a more nationalist stance in order to avoid losing votes to its conservative rival, Res Publica (Lagerspetz and Vogt 2013; Nakai 2014).The EK still represented the main opposition to the national-neoliberal camp, both in socio-economic terms and with regard to the ethnic divide. However, this opposition in the party system was less pronounced than in the population. Despite their general criticism of the ER, the EK and the SDE also supported neoliberal economic policies when in government (Reetz 2011; Lagerspetz and Vogt 2013). Thus, a significant sector of the population, particularly the Russian-speaking minority, found itself underrepresented at the national level (Lagerspetz and Vogt 2013; Pettai 2019). Pettai (2019, 175) concludes that "there are a number of important divides between the two ethnic communities, including major socioeconomic disparities, which may not always be addressed by dominantly ethnic Estonian political parties. Public opinion polls show that ethnic Russians tend to hold more centre-left views; still, because they are less present in electoral politics, the electoral landscape is by default shifted to the right." Hence, when EKRE entered parliament in the 2010s, the Estonian party system remained rather unipolar. There was an opposition between the national-neoliberal camp and the centrist EK, but the polarisation between them remained moderate and coalitions of parties from both sides were possible.


*Table 6.3: Bipolar opposition in the Estonian party system*

Source: Own compilation.

## 6.1.3 Characteristics and preferences of Estonian radical right parties

## Parliamentary strength

The ERSP received 8.8 per cent of the vote in the 1992 parliamentary elections, which earned it ten of the 101 seats in the Riigikogu (see Table 6.4). Even though the radical right was only the fifth strongest of the seven parties in parliament, its contribution of almost ten per cent was quite substantial in the highly fragmented Estonian party system.


*Table 6.4: Election results and parliamentary strength of radical right parties in Estonia*

Source: Nordsieck 2021.

EKRE achieved a similar result at the polls in its electoral breakthrough in 2015. However, the party's vote share of 8.1 per cent won it only seven seats, making it the smallest party in parliament. In the following parliamentary elections of 2019, EKRE won an impressive 17.8 per cent of the vote and almost tripled its seat share to 19 seats. As a result, it became the third strongest party in parliament and had much more bargaining power than in the previous legislature.

#### Ideological distance to the formateur

#### **ERSP and Pro Patria**

Research on party politics and the radical right in the Baltic states is even rarer than in other Central and Eastern European countries which makes evaluating the positions of the ERSP and its competitors relatively difficult (Auers and Kasekamp 2009, 242; Mudde 2018, 260). Only the recent success of EKRE and the National Alliance in Latvia brought more attention to this region. There is, nonetheless, little doubt that the ERSP was positioned clearly in the nativist, ultranationalist spectrum of Estonian politics. Pettai and Kreuzer (1998, 168) state that the "opposition to a liberal citizenship law has been the raison d'être" for the ERSP.

Similar to the majority of Estonian parties, the national-conservative alliance Pro Patria agreed with the radical right ERSP on the most important socio-cultural issue of the time—the rights of the Russian-speaking minority in the country. Both parties, like most of their competitors, supported the emigration of Russian-speaking Estonians (Bugajski 2002, 72–73; Kasekamp 2003, 404; Poleschuk 2005, 60). The main difference between the ERSP and Pro Patria was the latter's more moderate tone.Thus, this study places both parties clearly at the TAN end of the socio-cultural dimension. The slight difference in their position is reflected in scores of 9.50 and 8.50 for the ERSP and Pro Patria, respectively (see Table 6.5).


*Table 6.5: Socio-economic and socio-cultural distance between radical right parties and formateurs in Estonia*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions. \* EKRE was not included in the 2014 CHES wave. Therefore, the 2019 placements are used in 2015 and 2016 as well.

On the socio-economic dimension, which was of secondary importance to the Estonian party system in the early 1990s, information on the positions of the two parties is even scarcer and more general. Overall, both the ERSP and Pro Patria clearly supported the introduction of a free market economy. The parties advocated for the country's quick accession to NATO and the EU in order to limit Russian influence. Thus, the parties' support for rapid economic transformation also reflected their scepticism towards the Russian-speaking minority and their kin-state, Russia. While socio-economic issues hardly played a role in ERSP's programmatic documents, the literature states that the majority of the party's members expressed a clear preference for privatisation and a market economy (Pettai 2012, 84). Consequently, Pettai and Kreuzer (1998, 154) conclude that "Estonia got the shock-therapy market reform it had been promised by the nationalist parties" (see also Pettai 1993, 118). In light of the limited information available, both parties shall be placed on the liberal side of the socio-economic spectrum. The ERSP receives a score of 6.00, and Pro Patria is assigned a score of 7.00, reflecting the party's slightly more liberal position.

#### **EKRE, the Reform Party, and the Centre Party**

The core of EKRE's ideology is the concept of a *völkisch*, ethnic Estonian nation. The party's white supremacist ideology is evident in statements such as, "If you are black, go back!", made by Martin Helme, one of the party's leading figures, when talking about the ideal Estonian nation (Helme 2013, in Winkelmann 2018, 18). In its 2012 manifesto, EKRE also advanced a traditional Christian concept of the family, thus excluding the LGBTIQ+ and Muslim communities (Wierenga 2017; Winkelmann 2018).

Initially, the party's main enemy was the Russian-speaking minority and its kinstate, Russia. EKRE has accepted that many ethnic Russians and Russian speakers live in the country; however,it advocated for the introduction of Estonian as the only national language and rejected dual citizenship. The party also treated those members of the minority who remained stateless like a fifth column. Additionally, EKRE's programmatic documents included irredentist claims against Russia, seeking to restore the borders of the interwar period (Winkelmann 2018, 18). The Bauska Declaration, a joint manifesto of radical right parties from Estonia, Latvia, and Lithuania signed in 2013, highlighted the importance of cross-border cooperation and shared anti-Russian sentiments among the radical right in the Baltics. Here, EKRE and its Baltic brethren, diverged from the many European radical right parties that cooperated with Putin's Russia (Wierenga 2017; Braghiroli and Petsinis 2019).

Similar to many other radical right parties in Central and Eastern Europe, EKRE emphasised immigration as a core issue beginning in the mid-2010s. Large parts of the Russian-speaking minority in Estonia shared the dislike for refugees and Muslims. This common position acted as a bridge, connecting the radical right and the Russian-speaking minority, thus making cooperation between these two camps possible for the first time (Wierenga 2017; see also Petsinis 2019). At this point, Wierenga (2017, 16) describes EKRE as "an ethnic nationalist party that wishes to conserve the ethnic makeup of Estonia, but would consider Russian-speaking Estonians who speak Estonian and pledge their allegiance to Estonia as allies in light of the refugee crisis".

On the socio-economic dimension, EKRE tends to take protectionist positions and puts the economic interests and social welfare of the imagined Estonian community first. For instance, the party called for an increase in agricultural production in order to make the country independent of exports, demanded restrictions on land acquisition by foreign investors (Winkelmann 2018, 17), and favoured placing higher taxes on foreign capital (Petsinis 2019, 220). EKRE's programme also contained proposals to strengthen the welfare state. In line with the neoliberal consensus in the country, however, the party also emphasised that economic growth and support for the Estonian economy were key elements for achieving social prosperity (Braghiroli and Petsinis 2019, 439n28). Thus, EKRE's socio-economic positions provide another example of the social-national economics typical of (Central and Eastern European) radical right parties. Unlike many other members of the radical right party family in the region, however, EKRE places less emphasis on redistributive policies and is thus located only slightly left of centre on socio-economic issues in the CHES.

EKRE was present in parliament for three consecutive government formations. In 2015, the ER acted as the formateur, while the EK played that role in 2016 and 2019. The ER has been the most vocal proponent of neoliberal economic policies in the country (Lagerspetz and Vogt 2013; Saarts and Saar 2020). Following the model of Southeast Asian states, the party regards Estonia as the "Nordic Tiger" and places more importance on economic growth than on social security (Jakobson et al. 2012, 64–65).

Initially, the ER's socio-cultural profile had been less nativist and less anti-Russian than other centre-right parties in Estonia. Since the second half of the 2000s, however, nationalist tendencies have become ascendant, as illustrated by its uncompromising positioning during the conflict over the relocation of the statue of the Soviet soldier in Tallinn (Nakai 2014, 78–79). The party has taken a liberal position on other socio-culturalissues, such as supporting same-sexmarriage, however(Mölder 2016). While the CHES data reflects ER's neoliberal socio-economic positions quite well, the party's GALTAN placement in the 2016 wave (3.13) seems a bit too liberal. The party held liberal views on various socio-cultural issues, but it took tough positions towards the Russian-speaking minority, which reflects the presence of nationalist currents in ER's ideological platform. Therefore, the party's GALTAN position is slightly adjusted to 4.00.

The EK takes centre-left positions on the socio-economic dimension, even though the party has not consistently enacted this type of legislation when in power (Reetz 2011; Lagerspetz and Vogt 2013). In its programmatic documents, however, the party advocates for progressive taxation, an increase of pensions, and more extensive social benefits, in particular for families (Jakobson et al. 2012, 64–69; Mölder 2018, 91). The EK takes a rather favourable stance towards the Russianspeaking minority (Nakai 2014), although it has never completely rejected an ethnocultural concept of the nation (Jakobsen et al. 2012: 61). Otherwise, the literature describes the party as culturally conservative (Walker 2019; Mölder 2020, 119), for instance regarding issues, such as LGBTIQ+ rights (Mölder 2016, 90). Peiker (2016, 115) also mentions an "illiberal top-down style of government". The 2014 CHES wave reflects these descriptions of the party's socio-cultural positions and places it at the moderate end of the GALTAN dimension (6.63), which seems more adequate than the score of 4.62 in 2019. Hence, EK's socio-cultural position of 2014 will also be used in 2019. The party's socio-economic positions in the left centre are well represented in both relevant CHES waves.

#### 6.1.4 Summary

The radical right ERSP entered the Estonian parliamentin 1992 and wasimmediately includedin government. In a nascent party system that was dominated by the ethnolinguistic divide between the Russian-speaking minority and the Estonian majority, the coalition of three nationalist parties cemented the ethnic democracy model in the country with restrictive citizenship and language laws.Thus, the radical right in Estonia has successfully contributed to curbing the rights and the political representation of the Russian-speaking minority from the very beginning. ERSP's participation in government, and its merger with the national-conservative Pro Patria in 1995, illustrates a general openness to the radical right's nativist ultranationalism in Estonian politics during the early 1990s.

The circumstances were different when EKRE made its electoral breakthrough in 2015. The new radical right party was initially ostracised by its mainstream competitors, but the cordon sanitaire did not hold for long. Although EKRE continued to face fierce opposition from some parties and parts of the population, the EK nevertheless decided to form a coalition with the radical right in 2019. The cooperation between these parties is even more remarkable given that EKRE has been the most outspoken, anti-Russian party in the country, and EK has enjoyed the greatest support among the Russian-speaking minority. Cooperation between the two parties was facilitated by their shared opposition to refugees and Muslims in the context of the "migration crisis" (Wierenga 2017, 14). Even though this coalition did not last the entire term and EKRE found itself on the opposition bench after less than two years in power, the inclusion in government lent additional credibility to the young radical right party.

## **6.2 Latvia**

#### 6.2.1 Government formation with radical right parties in Latvia

The radical right National Alliance (NA) entered the Latvian parliament for the first time in 2010, and since then,it has consolidated its position within the Latvian party system. The party remained in opposition in 2010, although future prime minister, Valdis Dombrovskis, seriously considered including the party in his cabinet. Dombrovskis, leader of the victorious electoral alliance Unity, intended to form an oversized coalition with the Union of Greens and Farmers (ZZS) and the NA. However, one of Unity's members, the Society for Political Change, vetoed the proposal to include the radical right, thus putting an end to the NA's hopes of entering government. Ultimately, the short-lived Dombrovskis III government of 2010 was backed only by the minimal winning coalition consisting of the Unity alliance and the ZZS (Auers 2011).

The 2011 snap election marked the first victory of the centre-left Harmony Centre (SC), the main representative of Latvia's Russian-speaking population. This party, however, did not succeed in forming a coalition, so the mandate passed to former president Zatlers' Reform Party (ZRP) which had edged out the Unity alliance for second place. Both ZRP and Unity wanted to cooperate, but they were short of a parliamentary majority by nine seats, leaving Zatlers to choose between inviting the NA into a three-party coalition and forming a grand coalition with the SC. Initially, he preferred the second option, but the prospect of governing with the Russophile SC upset large parts of the Latvian majority in the country, including members of the ZRP who threatened to withdraw their support in the investiture vote. As a consequence of the public protest and ZRP's intra-party revolt, Unity's incumbent prime minister Dombrovskis retained power by forming a coalition of Unity, ZRP, and NA (Auers 2012, 7). As six MPs left ZRP's parliamentary group almost immediately after the elections, the coalition was formally a minority government, controlling only 50 of the 100 seats in the Saeima (Ikstens 2012).

At the end of 2013, however, Dombrovskis resigned after a supermarket collapsed in Riga, causing 54 people to lose their lives (Ikstens 2014). Dombrovskis' fellow party member, Laimdota Straujuma, was sworn in as the new prime minister in early 2014. This transition also involved a change in the composition of the coalition, as Straujuma invited president Andris Bērziņš' ZZS to become a junior partner. This extended coalition now controlled a majority in parliament, even without the support of the six independents (Ikstens 2015).

The SC emerged victorious from the 2014 parliamentary elections, but it again failed to find coalition partners. The right to form a government passed to the incumbent prime minister, who decided to continue cooperation with the previous junior partners. Since the ZRP had disintegrated and joined Unity, the Straujuma II government consisted of three parties—Unity, ZZS, and NA (Ikstens 2015). In 2016, however, the chief executive was challenged by her internal rival, Unity party leader Solvita Āboltiņa, who sought to replace Straujuma as prime minister. However, Āboltiņa's candidacy was not supported by a majority in parliament. Thus, the ZZS took the chance and nominated its own candidate, Māris Kučinskis, who was ultimately elected as the new prime minister with the support of Unity and the NA (Auers 2016).

The 2018 parliamentary elections brought a significant change in the Saeima's composition. The collapse of Unity led to a reorganisation in the centre-right camp, including the emergence of various new parties (Ījabs 2018). The SC was the strongest party in the seven-party parliament, despite winning less than 20 per cent of the vote. Since a coalition capable of crossing the ethno-linguistic divide was out of the question (see below), it took three attempts for Arturs Krišjānis Kariņš, leader of New Unity (JV), the smallest parliamentary group and Unity's direct successor, to cobble together a five-party majority government. Besides JV, this coalition included the populist party Who owns the State? (KPV-LV), the New Conservative Party (JKP), the liberal party Development/For! (AP!), and the radical right NA.

Except for the 2011 minority coalition, all governments formed in Latvia while the radical right NA was present in parliament were minimal winning coalitions (see Table 6.6). In order to achieve majority status, the governments often required a large number of parties, most notably the five-party coalition of 2018. But even the Straujuma I government was supported by four parties. The analysis of the socioeconomic and socio-cultural range of these coalitions reveals an interesting pattern. All five coalitions thatinclude the NA are characterised by a high degree of socio-economic homogeneity. Three of them are even minimal connected winning and minimal range coalitions on the socio-economic dimension. This would also be true of the Dombrovskis IV government if the independent MPs were counted as members of ZRP's parliamentary group. Even the five-party coalition of 2018 is quite homogeneous, having a socio-economic range of just over two points. On the socio-cultural dimension, however, the coalitions that included the NA were quite heterogeneous. The most striking example is the Kariņš government of 2018. Here, more than six points separated the coalition partners on the GALTAN dimension. The only coalition in the 2010s that did not include the NA has a small ideological range on both ideological dimensions, even though it is neither a minimal range nor a minimal connected winning coalition. Based on a single case, however, it is impossible to determine whether or not this difference between coalitions with and without radical right parties is part of a general pattern.


Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\*Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in parentheses indicate the smallest possible ideological distance of a majority government.

*Table*

*6.6:*

*Format*

*and*

*ideological*

*range*

*of*

*governments*

*in*

*Latvia*

#### 6.2.2 The configuration of the Latvian party system

#### Fragmentation

The fragmentation of the Latvian party system was relatively high in the first two decades after the fall of Communism, even in comparison with other Central and Eastern European countries (Enyedi and Casal Bértoa 2018, 443). Low requirements for founding new parties and legal regulations that make political campaigning highly dependent on external financial resources have contributed to the constant emergence of new parties, opening the door to a personalisation and oligarchisation of Latvian party politics (Auers 2013). The low fragmentation of the party system in 2010, when it comprised only 3.9 effective parliamentary parties (see Table 6.7), is an exception and the lowest value in Latvia's three post-Communist decades (Casal Bértoa 2021). The low number of parties, and the uneven distribution of parliamentary seats, left them with a choice of only three possible minimal winning coalitions.

Since then, however, the fragmentation has been increasing consistently,mostly as a result of the disintegration of existing parties and the formation of new ones to replace them. The 2011 Saeima consisted of five parties, but their seats were more evenly distributed than in 2010. The effective number of 4.5 parliamentary parties points to a more complex bargaining situation, in which the parties could form five minimal winning coalitions, and each party had at least two options to enter government.<sup>1</sup> In 2014, the effective number of parliamentary parties in Latvia rose to 5.1, which indicates a further increase in complexity. The parliament consisted of seven parties, and no two-party coalition could achieve a majority. Fragmentation peaked at 6.4 effective parliamentary parties after the 2018 elections, marking the highest value across all countries. The complexity of the bargaining situation is reflected by the fact that it took three rounds and required five parties to form a majority government.

<sup>1</sup> This number does not include the six independent MPs as a separate parliamentary party. The count of possible minimal winning coalitions in this legislature considers them members of their original party, ZRP.


*Table 6.7: Fragmentation of the Latvian party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

## Bipolar opposition

The ideological configuration of the Latvian party system has been relatively stable over the last three decades (see Table 6.8). After gaining independence in 1991, Latvia introduced strict citizenship and language laws, especially in relation to public administration, which resembled the Estonian ethnic democracy model. Similar to Estonia, the proportion of ethnic Russians in the Latvian population had risen from about ten per cent in 1935 to over one-third in 1989 (Auers 2013, 96). Hence, the ethno-linguistic divide was essentially connected to the regime divide in Latvia as well (Duvold, Berglund, and Ekman 2020). While Estonia introduced gradual reforms to reduce the nativist elements in its political system, Latvia's minority policies have remained relatively strict and exclusive over the past three decades (Nakai 2014). The ethno-linguistic divide between Latvian-speaking and Russian-speaking parts of the population dominated party competition in the country. Nativism has been an essential part of ethnic Latvian parties' DNA since the 1990s (Mudde 2007, 53–54), and it remains significant (Auers 2013, 95–101; Braghiroli and Petsinis 2019, 438).

The SC represented one side of the ethnic divide. The party enjoyed the support of the majority of Latvia's Russian-speaking electorate, took a positive stance towards Russia, and even signed a formal agreement of cooperation with Putin's United Russia in 2009 (Ikstens and Balcere 2019, 258).The ethnic Latvian camp comprised various—and changing—parties. The parties within this camp can be distinguished between moderate and radical nationalists, the latter category including the radical right NA (Auers 2013; see also Reetz 2011; Ikstens and Balcere 2019). Socio-economic conflicts also played a role in the Latvian party system, but they were clearly secondary to the ethno-linguistic divide. Both conflict dimensions were reinforcing each other (Saarts 2011, 96–97; Auers 2013).The SC combined its pro-Russian stance with a centre-left socio-economic profile. Being the only supplier of left-wing economic policies in the country, the SC used these policy positions to mobilise support from ethnic Latvian voters who shared these preferences (Auers 2013). The liberal socio-economic positions of the ethnic Latvian parties differed only in degree, leaving the centre-left entirely to the oppositional SC. Hence, the national-neoliberal camp introduced to describe the parties of the ethnic majority in the Estonian party system(Lagerspetz and Vogt 2013) provides an adequate description of the ethnic Latvian parties as well.<sup>2</sup>


*Table 6.8: Bipolar opposition in the Latvian party system*

Source: Own composition.

The polarisation between the SC and the ethnic Latvian camp was so intense that coalitions across camps were impossible. Braghiroli and Petsinis (2019, 438) even conclude that "the major concern of the mainstream centre-right and conservative parties […], following the outcome of the 2010, 2011, and 2014 elections, was to prevent Harmony from forming a government. Therefore, a cordon sanitaire was built around Harmony".The situation did not change in 2018.The gap between the camps widened even further when the SC maintained its pro-Russian stance after Russia's

<sup>2</sup> Since the mid-2000s, the oligarchisation of Latvian party politics has caused an increasing politicisation of corruption. However, this issue has remained largely confined to the ethnic Latvian parties and has not been able to facilitate alliances across the ethnic divide. This has resulted in the further differentiation of ethnic Latvian parties into "oligarchic" and "corruption fighting" parties (Auers 2013, 92–95).

annexation of Crimea in 2014 (Ījabs 2018). Hence, the bipolar opposition in the Latvian party system prevented coalitions across camps during the entire period covered by this study.

## 6.2.3 Characteristics and preferences of Latvian radical right parties

### Parliamentary strength

In its electoral breakthrough in 2010, the NA won only a moderate 7.7 per cent of the vote, making it the smallest party in the Saeima (see Table 6.9). Unlike many other radical right parties in Central and Eastern Europe, however, the NA persisted and became one of the most consistent members of the party family in the region. The party secured well over ten per cent in each of the three parliamentary elections since 2010. Even in 2018, when the party lost ground for the first time, it still won 11 per cent of the vote. As one of Central and Eastern Europe's most successful radical right parties at the polls, the NA often contributed a substantial number of seats to the ruling coalition.


*Table 6.9: Election results and parliamentary strength of radical right parties in Latvia*

Source: Nordsieck 2021.

#### Ideological distance to the formateur

Unity served as the formateur of the four governments between 2010 and 2014, and its successor, JV, formed the 2018 coalition. During the mid-term government re-formation in 2016, the ZZS acted as the formateur. The NA advanced a nativist, ultranationalist concept of the Latvian nation based on ethnicity, culture, and language. The party's ethno-cultural nationalism is mainly directed against the Russian-speaking minority and its kin state, Russia. Even after the 2015 "migration crisis", the opposition to immigration and refugees remained secondary to its antiRussian platform. The NA campaigned for making Latvian the only language of instruction in public schools and increasing funding for traditional Latvian culture, thus trying to curb the influence of the Russian-speaking minority (Auers 2012, 6; Auers and Kasekamp 2015, 143). More radical forces in the party also regularly attended and supported the annual rally commemorating the Latvian SS Legionnaires who fought the Soviet forces in 1944 (Auers and Kasekamp 2013, 240–42). Even though the party was less outspoken against immigration than other radical right parties, particularly when in government, it still opposed the European quota system and framed immigration as a threat to national security. Framing immigration as a national security issue made it possible to connect it to the Russianspeaking minority and Russia (Braghiroli and Petsinis 2019, 442–43). In light of these positions, the party's GALTAN placements in the CHES (6.67, 8.09, and 8.11) are rather low, particularly in the 2010 wave, and are therefore adjusted to 8.50, which is still a moderate score for a radical right party (see Table 6.10).

On the socio-economic dimension, the NA combined liberal and protectionist positions. The national protectionism typical of radical right parties can be found in the Bauska Declaration: "We consider the independence of our economies to be just as important as our political independence. We are ready to combat the foreign financial influence in our countries and we see only Estonian, Latvian and Lithuanian capital as a basis of our national prosperity" (National Alliance 2021). The NA's national-conservative predecessor, For Fatherland and Freedom/Latvian National Independence Movement (TB/LNNK), however, was clearly positioned on the liberal end of the socio-economic dimension. The merger with the radical right All for Latvia! (VL!) added an element of national protectionism, but it did not mute the liberal economic currents in the NA. Hence, the socio-economic positions of the NA are more liberal than those of many other radical right parties in Central and Eastern Europe (Wierenga 2019, 143), which is also reflected in the party's CHES placements between 5.11 and 6.09.

The formateurs that the NA faced during the 2010s were also members of the neoliberal, ethnic Latvian camp. Unity, for instance, has been held responsible for the country's "fairly orthodox pro-market policies" (Duvold, Berglund, and Ekman 2020, 61). The party, and in particular prime minister Dombrovskis, fiercely defended Latvia's drastic austerity measures in the context of the economic crisis in the late 2000s (Sommers 2014; see also Duvold, Berglund, and Ekman 2020). On the socio-cultural dimension, Unity's version of ethno-linguistic nationalism was more moderate than that of the NA. Regarding other socio-cultural issues, the party held relatively liberal views. It even supported the European quota for refugees in 2015 (Auers 2016).Thus, the CHES places Unity slightly on the GAL side of the sociocultural spectrum, with scores of 4.00 in the 2014 and 4.82 in the 2019 wave, despite its moderate nationalism. In the 2010 wave, however, the party received a GALTAN score of 5.25, which is not far away from its later placements. Due to the qualitative

difference between a value below or above 5.00 (see Chapter 7), the party's GALTAN score in 2010 is adjusted to 4.50 to reflect the party's liberal leaning.


*Table 6.10: Socio-economic and socio-cultural distance between radical right parties and formateurs in Latvia*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions.

The ZZS is also located right of centre on the socio-economic dimension, but its positions are more centrist than Unity's. Even though the ZZS uses the label "Green" in its name, its socio-cultural positions are far from typical for the Green party family. The party and its leader, Aivars Lembergs, the mayor of Ventspils, have a conservative and nationalist outlook. They claim to protect Latvian national interests and national identity against alleged external enemies, such as the EU or George Soros (Galbreath and Auers 2010; Auers 2018, 352). Consequently, Galbreath and Auers (2010, 67) describe ZZS' rhetoric as "populist nationalist anti-liberal", which is also reflected in the party's CHES placement at 7.64 on the GALTAN dimension.

#### 6.2.4 Summary

During the last decade, the NA has established itself as a strong, radical right party in the Latvian party system. The party entered government almost immediately after its electoral breakthrough in the 2010 parliamentary elections. While a liberal party in the Unity alliance prevented the radical right's participation in government in 2010, the NA has been a junior partner in all ruling coalitions since 2011. These coalitions were almost exclusively minimal winning coalitions, characterised by a high degree of socio-economic homogeneity and socio-cultural heterogeneity.

The bipolar opposition in the Latvian party system has imposed limitations on government formation which have helped Unity and the NA to cooperate despite their socio-cultural incongruence. The SC's inability to find willing partners with whom to govern forced the remaining ethnic Latvian parties to accept substantial ideological differencesin order to form a majority government within this camp.The protests against Zatlers'attempt to overcome the bipolar oppositionin 2011illustrate how costly it was for the ethnic Latvian parties to consider forming a coalition with the pro-Russian SC.

#### **6.3 Bulgaria**

#### 6.3.1 Government formation with radical right parties in Bulgaria

In 2005, Ataka was the first radical right party to enter the Bulgarian parliament.The party was joined by a second radical right party, the Patriotic Front(PF)in 2014.Three years later, both parties merged to form the United Patriots (UP) and contested the parliamentary elections together.When Ataka first entered parliament in 2005, Bulgaria was working towards EU membership. In this situation, the three largest parties in parliament decided to form an oversized coalition with the goal of preparing the country for accession to the EU. This coalition included the Communist successor, the Bulgarian Socialist Party (BSP), the populist NationalMovement for Stability and Progress (NDSV) of the former Bulgarian tsar Simeon II Sakskoburggotski, and the Movement for Rights and Freedoms (DPS), which represented the interests of the Turkish minority (see Table 6.11). As the formateur, the BSP never considered inviting the Eurosceptic Ataka to join the government (Spirova 2006; Marinov 2008, 94–95).

Four years later, government formation took place under different circumstances. A new centre-right party, Citizens for European Development of Bulgaria (GERB), emerged victorious from the 2009 parliamentary elections. Contrary to expectations, it did not form a majority coalition with other conservative parties. Instead, two conservative parties, the Blue Coalition (SK) and Order, Law, and

Justice (RZS), as well as the radical right Ataka, supported a GERB minority government, but Ataka was the only support party to sign a written agreement (Spirova 2010; Avramov 2015, 301).

In May 2013, the next parliamentary elections were held early after the Borisov I government resigned following massive anti-government protests during the so-called "winter of discontent" (Avramov 2015, 299). However, only four parties entered parliament, and these organised themselves into two alliances, including GERB and Ataka on one side and the BSP and the DPS on the other. Each camp controlled 120 of the 240 seats in parliament, and the three challengers signalled during their campaigns that they would not support another minority government led by GERB. Ultimately, a BSP-DPS government under the leadership of prime minister Plamen Oresharski was sworn in—and it was Ataka's leader, Volen Siderov, who tipped the scales in favour of the minority coalition. He was the only member of the opposition to register for the investiture vote and thus enabled the government to secure the required quorum. All other deputies of GERB and Ataka deliberately abstained from the vote in order to boycott Oresharski's election (Karasimeonov 2013a, 2–3; Kostadinova and Popova 2014; Avramov 2015, 303–4). Even after this investiture vote, Ataka backed the Oresharski government on various occasions and did not support a vote of no confidence initiated by GERB (Karasimeonov 2013b, 8; Kostadinova and Popova 2015). Nevertheless, Ataka is not considered an official support party for the minority government because there was no formal public agreement between Ataka and the ruling parties.

The Oresharski government did not last long, however. The early elections of 2014 were again won by GERB, and Borisov received the mandate to form a government. Even though he was rather sceptical about renewing relations with Ataka, he still invited the party to exploratory talks, as he did with every party in parliament. Ataka, however, declined this invitation and decided to remain in opposition. Instead, Borisov formed a minority coalition with the centre-right Reformist Bloc (RB) and the nominally social democratic Alternative for Bulgarian Revival (ABV). This coalition was officially supported by the other radical right party in parliament, the PF (Karasimeonov 2014b, 2–3; Kostadinova and Popova 2015). In the run-up to the 2016 presidential elections, the government had lost significant public support, prompting the ABV and parts of the RB to withdraw from the coalition.The remaining coalition of GERB and RB still enjoyed the support from the radical right PF. Borisov intended to invite the PF to become a junior partner after the presidential elections later that year, but this plan never came to fruition. After GERB's candidate lost the presidential election, Borisov submitted his resignation in 2016, which resulted in yet another early election in 2017 (Kolarova and Spirova 2017).

GERB once again emerged as the winner of the parliamentary elections in 2017. Borisov held talks with several parties, but he quickly decided to form a majority coalition with the radical right UP. In May 2017, the coalition of GERB and UP was



Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

\*Partiesinparentheses=supportpartyofaminoritygovernment.

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in parentheses indicate the smallest possible ideological distance of a majority government.

sworn in. The government controlled a majority in parliament, and did not depend on the support of other parliamentary parties (Spirova 2018).Thus, after supporting minority governments in 2009 and 2013, the radical right finally received seats at the cabinet table for the first time in 2017.

The dominant format of government in Bulgaria during the period under investigation are minority governments.The Borisov IV government of GERB and UP was the only minimal winning coalition. None of the six Bulgarian governments since 2005 meets the criteria for minimal range or minimal connected winning coalitions. This is also due to the ideological distance between the ruling parties on the socio-economic and socio-cultural dimensions. The only exception is the Oresharski government of 2013, which was socio-economically and socio-culturally homogeneous, but it did not control a majority in parliament. These programmatic differences may have led parties to form minority governments with support parties instead of formal majority coalitions. The only minimal winning coalition, the 2017 GERB-UP government, supports this interpretation, since its ideological range was relatively small, at least compared to the previous governments.

## 6.3.2 The configuration of the Bulgarian party system

### Fragmentation

The Bulgarian party system was quite stable in the first decade following the fall of Communism. It was dominated by two large parties—the Communist successor party, BSP, and the oppositional SDS.The DPS was the only other party that had continuous electoral success.The disintegration of the SDS in the early 2000s, however, resulted in the emergence of various new liberal-conservative parties (Karasimeonov 2010). These changes led to high levels of volatility and growing fragmentation in the Bulgarian party system during the 2000s (Enyedi and Casal Bértoa 2018, 438). The Bulgarian party system remained structurally unstable in the 2010s, witnessing considerable electoral swings and the entrance and exit of new parties.

Despite this instability, the fragmentation scores in 2009, 2013 and 2017, were relatively low (see Table 6.12). In all three instances, the complexity of the bargaining situation was manageable from a purely mathematical perspective. In 2009, GERB controlled 116 of the 240 seats in parliament and could thus form a majority coalition with any one of the five small parties. The 2013 National Assembly consisted of only four parliamentary groups, which left the parties to choose between one of three minimal winning coalitions. In 2017, there were five parties present in parliament, but the constellation of two large and three small parties allowed for only four possible minimal winning coalitions.


*Table 6.12: Fragmentation of the Bulgarian party system*

Source: Own compilation based on data from Casal Bértoa 2021; Nordsieck 2021.

The fragmentation was significantly higher in 2005 and 2014. The parliament consisted of seven and eight actual parties, respectively. In both years,one party controlled significantly more seats than the others, but not enough to establish a twoparty majority coalition with most of the other parties. Since most of the possible minimal winning coalitions involved at least three parties, the bargaining situation was rather complex in these legislatures.

### Bipolar opposition

The structural changes in the Bulgarian party system also affected the oppositional constellations. Party competition in the first post-Communist decade was strongly influenced by the regime divide. The BSP represented the old regime and the SDS the oppositional forces, while the DPS was ready to cooperate with either side and thus enjoyed a pivotal role in government formation (Autengruber 2006, 80; Karasimeonov 2010, 2). When Simeon II arrived on the Bulgarian political scene in the early 2000s, he positioned himself, and his party the NDSV, as an alternative to the existing elites, although he remained open to cooperating with the established parties (Karasimeonov 2010, 2–4; see also Avramov 2015). Even though the NDSV was founded shortly before the 2001 elections, it immediately became the BSP's main competitor. However, despite forming a majority coalition with the DPS in 2001, Simeon II decided to include two BSP ministers in his cabinet, which underlines the absence of a deep bipolar opposition at that time (Karasimeonov 2010, 4). When a radical right party entered the Bulgarian parliament for the first time in 2005, the configuration of the Bulgarian party system looked much like it had at the beginning of the decade. Despite the emergence of new parties, including Ataka, polarisation between different camps was mild and most parties shared the goal of securing Bulgaria's accession to the EU (Keudel-Kaiser 2014, 86; see also Karasimeonov 2010).

Persistent clientelism, and a growing alienation of the political parties from (civil) society, contributed to another transformation of the Bulgarian party system, which became visible in the 2009 parliamentary elections. The NDSV failed to enter parliament, which left room for new parties on the centre-right to gain prominence. The most successful of these parties was the populist GERB, which presented itself as an anti-corruption party and won 40 per cent of the vote in the first parliamentary election it contested (Karasimeonov 2010, 25–26). After GERB's electoral breakthrough in 2009, the party became the BSP's chief competitor. Since the DPS had by then sided with the BSP, and most of the new conservative and radical right parties coalesced around GERB, the 2009 parliamentary elections marked the return of a bipolar opposition to the Bulgarian party system (see Table 6.13). The polarisation between GERB and the BSP was much more affective than it was ideological. The parties held different views, in particular on socio-economic policies (see below), but these disagreements do not explain the deep rift between these parties, and their respective camps, which ultimately prevented them from forming cross-camp coalitions by 2009. The polarisation continued to deepen further and has remained a characteristic feature of Bulgarian politics until the end of the 2010s. On some occasions, GERB and BSP even attempted to boycott parliamentary proceedings when the other party won the election (Karasimeonov 2019).



Source: Own compilation.

### 6.3.3 Characteristics and preferences of Bulgarian radical right parties

#### Parliamentary strength

In its electoral breakthrough in 2005, Ataka entered parliament with 8.1 per cent of the vote (see Table 6.14). The party improved its result slightly in the 2009 parliamentary elections, receiving 9.4 per cent, totalling the best result in the party's history. Since then, Ataka constantly lost at the polls, at least in part because some supporters disapproved of the party's support for the GERB minority government. It even looked like Ataka might fail to return to parliament in 2013. However, the anti-government protests revived the party, which allowed Ataka to pass the four per cent threshold comfortably (Avramov 2015). Because only four parties entered parliament after the 2013 elections, the 7.3 per cent of the vote won by Ataka resulted in more seats than the previous two terms. Ataka's partial support of the Oresharski government caused it to lose further credibility among its radical right electorate, however. Thus, in 2014 the party barely passed the four per cent threshold.


*Table 6.14: Election results and parliamentary strength of radical right parties in Bulgaria*

Source: Nordsieck 2021.

The PF outscored Ataka in its electoral breakthrough in 2014, gaining 7.3 per cent of the vote. Thus, the 2014 Bulgarian parliament included two relatively small radical right parties. After this experience, the PF and Ataka contested the 2017 election together as UP, which won 9.1 per cent of the vote. The UP fell short of Ataka and PF's combined result from 2014 (11.8 per cent), but they did secure a double-digit seat share (11.3 per cent) for a radical right party in the National Assembly for the first time.Ultimately, this showing was enough to form a majority coalition together with GERB.

## Ideological distance to the formateur

#### **The radical right: Ataka, the Patriotic Front, and the United Patriots**

In Bulgaria, several radical right parties have entered parliament. Ataka emphasised social-national economic policies. In its campaign materials from 2005, Ataka formulated the overarching goal of "making sure that the Bulgarian economy served the interests of the Bulgarian people" (Ghodsee 2008, 30).The party favoured, for instance, a strong welfare state, state investments, state ownership of key industries, as well as an increased minimum wage and progressive taxation (Avramov 2015, 308; Pirro 2016, 62–63; Popova 2016, 262–63). Ataka's socio-economic profile was thus not only more specific than that of many other radical right parties in Central and Eastern Europe, but it was also decidedly more leftist. The party's CHES placements on the socio-economic dimension reflect its position on the left (see Table 6.15). However, the score of 1.44 in the 2014 wave seems very low and is therefore slightly adjusted to 2.50. Although Ataka's 2013 election programme indeed emphasises opposition to neoliberal policies, there is no indication of more radical positions than before (Kostadinova and Popova 2014).

On the socio-cultural dimension, Ataka's profile is dominated by a strong antiminority platform.The party's policies and rhetoric are mainly directed against minority groups, such as Roma and Turks, and the party which represents them, the DPS. Anti-Semitism and verbal attacks against the LGBTIQ+ community also belong to the party's repertoire. Ataka's agenda has even included irredentist positions, although not as prominently as in other Central and Eastern European radical right parties (Karasimeonov 2010, 20; Cholova and Waele 2011, 34; Todorov 2013, 3; Pirro 2016, 62–66). Since the mid-2010s, the party has adopted the immigration issue, spreading racism and xenophobia while opposing the admission of refugees (Karasimeonov 2019, 7). Hence, Ataka displays all of the characteristics associated with the exclusionary, nativist ultranationalism of a radical right party, which is also reflected in the party's GALTAN placements between 9.17 and 9.65.

The two parties that merged to form the PF in 2014, the Internal Macedonian Revolutionary Organisation (VRMO) and the National Front for the Salvation of Bulgaria (NFSB), appeal to a similar electorate as Ataka. NFSB's leader, Valeri Simeonov, was previously amember of Ataka but left the party to form the NFSB due to personal differences with Volen Siderov (Todorov 2013, 3; see also Karasimeonov 2014a, 2019; Krasteva 2016). VMRO was re-established in 1990 and since then has been part of the nationalist spectrum of Bulgarian politics (Krasteva 2016, 179). Similar to Ataka, the PF identifies minorities and immigrants as its main enemies (Krasteva 2016),

but it tried to position itself as a more moderate, radical right alternative (Kostadinova and Popova 2015). At the same time, however, Simeonov criticised Ataka's cooperation with the mainstream when it supported the GERB minority government (Krasteva 2016).While VMROisindeedmoremoderate than Ataka, the NFSB resembles a "clone formation" of Ataka (Avramov 2015, 300). Simeonov underlined NFSB's similarities with Ataka, for instance, when insulting Roma in a public session of the parliament in 2014, for which he was later convicted in court (Karasimeonov 2017a). The CHES includes only the constituents of the PF, VMRO and the NFSB.Therefore, the PF's CHES scores are calculated by taking the mean of the two party's scores on the respective dimension. Thus, PF's GALTAN score of 8.31 is indeed more moderate than Ataka's. In 2017, all three radical right parties, Ataka, the NFSB, and VMRO, contested the parliamentary election as United Patriots. Therefore, the UP's GAL-TAN score of 9.34 is the mean of the parties' placements in the 2019 CHES wave.

The socio-economic positions of Ataka and the PF differ only slightly. In the 2014 election campaign, the PF campaigned for an expansion of the welfare state and wage increases (Kostadinova and Popova 2015). In contrast to Ataka, however, the PF's positions are more pro-market, which is also reflected in the LRECON scores of VMRO and the NFSB in the 2014 CHES wave (4.00). This score places the PF on the socio-economic left, but closer to the centre than Ataka, which received an adjusted score of 2.50.

The socio-economic positions of the three radical right parties remained largely unchanged in the run-up to the 2017 elections (Karasimeonov 2016, 9). The 2019 CHES wave, however, places both VMRO and the NFSB much closer to the centre of the LRECON dimension. The NFSB even receives a score of 5.50, locating the party slightly on the liberal end of the socio-economic dimension. Since this placement does not reflect the party's policy positions, the UP's LRECON score is calculated using VMRO's and NFSB's placement in the 2014 CHES wave and Ataka's in the 2019 wave. This results in a mean LRECON score of 3.84, which realistically describes the UP's left-leaning socio-economic profile.


*Table 6.15: Socio-economic and socio-cultural distance between radical right parties and formateurs in Bulgaria*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions.

\* The PF's CHES scores are calculated by taking the mean of the NFSB and VMRO.

\*\* The UP's CHES scores are calculated by taking the mean of Ataka in the 2019 wave, and the NFSB and VMRO in the 2014 wave.

#### **The formateurs: BSP and GERB**

The BSP won the 2005 parliamentary elections and was thus tasked with government formation.The BSP is the successor of the Bulgarian Communist Party, which, similar to Ceauşescu in Romania (see below), adopted a national Communist ideology. In the first post-Communist decade, the BSP continued this tradition.The party did not shy away from nationalist rhetoric and had some reservations about Bulgaria's integration into NATO and the EU (Spirova 2008; Genov 2010). Around the turn of the millennium, however, the BSP reformed its ideological profile and developed into a social democratic, centre-left party committed to a capitalist market economy, but favouring a comprehensive welfare state as well as limited privatisation and state interventionism (Smilov 2008, 15; Spirova 2008, 491). Moreover, the party also toned down its nationalist rhetoric and declared its support for European human rights standards (Spirova 2008, 491; Vachudova 2008, 870). The 2006 and 2014 CHES waves reflect the BSP's socio-economic and socio-cultural positions quite well. The party receives LRECON scores of 3.25 and 3.47, and GALTAN scores of 5.83 and 5.94, respectively.

The other formateur, GERB, has been the BSP's main opponent since the late 2000s. GERB's socio-economic agenda includes classic neoliberal policies, such as a flat tax and general tax reductions, cuts to the welfare state, and austerity measures (Karasimeonov 2009, 2014a; Kostadinova and Popova 2015). Similar to Ataka, the party started with a populist anti-corruption platform, but soon lost credibility in this area. GERB's chairman, Boyko Borisov, and his party are also decidedly pro-European. Regarding socio-cultural issues, however, GERB used sexist and homophobic rhetoric, and the party adopted some of Ataka's nationalist and anti-minority positions (Avramov 2015, 311–12; see also Pirro 2016). Krasteva (2016, 176) thus describes GERB's ideology as "moderate nationalism". GERB's GALTAN score of 6.91 in the 2010 CHES wave better reflects the party's position on the socio-cultural dimension than the centrist placements in the later waves (5.12 and 5.76, respectively). Therefore, this study uses the party's 2010 GALTAN score in 2013, 2014 and 2017 as well.

#### 6.3.4 Summary

In Bulgaria, it took 15 years for radical right parties to enter parliament for the first time after the fall of the Iron Curtain. Since 2005, however, the radical right has been present in parliament without interruption. For much of this time, Ataka, the PF, or their alliance, the UP, have participated in government, despite never winning more than ten per cent of the vote. Initially, Ataka and the PF served as support parties for GERB-led minority governments. In 2017, however, GERB invited the UP to become a junior partner.While scholars have argued that radical right support parties of minority governments gain certain benefits, such as influence over government policies while maintaining an oppositional appeal (Zaslove 2012; see also Albertazzi and McDonnell 2005), this strategy has not paid off for the Bulgarian radical right. Ataka in particular lost at the polls after formally supporting the Borisov I government and backing the oppositional Oresharski government on various occasions, including the investiture vote.

The majority of governments in Bulgaria, including those with radical right parties, were characterised by a relatively broad socio-economic and socio-cultural range. Borisov in particular tended to form multi-party governments with ideologically distant junior partners and support parties. The GERB-UP government of 2017 was an exception, as it was a minimal winning coalition that consisted of only two, ideologically proximate, parties. This coalition was also significantly more stable than the previous governments, as it lasted the full term, despite some quarrels between GERB and the radical right, as well as disputes within the UP itself (Karasimeonov 2017b, 16).

## **6.4 Romania**

## 6.4.1 Government formation with radical right parties in Romania

The radical right repeatedly entered the Romanian parliament during the 1990s and 2000s. Two radical right parties, the Greater Romania Party (PRM) and the Party of Romanian National Unity (PUNR), won seats in the first free elections in 1992.<sup>3</sup> The Democratic National Salvation Front (FDSN), main successor of the Romanian Communist Party, won the 1992 parliamentary elections. Unlike in 1990, however, the party fell short of a majority,in part because a reform-oriented wing, led by Petre Roman, formed the Democratic Party (PD) (Gabanyi 1998, 251). The lower house of the Romanian parliament comprised seven parties and electoral alliances, as well as 13 representatives from different minorities. Neither the PD, nor the parties from the anti-Communist opposition in the Romanian Democratic Convention (CDR), wanted to cooperate with president Iliescu's FDSN. Thus, the FDSN's choice of potential coalition partners was rather limited, mostly to the radical right parties, which had already supported Iliescu's presidential candidacy (Gallagher 1995, 219). The FDSN ultimately formed a single-party minority government under prime minister Nicolae Văcăroiu, which was officially supported by the PRM, the PUNR, and the orthodox Communist, Socialist Party of Labour (PSM). This government underwent a re-formation in 1994, when the PUNR became a junior partner and received seats in the cabinet (Gallagher 1994, 30–32; Shafir 1999, 216; Autengruber 2006, 70–71).

The 1996 general elections removed the old elites from power. The CDR won the election, receiving 30 per cent of the vote. The FDSN, now renamed Party of Social Democracy in Romania (PDSR), scored 21.5 per cent and came in second. The CDR comprised several liberal and conservative parties, the largest of these being

<sup>3</sup> Whether the 1992 Romanian general elections can be considered fair, free, and democratic is disputed. Observers noted some irregularities (Carey 1995), but there was no conclusive evidence of systematic manipulation (Autengruber 2006, 70).

the Christian Democratic National Peasants' Party (PNŢCD). Thus, PNŢCD candidate Victor Ciorbea became the formateur and designated prime minister. He led an oversized coalition with other parties from the CDR, such as the National Liberal Party (PNL) and the National Liberal Party – Democratic Convention (PNL-CD), as well as Petre Roman's PD, the Social Democratic Party of Romania (PSDR, not to be confused with the PDSR), and the Hungarian minority party, Democratic Alliance of Hungarians in Romania (UDMR). The formateur never considered a coalition with the radical right parties, PRM and PUNR (Autengruber 2006, 72–74).

This ideologically heterogenous, multi-party coalition re-formed several times during the legislature, including two changes of the prime minister. The PNL-CD and the PD, for instance, left the coalition in 1997 and 1998, respectively. Also in 1998, Ciorbea was replaced as prime minister by another PNŢCD politician, Radu Vasile. In 1999, the independent Governor of the National Bank of Romania, Mugur Isărescu, was elected prime minister, and the PD rejoined the coalition (Autengruber 2006, 74–75; Ștefan 2019, 407). Overall, the four-year term saw three different prime ministers and seven government coalitions, each with a slightly different partisan composition.The radical right parties, PRM and PUNR, were constantly in opposition. In order to avoid inflating the number of cases from a single country, the study includes only the Ciorbea I and the Isărescu I governments from this period.

In the 2000 general elections, the incumbent parties were punished for their inconsistent behaviourin office.The PDSR emerged as the clear winner of the elections with 36.6 per cent of the vote. The party secured a comfortable advantage over the radical right PRM, which came in second with 19.5 per cent—one of the best results of a radical right party in Central and Eastern Europe to date.The PUNR,in contrast, dropped out of parliament and never recovered from this defeat. Even though the PDSR and the PRM controlled a comfortable majority in parliament, they did not renew their cooperation from the early 1990s (Autengruber 2006, 74; Pop-Eleches 2008, 470). In fact, the PDSR joined the liberal and conservative parties in their cordon sanitaire vis-à-vis the radical right (Cinpoeş 2015, 288). Instead of cooperating with the PRM, PDSR's Adrian Năstase formed a minority government that was supported by the Hungarian minority party, UDMR, and the liberal PNL (Popescu 2003, 332; Gabanyi 2005).

In 2004, government formation was strongly influenced by the presidential elections, which were held together with the parliamentary elections. Traian Băsescu, joint candidate of the PNL and the PD, edged out former PDSR prime minister Năstase in the second round of the presidential race, whereas the PDSR, again renamed Social Democratic Party (PSD), came in first in the parliamentary elections. However, Băsescu used his constitutional powers to nominate PNL's leader, Călin Popescu Tăriceanu, to be the formateur of the new government. Popescu Tăriceanu successfully forged a minority coalition consisting of the PNL, the PD, the UDMR, and the Romanian Humanist Party (PUR) which initially contested the election in an alliance with the PSD (Gabanyi 2005, 4–5; Stan and Zaharia 2007). After the PNL and the PUR left the coalition due to internal conflicts in 2007, the remaining twoparty minority coalition continued in office until the end of the term, relying on the informal support from the PSD (Stan and Zaharia 2008). The PRM continued to be ostracised by the other parliamentary parties, and spent its last term in parliament on the opposition bench (Cinpoeş 2015).

The governments that were formed in Romania when radical right parties were present in parliament, were predominantly minority governments and often included a relatively large number of parties (see Table 6.16). This observation is true for governments with and without radical right parties, and it continues after the PRM dropped out of parliament in 2008 (Ștefan 2019). None of the three coalitions that were formed without radical right participation between 2000 and 2008, when data on party ideology is available, were minimal connected winning or minimal range coalitions on the socio-economic or socio-cultural dimensions. They did not fulfil the majority criterion required for both formats, and in 2000 and 2004, there were socio-economically and socio-culturally more homogeneous coalitions available to the parties. The qualitative data on the ideological positions of the FDSN/PDSR and the two radical right parties indicates that the ideological range of the 1992 and 1994 governments was rather small (see below). However, due to the lack of positional data for all parliamentary parties, it is impossible to determine whether these governments meet the ideological criteria for minimal connected winning or minimal range coalitions. However, since both of them are minority governments, they do not fulfil the majority criterion required for these formats.

## 6.4.2 The configuration of the Romanian party system

## Fragmentation

Table 6.17 shows that the fragmentation of the Romanian party system had been constantly declining between 1992 and 2008 (Enyedi and Casal Bértoa 2018). In 1992, the effective number of parliamentary parties was still relatively high (4.8). From a purely mathematical perspective, the bargaining situation after the 1992 elections was quite complex, because the seven parties in parliament could form ten different minimal winning coalitions.<sup>4</sup> In 1996 the number of parliamentary parties decreased to 4.2, but the six parties and electoral alliances in parliament still faced a bargaining situation of moderate complexity. Due to the relatively even distribution of seats, none of them was close to a majority, and most of the possible minimal winning coalitions comprised three or more parties.

<sup>4</sup> This number is based on a conservative count, considering the CDR as a single entity and excluding the 13 minority representatives in the legislature.


*Table6.16:FormatandideologicalrangeofgovernmentsinRomania*


Source: Own compilation based on data from Casal Bértoa 2021; Bergman, Ilonszki, and Müller 2019a; Jolly et al. 2022.

Partiesinparentheses=supportpartyofaminoritygovernment.

\*

 \*\* Abbreviations: MWC = minimal winning coalition, MinSP = single-party minority government, MinC = minority coalition government, Surplus = oversized coalition.

\*\*\* Values indicate the ideological range between the most distant parties in government, including support parties of minority governments. Values in parenthesesindicatethesmallestpossibleideologicaldistanceofamajoritygovernment.

188 Pariahs or Partners?


*Table 6.17: Fragmentation of the Romanian party system*

Source: Own compilation based on data from de Nève 2010; Casal Bértoa 2021; Nordsieck 2021.

In 2000 and 2004, the fragmentation drops significantly to 3.6 and 3.4 effective parliamentary parties, respectively. This reflects a new situation in which the number of actual parties in parliament decreased further, resulting in a more manageable number of possible minimal winning coalitions. However, electoral alliances complicate coalition bargaining in Romania because they are mainly vote-winning instruments and do not necessarily entail a commitment to cooperate in government after elections.Government formationin 2004is a casein point. If the PSD and the PUR had continued their alliance in parliament, they could have joined together with almost any other party, and this coalition would have controlled a majority. Instead, the PUR broke away and decided to enter a four-party minority coalition led by the oppositional PNL.

### Bipolar opposition

The Romanian party system of the early 1990s was structured by bipolar opposition rooted in the regime divide (Ștefan 2019, 397). In 1990, the oppositional forces and their electoral alliance, CDR, stood no chance against the National Salvation Front (FSN), the predecessor of the FDSN. After the violent overthrow of Communism in 1989, the FSN formally distanced itself from Ceauşescu but remained in control of the former regime's resources (Autengruber 2006, 146). The FSN's structural advantage won the party a landslide success in 1990, which is the main reason why these elections are considered neither free nor fair. In the run-up to the 1992 elections, however, the FSN split—the hardliners formed the FDSN and the reform-oriented members established the PD. The latter credibly distanced themselves from the Communist regime and sided with oppositional camp. The bipolar opposition that ran between CDR and PD on one side and the FDSN and some small

parties rooted in the former regime, like the radical right PRM, on the other, remained deeply polarised until the late 1990s (Gabanyi 1997, 194; Autengruber 2006, 147; Pop-Eleches 2008, 468–69) (see Table 6.18).

This regime divide was reflected in various political conflicts, such as the issue of the country's economic transformation. Here, FDSN, PRM, and PUNR preferred incremental reforms, while the opposition camp favoured a swift transition to a capitalist market economy.The opposition between the two camps also entailed an ethnic divide over the rights of the Hungarian minority in the country and Romania's relations with neighbouring Hungary.The party of the Hungarian minority, UDMR, was also a part of the opposition camp, while the national Communist camp, and most importantly the radical right parties, held anti-Hungarian views (Gabanyi 1997, 194–98; Autengruber 2006, 147; Cabada, Hloušek, and Jurek 2014, 97; Cinpoeş 2015, 287). Thus, socio-economic and socio-cultural divides were aligned in the Romanian party system.


*Table 6.18: Bipolar opposition in the Romanian party system*

Source: Own compilation.

Due to the moderation of the FDSN/PDSR, which began distancing itself from its Communist past and its nationalist rhetoric in the late 1990s, this bipolar opposition has gradually waned. After its electoral victory in 2000, the party continued the economic reforms initiated by the previous government, indicating a reduction in tensions between camps, as well as a centripetal thrust in the party system (de Nève 2002, 309; Fesnic and Armeanu 2010). In 2000, the PDSR minority government was already tolerated by the oppositional UDMR. Hence, while coalitions across camps were impossible for most of the first post-Communist decade, a gradual thawing

of inter-camp relations began at the end of the 1990s, stimulated by parties' shared interest in joining the EU and NATO (Ștefan 2019, 397).

## 6.4.3 Characteristics and preferences of Romanian radical right parties

## Parliamentary strength

In the 1992 general elections, the two radical right parties, PRM and PUNR, received more than ten per cent of the total votes and a corresponding number of seats in the chamber of deputies (see Table 6.19). The PUNR was clearly the stronger of the two parties, winning 7.7 per cent of the vote compared to 3.9 per cent for the PRM. The Romanian electoral system included only a three per cent threshold, so the PRM's vote share, although low, was sufficient to secure 16 parliamentary seats. With seat shares of 4.7 and 8.8 per cent, however, the radical right parties had limited opportunities to contribute to a majority coalition in parliament. In 1996, the PUNR suffered substantial losses, whereas the PRM improved marginally, but neither one gained more than five per cent of the vote, making them the smallest parties in the Chamber of Deputies.


*Table 6.19: Election results and parliamentary strength of radical right parties in Romania*

Source: Nordsieck 2021.

The PUNR's downward trend continued, and the party dropped from parliament in 2000. The PRM, in contrast, won almost 20 per cent of the vote and 25 per cent of the seats, good for second place in the 2000 parliamentary elections.The PRM could have made a sizeable contribution to a parliamentary majority; however, the established parties placed it behind a cordon sanitaire making government participation impossible. In 2004, the PRM won 13 per cent of the vote and 14.4 per cent of the seats in parliament. However, this result marked the beginning of the party's electoral decline.The PRM failed to pass the threshold of representation in 2008 and has not returned to parliament since.

#### Ideological distance to the formateur

#### **The radical right: PRM and PUNR**

The main feature of the PUNR's ideology was its outright hostility towards the Hungarian minority in Romania, which included verbal attacks and even calls for violence during the ethnic tensions of the early 1990s. The PUNR wanted to curb the rights of the Hungarian minority in the education sector and to restrict the use of the Hungarian language. The party also supported banning the UDMR and resettling Hungarians in their kin state. Anti-Semitism and racism against Roma were secondary to the party's anti-Hungarian agenda but still part of its ideological platform. The party even made irredentist claims, for instance proposing to re-annex parts of Bukovina and Bessarabia (Gallagher 1995, chap. 6; Shafir 1999, 214–17; Andreescu 2003, 30–31; Adamson, Florean, and Thieme 2011, 319–20).

The PRM held similar socio-cultural positions, but prioritised them differently. Irredentism was more prominent in the party, as its name—Greater Romania Party—already suggests. Moreover, the PRM targeted Roma and Jews rather than Hungarians. Party leader Corneliu Vadim Tudor, often referred to as Ceauşescu's "court poet" (Shafir 1999, 214), was notorious for his anti-Semitic writings and hate speech. He called for the deportation of Roma to labour camps, denied the Holocaust publicly, and supported the rehabilitation of Marshall Antonescu, the leader of Romania's fascist Iron Guard which ruled the country from 1940 to 1944 (Shafir 1999, 214–16; Țurcanu 2010, 5–7; Adamson, Florean, and Thieme 2011, 320–22).

Both radical right parties were situated left of centre on the socio-economic dimension, but the PRM was clearly the more anti-reformist and protectionist of the two.The PRM was rather sceptical towards the transformation of the economic system to a capitalist market economy and called for a comprehensive welfare state. If privatisation had to take place at all, the party wanted it to benefit Romanian, or at least post-Soviet, investors. Even more racist and nationalist was the party's demand to expropriate Hungarian and Jewish-owned businesses (Gabanyi 1997, 222; Adamson, Florean, and Thieme 2011, 321).

The PUNR was generally more moderate but also supported expropriating Hungarians and favouring Romanian investors during the process of privatisation, particularly the Communist nomenklatura (Gabanyi 1997, 231). Gallagher (1997, 31) describes the party's socio-economic platform as "inconsistent" compared to other Romanian parties at this time,and this resonates with the party's perception as an anti-Hungarian, single-issue movement (Andreescu 2005, 186). Moreover, the PUNR's somewhat blurry socio-economic platform might also be a result of the party's focus on its stronghold in Transylvania, where it attempted to ensure that party members and affiliates profited directly from political and economic transformation processes.

Both Romanian radical right parties have been clearly influenced by the national Communist legacy of the Ceauşescu regime, which created "a cultural system where extreme nationalist themes, symbols andideas occupied a prominent position"(Cinpoeş 2015, 286). The PRM in particular invoked this legacy as a justification for both its nationalism and its reluctance towards economic and political transformation. But PUNR's leader, Gheorghe Funar, also adopted the clientelistic practices and ideological positions of the old regime (Gallagher 1995, chap. 6). In sum, the PRM and the PUNR are clearly located on the nationalist-authoritarian end of the GALTAN dimension and on the left side of the socio-economic spectrum. Since the PUNR is slightly more moderate on both dimensions, the party receives a GALTAN score of 9.0 and an LRECON score of 3.0, whereas the PRM is placed at 9.5 and 2.0 (see Table 6.20).

#### **The formateurs: FDSN/PDSR, PNŢCD, and PNL**

The governments in the 1992 and 2000 legislatures were formed by the FDSN, later renamed PDSR. As a successor of the Romanian Communist Party, the FDSN was deeply rooted in the national Communist Ceauşescu regime. In the early 1990s, the party favoured slow and gradual economic reforms at the "lowest possible social cost" (Văcăroiu 1993, in Ionescu 1993, 17). They hoped to achieve this goal by using measures such as very limited privatisation and state subsidies for key industries. The PDSR only reorganised itself into a centre-left, social democratic party, fully accepting Romania's transformation to a capitalist market economy and committed to integrating into NATO and the EU in 1997, after being replaced in government. However, the PDSR never abandoned the goal of creating a strong welfare state (Gabanyi 1997, 224; Bugajski 2002, 846–47; de Nève 2002, 66–67; Pop-Eleches 2008, 470; Vachudova 2008, 871).


*Table 6.20: Socio-economic and socio-cultural distance between radical right parties and formateurs in Romania*

Source: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author. Values in parentheses indicate author's placement based on a qualitative assessment of party positions.

The socio-cultural positions of the FDSN in the early 1990s were quite close to the radical right. In addition to the party's authoritarian style of government, "nationalism was the tool of choice" for the FDSN (Cinpoeş 2015, 287); however, liescu and his party were more moderate than the PRM and the PUNR in this regard (Gallagher 1995, chap. 4; Andreescu 2003, 2005). By the end of the decade, the party toned down its nationalist rhetoric markedly and distanced itself from the radical right (Bugajski 2002, 843; Pop-Eleches 2008, 470; Vachudova 2008, 871).

In light of these positions, the FDSN/PDSR receives a score of 2.50 on the socioeconomic dimension and 7.50 on the GALTAN dimension in 1992 and 1994.The 2002 CHES wave covers the party's positions in 2000, but its placement at 2.45 on the socio-economic dimension seems a bit low, given the reforms it made since the late 1990s. Its LRECON score is therefore adjusted to 3.00 in 2000.

The CDR came to power after the 1996 general elections. The PNŢCD was the strongest party in this alliance and thus also the party of the prime minister and the formateur. The PNŢCD is also considered the formateur of the 1999 government because it continued to be the largest individual party in the coalition. The party dates back to the pre-Communist period, making it one of the few successful historical parties in Central and Eastern Europe after 1989 (Bugajski 2002, 839, 852–853). The PNŢCD is Christian democratic in name and ideology. After the tentative economic reforms of the previous government, the party aimed at accelerating Romania's transformation to a market economy, including more extensive privatisation. Overall, however, the PNŢCD adopted a social market economy, seeking to balance free market economics with social security (Gabanyi 1997, 219–20; Bugajski 2002, 852–53). The party's conservative profile was reflected by its preference for a constitutional monarchy and an "enlightened patriotism [*aufgeklärter Patriotismus*]" (Gabanyi 1997, 218), which valued national identity but differed from the exclusive nationalism of the radical right. The PNŢCD remained committed to minority rights and sought reconciliation with the Hungarian minority in the country (Gabanyi 1997, 219–20). It also stood for democratic values, the rule of law, and Western integration. Thus, the party receives centre-right LRECON and GALTAN scores (6.00 and 5.50, respectively), which are typical for the Christian democratic party family.

Due to the active intervention of President Băsescu, the PNL became the formateur of the 2004 government and maintained this role in the 2007 re-formation. The party was clearly positioned on the liberal end of the socio-economic dimension. Once in government, the PNL liberalised the tax code and introduced a flat tax (Gabanyi 2005, 6; W. M. Downs and Miller 2006). Despite its liberal economic programme, the PNL did not want to abandon the welfare state completely because the majority of the Romanian electorate was rather left-leaning.The party also held relatively liberal socio-cultural views. More than other Romanian parties, it supported minority rights and criticised the dominant role of the Romanian Orthodox Church in politics and society (Grecu et al. 2003). These positions are also reflected in the PNL's CHES scores, which place the party at 7.50 on the socio-economic and 4.10 on the socio-cultural dimension.

## 6.4.4 Summary

The radical right entered governmentin Romaniain the early 1990s,when the regime divide structured party politics. Using the resources of the old regime, the FDSN won the 1992 elections comfortably. Unlike in 1990, however, the party was no longer capable of winning an absolute majority on its own. Given the deep regime divide, the FDSN depended on the support of other parties from the national Communist camp in order to retain executive power. Notably, these parties included the two radical right parties, PRM and PUNR.Thus, as members of the governing coalition, the radical right parties contributed to slowing down the transformation process in Romania.

In 2000, the PUNR dropped out of parliament while the PRM tallied almost 20 per cent of the vote. Compared to the early 1990s, however, the Romanian party system had changed markedly. The regime divide was waning and the PDSR had distanced itself from its former radical right ally in order not to jeopardise Romania's integrationinto NATO and the EU.Thus, the PRM founditself ostracised by the other parties at the peak of its electoral success and had no other option but to remain in opposition. Due to the inability of its leader to deal with the cordon sanitaire, the PRM continuously lost support and dropped out of parliament in 2008 (Adamson, Florean, and Thieme 2011, 322; Cinpoeş 2015, 288–89). Neither the PRM nor any other radical right party has entered the Romanian parliament since. This does not mean, however, that radical right personalities or politics are absent from the Romanian party system. Indeed, there is a widespread trend of "political cruising"and "casual intolerance" (Cinpoeş 2015, 290–91), meaning that politicians, including radical right ones, frequently switch party allegiances and that Romanian parties struggle to distance themselves from intolerance and discrimination.

## **7. Calibration: Preparing the data for the crossnational analysis**

The aim of this chapter is to transform the raw data presented in the previous two chapters into a binary dataset, a prerequisite for analysis using csQCA. To do so, it is necessary to define thresholds of set membership for all conditions and the outcome, in particular the so-called threshold of indifference, which demarcates set membership from non-membership. In the dataset, membership and non-membership are coded as 1 and 0, respectively.

#### **7.1 The outcome: Government participation**

In most cases,it is rather obvious whether a party is in government or in opposition. The definition of government coalitions presented in Chapter 4, however, shows that the situation can be less clear when it comes to support for minority governments. This study has established that parties which endorse minority governments should also be considered as members of the government if their support is permanent, publicly acknowledged, and mutually agreed upon. Hence, the outcome (GOVPART) is coded as present if radical right parties are either junior partners in a government coalition or they meet these criteria while supporting aminority government. If they do not fulfil these criteria, then radical right parties are considered non-members of the set of radical right parties in government, regardless of whether or not they back the government in individual parliamentary votes. Table 7.1 provides an overview of radical right parties' participation in government in Central and Eastern Europe and the corresponding membership score.

Between 1990 and 2020, the radical right has entered government in almost every country where it gained representation in parliament. The only exception is the Czech Republic, where none of the three radical right parties have participated in a coalition so far. In 22 out of a total of 48 cases, the radical right is in the set of government participants. Thus, once Central and Eastern European radical right parties gain seats in parliament, they participate in government almost half of the time. These data dispel the common knowledge that radical right parties are pariahs.





Source: Own compilation.

## **7.2 The context factors: Bipolar opposition and fragmentation in the party system**

Bipolar opposition describes a situation where two political camps are so deeply divided that including a party from the opposing camp in a government coalition would be impossible. The impact of bipolar opposition on government formation is determined by whether or not the radical right and the formateur are in the same camp. The condition SAMESIDE combines both aspects. If bipolar opposition is present and radical right parties are in the same camp as the formateur, then they are coded as members of the SAMESIDE set. Thus, if there is no bipolar opposition or if bipolar opposition exists but the radical right party is not in the formateur's camp, then the set membership is 0. Empirically, the calibration of this condition builds on the qualitative assessment of party system configurations derived from the secondary literature discussed in the country case studies (see Chapters 5 and 6).

The second contextual condition accounts for the fragmentation of party systems (FRAG).The data for this condition is available on a metric scale, but there is no commonly used, qualitative threshold to delineate fragmented party systems from compact ones. In such cases, the calibration of set membership must resort to using the empirically observed data, which is the least preferable option. The distribution of the data as well as the mean or median can serve as starting points for the calibration. They should not be used as the only source for defining the threshold of indifference, however, since this would make thresholds very sensitive to case selection. Instead, the empirical data should be examined in light of the concept it is supposed to measure (Berg-Schlosser and Cronqvist 2012, 197–98; Schneider and Wagemann 2012, 33–35).

In this study, the fragmentation of party systems serves primarily as an indicator of the complexity of the bargaining environment from a numerical-structural perspective.The fragmentationin the empirically observed cases ranges from aminimum of 3.2 to a maximum of 6.4 effective parliamentary parties (see Table 7.2).The median and the mean are both at 4.4, which is a rather high value in the context of European party systems (Enyedi and Casal Bértoa 2018). This observation suggests that the line between fragmented and compact party systems should be drawn below, rather than above, this value. The data shows that there are only two cases with a value near four effective parties. Apart from these two cases, there is a relatively large gap between 3.6 and 4.2 effective parliamentary parties in the data. The complexity of the bargaining situation in the cases in this area shall be examined in more detail. This investigation will assess the bargaining situation from a purely numerical perspective based on the effective and total number of parliamentary parties. The analysis starts with the cases that constitute the lower and upper boundaries of the gap.These include Poland in 2001 and Romania in 2000, as well as the Czech Republic in 1996 and Estonia in 2019. It then turns to the cases within the gap, Latvia and Slovakia in 2010.

The 2001 Polish parliament and the 2000 Romanian parliament show a fragmentation score of 3.6 effective parliamentary parties.The winner of the 2001 Polish parliamentary elections came very close to a majority and could have chosen between any of the five other parties to form a two-party majority coalition. A majority government without the dominant party would have required all of the other parties to cooperate.The result of the 2000 Romanian parliamentary election created a similar bargaining situation.The five-party parliament was dominated by a single party that controlled almost twice the number of seats as the runner-up and could have formed a majority coalition with any of the other parliamentary parties. Despite five and six potential minimal winning coalitions, respectively, the value of 3.6 effective parliamentary parties reflects a moderately complex bargaining situation in both cases. Hence, the threshold of indifference should be set above this value.

The two cases with 4.2 effective parties, the Czech Republic in 1996 and Estonia in 2019, constitute the upper boundary of the gap. The 1996 Czech parliament consisted of six parties, two of which were relatively strong and able to form a majority coalition, either with each other, or with at least two of the smaller parties. In total, the parties could have formed eight different minimal winning coalitions, leaving every party with two or more options to enter government. In the Estonian case, there were only five parties in parliament but the seats were distributed more evenly with 34, 26, 19, 12 and 10, respectively. Thus, the strongest party could have formed a majority coalition with either the second or third strongest party. The parties in second and third place, however, did not control enough seats to form a majority without one of the two small parties. Thus, even though there were only five possible minimal winning coalitions, most parties had more than two options for getting into power. Thus, the fragmentation score of 4.2 reflects rather complex bargaining situations, suggesting that the threshold should be set below this value.

Therefore, the focus turns to the two cases that lie within the gap. In 2010, Latvia elected a five-party parliament with a fragmentation score of 3.9. Three of the parliamentary parties were relatively strong, holding 33, 29 and 22 seats out of a total of 100, while the other two were small with eight seats each. This constellation creates a moderately complex bargaining situation with only three possible minimal winning coalitions: The strongest party could form a majority coalition with either the second or the third strongest party, or the second and third strongest parties could govern together. All other coalitions are either short of a majority or oversized. Even though the distribution of seats is somewhat similar to the case of Estonia in 2019, the bargaining situation in Latvia is less complex because, from a purely office-seeking perspective, the two small parties areirrelevant for government formation.Thus, the fragmentation score of 3.9 effective parliamentary parties reflects a moderately complex bargaining situation.

In 2010, Slovakia had 4.0 effective and six actual parliamentary parties. One of the six parliamentary parties was relatively close to a majority, holding 62 of the 150 seats, and one very small party held only nine seats. All of the remaining parties were large enough to form a two-party majority coalition with the strongest party. Thus, the strongest party could choose between four possible partners to form a majority government. Alternatively, a coalition of all four medium-sized parties could have controlled a majority, thus allowing for a total of five possible minimal winning coalitions, which makes the bargaining situation somewhat more complex than in the Latvian case. However,it resembles the situation in Romania and Poland in 2000 and 2001, respectively, where coalition bargaining involved two scenarios and thus only a moderate level of complexity, suggesting that the Slovak case should not be considered as a member of the set of fragmented party systems.

Hence, cases are members of the set of fragmented party systems (FRAG), if the number of effective parties exceeds 4.0. Cases with four or less effective parliamentary parties are considered non-members of the set. Table 7.2 presents the data and the membership scores of the individual cases in the sets of the two context conditions (SAMESIDE and FRAG). The table shows that the calibration has resulted in a rather uneven distribution of both conditions. Radical right parties are on the same side of a bipolar opposition as the formateur in only 15 of the 48 cases, and the party systems show a low level of fragmentation in only 13 cases. Nevertheless, the diversity of the observed cases with regard to both conditions is still high enough to include them in the analytical model.

The table also reveals that there are somewhat stable patterns in the configuration of the party systems in some countries. All four Czech cases, for instance, are characterised by the absence of a bipolar opposition and high levels of fragmentation, even though they refer to different radical right parties in different periods.The Estonian party system shows a similar configuration, even though the absence of a bipolar opposition is based on different reasons than in the Czech Republic. Almost all Latvian cases, in contrast, are characterised by high levels of fragmentation and the radical right is located on the same side of the bipolar opposition in the party system as the formateurs. Hence, it might well be that the explanatory patterns to be analysed in the following chapters reflect these country-specific patterns to some degree (Müller, Bergman, and Strøm 2008, 19–20).


*Table 7.2: Configuration of the party system: Bipolar opposition and fragmentation*




Source: Own compilation; for data on bipolar opposition, see Chapters 5 and 6; data on fragmentation: Casal Bértoa 2021, amended by the author.

## **7.3 Characteristics and preferences of radical right parties**

The last step is to calibrate the party-level factors, which include the socio-economic and socio-cultural distance between radical right parties and the formateur, and radical right party's parliamentary strength.The seat shares of the radical right parties in the dataset range from 3.6 per cent for the Hungarian MIÉP in 1998 to almost 25 per cent of the seats in case of the Romanian PRM in 2000 (see Table 7.3).The median is 8.7, the mean 9.5 per cent of the seats, which reflects the radical right's modest, average electoral results in Central and Eastern Europe.The literature identifies single- versus double-digit electoral results as a criterion for distinguishing large radical right parties from their less successful counterparts (Fagerholm 2021).While ultimately one seat more or less may not make much difference from a mathematical perspective, crossing this barrier serves as a certain landmark for small parties and strengthens their bargaining position at least psychologically.Therefore, radical right parties are considered to be members of the set of large parliamentary parties (SEATS) if they control at least ten per cent of the seats in parliament. Below that threshold, they are not considered to be members of this set.

The socio-economic and socio-cultural positions of radical right parties and formateurs are measured with a similar indicator—the LRECON and GALTAN party scores in the CHES (see Chapter 4.4). Therefore, the calibration of the two conditions of socio-economic (LRECONPROX) and socio-cultural proximity (GALTAN-PROX) between radical right parties and the formateur can be discussed together. The CHES scores range from zero to ten and entail a qualitative threshold at a value of five, which separates the socio-economically left from the right and the socio-culturally liberal from nationalist-authoritarian positions, respectively. Therefore, one option would be to consider radical right parties and formateurs ideologically proximate, if they are positioned on the same side of this qualitative threshold. When doing so, however, even a difference of almost five points between a radical right party and a formateur could still be regarded as ideological proximity if the formateur is positioned just above five and the radical right party close to ten. Conversely, two parties that hold centrist positions but are situated on either side of five would not be regarded as members of the set of ideologically proximate parties. Hence, calibrating set membership only on the basis of this threshold could lead to membership scores that do not adequately reflect theideological distance between radical right parties and the formateur.






Source: Own compilation; data on seat share: Nordsieck 2021; data on ideological positions: Chapel Hill Expert Survey (Jolly et al. 2022), amended by the author's qualitativeassessmentbasedonsecondaryliterature(seeChapters5and6).

 \* Bold print indicates that the radical right party and the formateur occupy positions on opposite sides of the socio-economic or socio-cultural dimension, respectively.

Therefore, in addition to this qualitative threshold, the distance between these parties will be used to calibrate the LRECONPROX and GALTANPROX sets. If the radical right and the formateur are positioned on the same side of the socio-economic or socio-cultural dimension, respectively, they are considered to be members of the set of proximate parties if no more than 2.5 points separate them. If they are not positioned on the same side, the relative distance required to speak of ideological proximity is reduced to a maximum distance of 1.5 points in order to ensure that both parties indeed occupy centrist positions. Following this logic, the smaller range also applies if the radical right party, or the formateur, are positioned exactly at 5.00.

The distribution of cases in the SEATS set reflects the relative electoral weakness of radical right parties in Central and Eastern Europe. Only in one-third of the cases did a radical right party control at least ten per cent of the seats in parliament. With the exception of Poland and the individual case in Hungary, set membership also varies within countries over time, even though certain country-specific patterns emerge.The Latvian NA, forinstance, has controlled a large number of seats formost of its time in parliament, whereas radical right parties in Bulgaria usually remained below this threshold.

Membership in the LRECONPROX and GALTANPROX sets is more evenly distributed. Socio-economic proximity between the radical right and the formateur exists in 28 out of 48 cases and socio-cultural proximity in 19 out of a total of 48 cases. Given the radicalisation of mainstream parties and the dissemination of radical right politicsin Central and Eastern Europe (Minkenberg 2013, 2017;Mudde 2018; Pytlas 2018), the relatively low number of casesin the GALTANPROX set is somewhat surprising. The membership scores in Table 7.3 also indicate that ideological proximity on one dimension does not necessarily coincide with proximity on the other, supporting the argument that these dimensions are best studied separately.

Now that the calibration of set membership is complete, the study can continue with the comparative causal analysis using QCA. As discussed in Chapter 4, this analysis will be performed separately for the period before and after the first thirdgeneration elections.

## **8. Government formation with radical right parties in the nascent post-Communist party systems**

This chapter analyses government formation with radical right parties in Central and Eastern Europe during the time before the first third-generation elections (Pop-Eleches 2010; see also Chapter 4), a period which largely corresponds to the first post-Communist decade. In accordance with good practice in QCA, the investigation contains separate analyses of radical right parties' inclusion in, and exclusion from, government, each of which examines the necessary and sufficient conditions for the respective outcome.

## **8.1 Explaining the government participation of radical right parties**

### 8.1.1 Analysis of necessity

The first step of a comparative analysis with QCA, the search for necessary conditions, starts with an examination of the consistency, coverage, and relevance (RoN) of the five conditions(and their negations) specifiedin the analyticalmodel. Table 8.1 presents the parameters of fit for the period before the first third-generation elections. These indicate that proximity to the formateur on the socio-economic (LRE-CONPROX) and the socio-cultural (GALTANPROX) dimensions qualify as necessary conditions for the inclusion of radical right parties in government.These conditions have the highest possible consistency score (1.00), indicating that ideological proximity to the formateur on both dimensions was present in all seven cases in which the radical right entered government in the transformational decade. High RoN and coverage scores establish that neither factor constitutes a trivial necessary condition.

None of the other conditions or their negations reach a consistency of 0.9, the minimum requirement for necessary conditions. High fragmentation of the party system (FRAG) and a small seat share of the radical right (∼SEATS) are the only conditions that come close to passing this threshold.Their low RoN score, however, signals that the relatively high consistency results from both conditions being present in almost all instances of government formation with radical right parties before the first third-generation elections, regardless of whether these parties enter government or remain in opposition. Hence, no factors other than the socio-economic and socio-cultural proximity between radical right parties and the formateur qualify as necessary conditions. In order to confirm whether the causal relationship indicated by the parameters of fit exists, the following section assesses these conditions in more detail.


*Table 8.1: Parameters of fit necessity: Government participation of radical right parties (before first third-generation elections)*

Source: Created with QCA Package for R (Duşa 2019).

In the case of the ERSP's participation in the Estonian government of 1992, the issue of Russian-speaking minority rights dominated the political debate. In the newly independent country, the parties of the Estonian majority had already introduced an electoral law which stipulated that only the citizens of the inter-war Estonian state, and their descendants, had the right to vote in the 1992 parliamentary elections. Other residents of the country could only obtain citizenship after a threeyear naturalisation process (Raun 1994, 74). Because these requirements were impossible for most non-ethnic Estonian residents to fulfil prior to the election, they could neither run for office nor vote in 1992.Thus, a large part of the Russian-speaking population was excluded from electoral politics and, consequently, the parliament consisted only of parties that favoured an ethnic model of democracy (Lagerspetz and Vogt 2013, 66; Braghiroli and Petsinis 2019, 438). Under these circumstances, the inclusion of a party that disagreed on the pivotal ethnic issue would be hard to imagine. Hence, socio-cultural proximity between the radical right ERSP and Pro Patria, the formateur of the 1992 coalition, was indeed a necessary condition for their cooperation.

In addition to similar positions on the ethnic question, all parties which entered the Estonian parliament in 1992, including the radical right ERSP, shared the desire to rapidly establish a market economy (Pettai 1993; Pettai and Kreuzer 1998). Thus, the ERSP's socio-economic proximity to Pro Patria also facilitated the party's participation in government. Moreover, demarcating the Estonian majority from the Russian-speaking minority, and their kin state Russia, united the socio-cultural and the socio-economic dimensions, because this issue entailed both support for an exclusionary ethnic construction of nationhood in politics and society as well as the immediate introduction of a market economy, which promised economic cooperation and security through integration into Western alliances. The alignment of the socio-economic and socio-cultural dimensionin the Estonian party system suggests that the simultaneous proximity between ERSP and Pro Patria on both dimensions was a necessary condition for the government participation of the Estonian radical right in 1992.

In Romania the issue of state- and nation-building, and national identity, particularly with regard to the rights of the Hungarian minority in the country, figured prominently in the 1992 general election campaigns. The two radical right parties, PUNR and PRM, were extremely hostile to the Hungarian minority. These parties, and organisations close to them, even orchestrated, or condoned, violent incidents (Shafir 1999).The Communist successor party, FDSN, formateur of the governments with radical right parties in Romania in the early 1990s, was slightly more moderate in this respect. Due to international pressure, the party quickly distanced itself from acts of physical violence (Vachudova 2005, 101–2). Otherwise, however, it shared many of the radical right's positions. The importance of socio-cultural proximity for government participation is further illustrated by the early termination of the government in 1996. The coalition disintegrated after President Iliescu of the PDSR—formerly the FDSN—signed a neighbourhood treaty with Hungary, which caused the PUNR to leave the coalition (de Nève 2002, 335).

The ideological proximity between the radical right parties and the FDSN also entailed redistributive and protectionist socio-economic positions. For example, in the early 1990s the FDSN prioritised slowing down the economic transformation. The PRM shared the FDSN's scepticism towards privatisation and free market economy. The PUNR was more moderate in this regard but it did not favour rapid economic transformation either (Gabanyi 1997; Gallagher 1997; Shafir 1999; Bugajski 2002; Pop-Eleches 2008; see also Chapter 6.4).

Overall, the Romanian party system of the 1990s was characterised by a strong bipolar opposition based on the legacy of Ceauşescu's specific brand of national Communism. The successors of the old regime, including the FDSN, the PRM and, to a lesser degree, the PUNR (Shafir 1999, 214; Grün 2002; Pop-Eleches 2008), adhered to economic protectionism and nativism, while the oppositional camp held rather liberal socio-economic and socio-cultural views. Thus, similar to the Estonian case, cooperation between FDSN and the radical right in Romania also required simultaneous socio-economic and socio-cultural proximity.

In Slovakia, the radical right SNS became a junior partner in two coalitions led by Vladimír Mečiar's HZDS in 1992 and 1994. In 1992, the independence of the Slovak state was at the centre of the political debate. Mečiar and his party campaigned for greater Slovak autonomy because of their opposition to the neoliberal economic policies imposed by the federal government in Prague, along with broader centre-periphery conflicts. The radical right SNS voiced ethnic nationalist resentment against the Czech population and demanded the complete dissolution of the Czechoslovak federation. Later on in the campaign, Mečiar adopted the radical right's narrative, as well as the demand for an independent Slovak state, in order to win popular support (Szomolányi and Mesežnikov 1997; Fisher 2006; Mesežnikov 2008). Thus, HZDS and SNS agreed on the key socio-cultural issue of that time, which facilitated the formation of a joint government.

Although the 1994 elections took place in an independent Slovak state, the salience of state- and nation-building remained high. On the one hand, Hungary's nationalist policies fuelled increasing scepticism, and even open hostility, towards the Hungarian minority in the country. On the other hand, the two governing parties set out to build an autocratic and clientelistic state that seriously endangered Slovakia's democratic consolidation and integration into Western alliance systems, particularly the EU. While the SNS focused on the former aspect, Mečiar's HZDS concentrated on the latter. In principle, however, the parties continued to agree on crucial socio-cultural issues, and these shared positions constituted the foundation of their renewed cooperation in the 1994 government.

The ideological platforms of the SNS and the HZDS combined nationalism with economic protectionism.The HZDS attempted to slow down the economic transformation set in motion by the central government after Czechoslovakia's Velvet Revolution. The SNS also supported national-protectionist economic policies in principle, but somewhat less vehemently than the HZDS or their Romanian brethren (Szomolányi and Mesežnikov 1997, 143; Gyárfášová and Mesežnikov 2015, 229–30; see also Pirro 2016). Unlike in Romania, the SNS was not the only potential junior partner that shared the formateur's fundamental socio-economic position. The socio-economic platform of the reformed Communist successor party, SDĽ, was similar to that of the HZDS as well. However, Mečiar preferred a coalition with the SNS, which was ideologically closer to his party on the socio-cultural dimension and ultimately enabled his autocratic rule.Thus, the Slovak case further substantiates the assumption that both ideological proximity conditions are necessary for government participation, even in a party system where the two dimensions are not reinforcing each other.

These empirical observations corroborate that small socio-economic and sociocultural distances are of great importance for the participation of radical right parties in government during the period before the first third-generation elections (Hypothesis 2a).The analysis demonstrates that the concurrence of socio-economic and socio-cultural proximity is a necessary condition for government participation.The parameters of fit for the conjunction LRECONPROX\*GALTANPROX are equal to 1.00 and therefore support this conclusion. Hypothesis 2a also includes the possibility that radical right parties could enter government regardless of their ideological distance to the formateur if they are situated on the same side of a bipolar opposition. Because the joint presence of socio-economic and socio-cultural proximity is a necessary condition for government participation, however, this implies that radical right parties never entered government if their ideological positions differed substantially from those of the formateur.Thus, bipolar opposition in the party system never overshadowed ideological distance, indicating that socio-economic and socio-cultural proximity are more important for government participation of radical right parties than initially expected.

#### 8.1.2 Analysis of sufficiency

The analysis of sufficient conditions, the second step in a comparative analysis with QCA, begins by compiling the truth table from the 14 cases of government formation with radical right parties in the transformational decade. Table 8.2 shows that only four of the 32 possible combinations of the five conditions have been empirically observed in this period. The truth table includes no contradictions, which means that each row covers only cases in which radical right parties either entered government or remained in opposition. All rows that contain cases of government participation of radical right parties have a perfect consistency of 1.00, so this value is used as the consistency cut-off value for the computer-assisted minimisation process. The outcomes in rows 1 – 3 are coded 1,indicating that they are cases of radical right government participation, while row 4 covers all cases in which the radical right did not enter government, and they are coded 0.<sup>1</sup>

The empirically observed cases are clustered in very few truth table rows, so the conservative solution yielded by logical minimisation with the fsQCA software is the

<sup>1</sup> Another criterion for coding the outcome is the number of cases represented in a truth table row. As this study involves a rather small number of cases, the frequency cut-off is set to 1, which means that every row that represents at least one empirically observed case is not considered a logical remainder (Schneider and Wagemann 2012, 152–53).

 

product of a single minimisation step:<sup>2</sup>

$$\begin{Bmatrix} \text{LRECONPROX} \ast \text{GALTAANPROX} \ast \sim \text{SEATS} \ast \text{FRAG} \\\\ \text{LRECONPRX} \ast \text{GALTAANPROX} \ast \text{SEATS} \ast \sim \text{FRAG} \ast \sim \text{SAMESIDE} \end{Bmatrix} \to \text{GOVPARFT} \ast \text{BIFF}$$

Because this solution is parsimonious enough to allow for a reasonable interpretation of the results, and the primary goal of this study is to explain the government formation of radical right parties in the empirically observed cases (see Chapter 4.1), there is no need for further minimisation with the help of logical remainders.<sup>3</sup>

<sup>2</sup> The standard analytical procedure in the fsQCA software yields three solutions, depending on the inclusion or exclusion of different types of logical remainders (see Chapter 4).

<sup>3</sup> The parsimonious solution, including all logical remainders that contribute to parsimony, is either *LRECONPROX → GOVPART* or *GALTANPROX → GOVPART*, depending on the choice between these two tied prime implicants (Ragin 2018). Neither of the two solutions includes the complete necessary condition identified above, which highlights that the parsimonious solution is often based on untenable assumptions (Schneider and Wagemann 2012). This observation further supports the author's choice to build the analysis on the conservative solution.





Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The solution coverage and consistency reach the maximum value of 1.00 (see Table 8.3). Thus, the solution explains all of the empirically observed cases of government participation of radical right parties in this period, and none of the cases covered by the solution refers to a radical right party that remained in opposition. Hence, these parameters of fit suggest that the conservative solution term qualifies as a sufficient condition for the government participation of radical right parties in the period before the first third-generation elections. In order to substantiate this claim, the remainder of this section goes back to the cases and examines in more detail whether the two sufficient paths that constitute the solution term offer theoretically sound explanations for government participation of radical right parties.

Before doing so, the two previously identified necessary conditions (LRECON-PROX and GALTANPROX) can be factored out, as they are present in both of the sufficient solution paths.This operation results in the following term, the content of which is identical to the term above and in Table 8.3:

*LRECONPROX\*GALTANPROX (*∼*SEATS\*FRAG + SEATS\**∼*FRAG\**∼*SAMESIDE) GOVPART* →

This term reveals that LRECONPROX and GALTANPROX are indeed necessary parts of both solution paths but that these two conditions alone are not sufficient for the outcome to occur. Since the role of socio-economic and socio-cultural proximity between radical right parties and the formateurs has already been discussed, the following remarks focus on the additional conditions in the two solution paths.

*Table 8.3: Sufficient conditions for the government participation of radical right parties (before first third-generation elections)*


Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The first solution path (LRECONPROX\*GALTANPROX\*∼SEATS\*FRAG) explains six of the seven instances of government participation. Here, ideological proximity to the formateur on the socio-cultural and socio-economic dimension is accompanied by a small seat share of the radical right and high levels of fragmentation in the party system. In this situation, the formateurs needed to cooperate with at least two, and in Romania even three, other parties in order to secure a parliamentary majority for the cabinet.

The interplay between the ideological and numerical factors is best illustrated in the Slovak and Romanian cases. After the 1994 Slovak parliamentary elections, the HZDS won more than 40 per cent of the seats in a parliament consisting of an effective number of 4.4 and a total number of six parties. Being so close to a majority, the party could have formed a two-party majority coalition with four of the other six parties in parliament. The result, however, was a government of the HZDS with the two smallest parliamentary parties, one of them being the radical right SNS. Mečiar initially negotiated with the SDĽ and the Christian democratic KDH about forming a coalition, but ideological differences prevented these negotiations from succeeding.Only afterwards did Mečiar turn to the SNS and the ZRS, with whom he reached an agreement (Malová 1995). Due to the simultaneous socio-cultural and socio-economic proximity between the formateur and the radical right party, as well as the availability of another small party in the fragmented Slovak parliament, the SNS entered government in 1994 despite its low seat share. Thus, a small number of seats did not exclude the SNS from government, but it should not be considered an unequivocal advantage either.The sequence of coalition talks suggests that formateurs prefer coalitions with larger junior partners, even in fragmented party systems, and they turn to smaller parties only if it results in an ideologically homogeneous government.

In Romania, the Communist successor party, FDSN, won more than one-third of the seats in the chamber of deputies in the 1992 general elections. Both governments that formed during this legislature included two radical right parties, PUNR and PRM. In 1992, the Romanian party system was even more fragmented than the Slovak one in 1994, featuring a total of seven parties and an effective number of 4.4 parties in parliament. From a purely office-seeking perspective, there was only a single two-party coalition that would have controlled a majority of seats.This hypothetical coalition of the FDSN and the CDR, the electoral alliance that emerged from the forces outside the Communist Party that fought the Ceauşescu regime, was ruled out due to the regime divide which structured party competition and shaped the parties' policy positions at the time. Therefore, the FDSN depended on ideologically compatible parties in its own camp to form a parliamentary majority, despite their low seat share. In the highly fragmented parliament, all four parties in this camp were required to form a majority. Even the support of the PRM, which held less than five per cent of the seats in parliament, was vital for the FDSN minority government to assume office.The PRM remained a support party until the government collapsed in 1996, whereas the PUNR, which was ideologically closer to the PDSR and controlled a larger number of seats, received cabinet posts the government re-formed in 1994 (Gallagher 1994, 30–32; Shafir 1999, 216; Autengruber 2006, 70–71). This case provides further support for the argument that majority governments require more parties in fragmented than in compact party systems, which increases the chances for small parties to gain executive power. Hence, only the interplay of all four conditions in the first solution path explains the government participation of the PRM and the PUNR.

The last case covered by the first solution path is the government participation of the ERSP in 1992. With 5.9 effective parliamentary parties, the fragmentation in the Estonian party system was even higher than in Slovakia and Romania. Pro Patria emerged victorious from the parliamentary elections but controlled less than 30 per cent of the seats in parliament. The party thus needed at least two more parliamentary parties to form a majority government. The seven parties represented in the 1992 Estonian parliament, including the ERSP and Pro Patria, held quite similar views on the economic transformation of the country and the rights of the Russianspeaking minority. Alongside this socio-economic and socio-cultural proximity, the fragmentation of the party system contributed to ERSP's government participation. The formation of a coalition that did not involve more than three parties was helped by ERSP's seat share.The party controlled 9.9 per cent of the parliamentary seats and could therefore make a more substantial contribution to the parliamentary majority of the government than the radical right parties in Romania or Slovakia. Hence, this case further illustrates that governments in fragmented party systems include a larger number of (small) parties. At the same time, however, Pro Patria formed a government that consisted of no more than three parties, suggesting that formateurs still try to keep the number of junior partners as low as possible. Thus, the radical right ERSP benefitted from the fragmentation of the Estonian party system, and its near-ten per cent seat share further improved the party's bargaining position.<sup>4</sup>

The second solution path (LRECONPROX\*GALTANPROX\*SEATS\*∼FRAG\* ∼SAMESIDE) covers the government participation of the SNS in 1992. The unique coverage of this path indicates, however, that this case is not covered by the first solution path (see Table 8.3). Hence, the second path contributes to the understanding

<sup>4</sup> Lowering the threshold for large radical right parties from 10.0 to 9.9 per cent, so that the ERSP's set membership changes from 0 to 1, would alter the result only marginally. The result of the minimisation would then be *LRECONPROX\*GALTANPROX\**∼*SEATS\*FRAG + LRECONPROX\*GALTANPROX\*SEATS\**∼*SAMESIDE → GOVPART*, whereby the case of the ERSP would now be covered by the second solution path together with the SNS in 1992. The fact that such a minor change of the threshold does not lead to substantial changes in the QCA solution is an indicator of the robustness of results in QCA (Schneider and Wagemann 2012, 287–91).

of government formation with radical right parties in the early years of the post-Communist transformation, even though it describes only one individual case. This case differs from the previous six in that it represents the only example of a radical right party with a large seat share entering government. Unlike in the following election, the SNS won a relatively high number of seats in the 1992 Slovak parliament (10.0 per cent). Moreover, the parliament consisted of only five actual and 3.2 effective parties, which was unusually compact at that time. Of the two numerical factors, however, the low fragmentation of the party system was more decisive in this specific case than the SNS' high seat share.The HZDS, as the formateur and the strongest party, held 74 of the 150 seats in parliament.Thus, a two-party coalition of HZDS and any of the other parliamentary parties would have controlled a majority. Due to the low fragmentation, however, Mečiar's choice of potential junior partners was much smaller than in 1994. Among the four available parties, the SNS turned out to be the ideologically most compatible partner, because it largely agreed with the HZDS on the central socio-economic and socio-cultural policies, such as an incremental and clientelistic transformation of the economy and Slovak independence based on nativism. The negation of the SAMESIDE condition signals that the formation of the 1994 Slovak government was not constrained by strong bipolar opposition. Unlike in Romania, for instance, the regime divide did not structure Slovak politics in the early 1990s (Grzymała-Busse 2001, 98).

Based on this discussion of the cases, the two solution paths can indeed be considered sufficient conditions for the government participation of radical right parties in Central and Eastern Europe before the first third-generation elections. In addition to the crucial importance of socio-cultural and socio-economic proximity, the results illuminate the role played by the two numerical factors. For instance, the high fragmentation of Central and Eastern European party systems helps electorally weak radical right parties come to power.The case of the SNS in 1992 shows that radical right parties can also gain cabinet posts in less fragmented party systems. Since the HZDS came very close to a majority of its own, the ideological proximity of the two parties and the low fragmentation of the party system that enabled the formation of a two-party majority coalition turned out to be more relevant explanations for the government participation than the SNS' large seat share in this case.

These findings confirm the theoretical assumptions regarding the numerical factors in the transformational decade. With one exception, it is indeed electorally weak radical right parties in fragmented party systems that entered government in this period (Hypothesis 1b). The coalition involving the Slovak SNS in 1992 diverges from the dominant pattern but it does not fundamentally contradict the theoretical expectations regarding the connection between parliamentary strength and the fragmentation of party systems, because the only configuration that should prevent radical right parties' from entering government is a small seat share in compact party systems (Hypothesis 1a).

Hypothesis 2a posits that radical right parties should enter government if they hold similar socio-economic and socio-cultural positions and/or are on the same side of a bipolar opposition in the party system as the formateur. The simultaneous presence of socio-economic and socio-cultural proximity even constitutes a necessary condition for government participation. The presence of the SAMESIDE condition, however, which indicates that the radical right party and the formateur are on the same side of a bipolar opposition—in this period most likely rooted in the regime divide—does not occur in either of the two solution paths.Thus, Central and Eastern European radical right parties did not have to be on the same side of the regime divide as the formateur in order to enter government.

However, the analysis demonstrates that the regime divide still affected government formation with radical right parties in the 1990s indirectly as text and context (Minkenberg 2009; see also Chapter 2). In case of the government participation of the PRM and PUNR in Romania, for instance, the party system was shaped by strong bipolar opposition based on the regime divide and the ideological positions of FDSN, PRM, and PUNR were strongly influenced by the legacy of the national Communist Ceauşescu regime. In Estonia, the regime divide was intertwined with the ethno-linguistic divide. It provided a reference for central socio-economic and, in particular, socio-cultural positions of political parties. Due to the restrictive electoral law introduced before the 1992 elections, however, the regime divide did not produce bipolar opposition in the party system that constrained government formation. Since large parts of the Russian-speaking minority did not receive active and passive voting rights after Estonia's independence from the Soviet Union, this pole of the divide was not represented in parliament (Raun 1994; Saarts 2011; Lagerspetz and Vogt 2013). Hence, the constraining effect of the regime divide on party competition in Estonia unfolded before the post-electoral stage. The regime divide was much less prominent in Slovakia. Here, the confrontation between the Communist regime and the oppositional forces was less violent than in the Czech lands. In addition, the Slovak Communist successor party, SDĽ, undertook credible reforms and the representatives of the former regime joined various parties,including the HZDS, which at the time was a successor organisation to the oppositional alliance, Public Against Violence (Grzymała-Busse 2001, 98–100; Bugajski 2002, 311).

Further conclusions can be drawn regarding the composition of governments. The analysis confirms Hypothesis 3a, showing that radical right parties always entered government as junior coalition partners, or as support parties of minority governments, but never as members of an oversized coalition. In Romania, both the PRM and the PUNR were involved in a minority government as support parties after the first free elections. The PUNR later became a full-fledged junior partner in the coalition and received ministerial posts. In Estonia in 1992 and in Slovakia in 1992 and 1994, the ERSP and the SNS, respectively, served as junior partners in minimal winning coalitions.

The format of governments with radical right parties corresponds to the dominant type of coalitions in the respective country. In Romania,minority governments are a frequent phenomenon, whereas in Estonia and Slovakia, minimal winning coalitions are the most common coalition format between 1990 and 2014 (Bergman, Ilonszki, and Müller 2019a). This pattern is somewhat less clear when considering only the governments formed before the first third-generation elections. In Romania and Slovakia, the format of governments with radical right parties does not diverge much from the overall pattern in these countries in the 1990s. In Estonia, however, the 1992 Pro Patria-ERSP government is the only minimal winning coalition during the first post-Communist decade. Minimal winning coalitions emerge as the dominant type of government in Estonia only after the turn of the millennium.<sup>5</sup> However, a single government of this type does not warrant the conclusion that the format of governments with radical right parties diverges from the general pattern in the country.

## **8.2 Explaining the exclusion of radical right parties from government**

In QCA, the analysis of the negative outcome—the exclusion of radical right parties from government—is carried out separately because, as a set-theoretic method, it is fundamentally based on the assumption of asymmetric causality (Schneider and Wagemann 2012, 81–83; see also Chapter 4). This can be illustrated with the help of the two ideological dimensions. The first half of the analysis confirms that the combination of socio-economic and socio-cultural proximityis a necessary condition for the participation of the radical right in government. However, this result does not imply that ideological proximity on both dimensions must be absent in cases where radical right parties failed to enter government. Rather, as the positive outcome occurs only in the joint presence of both factors, the absence of either one could prevent it from happening. Put more generally, asymmetric causality means that the explanation of the negative outcome is not necessarily the exact opposite of the explanation of the outcome.

<sup>5</sup> The formats of the governments formed before the first third-generation elections are distributed as follows in the three countries: Estonia: 1 minimal winning coalition, 3 minority governments and 2 oversized coalitions; Romania: 2 minimal winning coalitions, 5 minority governments and 5 oversized coalitions; Slovakia: 3 minimal winning coalitions, 2 minority governments, 1 oversized coalition (Bergman, Ilonszki, and Müller 2019a).

## 8.2.1 Analysis of necessity

The analysis of the negative outcome begins with the search for necessary conditions. Table 8.4 contains the parameters of fit necessity to determine whether or not the individual conditions and their negations qualify as necessary conditions. Five conditions, FRAG,∼LRECONPROX,∼GALTANPROX,∼SEATS, and ∼SAMESIDE, show consistency scores above 0.9. However, the RoN values for high party system fragmentation (FRAG) and low seat share (∼SEATS) are so low that they must be considered trivial necessary conditions. They do not develop any causal traction for explaining the negative outcome because they are present in most of the cases, regardless of whether radical right parties entered government or remained in opposition in the first post-Communist decade. ∼SAMESIDE reaches higher coverage and RoN scores, but they are still too low for this factor to qualify as a non-trivial necessary condition.


*Table 8.4: Parameters of fit necessity: Exclusion of radical right parties from government (before first third-generation elections)*

Source: Created with QCA Package for R (Duşa 2019).

The negations of socio-economic (∼LRECONPROX) and socio-cultural proximity of the radical right party to the formateur (∼GALTANPROX), however, both clearly qualify as non-trivial necessary conditions with a coverage and RoN of 1.00. This finding corresponds to the directional expectations with regard to these two conditions in the period before the first third-generation elections. Hypothesis 2a implies that ideological distance on the socio-economic or the socio-cultural dimensions should prevent radical right parties from entering government. The data show, however, that all radical right parties who remained in opposition held distant socio-economic and socio-cultural positions to the formateur. However, in light of the necessary condition for government participation (LRECONPROX\*GAL-TANPROX) and the corresponding theoretical assumptions, here the negations of both individual factors are considered as necessary conditions for the exclusion of radical right parties from government. The interplay of these two conditions and the other explanatory factors will be examined in more detail in the analysis of the sufficiency.

## 8.2.2 Analysis of sufficiency

The analysis of sufficient conditions for the exclusion of radical right parties from government is based on the same truth table used previously (see Table 8.2 above), but the outcome is different (∼GOVPART). This outcome can be coded 0 in rows 1 – 3 and 1 in row 4. Because all observed cases of the exclusion of radical right parties from government in this period are clustered in a single truth table row, no minimisation is possible without using logical remainders. Thus, the configuration of this truth table row also represents the conservative solution, which is reported in Table 8.5. Since this row contains no contradictory cases and covers all observed instances of radical right parties that remained in opposition, the coverage and consistency of this solution equals 1.00. Due to the clustering of cases and because a solution term with only one path can be interpreted easily, the conservative solution serves as the basis for the interpretation.<sup>6</sup>

The solution indicates that radical right parties were excluded from government if they had a small seat share in a highly fragmented parliament, and they were neither socio-economically nor socio-culturally proximate to the formateur. ∼SAME-SIDE indicates that the radical right party was either not in the same camp as the formateur in the presence of bipolar opposition or that there was no bipolar opposition in the party system that constrained government formation to coalitions within one camp. The two numerical factors, small seat share and high fragmentation, appear in the same configuration that was observed when most radical right parties entered government in this period, reflecting the fluidity of party systems in the nascent post-Communist democracies. The analysis of government participation demonstrates that a small seat share does not constitute a substantial disadvantage for radical right parties in fragmented party systems when they are ide-

<sup>6</sup> The parsimonious solution generated with the fsQCA software is either ∼*LRECONPROX →* ∼*GOVPART* or ∼*GALTANPROX →* ∼*GOVPART*, depending on the researcher's choice between these two tied prime implicants.

ologically proximate to the formateur. If they are not, however, as in all the cases observed here, radical right parties remain on the opposition bench.

On the socio-economic dimension, all seven cases show a medium ideological distance of 3.5 to four points, but the radical right party and the formateur are always located on opposite sides of the spectrum. SPR-RSČ, SNS, PRM, and PUNR all hold—more or less distinct—national-protectionist positions. When they were excluded from government, these parties faced formateurs that emerged from the anti-Communist opposition camp in their countries, such as the Czech ODS, the Slovak SDK, and the Romanian CDR. These parties and coalitions unanimously advocated liberal pro-market economic policies.

*Table 8.5: Sufficient conditions for the exclusion of radical right parties from government (before first third-generation elections)*


Source:CreatedwithfsQCA(RaginandDavey2016).

 3.0

On the socio-cultural dimension the formateurs occupied rather centrist positions. The ODS and the SDK leaned slightly towards the green-alternative-libertarian (GAL) end of the scale. Within the Czech ODS, the conservative forces that were to set the tone for the party in later years had not yet gained the upper hand (Bugajski 2002). The situation of the Slovak SDK was quite similar. Here, the liberal, prodemocratic and pro-European forces also dominated the conservative ones in the late 1990s. After all, they had formed this alliance explicitly in opposition to the illiberal and autocratic Mečiar regime (Bugajski 2002, 301; Fisher 2006, 162–64). Thus, the distance between ODS and SPR-RSČ in the Czech Republic in 1992 and 1996 and between SDK and SNS in Slovakia in 1998 was substantial, including positions on opposite sides of the GALTAN spectrum.

The Romanian CDR faced rather towards the traditional-authoritarian-nationalist (TAN) end. However, the distance between the Christian democratic PNŢCD, the strongest member of the alliance, and the two radical right parties, PRM and PUNR, clearly exceeded 2.5 points, the threshold for ideological proximity. Even though nationalism was not completely absent within the CDR in general, and the PNŢCD in particular, it differed in kind from the aggressive nativist positions of the radical right parties. This difference was most evident in the parties' positions towards the Hungarian minority in Romania. For the PUNR and the PRM, the Hungarian minority and their kin state served as arch enemies, whereas the PNŢCD sought national reconciliation and even included the Hungarian minority party, UDMR, in the coalition formed after the 1996 election (Gabanyi 1997, 218–20).

In addition to these two party-level ideological factors, the solution includes the negation of the SAMESIDE condition. Yet, there are significant differences between the cases with regard to the two alternative sources of ∼SAMESIDE. The formation of the governments in Romania in 1996 and in Slovakia in 1998 was constrained by strong bipolar opposition in the party system. In Romania, the regime divide structured party competition in 1996. Here, the camp of the former opposition to the Communist regime, joined by the reformed Communist successor party, PD, managed to gain power for the first time since the fall of the Ceauşescu regime (Autengruber 2006, 72–74; see also Gabanyi 1997; Pop-Eleches 2008; Ștefan 2019). While the intensity of the regime divide was decreasing by the end of this legislature, cross-camp cooperation among political parties was absolutely impossible in the context of the 1996 Romanian general elections and the subsequent government formation.

In the context of the 1998 Slovak elections, the gulf between the pro-democratic alliance of SDK, SDĽ, and SMK on the one side, and the autocratic nationalist camp of the HZDS and SNS on the other, was at least as deep. Despite different political views, the oppositional alliance was united in their goal of toppling the Mečiar government—including the radical right SNS—and putting Slovakia back on track towards democracy, the rule of law, and EU membership (Pridham 2002; Hloušek and Kopeček 2008; Vachudova and Hooghe 2009). Hence, in these two instances of government formation, ∼SAMESIDE refers to a situation in which bipolar opposition structures the party system, but the radical right party does not belong to the same camp as the formateur. Under these circumstances, even the fragmentation of the Romanian and Slovak party systems in the late 1990s did not help the electorally weak radical right parties. Although both the 1996 Romanian and the 1998 Slovak government included a high number of parties, there was neither an ideological nor a numerical incentive for the formateurs to consider the radical right parties as potential partners in government.

In the Czech Republic in 1992 and 1996 and in Romania in 1999 ∼SAMESIDE refers to a different situation. Here, the party system is not characterised by a bipolar opposition that constrains government formation to alliances within one camp. Party competition in the Czech Republic was dominated by socio-economic divides, but until the mid-1990s, there were also other salient, cross-cutting divides. Hence, at that time, the Czech party system features multi-polar oppositions rather than a clear-cut bipolar one (Kitschelt et al. 1999, 226–30; Grzymała-Busse 2001; Vodička 2005; see also Balík and Hloušek 2016; Mansfeldová and Lacina 2019). In Romania, the regime divide had cooled considerably towards the end of the transformational decade, not least due to a reorientation of the PDSR. The party increasingly distanced itself from national Communism in the second half of the 1990s in order not to jeopardise Romania's accession to the EU, which was popular among Romanian voters(Pop-Eleches 2008, 470; Vachudova 2008,871; see also Ștefan 2019).These changes led to an erosion of the barriers between the PDSR and its former allies in the PD, and eased the relation between the PDSR and the constituent parties of the CDR as well.The minority government that formed after the 2000 general elections under the leadership of the PDSR was already supported by the oppositional PNL and the UDMR (Popescu 2003, 332). In these three cases, the ideological distance between radical right parties and formateurs was not reinforced by membership in opposite camps. In both countries, party systems were quite fragmented and governments comprised three or more parties. However, cooperating with small ideologically distant parties of the radical right was not an option for the formateurs, who preferred coalitions with other less radical parties. In fact, the positions of the SPR-RSČ and the PRM were considered so radical that all of their competitors had come to rule out cooperation with them (Čakl andWollmann 2005, 48; Cinpoeş 2015, 288).Hence, while there was no bipolar opposition that constrained coalition formation in these cases, the cordon sanitaire constituted a serious constraint for the two radical right parties in these countries.

In sum, a small seat share did not prevent radical right parties from entering government in the fragmented party systems of Central and Eastern Europe, but it did not help them either if they were ideologically distant from the formateur on the socio-cultural and the socio-economic dimensions. In some cases, the exclusion of

radical right parties from government is further aided by the fact that the party system was structured by bipolar opposition, and they were not in the same camp as the formateur. In others, the ideological distance was not reinforced by bipolar opposition, indicating that the two necessary conditions, ∼LRECONPROX and ∼GAL-TANPROX, have the greatest impact on the exclusion of radical right parties from government, even though these conditions alone are not sufficient for this outcome.

#### **8.3 Summary**

The empirical analysis provides support formany of the hypotheses on the formation of governments with radical right parties in the period before the first third-generation elections. With only one exception, small radical right parties in fragmented party systems entered government in the emerging post-Communist democracies of Central and Eastern Europe (Hypotheses 1a and 1b). However, this same configuration of factors has been observed for radical right parties which remained in opposition. Hence, it can be concluded that due to the high fragmentation of Central and Eastern European party systems during this transformational decade, the poor electoral performance of radical right parties did not prevent them from joining government coalitions.

Ultimately, however,ideological factors play the more important part in explaining why radical right parties make it into government or not. More precisely, simultaneous ideological proximity between radical right parties and the formateur on the socio-economic and socio-cultural dimensions is necessary for them to be included in government. Alternatively, radical right parties that remained in opposition always lacked ideological proximity on both dimensions (Hypothesis 2a). The presence, or absence, of these conditions constitutes a necessary condition for radical right parties' inclusion in, or exclusion from, government. Thus, the findings even suggest a linear causal relationship between the government participation of radical right parties and their socio-economic and socio-cultural proximity to the formateur in the period before the first third-generation elections. If party systems are characterised by bipolar opposition, then the position of radical right parties in the same, or opposite, camp can reinforce ideological proximity, or distance, respectively. Bipolar opposition, and particularly the regime divide, constrained government formation to coalitions within one camp less frequently than expected. However, the regime divideinformed parties' socio-cultural and socio-economic positions and thus affected the necessary conditions for radical right parties' inclusion in, and exclusion from, government. From a broader perspective, the crucial role of the party-level ideological factors in the process of government formation also supports the argument that ideological party competition already existed in the

early phase of the transformation in Central and Eastern European party systems (Hloušek and Kopeček 2010; see also Chapter 2.2).

As regards the composition of governments with radical right parties, the empirical observations confirm Hypothesis 3a: Radical right parties were never involved in oversized coalitions—they only served as junior partners in minimal winning coalitions or as support parties of minority governments. The format of governments with radical right parties tends to correspond to the dominant patterns within each respective country. Whether this initial finding applies to governments with radical right parties in Central and Eastern Europe more generally, however, will be evaluated in the analysis of the consolidating decades in following chapter.

## **9. Government formation with radical right parties in the consolidating democracies of Central and Eastern Europe**

After analysing government formation with radical right parties prior to the first third-generation elections, this chapter turns to government formation in the following two consolidating decades. During this period, the hypotheses state that radical right parties should be included in government if party systems are less fragmented and the radical right has been successful at the polls (Hypothesis 1c), and if they are socio-culturally proximate to the formateur and/or situated on the same side of a bipolar opposition as the formateur (Hypothesis 2b). Furthermore, radical right parties are not expected to be involved in oversized coalitions (Hypothesis 3a) and governments with radical right parties should be ideologically homogeneous, in particular on the socio-cultural dimension (Hypothesis 3b).

## **9.1 Explaining the government participation of radical right parties**

### 9.1.1 Analysis of necessity

The analysis of radical right government participation during the consolidating decades begins with the search for necessary conditions. Table 9.1 shows the parameters of fit for all conditions and their negations. Unlike in the earlier phase, none of these reach the required consistency threshold of 0.9; therefore, they cannot be considered necessary conditions. The consistency of high fragmentation falls just below this threshold, but even if it were slightly higher, the coverage and RoN are too low, indicating that this condition would be trivially necessary. Therefore, typical of much QCA research, the analysis produces no necessary conditions for the government participation of radical right parties in this period.

## 9.1.2 Analysis of sufficiency

The analysis of sufficient conditions begins by converting the raw data into a truth table (see Table 9.2). In contrast to the transformational decade, this truth table reveals that the number of logical remainders is significantly lower, which is partly due to the higher number of cases (N = 34). In the period after the first third-generation elections, 19 of the 32 possible combinations of the five conditions have been empirically observed. Yet, the truth table includes a contradictory configuration: The Estonian EKRE in 2019 and the Czech SPD in 2017 both share the same configuration of conditions found in row 8. However, EKRE participated in government, while the SPD remained in opposition.


*Table 9.1: Parameters of fit necessity: Government participation of radical right parties (after first third-generation elections)*

Source: Created with QCA Package for R (Duşa 2019).

When such a contradiction cannot be eliminated, despite engaging in the QCAspecific iterative process between ideas and evidence (Ragin 1989, chap. 9; see also Rihoux and Lobe 2009; Schneider and Wagemann 2012), the investigator may either include or exclude it from the analysis. Exclusion will result in lower solution coverage, whereas inclusion reduces the solution consistency, as it is also based on cases in which the outcome does not occur. The latter strategy is a viable option if the distribution of contradictory cases is uneven, for instance if the outcome occurs in nine out of ten cases (Schneider and Wagemann 2012, 122).<sup>1</sup> As this is not the case here, row 8 will be excluded from the minimisation. Instead, the two contradictory cases will be subjected to a more detailed examination at the end of the analysis.

After excluding the contradictory configuration, the highest possible consistency cut-off of 1.00 can be applied for the inclusion of truth table rows in the minimisation. The logical minimisation yields a much more complex conservative solution than in the earlier period consisting of four paths, each with four conditions (see Table 9.3). Including logical remainders in the minimisation process produces a more easily interpretable and parsimonious solution. The selection of logical remainders for further minimisation rests on counterfactual claims, or, in other words, theoretically and empirically grounded expectations about the outcome caused by the configuration in the respective truth table row. Here, only easy counterfactuals will be taken into consideration. Easy counterfactuals neither contradict the theoretical assumptions nor the empirical observations and contribute to a more parsimonious solution (Ragin and Sonnett 2005; Ragin 2008, chap. 8; Schneider and Wagemann 2012, chaps. 6 and 8; see also Chapter 4).<sup>2</sup>

<sup>1</sup> Such a distribution of cases would also be reflected in a high consistency of the truth table row despite the contradiction.

<sup>2</sup> Since the hypotheses in this study involve the interplay of multiple explanatory factors, Schneider and Wagemann's (2012, chap. 8) Enhanced Standard Analysis is preferred over the standard analysis in the fsQCA software because it allows for conjunctional directional expectations.





Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

Next, it is necessary to determine which of the 13 logical remainders in the truth table qualify as easy counterfactuals. It has been hypothesised that radical right parties predominantly enter government in the consolidating decades if they hold a large seat share and the fragmentation of party systems is low (SEATS\*∼FRAG).When the fragmentation is low and radical right parties are small (∼SEATS\*∼FRAG), they should remain in opposition. The other two combinations of these numerical conditions—small and large radical right parties in fragmented party systems (∼SEATS\*FRAG and SEATS\*FRAG)—could also lead to government participation of radical right parties. However, they are not expected to be characteristic of this period, due to the decreasing average fragmentation of Central and Eastern European party systems during the three post-Communist decades (see Chapter 4, esp. Table 4.2).

The empirically observed cases of government participation (truth table rows 1 – 8) show that all four combinations of the two numerical conditions are present when radical right parties enter government. The truth table reveals that SEATS\*∼FRAG is not the predominant configuration in this period. It occurs only in one of the 15 cases of government participation (row 6). Rather, the predominant configuration, observed in nine cases (rows 1, 2, 7, and 8), is SEATS\*FRAG.<sup>3</sup> The configuration ∼SEATS\*FRAG, found in most cases of government participation in the transformational decade, is also present in four cases (rows 3 and 4). Thus, logical remainders that include any of these three configurations can be considered easy counterfactuals. Ataka's participation in the 2009 Bulgarian government (row 5) shows that radical right parties can even enter government under the theoretically unfavourable conditions of a low seat sharein a party system with low fragmentation (∼SEATS\*∼FRAG). However, a single outlier is not enough to discard the respective hypothesis completely and consider logical remainders with this configuration easy counterfactuals.

The hypotheses further suggest that radical right parties enter government in the consolidating party systems of Central and Eastern Europe if they are proximate to the formateur on the socio-cultural dimension and/or on the same side of a bipolar opposition in the party system. The data in the truth table support this hypothesis, since all rows that lead to government participation include the GALTANPROX and/or SAMESIDE conditions.The only exceptionis the contradictory configuration in row 8. Therefore, all logical remainders containing GALTANPROX and/or SAME-SIDE qualify as easy counterfactuals.

<sup>3</sup> Row 8 also includes the case of the Czech SPD, which remained in opposition.

*Table 9.3: Sufficient conditions for the government participation of radical right parties (after first third-generation elections)*



Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

These criteria rule out three of the 13 logical remainders (rows 20 – 22). Further minimisation that includes the other 10 logical remainders, however, not only involves using a large number of unobserved configurations, but it also results in an intermediate solution that is less parsimonious than the conservative one. Identifying those logical remainders that are more likely to lead to the outcome helps to reduce their number further. In light of the empirical observations, for instance, remainders that include the conjunction SEATS\*FRAG, the most frequent configuration of numerical factors when radical right parties entered government during the consolidating decades, should be most favourable. Moreover, while the presence of either GALTANPROX or SAMESIDE is theoretically and empirically sufficient for government participation, the truth table suggests that, with the exception of the contradictory configuration in row 8, radical right parties enter government only if at least two of the three favourable ideological factors (LRECONPROX, GALTAN-PROX, and SAMESIDE) occur simultaneously.

These criteria eliminate rows 23, 25 and 27, because they contain neither the configuration SEATS\*FRAG nor the simultaneous presence of two favourable ideological conditions. The remainder in row 32 is the only one that fulfils both criteria and will therefore beincludedin theminimisation to craft theintermediate solution.The six remaining truth table rows (24, 26, 28, 29, 30, and 31) meet only one of the two criteria and will therefore be subjected to a more detailed counterfactual analysis (Schneider and Wagemann 2012, chap. 8), comparing them to empirically observed cases of government participation that differ in only one condition. If the remainder's configuration is more favourable than in the observed case, it should lead to the outcome.

The remainder in row 24, for instance, is very similar to row 2, which covers the government participation of the Slovak SNS in 2006 and twice in 2016. The configurations differ only with regard to party system fragmentation. In the three Slovak cases, a large radical right party entered government in a fragmented party system (SEATS\*FRAG), whereas the remainder refers to large radical right parties in compact party systems (SEATS\*∼FRAG). While Hypothesis 1c suggests that SEATS\*∼FRAG should be the predominant configuration of the two numerical factors in this period, the configuration SEATS\*FRAG has been observed most frequently in the empirical data. Hence, the observation described in row 24 is not more favourable for government participation than the configuration in row 2. However, this remainder will still be included in the minimisation, because exactly the same configuration led to the participation of the SNS in the 1992 Slovak government.

The remainder in row 26 also differs in one condition from the configuration in row 2. While the SNS and the formateur of the 2006 and 2016 governments shared similar socio-economic positions, the remainder in row 26 contains the condition ∼LRECONPROX. Because ideological proximity should favour government participation rather than ideological distance, the remainder's configuration is not more likely to lead to the outcome than the configuration in row 2. Unlike in the previous example, there is no additional evidence that would justify the inclusion of this remainder in the minimisation.

Rows 28, 30 and 31 differ in one condition from row 1, which describes four of the Latvian NA's government coalitions. In these instances, the NA controlled a large seat share in a fragmented parliament (SEATS\*∼FRAG), while the remainder in row 28 describes large radical right parties in compact party systems (SEATS\*∼FRAG), and row 30 refers to small radical right parties in compact party systems (∼SEATS\*FRAG). Both remainders are excluded from the minimisation because these conditions are similarly, or even less, favourable for the radical right than row 1, which includes the most frequently observed configuration of the two structural-numerical factors when radical right parties entered government in the consolidating decades. Row 31 differs from row 1 in the LRECONPROX condition. When the NA entered government,it was socio-economically close to the formateur. The remainder lacks socio-economic proximity and is therefore not considered for minimisation.

Row 29 also differs only in the LRECONPROX condition from the configuration in row 6, which covers the government participation of the Bulgarian UP in 2017. The UP and GERB, the formateur of the 2017 government, were on the same side of the bipolar opposition in the Bulgarian party system and close to each other on the socio-cultural but not the socio-economic dimension. The remainder in row 29 includes socio-economic proximity between the radical right party and the formateur. Since there is little doubt that the presence of this condition should support government participation, this logical remainder will be used for crafting the intermediate solution.

Based on these considerations, further minimisation using the logical remainders in rows 24, 29 and 32 yields the intermediate solution reported in Table 9.3.This solution still contains four solution paths, but three of them now include fewer conditions than in the conservative solution, which makes the intermediate solution somewhat more parsimonious and easier to interpret.<sup>4</sup> Factoring the solution term further eases the interpretation of the intermediate solution. In light of the argument that bipolar opposition in the party system can potentially overshadow sociocultural and socio-economic proximity and that the SAMESIDE condition is present

<sup>4</sup> The parsimonious solution generated with the fsQCA software is ∼*LRECONPROX\*GALTAN-PROX + GALTANPROX\*SEATS + LRECONPROX\*FRAG\*SAMESIDE → GOVPART*. The consistency and coverage of the solution is 1.00 and 0.93, respectively.

in three of four solution paths, this condition can be factored out to read:

$$\left\{ \begin{array}{c} \text{LRECONPROX} \ast \text{GALTANPROX\*SEATS} \\ + \\ \text{SAMESIDE\* } \text{(LECCONPROX\*SERS\*FRAG + \sim LRECONROX\*GALTANPROX)} \\ + \text{GALTANPROX\*FRAG)} \end{array} \right\} \to \text{GOVPAR}$$

This factorised term, which is logically equivalent to the intermediate solution, better illustrates the conditions under which radical right parties enter government in the presence, or absence, of bipolar opposition in the party system.The branching diagram in Figure 9.1 provides a graphical illustration of this solution term, highlighting the differences and similarities of the four solution paths. The case-based interpretation of these solution paths will be structured accordingly.

*Figure 9.1: QCA solution for the government participation of radical right parties (after first thirdgeneration elections)*

Source: Own composition.

## Path 1: Ideological proximity and parliamentary strength

The first solution path (LRECONPROX\*GALTANPROX\*SEATS) is the only one that does not contain either of the two party system conditions. It covers a total of four cases: Three government coalitions with the Slovak SNS in 2006 and 2016, as well as the Latvian NA in 2016.

By 2006, the intensity of the bipolar opposition in the Slovak party system had weakened considerably and electoral support for Smer had been growing since the early 2000s. Smer started out as a third way party in the late 1990s, suggesting programmatic ties to New Labour in Britain as well as the attempt to take a middle path between the autocratic Mečiar camp and the democratic opposition (Haughton and Rybář 2008, 244). However, the party's centre-left socio-economic profile and its nationalist leaning placed it in the position that was previously occupied by Mečiar's declining HZDS.Thus, even though the opposition between competing camps in the Slovak party system was less polarised than in previous years, the general alignment of socio-economic and socio-cultural divides remained stable.

In the run-up to the 2006 parliamentary elections, Smer emphasised socioeconomic issues over socio-cultural ones. The SNS—with occasional support from Smer—ensured that nativism, mostly directed against the Hungarian and Roma minorities in the country, remained a salient, though secondary, issue in the campaign (Haughton and Rybář 2008; Mesežnikov 2008; Pytlas and Kossack 2015; Pytlas 2016). As regards socio-economic issues, the SNS' 2006 manifesto was slightly more centrist than the party's usual positions (Pirro 2016). Nevertheless, the SNS and the HZDS were the only parties with compatible centre-left socio-economic positions with whom Fico's Smer, the 2006 government formateur, could negotiate. In turn, Fico's accommodative strategy on the socio-cultural dimension facilitated cooperation with the SNS. Smer's nationalist leaning was not (yet) clearly visible in the 2006 election campaign, but it became more obvious during the subsequent period of government, for instance in the restrictive amendment to the Slovak language law and frequent references to the Hungarian minority as a threat to national identity (Pytlas 2016, chap. 5). The ideological range of the coalition as a whole also provides valuable information about the importance of ideology in this case. Out of the five coalitions formed in Slovakia between 2006 and 2020, the 2006 coalition of Smer, SNS, and HZDS is the only one that qualifies as a minimal connected winning coalition on both ideological dimensions. The other coalitions in the country were more heterogeneous, particularly on the socio-cultural dimension.

The SNS' surprisingly large seat share also helped the party to get into office. After failing to enter parliament in 2002, the SNS won 13.3 per cent of the seats, third most, in 2006. Smer controlled only one-third of the 150 seats and was thus well short of a majority. Hence, Fico required large junior partners, such as the SNS. Moreover, minority governments are rather uncommon in Slovakia and the few that formed after the Velvet Revolution resulted from defections during a legislative term (Bergman, Ilonszki, and Müller 2019a). Hence, it is unlikely that Smer seriously considered the option of forming a minority government after the 2006 election, which improved the bargaining position of a relatively large and ideologically compatible party, such as the SNS, even further.

When the SNS came to power again ten years later, the configuration of the Slovak party system had changed substantially. Socio-cultural and socio-economic divides remained relevant, and party system polarisation had grown due to Fico's controversial personality. However, the emergence of various populist anti-establishment parties that accused Smer and the SNS as well as the SDKÚ-DS of corruption, introduced a new cross-cutting divide. Thus, the conflict structure of the Slovak party system in 2016 could best be described as multi-polar (Havlík et al. 2020, 230).

Smer once more emerged victorious from the 2016 parliamentary elections and was thus in charge of forming a government. The party had incorporated the SNS' exclusionary conception of Slovak national identity, and during the course of the "migration crisis" in the mid-2010s, Fico's attacks against immigrants and refugees were even more extreme than the radical right party's (Rybář and Spáč 2016; Androvičová 2017; Baboš and Malová 2017). Now that the HZDS had completely vanished from the political arena, the SNS was the only party in parliament with socioeconomically and socio-culturally proximate positions to the weakened formateur. Therefore, it was no surprise the SNS was Fico's first choice as a coalition partner in 2016.

The SNS again received a respectable electoral result, yielding ten per cent of the seats in a fragmented eight-party parliament. However, the combined seat share of the two parties was still far away from reaching a majority. Due to the absence of other ideologically compatible coalition partners, however, Smer struggled to find additional junior partners. The eventual formation of an ideologically heterogeneous, four-party majority government with the Hungarian minority party, Most-Híd, and the newly founded conservative party, Sieť, was aided by two factors: First, many parties agreed that Slovakia's upcoming role as president of the Council of the European Union should not be managed by a caretaker government (Baboš and Malová 2017) and, second, the country's political culture included a general scepticism towards minority governments. The coalition that was ultimately cobbled together was exceptional in that it included both the anti-Hungarian SNS and Most-Híd, the representative of the Hungarian minority in Slovakia. However, the hostility between these parties lost some momentum in the mid-2010s after the SNS changed its leadership and the politicisation of the immigration issue became more intense (Baboš and Malová 2017; Harris 2019; Rybář 2020). In addition, the Hungarian minority was now represented by Most-Híd, not by the SMK, which had been the SNS' main opponent during the late 1990s when the polarisation in the Slovak party system peaked.

After a few months, however, Sieť's parliamentary group dissolved, which resulted in a change of the partisan composition of the coalition and, thus, a new government according to the counting rules applied in this study. The new government still held a majority because some Sieť deputies joined Most-Híd's parliamentary group (Baboš and Malová 2017). This re-formation of the coalition, however, underlines the importance of the SNS' relatively large seat share. Since not all Sieť MPs continued to support the government coalition, SNS' seats proved vital for retaining the majority status of the coalition.

While these three Slovak coalitions are uniquely covered by the first solution path, the government participation of the Latvian NA in 2016 is covered by the first (LRECONPROX\*GALTANPROX\*SEATS), second (LRECON-PROX\*SEATS\*FRAG\*SAMESIDE), and fourth path (GALTANPROX\*FRAG\*SAME-SIDE). Similar to the SNS, the NA also remained in power following a mid-term re-formation of the government coalition in 2016. Here, however, the partisan composition of the coalition remained the same but the prime minister's party changed from the liberal Unity to the Union of Greens and Farmers (ZZS) led by the oligarch, Aivars Lembergs. This case diverges from the other instances of the NA's government participation because the radical right was not confronted with Unity as a formateur but with the ZZS. Unlike Unity, the ZZS was not only on the same side of the bipolar opposition in the Latvian party system, which mainly originated from the ethno-linguistic divide, but the parties shared similar positions on other socio-cultural issues beyond that divide as well (Galbreath and Auers 2010).

After a failed attempt by an internal rival to replace Unity's incumbent prime minister, Straujuma, in 2016, the ZZS became the formateur. The party took advantage of this intra-Unity power struggle by nominating its own candidate, who was eventually elected prime minister by the three coalition parties (Auers 2016). Regardless of these machinations, the NA's coalition membership was never in doubt. Hence, the additional proximity to the formateur on socio-cultural issues other than the ethno-linguistic divide, facilitated cooperation with the ZZS, but it is unlikely that the radical right would have left the ruling coalition if a new, socio-culturally incongruent Unity prime minister had taken over. The large seat share, however, helped the NA to remain in power in 2016. It had become one of the strongest and most stable members of the radical right party family in Central and Eastern Europe over the last decade. When the government re-formed in 2016, the party controlled an impressive 17 per cent of the seats in parliament. It could thus contribute much more to the parliamentary majority than the two remaining parties in the ethnic Latvian camp, which were substantially smaller than the NA.

In sum, all four instances of government participation covered by the first solution path follow a similar pattern. Regardless of the configuration of the party system, radical right parties entered government because they were socio-economically (LRECONPROX) and socio-culturally proximate to the formateur (GALTANPROX), and because they controlled a large number of seats in parliament (SEATS). However, the explanation for the NA's government participation diverges slightly from the three Slovak cases. The three party-level factors also facilitated the NA's participation in the 2016 government. Unlike in Slovakia, however, this government was formed in the context of a deep bipolar opposition in the party system that was based on the ethno-linguistic divide. Here, other socio-cultural issues beyond this divide, as reflected in the GALTAN dimension, were less important. Therefore, the sociocultural proximity to the formateur helped the NA, but it was less decisive for the party's inclusion in government than for the SNS. These observations, and the fact that this case is also covered by two paths that include the SAMESIDE condition, imply that bipolar opposition was more important in the Latvian party system than this solution path suggests.

## Paths 2 – 4: The impact of a bipolar oppositions in the party system

Solution paths 2 – 4 differ from the first one in that they describe situations in which the radical right party and the formateur are situated on the same side of a bipolar opposition in the party system (SAMESIDE). All three paths, however, include additional conditions that are needed to explain the government participation of radical right parties.

The second solution path (LRECONPROX\*SEATS\*FRAG\*SAMESIDE) is the only one in the intermediate solution that does not include socio-cultural proximity (GALTANPROX). It covers all cases of government participation of the Latvian NA, including the previously discussed ZZS-led coalition. A deep bipolar opposition rooted in the ethno-linguistic divide between the Latvian majority and the large Russian-speaking minority has been an essential feature of the Latvian party system since the country's independence from the Soviet Union. This opposition was reinforced by a salient socio-economic divide (Saarts 2011; Auers 2013). In this context, Harmony Centre (SC) not only represented the Russian-speaking minority in the Latvian party system, but also the left socio-economic pole. All ethnic Latvian parties, in contrast, held either centrist or liberal socio-economic positions. Although SC emerged victorious from most parliamentary elections in the 2010s, the lack of junior partners in the Russian-speaking camp always prevented the party from forming a government.

Since 2011, the radical right NA has participated in five government coalitions, all including only ethnic Latvian parties. Thus, being part of this camp was clearly key to the NA's government participation. With the exception of the 2016 coalition, the socio-economically and socio-culturally liberal Unity was tasked with the formation of government. Regarding other socio-cultural issues, such as gender equality or immigration, however, Unity held decidedly more liberal views than the radical right (Auers 2016, 3), which is reflected by a distance of around four points between these two parties on the GALTAN dimension.The NA differed from most Central and Eastern European radical right parties in its centre-right socio-economic positions. The party advocated socio-economic policies that were not so different from the liberal ones that Unity and the other parties in the ethnic Latvian camp supported (see Chapter 6.2).This solution path and the ideological range of Latvian governments in the 2010s suggest that the parties in the ethnic Latvian camp paid attention to the socio-economic proximity of their partners, whereas they were willing to overlook different positions on the GALTAN dimension. All four governments since 2011 were quite homogeneous along the socio-economic dimension. Even the five-party coalition of 2018 features a relatively low socio-economic range of only 2.18 points (see Table 6.6). The socio-cultural range of these coalitions, in contrast, has always been quite broad, ranging from 4.50 to 6.18 points.

This solution path further illustrates that the NA's parliamentary strength also contributed to its inclusion in government. The configuration highlights that the large seat share (SEATS) proved to be an asset for the party in Latvia's fragmented party system (FRAG). Since the whole ethnic Latvian camp shared similar socio-economic positions and placed less importance on compatible socio-cultural ones (outside the ethno-linguistic divide), government formation could almost be reduced to calculating each party's potential contribution to a parliamentary majority. Due to the high fragmentation, all Latvian governments between 2011 and 2018 included at least three parties. With seat shares ranging from 13 to 17 per cent, the NA was among the larger parties in the ethnic Latvian camp, making it harder to form stable majorities without it. Thus, while the previous analysis shows that a small seat share is not necessarily a disadvantage, the Latvian cases demonstrate that a large seat share in a fragmented party system can create an actual advantage for radical right parties.

These observations demonstrate that the interplay of two ideological factors, being on the same side of bipolar opposition in the party system (SAMESIDE) and being proximate to the formateur on the socio-economic dimension (LRECONPROX), as well as two structural-numerical factors, having a large seat share (SEATS) in a fragmented party system (FRAG), explain NA's government participation during the 2010s. This result confirms that the party's socio-cultural proximity to the ZZS in 2016 was helpful, but less important than the socio-economic proximity to the formateur and its large seat share. Hence, there is a clear, country-specific pattern that explains the government participation of the radical right NA in Latvia. In the Latvian party system, conflicts over national identity, and thus the core issue of the radical right, manifested in a deep bipolar opposition in the party system rather than in the broader GALTAN dimension.

In the third solution path (∼LRECONPROX\*GALTANPROX\*SAMESIDE), socio-cultural proximity (GALTANPROX) and socio-economic distance (∼LRECON-PROX) between the radical right party and the formateur accompany the SAME-SIDE condition. This path covers the 2009 (Ataka), 2014 (PF), 2016 (PF), and 2017 (UP) Bulgarian government coalitions. In the late 2000s, contestation between two camps had emerged in the Bulgarian party system. The bipolar opposition between the Communist successor party, BSP, joined by the unofficial party of the Turkish minority, DPS<sup>5</sup> , on one side, and the conservative camp, led by GERB, on the other, ran so deep that the BSP and GERB sometimes boycotted parliamentary work altogether, rather than assuming their role as a constructive democratic opposition

<sup>5</sup> The formation of ethnic parties is prohibited by law in Bulgaria.

(Karasimeonov 2019). The socio-economic dimension was home to the most salient divide in the Bulgarian party system during this period (Rohrschneider and Whitefield 2009, 290)<sup>6</sup> , but the antagonism between BSP and GERB rested on affective rather than ideological polarisation.

Ataka's support for the GERB minority government in 2009 represents the first time that a radical right party participated in government in Bulgaria. Situated clearly on the left side of the socio-economic dimension, Ataka is a prime example of radical right parties in Central and Eastern Europe supporting social-national economic policies (Avramov 2015, 308; Pirro 2016, 62–63; Popova 2016, 262–63). Hence, the party was quite far away from the socio-economically liberal GERB. On socio-cultural issues, however, the distance between these two parties was smaller. Although GERB did not incorporate radical right narratives into its own platform as extensively as other mainstream competitors in the region, Boyko Borisov and his party still approached some of Ataka's nativist positions and employed a "moderate nationalism" (Krasteva 2016, 176; see also Avramov 2015; Pirro 2016). In return for the formal support of GERB's single-party majority government, Ataka expected concessions regarding its socio-cultural core issues. Hence, socio-cultural proximity (GALTANPROX) was likely more important to Ataka than it was to GERB.

The composition of the 2009 Bulgarian government provides additional insights into the role played by ideological factors. GERB controlled 117 of the 240 seats in parliament and thus missed out on the majority by only a narrow margin. Even though any one of the smaller parties in the conservative camp could have provided GERB with a majority, Borisov struck a deal with all four of them, including the radical right Ataka.The government resembled an "oversized minority government" because all four parties agreed to support a single-party GERB cabinet. Since none of the support parties were simultaneously socio-economically and socio-culturally proximate to GERB, Borisov might have opted to include all of them in order to ensure that he would always have a working majority, whether the parliament was voting on socio-economic or socio-cultural policies.

In the 2014 Bulgarian parliamentary elections, another radical right party, the PF, was elected to parliament, and it was subsequently included in the government. GERB won the election and Borisov was again tasked with forming a government. The situation was different than in 2009, however, because GERB controlled only

<sup>6</sup> Since the 2014 wave, the CHES provides data on the salience of the socio-economic and sociocultural dimensions for the individual parties in addition to their positions (Jolly et al. 2022). The salience ranges from 0 (low) to 10 (high). The salience of each dimension in the party system can be obtained by weighting the salience of the respective dimension by the parties' electoral strength. In Bulgaria, the salience of the socio-economic dimension is at 6.7, while that of the socio-cultural one is only at 4.9 (see Appendix II).

one-third of the seats in parliament.This time, Borisov formed a three-party minority coalition, which was officially supported by the PF. Unlike in 2009, the support of the radical right was vital for maintaining the government's majority in parliament, as it was the coalition's only support party. In 2016, one of the junior partners left the coalition, which led to a change in its partisan composition, and thus, the formation of a new government. Despite continuing support from the radical right PF, the remaining coalition parties no longer controlled a parliamentary majority. In this situation, Borisov considered including the PF as junior partner in the minority coalition. Later that year, and before this plan had been put into practice, however, Borisov and his government resigned after the GERB candidate was defeated in the presidential elections (Karasimeonov 2014b, 2–3; Kostadinova and Popova 2015; Kolarova and Spirova 2017).

The PF held slightly more moderate positions than Ataka, resulting in a smaller socio-economic and socio-cultural distance to GERB. However, ideological proximity between GERB and the PF existed only on the socio-cultural dimension (GAL-TANPROX). In the socio-economic sphere, the PF was situated left of centre and thus qualitatively and quantitatively rather distant from GERB's liberal positions (∼LRECONPROX). The PF's participation in the 2014 and 2016 government was assisted by the bipolar opposition in the party system which limited GERB's choice of potential partners to parties within its own camp (SAMESIDE). The ideological heterogeneity of the 2014 and 2016 governments shows that GERB was primarily concerned with forming a coalition with parties in its own camp. For that purpose, Borisov accepted support from parties with a wide array of positions on the socioeconomic and socio-cultural dimensions (see Table 6.11). Similar to the cooperation with Ataka in 2009, GERB included the radical right PF despite socio-economic differences, while the party's moderate nationalism provided enough of an incentive for the PF to support the GERB-led coalition.

After the 2017 elections, the re-elected GERB formed a coalition with the radical right UP, an alliance of Ataka and the PF.While Ataka and the PF had already participated in government as support parties, this coalition was the first to provide their leaders with cabinet posts. As the UP consists of the aforementioned radical right parties, their socio-cultural and socio-economic proximity to the formateur is similar to the aforementioned cases. Hence, the bipolar opposition is the primary constraint on GERB's potential coalition partners. Within this camp, GERB favoured a coalition with the UP, which from the radical right's point of view, was again helped by GERB's nationalist-leaning position on the socio-cultural dimension. In 2017, the UP was also in a stronger bargaining position than its predecessors in 2009, 2014, and 2016. Apart from the oppositional parties, BSP and DPS, GERB, the UP and the new national-conservative party, Volya, were the only parties that passed the threshold of representation in the 2017 parliamentary elections. Volya was ideologically more proximate to GERB than the UP, but unlike the radical right alliance, the party

did not control enough seats to form a majority government with GERB. Hence, the UP was ultimately the only party in the conservative camp that could provide GERB with a majority in parliament.

In sum, the third solution path describes a clear pattern that explains the government participation of Bulgarian radical right parties. Here, the bipolar opposition in the party system, and the socio-cultural proximity between the formateur and the different radical right formations, were more important than incongruence on the socio-economic dimension. The absence of socio-economic proximity between radical right parties and the formateur (∼LRECONPROX) suggests that the affective polarisation between the two oppositional camps overshadowed ideological distance on this most salient dimension in Bulgarian politics. The ideological heterogeneity of the four Borisov governments shows that GERB was willing to cooperate with junior partners in his camp, regardless of their ideological positions. Socio-cultural proximity still played a role in the formation of these governments, however. The radical right parties' decision to support several GERB minority governments cannot be explained without their proximity to the formateur on the GAL-TAN dimension and, thus, the prospect of implementing some of their preferred nativist policies. Given that they did not gain access to cabinet posts, it is hard to imagine that Ataka and the PF would have supported Borisov had such ideological incentives been absent.

The fourth solution path (GALTANPROX\*FRAG\*SAMESIDE) covers two of the cases already discussed in the previous path, the Bulgarian PF in 2014 and 2016. Therefore, the unique coverage of the third and fourth path, respectively,is relatively low. This observation, and the fact that both paths differ only in one condition, also suggests that they describe different varieties of a similar explanatory pattern. This solution path adds that the PF entered government in a highly fragmented party system (FRAG). In this context, GERB needed three other parties to control a parliamentary majority.Thus, the fourth solution path points out that high fragmentation necessitated that several small parties participate in government, while bipolar opposition limited the number of coalition partners available to the formateur. Since Borisov depended on the support of several parties from the conservative camp, the PF's small seat share proved not to be a decisive disadvantage. Because the bipolar opposition in the party system and socio-cultural proximity to the formateur are consistent characteristics in all instances of government participation of the radical right in Bulgaria, these two paths can be regarded as different varieties of a similar pattern explaining government participation of radical right parties in Bulgaria.

The NA's participation in the 2016 Latvian government has been discussed in the context of the first two solution paths. This solution path corroborates that bipolar opposition helped the NA to enter the coalition and their large seat share served as an advantage in the context of high party system fragmentation. Contrary to the above observation, however, this path does not include socio-economic proximity

between the radical right party and the formateur, which suggests that this condition is causally irrelevant. Instead, it includes socio-cultural proximity, which was found to be less decisive. This solution path provides a logically correct description of the configuration of the individual explanatory factors in the case of the Latvian NA in 2016. All three conditions support NA's participation in government. However, the previous discussion shows that socio-economic proximity (LRECONPROX) was also important for NA to enter government. Thus, analytically, the second solution path, which includes the LRECONPROX condition, provides a more adequate explanation of NA's government participation in 2016 than this one. Because the LRE-CONPROX condition is absent from the fourth solution path, and does not appear in its negation, this path does not fundamentally contradict this conclusion.

The only two cases that are uniquely covered by the fourth solution path refer to the government participation of the Polish LPR in 2005 and 2006. In the run-up to the 2005 Polish parliamentary elections, everything was set for a coalition including the two main parties, PO and PiS, which had emerged from the post-Solidarność camp. However, due to the fierce competition between their candidates during the presidential elections, scheduled only two weeks after the parliamentary elections, the rift between these parties deepened so quickly and dramatically that the envisioned coalition was no longer possible. Even though this bipolar opposition became increasingly charged with socio-economic and socio-cultural conflicts, polarisation was primarily affective in this initial phase (Szczerbiak 2007; Pytlas and Kossack 2015; Pytlas 2016).

This bipolar opposition prevented cooperation between PiS and PO and, thus, paved the way for the LPR,which had sided with PiS(SAMESIDE), to join the governmentin 2005. LPR'sinclusion was further helped by the party'sideological proximity to PiS, particularly but not exclusively on the socio-cultural dimension (GALTAN-PROX), which was most evident in their shared support for the idea of a Fourth Republic. "The Fourth Republic would experience moral cleansing through deep lustration, anti-corruption measures, and reaffirmation of Catholic values; its new Constitution would repair the state; it would heal society with a social contract, including fundamental changes in social and economic policy" (Millard 2010, 127). Hence, once the PO-PiS coalition fell apart, the LPR became the most obvious partner for the PiS from an ideological perspective.

The Polish party system was quite fragmented after the 2005 election (FRAG) and PiS controlled only one-third of the seats in the Sejm, so like GERB in the Bulgarian election of 2014, it also depended on support from several parties. After the only possible two-party majority government had been ruled out, PiS required at least two of the four remaining parliamentary parties to reach a majority. Of these four, PiS chose the two with the greatest ideological proximity on the GALTAN dimension—LPR and the populist Samoobrona (SO).

When PiS, LPR, and SO decided to cooperate after the 2005 elections, the parties did not form a coalition government—despite substantial agreement in most policy areas. Instead, the two smaller parties supported a PiS minority government. Sections of the PiS electorate held critical views towards the radical right, and these voters may have motivated this decision. In 2006, however, Kaczyński decided to formalise the cooperation in an attempt to stabilise the conflict-ridden government (Millard 2010, 144). This gambit failed spectacularly, though, and ongoing conflicts between the coalition partners ultimately resulted in the early termination of the government in 2007. The change in the status, as the two small parties shifted from being support parties of a minority government to junior partners in a majority coalition, marks a new government according to the counting rules applied in this study. Yet, the explanation of LPR's participation in the 2005 government also applies to its re-formed version in 2006.

The analysis of this solution path shows that, with the exception of the Latvian case, paths three and four represent different varieties of a similar underlying pattern. All instances of government participation of radical right parties in Bulgaria and Poland were aided by the constraining effect of a bipolar opposition in the party system based on affective polarisation as well as the socio-cultural proximity between the radical right and the formateur. This pattern varies slightly in two ways. First, while the radical right in Bulgaria entered government despite socioeconomic distance to the formateur, the LPR and PiS held very similar positions on this dimension. Second, some variation exists regarding the fragmentation of party systems. While Ataka in 2009 and the UP in 2017 entered government in relatively compact party systems, when the formation of governments with few parties was possible, the PF and the LPR were further assisted by the fact that the formateurs needed more than one junior partner, or support party, from within their camp to secure a majority. Interestingly, and contrary to Hypothesis 1a, Ataka was even included in the 2009 government despite its small seat share in a compact party system.

#### Summary

The case-based analysis determines that there are different patterns for explaining the government participation of Central and Eastern European radical right parties in the period after the first third-generation elections. These explanatory patterns differ depending on the presence, or absence, of bipolar opposition in the party system. If no bipolar opposition existed, all three party-level conditions—socioeconomic proximity, socio-cultural proximity, and a large seat share—needed to be present for radical right parties to enter government.

In 12 of the 15 cases of a government participation of radical right parties in this period, however, party systems were characterised by bipolar opposition. Yet, the fact that radical right parties and the formateur were in the same camp alone was not sufficient for these parties to enter government.The explanatory patterns within this group of cases differ according to the nature of the bipolar opposition. In Bulgaria and Poland, the bipolar opposition that existed when the radical right entered government resulted primarily from affective polarisation between the largest parties in the country and not so much from policy-oriented divides. In this environment, the radical right party and the formateur needed to share similar positions on the socio-cultural dimension and, thus, on the dimension that concerns the ideological core of the radical right. In Latvia, however, the bipolar opposition itself was deeply rooted in ideological polarisation based on the ethno-cultural divide. Here, ideological proximity on the broader socio-cultural dimension played a subordinate role. Instead, radical right parties needed to be part of the ethnic Latvian camp, be socio-economically close to the formateur, and control a large seat share in parliament in order to enter government.

## **9.2 Explaining the exclusion of radical right parties from government**

### 9.2.1 Analysis of necessity

After analysing the government participation of Central and Eastern European radical right parties in the period after the first third-generation elections, this section turns to the conditions under which these parties remained in opposition. The first step is to test for necessary conditions. Table 9.4 shows that none of the conditions or their negations meets the criteria for necessity. With a consistency score of 0.84, the absence of socio-cultural proximity (∼GALTANPROX) comes closest to the required minimum of 0.9, but all of the other factors fall well below this threshold. Hence, there is no necessary condition for the exclusion of radical right parties from government in this period.


*Table 9.4: Parameters of fit necessity: Exclusion of radical right parties from government (after first third-generation elections)*


Source: Created with QCA Package for R (Duşa 2019).

## 9.2.2 Analysis of sufficiency

In the absence of necessary conditions, the investigation proceeds with the analysis of sufficient conditions for the negative outcome (∼GOVPART). Rows 9 – 19 in the truth table above (see Table 9.2) show a consistency of 1.00 for the negative outcome, so they will be included in the minimisation. The contradictory configuration in row 8 is again excluded. The resulting conservative solution generated by the minimisation comprises five solution paths that consist of three to four conditions each (see Table 9.5). Therefore, further steps will be taken to arrive at an intermediate solution that is easier to interpret.The analysis of radical right parties' participation in government during this period highlights the relevance of radical right parties and the formateurs being on the same side of a bipolar opposition in the party system. Four of the five paths in the conservative solution include the condition SAMESIDE, or its negation, indicating that this condition plays an important role in the exclusion of radical right parties from government as well. Therefore, the intermediate solution here is not crafted by further minimisation with the help of logical remainders but by undoing the minimisation step in the fourth solution path (∼LRECONPROX\*∼GALTANPROX\*∼SEATS\*FRAG), the only path that does not include the SAMESIDE condition.<sup>7</sup> The fourth solution path is based on the minimisation of truth table rows 11 (∼LRECONPROX\*∼GALTANPROX\*∼SEATS\*FRAG\*∼ SAMESIDE) and 14 (∼LRECONPROX\*∼GALTANPROX\*∼SEATS\*FRAG\*SAME-SIDE). Undoing the minimisation of these two rows adds an additional solution path to the conservative solution, making it even more complex.

<sup>7</sup> This solution is intermediate because it results from an intermediate step in the minimisation procedure. It should not be confused, however, with the intermediate solution in the standard analysis in the fsQCA software, which involves logical remainders (see Chapter 5). The parsimonious solution generated by the fsQCA software is ∼*SEATS\**∼*SAMESIDE +* ∼*LRECONPROX\**∼*GALTANPROX + LRECONPROX\**∼*FRAG →* ∼*GOVPART*. The consistency and coverage of the solution is 1.00 and 0.95, respectively.




Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

One of these new paths covers only the Czech Úsvit in 2013 and the Estonian EKRE in 2015. These cases are also covered by the second solution path (∼LRECONPROX\*∼GALTANPROX\*∼SAMESIDE), indicating that the unique coverage of this new path is zero. Therefore, this redundant path can be dropped, because the solution contains the same logical information without it. This results in the intermediate solution found in Table 9.5. It still contains five solution paths and is thus similarly complex as the conservative solution. With regard to bipolar opposition in the party system, however, the intermediate solution is easier to interpret than the conservative one, which is also illustrated in the branching diagram in Figure 9.2.

The diagram shows that the solution includes two paths for radical right parties that remained in opposition when they were on the same side of a bipolar opposition as the formateur and three paths when they were not. In each of these paths, of course, the SAMESIDE condition or its negation is accompanied by further explanatory conditions. Moreover, the consistency (1.00) and coverage scores (0.95) of the intermediate solution indicate that it is a perfect subset of the negative outcome and covers most of the cases in which radical right parties remained in opposition during the consolidating decades. The remainder of this section first investigates the first three solution paths, which cover the 14 cases in which radical right parties were not in the same camp as the formateur (∼SAMESIDE) when they remained in opposition. It then turns to the two solution paths, and four cases, where the radical right was excluded from government despite being in the same camp as the formateur.

#### Paths 1 – 3: Not in the same camp and not in government

The first three paths of the intermediate solution refer to situations in which radical right parties are not in the favourable position of being on the same side of a bipolar opposition as the formateur. The first two paths are somewhat similar; both include ∼SAMESIDE, ∼GALTANPROX and one additional condition, ∼FRAG and ∼LRECONPROX respectively. Moreover, these two paths cover three of the same cases—the Romanian PRM in 2004 and 2007, and the Slovak SNS in 2010. The third solution path, instead, differs significantly from the previous two and uniquely covers four cases.

Before analysing the first three solution paths in more detail, it is important to remember that ∼SAMESIDE can describe two different situations, that is, the absence of a bipolar opposition in the party system on the one hand, and the position of radical right party and formateur on opposite sides of a bipolar opposition on the other (see Chapter 8). In 12 of the 14 cases covered by the first three solution paths, there was no bipolar opposition in the party system.Only the Bulgarian Ataka and the Slovak SNS in 2013 and 2010, respectively, were not in the formateur's camp while bipolar opposition existed in the party system.

*Figure 9.2: QCA solution for the exclusion of radical right parties from government (after first thirdgeneration elections)*

Source: Own compilation.

The first solution path (∼GALTANPROX\*∼FRAG\*∼SAMESIDE) covers seven cases, including the two ones in which ∼SAMESIDE indicates that the radical right party and the formateur are in opposite camps. In the run-up to the 2013 Bulgarian parliamentary elections, Ataka's support was in decline because it had been supporting the GERB minority government since 2009. Participation in government alienated the party's hardcore supporters in particular (Avramov 2015). As a result, the party tried to distance itself from GERB, which marginally affected the bipolar opposition in the party system. The 2013 parliamentary elections resulted in a hung parliament that consisted of only four parties. The 240 seats were split 120–120 between GERB and Ataka, on one hand, and BSP and DPS, on the other. While supporting a GERB-led government was not an option for BSP and DPS, the stalemate between the two camps saved Ataka from completely dissociating from its former ally, since the two parties were one seat short of a majority. Ultimately, a difficult formation process resulted in a minority coalition being formed by BSP and DPS, thus placing Ataka in the opposite camp of the bipolar opposition (∼SAMESIDE). However, this government assumed office because Ataka's leader, Volen Siderov, was the only member of the opposition to remain in parliament during the investiture vote. In order to prevent the government from reaching the necessary quorum, all of the other GERB and Ataka MPs had left the legislature. Ataka supported the government in other parliamentary votes, but not on the basis of a formal agreement (Kostadinova and Popova 2014, 2015; see also Karasimeonov 2013a, 2013b; Avramov 2015). Apparently, BSP and Ataka were not ready for closer

cooperation, which would have meant overcoming the bipolar opposition in the Bulgarian party system.

The other two conditions in the first solution path also contributed to Ataka's exclusion from government. In the divided Bulgarian party system of 2013, the low number of parliamentary parties (∼FRAG) severely limited the available parties for coalitions within each camp. None of the four parties was ready to cross the gulf between the camps and formally cooperate with the other camp. Ataka went furthest and enabled the formation of a government. Ultimately, however, the bipolar opposition remained intact, which was aided further by the lack of socio-cultural proximity between Ataka and the BSP (∼GALTANPROX). Thus, joining a BSP-DPS government did not promise Ataka substantial policy concessions regarding the party's socio-cultural core issues.

The 2010 parliamentary elections in Slovakia were won by Smer, the prime minister's party from the incumbent coalition which had also included the radical right SNS and Mečiar's HZDS. Under the rule of this coalition from 2006 to 2010, a strong bipolar opposition had resurfaced in the Slovak party system, similar to the one from the late 1990s. This opposition was primarily based on ideological polarisation along reinforcing socio-economic and socio-cultural divides (Haughton, Novotná, and Deegan-Krause 2011; Deegan-Krause and Haughton 2012; see also Chapter 5.4). Even though the SDKÚ-DS came in second in the 2010 elections, the party became the formateur of the government that was to assume office in the same year. Here, the policy-based bipolar opposition(∼SAMESIDE) that coincided with the absence of socio-cultural (∼GALTANPROX) (and socio-economic) proximity between SDKÚ-DS and SNS ruled out cooperation between these two parties. Moreover, the 2010 elections saw only 4.0 effective parties enter parliament (∼FRAG), which also limited the number of potential coalition partners. In particular, the HZDS failed to cross the five per cent threshold for the first time since 1992, leaving the two remaining parties in the national-protectionist camp, Smer and SNS, short of a majority.

In the remaining five cases covered by this solution path (the Polish LPR in 2001 and 2003, and the Romanian PRM in 2000, 2004, and 2007), ∼SAMESIDE denotes the absence of bipolar opposition in the party system. Prior to LPR's electoral breakthrough in 2001, the Polish party system had been structured by the regime divide. After the 2001 parliamentary elections, however, this divide had receded to such an extent that coalitions between the two camps became possible for the first time (∼SAMESIDE) (Millard 2010, 113–14). The 2001 Polish parliament consisted of only 3.6 effective parties, which was mainly due to the strong position of the SLD (∼FRAG).The party held 43.5 per cent of the seats and was so close to a majority that even the seats of the LPR—the smallest parliamentary group—would have sufficed. Despite the relatively similar socio-economic positions of the two parties, such a coalition was never an option because on the GALTAN dimension, the parties were separated by almost eight points, the largest distance between a radical right party

and a formateur in all of the cases covered by this study (∼GALTANPROX). This huge socio-cultural distance between the SLD and the LPR, together with the strong position of the SLD in a relatively compact parliament, prevented the radical right from being included in government in either 2001 or 2003.

Electorally, the Romanian PRM was in a much better position than the LPR in the 2000s. The PDSR won the 2000 general elections in Romania, but the PRM reached its all-time high and secured almost a quarter of the seats in parliament. Hence, the former allies could have formed a two-party majority government quite comfortably. Yet, compared to the early 1990s when both parties governed together, the PDSR had moderated its nationalist positions and accepted Romania's transformation into a market economy in order not to jeopardise the country's EU accession (Pop-Eleches 2008, 470; Vachudova 2008, 871; see also Ștefan 2019).These positional shifts increased the socio-cultural distance between the PRM and the PDSR (∼GAL-TANPROX), although both parties were still located in the nationalist end of the GALTAN spectrum. Furthermore, the PDSR joined the cordon sanitaire intended to separate the radical right party from participating in government (Cinpoeş 2015, 288). This decision was certainly helped by the erosion of the bipolar opposition, which no longer limited the parties' potential partners to the radical right, as it had in the 1990s (∼SAMESIDE). In fact, the Hungarian minority party, UDMR, and the liberal PNL ultimately supported a PDSR-led minority government. Moreover, the 2000 Romanian parliament comprised only 3.6 effective parties (∼FRAG) and the electoral alliance led by the PDSR held so many seats that any one of the four remaining parties could have provided the government with a parliamentary majority.

In the 2004 elections, the PRM suffered heavy losses but remained relatively strong. Bipolar opposition was still absent from the Romanian party system (∼SAMESIDE), and the fragmentation fell to 3.4 effective, and four actual, parliamentary parties—or electoral alliances (∼FRAG). However, the bargaining situation was more complex than these numbers suggest because the electoral alliance of PDSR and PUR formed separate groups in parliament. Therefore, the number of possible minimal winning coalitions was higher than it would have been with only four parliamentary groups, and the parties ultimately formed a four-party minority coalition, a rather uncommon format in compact party systems. In 2007, two parties dropped out of government, leaving a two-party minority coalition in office following this mid-term re-formation. In both 2004 and 2007, the National Liberal Party (PNL) served as the formateur. The PNL had distanced itself from the PRM in the 1990s, when it was part of the oppositional alliance, CDR, and the party had no intention of changing this course one decade later. In addition, the ideological distance between the PNL and the PRM regarding socio-cultural (and socio-economic) issues was much more pronounced than between PDSR and PRM (∼GALTANPROX). Hence, the PRM's exclusion from government in the 2000s is mainly a result of the party's the socio-cultural distance to the formateur that

culminated in the cordon sanitaire, which, in turn, had been enabled by the erosion of the regime divide and its constraining effect on government formation.

The second solution path (∼LRECON\*∼GALTANPROX\*∼SAMESIDE) involves only ideological factors. As for the cases that are also covered by the first path, the PRM in 2004 and 2007 and the SNS in 2010, this configuration underlines that the distances on bothideological dimensionsinhibited the radical right from joining the government. The Slovak SNS in 2010 lacked ideological proximity to the SDKÚ-DS on the socio-economic (∼LRECONPROX) and socio-cultural dimensions (∼GAL-TANPROX). In a party system characterised by bipolar opposition based on the congruent alignment of these two conflict dimensions, this path highlights that the SDKÚ-DS would not have cooperated with the ideologically distant radical right party from the opposite camp (∼SAMESIDE) regardless of either the party system's level of fragmentation or the party's parliamentary strength.The situationin the two Romanian cases was somewhat similar: Ideological proximity between the PRM and the PNL existed neither on the socio-economic (∼LRECONPROX) nor on the sociocultural dimension (∼GALTANPROX), but in the absence of a bipolar opposition in the party system (∼SAMESIDE), the ideological distance was further reinforced by a cordon sanitaire vis-à-vis the PRM.

In the remaining three cases uniquely covered by this solution path, the mainstream political parties had established a cordon sanitaire against the radical right. In the Czech Republic, the cordon sanitaire had existed since the fall of Communism and, thus, already contributed to the exclusion of the SPR-RSČ from government in the 1990s (see Chapter 8). The same happened to Tamio Okamura's new radical right party, Úsvit, in 2013. Due to multi-polar oppositions in the party system, the formation of the 2013 government was not limited to coalitions within two competing camps (∼SAMESIDE). The social democratic ČSSD won the election and subsequently led the process of government formation. In contrast to the radical right Úsvit, the social democrats positioned themselves clearly on the left in terms of socio-economic policies (∼LRECONPROX) and on the GAL end of the socio-cultural dimension (∼GALTANPROX) (Havlík 2014, 46; Koubek and Polášek 2017, 16). Thus, in case of the ČSSD, the cordon sanitaire is the logical consequence of this ideological distance, particularly on the socio-cultural dimension. Moreover, Úsvit stayed true to its anti-establishment appeal and ruled out participation in government itself (Hloušek, Kopeček, and Vodová 2020, 132).

After its electoral breakthrough in the 2015 parliamentary elections, the Estonian radical right party EKRE faced the neoliberal ER as the formateur. After the turn of the millennium, the ethno-linguistic divide in the country had become less polarised (∼SAMESIDE) (Lagerspetz and Vogt 2013). Although EKRE and the ER were mostly elected by the Estonian majority in the country, the parties' positions regarding the ethno-linguistic divide and other socio-cultural issues, such as immigration or gender equality, diverged significantly (∼GALTANPROX). On the socioeconomic dimension, the ER's neoliberal economic programme was not compatible with EKRE's national-protectionist positions either, although they were less radical than those of other radical right parties in Central and Eastern Europe (∼LRECON-PROX) (see Chapter 6.1). As in the Czech Republic, the ideological distance between the radical right party and the formateurinhibited cooperation between the two parties in government, and the presence of a cordon sanitaire vis-à-vis EKRE ensured the party's exclusion from government (Braghiroli and Petsinis 2019, 439).

In the last case covered by this solution path, Marian Kotleba's ĽSNS faced the populist anti-establishment party OĽaNO, led by Igor Matovič, as the formateur after the 2020 parliamentary elections.The bipolar opposition in the Slovak party system in 2010 had splintered into multi-polar oppositions by 2020, and this was essentially due to the rise of anti-establishment parties, including OĽaNO and the ĽSNS (∼SAMESIDE) (Rybář and Spáč 2016). While these two parties were united in their anti-establishment appeal, they were not particularly close on either the socio-economic (∼LRECONPROX) or the socio-cultural dimensions (∼GALTANPROX). Matovič and his party expressed support for liberal socio-economic policies, whereas the ĽSNS advocated a clear national-protectionist course. On the socio-cultural dimension, OĽaNO rather belonged on the conservative end of the GALTAN dimension, but the party kept a distance to Kotleba's racist and nativist positions (Mesežnikov and Gyárfášová 2017; Havlík et al. 2020, 218; Hloušek, Kopeček, and Vodová 2020, 107–8). Matovič also endorsed the cordon sanitaire, which can be seen as the culmination of the ideological distance of all parties from the ĽSNS, particularly on the socio-cultural dimension (see Chapter 5.4).

Even though the first two solution paths are quite similar, they refer to different patterns that explain the exclusion of radical right parties from government. The case-based analysis reveals that there are not only differences between, but also within, these two solution paths. This variation stems primarily from the different situations captured by ∼SAMESIDE. If this condition refers to radical right parties competing in party systems characterised by bipolar opposition, then this represents a clear disadvantage for the radical right, as did their socio-cultural distance to the formateur. These two factors, in conjunction with socio-economic distance and/or low levels of fragmentation, explain the exclusion of Ataka in 2013 and the SNS in 2010. In the Slovak case, the SNS' socio-cultural and socio-economic distance from the SDKÚ-DS is deeply intertwined with the parties' membership in antagonistic camps. Here, the combination of ∼SAMESIDE, ∼GALTANPROX, and ∼LRE-CONPROX clearly reflects the ideological polarisation in the party system. In the Bulgarian case, where polarisation between oppositional camps is rather affective, the lack of socio-cultural proximity is not so closely related to the bipolar opposition. Hence, these two cases confirm the above conclusion that the nature of bipolar opposition affects its relation to ideological proximity and, thus, its impact on government formation. If ∼SAMESIDE refers to the absence of a bipolar opposition in the party system, then this condition does not constrain government formation, as has been illustrated in the Romanian cases. In such situations, socio-cultural distance between radical right parties and the formateur played the most pivotal role in preventing the radical right from entering government.

The second variation of explanatory patterns within these two solution paths concerns the presence of a cordon sanitaire, which existed in all cases except the Polish ones. In Romania the emergence of the cordon sanitaire was facilitated by the erosion of the regime divide, enabling the Communist successor party, PDSR, to form coalitions with parties of the former opposition and to distance itself from the radical right. Hence, the absence of a bipolar opposition in the party system had an impact on the exclusion of the radical right from government, even though it did not constrain government formation to coalitions within the same camp.

The third solution path (LRECONPROX\*∼SEATS\*FRAG\*∼SAMESIDE) is quite different from the previous two. The GALTANPROX condition is absent from this path, but with the presence of socio-economic proximity, it includes an ideological condition that, individually, should favour government participation. The two numerical factors refer to small radical right parties (∼SEATS) in fragmented party systems (FRAG). ∼SAMESIDE here refers exclusively to the absence of bipolar opposition in the party system.

The Bulgarian Ataka was only a few months old when entering parliament in 2005. By the mid-2000s, the regime divide that had characterised the Bulgarian party system in the first post-Communist decade had already eroded substantially and no longer constrained coalition formation (∼SAMESIDE). The absence of antagonistic camps enabled the victorious Communist successor party, BSP, to form an oversized coalition with the other two major parties in parliament, the NDSV founded by the former Tsar, Simeon II Sakskoburggotski, and the DPS. All partners in this oversized coalition shared the primary objective of passing the votes required to conclude Bulgaria's accession to the EU with a broad majority (Karasimeonov 2010). Ataka held only a small seat share (8.6 per cent, ∼SEATS) in the fragmented Bulgarian parliament (4.8 effective parliamentary parties, FRAG), which made the party unattractive to a formateur that sought broad political majorities and could draw on larger parties. Hence, Ataka's small seat share and the absence of the bipolar opposition in a fragmented party system prevented the party's inclusion in government.<sup>8</sup> Even Ataka's socio-economic proximity to the BSP (LRECONPROX) could not compensate for this disadvantage.

<sup>8</sup> Ataka's anti-minority and Eurosceptic positions were rather unfavourable for the party's participation in government (Spirova 2006; Marinov 2008; Avramov 2015). Due to the logical minimisation, however, the lack of socio-cultural proximity to the formateur (∼GALTANPROX) is not found in the solution path.

The second case covered by this solution path deals with the formation of the 2016 government in Estonia and the radical right EKRE. Here, a government re-formation took place after the previous government, formed a year earlier by the neoliberal ER, the national-conservative Isamaa and the social democratic SDE, fell. More precisely, the incumbent prime minister party, ER, was replaced by its main competitor, the EK while both junior partners remained in power (Mölder 2017). This replacement illustrates that bipolar opposition did not constrain government formation to parties from two opposite camps in Estonia (∼SAMESIDE). The EK and EKRE shared similar, centre-left, socio-economic positions (LRECONPROX), but they were not natural allies because EK is the main representative of the Russian-speaking minority in Estonian party politics, whereas enmity towards the Russian minority is one of EKRE's main ideological features. EKRE's small seat share (∼SEATS) created another disadvantage. The radical right party was the smallest in parliament and held such a small number of seats that the EK would have needed two additional parties to form a majority with it. In a fragmented party system, such as the Estonian one in 2016 (FRAG), a small seat share does not exclude radical right parties from government per se. What does, however, is the aforementioned cordon sanitaire against EKRE that was still in place one year after the party's electoral breakthrough.

The last two cases covered by this solution path concern the government formations following Marian Kotleba's electoral breakthrough on the national level in the 2016 Slovak parliamentary elections.Of the two radical right parties electedinto parliament in 2016, the SNS entered government and Kotleba's ĽSNS remained in opposition. As has been discussed earlier in this chapter, the 2016 Slovak parliamentary elections resulted in a very fragmented parliament (5.7 effective parties, FRAG) which created a complex bargaining situation. Ultimately, Fico's Smer managed to form an ideologically heterogeneous four-party coalition with the SNS, Most-Híd and Sieť. Shortly thereafter, however, the Sieť faction collapsed, which left the remaining three-party coalition with a somewhat smaller ideological range, particularly on the socio-economic dimension. Kotleba's ĽSNS was close to both Smer and the SNS in socio-economic (LRECONPROX) and socio-cultural terms. Despite the ĽSNS' small seat share (∼SEATS), these three parties could have formed an ideologically homogeneous minimal winning coalition. Under these favourable circumstances, the cordon sanitaire against the ĽSNS' was the main obstacle for its participation in government.

The third solution path also contains different explanations for the exclusion of radical right parties from government.Ataka remainedin opposition due toits small seat share (∼SEATS) in a fragmented party system (FRAG) that was not constrained by bipolar opposition (∼SAMESIDE) and, thus, left the formateur with better alternatives. EKRE and the ĽSNS, in contrast, did not enter government because of a cordon sanitaire. While the explanation of the Bulgarian case differs significantly

from the patterns identified in connection with the first two solution paths, the importance of the cordon sanitaire in the Slovak and Estonian cases resembles the explanation for Úsvit's and the PRM's exclusion from government. All these cases are characterised by the absence of bipolar opposition in the party system and the existence of a cordon sanitaire vis-à-vis the radical right.

#### Paths 4 – 5: In the same camp but not in government

Solution paths 4 and 5 describe situations in which radical right parties are in the same camp as the formateur in a party system characterised by bipolar opposition (SAMESIDE). The following discussion of these solution paths examines which factors prevent them from entering government despite this theoretically favourable condition. The fourth solution path (∼LRECONPROX\*∼GALTANPROX\*∼SEATS\* FRAG\*SAMESIDE) covers Ataka's exclusion from the 2014 and 2016 governments, which included its radical right competitor, PF.GERB won the early elections in 2014 and formed a three-party minority coalition that was supported by the PF. In an attempt to distance itself from its former ally, GERB, Ataka adopted slightly more radical positions, which is reflected in an increase in the ideological distance between Ataka and GERB on the GALTAN dimension. Hence, in 2014 and 2016, both parties were neither socio-culturally (∼GALTANPROX) nor socio-economically (∼LRE-CONPROX) proximate to one another despite being members of the same camp (SAMESIDE).The Bulgarian party system was highly fragmented after the 2014 elections (FRAG), leaving GERB with a choice of several junior partners in the conservative camp, including the radical right parties, Ataka and PF. The PF was the more obvious choice for two reasons: First, it was slightly more moderate than Ataka and, second, Ataka controlled less than five per cent of the seats in parliament (∼SEATS). In 2014, the GERB-led minority coalition was only two seats short of a majority. Here, Ataka's contribution would have been sufficient for a majority, but when parts of the Reformist Bloc and the ABV left the coalition in 2016, Ataka could not compensate for their votes, even if GERB had considered cooperating with it.

The fifth and final solution path (LRECON\*∼SEATS\*∼FRAG\*SAMESIDE) also covers two cases uniquely.The Latvian parliamentary elections in 2010 were the only time when the radical right NA entered parliament but remained in opposition.The NA's ideological position in 2010 was not different than in the following years when the party was included in government. The party was socio-economically close to Unity, the formateur of the 2010 government (LRECONPROX), and positioned on the same side of the bipolar opposition in the party system, which was mainly based on the ethno-linguistic divide (SAMESIDE). This formation of government differed markedly from later ones in the decade with regard to the two numerical factors. After the 2010 elections, the fragmentation of the party system was fairly low by Latvian standards, with only 3.9 effective parliamentary parties (∼FRAG). Moreover, in its electoral breakthrough, the NA received only eight of the 100 seats in the Saeima

and was thus one of the smallest parliamentary groups (∼SEATS). Nevertheless, the formateur and future prime minister, Valdis Dombrovskis, considered an oversized coalition that included the NA as an additional partner. Yet, one of the constituent parties of the Unity alliance ultimately vetoed this coalition because it viewed the NA's nativist positions as too radical (Auers 2011). Due to the NA's electoral weakness, Unity and ZZS were still able to form a majority government without the radical right in 2010. In the following years, the NA improved its electoral results, thus making it more difficult for the other parties in the ethnic Latvian camp to form majorities without it. Only then did the other parties come to accept the NA as a viable coalition partner, even though its participation led to significantly more heterogeneous coalitions on the socio-cultural dimension.

Following the 1998 Hungarian parliamentary elections, ideological factors clearly favoured including the radical right MIÉP in government. The party was not only on the same side of the bipolar opposition in the Hungarian party system as the formateur, Fidesz (SAMESIDE), but it was also close to that party in socio-economic (LRECONPROX) and socio-cultural terms. However, MIÉP's parliamentary group was the smallest in the relatively compact Hungarian parliament (∼SEATS and ∼FRAG). Compared to the other parties in the national-conservative camp, MIÉP was thus in a weak position, as it could contribute very little to a parliamentary majority. Similar to the Latvian case, Fidesz and the other parties from that camp could comfortably reach a majority without the radical right party. When taking into consideration, however, that Viktor Orbán ultimately decided to form an oversized coalition, the ideological factors come back into play. In the nationalconservative camp, MIÉP was without a doubt the most radical party. Given the comfortable majority of this camp, it is not surprising that Orbán decided to limit even an oversized coalition to his more moderate allies, excluding the most radical party from government.

Even though solution paths 4 and 5 differ in some details, they point to a similar explanatory pattern. In all four cases covered by these paths, the radical right parties were not large enough to make a decisive contribution to the majority of their camp. As their camp had a majority even without the radical right, the formateurs preferred to form governments with the more moderate, ideologically proximate parties. Thus, in the cases examined here, the radical right parties were ultimately too small and too radical to enter government.

#### Summary

The analysis of the negative outcome demonstrates that the explanations for the exclusion of radical right parties from government are quite diverse. It is possible, nonetheless, to identify some more general explanatory patterns. First, the presence or absence of bipolar opposition in the party system was critical for distinguishing between broader explanatory patterns. If bipolar opposition existed and radical right parties were not on the same side as the formateur, then they were not included in government. In the observed cases, this constellation was further reinforced by the ideological distance between the radical right and the formateur, in particular on the socio-cultural dimension. Sometimes, however, radical right parties were even excluded from government if they were on the same side as the formateur. This occurred primarily when the parliamentary seats of the radical right party were not required to establish a majority. In this case, formateurs preferred to form governments with ideologically more moderate parties from their camp.

If there was no bipolar opposition in the party system, radical right parties were excluded from government when they faced a cordon sanitaire regardless of whether or not other favourable conditions existed. If neither a cordon sanitaire nor bipolar opposition existed in the party system, the lack of socio-cultural proximity to the formateur was the main reason that prevented radical right parties from entering government, partly aided by their small seat share in a compact party system.

#### **9.3 Explaining the contradictory configuration**

The last part of the empirical analysis investigates the contradictory configuration (LRECONPROX\*∼GALTANPROX\*SEATS\*FRAG\*∼SAMESIDE) that covers the Estonian EKRE in 2019 and the Czech SPD in 2017. Both parties controlled a large seat share in parliament and they were socio-economically, but not socio-culturally, proximate to the formateur. In both cases, government formation took place in a fragmented party system that featured no bipolar opposition.The two cases are also similar with regard to the existence of a cordon sanitaire, because they mark the point when a previously existing cordon sanitaire in the country fully eroded. The following comparative case studies explore why, despite the same configuration of the explanatory conditions, EKRE entered government but the SPD did not. The analysis approaches this question from two different perspectives: Can qualitative differences in the explanatory factors used in this study account for the different outcomes, or are additional factors that are not included in the analytical model needed?

In the 2019 Estonian parliamentary elections, only five parties entered parliament. The fragmentation of the party system was at 4.2 effective parliamentary parties (FRAG) because the parties' seat shares were relatively evenly distributed. EKRE showed strongly, winning 18.8 per cent of the seats (SEATS). Most observers expected either a grand coalition between the victorious ER and the second place EK, or a three-party coalition consisting of the ER, SDE, and the national-conservative Isamaa. EKRE's inclusion in government seemed unlikely, because ER and SDE categorically ruled out any cooperation with the radical right (Rankin 2019; Hartleb 2019). The possibility of a grand coalition illustrates that there was no bipolar opposition constraining the formation of the 2019 Estonian government (∼SAMESIDE). However, incumbent EK prime minister, Jüri Ratas, desperately clinging to power, rejected an offer to become the grand coalition's junior partner. Instead, he began unofficial coalition negotiations with Isamaa and EKRE, while ER leader, Kaja Kallas, was still officially tasked with government formation by the head of state. Ratas' negotiations with EKRE sparked criticism from his political competitors and large parts of the Estonian public. Moreover, they also caused massive tensions within his own party. Several party members strongly opposed cooperation with the radical right and one prominent figure even left the party when the coalition agreement was concluded (Vahtla 2019). The prospect of cooperating with the radical right EKRE was particularly difficult for some EK members, such as Yana Toom, who also served as the spokesperson for the Russian-speaking minority (Ehin and Talving 2019).

Apart from the important ethno-linguistic divide, however, the positions of EKRE and EK overlapped in several respects, both on the socio-economic (LRE-CONPROX) and the socio-cultural dimensions.The government coalition including the EK, EKRE, and Isamaa has been justifiably characterised as a conservative coalition (Walker 2019; Mölder 2020, 119). All three parties are clearly located on the TAN end of the socio-cultural dimension, although EKRE and EK are still separated by almost three points and are therefore not considered socio-culturally proximate to one another according to the standards used in this study (∼GALTANPROX). Below this level, however, the socio-economic and socio-cultural connectedness of their coalition indicates a certain ideological proximity between the three parties. Indeed, on the socio-cultural dimension, this government was a minimal range coalition.

In the Czech Republic, the party system has been in flux since the beginning of the 2010s. The bipolar opposition between the conservative ODS and the social democratic ČSSD that shaped Czech politics in the 2000s had vanished (∼SAME-SIDE). Both parties continuously lost at the polls and this made way for various new parties in parliament (Balík and Hloušek 2016; Mansfeldová and Lacina 2019). Many of these newcomers were populist anti-establishment parties, including the winner of the 2017 parliamentary election, Andrej Babiš' ANO.The radical right SPD—Tamio Okamura's second attempt to gain a footholdin Czech politics—also belonged to this category.

The Czech party system comprised 4.8 effective parliamentary parties and was thus highly fragmented after the 2017 elections (FRAG). In fact, a total of nine parties entered parliament but, with the exception of ANO, none of them controlled more than 12.5 per cent of the seats. Winning 11 per cent of the seats, the radical right SPD was among the strongest of these small parties (SEATS). Due to various accusations of corruption against Babiš, several parties had already ruled out a coalition with ANO during the election campaign (Kudrnáč and Petrúšek 2018). Since the SPD was not among them, Babiš and Okamura held exploratory talks, which marked the end of the cordon sanitaire vis-à-vis radical right parties that had existed since the fall of Communism (Hloušek, Kopeček, and Vodová 2020, 158).<sup>9</sup> The socio-economic platforms of both ANO and the SPD were rather vague but centrist (LRECONPROX). Babiš kept a low profile on the socio-cultural dimension as well. He limited his programmatic appeal to the populist anti-establishment claim that he would run the state like a firm (Buštíková and Guasti 2019; Hanley and Vachudova 2018; Kudrnáč and Petrúšek 2018). Okamura and the SPD, in contrast, fully embraced their anti-Roma and anti-immigration agenda (∼GALTANPROX). Apparently, though, the shared socio-economic positions and anti-establishment appeals did not provide sufficient common ground for the parties to cooperate. Instead, after failing to form a single-party minority government, Babiš ultimately forged a minority coalition with the ČSSD, which had also ruled out cooperation with ANO initially. The coalition of ANO and ČSSD that assumed office in 2018 did not control a majority in parliament, but it was supported by the Communist successor party, KSČM, which had been, until then, considered a pariah in Czech politics (Hloušek, Kopeček, and Vodová 2020, 159–61). Without the KSČM, the minority coalition was ideologically relatively homogeneous. When including the support party, however, theideological rangeincreases significantly and would not have beenmuch different if the parties had opted for the radical right SPD instead.

Both cases are very similar with regard to ideological proximity. The EK and EKRE were slightly closer to each other on the GALTAN dimension than ANO and the SPD, but EK and EKRE disagreed on the ethnic divide. This divide was not as deep in Estonia as it was in neighbouring Latvia, but it was still an essential sociocultural issue in country. Both EKRE and the EK as well as ANO and the SPD also shared centre-left socio-economic positions, even though they were much vaguer in the case of the two Czech parties. Moreover, despite the absence of bipolar opposition in the party system, there were constraints on coalition formation in both countries, because some parties ruled out cooperation with specific competitors during the respective campaigns.

Greater differences can be found in the numerical factors. Even though fragmentation lay above the threshold of 4.0, which marks the distinction between complex and rather straightforward bargaining situations (see Chapter 7), the Czech party system in 2017 was certainly more fragmented than the Estonian one in 2019. This difference is best illustrated when comparing the actual number of parliamentary parties. The Estonian parliament consisted of 4.2 effective and five actual parties, the Czech parliament of 4.8 and nine, respectively. The seat shares of the two

<sup>9</sup> The erosion of the cordon sanitaire had already begun earlier when the SPD was included in several governments at the local level following the 2016 regional elections (Hloušek, Kopeček, and Vodová 2020, 158).

radical right parties showed a similar gradual difference.Both parties were relatively successful at the polls, but EKRE secured 18.8 per cent of the seats in parliament and was, thus, clearly stronger than the SPD, which barely achieved a double-digit result. Hence, from a purely numerical perspective, EKRE's bargaining position was better than the SPD's. The five Estonian parties that entered the Riigikogu in 2017 could form only five minimal winning coalitions, three of them including EKRE. In the highly fragmented Czech parliament, the SPD held enough seats to make a sizeable contribution to a majority, but the fragmentation was so high that more than 20 minimal winning coalitions could be formed, many of them without the radical right.

The cases also differ in the format and ideological range of the coalitions that were formed.The conservative coalition In Estonia was ideologically much more homogeneous than the minority government in the Czech Republic, particularly when including the support party. The format of these governments also corresponds to the dominant patterns in their respective countries. In the Czech Republic, the formation of minority governments is part of the country's political culture and was thus also a viable option for Andrej Babiš after the 2017 elections. In contrast, "the overwhelming preference of Estonian politicians has been to form minimal winning coalitions, and resort to minority cabinets only when absolutely necessary" (Pettai 2019, 185).

In Estonia, ER leader and official formateur, Kaja Kallas, never seriously considered the option of an ER-SDE minority government (Whyte 2019b). Hence, EKRE's strong bargaining position played an important role. Since ER and SDE had ruled out cooperation with EKRE, only two possible majority coalitions remained viable after the EK turned down the invitation to join the grand coalition—the conservative coalition of EK, EKRE, and Isamaa or an alliance of ER, SDE, and Isamaa.Thus, a pivotal role fell to Isamaa, as it was the only party represented in both coalitions. Isamaa had office- and policy-related reasons for choosing the conservative coalition with the EK and EKRE. As a national-conservative party, it was socio-culturally close to EKRE and had previously expressed a preference for a right-wing government. Moreover, the EK offered Isamaa five of the 15 available ministries despite having a seat share of only 12 per cent, reflecting the party's pivotal position (Whyte 2019a; Mölder 2020). Hence, Isamaa also had a strong incentive to join the conservative coalition from an office-seeking perspective.

In the Czech Republic, the ideological preferences of the junior partner, ČSSD, did not play such a vital role, since the parties did not necessarily require another formal coalition partner, given that they had the option of forming a minority government. Yet, if the ČSSD had had a say in the choice of another junior partner, or the support party, it is unlikely that it would have opted for the SPD, since both parties held quite different socio-cultural positions (see Chapter 5.1).

In sum, gradual differences within the explanatory factors included in this study's analytical model and an additional factor, the ideological preference of the junior partner, explain why EKRE entered government and the SPD remained in opposition. EKRE's larger seat share, and the lower fragmentation of the Estonian party system, put the party in a more favourable bargaining position than the Czech SPD. Ultimately, however, the more decisive factor was the preference of the other junior partner, aided by the political culture of government formation, in the respective country.

#### **9.4 Summary**

The analysis of government formation after the first third-generation elections demonstrates that none of the conditions, or their negations, was singularly necessary, or sufficient, for either radical right parties' inclusion in or exclusion from government. The explanatory patterns of government formation clearly became more diverse in the consolidating decades, and the findings in this period confirm some of the hypotheses and results from the previous analysis, but they also illustrate the need for qualifications and revisions in several respects.

With regard to the two numerical factors—the seat share of radical right parties in parliament and the level of party system fragmentation—the results support the hypothesis that small radical right parties should not enter government in compact party systems (Hypothesis 1a). Ataka's government participation in 2009 shows, however, that even this configuration does not always prevent radical right parties from government participation. If a formateur comes very close to a majority in parliament and considers the radical right a viable coalition partner, even a small seat share can be sufficient for entering government.Contrary to the theoretical expectations, however,it was not predominantly large radical right parties in compact party systems that entered government in the consolidating decades. Radical right parties with large (and even small) seat shares in fragmented party systems were included in government much more frequently. Hence, Hypothesis 1c has not been confirmed.

These results underline that the fragmentation of Central and Eastern European party systems is decreasing, but rather slowly and not consistently across all of the countries in the region. In some states, fragmentation remains above the threshold of four effective parliamentary parties consistently (Enyedi and Casal Bértoa 2018; Casal Bértoa 2021). When fragmentation is high, bargaining situations are more complex and majority governments usually require at least three parties. Hence, even parties with a low seat share, including radical right ones, have a chance to be considered junior coalition members or support parties for a minority government in fragmented party systems. Radical right parties with a large seat share, however, are in a stronger position still. Thus, a large seat share gives radical right parties an advantage in government formation, but moderate election results do not necessarily create a decisive disadvantage if party systems are highly fragmented.

With regard to the ideological factors, the analysis largely confirms that the radical right must be socio-culturally close to the formateur and/or on the same side of a bipolar opposition in the party system to enter government during the consolidating decades (Hypothesis 2b). The only outlier was EKRE's participation in 2019 Estonian government. Beyond this confirmation, however, the analysis points to various qualifications of the hypothesis. In the absence of a bipolar divide, for instance, socio-cultural proximity alone was not sufficient for radical right parties to enter government. They also needed to be socio-economically close to the formateur and to control a large seat share in parliament. If bipolar opposition in the party system existed, then the explanatory patterns for government participation differed mainly with regard to the nature of that opposition. If the opposition was based on affective polarisation, the radical right party needed to be in the same camp and hold similar GALTAN positions as the formateur in order to enter government. If the bipolar opposition was result of a socio-cultural divide, as was the case in Latvia, the ideological proximity on the GALTAN dimension became less important. Instead, the government participation of the radical right was aided by socio-economic proximity to the formateur and a high seat share in fragmented parliaments.

These findings highlight that radical right parties and formateurs must share fundamentally similar positions on socio-cultural issues in order to govern together during the consolidating decades. These positions may take the form of ideological proximity on the GALTAN dimension or a position on the same side of a socioculturally rooted, bipolar opposition in the party system. Researchers refer to concepts that entail multiple expressions of one overarching phenomenon as "higher order concepts" (Schneider 2019; see also Mahoney, Kimball, and Koivu 2009). The SOCCUL condition captures the higher order concept related to parties' similar socio-cultural positions in a single set. This condition is present if GALTANPROX is present and/or SAMESIDE is present and based on ideological polarisation that originates from socio-cultural divides. An amended analysis of necessity shows that this SOCCUL condition is a non-trivial necessary condition for government participation in the consolidating decades and over the entire 30-year period covered by this study.<sup>10</sup> The case of EKRE in 2019, which does not fulfil either criterion, prevents the consistency score from reaching 1.00. Even here, however, all three members of the conservative coalition were on the same side of the GALTAN dimension. Thus, this outlier does not fully contradict the set relation of necessity.

<sup>10</sup> In the consolidating decades, this condition reaches a consistency of 0.93, a coverage of 0.78 and a RoN of 0.80. In the whole dataset, these parameters of fit are even higher (consistency: 0.96, coverage: 0.84, RoN: 0.85). For the calibration of this condition, see Appendix III.

Additionally, socio-economic proximity between the radical right and the formateur played a more prominent role in the explanations for their government participation than initially expected during this period. Particularly in party systems where socio-economic issues were highly salient, such as in Latvia or Slovakia, these were often included in the sufficient condition for government participation.

The explanations for the exclusion of radical right parties from government also depended on the presence or absence of bipolar opposition in the party system. In the absence of bipolar opposition, cordons sanitaires and radical right parties' sociocultural distance to the formateur, sometimes further aided by a small seat share in parliament, explain why they remained in opposition. In fact, the qualitative casebased analysis revealed that the existence of a cordon sanitaire frequently prevented radical right parties from entering government, often under otherwise favourable conditions. Despite the widespread openness to radical right politics in the political mainstream of Central and Eastern Europe, radical right parties faced at least a temporary cordon sanitaire in no less than nine of the 19 cases in which they remained in opposition.

If bipolar opposition existed and radical right parties were not in the same camp as the formateur,which happened only twice, they were notincludedin government. In some instances, radical right parties were even excluded from government despite their favourable position in the formateur's camp.This was the case if they were too small and too radical or,in other words,if the more moderate parties in the camp controlled enough seats in parliament to form a majority government without the radical right. These results support the argument that parties prefer ideologically close coalition partners within their own camp (Grzymała-Busse 2001).

The final remarks in this chapter concern the hypotheses about the composition of governments with radical right parties. Hypothesis 3a posits that radical right parties should not be included in oversized coalitions, and this can be confirmed by the analysis. They were predominantly junior partners in minimal winning coalitions and, somewhat less frequently, support parties for minority governments. In Estonia and Slovakia, the junior partnership of radical right parties in minimal winning coalitions corresponds to the dominant format of government in these countries, as does the government participation of the radical right in Poland, where minority governments and minimal winning coalitions are the dominant types. In Bulgaria and Latvia, however, the format of governments with radical right parties deviates from the usual pattern. Minority governments were the most frequent type of governments with radical right parties in Bulgaria, but are otherwise rare in the country. In Latvia, the radical right was only involved in minimal winning coalitions, whereas minority governments and oversized coalitions have been more usual in general (Bergman, Ilonszki, and Müller 2019a).

The analysis of the ideological range of governments generates mixed results. Hypothesis 3b holds that radical right parties should be included in ideologically homogeneous governments, particularly regarding the socio-cultural dimension. However, only the 2006 coalition with the LPR in Poland qualifies as a minimal range and minimal connected winning coalition on both ideological dimensions. The coalition with the radical right SNS that was formed in Slovakia in the same year was both socio-economically and socio-culturally connected, but it does not meet the criteria of the minimal range theory. The coalition that includes the SNS in 2016 is socio-economically connected, but only after the defection of Sieť. In Latvia, the coalitions with the radical right NA are socio-economically, but not socio-culturally, homogeneous. The opposite is true for the 2019 Estonian government that includes the radical right EKRE, which is socio-culturally, but not socio-economically, connected. The governments with radical right parties in Bulgaria are heterogeneous, if the support parties are included. Even the two-party majority coalition of 2017, in which the Bulgarian radical right received cabinet posts for the first time, satisfies the criteria for neither a minimal range nor a minimal connected winning coalition. In light of these observations, Hypotheses 3b cannot be confirmed. Moreover, the question whether or not the ideological range of coalitions with radical right parties differs from the usual pattern in the respective country must also remain unanswered because there is no comparative data on the ideological range and connectedness of governments in Central and Eastern Europe during the 1990s, and it is beyond the scope of this study to generate such an extensive dataset.

## **10. Conclusions**

This chapter summarises the empirical findings on government formation with radical right parties in Central and Eastern Europe, paying particular attention to the comparison of explanatory patterns in the two periods under investigation. It discusses the implications of these findings for future research on radical right parties and government formation in European democracies, but not exclusively in the eastern EU member states. The concluding remarks also relate the findings to the role of radical right parties in the development of democracy in Central and Eastern Europe.

This study has argued that there is a fundamental functional equivalence between the patterns of party competition, and the role of radical right parties, in Central and Eastern and Western Europe. Furthermore, there are context-specific features that account for different patterns of government formation in the two regions.Thus, the central argument of the study holds that, similar toWestern Europe, ideological preferences of Central and Eastern European radical right parties and their competitors, the electoral fortunes of the radical right, and the configuration of the party systems in which they operate, explain why radical right parties enter government or remain in opposition. However, the specific features of the Central and Eastern European context should result in differences between the two regions of Europe. The importance of the socio-economic transformation, the salience and specific nature of nationalism in the region, the regime divide, and the fluidity of party systems should lead to different patterns of government formation in Central and Eastern Europe, particularly in the first decade after the fall of the Iron Curtain.

#### Comparative results

In the period before the first third-generation elections, the explanations for radical right parties' inclusion in, and exclusion from, government were relatively similar across Central and Eastern Europe, indicating the post-Communist transformation's impact on the entire region. The transformational decade was characterised by fragmented party systems and small radical right parties in parliament. Due to high levels of fragmentation, majority coalitions usually comprised a relatively large number of parties, which increased the chances of small parties entering government. Thus, the lack of parliamentary strength did not create a decisive disadvantage for the radical right. However, because almost all empirical cases in this period included small radical right parties in fragmented party systems, ideological factors ultimately made the critical difference. Simultaneous proximity on the socioeconomic and socio-cultural dimensions was a necessary condition for the radical right to enter government in this period. The absence of socio-economic and sociocultural proximity, in turn, constituted a necessary condition for the exclusion of radical right parties from government.Thus, whether the small radical right parties entered government or remained in opposition during the first post-Communist decade ultimately depended on the presence or absence of ideological proximity to the formateur on both dimensions.

The regime divide generated some variation within this explanatory pattern. In some countries, it produced a bipolar opposition in the party system that ran so deep, parties from competing camps found it impossible to cooperate with one another when forming a government. Here, the position of radical right parties in the same, or oppositional, camps reinforced their ideological proximity, or distance, when forming coalition governments. In others, the regime divide did not result in a clear-cut bipolar opposition in the party system.Often, however,it still affected government formation indirectly, because it was a vital source for parties' ideological positions. Slovakia is the only country in this study where the regime divide played a subordinate role in determining coalition participation.

The patterns of government formation with radical right parties in Central and Eastern Europe change considerably, and become more diverse, in the consolidating decades. For example, no individual explanatory factor qualifies as a necessary condition for radical right parties' inclusion in, or exclusion from, government. In this period, the explanations for the participation of radical right parties in government differ primarily with regard to the presence or absence of bipolar opposition in the party system. If no bipolar opposition exists, to enter government, radical right parties must control a large seat share in parliament and be ideologically proximate to the formateur on the socio-economic and the socio-cultural dimensions. In the vast majority of cases, however, party systems featured a deep bipolar opposition, and the government participation of radical right parties was facilitated by being in the same camp as the formateur. This condition alone, however, was not sufficient for explaining their inclusion in government.The additional factors needed to explain why radical right parties enter government depended on the nature of the bipolar opposition. If this opposition resulted from affective polarisation between the largest parties in a country rather than ideological divides, radical right parties needed to share similar socio-cultural positions with the formateur. Thus, GALTAN positions, related to the nativist ideological core of the radical right, played a crucial role even if party competition was strongly affective. If a specific socio-cultural conflict, such as the ethno-linguistic divide in Latvia, split the party system into two

competing camps, ideological proximity on the broader GALTAN dimension played a subordinate role in government formation. In this context, radical right parties needed to control a large seat share in parliament and be socio-economically proximate to the formateur to be included in government.

Despite these differences, fundamentally similar socio-cultural positions of radical right parties and formateurs are a part of all explanatory patterns of government participation in the consolidating decades.These similar positions can take the form of ideological proximity on the GALTAN dimension or a position on the same side of a bipolar opposition in the party system that originates from a socio-cultural divide. When calibrating a new condition that is true if at least one of these forms of similar socio-cultural positions is present, this condition is a necessary condition for government participation.

There were also multiple explanations for the exclusion of radical right parties from government in the consolidating decades. Again, these explanations differed depending on the presence or absence of bipolar opposition. Radical right parties remained in opposition, for instance, if bipolar opposition existed and they were not in the same camp as the formateur, although this condition alone was not sufficient to produce this outcome. Some radical right parties did not enter government, even though they were in the same camp as the formateur. This was the case if they were too small and too radical, or,more precisely,if the other more moderate parties in the formateur's camp could form a parliamentary majority without the seats of the radical right. If the party system was not structured by bipolar opposition, then cordons sanitaires often prevented radical right parties from entering government, sometimes despite otherwise very favourable conditions. If government formation was neither constrained by a cordon sanitaire nor by bipolar opposition, then radical right parties remained in opposition when they controlled only a small share of seats in parliament and lacked ideological proximity to the formateur on the sociocultural dimension.

These results show that the explanatory patterns clearly differed in the periods before and after the first third-generation elections. However, there were also commonalities across all three decades. For instance, in order to govern together, radical right parties and formateurs needed to share similar socio-cultural positions. In fact, the higher-order condition that entails both forms of socio-cultural similarity qualifies as a necessary condition for government participation of Central and Eastern European radical right parties over the entire 30-year period covered by this study. The socio-economic dimension was also relevant for explaining government formation with radical parties in all three decades. Of course, the socio-economic proximity of radical right parties to formateurs was more vital during the transformational decade. However, in several countries, particularly those where the socio-economic dimension was highly salient, this factor featured prominently in the explanations for government participation also after the turn of the millennium. Moreover, the fragmentation in many Central and Eastern European party systems remained so high that low seat shares did not necessarily constitute a decisive disadvantage for radical right parties in the early phase of the transformation as well as the consolidating decades. Another similarity throughout the 30 years covered by this study concerns the format of governments with radical right parties.These parties entered government either as junior partners in minimal winning coalitions, support parties of minority governments or, in rare cases, as junior partners in minority coalitions, but they were never involved in oversized coalitions.

These empirical findings show that the analytical model developed in this study goes a long way in explaining why Central and Eastern European radical right parties enter government or remain in opposition.Moreover, the case-based configurational approach reveals crucial insights into the complex interplay of the individual explanatory factors. However, the empirical analysis determines that the existence of a cordon sanitaire has a greater impact on the exclusion of radical right parties from government than initially expected, and therefore, it should be added to the analytical model. Hence, the analysis suggests a minor modification of the initial model of government formation with radical right parties in Central and Eastern Europe (see Figure 10.1).

*Figure 10.1: Modified model of government formation with radical right parties in Central and Eastern Europe*

Source: Own composition.

### Implications for studying radical right parties and party competition in European democracies

The present study corroborates existing knowledge about government formation with radical right parties, which stems mostly from research on Western European democracies. Similar to the works of Bale (2003), de Lange (2008, 2012), and Fagerholm (2021), it shows that ideological proximity between radical right parties and the formateur on the socio-cultural dimension, as well as party system polarisation, are crucial for explaining why radical right parties enter government or remain in opposition. Overall, the importance of shared socio-cultural positions is the most striking similarity across European democracies. The study also finds that radical right parties' socio-economic proximity to the formateur impacts their participation in government. This factor is essential in the early transformational period but it remains important even after the turn of the millennium, particularly in party systems with salient socio-economic divides. Whether this is a distinct feature of the Central and Eastern European context, or it applies to the rest of the continent as

well, requires further research since the existing literature does not address socioeconomic issues in post-electoral party competition with Western European radical right parties.

These findings provide clear empirical support for the existence of policy-based party competition in Central and Eastern Europe. Government formation in the region is not only a matter of parties' pursuit of public office, it also depends on their ideological preferences (Savage 2014; see also Fagerholm 2021). Despite the widespread absence of detailed party programmes in the early phase of the post-Communist transformation, even in that period parties held core ideological positions that provided guideposts for voters and competing parties (Hloušek and Kopeček 2010, 9–10). In order to arrive at this conclusion, it is necessary to conceptualise the policy space with measures other than classic left-right dimension (Savage 2014, 550).The present study applies a two-dimensional concept using a socio-economic and a socio-cultural dimension. This is not the only viable approach, but in contrast to developing a country-specific left-right dimension (Savage 2014) or focusing solely on socio-cultural issues (Fagerholm 2021), the two-dimensional approach helps to better capture the different elements of a party's ideology, as well as their interaction (see also Spies and Franzmann 2011). The two-dimensional concept of the policy space also provides a promising approach for pan-European research on government formation and party competition. It accounts for the "specificity of the East European context where left-right ideological differences are often blurred" (Minkenberg et al. 2021, 664) and, in contrast to a country-specific left-right dimension (Savage 2014), it applies the same categories to all countries.

With regard to the impact of electoral success on radical right parties' participation in government, the results of this study diverge from the existing literature. While previous studies suggested a linear effect (de Lange 2008), or an advantage for medium-sized radical right parties (Fagerholm 2021), this study demonstrates that party system fragmentation mitigates the impact of parliamentary strength. In the fragmented party systems of Central and Eastern Europe, radical right parties often enter government despite meagre electoral results, particularly in the first post-Communist decade. This finding should be of interest to scholars of radical right parties and party competition across Europe because party system fragmentation in Western Europe is on the rise (Enyedi and Casal Bértoa 2018, 440).Thus,increasingly complicated bargaining situations due to the introduction of many new parliamentary parties could soon help radical right parties to join coalitions in this part of the continent, where parliamentary strength has been a key explanatory factor for their participation in government (de Lange 2008; see also Bale 2003; de Lange 2012).

The most important conceptual contribution of this study stems from the temporal distinction. The analysis shows that the patterns of government formation with radical right parties differ substantially in the two periods analysed. In the early transformational phase, the explanations for why radical right parties enter government or remain in opposition, respectively, are fairly similar across Central and Eastern Europe. This result indicates that the challenges of the triple transition affected party competition in the entire region in comparable ways, despite all of the idiosyncratic regime changes and particular post-Communist adaptations observed in the individual countries (von Beyme 1996; Linz and Stepan 1996; Kitschelt et al. 1999). In this period, government formation with radical right parties had some traits in common withWestern Europe, such as the importance of socio-cultural positions, but it followed a clear Central and Eastern European, or transformational, pattern. This pattern includes electorally weak radical right parties in highly fragmented party systems and the influential impact of both socio-economic and sociocultural positions on government participation. Since these characteristics are a result of the triple transition, similar patterns are likely to be observed in other areas of politics and party competition in Central and Eastern Europe during the 1990s as well.

In the consolidating decades, the explanations for government participation became more similar to those in Western Europe, even though the regime change and the transformational legacies still informed the ideological platforms of parties in the region to a certain degree. Moreover, the development of party competition in Central and Eastern Europe has not followed a universal or linear trajectory. While party systems show increasing closure in some countries (Enyedi and Casal Bértoa 2018), others witness massive changes, such as the rise of new populist anti-establishment parties in the post-transformational phase (Hanley and Sikk 2016; Engler, Pytlas, and Deegan-Krause 2019). Despite these differences, however, the crucialimportance of radical right parties' socio-cultural proximity to formateurs and the impact of bipolar oppositionsin the party system resembles the explanatory patterns of government formation in Western Europe.This is not merely a result of Central and Eastern Europe "catching up" with the West, however. Both parts of the continent are converging and they are facing similar challenges. While there is currently no Western European country where radical right and authoritarian governments endanger liberal democracy and the rule of law to the same degree as in Hungary and Poland, the coalition of the conservative Austrian People's Party (ÖVP) and the radical right Freedom Party of Austria (FPÖ) that governed Austria from December 2018 to May 2019 demonstrated that illiberal tendencies are not foreign to Western Europe either (Wodak 2019)—not to mention established democracies in other Western democracies, such as the US under the presidency of Donald Trump. Moreover, radical right parties remain strong in Western Europe, and the rise of new parties has also caused increasing fragmentation and more complex bargaining situation in this part of the continent (Bolleyer 2013; Emanuele, Chiaramonte, and Soare 2020). In the 2022 French parliamentary elections, for instance, Marine Le Pen's Rassemblement National won 89 of the possible 577 seats and became the largest individual opposition party facing Emmanuel Macron's La République En Marche. Due to the

electoral gains of the radical right the re-elected president fell short of a majority in a fragmented parliament with strong oppositional forces on the right and left. While such a complex bargaining situation is uncommon in France, it resembles many of the negotiations analysed in this study. Since the functional equivalence of party competition in Western and Central and Eastern Europe works in both directions, researchers and politicians can benefit from the insights provided here related to government formation in Central and Eastern Europe when evaluating bargaining situations in the established Western democracies (see also Grotz and Weber 2016).

Overall, it is important to note that the periodisation in this study served as a proxy for the development of party competition and the consolidation of the procedural rules of democracy in the region. Current developments in Hungary and Poland clearly show that democratic consolidation does not have to be a linear process and that illiberal turns which lead to a deterioration of democratic norms and institutions are possible, even after two or three decades (Buštíková and Guasti 2017; Vachudova 2020; see also Cianetti, Dawson, and Hanley 2018).These systems might also present new distinct patterns of government formation, such as a tendency towards one-party majority governments as an expression of the uneven playing field for the incumbent and the opposition parties (Levitsky and Way 2010). Moreover, the established Western European democracies have also seen several fundamental changes in the context of party competition over the past decades, such as the postmaterialist value change since the late 1960s, the fall of the Iron Curtain in 1989/90 or the economic crisis in the late noughties (Inglehart 1977; Ignazi 1992; Hernández and Kriesi 2016). Comparative research on coalition formation, however, has not yet systematically taken the impact of these changes into account, neither in Western Europe nor in Central and Eastern Europe.

Finally, from a research design perspective, this study demonstrates that it is possible to combine a configurational case-oriented approach with a rather extensive empirical design (Müller, Bergman, and Strøm 2008, 33–35). QCA has shown its potential as a method and a research strategy for identifying country- and casespecific explanations for the government participation of radical right parties, while still being able to identify cross-national patterns. This method also enabled the study to incorporate the temporal and spatial context (Ekiert and Hanson 2003b) into the comparative analysis of government formation, thus addressing issues that are often overlooked in medium-to-large-N comparative research in the field (Müller, Bergman, and Strøm 2008, 19–20). Hence, QCA-based research designs offer a promising path for analysing government formation, in particular for projects that focus on specific aspects of coalition politics, such as government formation in a specific region, participation of a specific party family in government, or the formation of minority governments (Keudel-Kaiser 2014; see also Fagerholm 2021).

#### Closing remarks

When present in parliament, Central and Eastern European radical right parties entered government almost half of the time. In contrast to Western Europe, there was never a period of normalisation leading up to the point when the political mainstream accepted them as viable coalition partners (de Lange 2008, 2012). In several Central and Eastern European countries, radical right parties entered government almost immediately after the fall of Communism, even without being particularly successful at the polls or inducing rightward shifts in the policy positions of their competitors. Yet, the frequency of the radical right's government participation in Central and Eastern Europe has increased during the last decade. Hence, their inclusion in government is not simply the result of party system fluidity and transformational politics immediately after 1989. When including the Fidesz governments in Hungary since 2010, the Czech Republic is the only one of all eight countries covered by this study where radical right parties have not (yet) gained executive power at the national level, and even there, the long-standing cordon sanitaire against the radical right is deteriorating.

These observations underline that radical right parties,in general, and their participation in government, in particular, are not exceptional to Central and Eastern European politics. Rather, radical right parties constitute a pathological normalcy in European democracies (Mudde 2010; see also Pytlas 2018). Research demonstrates that they use the access to power for implementing their illiberal policies (Minkenberg et al. 2021; Wierenga and Petsinis 2021) and for injecting their ideology into the political mainstream (Pytlas and Kossack 2015; Pirro 2016; Pytlas 2016; Minkenberg et al. 2021). Thus, they shrink the ideological distance between themselves and their mainstream competitors, which, in turn, improves their chances to enter government in the future.The direct and indirect impact of radical right parties results from both the agency of radical right parties themselves and the positive engagement of mainstream parties with them, both in electoral and post-electoral party competition (Bale 2003; de Lange 2008, 2012; Pytlas 2016; Minkenberg et al. 2021).

Conceptualising radical right parties and their ideology as a pathological normalcy presumes that they will never disappear completely.How much influence they gain, however, depends to a large extent on the strategic choices of the mainstream parties (Pytlas 2016, 224). In post-electoral party competition, mainstream parties are confronted with the choice of whether or not to form a government with the radical right in order to achieve their political goals or to gain access to political power. If they do, they choose not to contain the influence of radical right parties and their politics, and thus, these mainstream parties inadvertently contribute to an erosion of liberal democracy and its underlying values.

## **References**


Beyme, Klaus von. 1984. *Parteien in westlichen Demokratien.* München: Piper.


Bugge, Peter. 1994. "The Czech Republic." In *Political Parties of Eastern Europe, Russia and the Successor States*, edited by Bogdan Szajkowski, 149–73. Harlow: Longman.


Duşa, Adrian. 2019. *QCA with R.* Cham: Springer.


Laakso, Markku, and Rein Taagepera. 1979. "Effective" Number of Parties." *Comparative Political Studies* 12 (1): 3–27. https://doi.org/10.1177/001041407901200101.

Lagerspetz, Mikko, and Henri Vogt. 2013. "Estonia." In Berglund et al. 2013, 51–84.


Łapiński, Dariusz. 2004. *Das Weltbild und die Wirtschaftsauffassung des polnischen Rechtspopulismus.* Berlin: Weißensee.


Pirro, Andrea L. P. 2017. "Hardly Ever Relevant? An Appraisal of Nativist Economics Through the Hungarian Case." *Acta Politica* 52 (3): 339–60. https://doi.org/10.105 7/s41269-016-0032-y.

Poleschuk, Vadim. 2005. "Estonia." In Mudde 2005, 58–79.


Shafir, Michael. 1999. "The Mind of Romania's Radical Right." In Ramet 1999, 213–32.

Shulman, Stephen. 2002. "Challenging the Civic/Ethnic andWest/East Dichotomies in the Study of Nationalism." *Comparative Political Studies* 35 (5): 554–85. https:// doi.org/10.1177/0010414002035005003.

Sikk, Allan. 2005. "How Unstable? Volatility and the Genuinely New Parties in Eastern Europe." *European Journal of Political Research* 44 (3): 391–412. https://doi.org/1 0.1111/j.1475-6765.2005.00232.x.


## **Appendices**

#### **Appendix I. Robustness**

#### A) Fuzzy-set QCA

#### 1. Calibration of set membership

The fuzzy-set QCA conducted here applies 0.2 intervals. The calibration of fuzzy set membership follows the same considerations as the crisp set analysis, butit adds additional qualitative thresholds to produce gradual membership scores.The outcome is coded 1 if radical right parties were junior partners in coalition governments. Radical right parties that participated in government as support parties of a minority government receive a membership score of 0.8 if they were the only support party and 0.6 if they were one of several support parties (Fagerholm 2021). Radical right parties that remain in opposition generally receive a membership score of 0. There are three exceptions to this general rule, however. Even though there was no formal agreement between Ataka and the formateur of the 2013 Bulgarian government, the BSP, the radical right supported the government in the investiture vote and on several other occasions (Ilonszki 2019, 226).Therefore, Ataka receives a membership score of 0.4in 2013.The HungarianMIÉP also voted with the firstOrbán government in the 1998 legislature on various occasions but not as constantly as Atakain 2013 (Karasimeonov 2013a, 2013b; Kostadinova and Popova 2014, 2015; Avramov 2015).The set membership of this party is therefore 0.2. The Latvian NA receives the same membership score in 2010, though for a different reason. Future prime minister and formateur, Valdis Dombrovskis, had invited the NA to join an oversized coalition with his electoral alliance, Unity, and the ZZS. Yet, one party from the Unity alliance ultimately vetoed NA's government participation (Auers 2011).

Regarding the ideological proximity on the socio-economic (LRECONPROX) and socio-cultural (GALTANPROX) dimensions, the calibration of set membership makes a distinction according to whether radical right parties and formateurs are located on the same side of the respective ideological dimension or not (see Chapter 7). If they are located on the same side, the following thresholds will be applied: 0 – 1 point = 1, 1.01 – 1.75 = 0.8, 1.76 – 2.50 = 0.6, 2.51 – 3.25 = 0.4; 3.26 – 4.00 = 0.2, > 4.00 = 0. If radical right parties and formateurs are not located on the same side, the following thresholds will be applied: 0 – 0.5 points = 1, 0.51 – 1.00 = 0.8, 1.01 – 1.50 = 0.6, 1.51 – 1.75 = 0.4; 1.76 – 2.00 = 0.2, > 2.00 = 0.

Regarding seat share (SEATS), the threshold of indifference between set membership and non-membership is set at ten per cent of the seats in parliament. Moreover, radical right parties are considered full members of the set of large radical right parties if they hold at least 13 per cent of the seats, and full non-members of this set when their seat share is below seven per cent. These consideration result in the following thresholds of fuzzy set membership: ≥ 13 per cent of the seats in parliament = 1; 11.5 – 12.9 = 0.8, 10.0 – 11.4 = 0.6, 8.5 – 9.9 = 0.4, 7.0 – 8.4 = 0.2, < 7.0 = 0.

The SAMESIDE condition is included in the same dichotomous coding as in the crisp-set QCA (see Chapter 7). The fragmentation of the party system (FRAG) can be calibrated into fuzzy set membership scores. Based on the discussion of party system fragmentation and coalition formation found in Chapter 7, the following thresholds will be applied: ≥ 4.7 effective parliamentary parties = 1, 4.6 – 4.4 = 0.8; 4.3 – 4.1 = 0.6; 4.0 – 3.8 = 0.4; 3.7 – 3.5 = 0.2; < 3.5 = 0. The result of the calibration of fuzzy set membership is reported in Table A1.1.

*TableA1.1:Calibrationoffuzzysetmembership*




Source: Own compilation, based on data from Casal Bértoa 2021; Nordsieck 2021; Jolly et al. 2022.

## 2. Government formation with radical right parties before the first third-generation elections

#### 2.1 Government participation of radical right parties

The result of the analysis of necessity is reported in Table A1.2.Unlike in the crisp-set QCA,GALTANPROX falls just short of the minimum consistency required for necessary conditions. Hence, based on the fuzzy-set QCA, only socio-economic proximity between radical right parties and the formateur qualifies as a necessary condition for radical right government participation in the period before the first third-generation elections.


*Table A1.2: Parameters of fit necessity: Government participation of radical right parties (before first third-generation elections) (fsQCA)*

Source: Created with QCA Package for R (Duşa 2019).

The fuzzy set truth table (Table A1.3) is exactly the same as in the crisp-set QCA. Even the raw consistency of the four rows that cover the empirically observed cases indicates a perfect set relation despite the fuzzy membership scores.

*TableA1.3:Truthtable:Governmentparticipationofradicalrightparties(beforefirstthird-generationelections)(fsQCA)*



Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The conservative, intermediate, and parsimonious solutions resulting from the truth table analysis are reported in Table A1.4. In light of the theoretical assumptions, and because LRECONPROX is a necessary condition and GALTANPROX is very close to being one, the intermediate solution is based on the directional expectations that these two conditions are present. These directional expectations do not result in any further minimisation, though. Hence, the intermediate solution is exactly the same as the conservative solution. When calculating the parsimonious solution, the prime implicants LRECONPROX and GALTANPROX are tied. Both parsimonious solutions are based on difficult counterfactuals and are therefore not considered for further interpretation.The parsimonious solution reported in Table A1.4 which keeps both prime implicants, is therefore reported merely for illustrational purposes.

*Table A1.4: Sufficient conditions for the government participation of radical right parties (before first third-generation elections) (fsQCA)*



Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* The intermediate solution based on the directional expectations LRECONPROX (present) and GALTANPROX (present) is exactly the same as the conservative solution.

\*\* The prime implicants LRECONPROX and GALTANPROX are tied. The parsimonious solution includes both prime implicants.

The conservative solution is very similar to that in the csQCA and corroborates the above findings. The only difference is the lower coverage of the fsQCA solution. However, the lower coverage does not indicate that there is a case in which radical right parties entered government which is not covered by this solution, but it is a result of the fuzzy set membership scores which yield gradually different membership scores in the outcome set and the respective sufficient solution path. For instance, the membership score of the case of the Estonian ERSP in the set of the solution is only 0.6 while its membership in the outcome is 1.

#### 2.2 The exclusion of radical right parties from government

Table A1.5 shows that none of the conditions, nor their negations, qualify as necessary condition for the negative outcome. ∼LRECONPROX, ∼GALTANPROX, ∼SEATS and ∼SAMESIDE fall just below the required consistency threshold for necessary conditions. Of these four conditions, however, only ∼LRECONPROX and ∼GALTANPROX show coverage and RoN scores that are high enough to rule out triviality. In sum, however, the fsQCA yields no necessary conditions for the negative outcome in this period.


*Table A1.5: Parameters of fit necessity: Exclusion of radical right parties from government (before first third-generation elections) (fsQCA)*

Source: Created with QCA Package for R (Duşa 2019).

The analysis of sufficient conditions for the negative outcome is based on the same truth table as the analysis of government participation. Table A1.6 reports only those rows that cover empirically observed cases (for logical remainders, see Table A1.3). It shows that only row 4, which covers all cases in which radical right remained in opposition before the first third-generation elections, has a sufficiently high, raw consistency with the negative outcome (∼GOVPART).


*TableA1.6:Truthtable:Exclusionofradicalrightpartiesfromgovernment(beforefirstthird-generationelections)(fsQCA)*

334 Pariahs or Partners?

Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The conservative, intermediate, and parsimonious solutions resulting from the truth table analysis are reported in Table A1.7. Since all empirically observed cases are clustered in the same truth table row, the conservative solution is identical with the configuration of that row. Given the theoretical assumptions, and the fact that ∼LRECONPROX and ∼GALTANPROX come closest to being a non-trivial, necessary condition, the intermediate solution is based on the directional expectations that these two conditions are absent. Again, however, these directional expectations do not result in further minimisation. Hence, the intermediate solution is the same as the conservative solution. When calculating the parsimonious solution, the two prime implicants, ∼LRECONPROX and ∼GALTANPROX, are tied. Since it serves only illustrational purposes, the parsimonious solution reported in Table A1.7 again includes both tied prime implicants.

The results from the fsQCA of the negative outcome mirror those from the csQCA. It is striking however, that the coverage of the solution term in the fsQCA is significantly lower than in the csQCA, despite the fact that both solutions cover all empirically observed cases in which radical right parties did not enter government in the period before the first third-generation elections. Hence, even though the results from the csQCA and the fsQCA in this period are fairly similar, the low coverage of the solution term points at a certain degree of noise within the cases and solution paths, which is reflected in the discussion of the variation within these explanatory patterns (see Chapter 8).

### 3. Government formation with radical right parties after the first thirdgeneration elections

#### 3.1 Government participation of radical right parties

The results of the analysis of necessity are reported in Table A1.8. None of the conditions, or their negations, come close to the 0.9 consistency threshold required for necessary conditions. Hence, there are no necessary conditions for the government participation of radical right parties in the period after the first third-generation elections.

*TableA1.7:Sufficientconditionsfortheexclusionofradicalrightpartiesfromgovernment(beforefirstthird-generationelections)(fsQCA)*



Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* The intermediate solution based on the directional expectations LRECONPROX (absent) and GALTANPROX (absent) is exactly the same as the conservative solution.

\*\* The prime implicants∼LRECONPROX and∼GALTANPROX are tied. The parsimonious solution includes both prime implicants.


*Table A1.8: Parameters of fit necessity: Government participation of radical right parties (after first third-generation elections) (fsQCA)*

Source: Created with QCA Package for R (Duşa 2019).

The following analysis of necessity is based on the truth table in Table A1.9. The literature recommends a minimum consistency cut-off of 0.75 – 0.80 in fsQCA. In Table A1.9, there is a large gap between the first five rows, which show a perfect consistency score of 1.00, and the following ones with a consistency score of 0.67 or lower. Hence, a consistency cut-off of 1.00 can be applied.


*TableA1.9:Truthtable:Governmentparticipationofradicalrightparties(afterfirstthird-generationelections)(fsQCA)*



Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The solutions from the analysis of sufficiency, generated using the fsQCA software, are reported in Table A1.10. Since the theoretical assumptions do not allow for directional expectations regarding individual conditions, and the analysis of necessity did not yield any conditions that qualify as necessary, no directional expectations are made in the standard analysis in fsQCA. Therefore, the conservative and intermediate solutions are identical. The parsimonious solution includes two tied prime implicants (∼LRECONPROX\*GALTANPROX\*∼FRAG and ∼LRECONPROX\*∼FRAG\*SAMESIDE). Table A1.10 reports both prime implicants for illustrational purposes, but they are not subjected to further interpretation.


*TableA1.10:Sufficientconditionsforthegovernmentparticipationofradicalrightparties(afterfirstthird-generationelections)*


Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations about individual conditions have been made, the intermediate and conservative solutions are exactly the same. \*\* The prime implicants ∼LRECONPROX\*GALTANPROX\*∼FRAG and ∼LRECONPROX\*∼FRAG\*SAMESIDE are tied. The parsimonious solution includes both primeimplicants.

The intermediate solutions from the csQCA and the fsQCA differin two respects. First, the fsQCA solution includes only solution paths in which the SAMESIDE condition is present. However, a comparison between these solutions paths and those three in the csQCA solution, which also contain this condition, reveals that these sufficient solution paths are relatively similar. Secondly, the fsQCA solution has a much lower coverage than the csQCA solution. Unlike in the analysis of sufficiency of the negative outcome in the earlier period, here, the low coverage results from the fact that the solution does not cover several cases of government participation. More precisely, it covers only nine out of 14 cases, more than half of which concern the government participation of the Latvian NA.

3.2 The exclusion of radical right parties from government

The parameters of fit reported in Table A1.11 indicate that none of the conditions, or their negations, are necessary conditions for the negative outcome. They all fall well short of the consistency threshold of 0.9.


*Table A1.11: Parameters of fit necessity: Exclusion of radical right parties from government (after first third-generation elections) (fsQCA)*

Source: Created with QCA Package for R (Duşa 2019).

The abbreviated truth table in Table A1.12 (for logical remainders, see Table A1.9 above) shows relatively small gaps in the consistency of individual rows in the area above 0.75. Since all truth table rows with a raw consistency of at least 0.75 cover only cases in which radical right parties did not enter government, the consistency cutoff is set at 0.75, thus including rows 1 to 10 in the minimisation.


*A1.12:Truthtable:Exclusionofradicalrightpartiesfromgovernment(afterfirstthird-generationelections)(fsQCA)*

*Table*


Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The solutions from the analysis of sufficiency are reported in Table A1.13. Since neither the theoretical assumptions nor the results from the analysis of necessity allow for any directional expectations regarding individual conditions, the conservative and intermediate solutions yielded by the fsQCA software's standard analysis are identical. The parsimonious solution again serves only illustrational purposes and includes two tied prime implicants (∼LRECONPROX\*∼GALTANPROX and∼LRECONPROX\*∼SAMESIDE).The conservative solutionis very similar to the csQCA of the negative outcome in the same period. It only differs in one of the five solution paths, both of which include the condition ∼SAMESIDE, but in combination with different INUS conditions. The consistency and coverage are lower than in the csQCA, but not as significantly as in the analysis of government participation in this period. This results from the fact that there are three cases of radical right parties in opposition not covered by the solution.

Overall, the fsQCA yielded fairly similar results as the csQCA, though mostly with a lower consistency and coverage. These lower consistency scores indicate that some of the factors for radical right parties' inclusion in, and exclusion from, government during the first post-Communist decade which were necessary conditions in the csQCA do not qualify as such in the fsQCA. Because they are on the border to being necessary, however, this result does not fundamentally contradict the findings in the csQCA.The sufficient solution paths in the fsQCA and the csQCA are also quite similar. Hence, the fsQCA corroborates the robustness of the results in this study (Schneider and Wagemann 2012, chap. 11.2).


*TableA1.13:Sufficientconditionsfortheexclusionofradicalrightpartiesfromgovernment(afterfirstthird-generationelections)*

*(fsQCA)*



Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations about individual conditions have been made, the intermediate and conservative solutions are exactly the same. \*\* The prime implicants ∼LRECONPROX\*∼GALTANPROX and ∼LRECONPROX\*∼SAMESIDE are tied. The parsimonious solution includes both prime implicants.

## B) Temporal threshold: EU membership (csQCA)

## 1. Government formation with radical right parties before EU membership

In Chapter 4, the first third-generation elections were selected as the temporal threshold instead of the countries' accession to the European Union. Dividing the three post-Communist decades into before- and after-accession periods results in certain changes to the results.

1.1 Government participation of radical right parties

Table A1.14 shows that the RoN and coverage of socio-economic proximity (LRE-CONPROX) between radical right parties and formateurs decreases, so that it no longer qualifies as a necessary condition, rendering socio-cultural proximity (GAL-TANPROX) the only necessary condition for government participation.This finding corroborates the argument that the importance of socio-economic issues decreases in the consolidating decades.


*Table A1.14: Parameters of fit necessity: Government participation of radical right parties (before EU membership)*

Source: Created with QCA Package for R (Duşa 2019).

Extending the first period reduces the number of logical remainders to 22 (see Table A1.15), compared to 28 in the period before the first third-generation elections (see Chapter 8).


*A1.15:Truthtable:Governmentformationwithradicalrightparties(beforeEUmembership)*

*Table*


Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

The conservative solution that results from the minimisation of rows 1 to 3, remains the same as in the period before the first third-generation elections (see Table A1.16), because all additional cases concern radical right parties that remained in opposition.




Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations about individual conditions have been made, the intermediate and conservative solutions are exactly the same.

#### 1.2 The exclusion of radical right parties from government

Similar to the analysis of necessary conditions for government participation of radical right parties, the negation of the LRECONPROX condition does not qualify as a necessary condition for the exclusion of radical right parties from government when the time frame is extended until accession to the EU (see Table A1.17). The consistency of the socio-cultural distance between radical right parties and formateurs (∼GALTANPROX) is lower than in the period before the first third-generation elections, but it is still above the minimum consistency for necessary conditions. As the lower consistency results from a contradiction in kind (MIÉP 1998), however, it is questionable whether or not this condition can be considered necessary for the negative outcome.The negation of the SAMESIDE condition also passes the consistency threshold of 0.9. However, the RoN is rather low and∼SAMESIDE reflects two theoretically different concepts—the absence of a bipolar opposition in the party system and the radical right and the formateur in opposite camps. Hence, this condition is not considered necessary either.


*Table A1.17: Parameters of fit necessity: Exclusion of radical right parties from government (before EU membership)*

Source: Created with QCA Package for R (Duşa 2019).

The analysis of sufficient conditions for the negative outcome before the accession to the EU includes more truth table rows and is more complex than in the period before the first third-generation elections, when all cases in which radical right parties were excluded from government clustered in a single truth table row.Therefore, the solutions in Table A1.18 differ somewhat from the results of the original analysis. This observation indicates that country-specific party competition in Central and Eastern Europe began to diversify in the period between the first third-generation elections and the accession to the European Union.


*Table A1.18: Sufficient conditions for the exclusion of radical right parties from government (before EU membership)*


Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations about individual conditions have been made, the intermediate and conservative solutions are exactly the same. The prime implicants LRECONPROX\*∼GALTANPROX\*∼FRAG\*∼SAMESIDE and LRECONPROX\*∼GALTANPROX\*∼SEATS\*∼SAMESIDE are tied. The conservative/intermediatesolutionincludesbothprimeimplicants. \*\*

 The prime implicants ∼SEATS\*∼FRAG and∼FRAG\*SAMESIDE are tied. The parsimonious solution includes both prime implicants.

## 2. Government formation with radical right parties since EU membership

2.1 Government participation of radical right parties

The parameters of fit in Table A1.19 illustrate that none of the conditions in the analytical model, nor their negations, are necessary for government participation of radical right parties in the period since EU membership. This mirrors the result in the period after the first third-generation elections.


*Table A1.19: Parameters of fit necessity: Government participation of radical right parties (since EU membership)*

Source: Created with QCA Package for R (Duşa 2019).

The number of logical remainders in the truth table that covers the period after the countries' accession to the EU is slightly lower than in the period after the first third-generation elections, and the truth table contains the same contradictory configuration (see Table A1.20).


*A1.20:Truthtable:Governmentformationwithradicalrightparties(sinceEUmembership)*

*Table*



Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

Because changing the temporal threshold affects only cases in which radical right parties remained in opposition, the analysis of sufficiency in the period since EU membership yields the same conservative solution as in the period after the first third-generation elections (see Table A1.21).


*Table*

*A1.21:*

*Sufficient*

*conditions*

*for*

*the*

*government*

*participation*

*of*

*radical*

*right*

*parties*

*(since*

*EU*

*membership)*


 Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations about individual conditions have been made, the intermediate and conservative solutions are exactly the same.

2.2 The exclusion of radical right parties from government

Table A1.22 shows that there are no necessary conditions for the exclusion of radical right parties from government in the period after the countries' accession to the EU.

*Table A1.22: Parameters of fit necessity: Exclusion of radical right parties from government (since EU membership)*


Source: Created with QCA Package for R (Duşa 2019).

The analysis of sufficiency yields fewer, and slightly different, solution paths than in the period after the first third-generation elections because some cases have already been included in the period before EU membership (see Table A1.23). However, both approaches to periodisation result in similar findings. Here, again the absence of socio-cultural proximity between radical right parties and the formateur is included in three of the four solution paths, and the cases covered by the solution comprise a large number of parties that remained in opposition due to a cordon sanitaire, sometimes despite otherwise favourable conditions.

#### C) Recalibrating fragmentation

In Chapter 7, the set of fragmented party systems was calibrated to assign party systems with more than 4.0 effective parliamentary parties a membership score of 1. The two cases with a fragmentation of 3.9 and 4.0, Latvia and Slovakia in 2010, were characterised as moderately complex bargaining situations, because the level of complexity in these two cases was more similar to party systems with lower fragmentation. Hence, both cases were not included in the set of fragmented party sys-




Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations on individual conditions have been made, the intermediate solution is exactly the same as the conservative solution. \*\* The prime implicants ∼GALTANPROX\*∼FRAG and∼GALTANPROX\*∼SEATS are tied. The parsimonious solution includes both prime implicants.

tems. Lowering the threshold to 3.8 and thus including both cases in this set changes the results marginally. Since both cases concern the period after the first third-generation elections, the results of the earlier period are not affected.

## 1. Government participation of radical right parties

Table A1.24 shows that changing the calibration of the FRAG condition does not affect the conclusions regarding necessary conditions.The consistencies of FRAG and ∼FRAG remain constant, while the RoN and coverage scores change marginally.


*Table A1.24: Parameters of fit necessity: Government participation of radical right parties (after first third-generation elections)*

Source: Created with QCA Package for R (Duşa 2019); values in parentheses report parameters of fit in the original analysis.

The new truth table (see Table A1.25) shows that changing the calibration results in one additional logical remainder, since the case of the Slovak SNS in 2010 now has the same configuration as the Czech Úsvit in 2013 and the Estonian EKRE in 2015. The change neither resolves the previously existing contradiction nor does it create a new one.


*A1.25:Truthtable:Governmentparticipationofradicalrightparties(afterfirstthird-generationelections)*

*Table*



Source: Created with fsQCA 3.0 (Ragin and Davey 2016).

#### 376 Pariahs or Partners?

As the change in the calibration concerns only cases in which radical right parties remained in opposition, it does not affect the conservative solution yielded by the analysis of sufficient conditions for the participation of radical right parties in government in the period after the first third-generation elections (see Table A1.26). Additionally, the recalibrated cases are not covered by any of the logical remainders that were used for crafting the intermediate solution in the original analysis (see Chapter 9.1).


*TableA1.26:Sufficientconditionsforthegovernmentparticipationofradicalrightparties(afterfirstthird-generationelections)*


Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* Since no directional expectations on individual conditions have been made, the intermediate solution is exactly the same as the conservative solution. \*\*

 The prime implicants ∼LRECONPROX\*GALTANPROX, ∼LRECONPROX\*∼SEATS\*∼FRAG, and ∼LRECONPROX\*∼FRAG\*SAMESIDE are tied. The parsimonious solution includes all three prime implicants.

### 2. The exclusion of radical right parties from government

The recalibration of fragmentation results in marginal changes to the parameters of fit necessity for the negative outcome (see Table A1.27). However, there are still no necessary conditions for the exclusion of radical right parties from government in the period after the first third-generation elections.


*Table A1.27: Parameters of fit necessity: Exclusion of radical right parties from government (after first third-generation elections)*

Source: Created with QCA Package for R (Duşa 2019); values in parentheses report parameters of fit in the original analysis.

The conservative solution in Table A1.28 is somewhat different from the one resulting from the original calibration of fragmentation in Chapter 9 (see Table 9.5). In the original analysis of sufficiency, the intermediate solution was crafted by reversing the minimisation step that led to dropping the condition SAMESIDE from the first solution path. Thus, the intermediate solution better illustrates which conditions led to the negative outcome when the SAMESIDE condition was present or absent, respectively. The same procedure is used here. After this step, the two intermediate solutions are quite similar and illustrate that the recalibration does not result in a significant change of the explanatory patterns for the negative outcome.


*A1.28:Sufficientconditionsfortheexclusionofradicalrightpartiesfromgovernment(afterfirstthird-generationelections)*

*Table*




Source:CreatedwithfsQCA3.0(RaginandDavey2016).

 \* The prime implicants ∼LRECONPROX\*SAMESIDE and ∼LRECONPROX\*GALTANPROX are tied. The parsimonious solution includes both prime implicants.

## **Appendix II. Salience of socio-economic and socio-cultural issues in Central and Eastern European party systems**

Since the 2014 wave, the Chapel Hill Expert Survey (CHES) provides data on the salience of the socio-economic and socio-cultural dimensions for the individual parties, in addition to their positions (Jolly et al. 2022). The salience ranges from 0 (low) to 10 (high). The average salience of each dimension in the party system can be obtained by summing the salience of the respective dimension for each party and weighting it by their vote share. The mathematical formula for calculating the salience is as follows:

> Salience = ∑{(saliencei\*vote share<sup>i</sup> )/(∑vote share<sup>i</sup> )}

where i represents individual parties. Table A2.1 shows the average salience of the socio-economic and socio-cultural dimensionsin the countries covered by this study in the second half of the 2010s, which is based on the salience of the LRECON and GALTAN dimensions in the 2014 and 2019 CHES waves.


*Table A2.1: Salience of the socio-economic and socio-cultural dimension in the second half of the 2010s*

Source: Own compilation, based on data from the Chapel Hill Expert Survey (Jolly et al. 2022).

## **Appendix III. Calibration of the higher-order condition of fundamentally similar socio-cultural positions**

Table A3.1 displays the calibration of the higher-order condition of fundamentally similar socio-cultural positions of radical right parties and formateurs (SOCCUL). The SOCCUL condition is true if GALTANPROX is present and/or SAMESIDE is present and based on ideological polarisation that originates from socio-cultural divides.







 For

 see

 7.

# Social Sciences

 Anke Strüver, Sybille Bauriedl (eds.) Platformization of Urban Life Towards a Technocapitalist Transformation of European Cities

September 2022, 304 p., pb. 29,50 € (DE), 978-3-8376-5964-1 E-Book: available as free open access publication PDF: ISBN 978-3-8394-5964-5

#### Mihnea Tanasescu Understanding the Rights of Nature A Critical Introduction

February 2022, 168 p., pb. 40,00 € (DE), 978-3-8376-5431-8 E-Book: available as free open access publication PDF: ISBN 978-3-8394-5431-2

 Oliver Krüger Virtual Immortality – God, Evolution, and the Singularity in Post- and Transhumanism

2021, 356 p., pb., ill. 35,00 € (DE), 978-3-8376-5059-4 E-Book: PDF: 34,99 € (DE), ISBN 978-3-8394-5059-8

# Social Sciences

 Dean Caivano, Sarah Naumes The Sublime of the Political Narrative and Autoethnography as Theory

2021, 162 p., hardcover 100,00 € (DE), 978-3-8376-4772-3 E-Book: PDF: 99,99 € (DE), ISBN 978-3-8394-4772-7

 Friederike Landau, Lucas Pohl, Nikolai Roskamm (eds.) [Un]Grounding Post-Foundational Geographies

2021, 348 p., pb., col. ill. 50,00 € (DE), 978-3-8376-5073-0 E-Book: PDF: 49,99 € (DE), ISBN 978-3-8394-5073-4

 Andreas de Bruin Mindfulness and Meditation at University 10 Years of the Munich Model

2021, 216 p., pb. 25,00 € (DE), 978-3-8376-5696-1 E-Book: available as free open access publication PDF: ISBN 978-3-8394-5696-5