Carl Friedrich Bossert Power, Alliances, and Redistribution For Verónica and Emilia

Carl Friedrich Bossert

# Power, Alliances, and Redistribution

The Politics of Social Protection for Low-Income Earners in Argentina, 1943–2015

Budrich Academic Press Opladen, Berlin & Toronto 2021 © 2021 This work is licensed under the Creative Commons Attribution-ShareAlike 4.0. (CC-BY-SA 4.0). It permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you share under the same license, give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/4.0/

© 2021 Dieses Werk ist der Budrich Academic Press GmbH erschienen und steht unter der Creative Commons Lizenz Attribution-ShareAlike 4.0 International (CC BY-SA 4.0): https://creativecommons.org/licenses/by-sa/4.0/ Diese Lizenz erlaubt die Verbreitung, Speicherung, Vervielfältigung und Bearbeitung bei Verwendung der gleichen CC-BY-SA 4.0-Lizenz und unter Angabe der UrheberInnen, Rechte, Änderungen und verwendeten Lizenz.

This book is available as a free download from www.budrich.eu (https://doi.org/10.3224/96665028). A paperback version is available at a charge. The page numbers of the open access edition correspond with the paperback edition.

This book is based on a doctoral thesis with the same title handed in at the faculty of social sciences (FB05) at the University of Kassel. The defense took place on November 6th, 2019. The writing of this book was possible thanks to a scholarship of the Hans-Böckler-Foundation.

Includes bibliographical references and index.

ISBN 978-3-96665-028-1 eISBN 978-3-96665-998-7 DOI 10.3224/96665028 1. Social policy – Argentina. 2. Social policy – Latin America. 3. Inequality – Latin America. 4. Redistribution – Argentina. 5. Political economy – Argentina. 6. Power resources approach. 7. State theory.

Budrich Academic Press GmbH Stauffenbergstr. 7. D-51379 Leverkusen Opladen, Germany

86 Delma Drive. Toronto, ON M8W 4P6 Canada www.budrich.eu

A CIP catalogue record for this book is available from Die Deutsche Bibliothek (The German Library) (http://dnb.d-nb.de) Jacket illustration by Bettina Lehfeldt, Kleinmachnow, Germany – www.lehfeldtgraphic.de Picture credits: Cover image: Nicolas Pousthomis Typesetting by Anja Borkam, Jena, Germany – kontakt@lektorat-borkam.de Printed in Europe on acid-free paper by Books on Demand GmbH, Norderstedt

## Contents



## Tables



## Figures


## Acronyms




## Preface and Acknowledgements

I am convinced that behind every book there is a story worthy of being told. In the case of this book, *another world is possible!* might well be the title of that story. These four words have been the motto of the global justice movement that emerged with considerable strength around the turn of the century. At a time when inequality was on the rise and many governments around the globe were implementing deeply regressive social policies and tax reforms, this movement provided me with the opportunity to participate in a wide range of inspiring protests and conferences, among which the World Social Forum in Porto Alegre, Brazil, which boasted more than 100,000 participants, constituted one of the most important impetuses for my following scientific work. It was in this framework that I got to know more about several progressive political experiments that were under way in Latin America. After years of intensive political mobilization, left parties and coalitions had won elections in countries such as Brazil, Argentina and Uruguay and started redistributive reforms at a larger scale. While virtually all other regions of the world remained on the neoliberal track and inequalities continued to grow, Latin America increasingly swam against the current. However, my observations over the years were ambivalent. While social policy reform set into motion processes of significant redistribution, inequality and related human sufferings were and still are extreme. I realized that social policy cannot remedy all injustices and social problems, but that it is of enormous importance for increasing the well-being of the population—and in particular that of the poorest sectors. While many of my colleagues and friends were critical of state-led social policies, as these had often served to control, discipline and disorganize the beneficiaries, I always believed that social policy is neither *per se* divisive, regressive and disciplinary nor *per se* solidarity-enhancing, redistributive and emancipating. Rather, whether their effects incline to one side or the other depends on the design of social programs and the forces that are able to shape the process of development and implementation. Hence, the crucial question I became increasingly passionate about was how progressive social policy could be advanced and expanded. It was this passion that provided the drive for nearly a decade of intensive research and reflection on Argentinian and Latin American social policy. The outcome is this book—which I believe will provide important new insights in the long-term dynamics of redistributive struggles and social policy change.

While the writing of this book was driven by personal experiences and motivation, the outcome is as much the result of the inspiration, solidarity and support given by others.

In the first place, I want to thank Hans-Jürgen Burchardt of the University of Kassel for his committed supervision of my doctoral thesis. There are very few supervisors who invest as much time and thinking in such a task as he has. I am also deeply grateful to Fernando Groisman of the University of Buenos Aires who generously supported my work with his extensive knowledge of Argentine social policy and methodological approaches. Fernando Groisman was not only generous with his colleagues but also tirelessly committed to social justice. His death in 2016 leaves a big hole in the scientific community and in our hearts.

I also want to thank my friends and colleagues from the Graduate School of Global Social Policies and Governance at the University of Kassel who contributed through innumerable theoretical, methodological, and empirical discussions. I am particularly grateful to Nico Weinmann, Paul Hecker, Ezequiel Bistoletti, Jenny Jungehülsing, Johanna Neuhauser, Sebastian Matthes, Stefan Peters, Anne Tittor, Donna McGuire, and Philipp Fehling.

Properly understanding Argentine politics and social policy would furthermore not have been possible without frequent contact and interchange with a range of Argentine friends and colleagues. Especially, I would like to mention Diego Rozengardt, Laura Golbert, Aldo Isuani, Aldo Neri, Eugenia Sconfienza, Nicolás Damín, and Guillermo Alonso. They were all important for my understanding of highly complex political phenomena such as Peronism and were also extremely generous in providing me with crucial contacts, interview material, and literature.

I am deeply thankful to all 28 interview partners, who mostly occupied important political positions and nevertheless took the time to respond to my questions about a variety of important social policy reform processes. Their openness and often unpublished knowledge contributed significantly to my efforts to reconstruct the history of social policy change and helped decisively to identify causal relationships.

Moreover, the doctoral scholarship of the Hans-Böckler-Foundation, as well as the provision of an excellent network, workplace, and seminars by the Graduate School of Global Social Policies and Governance were enormously helpful.

However, most importantly, I am endlessly thankful to my family and friends for supporting me throughout the many years of hard work on the thesis. My parents Elisabeth and Tobias, as well as my brothers Gottlieb and Leopold, always supported me when I was in need: ranging from IT problems, over financial bottlenecks to the simple necessity of company and heartiness. The same I can say for Berti, Silvia, Andrés, Pablo and Roxana who constitute something like my second family in Argentina. Nobody was, nevertheless, more vital for my ability to write this thesis than my wife Verónica. She is not only the best *compañera de vida* and emotional support I could have but also an outstanding political scientist. There is nobody with whom I have discussed more issues of theoretical and empirical importance. Her sharp and creative analyses have contributed hugely. Furthermore, she took a lot of effort to keep my back free while I was struggling to finish the thesis. For all this I am infinitely grateful. And finally, our daughter Emilia has made the last seven months of my working on the thesis the most cheerful and beautiful of my life.

## 1 Introduction

Without any doubt, the proliferation and expansion of social policies over the last hundred years has been a macro-political phenomenon of enormous social, economic, and political importance for Latin American societies. But while the rapid growth of social programs in industrialized countries received growing scientific attention from the 1970s on, similar processes in Latin America were subject to a comparatively limited number of studies until the early 2000s.<sup>1</sup> When in the late 1970s Ian Gough (1979, 1) went as far as to say of the industrialized countries that the "twentieth century, and in particular the period since the Second World War, can fairly be described as the era of the welfare state," several Latin American countries could look back at over seven decades of social insurance development and even surpassed the United States and other industrialized countries in programs such as health-maternity insurance and family allowances (Mesa-Lago 1989, XV). When Esping-Andersen ([1990] 1998, 1) observed for the industrialized countries that what "once were nightwatchman states, law-and-order states, militarist states, or even repressive organs of totalitarian rule, are now institutions predominantly preoccupied with the production and distribution of social well-being," many Latin American states devoted about 40% of their expenditures to social policies and in some countries, such as Argentina, Uruguay, and Costa Rica, this number even ascended to 50% or 60% (Segura-Ubiergo 2007, 14).<sup>2</sup> The importance of the expansion of social policies, however, not only rested on the devotion of significant resources to providing such elemental things as access to health care, education, food, and income security. It rested as much on shaping the social stratification of the society, family and gender relations, the distribution of political power and the basic dynamics and rules of the economy (e.g., Barrientos 2004; Esping-Andersen [1990] 1998; Filgueira 2005; Huber and Stephens 2001; Lewis 1992; Martínez Franzoni 2008; Orloff 1996).

Already towards the end of the 20th century, Carmelo Mesa-Lago (1989, XV) had concluded that "in terms of social security […] Latin America is a leader in the Third World."

<sup>1</sup> A significant growth of studies on Latin American social policy commenced with the proliferation of neoliberal privatization and retrenchment measures during the 1990s and further intensified with the political left-turn in the region during the 2000s.

<sup>2</sup> According to Segura-Ubiergo (2007, 14), during the 1973–2000 period only four industrialized countries devoted a higher share of their public expenditure to social policies: the Netherlands, Switzerland, France and Germany.

## **A Latin American Paradox: Significant Social Expenditure Without Significant Redistribution**

In the context of high world market prices for Latin American commodity exports, high economic growth rates and the election of left-of-center governments in a significant proportion of the region, social policy expansion received another decisive push during the first one and a half decades of the 21st century. Many of the reform initiatives during this period, such as the expansion of social services, universal cash transfers, targeted social assistance programs, and the easing of access criteria for social insurance benefits, had a clearly redistributive orientation (Barrientos and Santibáñiez 2009; Cruz-Martínez 2019; Lustig 2015; Lustig, Pessino and Scott 2014). In 2015, the average public social expenditure in the region reached 14.6% of GDP and was hence not far from the 19.0% average spent by OECD countries.<sup>3</sup> Some countries, such as Argentina, Brazil, and Costa Rica, even spent over 23% of their GDP and surpassed, in relative terms, highly developed welfare states, such as the Netherlands, the United Kingdom or Iceland (OECD 2019; CEPAL 2017a, 123).

However, despite this stunning expansion of social policies and the progressive character of the recent reform cycle, Latin America is still characterized by extreme inequalities.<sup>4</sup> These concern not only the distribution of incomes but pervade nearly every aspect of social and economic life, such as access to health care, education, labor markets, land, housing, and sewage treatment (Burchardt 2012; Ferranti et al. 2004; Peters 2013; Tittor 2012). Inequality expert Nora Lustig (2015, 14) recently pointed out that while Latin America is home to about "5 percent of the world's billionaires, the poor are strikingly poor. Infant mortality and malnutrition in rural areas and shantytowns, and among disadvantaged groups in Latin American middle-income countries, are much the same as in notably poorer nations." Hence, there is a pressing question: How can Latin America be at the same time leader in the Third World in terms of social policy *and* in terms of inequality?

An important part of the answer lies *in the way* social policy systems evolved in Latin America. While social policy transfers and taxes in OECD countries reduced income inequalities measured with the Gini coefficient by an estimated average of 36% during the early 2010s, they did so by a meager

<sup>3</sup> The numbers are simple un-weighted country averages. The average provided by CEPAL (2017a, 123) for Latin America and the Caribbean covers only central government spending for many of the countries. Therefore, the already considerable social spending share of 14.6% of GDP constitutes still an underestimation of the true overall public social expenditure.

<sup>4</sup> As a socialist country, Cuba constitutes an exception to this as well as to the following observations regarding the relationship between social policies and inequality.

6% in Latin America (Hanni, Podestá and Martner 2015, 13).<sup>5</sup> Of course, neither among Latin American countries nor among industrialized countries are social policy regimes uniform. Since the publication of Gøsta Esping-Andersen's influential book *The Three Worlds of Welfare Capitalism* in 1990, a growing number of comparative studies have clustered welfare regimes according to a range of aspects, such as their effects on social stratification, labor markets, gender relations, and income inequality .<sup>6</sup> However, even if we compare leading countries of the most developed and redistributive social policy regime types in Latin America, such as Argentina, Uruguay, and Costa Rica, with countries of the least egalitarian liberal regime type in the industrialized countries, such as Australia, Canada or the United Kingdom, the latter achieve significantly stronger redistributive effects. Around 2010, social policy transfers reduced income inequality measured with the Gini coefficient by approximately 9% in Argentina, 5% in Costa Rica and 11% in Uruguay (Hanni, Podestá and Martner 2015, 11). In comparison, the same effect was about 18% in Australia, 14% in Canada, and 21% in the United Kingdom (Joumard, Pisu and Bloch 2012, 13). Significantly, Argentina, Costa Rica, and Uruguay actually devoted bigger shares of their GDPs to social spending than did Australia, Canada and the United Kingdom.<sup>7</sup>

This stunning paradox of significant social spending without comparably significant redistribution raises at least two important technical and political questions: First, what are the peculiarities in the design and the rules of Latin American social policy regimes that explain this paradox? And second, what are the underlying social, economic, and political reasons that led to the development of these peculiarities?

The first question has motivated the undertaking of several technical studies with detailed analyses of the redistributive effects of virtually all major social policies in Latin America (e.g., Ferranti et al. 2004; Goñi, López and Servén 2008; Lindert, Skoufias and Shapiro 2006). The findings show that the key reason for the lack of redistribution is the truncated character of most social programs, which means that they either exclude or strongly disadvantage low-

<sup>5</sup> The numbers refer to simple un-weighted country averages. In both regions cash transfers had a significantly stronger impact on income inequality than taxation (OECD 2007, 53). In Latin America, Hanni, Podestá and Martner (2015, 13) estimate that cash transfers account for over 60% of reduction of the Gini coefficient.

<sup>6</sup> For an overview of different social policy systems in Latin America see e.g., Barrientos (2004); Burchardt, Tittor and Weinmann (2012); Filgueira (2005); Gough (2013); Gough and Wood (2004); and Martínez Franzoni (2008). For an overview of different social policy systems in industrialized countries see e.g., Ebbinghaus (2012), Esping-Andersen ([1990] 1998); Huber and Stephens (2001); and Orloff (1996).

<sup>7</sup> According to CEPAL (2013, 174) Argentina spent 27.8%, Costa Rica 22.7% and Uruguay 24.2% of their GDPs on social policies, while according to the OECD (2019) Australia spent 16.6%, Canada 17.5% and the United Kingdom 22.4%.

income earners, precisely those who suffer from the most urgent social needs.<sup>8</sup> While big shares of the social spending are used to finance occupationally based social insurance schemes, most Latin American low-income earners are peasants, informal workers or unemployed workers and are therefore deprived of access to these programs. In practice, this often means exclusion from a whole range of social protections, such as family allowances, pensions, unemployment allowances, health care, work accident compensation, and wage payment during sickness. According to the International Labor Organization (ILO), in 2015, approximately 140 million persons or 53% of all workers in Latin America were informal workers (Salazar-Xirinachs and Chacaltana 2018, 22). Furthermore, social insurance systems are highly fragmented in most Latin American countries, which implies that, within the population of formal workers, members of high-income groups often receive disproportionally better benefits than members of low-income groups. Even spending on non-contributory social policies, such as public health care and education, often benefits the better-off more than poorer sectors, as the latter often live in rural or marginalized urban areas in which public social services are either lacking or provided in inferior quality (Burchardt 2012; Ferranti et al. 2004; Goñi, López and Servén 2008; Lindert, Skoufias and Shapiro 2006).

## **The Research Focus**

While several of the above mentioned studies contain detailed analyses of the social policy designs that disadvantage or exclude low-income earners and elaborate recommendations for rendering Latin American social policy regimes more inclusionary (Barrientos and Hulme 2008; Ferranti et al. 2004; Goñi, López and Servén 2008; Lindert, Skoufias and Shapiro 2006), we know much less about the politics that led to the proliferation of such inegalitarian regulations. Social policies do not simply develop out of technical debates and recommendations but are always the outcome of political processes. This does not mean that such recommendations do not influence social policy development, but rather that whether and what recommendations are made at a certain moment and even more so which of them enter the governmental agenda-setting process and are implemented depends decisively on political factors.

The main goal of this book is, therefore, to contribute to a better understanding of the politics of social protection for low-income earners in Latin America and to examine the underlying structures, constellations of actors, and mechanisms. To do so, it engages in a profound long-term analysis of the paradigmatic Argentinian case.

<sup>8</sup> For the purpose of this study, the group of low-income earners is defined as the 40% of the population with the lowest income.

#### **Understanding the Politics of Social Protection for Low-Income Earners in Latin America: Existing Literature, Empirical Gaps, and Remaining Theoretical Puzzles**

While some recent studies have focused on the political processes underlying the expansion of social protections for low-income earners during the 2000s (Garay 2010; Pribble 2013), there are still important gaps regarding the preceding decades. So far, no study has explicitly focused on the long-term politics of social protection for low-income earners. This means that we lack not only empirical information on many reform processes but also a coherent theoretical framework that takes into account the influence of major economic, social and political transformations that unfold over longer periods of time.

Most studies that adopted a long-term perspective, in turn, focused their attention on general social policy regime development. This way, they provided significant insights into the history and political economy of social policy in Latin America. Carmelo Mesa-Lago (1978; 1989) argued that social policy expansion during the 20th century was a gradual process in response to the growing power of occupationally based pressure groups, such as military men, civil servants, and formal workers. Alex Segura-Ubiergo (2007) found that economic development, Import Substitution Industrialization (ISI), the power of left and labor movements, and democracy fostered the growth of social spending. In addition to these factors, Fernando Filgueira (2005) and Stephan Haggard and Robert Kaufman (2008) showed that social policy regime development in Latin America was also often shaped by critical realignments during which powerful elites sought to co-opt or control labor movements, which besides repression usually involved significant social policy concessions.<sup>9</sup> Evelyne Huber and John D. Stephens (2012) and James McGuire (2010) further contributed to the literature by showing that democracy and the power of the political left historically contributed to the implementation of more redistributive social policies.<sup>10</sup> However, all these studies compared a relatively large number of countries and did not focus on the evolution of social protections for low-income earners. Thus, existing studies have improved our understanding of the overall growth of social policy systems in Latin America, but we still lack a deeper understanding of why these systems evolved in a truncated way that disadvantages or even excludes low-income groups from social protection.

<sup>9</sup> Although not focused on welfare regime development, Ruth and David Collier ([1991] 2009) formulated a similar argument in their book on critical junctures and labor incorporation in Latin America.

<sup>10</sup> In congruence with these arguments, Haggard and Kaufman (2008, 79–113) observed that democratic governments in Latin America between 1945 and 1980 were more likely to expand social protections to hitherto uncovered groups while authoritarian governments tended to increase benefits for already protected groups.

Notwithstanding, it is possible to derive a range of hypotheses from the existing literature. Most of the cited historical studies of social policy development in Latin America either explicitly or implicitly assumed that the expansion of social programs was biased towards the middle and formal working classes, because these sectors were better organized, occupied strategically more important positions in the state and the economy and were hence more powerful than low-income groups, such as peasants, the unemployed and informal workers (Barrientos and Santibáñez 2009; Filgueira 2005; Haggard and Kaufman 2008; Mesa-Lago 1978; 1989; Segura-Ubiergo 2007). If we consider, however, that it is also true in most industrialized countries that lowincome groups were weakly organized and lacked political power (Huber and Stephens 2001, 18–19), it becomes evident that the weakness of social protections for low-income earners in Latin America cannot solely be explained by the unequal distribution of power between different popular classes.11 Furthermore, this line of argument has difficulties in explaining why governments in several countries indeed undertook strong efforts to introduce and expand social protections for low-income earners during certain periods. In Argentina, for example, the Peronist governments of the 1940s, 1950s, and 1970s massively expanded social assistance policies, eased access criteria so that most elderly low-income earners could access relatively generous pensions, and built up a universal and free public health system . In Costa Rica, the PLN governments during the 1970s nearly universalized health insurance coverage, significantly extended pension coverage, and set into motion massive social assistance programs (Martinez Franzoni and Sánchez-Ancochea 2013). Under the governments of Eduardo Frei and Salvador Allende in the 1960s and 1970s, Chile massively strengthened the provision of basic health services, education, and housing policies in favor of low-income groups (Arellano 1985). Similar processes also took place in other Latin American countries (Huber and Stephens 2012, 73–102). What were the underlying factors and political dynamics that enabled the expansion of social protections for low-income earners during these periods? And what were the reasons why these processes were unable to lastingly transform the respective social policy regimes into inclusionary institutions?

In their pathbreaking book *Democracy and the Left: Social Policy and Inequality in Latin America*, Evelyne Huber and John D. Stephens (2012) make an important contribution to the effort to resolve these puzzles. Through a mixture of statistical methods and five compact case studies they test a large variety of hypotheses and provide strong evidence in support of the key argument

<sup>11</sup> Largely in accordance with Collier and Collier ([1991] 2009, 788) the term popular classes refers to the urban and rural lower classes, including formal and informal labor, precarious self-employed workers, peasants, the unemployed and the lower middle class. The lower middle class has been included as in Argentina their members often form part of the organized labor movement.

of the power resources approach by showing that left-of-center political parties were pivotal promoters of redistributive social policy. Huber and Stephens (2012) furthermore show that the length of democracy decisively influenced the redistributive characteristics of social policy regimes, as it provided a crucial condition for the development of left party strength and electoral competition for low-income voters. In contrast, right-wing political actors and economic elites pursued more regressive social policy agendas and repeatedly stopped processes of progressive change or even reversed them through posterior retrenchment and privatization. An expansion of social protections for low-income earners can hence be assumed most likely under democratic rule, when left parties attain governmental power or when they are strong enough to generate legislative pressures and to cause significant electoral competition for the vote of low-income earners.

Despite these significant advances for our understanding of the politics of social protection for low-income earners in Latin America, there nevertheless remain several theoretically puzzling questions. One first and broad theoretical puzzle regards the involved constellations of actors and interests, in general, and the formation of popular class identities and actors, in particular. Based on the analysis of industrialized countries, power resources theorists have argued that redistributive and inclusionary welfare states are crucially the outcome of strong working classes represented by trade unions and left-of-center parties. While Huber and Stephens (2012) and Pribble (2013) indeed found clear evidence for a similar role of left-of-center parties in Latin America, the role of trade union movements and non-left popular class-based parties remains still far less clear. McGuire (2010) found that Latin American unions played a rather ambiguous role, at times supporting social policy expansion to low-income earners and at times opposing it. Mesa-Lago (1978) and Haggard and Kaufman (2008) even argued that trade unions mostly defended specific group interests and therefore contributed to the development of exclusionary social policy. In a similar vein, non-left popular class-based parties, such as the Partido Justicialista (PJ) in Argentina or the Partido Revolucionario Institucional (PRI) in Mexico, have led both the progressive expansion of social protections and the posterior regressive retrenchment and privatization (Levitsky 2003). The fact that such non-left parties were the dominant political representation of labor and other popular classes during prolonged periods in several Latin American countries raises the question of whether or in what way the left–right divide can explain the development of social protection in countries where the main parties organized primarily around other divides, such as for example Peronism and Anti-Peronism in Argentina. Furthermore, economically marginalized sectors, such as peasants, informal and unemployed workers are much more numerous in Latin America than in the industrialized countries, yet their role and that of actors representing them remain largely unclear. Regarding the recent process of social policy expansion, Garay (2010) argued that protests and movements of the lower popular classes<sup>12</sup> were indeed important drivers. This puts into doubt the hypothesis that these sectors were generally powerless and unable to influence social policy-making. At the same time, it raises questions regarding their role in earlier processes of inclusionary social policy reform. How can we make sense of these complex constellations of actors and interests and their relations to social protections for low-income earners? Which actors supported and which opposed such policies? Why and under what conditions did they do so?

A second major puzzle concerns the role of contextual factors in the politics of social protections for low-income earners. Here we can identify hypotheses concerning the effects of specific factors in the wider literature, but we lack a coherent theoretical framework which ties together the most relevant factors and points out how these interact and shape the political process. Regarding the politico-institutional context, the controversy turns mainly around the effects of regime types and path dependencies. Huber and Stephens (2012) and McGuire (2010) found that democracy has significant long-run positive effects on the progressiveness and inclusiveness of social policy, while semiauthoritarian and authoritarian regimes favor retrenchment or the development of regressive insurance-based systems.<sup>13</sup> In a similar direction, Garay (2010) and Pribble (2013) argued that democracy was a decisive precondition for the recent expansion of social protections for low-income earners during the 2000s. In contrast, Hans-Jürgen Burchardt (2008; 2010) directed our attention to the fact that this relation cannot be observed during the 1980s and 1990s as formal re-democratization did not coincide with more progressive social policy nor a reduction of inequality. The question of why democracy was associated with redistributive policies during some periods while not during others remains a puzzle.

The role of different structural factors is also contested. Based on Huber and Stephens (2012) it could be assumed that industrialization had a positive indirect effect on social protection for low-income earners, as it facilitated the growth of labor unions and left parties which then pressed for redistributive policies. In contrast, Barrientos and Santibáñez (2009) and Mesa-Lago (1978) argued that industrialization in Latin America led to the development of labor movements that represented only small portions of the overall population and hence contributed to the development of truncated, exclusionary social policy regimes. Huber and Stephens (2012) furthermore argued that higher degrees of globalization and bad economic performance restricted the capacities of governments to implement progressive social policies. High degrees of

<sup>12</sup> The term lower popular classes refers to the economically marginalized segments of the popular classes, which have usually not been represented by the organizations of the trade union movement. These segments include urban and rural informal workers, precarious self-employed workers, peasants and the unemployed.

<sup>13</sup> A discussion and definition of different regime types can be found in chapter 2.

globalization, it is argued, strengthen capital vis-à-vis labor and the state, while bad economic performance limits the resources available for redistribution.<sup>14</sup> At the same time, however, the recent phase of social policy expansion for lowincome groups took place in a context of relatively open markets. While Huber and Stephens (2012) also argued that a high dependence on primary goods exports has negative indirect effects because it strengthens an often aggressively anti-redistributive landed oligarchy , countervails industrialization and increases external economic vulnerabilities; a variety of recent studies that focused on the political and social consequences of resource extraction in Latin America argued that the expansion of social protections for low-income earners during the 2000s was closely related to the increase of primary goods production and exports, which both generated fiscal revenues and required popular legitimization (Burchardt 2016; Gudynas 2012; Peters 2016).

Although these controversies allow for identifying a range of potentially important contextual factors, so far there exists no study that systematically analyzes whether and how these different factors influenced the politics of social protection for low-income earners in the long run, which of the many contextual factors were decisive and how they interacted. Even the few studies that explicitly examine the politics of social protection for low-income earners were limited in their dealing with contextual factors (Garay 2010; Pribble 2013). Due to their focus on the recent phase of social policy expansion, their relatively short timeframes did not allow them to evaluate the effects of institutional, structural, and discursive transformations that occurred over longer periods. Hence, making sense of this multiplicity of potentially influential factors and evaluating their empirical importance still requires substantial research and theory development. What were the contextual factors that contributed to the expansion of social protections for low-income earners during certain periods? What factors allowed for their posterior retrenchment? Which contextual factors were common to different periods of expansion? Which factors consistently differentiated periods of expansion from periods of stagnation or retrenchment? And in what ways did these factors interact and matter in concrete political processes of social policy reform?

The third theoretical puzzle concerns the main mechanisms that characterize the politics of social protections for low-income earners in Latin America. Due to the broad focus on social policy regimes in general and the comparison of a high number of cases, existing long-term studies did not engage in detailed and systematic analyses of the political processes that led to the expansion or retrenchment of social protections for low-income earners (e.g., Filgueira 2005; Haggard and Kaufman 2008; Huber and Stephens 2012; Mesa-Lago 1978; 1989; Segura-Ubiergo 2007). However, such analyses are necessary if

<sup>14</sup> Haggard and Kaufmann (2008) and Segura-Ubiergo (2007) make similar arguments with regard to the overall level of social spending, yet without focusing on the specific issues of redistribution and social protection for low-income earners.

we want to refine our theoretical understanding of *how* different factors shape the political struggle and the subsequent social policy outcomes for the poorest sectors. In contrast, the recent short- and medium-term studies of Jennifer Pribble (2013) and Candelaria Garay (2010) explicitly focused on the expansions of social protections for low-income earners and identified several mechanisms that structured these processes. They agree that democracy was an important driver behind the expansion. The first mechanism was that democracy allowed for the emergence of significant electoral competition for low-income voters and hence provided incentives even for non-left parties to respond, although with minor concessions, to demands for social protection. A second mechanism observed by Pribble (2013) was that democracy provided opportunities for a strengthening of the political left, which in several cases enabled progressive coalitions to reach majorities in parliaments, to lead governments and on that basis implement inclusionary social programs.15 Pribble (2013) furthermore found that programmatic left parties with close ties to base-level social movements more consistently expanded universal social policies than pragmatic and personalistic left parties. The key mechanism identified here was that binding programs and strong civil society linkages reinforce the party's ideological commitment to equality and protect it from erosion in the face of countervailing political and economic pressures. Garay (2010) identified the growth of lower popular class protests and movements as another key driver behind the expansion of social protection for low-income earners. Thereby she observed two main mechanisms. One mechanism was that governments responded to these pressures from below with social policy concessions in order to re-establish social peace or to co-opt, divide or weaken the protesting movements. The other mechanism was that the formation of low-income earners' movements enabled these sectors to participate in the policy-making process through alliances with partisan political forces. Despite these advances in identifying political mechanisms related to the expansion of social protections for low-income earners, there remain puzzling questions. To what extent did such mechanisms also characterize earlier processes of social policy expansion? How did major societal transformations, such as the expansion and decline of industrial employment or the rise and fall of neoliberal thinking, affect the operation of such mechanisms? What mechanisms operated when social protections were not expanded but dismantled or retrenched? And why did several of the above-mentioned mechanisms not work during the 1990s when left or popular class-based parties were strong in several countries and democracy consolidated?

In sum, a variety of hypotheses can be drawn from the existing literature. But there remain puzzling questions and, most importantly, there is still a need for the development of a coherent and empirically sustained theoretical frame-

<sup>15</sup> Huber and Stephens (2012) make a similar argument regarding the mechanisms that shaped general welfare policy expansion in Latin America.

work that ties together the roles and interactions of the main actors, contextual factors and mechanisms shaping the long-term politics of social protection for low-income earners in Latina America.

## **Why, How and What We Can Learn from the Argentinian Case: Case Selection and Research Strategy**

Comparative studies on Latin American social policy regimes have been strong in testing hypotheses derived from existing welfare state theories and in generalizing the findings by showing which arguments possess explanatory power throughout the region and over time (Haggard and Kaufman 2008; Huber and Stephens 2012; Mesa-Lago 1978; Segura-Ubiergo 2007). However, as Esping-Andersen (1990, 2) pointed out, the trade-off related to this effectiveness in theory testing and generalizability is that large-scale comparisons "prohibit detailed treatments of individual countries." Considering that most welfare state theories were once developed based on European experiences, and hence against the background of significantly different historical, economic, social, cultural and political contexts, it is not surprising that after testing these theories in Latin America there remain several puzzling questions and mismatches. In order to solve these puzzles and to overcome these mismatches, it is necessary to move from theory testing to theory building, modification, and specification. This is even more the case for the specific issue of social protection for low-income earners, as this aspect of social policy regime development has not been the explicit focus of any of those theories.

For the aim of theory building, modification, and specification, the indepth analysis of a single but paradigmatic country-case can provide important advantages. While such studies face clear limitations regarding the generalizability of their findings, they can draw great strength from their close focus on and the rich treatment of political processes, which facilitates the discovery of novel explanations, the testing of complex arguments and the refining of existing theoretical claims (Collier and Collier [1991] 2009, 20; Mahoney and Rueschemeyer 2003, 13; Rueschemeyer 2003, 305–336). Theoretically puzzling questions like why labor unions and popular class-based parties supported or opposed social protections for low-income earners at different historical moments, why democracy was in the long-run associated with more inclusionary social policy but why such effects could not be observed during the 1980s and 1990s, and how different ideational influences entered and shaped political processes, can be dealt with much better in a single country study that permits a detailed qualitative analysis of such phenomena. In contrast to large-scale comparisons, the present study will be able to closely trace political processes of social policy reform by reconstructing how events unfolded over time, how different factors and actors triggered or influenced such events and how these were in the end causally related to increases or reductions of the level of social protection for low-income earners. A long timeframe furthermore permits comparisons of different reform periods and discovery of the contextual factors, constellations of actors and political mechanisms that were common to periods of expansion and which consistently differentiated them from periods of retrenchment. Thus, by focusing on one country, Argentina, over a long period of time, 1943 to 2015, the present study aims to further develop and modify existing theories—in particular the power resources approach—and to construct a consistent theoretical framework for the analysis of the politics of social protection for low-income earners in Latin America that can subsequently inform future research on other countries as well as comparative studies.

The Argentinian case is particularly well suited to serve this aim due to both methodological and theoretical reasons. The high frequency and intensity of change regarding social protections for low-income earners and the related political struggles provide a fertile basis for the identification of patterns and causal relations through the application of process tracing and period comparison methods.<sup>16</sup> At the same time, Argentina constitutes a paradigmatic case concerning the theoretical puzzles identified above and is therefore especially suited to contribute to the academic controversy.

First, the political landscape of Argentina contains constellations of actors and social policy interest that cannot be entirely understood with existing welfare state theories and therefore serve to solve a range of related puzzling questions. The Peronist PJ constitutes an emblematic example for a programmatically flexible and ideologically heterogeneous popular class-based party, and therefore Argentina is well suited to explore the role of such parties and to conceptualize the conditions and influences under which these tend to support or to oppose social protections for low-income earners.17 Argentina furthermore has a strong trade union movement that was frequently involved in social policy reform processes and shares with most of its Latin American counterparts that its social bases are nearly exclusively composed of formal workers. Therefore, Argentina allows for digging into the question under which conditions Latin American union movements tended to support or to oppose social protections for low-income earners. Similarly, the country has seen the development of social movements representing non-unionized lower popular classes, such as peasants, the unemployed, and informal workers. The fact that such social movements have been actively or passively involved in social policy reform processes at several historical moments provides a fertile ground

<sup>16</sup> A detailed description of the methodological approach can be found in chapter 3.

<sup>17</sup> According to Levitsky (2003) and Levitsky and Roberts (2011) other examples of such programmatically flexible and ideologically heterogeneous popular class-based parties in Latin America are the Movimiento Nacionalista Revolucionario (MNR) in Bolivia, the Partido Revolucionario Institucional (PRI) in Mexico and the Alianza Popular Revolucionaria Americana (APRA) in Peru.

for the analysis and conceptualization of their role in the politics of social protection for low-income earners. The repeated installation of military governments furthermore allows us to not only contrast the effects of democratic and authoritarian regime types but to also explore the role of the military itself as an active participant in social policy-making.

Second, Argentina can be considered an emblematic case to explore the role of a range of contextual factors that are either still subject to academic controversy or have so far remained outside the scope of research. The country is one of the leading economies of the region, has reached among the highest levels of industrialization and developed extensive state structures which comprise an ample and cost-intensive set of social policies. Therefore, based on existing welfare state theories, Argentina could be considered one of those countries in Latin America in which due to a range of contextual factors broad social protections for low-income earners were most likely to develop. Nevertheless, despite significant progress since the beginning of the 21st century, Argentinian social policies severely disadvantage low-income earners in their access to cash transfers and essential social services, such as health care and education. At the same time, Argentina shares a broad range of contextual factors with the rest of the region, such as widespread informal employment, high inequality, dependence on primary goods exports, the application of ISI development strategies and repeated military coups, among others (Burchardt 2006; Collier and Collier [1991] 2009). Such contextual factors distinguish the region from the industrialized countries that informed existing welfare state theories. Hence, the Argentine case allows us to explore whether and how these specific characteristics of Latin American societies shaped the politics of social protection for low-income earners in order to modify existing theories and to develop new theoretical arguments. The fact that Argentine society experienced comparatively dramatic transformations regarding most of these contextual factors makes the country a particularly favorable case to observe and trace whether and how these transformations affected actors, interests, and political dynamics.

And third, the extraordinarily high number of social policy reform processes that profoundly affected the level of social protection for low-income earners, the intensity of the associated political disputes and the comparatively broad mediatic and academic documentation provide a fertile base for the identification of political mechanisms. The Argentine case is therefore not only well suited to improve our understanding of *what* factors and actors influenced the level of social protection for low-income earners, but also to explore *how*  and *why* they did so.

In addition to the contribution to the theoretical debate, the present study aims to close an important empirical gap in the literature on Argentine social policy. There exists a high number of excellent studies on social policy development in the country.18 The large majority concentrate on single reform processes, specific policy areas or short timeframes. Rubén Lo Vuolo and Alberto Barbeito (1998a), Aldo Isuani (2008) and Laura Golbert (2010) provide outstanding accounts of the long-term evolution of the Argentinian social policy regime. However, as has been shown above, general social policy regime development did not linearly correlate with the evolution of social protections for low-income earners. Indeed, the paradox of social policy expansion during the 20th century in Latin America lies in the fact that social policy expanded without improving the level of social protections for low-income earners in any comparable measure. However, a detailed empirical account of both the evolution of the level of social protections for low-income earners in Argentina as well as the underlying political processes is still missing—a gap that the present study aims to fill.

## **The Structure of the Study**

Chapter 2 starts with an extended theoretical literature review, which complements the short overview given in the introduction by providing additional detail and incorporating theoretically relevant literature on non-Latin American countries. In its main section, the chapter develops a theoretical framework for the analysis of the politics of social protection for low-income earners in Latin America. Chapter 3 describes the methodological approach, the application of process tracing and period comparison, as well as the main sources of primary and secondary information. Chapter 4 constitutes an extensive and systematic reconstruction of the politics of social protection for low-income earners in Argentina between 1943 and 2015. The text follows a coherent structure specifically designed to reveal and analyze contextual factors, constellations of actors, and political mechanisms related to changing levels of social protection for low-income earners. For this reason, the chapter is divided into seven subchapters that each deal with a period of expansion or retrenchment. Each subchapter, in turn, is divided into different sections that outline the economic, social and political context of the period, trace the political processes of social policy change and assess the resulting policy outcomes regarding social protections for low-income earners. Chapter 5 engages in a structured comparison of the different subperiods and reform processes and thereby works out recurrent patterns, causal relations, and political mechanisms. And finally, chapter 6 provides an overview of the main findings, a concise assessment of the initial hypotheses outlined in chapter 2 and a discussion of the implications of these

<sup>18</sup> Given the high number of studies and authors I refrain from citing them here, because any selection would unavoidably appear arbitrary. However, as the present study heavily draws on the empirical information provided by this ample spectrum of secondary literature, a look into the literature list at the end of the book will suffice to grasp the extensive work done mostly by Argentine researchers.

findings for the wider theoretical debate and future research on social policy in Latin America.<sup>19</sup>

<sup>19</sup> The whole study benefits from and draws on my own previous research on social policy development in Argentina (Bossert 2011) as well as on two articles that have been written in parallell to working on the present study and that present some of the information and arguments of the present study in a somewhat different and much more compressed form (Bossert 2016; Weinmann, Bossert and Hecker 2016).

## 2 Theorizing the Politics of Social Protection for Low-Income Earners in Latin America

The political processes behind the inclusion of the lowest income sectors in policies of social protection in Latin America, and, more often, their exclusion from such policies, has seldom been an explicit focus of academic discussion. Further empirical research and theory development are therefore needed to improve our understanding of these processes.

Several studies on Latin American welfare states and political economies have, in fact, dealt with the issue indirectly. Furthermore, different causal claims regarding the level of social protection for low-income earners can be identified in the theoretical debates on both the state and welfare policy in the industrialized countries. These claims are often implicit and not extensively elaborated. In the following chapter, a variety of relevant theoretical approaches will be reviewed. These approaches can be roughly grouped into four categories: actor-centered, structuralist, institutionalist, and ideational approaches.

On that basis, a concrete theoretical framework for the analysis of the political economy of social protection for low-income earners will be established. This theoretical framework elaborates on the main causal assumptions and analytical categories made before the systematic historical analysis of the Argentine case. These assumptions and categories are in part derived from the preceding review of the relevant theoretical literature, but also include proposals for theoretical innovation inspired by contrasting existent theoretical approaches with previous empirical knowledge stemming from several years of studying welfare states in Latin America and Europe.

## **2.1 Review: Welfare State Theory and Social Protection for Low-Income Earners**

The following review of welfare state theories focuses on relevant arguments regarding the level of social protection for low-income earners in Latin America. It will be organized according to the various emphases of different theoretical schools.<sup>20</sup> Due to their influential position in the debate about the

<sup>20</sup> Several authors have previously elaborated overviews of the academic debate on the political economy of the welfare state and distinguished different schools of thought (Bossert 2011; Schmidt et al. 2007; Siegel 2006; Starke 2006). The following review

political economy of Latin American social policy, the review will start with actor-centered approaches. Then approaches emphasizing the role of social and economic structures, institutions and ideas will be discussed.

## **Actor-Centered Approaches**

One of the first and best known systematically comparative studies on Latin American social policy is Carmelo Mesa-Lago's (1978) *Social Security in Latin America.* This account argues that the unequal access to social security has been historically shaped by the differing influences of a variety of pressure groups. The main actors in his analysis are five different occupational groups: military men, civil servants, white-collar workers, blue-collar workers, and peasants. Like other pluralist accounts (see Smith 2006, 23–24), Carmelo Mesa-Lago (1978, 333) assumes that modernization and industrialization processes led to pluralist and increasingly complex societies in Latin America. These socio-economic developments then constituted the social basis for the emergence of multiple pressure groups in the field of social policy. The socioeconomic structure of the society is thereby assumed to shape the power of these groups, and therefore, their ability to attain social protection. Hence, as will be seen later, the pluralist pressure group perspective implicitly builds on socio-economic arguments similar to those made by proponents of the Logic of Industrialism.21 From this perspective, scarce social protections for the lowest income sectors are the result of the lack of power of those subgroups of the labor force that have been weakly organized, work in sectors of low strategic importance and have faced restrictions on electoral rights over prolonged periods (Mesa-Lago 1978, 14). However, the pressure group perspective lacks a broader theoretical argument about the constellations of power and interest in the field of social policy. One of the most problematic aspects of Mesa-Lago's pressure group perspective is that it does not consider the role of capital owners, such as big landowners and the industrial bourgeoisie, in the development of welfare states. In fact, these groups appear to be very powerful and historical evidence from Latin America suggests that they played an important role in welfare state development.<sup>22</sup> Furthermore, in practice, the pressure groups identified by Mesa-Lago seldom acted as independent actors in the politics of social protection. More often they were involved in broader alliances with other groups of workers or even other social classes.

draws on these texts, but elaborates a significantly different perspective that is specifically focused on the issue of social protection for low-income earners in Latin America.

<sup>21</sup> The main arguments associated with the Logic of Industrialism will be reviewed in the following subchapter.

<sup>22</sup> See the following subchapter for the different economic and political functions of the welfare state for different social and economic interests.

A variety of approaches based on Marxist premises, in contrast, regard the politics of social policy as crucially shaped by the dynamics of class struggle (Block 1977; Huber 1996; Huber and Stephens 2001; Korpi 1983; Piven and Cloward [1971] 1972; Poulantzas [1978] 2000; Stephens 1979) Frances Fox Piven and Richard A. Cloward ([1971] 1972, XIII–XIV) argue that the initiation or expansion of social policies can be understood as a consequence of civil disorder and popular mobilization. Claus Offe (1972a; 1984; see also Borchert and Lessenich 2006, 18–19) relates the expansion of social policy to the necessity of state actors to achieve a certain level of legitimacy in the eyes of the popular masses and Nicos Poulantzas ([1978] 2000, 184–185) writes that social policies "directly depend, both for their existence and for their rhythms and modalities, on the intensity of popular mobilization: whether as the effect of struggles or as an attempt to defuse struggles on the part of the state." As a middle-range theory focused mainly on the political economy of social policy, proponents of the power resources approach have elaborated in more detail on the question of how class struggle affects welfare state development and also presented extensive empirical evidence (e.g., Korpi 1983; Huber 1996; Huber and Stephens 2001; Huber and Stephens 2012; Stephens 1979). Therefore, in the following, I will focus on this approach.

Unlike the pluralist approaches, which assume that pressure groups compete on a roughly leveled playing field, the plural society, power resources theorists contend that politics has to be analyzed against the background of structures that provide very unequal initial conditions for the development of power and influence. Due to the dynamics of capitalist systems and the concentration of wealth and the control over the means of production in the hands of a relatively small economic elite, capitalist interests have a decisive advantage and depend much less on organization for political articulation than do the interests of other classes (Huber and Stephens 2001, 13).

Different strands of Marxist literature have theoretically elaborated in more detail on this structural advantage of capitalist interests in shaping state policy. Three main factors can be identified in this literature. First, the existence of close social ties and networks between the economic and political elites provides direct and non-transparent channels for interest articulation (e.g., Miliband 1969).23 Second, the institutional forms of the state and its apparatuses contain biases that favor the prevalence of capital interests over those of other social classes (e.g., Offe 1972a; 1984; Poulantzas [1978] 2000; Jessop 1990; 1999).24 And third, state managers, including both politicians and bu-

<sup>23</sup> It should be noted here, that although Miliband pays close attention to the close ties between political and economic elites, his argument is not limited to this channel of influence. In several parts of the book *The State in Capitalist Society* he also mentions the role of ideological biases and structural constraints.

<sup>24</sup> Examples might be the institutional rules for campaign financing or the control of mass media.

reaucrats, have self-interests in facilitating capital accumulation and economic growth because state revenues—and in the last instance also public support depend to a significant degree on these (e.g., Block 1977, 15; Hay 2006).<sup>25</sup>

For power resources theorists, the varying degrees of advantage of capitalist interests are importantly related to the varying success of labor and other subordinate classes in maximizing their power through organization and the long term effects that this has on the structuring of the state and the economy (Huber and Stephens 2001; Korpi 1983; Stephens 1979). The relationship between the economic and the political elite, the institutional forms of the state and the role of the state in the economy are all importantly related to the capacities of the working class and other subordinate classes to organize and, through this means, to exercise influence on the long term evolution of political, economic and social structures. Hence, through the organization of the subordinate classes, state action becomes "less dependent on dominant, and more responsive to subordinate class interests" (Rueschemeyer, Huber Stephens, and Stephens 1992, 65).

Because the power resources approach assumes that welfare states play a major role in shaping the distribution of wealth and power in society, it is the political strength of the popular classes, and particularly the working class, visà-vis the economic elite, that is related to the development of extensive and redistributive welfare states which provide egalitarian access to social protection (Esping-Andersen [1990] 1998). The political power of the working class is thereby associated with the construction of strong left parties, a centralized union movement and high degrees of unionization. Union centralization appears to be of particular importance for the development of social protection for low-income groups as according to power resources theorists (Huber and Stephens 2001; Stephens 1979) it not only concentrates the resources of the labor movement but also forces the leadership to take a more class-wide view, and hence to demand rather egalitarian social policies that favor the broader working class and their allies and not just certain particularly well-organized or structurally powerful sectors. In contrast, inegalitarian, fragmented and highly stratified welfare systems are assumed to be related to politically weaker, divided working classes and a relative strength of conservative parties and ideologies (Esping-Andersen [1990] 1998). The division of the working

<sup>25</sup> Nevertheless, the advantage of capital may differ significantly over time and country. So, for example, might the strength of the social ties between economic and political elites vary under governments of different colors. The biases of institutions might furthermore vary because of different rules for campaign financing, lobbying, and so on. In addition, the dependence of state revenues on private sector capital accumulation can be lower in countries where historical developments have resulted in a wide range of industries and services under public control and tight capital flow regulations. The same might be said about states that were able to retain control over the exploitation of valuable natural resources such as oil.

class and other subordinate classes into different, competing pressure groups, as assumed by Mesa-Lago (1978), is interpreted from this perspective as the result of a lack of popular class solidarity and organization, and therefore as part of a wider balance of class power favorable to capital.

While recent comparative studies of Latin American welfare states have provided considerable evidence in support of the power resources approach (Huber et al. 2006; Huber, Mustillo and Stephens 2008; Huber and Stephens 2012), other studies have questioned the approach on the grounds that trade unions and labor-based parties often played an ambiguous role with regard to social protection for low-income earners (McGuire 2010).

Feminist welfare state research has emphasized that social policy not only affects class relations and income distributions but also gender relations (e.g., Lewis 1992; Orloff 1996; Sainsbury 1996). The work of Jane Lewis shows that different welfare states influence gender relations in different ways and that some welfare states fulfill a stronger role in reproducing the male breadwinner model than others (Lewis 1992). Furthermore, feminist scholars have argued that welfare state development itself is shaped by gendered relations of power and feminist mobilization. Several empirical studies provide empirical leverage for this argument. Dana Hill's and Leann Tigges' (1995) quantitative study indicates that popular class power is positively related to women's access to welfare benefits, and that gender egalitarian social policy is additionally enhanced by a high share of women in labor unions.<sup>26</sup> Barbara Hobson's and Marika Lindholm's (1997) qualitative analysis of social policy during the 1930s in Sweden shows that the successful mobilization of power resources by feminist actors, both within and outside the labor movement, was crucial to the implementation of gender egalitarian policy during that decade. And Merika Blofield (2006, 10–16) provides evidence for South American and Southern European countries that indicates that left parties are much more likely to become driving forces for the implementation of gender egalitarian policy when feminist movements are strong. Furthermore, women's movements have been found to be overwhelmingly progressive regarding social protection for lowincome groups (Huber and Stephens 2001, 21). This has been related to the circumstance that gender egalitarian social policies, such as universal income transfers, health care, and childcare provision, also benefit the lowest income sectors (Arza 2012c: 11; Huber 2006, 301; Huber and Stephens 2001, 21; Sainsbury 1996, 127–169). Hence, a variety of feminist approaches to welfare state analysis indicate that the level of social protection for low-income earners depends not only on the balance of class but also of gender power, and on the particular relation between popular class and feminist actors.<sup>27</sup>

<sup>26</sup> Based on qualitative research, Diane Sainsbury (1999, 268) comes to similar conclusions.

<sup>27</sup> The relations between popular class and feminist actors might take different forms: a) an actor might be both at the same time, b) popular class and feminist actors might be

#### **Structuralist Approaches**

A variety of theoretical approaches from different academic traditions emphasize the role of economic and social structures in the analysis of the political economy of the welfare state. Despite the common emphasis on structural factors, the literature within this strand is very diverse, ranging from pluralist accounts, such as Harold Wilensky's (1975) *The Welfare State and Equality,* to Marxist accounts, such as James O'Connor's ([1973] 2002) *The Fiscal Crisis of the State,* contemporary contributions to the Varieties of Capitalism (VoC) literature (e.g., Hall and Soskice 2001) and arguments about the effects of globalization processes on the welfare state (e.g., Scharpf 2000; Segura-Ubiergo 2007). Many of these accounts assume a more or less direct determination of social policy by social and economic structures and hence represent a rather functionalist logic. In the following, I will briefly outline these theoretical positions and discuss the importance of some of their arguments for the analysis of the political economy of social protection for the lowest income groups in Latin America.

One of the most influential strands of structural-functionalist thinking in welfare state research up to the 1970s was the so-called Logic of Industrialism. Theorists adhering to this line of argument interpreted social policy as a more or less automatic response to the socio-economic development of the underlying society (Wilensky 1975; for reviews, see Myles and Quadagno 2002; Quadagno 1987). While economic growth and industrialization create new social needs through demographic change and the weakening of traditional family structures, they also generate new wealth and therefore increase the capacity of the state to provide public social security (Schmidt and Ostheim 2007a, 29 – 39). Harold Wilensky (1975), at that time one of the most important advocates of this argument, found in his very influential quantitative comparison of 64 countries a strong correlation between the level of economic development, the relative size of the aged population, the age of the social security system and the level of social spending. On this basis, he argued that the "economic level is the root cause of welfare-state development" (Wilenksy 1975, 47). The state is thereby seen as a neutral entity responding quasi-automatically to the demands of the pressure groups that emerge in the process of economic development due to the spreading of new social risks (Quadagno 1987). This argument stands in sharp contrast to the findings of numerous analyses of Latin American welfare states, which show that these tend to exclude particularly those parts of the population that face the most severe social risks (Goñi, López, and Servén 2008; Lindert, Skoufias and Shapiro 2006; Mesa-Lago 1978; 1989). Hence, the causal logic proposed by Harold Wilensky seems to

largely independent from each other but cooperate closely or c) they might be independent actors without closer ties.

contribute little to our understanding of social protection for the poorest in Latin America.

The analyses of informality undertaken by PREALC, the regional employment program of the ILO, contain a different theoretical argument establishing a direct causal relationship between the level of industrialization and the level of social protection for the lowest income groups in Latin America. According to PREALC's dualistic perspective on informality, the expansion of urban industrial employment in Latin America was insufficient to absorb the rapid growth of the workforce due to the accelerated migration of workers from the countryside. This, so the argument goes, led to a persistent surplus of labor in the urban centers and the emergence of a dual economy with a capital-intensive high-productivity formal sector guided by the logic of capital accumulation and an informal sector in which the marginal masses produce with scarce capital and low productivity in order to secure their own survival. The earnings in the low-productivity sector, it is argued, are so low that social security, and hence the vast part of public social protections, is commonly not accessible to these workers (Portes and Schauffler 1993; Tokman 1987). Hence, from this perspective, the level of industrial employment appears to be directly related to the coverage of the systems of social protection. Although this presents a more plausible argument for a direct causal relation between industrialization and social protection for the lowest income groups, it continues to be problematic. More recent studies show that in Latin America workers in big industrial enterprises also often lack access to social security systems (e.g., Beccaria and Groisman 2009a). Furthermore, a comparison of different countries shows that, depending on the precise rules of access, social security coverage among workers in non-industrial sectors varies significantly (CEPAL 2006, 45; Rofman, Lucchetti and Ourens 2008). Governments are theoretically able to introduce universal social policies that facilitate access considerably. Hence, there is no automatism that directly links the level of industrial employment to the level of social protection for the lowest income sectors. Coverage depends importantly on governmental efforts to formalize employment relations and on the concrete design and implementation of social policies. In light of this, it appears more fruitful to consider the plausible indirect social policy effects that stem from the level of industrialization due to its role in shaping the constellations of power and interest between influential actors.

While the Logic of Industrialism assumes a functional fit between the socio-economic structure and the welfare state, some early Marxist (e.g., J. R. O'Connor [1973] 2000; Offe 1972b) and several contemporary VoC (Iversen and Soskice 2001; Mares 2001; Schneider and Soskice 2009) analyses of social policy assume a relative functional fit between the structure of the capitalist economy and the welfare state. It must be mentioned that neither of these authors proposed a purely structural-functionalist argument. Both Offe (1972b)<sup>28</sup> and J. R. O'Connor ([1973] 2000) acknowledge a certain importance and contingency of class struggle and politics when analyzing the legitimizing role of the welfare state. Iversen and Soskice (2001) and Schneider and Soskice (2009), in turn, consider the role of different electoral systems and voter preferences. Nevertheless, a functionalist logic prevails in the conceptual frameworks of these studies.<sup>29</sup>

Early Marxist structural-functionalist accounts argued that social policy fulfills several necessary functions for the process of capital accumulation (J. R. O'Connor [1973] 2000; Offe 1972b). James O'Connor ([1973] 2000, 138) went so far as to argue that "the fundamental intent and effect of social security is to expand productivity, production, and profits." Among others, these functions include the assurance of the existence of a future workforce through the maintenance and regulation of the non-working and poor population, the legitimization of the process of capital accumulation and the equipment of the workforce with the required skills (Blau 1989, 33–35). If these are necessary functions, then we might ask why different states fulfill these functions to very different degrees and in very different ways. And why do Latin American welfare states provide much less access to public social protection to the unemployed and poor populations than most European welfare states do? A possible hypothesis stemming from these arguments would be that capital accumulation in most Latin American countries is less dependent on social policies that protect low-income earners due to the greater reliance on the exploitation of natural resources, a structural over-supply of labor and lower skill requirements of most Latin American industries.

In this regard, recent contributions to the VoC literature appear interesting. While the outlined Marxist approaches analyze the role of social policy from a macro-perspective centered on the requirements of the wider process of capital accumulation, the VoC literature is more micro-level oriented and focuses on the requirements and preferences of different firms and economic sectors. Isabela Mares (2001) and Ben Ross Schneider (2009) argued that different economic sectors and types of firms have significantly different interests with regard to social policy. Skill-dependent business sectors require extensive educational systems, and generous social security systems provide the necessary incentives for employees to invest in specific skills that might be significantly devalued in the case of job loss (Hall and Soskice 2001; Iversen and Stephens 2008). Furthermore, the reduction of social inequality through egalitarian welfare states has positive indirect effects on the skill levels of the population, particularly regarding lower income groups (Huber and Stephens 2001). As a

<sup>28</sup> In later essays Offe moved towards a stronger emphasis on the role of class struggle (see e.g. Offe 1984).

<sup>29</sup> For similar assessments made by other authors, see Myles and Quadagno (2002, 37) and Quadagno (1987, 114).

result, economic sectors and types of firms that depend on an extensive skilled workforce might benefit more from social protection for the poorest parts of the population than those with low skill requirements. On the basis of such arguments, it could be assumed that different economic structures and positions in the global division of labor are related to considerably different constellations of capitalist requirements with regard to social policy. Ben Ross Schneider's (2009) attempt to apply the VoC approach to the Latin American context supports this assumption. He argues that the compositions of Latin American economies, as well as the ways firms relate to each other and to the world market, are associated with considerably lower skill requirements than in Europe, North America, and East Asia. Considering moreover the persistent excess of labor (Portés and Schauffler 1993; Tokman 1987), Latin American capitalism seems to depend much less on welfare state functions such as the provision of quality education, health care and income maintenance to the lowest income groups in order to ensure the maintenance of a sufficiently abundant and skilled workforce.

The capital-analytical perspectives of the mentioned neo-Marxist scholars and the VoC approach provide some important insights into the role of social policy for the process of capital accumulation. Interestingly, with regard to Latin America, recent studies seem to coincide with the power resources approach in assuming that employers tend to have interests opposed to egalitarian social policies (Esping-Andersen [1990] 1998; Schneider 2009). However, due to their strong focus on the functioning of capitalism, they tend to underestimate the ability of popular struggles to influence state policy (Poulantzas [1978] 2000; Stephens 1979). Furthermore, because of the neglect of agency, the functionalist approaches lack convincing arguments of how the functional requirements of capitalism translate into concrete policy decisions (Quadagno 1987, 115).

Another heterogeneous strand of literature, which could also be broadly labeled capital-analytical, has focused on economic globalization and its effects on social policy. Economic globalization has commonly been associated with three main processes: increases in international trade, increases in the mobility of production, and increases in international financial flows (Green-Pedersen 2004). These processes, it has been argued, have led to a growing exposure of national economies to international competition. This, in turn, has been supposed to generate pressures on national governments to adapt their welfare states to the new context of more globalized forms of capital accumulation in order to not endanger economic growth. However, there is striking disagreement on whether and how welfare states affect the standing of national economies in the global markets (Gough 2000). Some authors have argued that globalization made generous and redistributive welfare states unsustainable. According to their logic, welfare states increase costs for firms and therefore reduce the competitiveness of locally produced goods and services; they provide incentives to move production to cheaper countries; and they potentially contribute to unbalanced budgets, which are associated with negative reactions on the global financial markets (e.g., Rhodes 2000; Scharpf 2000; Strange 2003). Others have, however, claimed the opposite to be the case. Economic openness, it has been asserted, tends to increase economic insecurity for workers, and hence potentially fosters social conflict and strengthens the appeal of trade unions and left parties (Green-Pedersen 2004). Therefore, in order to sustain the necessary stability of both work relations and consumption, open economies have to invest more in social policy than relatively closed, domestic market-oriented economies. Furthermore, international competitiveness, particularly on the technological and reliability levels, benefits from investment in education and social security. It is thus argued that it is no coincidence that the open Scandinavian economies possess generous and egalitarian welfare states and are at the same time highly successful on the world market. Another group of scholars has argued that there is no conclusive empirical evidence for either of the two arguments outlined above (Kite 2004; Swank 2003). Referring directly to Latin America, Alex Segura-Ubiergo (2007, 264–265) observes a direct negative relationship between the level of foreign trade and the level of social spending. In his theoretical interpretation of this observation, he largely supports the first of the arguments outlined above, claiming that in the face of globalization generous welfare states have become increasingly difficult to sustain in Latin America. I am, however, highly skeptical about his quantitative analysis and the interpretation of the results. First, his quantitative analysis includes only a limited number of control variables and covers a period (1973–2003) for which a variety of alternative variables constitute likely causes of the decline or stagnation of social expenditure. It would be equally plausible to relate the stagnation of social spending during this period to, for example, the hegemonic status of neoliberal thought, the reiterated economic crises, or the pressures of International Financial Institutions. Second, his regression analysis might possibly have yielded very different results had it been undertaken a few years later, as during the last decade both social expenditure and foreign trade increased significantly in most Latin American countries.<sup>30</sup>

The contradictory evidence regarding the relationship between various quantitative measures of trade and financial openness, on the one hand, and social expenditure, on the other, indicates that the relationship between globalization and the welfare state might be too complex to grasp by simply putting quantitative measures in relation to each other. It is plausible that it is the qualitative aspect of world market participation, meaning the kind of integration into the international division of labor, that influences political processes in different ways and thereby leads to social policies of different kinds and magnitudes. Increases in foreign trade might have completely different effects on

<sup>30</sup> See CEPALSTAT and World Development Indicators.

the political economy of social policy when they are based on high-tech exports and primary good imports, as has for example been the case in Sweden, than when they are based on the export of primary products and the import of industrial goods, as has been the case in several Latin American countries. Although dependency theorists have not been interested so much in the political economy of social policy, they have explicitly dealt with the qualitative nature of the international division of labor and their political, economic and social consequences in Latin America (e.g., Cardoso and Faletto [1969] 1977; Senghaas 1974; Serra and Cardoso 1978). Their arguments inspire therefore a more complex understanding of the relationship between global economy and national social policy towards the lowest income groups. The different kinds of insertion into the international division of labor of Latin American and of countries of the center has been associated with the development in these countries of very different socio-economic structures *and* sectoral compositions of their economies (Cardoso and Faletto [1969] 1977; Senghaas 1974). These structural differences are in turn likely to result in different constellations of power and interest regarding social protections for low-income earners. The arguments of the dependency approach can be combined with the previously discussed arguments of structuralist Marxist and power resources theories, thereby helping to overcome the limitation to the national level. On the one hand, economic sectors focused on the extraction of primary goods and lowtech consumer industries have a bigger weight in peripheral economies than in countries of the center, which might reduce pressures on peripheral country governments to extend social policies aimed at the maintenance and education of the labor force. On the other hand, in part due to their role as primary goods suppliers on the international level and their correspondingly limited degrees of industrialization, peripheral countries tend towards an over-supply of labor and a high level of fragmentation of the popular classes, which constitutes a significant barrier to the construction of popular class power and solidarity between different segments of it (Senghaas 1974).

## **State- and Institution-Centered Approaches**

A variety of different theoretical accounts emphasize the role of institutions and state actors in shaping the politics of social policy. Influential arguments have been formulated regarding the role of institutional veto points (Immergut 1990; 2010), the role of regime type (Burchardt 2008; 2010; Haggard and Kaufman 2008; Huber and Stephens 2012; Segura-Ubiergo 2007), the role of past policies in path dependent social policy trajectories (Collier and Collier [1991] 2009; Pierson 2001), and the role of state bureaucracies as actors in shaping the direction of social policy (Martínez-Franzoni and Sánchez-Anconchea 2012; Rosenberg 1979; 1983).

That the number of institutional veto points plays a central role in the politics of the welfare state has become a widespread argument in the literature. The original argument, developed by Ellen M. Immergut (1990), was that different institutional structures in France, Sweden and Switzerland provided medical organizations with varying capacities to veto healthcare reform. Hence, due to different institutional structures, certain actors were able to effectively veto reform in some countries but not in others. Evelyne Huber, Charles Ragin and John D. Stephens (1993) shortly later incorporated variables for potential veto points, such as federalism and bicameralism, in their quantitative analysis and attempted to develop on this ground more general arguments about the role of the number of veto points in the politics of social policy. Their regression analysis indicated that the considered veto point variables were associated with a slower expansion of social policies in the long run. This finding was explained with the causal argument that the more veto opportunities the institutional structure provides, the more difficult is the passing of social legislation. Following this argument, the lack of social protection for the lowest income sectors in Latin America could potentially be the result of a high number of veto points. However, the opposite might also be assumed to be the case: veto points could also slow down or impede cutbacks and regressive restructuring of social policy. In a newer article, Ellen M. Immergut (2010, 240) states that during the last two decades veto points have weakened or slowed down welfare state retrenchment in Latin America but not in Europe. It is my position that these general arguments are not particularly helpful for the subject of this study. It is theoretically meaningful that more veto points tend to slow down political change. But without an argument regarding the kind of bias produced through these veto points, there is little reason to assume that they are in the long-term associated with either more or less social protection for the lowest income groups. Indeed, quantitative estimates of Jennifer Pribble (2008, 57) indicate that federalism and bicameralism have no significant effect on social protection for the lowest income groups in Latin America. This supports the argument that what matters is not simply the number of potential and actual veto points but how concrete institutional settings influence the constellations of power and interest and the choices of actors with regard to social policy. It is thus the question of which institutional structures and practices systematically favor or disadvantage actors and interests associated with an extension of social protection to the poorest sectors that we must look at.

More theoretically meaningful in this sense is what has been referred to as regime type in Stephan Haggard and Robert R. Kaufman's (2008) book about welfare state development in 22 developing countries. Regime type refers to the basic question of whether a political regime is democratic, authoritarian or something in between. There are several theoretical reasons why democratic regimes can be assumed to be more likely to respond to the social policy interests of the lowest income groups than their authoritarian equivalents. Democracy is commonly associated with the existence of certain civil liberties that provide opportunities for underprivileged sectors to legally organize in parties, trade unions, and social movements. These organizational structures then increase the capacities of the subordinate classes to pressure for the expansion of social protection. Even in the case that such processes of organization do not occur to a bigger extent, democracy nevertheless provides incentives for politicians to use the extension of social policies to win the votes of the numerically strong poor sectors of the population (Haggard and Kaufman 2008, 362; Pribble 2008, 36). As Rueschemeyer, Huber Stephens, and Stephens (1992, 66) point out, whether a government is legitimate or not in the eyes of the popular classes only makes a difference to those who rule to the degree that the subordinate classes possess some power. The popular classes can achieve such power through organization, but also through democratic institutions that provide them with a certain political weight due to their numerical strength. However, due to the varying preconditions for organization, associational power tends to be unequally distributed among and within the popular classes in Latin America. Industrial and public sector workers have reached considerable associational strength in most countries while informal, unemployed, and agricultural workers have most often lacked strong organizational representation. In the face of these imbalances of power within the popular classes, authoritarian regimes in Latin America felt the need to achieve a certain degree of legitimization only in the eyes of some segments of the working class. This was often pursued through the implementation or deepening of social insurance policies covering those segments. At the same time, authoritarian governments usually felt little need to legitimize their rule in the eyes of weakly organized lowincome groups, and hence these segments remained frequently excluded from the fruits of these social policy concessions (Pribble 2008, 36–38). While such imbalances of organizational power tend to persist under democratic rule, the right to vote provides the lower income groups with at least a certain degree of electoral power. Furthermore, democracy provides organized labor with additional incentives to seek the support of other popular classes, such as peasants and self-employed workers, and less organized segments of the working class in order to increase their chances of reaching a majority of the votes (Esping-Andersen [1990] 1998, 67; Przeworski 1985, 74–75).<sup>31</sup>

In sum, there are important theoretical reasons to assume that democracy affects both the constellations of power and the constellations of interest in ways favorable for the expansion of social protection for the lowest income groups. Empirical findings from Latin America are nevertheless mixed. On the

<sup>31</sup> As Esping-Andersen ([1990] 1998, 67) argues, it is no coincidence that the breakthrough of democracy and the pursuit of universalist social policy on the part of labor parties coincided in many Western countries. My previous research on the Argentine welfare state also indicates that democracy increases the probability that labor movements will try to win the support of the lowest income groups (Bossert 2011).

one hand, several long-term quantitative and qualitative studies found empirical evidence that democratic governments in Latin America were on average considerably more responsive to the social policy interests of the lowest income groups (Haggard and Kaufman 2008; Huber et al. 2006; Huber, Mustillo, and Stephens 2008; Huber and Stephens 2012; Pribble 2008; Segura-Ubiergo 2007). On the other hand, Hans-Jürgen Burchardt (2008; 2010) points out rightly that the spread of liberal democracies in Latin America during the 1980s did not coincide with a reduction of social inequality nor a significant expansion of social protection for the lowest income groups. As a consequence of neoliberal adjustment policies, inequality even increased in most countries during the 1980s and 1990s and social insurance coverage tended to decline. Only from the early 2000s on can a significant reduction of income inequalities be observed. Hence, there is at least no automatism that links formal democracy to an expansion of social protection for low-income groups. It is therefore necessary to further specify under what conditions democracy tends to lead to more egalitarian social policies and under what conditions such effects might be blocked or outplayed by cross-cutting pressures.

Path dependency approaches form another influential line of arguments with an important emphasis on institutions and policy legacies in the politics of social policy. The larger path dependence argument, as developed in David and Ruth Berin Collier's ([1991] 2009) influential book *Shaping the Political Arena*, focused on the broader long-term transformations of the political arenas in eight Latin American countries that were related to the political incorporation of labor during the first half of the 20th century. In more general terms, the argument is that periods of intense political change, so-called critical junctures that comprise both specific state policies and the reactions of different actors to it, can significantly shape long-term political developments. Therefore, politics cannot be considered as "simply fluid, constantly responding to socioeconomic change" (Collier and Collier [1991] 2009, 11). Rather, politics is also driven by "an autonomous logic and vested interests" that "may be resistant to such change over significant periods of time" (Collier and Collier [1991] 2009: 11). Hence, while Collier and Collier clearly acknowledge the importance of the economic system and the inherent class contradictions in shaping interest and power constellations, they emphasize the role of state policies and historical experiences of actors in determining the historically and spatially specific forms of class formations and political systems. Some scholars of the political economy of social policy in and outside Latin America took up these arguments and provided significant evidence that social policy development has a strong tendency towards path dependency. The central argument in the welfare state literature is that social policies, and in particular social insurance schemes, create powerful vested interests that significantly increase the political costs of diverging from the current social policy path for governments (Haggard and Kaufman 2008, 196–197; Huber and Stephens 2001, 22; Pierson 2001, 447; Pribble 2008, 5). Furthermore, as Christopher Abel and Colin Lewis (1993, 15) argue, social policies in Latin America have often had significant influence on the formation of the popular classes, and in particular the working classes, as political actors.

Recent social protest movements illustrate the politically and socially divisive uses of social security legislation and welfare provision. The practice of selective incorporation inhibited the development of working-class consciousness and solidarity. Pre-existing differentials along lines of occupation and nationality could be exacerbated by differential access to housing, food subsidies and recreational facilities. Welfare provision has functioned both as an instrument of an offensive against labor and as a mechanism in processes of the reconstitution of the institutional base of labor.

Hence, from the perspective of the path dependence approach, it can be hypothesized that the legacy of fragmented and divisive social policies in most Latin American countries constitutes a significant difficulty for the realization of egalitarian reforms that would extend social protection to the lowest income groups. Recent qualitative comparative studies of Juliana Martínez Franzoni and Koen Voorend (2009) and Evelyne Huber and John D. Stephens (2012, 255) find significant evidence in support of this hypothesis.

Another frequent line of argument in the welfare state debate emphasizes the role of state bureaucracies, and particularly welfare state bureaucracies, as influential and relatively independent actors in the politics of social policy (for a short and critical overview see Huber and Stephens 2001, 21). Such arguments have, for example, repeatedly been made regarding the expansion of social insurance coverage in Costa Rica (Martínez Franzoni and Sánchez-Anconchea 2012, 34–35; Rosenberg 1979, 129–130; 1983: 183–187). In this case, it has been argued that the expansion of social policy has occurred largely in the absence of popular pressures. Instead, the social insurance bureaucracy itself acted repeatedly as a pressure group to promote the extension of social policy coverage (Rosenberg 1979, 30; 1983, 184). Juliana Martínez Franzoni and Diego Sánchez-Anconchea (2012, 34–35) have furthermore proposed an argument explaining why the influence of bureaucrats on policy-making is likely to be high. First, politicians seldom drive the design of policy proposals, but usually leave this task to bureaucrats. And second, bureaucracies concentrate abundant technical knowledge and possess direct access to political leaders at the center of the policy-making process. Nevertheless, critics have argued that although bureaucrats might be present in the policy debate and undertake the actual writing of policies, the extent to which they can influence the medium- and long-term development of policy should not be overestimated. Factors such as the aims of the respective governments and the distribution of seats in the parliament profoundly limit the room for maneuver of bureaucracies. In addition to that, a central problem of these arguments for the analysis of the politics of social protection is that they lack a theoretical assumption with regard to the direction of the bureaucracies' influence (Huber and Stephens 2001: 21). The specific biases of different welfare bureaucracies seem to be the product of the concrete historical development of the respective apparatuses and the resulting internal rules, routines, and personnel. Hence, the role played by bureaucracies might be an essentially empirical question with restricted possibilities of theoretical generalization.

## **Ideational Approaches**

Arguments about the important role of ideas have recently become more frequent in the academic debate on the political economy of the welfare state (e.g., Béland 2007; Bleses and Seeleib-Kaiser 2004; Blyth 2002; A. Deacon 2000; Wincott 2011).<sup>32</sup> Ideas, so the argument goes, may have a very direct impact on social policies by shaping what governments consider appropriate. They may also have indirect effects on policy by influencing the wider public or other actors who can exercise pressure on the government. While more materialist approaches have usually emphasized the role of structural factors on welfare politics, authors adhering to ideational approaches have argued that the effects of structures are discursively mediated by ideas that shape how these structures are perceived and what kinds of responses are thought to be possible and desirable. A simple example will help to clarify the different positions. Materialist approaches have, for example, argued that structural changes in the economy such as a decline of growth and related fiscal constraints (Haggard and Kaufman 2008; Huber and Stephens 2001; Pierson 2001) or the increase of foreign trade (Segura-Ubiergo 2007) were the driving forces behind welfare state retrenchment. In contrast, authors who stress the importance of ideational factors have argued that the perceptions of governments and the wider public that such trends constitute problems that require solution, and even more so perceptions that welfare state retrenchment would be the adequate or even the only viable response, are not exogenous to politics (Hay 2002, 114–115). Such perceptions are not given by nature but related to certain ideas of how the economy and society work. An example of this is the idea that redistributive tax and welfare systems reduce the competitiveness of a national economy. Such ideas, in turn, are socially constructed and tend to change over time and differ across space. Therefore, proponents of ideational approaches for the study of welfare state change argue that the impact of relevant ideas should not be omitted from the analysis.

Two types of studies can be roughly distinguished with regard to the consideration of ideational factors in welfare state politics. On the one hand, there is a variety of studies that were concerned with relatively specific ideas that

<sup>32</sup> The discussion of ideational factors here and later draws in parts on very similar arguments made in my previous work (Bossert 2011).

influence policy-makers, such as politicians and bureaucrats, and the wider public with regard to the design of particular social policies. Alan Deacon, for example, analyzed the influence of different ideas on New Labour's social policy agenda during the 1990s in the United Kingdom (2000). Two of these influences were the ideas of the conservative academic Charles Murray, who argued that income maintenance programs for unemployed persons and low-income earners provided perverse incentives, meaning that these kinds of programs keep people from working, and the ideas of Lawrence Mead, who argued that long-term unemployment destroys the work ethic of the affected persons and that therefore the state should tie social assistance to certain conditions. Corina Rodríguez Enríquez and María Fernanda Reyes (2006, 8–9) find that very similar ideas have influenced social policy makers in Argentina and shaped most targeted social programs between the second half of the 1990s and the early 2000s. Contrary to these conservative ideas that dominated social policy during the 1990s in Argentina, Claudia Danani and Alejandra Beccaria (2011, 147) find that pension reform since 2004 was more strongly influenced by ideas that social security should be universal rather than targeted and constitute a right rather than a premium for merit. In a similar vein, Juliana Martínez Franzoni, Diego Sánchez-Ancochea and Héctor Solano (2012, 35) argue that technocrats in several international organizations and a growing number of social security bureaucrats in Latin America have recently recognized the advantages of universalism.

On the other hand, a number of analyses have focused on the role of ideas at a higher level of generality. Studies of this type have not concentrated on specific ideas but on wider overarching ideational frameworks (Wincott 2011, 144). Such overarching frameworks have been referred to as policy paradigms (Hall 1993, 290) or discursive paradigms (Jessop 1999, 52). Paradigms function as lenses through which actors conceptualize their environment (Hay 2001). While studies focusing on more specific ideas commonly concentrated on one or very few crucial actors, studies concerned with discursive paradigms are usually more interested in the broader consequences of paradigms for the political economy. Discursive paradigms influence usually not only governments and bureaucracies but tend to influence the views of a wide array of actors such as parties, trade unions, employers' associations, and the media. The opinions and perceptions of the wider public are also subject to their influence, and hence discursive paradigms are likely to have a profound impact on electoral politics. Moreover, discursive paradigms are often not restricted to one policy area but provide logical frameworks for a variety of fields, such as economic, social and fiscal policy. A broad body of literature on Latin American social and economic policy suggests that discourses about development constituted such discursive paradigms (e.g., Cortes and Marshall 1999; Haggard and Kaufman 2008; Lewis 1993; Lo Vuolo and Barbeito 1998a; Riesco 2007; Sikkink 1991; Sottoli 2000; Svampa 2005). Susana Sottoli (2000, 11), for example, found in her cross-national comparative study that neoliberal thinking guided both economic and social policy reform in Latin America. Specific ideas such as those dealt with by the first type of ideational studies are usually not entirely independent of such broader paradigmatic influences. The latter frequently provide the basic logic within which more specific ideas develop. Many authors have argued, for example, that the particular ideas that influenced social policy reform in different areas during the 1990s, such as pension, health and assistance policies, have all been anchored in a wider neoliberal economic logic (e.g., Cortes and Marshall 1999; B. Deacon 2005; Kay 2000; Lloyd-Sherlock 2005; 2006; Lo Vuolo and Barbeito 1998a; López-Ruiz 2008). Most of these accounts argue that neoliberal influence has pushed social policy in an even less redistributive and more exclusionary direction.<sup>33</sup>

## **2.2 Theoretical Framework**

## *2.2.1 The Power Resources Approach and Latin America: Strengths, Weaknesses and Promising Modifications*

My theoretical framework draws on concepts from several of the above-reviewed theories. Nevertheless, the main theoretical contribution to the field of welfare state research will be to further develop, specify and modify the power resources approach. The theoretical propositions made in this part are inspired by both theoretical discussions and previous empirical work on social policy development in Latin America.<sup>34</sup>

## **The Power Resources Approach: A Framework for Analyzing Distributive Struggles**

Due to several reasons, the power resources approach is particularly suitable for the purpose of my study. First, among the reviewed approaches, it is the one most directly concerned with the question of which kinds of constellations favor social policy inclusion of low-income earners and which tend to produce

<sup>33</sup> Alesina, Glaeser and Sacerdote (2001, 247) argue in a similar vein for the USA that the dominance of a liberal worldview is one of the factors of crucial importance for understanding the lack of redistributive social policies.

<sup>34</sup> Important sources of inspiration in this regard have also been the various theoretical discussions with Hans-Jürgen Burchardt, Nico Weinmann, Paul Hecker, Ezquiel Bistoletti, Johanna Neuhauser, Philip Fehling, Jenny Jungehülsing and Sebastian Matthes at the Graduate School of Global Social Policies and Governance at the University of Kassel.

fragmented, inegalitarian outcomes. This is due to the strong focus of the power resources approach on explaining distributive and redistributive effects of social policy.<sup>35</sup> Second, another strength of the approach is that it pays great attention to both the actors and the societal structures that shape the political processes of social policy change. Hence, the approach is able to integrate a variety of concepts from the above discussed theoretical schools and to provide a framework for a comprehensive analysis. And third, the approach has proven capable of operationalizing the theoretical claim that social policy development is highly dependent on the balance of power in society. This implies a significant advantage over other power-centered approaches such as Poulantzas' state theory, which proposes similar arguments but contains few concrete instruments for an empirical analysis.

While the power resources approach in welfare state research is mainly focused on the political economy of the redistributive character of social policy, and therefore constitutes a middle range theory, it can be linked to the wider state theoretical debate and shares a number of assumptions with broader theories. Hence, the adaptation of the approach to the Latin American context will prove useful not only for future social policy research but also for analyzing a variety of other issues that are related to state policy and the distribution of resources and opportunities. In contrast to pluralist perspectives and in commonality with Marxist state theory, the approach assumes that politics in capitalism do not take place on a leveled playing field. Rather, the general dynamics of capitalism and the related concentration of wealth and control over the means of production provide capitalists with a systemic advantage for developing influence on policy (Huber and Stephens 2001, 13; Miliband 1969; Wright 1994, 94–95). Yet, power resources theory does not conceive the state as an instrument of particular capitalists (Korpi 1983, 9). While it acknowledges the significant capacity of capitalists and their representatives to mobilize resources in order to influence public policy, the approach assumes that "the primary mechanisms by which the rule of capital" is secured is "through the structural dependence of the state on capital, that is, the dependence of state policy makers on capital's willingness to invest, and through hegemony, ideological domination" (Huber and Stephens 2001, 327). At the same time, power resources theory argues that capital interests are by no means the only interests shaping public policy. Instead, it emphasizes that state policy is shaped by the *relations* of power between contending interests. This means that other interests, such as those of working-class organizations or women's movements, can also influence public policy when they are able to develop and mobilize sufficient power resources (Huber and Stephens 2001, 13). In this regard, the power

<sup>35</sup> By explicitly mentioning both distributive and redistributive effects, the approach refers to the fact that social policy not only shapes the way in which public expenditure is allocated but also the way labor markets function and the capacity of trade unions and individual workers to negotiate wages and other terms of the work relationship.

resources approach is largely congruent with Nicos Poulantzas' ([1978] 2000, 123–145) conceptualization of the state as "the condensation of a relationship of forces." With regard to the distributive and redistributive characteristics of social policy, the power resources approach argues, as Nicos Poulantzas ([1978] 2000, 184–185) and Fred Block (1977, 20–24) do, that it is the power and action of popular class actors and their allies that provide the major impetus for the implementation and expansion of progressive welfare policies.

#### **Critique and Further Development of the Power Resources Approach**

There are also weaknesses of the approach that I will address with modifications. The first weakness to be addressed is the Eurocentric bias of the traditional power resources approach in welfare state research. This bias is in large part due to the fact that the approach originated from reflections on European experiences of social policy development (see e.g. Stephens 1979, Korpi 1983). The European experience took place under structural and historical circumstances and with constellations of actors that were significantly different from those found in other parts of the world, and so several assumptions particularly about the role of different types of actors—should not be transferred uncritically to Latin America.

Due to the peripheral and dependent character of most Latin American economies, formal working classes have usually constituted a considerably smaller part of the overall population. This comparatively stronger socio-economic heterogeneity of the popular classes can be assumed to have impacted both the relative strength and the perceived interests of different popular class actors. In Latin America, labor organizations, particularly trade unions, have often represented only the formal working classes. Informal workers, and hence most of the low-income earners, have in contrast faced considerable difficulties in attaining trade union organization. Whereas social democratic or communist tendencies largely dominated working-class organization in Europe, these influences have been considerably weaker in many Latin American labor unions where different forms of populist movements often constituted the dominant force. In a comparative study, James McGuire (2010, 303–304) finds that most Latin American labor movements have in practice been rather contingent supporters of social policy expansion to low-income earners. Hence, the following theoretical framework will not assume that trade union movements *per se* support egalitarian social policy but elaborate on the conditions under which they do so.

The power resources approach faces a similar problem when dealing with Latin American party systems. Here, in contrast to most European cases, working and other lower classes were often represented by non-left populist parties. The case of Argentina is emblematic. The PJ traditionally captured most popular class support, yet ideologically speaking it cannot be categorized as a genuinely left or center-left party. In a similar vein, the second major party, the Unión Cívica Radical (UCR), which traditionally captured important parts of the middle-class vote, also oscillated between a center-left and center-right orientation. Hence, the framework will deal with the conditions under which these centrist parties tend to engage in welfare state expansion towards low-income groups.

Another important issue is that the focus on working-class organization of the traditional power resources approach is too narrow for the Latin American context where organization around other identities often played a key role in expanding social protections. At times, unemployed, indigenous, and peasant movements where key drivers of progressive policy and hence their role will be taken into account in my analytical framework.

 The more differentiated perspective on the role of working-class actors and the broadening of the view to consider other popular class actors also forms part of my attempt to address a second, more general critique that the approach deduces social policy interests too narrowly from the class structure and the general dynamics of the capitalist system (Kulawik 1996; Levitsky and Mainwaring 2006). Feminist scholars have rightly pointed out that the formation of working classes as political actors was frequently accompanied by exclusions along lines such as gender, race, nationality, religion, and skill (Kulawik 1996; O'Connor 1998; Orloff 2009; Rose 1997). Most labor organizations have been male-dominated, and at times, have pursued a gender-specific anti-universalism in social policy (Kulawik 1996, 58). Institutionalist analyses have furthermore stressed that state policies have an important influence on the formation of actors and their interests (Collier and Collier [1991] 2009; Levitsky and Mainwaring 2006). In order to address these critiques, my theoretical framework will consider a range of further factors that interact with class in the formation of social policy interest constellations, such as cross-cutting cleavages (e.g. gender, race, nationality, religion, skill), institutions, path dependencies, and ideological influences.

A third issue that I consider problematic is that the traditional power resources approach implicitly views the state as a more or less unitary block whose actions depend on the forces that are controlling government and parliament (see e.g., Esping-Andersen [1990] 1998; Huber and Stephens 2001). This does not fit well with my observation that, for example, the ministries of labor and economy often pursued strongly diverging social policy agendas in Latin America. Poulantzas ([1978] 2000, 132–145) argued that the state is traversed by internal conflicts which often lead to contradictory actions being pursued by different state apparatuses. Portantiero (1988, 113–114) developed a similar argument in his theoretical work on the specificities of the state in Latin America. The internal conflicts of the state, in turn, reflect societal conflicts and relations of power, yet not in an immediate and mechanical way. Due to path dependency effects, the adaptation of state apparatuses to changing societal power relations can occur with a significant delay (Poulantzas [1978] 2000, 134). I consider Portantiero's and Poulantzas' view as more accurate in this regard and hence propose to modify the power resources approach in this direction.

Finally, a fourth important weakness of the traditional power resources approach in welfare state research is its very narrow focus on associational power resources, which has proven unable to adequately capture certain shifts in the relations of power that underlay the transition from welfare state expansion to retrenchment in many countries from the 1970s onwards. Earlier studies by power resources theorists on post-war welfare states up to the late 1970s provided strong quantitative and qualitative support for the argument that labor's associational power resources in the form of a recurrent participation of left parties in governments constituted the key determinant of the distributional character of social policy (Huber, Ragin and Stephens 1993; Korpi 1983; Stephens 1979). However, newer studies have found much more moderate or even no partisan effects for the period between 1970 and 2000 in both Latin America and the OECD countries (Huber and Stephens 2001; Huber, Mustillo and Stephens 2008). Why is this so? My hypothesis is that there was a shift in power towards proponents of social policy retrenchment that was not merely based on the associational weakening of progressive actors, but also on a broader shift of the societal discourse towards neoliberalism, a shrinking of the industrial sector and profound transformations of the labor markets in most countries. In order to address this issue and to capture such shifts in the power relations, I conceptualize four types of power resources instead of just one: discursive, structural, institutional and associational power resources. This modification is strongly inspired by theoretical innovations that took place in the fields of labor studies and feminist social policy analysis (Brinkmann et al. 2008; Dörre 2011; D. McGuire 2013; Hobson and Lindholm 1997; Schmalz and Dörre 2013; 2014; Silver 2003; Urban 2010; 2013; Wright 2000).<sup>36</sup>

Taken together, these modifications will improve both the explanatory capacity of the approach in general and its applicability to the Latin American context in particular.

<sup>36</sup> Hans-Jürgen Urban (2013) rightly points out that there is a surprising independence between the power resources approach in welfare state research and the power resources approach in trade union revitalization studies. One of the theoretical innovations of my approach will be to break this independence and to adapt and incorporate new insights and categories from union revitalization studies into the power resources approach in welfare state research.

## *2.2.2 The Main Argument: Power Resources and Governing Alliances*

In accordance with the traditional power resources approach (Esping-Andersen [1990] 1998; Huber and Stephens 2001; Korpi 1983) and one strand of Marxist state theory (Poulantzas [1978] 2000), I assume that the redistributive character of public policy can be best understood as the outcome of relations of power.

I assume that the expansion of social protections for the lower popular classes becomes more likely when these sectors achieve considerable power resources, which enable them to exert pressures on policy-makers or to become themselves participants in the design and implementation of social policy. Moreover, even in the absence of active pressures, their power resources provide political elites with incentives to seek a certain level of legitimacy in their eyes. Hence, the power resources of these sectors are also likely to stimulate attempts on the part of political elites to use social policy expansion as a tool to increase their legitimacy, to win or co-opt supporters, or to contain conflicts.

However, the lower popular classes and their organizations alone have seldom achieved enough power resources to generate far-reaching social policy change. Such change requires the formation of broader alliances with other political forces. In accordance with the power resources approach, I assume that the most likely allies in the political realm are left<sup>37</sup> political parties, trade unions and social movements,38 which are inclined to support social policy expansion to low-income earners due to their egalitarian ideological orientation and their aspirations to organize these sectors. Non-left, populist labor-based parties and trade unions can also be inclined to such alliances, particularly in democracies, where the achievement of governmental control might be significantly facilitated through the electoral and associational support of the lower popular classes and their organizations. When such alliances are formed, the power resources of the lower popular classes and their organizations will influence their political weight compared to that of other allied forces. In sum, a significant extension of social protections towards low-income earners is assumed to become most likely when an inclusionary alliance is formed, in which left-wing and lower popular class actors have influence on the agenda-setting process.

<sup>37</sup> A definition of the left-right political spectrum can be found in the subchapter on the ideological orientation of actors*.*

<sup>38</sup> As will be described later in this chapter, social movements interested in social policy expansion for low-income earners can be very varied, ranging from women's, student, indigenous and unemployed workers' movements to human rights organizations to progressive sections of the church.

Furthermore, it is assumed that alliances can influence social policy much more strongly when they control government.<sup>39</sup> Strong oppositional alliances can also be able to extract significant social policy concessions through external pressure, but governing alliances are much more effective in translating their power resources and interests into policies and assuring their implementation. Therefore, relatively limited changes in the distribution of power resources can lead to significant changes in social policy when they enable the assumption of a new governing alliance. In the same way, it can be assumed that the incorporation or withdrawal of actors from alliances can lead to major changes in the direction of policy when it leads to a change of government.

Truncated social policy systems are assumed to result from conservative governing alliances that aim at fostering middle-class support and at controlling trade union agitation. In their composition, these alliances are exclusionary, meaning that they more often marginalize or exclude oppressed groups along lines such as social status, informality, gender, race and migratory background. Lower popular class, egalitarian left-wing and other progressive actors have little participation and influence in such alliances.

Once implemented, inegalitarian social policies, especially exclusive social insurance schemes, are assumed to create vested interests and hence path dependency effects. It is expected that such path dependency effects pose additional difficulties for progressive social policy reform as the resistance from employers and right-wing political actors might be reinforced by popular class actors who benefit from the respective policies.

In situations where popular class forces lack power resources across the board, social policy expansion is assumed to be unlikely at all.

## *2.2.3 Understanding Interest Constellations: An Analytical Framework*

The empirical literature on the role of different actors in the development of Latin American social policy regimes provides us with several theoretical puzzles. On the one hand, the array of actors participating in social politics in Latin American countries differs significantly from those of the industrialized, democratic countries during the post-World War II period that provided the empirical bases for most of the welfare state theories. In particular, the role of the military in shaping social policies in Latin America was significantly stronger due to recurrent military dictatorships. IFIs and social movements of peasants, indigenous communities and other marginalized groups have also played a

<sup>39</sup> An alliance that controls government will be referred to as governing alliance. In accordance with Guillermo O'Donnell (1977, 551), a governing alliance is defined as "an alliance that imposes, through the institutional system of the state, policies conforming to the orientations and demands of its members."

much more important role in Latin America. What interests did these actors pursue in the politics of social protection for low-income earners? And why did they do so? Furthermore, actors that played a similarly important role in Latin America as in most industrialized countries, such as political parties, trade unions, and employers' associations, developed quite distinct interests and agendas due to different structural, historical, and cultural contexts. For example, trade unions and labor-based parties in Latin America led processes of equality enhancing social policy reform in some periods, while during other periods they pursued particularistic social policy interests to the detriment of the lowest income sectors (Collier and Collier [1991] 2009; Golbert 2010; Huber and Stephens 2012; McGuire 2010).

Hence, in order to improve our understanding of the dynamics of Latin American social politics, it is crucial to analyze the historically specific development of interest constellations and the factors that exerted influence on whether actors tended to support the expansion of social protections for lowincome earners or not.

In the following, five groups of potentially influential factors will be discussed. The discussion is inspired by both empirical research and the theoretical controversy summarized in the literature review. The first two groups of factors are linked to actor-centered and ideational approaches and refer to the ideological orientations and social bases of actors. The discussion of the third group is inspired by structuralist approaches and focuses on the impact of social and economic factors on the formation of actors and alliances in social politics. The fourth and the fifth group are inspired by historical institutionalist approaches. The discussion of the fourth group of factors deals with the way in which social policy legacies and state interventions may influence the role of different actors in social politics and thereby create path dependencies. Finally, the discussion of the fifth group focuses on the influence of different regime types in shaping interest constellations.

The discussion of these five groups of potentially influential factors subsequently serves as a theoretical foundation for the formulation of hypotheses regarding the role of different types of actors in the politics of social protection for low-income earners in Latin America.

#### **The Ideological Orientation of Actors**

The ideological<sup>40</sup> positions<sup>41</sup> of actors are theoretically meaningful as these affect their goals and their perceptions of what constitutes an issue that deserves action and which kind of action might be best suited. Based on several analyses of welfare states in western countries, power resources theorists provided ample evidence that left, centrist, Christian democratic and right actors promote considerably different social policies. Left ideologies are related to egalitarian values and solidarity,<sup>42</sup> and hence it is the left that is associated with the most progressive social policy agendas. Christian actors of the political center and center-right are often influenced by Christian Social Teaching and therefore tend to support relatively broad welfare states, yet favor less egalitarian, highly stratified, status conserving social security systems. Secular right-wing and (neo-)liberal actors, in contrast, are assumed to have stronger confidence in the market and to prefer private insurance and residual public social protection over broader public welfare systems (Esping-Andersen [1990] 1998, chaps. 2, 3; Huber, Ragin, and Stephens 1993, 717–718; Huber and Stephens 2001, 19– 20).

An extensive study, which Micheal Coppedge (1997) elaborated in collaboration with 53 country experts, shows that Latin American party systems are

<sup>40</sup> In accordance with Stuart Hall (1986, 29) I refer to ideologies as mental frameworks which social actors "deploy in order to make sense of, figure out and render intelligible the way society works." As such frameworks, "ideologies allow people, as group members, to organize the multitude of social beliefs about what is the case, good or bad, right or wrong, *for them,* and to act accordingly" (Van Dijk 1998, 8).

<sup>41</sup> With ideological positions I refer to the position of an actor's ideology on the left-right spectrum. In accordance with Norberto Bobbio (1996) and Detlef Jahn (2011) I refer to the left-right divide as corresponding at its core to the difference between egalitarianism and inegalitarianism. The left considers inequality to be socially constructed, and hence its reduction or even elimination is perceived as being possible and desirable. The right in contrast considers inequality to be naturally given or an individual responsibility, so that a reduction or even elimination of inequality is seen as neither feasible nor desirable (Bobbio 1996, 67). Bobbio (1996, 66) also denotes that the terms left and right always describe a relation and are marked by gradual rather than absolute differences. Therefore, "when we say the left has a greater tendency to reduce inequalities, we do not mean that it intends to eliminate all inequalities, or that the right wishes to preserve them all, but simply that the former is more egalitarian, and the latter more inegalitarian" (Bobbio 1996, 66).

<sup>42</sup> Other authors on Latin American political economy have used similar conceptualizations of the left for the purpose of their analyses. Evelyne Huber and John D. Stephens (2012, 28): "What makes parties left or center-left in our assessment and in the assessment of experts is their ideological commitment to the values of egalitarianism and solidarity and their class appeals to subordinate classes." Kurt Weyland (2010, 5) conceptualizes left organizations "in ideological terms" as "characterized by the determined pursuit of social equity, justice, and solidarity as an overriding priority." And Diane Sainsbury (1999, 268) adds that this also applies to the dimension of gender relations.

very heterogeneous and differ from those of many western countries. Nevertheless, his and a variety of other studies also show that despite the manifold differences in the concrete party formations, Latin American political systems are also characterized by a left-right divide, with the left being more egalitarian and the right more inegalitarian (e.g., Alcántara Saez 2004; Colomer and Escatel 2005; Rosas 2005; Wiesehomeier and Doyle 2012; Zechmeister 2006; Zoco 2006). Hence, also for the study of Latin American welfare politics the left-right divide can be assumed to constitute a crucial category.

Power resources theorists have, in practice, treated ideological position as a static characteristic of parties, trade unions or movements. However, ideological positions of actors are subject to social construction, and thus might change over time and differ across space. Neoliberal hegemony, for example, has had a considerable effect on positions of left and labor parties, trade unions and social movements around the globe. This influence has usually been highly unfavorable for low-income groups in terms of social policy. At the same time, these ideological shifts have been associated with internal struggles and it can be assumed that it makes a difference for their social policy agendas whether left currents or neoliberal factions within these actors gain power. Such struggles might be of even greater importance in Latin America than in Europe, as many labor-based and populist parties as well as trade unions are ideologically very heterogeneous and lack long-standing ideological traditions (Levitsky 2003, 22).

Summing up, the hypothesis to be verified in the following empirical analysis is that left-of-center political actors are more likely to push for an expansion of social protections for low-income earners and that right-of-center actors are more likely to promote retrenchment or regressive reform. However, as many influential actors in Latin American politics such as populist parties and trade unions cannot be clearly categorized as left or right and changed political orientation over time, this hypothesis must be complemented by a second assumption. The second assumption is that the role of these actors in social policy reform processes also decisively depends on whether left or right internal currents manage to influence the agenda-setting on these issues. Again, the hypothesis is that political actors are more likely to support an expansion of social protections for low-income earners when left internal currents are strong and influential. For the assessment of the political orientation of different actors and internal currents I combine expert assessments found in literature with information collected through expert interviews.

#### **The Social Bases and Organizational Structures of Actors**

Power resources theorists assume that the role of different organizations in social politics rests not only on their distinct ideological orientations but also on the different compositions of their social bases. It is assumed that the popular class bases of left parties and trade unions make it more likely that these actors support redistributive and egalitarian social policy than other actors which rely stronger on middle- and upper-class support. The same argument can be extended to social movements, which mostly represent socially and politically oppressed groups in Latin America. Christian democratic parties, it is argued, build on a cross-class constituency, mobilizing support from parts of the working class as well as from farmers and the middle and upper classes. The crossclass character of these parties, in combination with the influence of Christian Social Teaching, is related to the less redistributive and more stratified character of their social policy agendas. Secular center and right parties' constituencies are more concentrated among farmers and the middle and upper classes and are assumed to be more hostile to public welfare provision. The same is assumed for the employers' associations due to their capitalist class bases (Esping-Andersen [1990] 1998; Huber, Ragin, and Stephens 1993, 741; Huber and Stephens 2001; Huber and Stephens 2012, 28; Korpi 1983).

With regard to the organization of the trade union movement, power resources theorists have developed another argument of theoretical importance for the question of social protection for the lowest income groups. While craft unions tend to fragment the working class and to pursue their own sectional interests, centralized union movements tend to represent the class as a whole. Moreover, centralized union movements tend to concentrate more resources on political issues, such as social policy, and to be more politically influential (Huber and Stephens 2001, 25; Stephens 1979, 45–46). Based on these arguments, it could be assumed that trade union centralization is an important factor favoring the extension of social protection to the lowest income groups. Nevertheless, concerning the situation in Latin America, this argument needs some qualifications. Trade union centralization can adopt different forms, and not all these forms seem to favor egalitarian social policy. First, centralized trade unions do not necessarily represent the whole working class. Many Latin American trade unions organize only formal workers. In these cases, the centralization of the union movement cannot necessarily be expected to improve the representation of informal workers. Second, it depends very much on the ideological orientation and internal structure of the centralized movement whether it tends to represent the lowest income sectors or not. Its ideological orientation and the organizational structure, in turn, depend very much on the initial process of centralization, which was vastly different in Sweden, the crucial country of reference for power resources theorists, than in countries such as Argentina.

This brings us to an issue that has so far received limited attention from theorists of the power resources approach. Different popular class organizations differ from each other regarding non-class characteristics of their social bases (Silver 2003, 22–23), and these differences might have important consequences for their role in the politics of social protection for low-income groups. As has been emphasized by feminist scholars, the building of labor and other popular class movements has often been associated with active exclusions of groups such as low-skilled workers or women (Cook 1981, 9; Hobson and Lindholm 1997, 480; Rose 1997, 147). Low-income as such seldom constitutes an explicit criterion for non-class boundary drawing in popular class movements, but exclusions along other lines such as gender, ethnicity, or skill often de facto operate as barriers against low-income groups. In Latin America, a formal work contract is often an "indispensable requirement for union affiliation" (Marshall and Perelman 2008, 8), and thus, informal workers, who constitute the vast majority of low-income earners, are frequently excluded from union organization. Such exclusions tend to result in narrowly defined organizational identities and favor the pursuit of particularistic interests, such as those of relatively well-off, white, and male workers (Kulawik 1996, 58). Diane Sainsbury (1999: 268), Dana Hill and Leann Tigges (1995, 115) find that union density among female workers and the share of women of the total union membership are positively related to gender egalitarian social policy, and hence, that women's influence within the labor movement has egalitarian effects. Similar positive effects can be expected from the involvement of informal workers and racially or ethnically oppressed groups. Organizations which exclude or marginalize underprivileged groups along lines such as gender, race, migratory background or informality, and whose social bases mainly consist of relatively better-off segments of the popular classes, are therefore assumed to be less likely to push for the extension of social protection for low-income groups than more inclusive organizations.

Furthermore, it must be considered that, in practice, the interests of the social bases of organizations do not translate into policy agendas on a one-toone basis. The question is not only whether and to what degree low-income earners, women and racially or ethnically oppressed groups are present in these organizations, but also whether and to what degree they can participate in the agenda-setting process. In this regard, the formal and informal organizational structures of parties, trade unions, and social movements play a decisive role. These structures partly determine how different sectors of the social base can influence the agenda. They constitute a particular relation between the leadership and the rank and file and cannot be seen as neutral channels for participation in agenda setting and strategy formation. Rather, they often provide considerable autonomy to the leadership, which can lead to the divergence of the interests of leaders from those of the rank and file. These self-interests of the leadership, in turn, can differ more strongly from the interests of one part of the constituency than another. Furthermore, these organizational structures and routines might favor some sectors and disadvantage others concerning the internal distribution of agenda-setting power. Hence, the organizational structures and routines of actors could be described as selective (Jessop 1990, 9– 10; 1999; Offe 1972a, 65–106). Recent empirical analyses provide evidence regarding the relevance of these selectivities for the development of social policy towards the lowest income sectors in Latin America. Lapp (2004), Pribble (2008) and Bossert (2011) have analyzed the effects of party organizational structures on land and social policies towards low-income earners. All three authors found that party structures significantly affect the possibilities of lowincome groups to make a difference in politics. This point appears to be of particular importance with regard to major popular class-based parties in several Latin American countries. The ideological positions of these parties have shown to be very flexible and their social bases tend to be very heterogeneous, including segments of the "urban and rural poor, the middle sectors, and, in some cases, the peasantry" (Levitsky 2003, 23). Bossert (2011) furthermore found that this also applies to trade unions and social movements involved in the politics of the welfare state. Democratic structures and bottom-up procedural practices are important for the ability of the economically weakest sectors of the rank and file to convert their numerical strength into agenda-setting power and to reduce the danger of co-optation of the leadership by economic and political elites. In contrast, top-down organizational structures and procedural practices provide considerable autonomy to the leadership and strengthen channels of direct influence that are mainly open to economic and political elites inside and outside the organization, such as lobbying, personal favors and co-optation (Bossert 2011). Although such kinds of internal structures severely limit the agenda-setting power of the lowest income groups, their continued support for these organizations can nevertheless be assumed to require that their interests are considered to some, albeit lesser, degree (Rueschemeyer, Huber Stephens, and Stephens 1992, 55).

In a similar vein, it can be assumed that the inclusion of low-income groups, such as unemployed and informal workers or peasants, in a broader political alliance makes it more likely that the respective alliance supports the implementation or expansion of social protections for low-income earners. The involvement of low-income earners can be assumed to require concessions from the other sectors of the alliance and to strengthen their capacity to participate in the policy-making process. Their involvement may, nevertheless, take very different forms that can be assumed to have different effects on both the magnitude and the quality of these concessions and their capacities to participate in the policy-making process. For example, they may be involved via electoral mobilization, membership in parties or trade unions, the formation of more or less independent social movements or a mixture of the aforementioned options. These different types of involvement may concede more active or passive roles and different levels of relative power to low-income sectors within the alliance. Due to their numerical strength, at least some kind of passive involvement becomes very likely during democratic periods while complete exclusion might prevail when authoritarian forms of government coincide with a lack of political organization and mobilization.

All in all, the hypothesis is that the political participation of the lower popular classes, women and ethnically oppressed groups in social movements, trade unions, parties, and political alliances increases the likeliness that these actors support the extension of social protections for low-income earners. This effect is furthermore assumed to be significantly greater when the formal and informal organizational and coalitional structures permit these groups to influence the agenda-setting process.

#### **Structural Context and the Social Construction of Actors**

The nature of concrete political actors, their overall strength, their unity, their racialized and gendered character, and whether and to what degree the lowest income groups form part of different organizations and political alliances is not only the result of these actors' choices and active boundary drawing but is also importantly shaped by the significantly different material preconditions for organization encountered by different segments of the society.

Based on Marxist premises, the power resources approach assumes that capitalism provides significantly unequal preconditions for the development of power in social politics. The control over the means of production provides the economic elite with important advantages regarding the incorporation of their interests into the social policy agendas of governments.<sup>43</sup> Popular classes, in turn, depend much more on their ability to organize, to build parties, trade unions, and social movements in order to effectively articulate their interests. Different social and economic structures across countries and periods have thereby provided different pre-conditions for processes of organization.

Due to their role as primary goods suppliers on the international level and their correspondingly limited degrees of industrialization, peripheral countries have tended towards an over-supply of labor and a high level of fragmentation of the popular classes, which has constituted a significant barrier to the construction of popular class power and the unity among different segments of it (Senghaas 1974). Power resources theorists have adapted their approach to the Latin American context in so far that they have considered that the structural preconditions were less favorable for the formation of strong labor movements than in most countries of the center (Huber and Stephens 2012, 24–25). Yet, the approach still requires specification concerning the question of how such structural differences might also have left their imprint on the nature of popular class formations that emerged in Latin America. My theoretical assumption is that the high level of heterogeneity of the popular sectors in Latin America has provided unfavorable conditions for the development of organizational unity

<sup>43</sup> See 2.1.1., Wright (1994), and Przeworski (1985, 139) for more detailed accounts of this argument.

and contributed to vastly different levels of organization among different segments of the popular classes.<sup>44</sup>

One of the most salient characteristics distinguishing Latin American labor markets from those found in the center is the magnitude of informal employment. While formal wage-earners constituted only about 35% of the economically active population in Latin America at the end of the 20th century, around 45% of the economically active population worked in different kinds of informal employment (Portes and Hoffman 2003, 46–49). Portes and Hoffman (2003, 50) statistically define informal employment "as the sum total of own account workers (minus professionals and technicians), unpaid family workers, domestic servants, and waged workers without social security and other legal protections in industry, services, and agriculture." Approximately half of the informal workers, depending on the country, work as wage-earners. Many of them are employed in small enterprises which supply low-cost goods and services to consumers and cheap inputs to formal sector companies. However, informal wage-earners can also be found in big companies and the agricultural sector. The remaining half of informal workers are urban self-employed workers or peasants (Portes and Hoffman 2003).<sup>45</sup>

Formal employment is associated with significantly higher average wages (Gasparini and Tonarolli 2009, 16). Moreover, in contrast to informal wageearners, formal workers possess better access to the social security systems and union representation and their work conditions are regulated by labor laws, which implies a certain degree of de-commodification.

The levels of organization have been extremely low in the informal economy, which means that most low-income earners have largely remained at the margins of traditional trade union organizations. In part, this can be linked to the much less favorable conditions for organization in the informal economy and certain barriers to interaction that are related to the formal-informal divide. Due to the structural oversupply of labor (Senghaas 1974, Tokman 1987), particularly in the segment with low education, informal wage-earners face fierce competition for jobs. The lack of access to de-commodifying social and labor policies further aggravates these competitive pressures, which makes them much easier to dismiss and limits their bargaining leverage vis-à-vis their employers. This ultimately makes participation in union organization and collective action appear much riskier. The shorter average duration of employment makes it furthermore difficult to establish organizational ties. Moreover, the vast majority of informal workers is dispersed among small enterprises with fewer than five workers, works in the agricultural sector or is self-employed

<sup>44</sup> The following subchapters will show that political and discursive factors were also important in this regard.

<sup>45</sup> According to data from CEPAL (2000, 65, 71), the peasantry constitutes around 10% of the Latin American workforce, which means slightly less than half of the total informal self-employed population (Portes and Hoffman 2003).

(Beccaria and Groisman 2009a; Gasparini and Tonarolli 2009). Compared to big industrial companies, which bring together "high concentrations of people with similar class interests and far-flung communications" (Rueschemeyer, Huber Stephens, and Stephens 1992, 58), these enterprises provide very unfavorable conditions for organization.

In addition to the different structural preconditions for organization encountered by formal and informal workers at their workplaces, joint organization is also hampered by limitations on interaction and mobility between them stemming from the socio-economic cleavage often separating them. These limitations do not affect all informal and formal workers in the same way. Empirical analyses indicate that there is considerable mobility between informal and formal salaried employment (Perry et al. 2007, 54–55). However, such studies also indicate that the level of mobility depends strongly on the respective educational level of the informal worker. The better educated informal wage-earners have significantly better chances of eventually accessing formal employment (Pagés and Stampini 2009, 398). In contrast, those with lower educational degrees suffer from a high probability of durably remaining in informal employment (Gasparini and Tornarolli 2009, Groisman 2010). In turn, informality is associated with a considerably lower average income and more recurrent phases of unemployment compared to that of formal workers. This cleavage in income level and security often means that informal workers and their families cannot afford to live in the same neighborhoods as formal workers. This phenomenon has been described as residential segregation, that is the perpetuation of poor neighborhoods in which there is a relative homogeneity of informal and unemployed workers and their families (Groisman 2010). Residential segregation poses limits to the day-to-day interaction between formal and informal workers. Taken together with the significantly different consumption capacities this might constitute a structural barrier to the development of common class identities and joint organizations.

This does not mean that the development of common political identities and trade union organizations nor that of autonomous social movements of informal and unemployed workers is impossible, however. Trade unions can make strategic choices and some components of the labor movement have at times demanded more active strategies to organize these sectors (Palomino 2005; Svampa and Pereyra 2003). Nevertheless, under such structural conditions popular class unity seems to be more difficult to attain.

At the same time, these structural preconditions might provide space for types of progressive popular class movements that are rarely observed in western countries. While residential segregation might reinforce barriers to informal workers' participation in formal worker organizations such as trade unions, it might at the same time constitute a precondition for organization based on the territorial dimension (Collier and Handlin 2009). The emergence of a variety of organizations of unemployed and informal workers in Argentina might exemplify this (Merklen 2005; Svampa and Pereyra 2003). The MST movement in Brazil also shows that mobilization and organization of peasants and landless agricultural workers are possible (Carter 2010; Souza Martins 2002). In countries such as Bolivia, Ecuador, Guatemala, and Mexico, strong progressive social movements have been built around indigenous identity (Yashar 1998). The emergence of both rural and urban social movements that mostly organize groups belonging to the lowest income strata shows that despite limitations imposed by social and economic structures, much depends on how agents react to these and what strategies they develop. Ruth Berins Collier and Samuel Handlin (2009, 6) even raise the question of whether these processes of organization and mobilization and the associated governmental responses have led to a new critical juncture comparable to that of the incorporation of organized formal labor into the political systems in Latin America during the first half of the 20th century.

The gendered and racialized nature of concrete class formations cannot be understood solely as a result of active boundary drawing by these organizations detached from the wider inequalities in the social and economic structure. Evelyne Huber and John D. Stephens (2001, 47) argue that increasing participation of women in the paid labor force "increases the level of organization of women through unionization and thus the potential for women's political mobilization. This process of economic and social mobilization of women also leads to increased political mobilization of women, in existing parties, in women's organizations, in new social movements, and so on." In turn, female, and particularly feminist, actors have been decisive promoters of egalitarian social policy agendas within and apart from popular class organizations (Hill and Tigges 1995, 115; Huber and Stephens 2001, 17; Huber et al. 2009, 31). As the preconditions for organization vary significantly between different economic sectors and types of employment, the power of women within the labor movement might furthermore be influenced by the economic sectors and kinds of employment in which female workers predominantly work. Amy Hite and Jocelyn Viterna (2005, 57, 78) argue that the weakening of largely male-dominated trade unions during the last quarter of the 20th century and the strengthening of new, often community-based social movements, has also changed the gendered nature of popular class mobilization in favor of a more influential role for women.

To sum up, while we cannot deduce the strength and nature of historically concrete popular class formations from wider social and economic structures, the consideration of these structures together with other factors can help us to understand them. The hypothesis is that different sectors of the popular classes face significantly different pre-conditions for the construction of political organizations, which subsequently fosters a very unequal distribution of power, in particular between relatively well-organized formal workers and largely unorganized informal workers. This inequality, in turn, is assumed to constitute a difficulty for the construction of popular class unity and increases the likelihood of exclusionary alliances.

## **State Intervention, Path Dependency, and the Role of Different Actors in Social Politics**

In Latin America, state actors often played a decisive role in the constitution and transformation of political forces representing the popular classes (Levitsky 2003, 22–23; Portantiero 1988, 137–138). Governments such as those of Juan Domingo Perón in Argentina and Getúlio Dornelles Vargas in Brazil profoundly shaped the party system and the labor movement. The effects of such state interventions tend to persist over time, and hence, to shape the long-term evolution of popular class actors, their identity, and their interests (Collier and Collier [1991] 2009). As Ben Ross Schneider (2004) argues, while state intervention shaped the development of centralized business associations in Mexico, Chile, and Colombia, it however failed to create stable centralized associations in Brazil and Argentina. Peter Birle (1995) comes to a similar conclusion for Argentina. The focus of my analysis will therefore be to examine whether and how state interventions influenced the way in which different popular class actors participate in the politics of social protection for low-income earners.

While theoretically all kinds of state policies can have such effects, social policy legacies are particularly pivotal for the politics of the welfare state for two reasons. First, as has been argued before, social policies shape not only the distribution of wealth but also the distribution of power. They can, for example, increase the independence of individuals from market participation and from other family members. Thus, social policies play a pivotal role in shaping power relations such as those related to gender and class. De-commodifying social policies increase the individual negotiation power of workers and the capacity of the subordinate classes to organize (Esping-Andersen [1990] 1998, 22). However, as has been argued before, different segments of the popular classes have encountered very different preconditions for the development of collective power. Egalitarian social policies can theoretically reduce such differences. However, as social policies in Latin America were often a reaction of political and economic elites to the growing strength of those segments of the popular classes that managed to organize, they had often limited coverage and tended to deepen asymmetries of power among and within the popular classes (Abel and Lewis 1993, 15). Indeed, fragmented social policies were often one of the means of conservative political and economic elites to divide and control the popular classes (Rueschemeyer, Huber Stephens, and Stephens 1992, 68; Segura-Ubiergo 2007, 260).

Furthermore, social policy legacies not only influence the distribution of power but also intervene in the constitution and transformation of actors and their perceived interests and alliance preferences (Esping-Andersen [1990] 1998, 109). The welfare state, as Esping-Andersen ([1990] 1998, 23) argues, is "a system of stratification," and hence, "an active force in the ordering of social relations." Social policies shape lived experiences and patterns of interaction. Therefore, they are likely to have an imprint on the patterns of political mobilization. While fragmented social insurance policies that only cover the better off segments of the popular classes might reinforce barriers to common perceptions of interest between these and low-income earners, egalitarian social policies might significantly reduce them.

Moreover, social policies tend to create vested interests. These include the beneficiaries, but also the administrators and contractors of a particular policy (Haggard and Kaufman 2008, 197; Pribble 2008, 62–63). Such groups are likely to defend the respective policy, and thereby make significant changes politically more difficult (Huber, Ragin and Stephens 1993, 722–723). By this means, vested interests contribute to the inertia of particular policy paths, which can be assumed to be the case for both egalitarian and exclusionary social policies. The kind of vested interests created by fragmented social insurance policies plausibly reinforce the drawing of non-class boundaries by relatively "privileged," 46 often predominantly male and white popular class factions, and therefore constitute a significant barrier to the construction of political alliances between different, unequally protected segments of the popular classes (Bossert 2011).

The opposite process, the crumbling of "privileges," can be assumed to increase the chances for a breakdown of such barriers. Beverly Silver (2003, 134), for example, relates that in the late 19th and early 20th century the attack on craft standards in Britain, France, and the United States "chipped away at the 'consent' of the 'labor aristocracy'" and induced "skilled workers to reach out to the growing ranks of unskilled workers."

For all these reasons, social policies have often served political purposes such as to construct and reinforce political alliances. While left governments have recurrently adopted universal social policy programs in order to mobilize support from different segments of the popular classes and to foster solidarity among them (Esping-Andersen [1990] 1998, 67), conservative governments have frequently used fragmented social policies to divide and control them (Rueschemeyer, Huber Stephens, and Stephens 1992, 68; Segura-Ubiergo 2007, 260). Government ambitions to extend control over popular class forces have also often influenced the particular design of social policies in a way that provided politicians with patronage resources that facilitated the co-optation of trade union and social movement leaders (Haggard and Kaufman 2008, 60).

To sum up, my empirical analysis will take into account the ways in which state policies affected the constitution and transformation of popular class

<sup>46</sup> I put the word privileges in quotes because while workers with access to social insurance in Latin America might appear privileged compared to most informal workers, they are nevertheless exploited in the process of production.

actors and how this influenced their medium- and long-run participation in social politics. The hypothesis is that, once implemented, social policies create path dependencies, which means that they significantly affect the constellations of interest in posterior reform processes. Fragmented social insurance legacies are assumed to constitute obstacles to later attempts to shift social policy in a more egalitarian direction, while universal policy legacies are assumed to facilitate the extension of social protection for low-income groups (Martínez Franzoni and Voorend 2009; Huber and Stephens 2012, 255).

#### **Regime Type and Popular Class Actors**

In his elaboration on the concept of the strategic selectivity of institutions, Bob Jessop (1990, 9–10; 1999) emphasizes that institutions not only grant differential access to the policy-making process but also have an impact on the process of identity, interest and strategy formation of actors.<sup>47</sup> In this sense, theories on welfare state development have argued that regime type is of central importance for the constitution of popular class actors and their coalition strategies in social politics (Esping-Andersen [1990] 1998; Huber and Stephens 2012; McGuire 2010). Regime type refers to the basic question of whether a political regime is democratic, authoritarian or something in between.

As has been described earlier, associational power tends to be unequally distributed among and within the popular classes in Latin America. While formal workers in certain industries have achieved considerable organizational strength, informal and unemployed workers, and hence most low-income earners, have most often lacked strong organizational representation. Democracy does not tend to radically alter this situation, but it nevertheless provides the lowest income groups with a certain degree of electoral power, and this significantly changes the attractiveness of different types of alliances. As electoral success requires broad support, democracy provides strong incentives to left and labor-based parties to conceptualize the working class in broad terms and "even to go beyond this broad concept by emphasizing similar life situations and 'parallel interests'" (Przeworski 1985, 74–75). It is therefore no coincidence that in most Western European labor movements the "principle of a broad popular universalism emerged in tandem with the extension and consolidation of democratic rights" (Esping-Andersen [1990] 1998, 67). In the Latin American context, this means that the possibility of participating in electoral

<sup>47</sup> Bob Jessop (1990, 10) writes: "Particular forms of state privilege some strategies over others, privilege the access of some forces over others, some interests over others, some time horizons over others, some coalition possibilities over others. A given type of state, a given state form, a given form of regime, will be more accessible to some forces than others according to the strategies they adopt to gain state power. And it will be more suited to the pursuit of some types of economic or political strategy than others because of the modes of intervention and resources which characterize that system."

contests provides left and labor parties with strong reasons to appeal to the numerically strong group of weakly organized low-income earners. Advocating the extension of social protection to agricultural, unemployed, or informal workers is thereby one way of doing so. James McGuire's (2010, 303–304) observation that in Latin America labor unions have tended to assume more favorable positions regarding the extension of health care services to low-income groups in countries with a longer democratic history supports this argument.

To sum up, my hypothesis is that democracy provides left and labor-based parties and alliances with additional incentives to seek the support of numerically strong low-income groups such as peasants and informal rural and urban workers in order to increase the chances of achieving majorities. Although this cannot be assumed to mechanically lead popular class organizations to adopt positions in favor of an extension of social protection to low-income groups, it can be assumed that it increases the likelihood of this happening. In semi-authoritarian regimes these incentives are significantly weaker and in authoritarian regimes they are largely absent, which increases the likelihood of popular class actors focusing on narrower social policy agendas in favor of their own associational bases.

## *2.2.4 The Role of Different Actors: Hypotheses*

Based on the discussion above and in accordance with the traditional power resources approach, I assume that left parties are mostly supporters of an expansion of social protections for low-income earners due to their ideological orientation and their aspiration to mobilize lower popular class support (Esping-Andersen [1990] 1998; Huber and Stephens 2012).

Non-left labor-based parties, such as the Argentine PJ, and centrist parties, such as the Argentine UCR, have usually been much less ideologically defined than social democratic or communist parties. Nevertheless, within such parties, left and right internal currents have coexisted and struggled for influence. I assume therefore, that these parties support an extension of social protections for low-income earners when left tendencies become influential and when democracy provides them with incentives to compete for the lower popular class vote.

In accordance with the traditional power resources approach and the theoretical discussion above, I assume that right of center parties tend to oppose the extension of social protections for low-income earners. Christian conservative parties are assumed to support certain types of social policies due to the influence of Catholic Social Teaching, yet favor less egalitarian, stronger stratified, and status conserving social security systems that result in unfavorable policies for most low-income earners. Secular conservative and liberal parties are assumed to have stronger confidence in the market and to prefer private insurance and residual public social protection over broader public welfare systems (Esping-Andersen [1990] 1998; Huber, Ragin and Stephens 1993, 717– 718; Huber and Stephens 2001, 19–20).

In contrast to the assumptions of the traditional power resources approach, trade unions have played an ambivalent role regarding social protections for low-income earners in Latin America. Based on the theoretical discussion above, my assumption is that this ambivalence is due to the fact that Latin American trade unions seldom organize or represent informal low-income workers and that their leaderships are in part centrist or conservative in political orientation. However, I assume that trade unions pursue more inclusionary social policy agendas when left internal currents attain leadership. I also assume that trade unions that organize significant numbers of low-income workers, or which at least aspire to do so, are more likely to support such social policies. Yet, when a more conservative leadership dominates the organization, trade unions tend to concentrate on the improvement of social protections for their own organizational bases among formal workers, which at times even results in opposition to broader and more egalitarian policies.

In agreement with the power resources approach (Esping-Andersen [1990] 1998, Huber and Stephens 2001; Korpi 1983) and one strand of Marxist state theory (Poulantzas [1978] 2000), I assume that capitalists and big landowners tend to oppose the expansion of social protections for low-income earners because of two reasons. First, such policies tend to conflict with their economic interests. They require fiscal resources that can potentially result in higher tax levels. Furthermore, access to social protection tends to reduce the individual worker's dependence on market income and hence it tends to provide the worker with additional bargaining power and potential space for emancipation. This, in turn, is likely to generate upward pressures on wage levels. Second, Latin America's economic elite has traditionally been dominated by conservative and market liberal thinking. This does not mean that individual capitalists or employers' associations cannot have an interest in such policies under certain circumstances; for example, when mass mobilization puts the viability of their businesses in danger (Paster 2013) or when a particular policy has results beneficial to certain capital factions (Swenson 2004).

In accordance with most feminist social policy research, I assume that feminist actors tend to favor the extension of social protections to low-income earners because many gender egalitarian social policies, such as the universal provision of child care, health care and access to legal and free abortion, benefit low-income earners and because most feminist organizations are left-of-center in their political orientation (Arza 2012, 11; Huber 2006, 301; Huber and Stephens 2001, 21; Sainsbury 1996, 127–169).

Furthermore, based on the theoretical arguments outlined above, I assume that indigenous, peasant, unemployed and informal workers' movements favor the extension of social protections for low-income groups, as their social bases are among the key beneficiaries of such policies and because their political orientation is overwhelmingly left-of-center.

The Catholic church is assumed to support stratified, status conserving social policies due to the influence of Catholic Social Teaching. Notwithstanding, certain progressive base-level currents, such as those related to liberation theology, are assumed to support more redistributive social policies due to their left-of-center political orientation and their aspiration to empower and represent the lower popular classes.

The social policy agenda of the military is assumed to include a preference for social security privileges for members of the armed forces and to depend on the ideological orientation of the military elite regarding the wider configuration of the welfare state. While leftist military factions are assumed to support egalitarian social policy, rightist factions are assumed to promote either status conserving or liberal market-oriented social policies. Since in most of the Latin American cases, military leaders have been right-wing, the overall influence of the military is assumed to be highly regressive.

In accordance with Bob Deacon's (1997; 2005) and Anne Tittor's (2012) analyses, I assume that distinct international organizations have adopted different positions on social policy. IFIs, such as the IMF, World Bank, and IADB, have pushed a neoliberal agenda of privatization and marketization (B. Deacon 2005; Tittor 2012). While they promoted targeted and scarcely funded social assistance policies for particular low-income groups, they also promoted social insurance reforms which tended to reduce general coverage of low-income earners significantly. Korpi and Palme (1998) related this kind of social policy agenda with the so-called paradox of redistribution, meaning that despite increases in targeted social assistance expenditure, this kind of agenda rather reduces than increases the overall amount of progressive income redistribution and hence produces unfavorable results for low-income earners. The WHO, and more recently the ILO as well, have in contrast shown interest in universal health care and income transfer policies which would benefit most low-income earners.

## *2.2.5 (Re-)Conceptualizing Power Resources in Social Policy-Making: Associational, Structural, Institutional, and Discursive Power Resources*

In accordance with the traditional power resources approach in welfare state research, I assume that in the long run it depends crucially on the balance of power whether progressive actors and alliances are able to construct and defend extensive and redistributive social policy systems that provide ample social protection to low-income earners. However, while the traditional power resources approach focuses tightly on working-class power as the main enabler of redistributive social policy, I assume that this focus is too narrow for the Latin American context. Instead, I assume that the power of other popular classes, such as peasants, informal and unemployed workers, also plays an important role. Therefore, my approach will deal more broadly with popular class power. This implies that besides trade unions and left parties, a range of other social movements and associations, as well as progressive currents within populist and centrist parties and the church, will be considered promotors or at least enablers of social protection for low-income earners.

## **Definition of Power**

For the purpose of this study, I use Susan Strange's (1996, 17) basic definition of power as "the ability of a person or group of persons so to affect outcomes that their preferences take precedence over the preferences of others." Susan Strange's definition resembles in many aspects the definition found in Max Weber's classic *Economy and Society*, where he writes that power can be defined as "the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests" (Weber 1978, 53). However, Strange's definition, as I interpret and use it in this study, is broader. While Max Weber emphasizes that power is the probability to enforce one's own will despite resistance, Strange's definition of power also encompasses situations where there is no resistance, such as those contemplated by Bachrach and Baratz (1962; 1975) and Lukes ([1974] 2006). Bachrach and Baratz (1962, 948) argue that power is also exercised when actors devote their "energies to creating or reinforcing social and political values and institutional practices that limit the scope of the political process to public consideration of only those issues which are comparatively innocuous" to them. Lukes ([1974] 2006) furthermore considers that power can be exercised by manipulating others. Lukes ([1974] 2006) and Bachrach and Baratz (1962) thereby emphasize that the exercise of power is very much related to institutional, discursive, and structural factors.

## **Four Types of Power Resources**

If power is defined as "the ability of a person or group of persons so to affect outcomes that their preferences take precedence over the preferences of others" (Strange 1996, 17), the question is what this ability depends on. The answer will inevitably vary with the subject under analysis and the range of actors that is involved. As discussed above, on the basis of a broad literature review and previous empirical research, the assumption of this study is that the introduction, expansion, and defense of social protections for low-income earners crucially depends on the power of popular class actors and alliances, and hence, on their ability to shape political outcomes in this policy area. In the following, I will argue that there are four crucially important power resources on which this ability depends.<sup>48</sup> 

The first power resource refers to associational strength and therefore focuses on empowering characteristics of these actors and alliances themselves. However, in addition to the assumptions of the traditional power resources approach, I assume that there are three more power resources that are crucial for understanding social policy change (Esping-Andersen [1990] 1998; Huber and Stephens 2001; 2012; Korpi 1983). In contrast to associational power resources, these are less related to characteristics of these actors and alliances themselves and more strongly tied to the broader societal structures, institutions and discourses surrounding them.

In the following I will elaborate this argument and define four different types of power resources: a) associational power resources, b) structural power resources, c) institutional power resources and d) discursive power resources.

#### **Associational Power Resources**

Associational power resources stem from organization processes that enable formerly dispersed individuals to aggregate their resources and influence, to unite behind common goals, and to become able to act collectively. Due to the dynamics of capitalist systems and the concentration of wealth and the control over the means of production, economic elites depend much less on organization as a means of political articulation than popular classes do (Huber and Stephens 2001, 13). On these grounds, the traditional power resources approach has focused on the associational power resources of the working class as the main explanatory variable for the redistributive character of the welfare state. Trade union density, trade union centralization, and left party strength measured as long-term vote, parliament and cabinet shares have been used as indicators (Esping-Andersen [1990] 1998; Huber and Stephens 2001; 2012; Korpi 1983; Stephens 1979). Although these indicators have been considered to be interdependent, the long-term partisan composition of governments has been found to be the indicator with the strongest effect on social policy (Esping-Andersen [1990] 1998; Huber and Stephens 2001; 2012). Esping-Andersen ([1990] 1998, 107) furthermore suggest that the effects of these indicators of working-class power also depend on the relative strength and unity of center and right-wing parties and employers' associations, as well as on the capacity of popular class organizations to build effective coalitions.

<sup>48</sup> Of course, the conceptualization of power resources cannot capture all aspects of power. As does any concept that is intended to be used for empirical research, it reduces complexity and is developed for a specific historical and spatial context. Only in this way can it help us to analyze the crucial dynamics of the power struggles behind social policy making in Latin America.

These indicators are also useful in accounting for at least a part of the associational power resources of the popular classes in Latin America. Yet, as has been argued above, it is necessary to broaden the scope of the analysis to include also the associational strength of other actors, such as unemployed and informal workers' movements, indigenous movements, peasant organizations, and progressive currents within populist and centrist parties as well as the church. As left parties, trade unions and social movements often represent different segments of the popular classes, I furthermore assume that not only the overall strength of popular class organization is decisive for explaining the level of social protection for low-income earners but also the relative strength of the different types of popular class organizations and the patterns of coalition among them. As peasants', indigenous peoples', unemployed workers', and informal workers' movements often represent the most oppressed sections of the popular classes, the achievement of associational power resources on their part is considered to be particularly favorable for the expansion of social protections for low-income earners.

## **Structural Power Resources**

Structural power resources stem from the characteristics of the economic system and its current dynamic. Through the mainly private control and ownership of the means of production, capitalist systems confer ample structural power resources to capital (Korpi 1983, 16–17). As Erik Olin Wright (1994, 94–95) summarizes, such power resources can provide significant leverage in struggles over public policy:

Through a wide variety of concrete mechanisms—financing politicians, political parties, and policy think tanks; financially controlling the main organs of the mass media; offering lucrative jobs to high-level political officials after they leave state employment; extensive lobbying—capitalists are in a position to use their wealth to shape directly the direction of state policies.<sup>12</sup>When combined with the dense pattern of personal networks which give capitalists easy access to the sites of immediate political power, such use of financial resources gives the bourgeoisie vastly disproportionate direct leverage over politics.

Furthermore, in capitalism, both state revenues and societal wealth are dependent on the process of capital accumulation (Wright 1994, 100). As Adam Przeworski (1985, 139) argues, if "capitalists do not appropriate profit (…), production falls, consumption decreases and no other group can satisfy its material interests." Such a decline in wealth, in turn, threatens not only to undermine the financial bases of the state itself but also its popular support bases. State managers, such as politicians and high-level officials, have therefore strong incentives to implement policies that are assumed to generate a favorable investment climate. Hence, capital's control over most of the investment constitutes a strong systemic source of power. In this sense, Gregg Olson and Julia O'Connor (1998, 21) argued that neoliberal globalization further strengthened the structural power resources of capital as it increased the mobility of goods and capital, and hence facilitated the exit option.

While power resources theorists in the field of welfare state research acknowledge that capitalist systems tend to confer ample structural power resources to capital, they have so far not considered that different dynamics of capitalism and positions in the international division of labor might also affect the structural power resources of the popular classes. The concept of structural power resources reached broader attention in the field of labor studies by virtue of Erik Olin Wright (2000, 962) who argued<sup>49</sup> that workers' structural power resources stem "from the location of workers within the economic system." They can result "from tight labor markets or from the strategic location of a particular group of workers within a key industrial sector." Beverly Silver (2003) elaborated further on the concept and used it as an analytical category in her influential book *Forces of Labor*. Silver (2003, 13) defines two subcategories of structural power: marketplace bargaining power and workplace bargaining power. Marketplace bargaining power arises from tight labor markets and "can take several forms including (1) the possession of scarce skills that are in demand by employers, (2) low levels of general unemployment, and (3) the ability of workers to pull out of the labor market entirely and survive on nonwage sources of income." Workplace bargaining power stems from the location of a group of workers in the wider system of production and "accrues to workers who are enmeshed in tightly integrated production processes, where a localized work stoppage in a key node can cause disruptions on a much wider scale than the stoppage itself." So, for example, a strike of railway or port workers can force a wide array of industries to stop production due to the impossibility of obtaining necessary supplies or of transporting produced goods to their potential buyers.

There are theoretically two levels on which the possession of different degrees of structural power resources can be supposed to affect the extent of public social protection for low-income groups. On the level of the workplace, the distribution of structural power resources provides wage-earners with varying capacities to demand decent work contracts from employers, which might include the payment of contributions to social insurance systems. On the macrolevel of the political system, on which this study focuses, differing structural power resources are associated with different capacities to exercise pressure on the government and other actors in order to influence social policy. Even when pressures are not actively exercised, the capacity to use structural power resources generates incentives for governments and other actors to respond to

<sup>49</sup> In the concerned article, the author actually focusses on the effects of associational power resources and refers to structural power resources rather briefly. However, his distinction between associational and structural power resources was later taken up by several authors in the field of labor studies (e.g., Dörre 2011, 277; Silver 2003, 13).

the interests of the respective group in order to achieve a certain degree of legitimacy in their eyes or to prevent costly conflicts.<sup>50</sup>

Conditions of low unemployment and underemployment and universal social security make it harder for employers to replace workers in the face of strikes , and hence the capacity of workers to impose costs on employers and the state increases. Furthermore, under such conditions, workers face lower risks of losing their jobs and incomes, which might reduce fears of undertaking collective action. In a similar vein, in more industrialized, capital intensive economies or sectors with tightly integrated production processes, wage earners tend to possess more power to impose significant economic costs on employers due to the high level of dependency of broader production networks on specific groups of workers. Thus, the distribution of both marketplace bargaining power and workplace bargaining power affects the potential of workers to obstruct capital accumulation, and hence to exert economic pressure on both employers and governments. On these grounds, the power resources approach in the field of the welfare state can be importantly enriched by incorporating structural power resources as an analytical category.

The distribution of structural power resources can vary significantly over time and between different countries with distinct positions in the global economy and different economic and socioeconomic structures. Therefore, when applying this concept to Latin American countries, it is important to consider their specific social and economic characteristics. Due to the lower degree of industrialization, lower average capital intensity and a structural surplus of labor, popular classes in peripheral countries tend to possess significantly lower structural power resources than their counterparts in the center. Furthermore, structural power resources tend to be very unequally distributed among different segments of the popular classes. Those with higher educational levels tend to possess relatively strong marketplace bargaining power while major parts of the populations tend to be exposed to an extremely competitive labor market. Whereas certain groups of industrial, transport and state workers have considerable workplace bargaining power (Collier and Collier [1991] 2009, 41; Silver 2003, 170–171), ample parts of the labor force work in agriculture, small enterprises, or are self-employed, sectors in which the macro-economic effects of strikes tend to be much more limited.

#### **Institutional Power Resources**

In the field of welfare state research, approaches focusing on the role of political institutions have usually been seen as alternatives to approaches that focus on the relationship of forces in the society (Schmidt and Ostheim 2007b, 63).

<sup>50</sup> Such responses to popular class power usually involve both material benefits as well as repression and divide and rule strategies.

This is, however, a short-sighted interpretation. Institutions are in several ways related to the distribution of power resources (Brinkmann et al. 2008; Dörre 2011). Indeed, political institutions play a pivotal role in what Peter Bachrach and Morton S. Baratz (1962, 948) call the second face of power. According to these authors, actors can exercise power by creating or reinforcing institutions "that limit the scope of the political process to public consideration of only those issues which are comparatively innocuous" to them (Bachrach and Baratz 1962, 948). The exercise of power is thus inscribed into specific selectivities of these institutions (Jessop 1990, 9–10; 1999; Offe 1972a: 65–106; Poulantzas [1978] 2000, 123–145, 154–160). Once established, these institutions play an important role in shaping the distribution of power resources by providing some actors with more capacity to influence social policy than others.

Comparative literature has analyzed the effects of a variety of institutional factors on social and distributive politics, but the majority of studies indicate that regime type is the most important institutional factor affecting the redistributive character of the social policy systems in Latin America (Filgueira 2005; Haggard and Kaufman 2008; Huber et al. 2006; Segura-Ubiergo 2007). This fits well with my assumption that an important mechanism through which institutions influence social policy outcomes is their impact on the distribution of power resources among contending actors and interests. As popular classes depend much more on collective action than do economic elites, the right to vote and to organize can be considered a decisive power resource for them. Although there is no automatism that guarantees that such strategies will be taken up (Hobson and Lindholm 1997), electoral appeals to the numerically strong group of low-income earners based on social policy promises constitute a plausible power strategy in democratic regimes, particularly for actors of the political left. Several empirical studies support these theoretical assumptions and find a positive association between democracy and social protection for low-income earners in Latin America (Haggard and Kaufman 2008; Huber et al. 2006; Huber and Stephens 2012; McGuire 2010; Segura-Ubiergo 2007). It will thus be assumed that the institutional power resources of the popular classes are more extensive in democracies than under authoritarian rule.

Nevertheless, it must also be considered that in practice there has been no simple dichotomy between democracy and authoritarianism in Latin America. Rather authoritarian regimes have sometimes allowed a certain degree of electoral competition and interest group organization, while rather democratic regimes have often contained important limitations such as the proscription of major parties, the presence of non-elected representatives of the army in the Senate, or the limitation of the right to vote to only male citizens or those who are literate (Haggard and Kaufman 2008, 15; Pribble 2008, 30). In most cases, such formal restrictions on democracy gave the political system a bias that reduced the level of institutional power resources of the popular classes. Restrictions on the right to vote often excluded women and the poor. The parties that were proscribed were usually those with left ideologies or considerable support from the working class and other popular classes.

It is important for me to emphasize here that although democratic regimes provide opportunities for power strategies that are based on the electoral mobilization of the popular classes in favor of an expansion of social protections for low-income earners, there is no guarantee that this will in fact occur or be successful. Hence, my argument runs against the widely used median voter model of Meltzer and Richard (1981), which postulates that democratic regimes quasi automatically lead to redistributive public policy in societies with high levels of inequality. In contrast to Meltzer and Richard and in accordance with Burchardt (2010, 44) and Huber and Stephens (2012, 269), I argue that high levels of inequality tend to strengthen the political power of the economic elite rather than the power of the mass of low-income earners. While formal democracy provides the popular classes with institutional power resources, high levels of inequality provide the elite with additional structural power resources and create difficulties for the construction of political unity among and within the popular classes. Hence, inequality is assumed to countervail rather than to strengthen the redistributive effects of democracy.

One way in which high inequality might hinder the successful mobilization of electoral power resources in favor of redistribution is that it tends to increase the elite's capacity to shape widely shared discourses on the risks and supposedly negative consequences of redistributive policies. In the following, I will deal with this issue.

#### **Discursive Power Resources**

My definition of discursive power resources rests on the assumption that the ability of actors and alliances to affect social policy outcomes also depends on whether the wider discursive context allows them to frame their particular interests as acceptable or even desirable for the society as a whole. Power resources theorists in welfare state research have begun to acknowledge the importance of discursive paradigms for the balance of power in society, but they have still not integrated this aspect into their analytical frameworks (e.g., Huber and Stephens 2001; Huber and Stephens 2012).51 The following elaboration on the discursive power resources aims at closing this theoretical gap.

The concept of discursive power resources builds heavily on ideational approaches that focus on the importance of broad paradigms (Béland and Cox 2013; Hall 1993; Hay 2001; Hobson and Lindholm 1997; Wincott 2011). I

<sup>51 &</sup>quot;(...) since the neoliberal view of the world is highly congruent with the interests of capital, its rise to hegemony has further strengthened the leverage that capital has been gaining from globalization vis-à-vis governments and unions, by way of legitimizing the claims of capitalists regarding constraints on social policy in the eyes of the mass public." (Huber and Stephens 2001, 40)

adopt Peter Hall's (1993, 279) definition of a discursive paradigm as a "framework of ideas and standards that specifies not only the goals of policy and the kind of instruments that can be used to attain them, but also the very nature of the problems they are meant to be addressing" (Hall 1993, 279). Although ideational contexts are always heterogeneous, different discursive paradigms tend to prevail in different historical moments and to become "widely shared ways of thinking about policy challenges that lead to broad consensus about the appropriateness of a policy response" (Béland and Cox 2013, 193).

It is of crucial importance that these discursive paradigms are not neutral towards different actors and interests. They significantly shape the distribution of power by enabling certain actors but not others to successfully frame their preferences as desirable for the society as a whole.52 The ability to frame particular preferences as generally desirable, in turn, is heavily dependent on the extent to which these preferences are compatible with the logic of the prevalent discursive paradigm.

In order to prepare the concept of discursive power resources for empirical use, it is necessary to first identify the relevant discursive paradigms regarding the policy field and the historical context of interest. As a second step, it is necessary to analyze the nature and logic of each paradigm and the degree to which it is compatible with different policy agendas.

Based on the discussion of ideational approaches in the literature review above, I assume that development paradigms are the most important discursive paradigms in shaping the macro-political struggle over social policy in Latin America since the mid-20th century. As social policy has been most of the time subordinated to the overriding goal of economic development, the appropriateness of different social policy options was frequently judged in terms of their supposed consequences for the economy. Thereby, these paradigms were not only influential among policy-makers but also among the wider public.

The reviewed literature indicates that three different large-scale development paradigms were prevalent during the last 70 years in Latin America: Import Substitution Industrialization (ISI), neoliberalism, and a recently emerged paradigm which shall be referred to as post-neoliberalism (e.g., Burchardt 2006; Haggard and Kaufman 2008; Kirby 2009; Riesco and Draibe 2007; Sheahan 2002; Silva 2007). Both the ISI and the post-neoliberal paradigms are characterized by a logic influenced by Keynesian ideas that is friendly to state intervention, while neoliberalism pushes for the expansion of the market.<sup>53</sup> My

<sup>52</sup> According to Fred Block (2008, 4), already Marx argued that "classes are able to exert greater influence when they claim to be fighting for interests that are universal rather than just for their own particular interests."

<sup>53</sup> In this sense, the long-term evolution of development paradigms in Latin America reminds of Karl Polanyi's ([1944] 2001) argument of the double-movement, that is that phases of expanding the market provoke a counter-movement that eventually leads to a phase of renewed regulation and restriction of the market.

central argument is that the logics of ISI and post-neoliberalism provide much more room for actors that favor egalitarian social policies to frame their demands as beneficial, and therefore desirable, for the society as a whole. In contrast, neoliberalism as a market-driven ideology is based on a logic that is highly averse to such a framing. Therefore, the discursive power resources of actors pushing for an extension of social policy for low-income groups are significantly greater in a context where ISI or post-neoliberalism is the prevalent development paradigm than in a context of neoliberal prevalence. In the following, this argument will be put on a solid basis by showing how the different logics of the three development paradigms provided or restricted opportunities to frame the extension of social protection for low-income earners in a positive manner. As it is not the aim of this text to undertake an extensive discussion of all the aspects of these paradigms, I will concentrate mainly on those major logics of them that have important implications for the perceived viability and desirability of social protection for low-income earners.<sup>54</sup>

All three development paradigms have in common that they prioritize economic growth as the central means for social progress (Bruton 1998, 904; Peck and Tickell 2002, 394; Sikkink 1991, 40).<sup>55</sup> In this logic, social policy is not only judged for whether it improves general welfare or not, but also for whether it hampers or promotes economic growth. This can be understood as one of the reasons why ministries of economy tend to have significantly more influence on issues that formally fall into the areas of responsibility of the ministries of labor or social security than the other way round (see e.g., Alonso 1997).

While all three paradigms prioritize economic growth, they differ strongly regarding what kind of social policies are perceived to foster or hamper economic growth. There are three issues with particularly strong implications for the perception of what kind of social policy is viable and desirable in the three paradigms: their perspectives on redistribution, on the size and role of the state in the economy, and the conceptualization of the relationship between the national and international economy.

Let me begin with the first issue: the perspectives of the three development paradigms on redistribution. The development of the ISI paradigm during the mid-20th century was influenced by several aspects of Keynesian thinking (Burchardt 2006, 34; Filgueira and Filgueira 2002, 134). Among these aspects was the importance attributed to domestic demand, and hence the view that redistributive policies are compatible with, or even foster, economic growth (Riesco and Draibe 2007, 47). Particularly national-populist versions of the ISI paradigm emphasized the idea that redistributive measures stimulate the ex-

<sup>54</sup> The parts of the following discussion that focus on the neoliberal paradigm build largely on my previous work (Bossert 2011).

<sup>55</sup> Some scholars have recently argued that the new post-neoliberal paradigm is associated with a slightly higher valuation of social progress independent of growth (Grugel and Riggirozzi 2012, 15).

pansion of domestic industries by raising demand for locally produced consumer goods (Sikkink 1991, 34). Later versions of the ISI paradigm, often referred to as developmentalism, had a stronger pro-business orientation and put less emphasis on the positive role of redistribution (Lewis 1993, 191–194; Sikkink 1991, 35, 40). All in all, the ISI paradigm provided a lens through which the redistributive effects that would come along with an extension of social protections for low-income earners appeared to have rather positive effects for both the affected groups and the wider society.

In this regard, the contrast with the neoliberal paradigm could not be more pronounced. For several reasons, neoliberalism renders "issues of redistribution and social investment as antagonistic to the overriding objectives of *economic* development" (Peck and Tickell 2002, 394). The neoliberal paradigm contains the idea that income inequalities provide incentives for economic activity and increase savings and investments. Through this lens, the redistributive effects associated with an expansion of social protection for low-income groups tend to hamper economic growth, which is argued to have negative effects for the society as a whole.

The perspective of the recently emerged post-neoliberal paradigm on redistribution is significantly more benevolent. It largely coincides with the ISI paradigm in that redistributive measures foster domestic demand and thereby contribute to economic growth. Intellectuals who promote the post-neoliberal paradigm have furthermore emphasized the benefits of redistributive social policies, particularly those in favor of the lowest income groups, in terms of their positive effects on human capital formation (Barrientos and Hulme 2008, 16; Riesco and Draibe 2007).

The second important issue on which the three development paradigms differ radically relates to the size of the state and its role in the economy. The ISI paradigm, both in its national-populist and its developmentalist version, rejects the idea of efficient free markets (Bruton 1998, 905–907). Economic development is assumed to require a strong state that protects, promotes, and, if necessary, even constructs industries—a state that develops the needed infrastructure and manages supply and demand through fiscal and monetary policy (Silva 2007, 67). This benign view of the state offers a favorable framework for the idea of a broad, centralized, state-led social security system (Belmartino 2005d, 103–105; Lewis 1993, 188–191; Sikkink 1991, 3–4).

In sharp contrast, the neoliberal paradigm aims "to minimize the state and return power to the private market" (Hackworth and Moriah 2006, 525). This view of the state is related to the "pervasive naturalization of market logics" in the neoliberal paradigm and the conviction that the market is the fairest and most efficient mechanism to allocate resources (Peck and Tickell 2002, 394). Therefore, it promotes the reduction and privatization of social protections (Burchardt 2006, 194; Riesco and Draibe 2007, 51), and constitutes a very unfavorable ideational context for the expansion of social protection for lowincome groups.

In contrast to the market optimism of the neoliberal paradigm, the postneoliberal paradigm promotes a significantly larger and more active role to the state (Grugel and Riggirozzi 2012, 11). However, compared to the ISI paradigm, the post-neoliberal perspective on the role and size of the state in the economy is more cautious, putting more emphasis on regulation than on public industry-building. Furthermore, the post-neoliberal paradigm favors fiscal discipline and comparatively cautious monetary policies (Grugel and Riggirozzi 2012, 4). But in contrast to neoliberalism, the new paradigm emphasizes the expansion of state revenue rather than austerity as the right means to achieve fiscal equilibrium. In the face of the sharp increase of primary product prices on the world market during the 2000s, this increase in revenue builds heavily on the taxation, and in some cases the nationalization, of primary goods exports (Burchardt 2016, 60–69, Grugel and Riggirozzi 2012, 8–9; Matthes 2012, 20– 27). In this sense, the emerging paradigm partially resembles the national-populist version of the ISI paradigm in which revenues from primary goods production were often captured by the state, in one way or another, and used to stimulate domestic demand through social expenditure and infrastructural programs (e.g., O'Donnel 1977, 533–546; Rofman and Romero 1997, 165–166). All in all, the benign perspective on state intervention in the post-neoliberal paradigm provides opportunities for a positive framing of social protection measures for low-income earners.

A third pivotal difference among the three paradigms is the way they conceptualize the relationship between global and local economy and the consequences that this conceptualization has for the perception of what kind of social policy is viable and desirable. The ISI paradigm was shaped by the idea that the "division of labor between the rich countries and the poor ones seemed to doom the latter to permanent poverty" (Bruton 1998, 905). In order to change this situation, it contained the view that peripheral countries needed to transform their productive structure by expanding their industrial sectors. In turn, this structural transformation was perceived to be only possible through a partial dissociation from the world market by protecting national industries from imports and by managing capital flows (Bruton 1998, 905; Silva 2007, 67). In Latin America, the paradigm was domestic market oriented, and hence the competitiveness of national industries on international markets constituted an issue of secondary importance, if any importance at all (Sikkink 1991, 3–4). All in all, the conceptualization of the relationship between the global and the national economy in the ISI paradigm provided local governments with a relatively free hand to design social policy.

The neoliberal view of this relationship differs radically. According to it, development is only possible by attaining competitiveness on the global market. Seen through this lens, countries are compelled to compete with each other to become the most attractive business location (Brand 2005, 9). Progressive taxation and redistributive welfare states are perceived to constitute a disadvantage in this competition as they are assumed to deter both investors and highly skilled employees. Moreover, the neoliberal paradigm suggests that welfare states should be cautious not to create upward pressures on wage levels through extensive job and income protections, as these "artificially high" wage levels are supposed to render the local economy less competitive, particularly in tradeable goods sectors. In sum, the neoliberal paradigm promotes the view that redistributive welfare states are reducing competitiveness and therefore it provides little space to frame the expansion of social protections for low-income groups in positive terms.

The emerging post-neoliberal paradigm is much less optimistic about free trade. Several post-neoliberal governments have reintroduced or reinforced tariffs, trade barriers, and capital flow regulations. At the same time, the new paradigm is more attentive to trade balances and the primary goods producing sectors than the ISI paradigm. While seldom pursuing open political alliances with primary product producers, Latin American governments of the left have enabled these sectors to considerably expand the volume of production (Brand 2016, 21–27; Matthes 2012, 62–73). Regarding social policy, the post-neoliberal paradigm emphasizes its potentially positive effects on competitiveness, for example through facilitating education and diminishing social conflict, and therefore constitutes a benign ideational framework for the expansion of social protections for low-income earners.

All in all, both the ISI paradigm and the post-neoliberal paradigm provide much more favorable ideational contexts to mobilize mass support for an expansion of social protections for low-income earners than the neoliberal paradigm. Hence, they provide significantly more discursive power resources to actors that favor such policies.<sup>56</sup>

#### **The Interdependence of Power Resources**

The distinction of these four types of power resources as separate categories serves to operationalize the extended version of the power resources approach developed in this study. The separation into different categories shall, however, not obfuscate the multiple ways in which these types of power resources interact with each other.

In any concrete political struggle over social policy, different actors can mobilize a combination of different degrees of associational, structural, institu-

<sup>56</sup> Although this is to my knowledge the first attempt to systematically integrate the role of development paradigms in shaping the power resources of different actors in the analysis of social policy in Latin America, former studies have repeatedly mentioned the impact of these paradigms on the distribution of power and opportunities among actors (e.g., Boris 2013, 148–149; Huber and Stephens 2001, 240)

tional, and discursive power resources. Their overall capacity to prevail in the policy process will finally depend on both the combined level of power resources compared to that of contending actors and the capacity to mobilize these resources in a strategically intelligent way. For example, democracy may provide left parties with the opportunity to appeal to the lower popular classes with a progressive social policy agenda, but the prevalence of neoliberal thinking and a lack of party membership might make it difficult to convince significant numbers of supporters of the viability and desirability of this policy agenda.

Furthermore, the different types of power resources have important feedback effects on each other. For example, the distribution of associational power resources may have medium- and long-term effects on the prevalence of different discursive paradigms because these associational power resources might be used to shape public opinion in a way favorable to the actor's interests. Another example is that high unemployment and low industrialization, and thus low structural power resources for working-class organizations, might contribute to a weakening of labor's associational power resources because the fear of unemployment might de-motivate attempts to organize and the dispersion of workers among small enterprises might make organizing strategies difficult to pursue. The list of examples could be extended much further. Yet, for the purpose of this analysis, it is neither possible nor necessary to discuss all of the multiple feedback effects among the different types of power resources. As these feedback effects are not at the center of this study, they will be mentioned in the empirical analysis where useful, but not systematically discussed.

## 3 Methodological Approach

My methodological approach is rooted in the tradition of historical political analysis*,* a mode of analysis interested in examining causal configurations,<sup>57</sup> mechanisms,<sup>58</sup> and processes<sup>59</sup> of macro-political, social, and economic change (Mahoney and Rueschemeyer 2003; Tilly 2008, 133–159). Historical political analyses draw great strength from their close focus on one or few cases over longer periods of time and their rich treatment of political processes, which allows the construction and testing of complex categorical variables (Collier and Collier [1991] 2009, 20; Rueschemeyer 2003, 305–336). By combining process tracing with structured and contextualized comparisons of episodes across or within countries, the approach can provide compelling empirical leverage even for theoretically complex arguments (Collier 1993; Collier and Collier [1991] 2009, 20; Mahoney and Rueschemeyer 2003; Tilly 2008, 136–140).<sup>60</sup>

The processes of macro-political change of interest in this analysis are those related to significant changes in the level of social protection for lowincome earners between 1943 and 2015 in Argentina. Specifically, I focus on two important areas of social protection: income transfers and health care. My aim is to identify medium and long-term policy patterns and to investigate the underlying causal configurations, mechanisms, and processes.

For the identification of the relevant policy patterns, two broad criteria will be applied. The first criterion refers to whether social policy reforms extend or reduce the coverage of income transfers and health care provision among low-

<sup>57</sup> Under causal configurations I understand the particular combinations and interactions of factors that correlate with the outcome under study. Based on the theoretical approach outlined in the last chapter, key factors are the constellation of actors and alliances engaging in social politics and the level of popular class power resources.

<sup>58</sup> In accordance with Tilly (2008, 139) mechanisms are defined as "events that alter relations among some specified set of elements." The attempt of actors to appeal to lower popular class voters, for example, is a mechanism that is assumed to have consequences for the process of social policy reform.

<sup>59</sup> In accordance with Tilly (2008, 139), processes are defined as "causal chains, sequences, and combinations" usually involving a variety of mechanisms.

<sup>60</sup> With regard to ontology and epistemology my analysis is rooted in the critical realist perspective. On the level of ontology critical realists assume that there exists a reality that is independent of our perceptions of it. On the level of epistemology, however, critical realists disagree with positivists as they do not believe in the ability to objectively perceive or measure reality (Furlong and Marsh 2010, 184–211). On these grounds critical realists argue that the sophistication of the methodological approach and its consistency with the theoretical framework are of central importance. Yet, in contrast to positivist research, critical realists do not aspire to reveal the truth, but aim at providing as much leverage as possible to support the plausibility of the argument.

income earners. A significant change of coverage among low-income earners can result from the implementation or abolition of social programs as well as changes in the rules of access to existing policies. The second criterion refers to whether the quality of benefits that can be accessed by low-income earners is improved or reduced by social policy reforms.<sup>61</sup> 

Because social policy change can only partly be measured in numbers and because the interaction between causal factors often follows complex patterns, a qualitative rather than quantitative approach is appropriate. In order to be able to undertake an intensive dialogue between theory and evidence, I focus on only one case, Argentina, but trace social policy development in depth. This makes it possible to go beyond mere hypothesis testing and opens space for the discovery of novel explanations and the refining of theoretical categories (Mahoney and Rueschemeyer 2003, 13).

In accordance with my theoretical framework, I have chosen a long timeframe for the analysis. Theoretically important factors, such as path dependencies, the distribution of power resources, and the composition of governing alliances tend to change and develop consequences over the long run, and hence are best studied over long periods. The analysis begins with the *coup d'état* of 1943, which initiated a period of social policy expansion and marked the passage from a system of fragmented social policies with low coverage to a broader welfare state. Furthermore, this initial expansion period lay important institutional foundations that would shape the posterior dynamics of social politics (see e.g., Collier and Collier [1991] 2009; Haggard and Kaufman 2008; Isuani 2008). The end of the period of study, 2015, has been chosen to include the last complete cycle of social policy expansion.<sup>62</sup>

Argentina has been chosen as a country case for four main reasons. First, as argued in more detail in the introduction, Argentina constitutes a paradigmatic case regarding several puzzling theoretical questions that have to be addressed if we want to gain a better understanding of why Latin American social policy regimes developed in a way that excludes or disadvantages low-income earners. Second, Argentina experienced several periods of significant expansion and retrenchment of social protection for low-income earners, so it is particularly suitable for the application of within case comparisons between different subperiods. These comparisons of different periods can not only shed

<sup>61</sup> In the case of income transfers a qualitative change mainly implies that the real value of a benefit increases or decreases. Important social policy reforms in this sense comprise, for example, increases or reductions of minimum pensions or of social assistance benefits. With regard to health care, qualitative change will be mainly assessed by evaluating the evolution of expenditure as a share of the GDP in relation to the covered population and by taking into account the analyses of health experts regarding the effects of different reforms on the quality of health care provision.

<sup>62</sup> The 2015 presidential elections led to the assumption of a new governing alliance with a considerably different social policy agenda.

light on the patterns of policy change, but also allow us to observe how different patterns of change correlate with different constellations of actors and socio-economic, institutional, and discursive environments. Third, the intensity of the distributive struggle in Argentina since the mid-20th century and the broad range of available primary and secondary sources to reconstruct these struggles facilitates the undertaking of process tracing. Through this method, whether and in what way different underlying constellations and contextual factors influenced social policy reform processes can be explored, as can what mechanisms constituted the causal linkage between them and the resulting policy outcomes. Fourth, Argentina shares a range of structural and politico-cultural characteristics with several other Latin American countries, such as the existence of a broad informal sector, dependency on primary good exports, the emergence of a populist popular class-based party and recurrent periods of authoritarian governments (Burchardt 2006; Collier and Collier [1991] 2009). Therefore it is reasonable to assume that theoretical abstractions generated on the basis of the Argentine case are also meaningful for other Latin American countries.<sup>63</sup> At the same time, these characteristics differentiate Argentina and other Latin American countries from most western countries, which makes Argentina an interesting case for theory development and specification off the beaten track of euro-centered welfare state research.

## **Two Methodological Strategies**

I will apply two distinct methodological strategies in my analysis, each of which contributes in a specific way to the generation of compelling empirical leverage: a most-similar systems comparison between different periods of social policy development in Argentina and process tracing.

## **Most-Similar Systems Comparison**

Most similar systems designs try to emulate experimental designs. Cases for comparison are selected based on similarity regarding as many factors of potential influence as possible. Differences in the outcomes of interest can therefore be assumed to be related to the few differing factors of potential influence in the selected cases (George and Bennett 2005, 164–165). Referring to country cases, Adam Przeworski and Henry Teune (1970, 34) describe the logic of this approach in the following way:

(1) The factors that are common to the countries are irrelevant in determining the behavior being explained since different patterns of behavior are observed among

<sup>63</sup> I agree with Mainwaring and Pérez-Liñán (2005, 18) that regions such as Latin America possess "distinctive dynamics and intra-regional influences" and that therefore "delimiting some generalizations by regions is useful."

systems sharing these factors. (2) Any set of variables that differentiates these systems in a manner corresponding to the observed differences of behavior (or any interaction among these differences) can be considered as explaining these patterns of behavior.

In practical terms, however, similar systems designs are difficult to achieve when comparing country cases. Although there are countries in Latin America that share a range of similarities, the set of differing factors nevertheless tends to be sufficiently large as to make it impossible to identify those factors that might in fact be causing different outcomes. Uruguay and Argentina, for example, share a range of factors that have the potential to influence social policy outcomes, such as social structure, cultural legacy, and economic development, but there remain many potentially important differences such as different party systems, regime types, social policy legacies, and trade union movements. The number of potentially causal differences is too great to permit analysts to clearly identify those factors that are related to different social policy outcomes. In sum, due to the limited number of countries available for case selection, country comparisons face severe difficulties for the construction of mostsimilar-systems designs (George and Bennett 2005, 164; Przeworski and Teune 1970, 34).

The problem of finding sufficiently similar cases can be overcome by comparing different periods within one country instead of comparing different countries (George and Bennett 2005, 166–167). For the purpose of this study, Argentina provides particularly appropriate conditions for the application of period comparisons. There have been periods of accelerated expansion, or at least significant attempts to achieve such expansion, and periods of reduction or stagnation of social protection for low-income earners. We can observe expansive phases in the 1940s and early 1950s, in the first half of the 1970s and since 2002 (Lo Vuolo and Barbeito 1998a; Belmartino 2005a; 2005d; Golbert 2004; 2008; Alonso and Di Costa 2013). The periods in between have been marked by either stagnation or significant backlashes (Golbert 2010; Mesa-Lago 1989; Vargas de Flood 2006). Hence, periods of expansion and retrenchment of social protection for low-income earners can be distinguished sufficiently clearly to permit comparisons of those periods to each other. This enables me to identify those factors of the underlying political economy that differ between periods of expansion and retrenchment and so might be causally related to these diverging trends in social policy.

#### **Process Tracing**

Comparative historical analyses have achieved growing recognition using process tracing in addition to comparative methods (see e.g., Collier and Collier [1991] 2009; Haggard and Kaufman 2008; Rueschemeyer, Huber Stephens, and Stephens 1992). Such a twofold strategy has helped to overcome limitations faced by either of the two strategies.

Process tracing aims to provide leverage for theoretical arguments through the engagement "in a close processual analysis of the unfolding of events over time" (Collier 1993, 115). As investigators are in practice not able to observe processes in their entirety, process tracing rests on the observation and description of key events pertaining to the process of interest (Collier 2011, 823). These events are brought into a chronological sequence, which enables the investigator to analyze how different events and factors added up to culminate in the outcome of interest (Bennett 2010, 208–209).

Through the application of process tracing, I will systematically look for answers to research questions64 that will enable me to assess, modify, and specify the previously elaborated theoretical categories. It also allows me to discover new elements of theoretical importance for the politics of social protection for low-income earners (Collier 2011, 823–824). Argentina is very well suited for the application of process tracing as there have been open struggles over the extension and retrenchment of social protections for low-income earners. A context related analysis of these struggles will provide insights into the social policy interests and strategies pursued by different actors. Furthermore, important long-term transformations occurred in the theoretical categories established in the last chapter. I will therefore be able to observe the impact of social, institutional, and discursive change on the interests and strategies pursued by these actors and their capacities to prevail in the policy process.

Process tracing combines excellently with comparative strategies. At least three important limitations of the latter can be addressed through process tracing. First, comparative strategies are always only approximations of the ideal types formulated by authors such as Mills, Przeworski and Teune. In practice, it is hardly possible to find two or more cases that share all but one potentially causal factor, as proposed by Mill's method of difference, or two or more cases that differ on all but one potentially causal factor, as proposed by Mill's method of agreement. Therefore, purely comparative designs must deal with what Przeworski and Teune (1970, 34) call overdetermination. Comparisons help to limit the range of factors that are plausibly related to the outcomes of interest, but difficulties remain for identifying which of these factors are indeed

<sup>64</sup> Which actors are initiating reform processes?

Why are they doing so and how far are their motivations influenced by such factors as vested interests created through policy legacies, the ideological and social bases of these actors, and regime type?

Which actors support such reform proposals, which actors oppose them and why are they doing so?

Which actors and alliances manage to prevail in the process?

Why do they prevail and what role do associational, structural, institutional and discursive power resources play in these processes?

important and which are not. Even when comparing periods within one country, we can usually observe changes in more than one factor that could plausibly be related to the policy outcome of interest (George and Bennett 2005, 166–167). Process tracing constitutes an effective means of dealing with this limitation. Through the reconstruction of the political process that leads to the policy outcome of interest, which of these factors played an important role can be assessed (Bennett 2010, 208–209).

Charles Tilly (1997, 44) points out a second limitation of purely comparative strategies when he argues that Mill's methods always remained "fatally vulnerable to the allegation that a hitherto-unsuspected cause was operating." The tracing of political processes through the reconstruction of verifiable causal stories tends to reveal such unsuspected causes and is therefore well suited to contribute to theory building and the conceptual elaboration of new analytical categories (George and Bennett 2005, 215; Tilly 1997, 48).

The third limitation of purely comparative strategies is that they provide little information about the interaction of different factors and the mechanisms through which these factors influence social policy outcomes. Process tracing can illuminate such mechanisms and interaction effects through the reconstruction of the sequence of events that leads to a social policy reform and the role of different factors that are involved in the unfolding of these events (Bennett 2010, 208–209).

#### **Operationalization of Theoretical Concepts, Methodological Strategies and Structure of the Empirical Analysis**

The main argument outlined in the theory chapter is that an expansion of social protections for low-income earners is most likely when popular class forces have extensive power resources that enable them to push for, improve, or defend such policies. The main mechanism through which popular class forces influence social policy is assumed to be the formation of governing alliances.

To systematically assess this argument through the application of period comparison and process tracing, the empirical analysis will be structured in a consistent format. A separate subchapter will be dedicated to each period of expansion and retrenchment of social protection for low-income earners. At the beginning of each subchapter, the social, economic and political context of the period will be outlined and the distribution of power resources analyzed.<sup>65</sup> At the end of each subchapter, the patterns and results of social policy change towards low-income earners will be summarized.<sup>66</sup> This format facilitates the

<sup>65</sup> For this purpose, see the headline *The Political Arena: Social and Political Context, Constellation of Actors, and Distribution of Power Resources* at the beginning of each subchapter.

<sup>66</sup> See the headline *Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects* at the end of each subchapter.

comparison of periods and the identification of correlations between the power resources of popular class forces and the expansion or retrenchment of social policies.

For the process tracing analysis each subchapter contains at its center an extensive part that is dedicated to the reconstruction of the political processes that led to the formation of differing governing alliances and the implementation of different policy agendas. This part reveals for each period the concrete mechanisms that linked the distribution of power resources, the formation of differing alliances and the policy outcomes. At the same time, it also shows which strategies actors and alliances applied, which types of power resources they mobilized to make their interests prevail in a certain context, and how this influenced the outcomes in terms of social policy.

To assess the strength of popular class forces I have defined four types of power resources, each of which needs to be operationalized through the selection of indicators for the empirical analysis.

The first type has been termed *associational power resources*. Associational power resources stem from the construction of organizations such as parties, trade unions, and social movements. In the case of parties, I will take parliament and cabinet shares as the main indicators of their associational power. In the case of trade unions and social movements, I will use estimates of their membership as an indicator for their associational strength.

The second type has been termed *structural power resources*. Structural power resources stem from the characteristics of the economic system and its current dynamic. In capitalist systems, economic elites have systemic advantages over popular class forces regarding structural power resources. Yet, under circumstances of tight labor markets and capital-intensive production, popular class actors' structural power resources are significantly enhanced and the advantage of the economic elite is diminished. In order to account for such variance, I will use estimates of unemployment and, where available, informality as indicators for the power resources that stem from tight labor markets. In order to account for structural power resources stemming from capital-intensive production, I will use the percentage of the workforce occupied in industries as an indicator. Low unemployment and informality and high degrees of industrialization are assumed to increase the structural power resources available to popular class actors.

The third type has been termed *institutional power resources.* Here I use regime type classification as the main indicator. There exists an abundant literature on the classification of regimes in Latin America (see e.g., Collier and Levitsky 1997; Dahl 1971; Rueschemeyer, Huber Stephens, and Stephens 1992; for an overview see Lauth 2010). Most studies are inspired in one or another way by Robert Dahl's (1971, 3) eight criteria for polyarchy (e.g., Haggard and Kaufman 2008; McGuire 2010).67 McGuire (2010) and Rueschemeyer, Huber Stephens, and Stephens (1992) combine several of these criteria to form three broader criteria which they use to operationalize democracy. The set of criteria used by Rueschemeyer, Huber Stephens, and Stephens (1992),<sup>68</sup> slightly modified, is adequate for my analysis because it fulfills the requirements set by my theoretical premises and has been successfully applied to several Latin American countries. Democracy is operationalized along the following three criteria: first, regular, free and fair elections with universal and equal suffrage; second, freedoms of association and expression and existence of alternative sources of information; and third, dependence of the key decision makers in the state apparatuses on electoral processes or responsibility to elected bodies.69 The first criterion is directly linked to the theoretical assumption that mobilizing electoral support of low-income earners through social policy proposals constitutes a plausible strategy for governments to reach legitimacy and for oppositional forces to broaden their political support base and increase their electoral chances. Where elections are not free or fair, or large shares of the low-income earners are excluded from suffrage, such power and legitimization strategies lose significantly in attractiveness to political actors. The second criterion is linked to the assumption that freedom of association and expression provides chances for low-income earners to organize, which constitutes a central means of building up power resources to influence state policy. In political regimes that ban some or all these organizations, influ-

<sup>67</sup> The eight criteria established by Dahl (1971, 3) are: "1. Freedom to form and join organizations; 2. Freedom of expression; 3. Right to vote; 4. Eligibility for public office; 5. Right of political leaders to compete for support; 6. Alternative sources of information; 7. Free and fair elections; 8. Institutions for making government policies depend on votes and other expressions of preference."

<sup>68</sup> According to Rueschemeyer, Huber Stephens, and Stephens (1992, 43) democracy "entails, first, regular, free and fair elections of representatives with universal and equal suffrage, second, responsibility of the state apparatus to the elected parliament (possibly complemented by direct election of the head of the executive), and third, the freedoms of expression and association as well as the protection of individual rights against arbitrary state action."

<sup>69</sup> This operationalization of democracy is a minimalist one, which leaves out a variety of issues that are dealt with in more complex conceptualizations. Notwithstanding, as regime type is only one among several factors considered in this study, its operationalization requires a reduction of complexity in order to focus on those elements crucial for the broader theoretical framework proposed in this study. Besides this, the operationalization is deliberately based on a formalist conceptualization of democracy which does not include issues such as the socio-economic requirements of real democratic participation. Despite my agreement with broader conceptualizations of democracy, the decision to adopt a formalist conceptualization has been taken consciously, because issues such as the impact of particular socio-economic structures on decision-making (and non-decision-making) processes are explicitly dealt with in separate research categories.

encing social policy through the construction of social movements becomes significantly more difficult. Finally, the third criterion is necessary to make sense of the first two criteria and the related theoretical assumptions. If parliaments or presidents are elected but have little influence because the key decisions are reserved to non-elected military officials, elections and their outcomes lose much of their importance. I will follow the example of Haggard and Kaufman (2008), McGuire (2010) and Rueschemeyer, Huber Stephens, and Stephens (1992) and define authoritarian regimes as those regimes that significantly violate the criteria for democracy established above. As the continuum of regimes that lie between the ideal types of democratic and authoritarian regimes is large, it is useful to define an intermediate category which I term semi-authoritarian regimes. These regimes partially comply with the above-mentioned criteria, such as free and fair elections, but violate other aspects, for example through veto-powers of military officials or the ban of major political forces. Each of these regime types is considered to cause a different distribution of institutional power resources. Democratic regimes are considered to provide high, semi-authoritarian regimes are considered to provide medium, and authoritarian regimes are considered to provide low institutional power resources to the popular classes.

The fourth type has been termed *discursive power resources*. This category rests on the assumption that actors have better chances to prevail in the policy process when they are able to frame their particular goals as viable or even desirable for the wider society. This ability, in practice, heavily depended on the prevalence of different development paradigms, as these had vast influence on the public opinion, expert communities, and political circles in Latin America. The ISI paradigm and the post-neoliberal paradigm provided much more favorable conditions for framing egalitarian social policies as viable and desirable for the society as a whole than the neoliberal paradigm. Therefore, the ISI and post-neoliberal paradigms provide actors proposing such measures with high and the neoliberal paradigm with low discursive power resources.

In accordance with Guillermo O'Donnell (1977, 551), a governing alliance is operationalized as "an alliance that imposes, through the institutional system of the state, policies conforming to the orientations and demands of its members." On this basis, the analysis of the social policy making processes in each subchapter will begin with the reconstruction of the negotiations, events and actions that resulted in the construction of the governing alliance that ruled during the respective period. This reconstruction will enable me to observe not only the composition of the alliance, but also to assess the relative weight and influence of different actors within it. Following this step, I will chronologically trace all major social policy reform processes during the period, examining the role of different actors and reconstructing the way in which the mobilization of different power resources affected policy outcomes.

### **Sources of Empirical Information and Modes of Data-Collection**

For the implementation of the two methodological strategies, I draw on several complementary sources of empirical information and modes of data collection. This way I am able to obtain different perspectives on the subject and to compensate for limitations of each source of information with the strengths of others (Pickel 2009, 517–522). Due to the diverse kinds of information required and the long periods of time under study, this is characteristic for most historical analyses of macro-political processes (see e.g., Collier and Collier [1991] 2009; Haggard and Kaufman 2008; Rueschemeyer, Huber Stephens, and Stephens 1992).

Before I turn to the description of these sources and modes of data collection, it is useful to outline the kind of empirical information required for the analysis. Three main categories of necessary information can be distinguished: a) information on the structural, institutional and discursive contexts within which struggles over social protection for low-income earners took place; b) information on concrete processes of social policy change and the main actors involved in them; and c) information on the resulting social policies and their consequences for low-income earners.

Secondary literature constitutes an important source for all three kinds of information. This is a common feature of most historical analyses of social policy change (see e.g., Haggard and Kaufman 2008; Huber and Stephens 2012; McGuire 2010). Only this way is it possible to achieve adequate contextual knowledge and to account for macro-political change over long periods of time. The information on a)<sup>70</sup> and c)<sup>71</sup> available in the secondary literature is abundant. Significantly less complete is the secondary literature on b), the political processes underlying social policy change. While some processes of reform are relatively well researched, there is a lack of sufficiently detailed information on others. In short, the secondary literature on the main political processes of social policy change is relatively abundant for the early period of social policy expansion between 1943 and 1955 and the retrenchment of the 1990s. Regarding the main social policy reform processes during the remaining years, primary sources of information have to be consulted more extensively.<sup>72</sup>

<sup>70</sup> For comprehensive analyses of the structural, institutional and discursive transformations in Argentina since the mid-20th century see for example Adamovski 2012; Basualdo 2010 and 2011; James 1988; McGuire 1997; and Sikkink 1991.

<sup>71</sup> For comprehensive analyses of social policy development in Argentina since the mid-20th century see for example Arza 2010; Belmartino 2005a; Golbert 2010; Feldman, Golbert, and Isuani 1988; and Lo Vuolo and Barbeito 1998a.

<sup>72</sup> As part of the generally immense academic interest in the early phase of Peronism, there exists considerable secondary literature on the process of social policy expansion between 1943 and 1955 (see e.g., Andrenacci, Falappa, and Lvovich 2004; Belmartino 2005a; Berrotarán, Jáuregui, and Rougier 2004; Collier and Collier [1991] 2009; Isuani

Besides secondary literature, another important source of information for a) and c) are statistical databases. The national statistics institute INDEC, the SEDLAC Database (CEDLAS and World Bank 2018), CEPAL, the ILO, the World Bank, and the Groningen Growth and Development Centre (Timmer, de Vries, and de Vries 2014) provide ample statistical material regarding the transformations of the social and economic structures during the last 70 years in Argentina. Datasets of the Ministry of Economy contain disaggregated public expenditure data that can be used for an assessment of changes in the size and structure of social expenditure. The social insurance institute ANSES and the ministries of labor, health, and social development furthermore provide data on social programs under their responsibility. The World Values Survey (Inglehart et al. 2014) provides survey data that allows us to track the rise and demise of neoliberalism as the dominant discursive paradigm in public opinion since the early 1990s.

Where secondary information on c) is unavailable, insufficient, or not reliable, I additionally analyze legal texts and official documents to assess programmatic change in social policy. This has often been necessary since most secondary literature is focused on welfare regime development in general and therefore frequently lacks specific information regarding the implications of different reform measures for low-income earners.

In sum, secondary sources, statistical databases, legal texts, and official documents provide ample information for a) and c). Information stemming from these sources for b) is sufficiently complete only for certain periods and reform processes. In order to fill this gap, I will consult different additional primary sources of information. One important primary source of information on those processes are 30 interviews with 28 experts that were conducted during research stays in Argentina in 2013 and 2014. Expert interviews are particularly useful for the reconstruction of political processes, relatively time efficient and can be adequately focused on missing pieces of information (see e.g., Lauth, Pickel, and Pickel 2009, 167; Mieg and Brunner 2004, 199, 212ff). For my purpose, experts are those persons that due to their participation or other reasons possess special knowledge about the actors and actions involved in processes of social policy change in Argentina (see, Meuser and Nagel 2009, 467ff; Mieg and Brunner 2004, 212). The interviewed persons are former

<sup>2008;</sup> Lewis and Lloyd-Sherlock 2009; Ramacciotti 2009; Torre and Pastoriza 2002). The political processes of later periods of expansion of social protections during the 1970s, the 2000s and the 2010s have been less often analyzed in the literature (see e.g., Arza 2012a; 2012b; Belmartino 2005a; Feldman et al. 1988; Golbert 2012). The political processes underlying social policy retrenchment during the different authoritarian regimes between 1955 and 1983 and the failed attempts of social policy universalization during the democratic government of the 1980s have been the focus of significantly fewer analyses (see e.g.; Belmartino 2005a). More literature exists, in turn, on the political processes of social policy retrenchment during the 1990s (see e.g., Alonso 1998; 2007; Lo Vuolo and Barbeito 1998a; Goldberg and Lo Vuolo 2006; Kay 2000).

ministers and ministerial employees, members of the parliament, political advisers, trade union leaders and representatives, social movement leaders<sup>73</sup> and representatives, and scientific experts. Most of these experts were selected because they have been directly involved in important reform processes since the 1970s. Their participation occurred on different levels of these processes. While some experts, such as former ministers and members of parliament, possess special knowledge about the governmental and parliamentary level of the decision-making process, others, such as gender secretaries in trade unions or representatives of political parties, possess special knowledge about the internal dynamics of agenda-setting within collective actors of major importance. Only a small number of the interviewed experts are uniquely scientific experts on the subject who have no further involvement in political processes.74 In general, the selection of experts aimed at obtaining different perspectives on the respective processes.<sup>75</sup>

For each interview, a guideline was carefully prepared that took into account both the special knowledge that the respective interviewee could be expected to possess and the theoretical and empirical interest of my analysis (Meuser and Nagel 2009, 472f).<sup>76</sup> Two of these interviews were completely transcribed and thematically analyzed. In accordance with Lauth, Pickel, and Pickel (2009, 177) and Meuser and Nagel (2009, 476 f), the transcription, excerption and content analyses of the other interviews was focused on thematically important parts.

In addition to expert interviews, campaign documents of collective actors and internal documents of the social security administrations were used to confirm and complement information accessed through expert interviews. Furthermore, introductory notes published with laws and decrees often contained

<sup>73</sup> Among these were leaders of feminist, indigenous, small farmers' and unemployed workers' movements.

<sup>74</sup> In practice, scientific social policy experts were often also participants in social policymaking processes. Aldo Isuani and Pablo Vinocur, for example, are both academic social policy researchers, but at the same time political consultants and even occupants of high ranking government positions (secretaries) during the Presidency of De la Rúa. Only the following six interviews were specifically focused on the interviewees' roles as scientific social policy experts: Guillermo Alonso (2014), Laura Golbert (2014), Ana Natalucci (2014), Daniel Nieto (2014), Sebastian Pereyra (2014) and Corina Rodriguez Enriquez (2014). However, even in these cases, all of them were in close contact with policy-makers, political leaders, parties and movements. Their important contribution therefore not only consisted in complementing the academic literature but also in providing numerous contacts to politicians that had been involved in crucial social policy reform processes.

<sup>75</sup> A detailed list of all the interviews can be found at the end of the study together with the literature list.

<sup>76</sup> Discussions with Hans-Jürgen Burchardt and colleagues from the Graduate School of Global Social Policies and Governance have significantly contributed to the elaboration of effective guidelines.

interesting information about what objectives the government was pursuing with the measures and which actors were involved in the elaboration processes. For more recent reform processes on which secondary information is particularly scarce, I resorted to the online archives of three national newspapers (*El Clarín*, *La Nación* and *Página 12*).

## 4 The Politics of Social Protection for Low-Income Earners in Argentina, 1943–2015

With the emergence and rise to power of the populist Peronist movement after 1943, Argentina initiated the passage from a system of until then relatively limited social programs to a mass-based welfare state (Isuani 2008). Today the array of social policies in Argentina comprises everything from public education, health care, and child allowances through social health, pension, and unemployment insurance to social assistance and housing programs. However, the Argentine welfare state has developed in a highly fragmented way and provides significantly different levels of social protection to different segments of the population (Mera 2010; Belmartino 2005b; Lo Vuolo 1998a, 39; Lloyd-Sherlock 2005; Maceira 2009). Over the last 70 years, social protections for low-income earners have been the subject of intense political struggles and different governments have pursued vastly different agendas. While during some periods social policy embarked on a decidedly progressive reform path, during others it underwent retrenchment and regressive remodeling.

In the following, the text will reconstruct this history and analyze the structures, actors and processes that led to the significantly varying outcomes regarding social protections for low-income earners. After a brief part on the historical background before 1943, each period of either expansion, stagnation or retrenchment will be examined in a separate subchapter. In order to systematically explore the structural foundations of policy change, each of these subchapters starts with an overview of the social, economic, and political context of the period. This provides us with an understanding of the political arena, that is, the constellations of actors and interests and the distribution of power resources. On that basis, each subchapter then traces the formation of governing alliances and the subsequent political struggles and processes of reform. The final part of each subchapter reviews the main patterns of social policy change, which facilitates the comparison with other periods.

As the periods differ sharply in duration and the character of their reform processes, the chapters will also differ in their length.

## **4.1 Historical Background and Social Policy before 1943**

Before 1943 public health, pensions and other income transfer policies were very limited. Low-income groups either lacked social protection or depended on the discretionary and often paternalistic support of charity organizations. Nevertheless, the small-scale social policy developments between the early 19th century and 1943 are important to understanding the particular political dynamics of the subsequent construction of a broader welfare state. During this period, the emergence and increasing power resources of middle- and workingclass actors evolved into a challenge to oligarchic rule. Against this background, the latter adopted a double strategy of repression and concessions, including charity and social insurance policies for selected groups of the society. Although these social policies covered a minimal share of the population, they contributed to the development of vested interests, which turned out to have an important impact on later welfare politics.

#### **Incipient Social Policy in the 19th Century**

The first very limited but state-sponsored social policies were implemented by provincial governments during the 1820s (Isuani 2008, 171). The Revolución de Mayo of 1810 initiated a process of independence from Spain and was followed by a series of civil wars that caused severe social and economic difficulties, impeded the consolidation of a national state for several decades and gave political predominance to the provincial governments. These were dominated by the landed oligarchy, who controlled most of the Argentine agrarian economy and concentrated enormous wealth (Rapoport 2006, 31; Rock 1987, 79–104). Against this background, provincial governments engaged in the financing of charity societies that aimed by and large at the legitimization of their rule, the disciplining of the poor and the avoidance of social unrest (Isuani 1985, 117–118; Isuani 2008, 171). The biggest of these societies was the Sociedad de Beneficencia, which was founded in 1823 in the Province of Buenos Aires. Similar associations were created in other provinces as well. Although these societies depended mostly on public funds, they were usually administrated independently by women of the economic and political elite (Golbert 2008, 19). They ran hospitals, orphanages, and emergency accommodations; later some societies also handed out small cash benefits to persons in extreme poverty (Moreno 2004, 73). However, the conditions imposed on those receiving help were severe and the living conditions in these institutions were so adverse that their inhabitants showed a significantly elevated mortality rate (Isuani 2008, 171). All in all, the social protection offered to low-income groups by these societies was very restricted, discretionary and often tied to disciplining conditionality and moral paternalism (Passanante 1987, 30–31).

With the consolidation of the national state towards the 1880s, public education was massively extended (Lo Vuolo and Barbeito 1998b, 117). One of the primary aims of this expansion was to foster national identity, which was perceived by the political elite as a necessary condition for the consolidation of the Argentine nation (Isuani 2008, 172). Cholera and yellow fever epidemics during the 1860s and 1870s contributed to the creation of a National Department of Hygiene and the extension of the network of publicly subsidized hospitals (Golbert 2010, 54–55; Isuani 2008, 172). Nevertheless, state involvement in the health care sector remained mostly limited to the prevention of epidemics (Isuani 2008, 172).

Another form of social protection against health risks emerged with the foundation of mutual benefit societies from the mid-late 19th century on. Nearly all of these associations were established by immigrant workers so that most of them constituted at the same time collective associations based on class belonging or profession and on national origin (Munck 1998, 577). Among other things, these societies frequently fulfilled the functions of a health insurance society (Golbert 2010, 57). Passanante (1987, 66–71) estimates that there were approximately 1,200 mutual benefit societies with over 500,000 members in 1914. According to Cortés, fewer than 10% of these mutual benefit societies belonged to trade unions around the turn of the century (cited in Lo Vuolo and Barbeito 1998b, 118). This situation would change radically towards the second half of the 20th century when union-controlled health insurance funds were consolidated as one of the main pillars of the Argentine health care system.

All in all, 19th-century state involvement in social policy was largely restricted to those basic interventions that were perceived to be necessary to create nationhood and to avoid fundamental threats, such as epidemics and social unrest.

## **Structural Change, the Emergence of New Actors, and the Selective Implementation of Social Insurance Funds**

During the late 19th and early 20th century, Argentina experienced a period of accelerated economic growth and urbanization (Hänsch and Riekenberg 2008, 60). In 1914, Buenos Aires reached a population of 1.5 million inhabitants (Collier and Collier [1991] 2009, 91). Other cities in the interior had also grown significantly, so that Argentina at that time constituted the most urbanized country in Latin America (Hänsch and Riekenberg 2008, 60). In combination with the consolidation of a public administration, a basic state infrastructure and an expansion of urban industries, these developments led to the emergence of comparatively broad middle and working classes (Hänsch and Riekenberg 2008, 60). The size of the industrial labor force increased from approximately 120,000 workers in 1894 to nearly 400,000 in 1914 (Merkx 1969, 91).<sup>77</sup> Not surprisingly, these developments provided a favorable ground for the emergence and growing strength of the labor movement as well as middle-class demands for social and political participation (Collier and Collier [1991] 2009, 91–92).

<sup>77</sup> The share of the manufacturing workers of the total economically active population increased from 7% to 12%.

Although these new actors were still in the process of formation, their growing strength increasingly led to the perception among the conservative political and economic elite that repression alone would no longer be sufficient to confront them (Feldman, Golbert, and Isuani 1988, 28–29; Isuani 2008, 172–173). Furthermore, experiences from Germany and other countries, where conservative governments used social legislation to assure the loyalty of public employees and to contain and divide the labor movement, were not unknown to Argentine elite circles (Golbert 2010; 36–38). These national and international developments contributed to a gradual adoption of a double strategy of repression and social legislation towards these emerging actors (Feldman, Golbert, and Isuani 1988, 28–29).


Table 1: Economic and social context, 1894–1943

\* 1935, \*\* 1943, \*\*\* 1895, \*\*\*\* 1947. Sources: Martínez (1998) for GDP growth; Rapoport (2006, 138 and 229) for industry share of the GDP; Merkx (1969, 91) for labor force in manufacturing industries and Golbert (2010, 33) for share of urban population.

As in many other countries, the first to receive public pension funds were privileged groups of public employees (Mesa-Lago 1978, 11). In 1904, coverage was extended to all employees of the public administration (Lo Vuolo and Barbeito 1998b, 116). While the loyalty of middle sectors in public employment was hence relatively early fostered through social legislation, workingclass demands were still responded to in a much more repressive way. Due to a variety of factors, Argentina had developed an exceptionally strong and militant labor movement since the mid-19th century (Collier and Collier [1991] 2009, 92). Among the structural factors that facilitated this development were the incipient process of industrialization, a rapid progress of urbanization and the absence of an extensive labor reserve due to low levels of unemployment and a small peasantry (Collier and Collier [1991] 2009, 91–92; Lo Vuolo 1998a, 36; O'Donnell 1977, 526–528). Furthermore, most industrial workers of that time were migrants and many of them had already had experiences with trade unionism in their countries of origin and were familiar with anarchist and socialist ideas (Collier and Collier [1991] 2009, 92). In 1901, socialists and anarchists formed the first trade union federation FOA (Federación Obrera Argentina). Under anarchist dominance, it soon changed its name to FORA (Federación Obrera Regional Argentina) (Golbert 2010, 39–40). Only one year later, in 1902, the first Argentine general strike inaugurated two decades of vigorous labor activism and fierce repression (Collier and Collier [1991] 2009, 92). Ten general strikes took place during the first decade of the 20th century alone. As many as 220,000 workers participated in the biggest May Day demonstrations. Between 1907 and 1910, an impressive total of 785 strikes were carried out in Buenos Aires (Collier and Collier [1991] 2009, 93). The government often responded with repression to these protests, which, however, contributed to further radicalization and intensification of labor activism rather than to a pacification of the situation (Golbert 2010, 36).<sup>78</sup> This increasingly aroused preoccupations among the oligarchic elite, who were not only concerned with the magnitude and radical character of labor militancy but also with the fact that intensive strike activity in strategically important sectors of the Argentine agri-export model imposed painful economic costs on them (Isuani 2008, 172–173). In particular, the prolonged strike of the well-organized private railway workers during 1912 led to significant discomfort among the conservative elite (Feldman, Golbert, and Isuani 1988, 29). Although the political elite, and the oligarchy79 that dominated it, ideologically opposed redistributive social policies,80 their inability to control labor militancy with repression alone led to the perception that social policies for selected and particularly dangerous sectors of the working class would constitute a necessary complement to repression (Golbert 2010, 36). Feldman, Golbert, and Isuani's (1988, 29) analysis of parliamentary debates on various initiatives for new social legislation during these years clearly shows that the aim of containing and controlling labor conflicts was the key motivator behind these elite initiatives. While several social policy initiatives failed due to intra-elite resistance, the railway workers strike of 1912 had been threatening enough to lead to the implementation of the first pension fund for a working-class occupational group in Argentina in 1915 (Feldman, Golbert, and Isuani 1988, 29). In order to disincentivize strike participation among railway workers, the law included an

<sup>78</sup> The repression of the May Day demonstration in 1909, for example, led to the organization of a further general strike and the assassination of the police chief of Buenos Aires by an anarchist (Collier and Collier [1991] 2009, 93).

<sup>79</sup> With the term oligarchy I refer to the relatively small and largely conservative elite of big landowners that dominated the Argentine agricultural, financial and commercial sectors until at least the early 1940s. Birle (1995, 68–79) provides a good discussion of the literature on the mechanisms through which the oligarchy also dominated the successive governments during this period.

<sup>80</sup> For a good analysis of the anti-redistributive positions of the employer associations during this period see Birle (1995, 68–78). For a historical reconstruction of employer resistance to redistributive policies during this period see Lvovich (2005).

article which determined that workers fired due to strike activities would completely lose pension rights, including all contributions made until that date (Isuani 1985, 122).

This pattern of a slow, gradual expansion of social insurance coverage to middle-class sectors and selected occupational groups of the working class continued until the early 1940s. In 1921 utility workers<sup>81</sup> and in 1923 bank employees were provided with pension funds (Golbert 2010, 53). In 1939, merchant seaman, print workers, and journalists obtained pension funds (Lewis and Lloyd-Sherlock 2009, 117). In sum, however, these fragmented social policies reached only small minority groups, so that by 1944 merely 7% of the economically active population were covered by pension insurance (Feldman, Golbert, and Isuani 1988, 31). A similar trend can be observed in the provision of public financing for mutual benefit societies, which provided health insurance to their members. Only societies linked to certain groups of public employees, such as police and military officers, and well organized and strategically located working-class sectors, such as the railway workers, received considerable financial support (Tobar 2001, 18).

Figure 1: Incorporation of additional occupational groups and evolution of legal pension coverage, 1904–1939

Sources: Arza (2010, 259).

Neither the democratizing reforms of 1912 nor the rise to power of the Unión Cívica Radical party (UCR) between 1916 and 1930 significantly altered this trend. One reason why democratization did not lead to fundamentally different

<sup>81</sup> Workers of the tramway, telephone, gas and electricity sector.

directions in social policy is that democratization itself was part of the same double strategy of repression and concessions that the oligarchy employed to consolidate its power in the face of emerging middle- and working-class pressures. In line with social insurance funds for public employees and other middle-class occupational groups, the introduction of the free, secret and universal male vote was part of a series of policies implemented by conservative elites in order to reinforce the loyalty of public employees and to win middle-class support as a subordinate force in the governing alliance (Rock 1975, 26–35). Although initially feared by parts of the oligarchy, the UCR governments finally turned out to embody precisely this alliance between a still dominant oligarchy and newly incorporated middle-class sectors (Birle 1995, 76; Rock 1975, 41).<sup>82</sup> Democratization provided institutional power resources for the working class and other marginalized groups as well, but this did not lead to their incorporation into the governing alliance for a variety of reasons. For one thing, popular class forces themselves were neither strong enough nor willing<sup>83</sup> to lead a governmental alliance. Attempts by populist currents within the UCR to incorporate popular class forces into the governmental alliance, in turn, failed due to a mixture of different factors. First, during the late 1910s, the UCR government was not only confronted with labor militancy but also with a wave of massive and violent conservative mobilization as a reaction to the former (Collier and Collier [1991] 2009, 146–147). Second, predominantly anarchist and syndicalist trade unions were largely unwilling to cooperate with the government (Isuani 1985, 118). Third, vast parts of the popular sector, such as women or recent migrants, were impeded from voting (Isuani 1985, 117– 118). Fourth, the violent conservative mobilization, the slow-down of the industrialization process and the rise of unemployment during World War I decimated the associational and structural power resources of the labor movement (Collier and Collier [1991] 2009, 146–147; Golbert 2010, 46, 63; Merkx 1969, 91). And fifth, the dominance of the liberal economic paradigm provided a hostile discursive context for proposals of redistributive social policy. On the grounds of this paradigm, the elite was successful in framing redistributive policy as undesirable for the society as a whole, due to its supposed negative effects on economic growth and work ethic (Lvovich 2005, 143).

After the UCR government intensified its attempts to reduce the influence of the oligarchy during the late 1920s, a right-wing military coup ended the democratic experience and led to a restoration of old mechanisms of oligarchic

<sup>82</sup> The unbroken influence of the oligarchy was reflected in both favorable policies and the presence of members of the oligarchy in pivotal positions of the UCR governments and party leadership (Birle 1995, 76; Hänsch and Riekenberg 2008, 61).

<sup>83</sup> This skepticism towards electoral politics was largely due to the predominance of anarchist and syndicalist currents within the labor movement (Collier and Collier [1991] 2009, 93).

rule, such as electoral fraud and corruption (Birle 1995, 77–78; Collier and Collier [1991] 2009, 149; Riekenberg 2009, 134–135).

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

During the 19th century, public social policy was limited to basic interventions that were perceived to be necessary to avoid fundamental threats and to provide the basis for the formation of an Argentine national identity. Economic growth and urbanization contributed to the emergence of working and middle classes and the successful constitution of collective actors. These actors began to challenge the dominant political role of the conservative oligarchy. Concerning the middle classes, the oligarchic elite pursued a strategy of co-optation and gradual concessions from the late 19th century on. Among these concessions was the extension of certain social benefits to middle-class sectors and the democratizing electoral reform of 1912. This strategy was relatively successful in safeguarding the oligarchy's political dominance by incorporating middleclass sectors into the governing alliance. In contrast, the popular classes remained excluded from this alliance and repression continued to constitute the primary response to their demands. Popular class actors were clearly too weak to dispute the elite–middle class alliance's control over the government. During the first half of the 20th century, The Socialist Party, and later the Confederación General del Trabajo (CGT), demanded universalistic and redistributive social policies but were far from reaching majorities for their proposals in the parliament (Golbert 2010, 64; Lvovich 2005, 140–142). However, some occupational groups that had attained considerable levels of associational and structural power resources, such as railway workers and merchant seamen, were capable of inflicting significant economic harm on the oligarchy. The fact that repression alone proved insufficient to contain the vigorous activism of these unions led the governing alliance to combine repression with fragmented and divisive social policy measures. These were deliberately limited to very small segments of the popular classes. Other segments, which lacked comparable structural and associational power resources, and hence most low-income earners, remained largely excluded from social protection.

The introduction of particularistic social policies for certain occupational groups furthermore created vested interests that gradually began to generate tensions within the labor movement. Although the Argentine labor movement remained left-wing and continued to promote universal and egalitarian social policies, the defense of existing benefits also began to narrow the array of universalization strategies that would find the approval of the most powerful sectors of the movement (Golbert 2010, 64; Lo Vuolo and Barbeito 1998b, 118; Lvovich 2005, 151)

## **4.2 Popular Class Incorporation and Social Policy Expansion, 1943–1955**

Between 1943 and 1955 social policy was massively expanded to cover hitherto excluded groups. This constituted a significant improvement of social protection for low-income earners, regarding both quality and coverage. In political terms, this process was closely related to the construction and consolidation of a new, more inclusionary governing alliance that successively incorporated the working class and the urban and rural marginalized sectors as crucial support bases. The successful formation of this governing alliance, in turn, depended on social, economic and political developments that gradually altered the distribution of power resources in favor of popular class actors and increased their willingness to cooperate with the government.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

The eruption of the world economic crisis in 1929 severely affected Argentina and set in motion significant transformations of the economic, social, and political landscape. The contraction of the global demand for agricultural products and the parallel fall of prices resulted in a reduction of nearly 40% of the export value, which triggered a balance of payments crisis, a forced reduction of imports and a 14% decrease of the GDP in only three years (Birle 1995, 79– 80). The social consequences were dramatic. Unemployment and poverty increased steeply, particularly among rural workers and small farmers (Golbert 2010, 63). Due to the lack of alternative employment opportunities in rural areas, this led to a significant intensification of the urbanization process and a profound deterioration of housing conditions in the urban centers (Birle 1995, 81; Golbert 2010, 63). At the same time, import restrictions imposed after the crisis fueled a process of industrial growth oriented towards the production of goods for the national market (Rofman and Romero 1997, 163–164). The initiation of World War II further reinforced this process, as many countries exporting industrial products to Argentina shifted their resources towards arms production (Rock 1987, 238). These developments had important effects on the social structure. Between 1935 and 1941, the number of workers in manufacturing industries increased from less than 400,000 to over 730,000 and unemployment declined considerably (Birle 1995, 82; Murmis and Portantiero 2004, 140). During the years of the Peronist government, these trends continued so that by 1946 the number of manufacturing workers increased to over a million and full employment was reached (Merkx 1969, 91; Torrado 1992, 175). By 1950, industrial84 employment accounted for nearly 30% of the workforce and over 36% of the GDP (Timmer, de Vries, and de Vries 2014). These developments significantly increased the structural power resources of the popular classes and also provided favorable conditions for the expansion of their associational power resources through unionization.


Table 2: Economic and social context, 1943–1955

Sources: Martínez (1998) for GDP growth 1943 and 1944; Rapoport (2013) for GDP growth and inflation 1945–1955; Timmer, de Vries, and de Vries (2014) for industry share of the GDP and share of industrial employment; Torrado (1992, 175) and Belini (2014) for unemployment.

As in other countries, the economic crisis of the early 1930s led to increasingly widespread doubts regarding the adequacy of the dominant liberal economic paradigm and the capacity of the market to solve the major economic and social problems of the time (Golbert 2008, 21). Nevertheless, despite the de facto occurring process of import substitution and the increasing discredit of the liberal economic paradigm, Keynesian ideas were still largely absent from the debates of the 1930s in Argentina (Lvovich 2005, 143).<sup>85</sup> Towards the late 1930s and early 1940s, this situation began to change. The recurrent failure of economic policies inspired by the liberal paradigm and the rather positive

<sup>84</sup> The numbers for the industrial sector here and later in the text include manufacturing, construction and utilities.

<sup>85</sup> Rather than breaking with the liberal economic paradigm, the governments of the 1930s and early 1940s tried to adapt the agri-export model to the new international framework (Birle 1995, 82).

results of the first attempts of more interventionist policies contributed to the rapid spread of Keynesian ideas around the world (Flier 2005, 197; Sikkink 1991, 41–42). In Latin America, Keynesian ideas merged with those of local economists surrounding the United Nation's Economic Commission for Latin America<sup>86</sup> to form the paradigm of Import Substitution Industrialization (ISI) (Bethell 1996, 216; Burchardt 2006, 34). By the mid-1940s, these ideas came to constitute the dominant economic paradigm in Argentina. They were shared by a wide range of Argentinian intellectuals and economists and for "the majority of the population, some form of import-substituting industrialization, whether national populist or developmentalist, was the most compelling option" (quote: Sikkink 1991, 3; see also: Andrenacci, Falappa, and Lvovich 2004: 90; James 1988, 19–20). While the most important Argentine employers' associations continued to advocate the liberal economic paradigm, it had clearly lost its popular support (Birle 1995, 89–110, Sikkink 1991, 3). The shifting predominance from the liberal economic paradigm to the ISI paradigm was accompanied by related ideational developments in specific social policy fields. In the area of social security, the ILO became globally influential in promoting state-led expansions of social insurance arrangements from the 1930s on (Flier 2005, 199–202; Gaggero and Garro 2004, 175).<sup>87</sup> During the 1940s, the ILO's major concern shifted from the expansion of social insurances to the promotion of broader concepts of social security that would encompass non-salaried parts of the population (Flier 2005, 206–207). The influence of these concepts was additionally reinforced during the 1940s through the widespread international attention to the universalistic prescriptions of the Beveridge Report (Belmartino 2005d, 107). The dramatic social consequences of the crisis and the incapacity of the existing systems of social protection to cope with them contributed to the fact that these international ideas soon became increasingly popular in Argentina (Belmartino 2005d, 123; Flier 2005, 198). Taken together these changes in the ideational context shifted significant discursive power resources to actors who favored the extension of social protections.

The transformations of the social, economic, and discursive context were accompanied by new developments in the realm of actors participating in social politics. The decline of foreign trade and the deterioration of the terms of trade reduced the economic weight of the hitherto dominant agricultural sector and increased the relative weight of the industrial sector (Thorp 1992, 183– 184). In 1943, the value of industrial production surpassed that of agricultural production for the first time (Rock 1987, 232). Within the industrial sector, the weight of local market-oriented small- and medium-size enterprises increased

<sup>86</sup> Elsewhere referred to with its Spanish acronym CEPAL.

<sup>87</sup> The foundation of the Inter-American Conference on Social Security within the ILO in 1942 furthermore served to reinforce such ideas at the regional level (Flier 2005, 199– 202).

with the process of import substitution. Although these industries would have benefited from a more protective state, this did not translate into a repositioning of the main industrial employer association, the UIA. Dominated by large enterprises, the UIA continued to adhere to the liberal economic paradigm and pursued interests quite similar to those of the agricultural, commercial, and financial employers' associations (Birle 1995, 85–86). An extension of social protections to low-income groups was rejected by employers' organizations in concert with conservative politicians, arguing that this would have adverse effects on the economy and the work ethic of the beneficiaries (Golbert 2010, 46; Lvovich 2005, 143).

The labor movement, in contrast, reinforced its commitment to broad and redistributive social policies. In 1930, the socialist and the syndicalist trade union confederations joined forces to form the CGT, which from 1935 was dominated by the more moderate socialist current (Golbert 2010, 64; Murmis and Portantiero 2004, 141). In 1931, the CGT presented its political platform, the Programa Mínimo, which included explicit demands for the establishment of a national social insurance system that would cover unemployment, health, maternity, invalidity, and pension benefits (CGT 1931, reprinted in Golbert 2010, 64). Both the CGT and the Socialist Party promoted universal social policies, which would have benefited the majority of the popular sector, unlike the selective concessions implemented by the exclusionary governing alliances before 1943. While the labor movement possessed strong associational and structural power resources in only a few specific branches of the economy, its ambition was nevertheless to represent the social policy interests of the popular sector as a whole (Gaggero and Garro 2004, 176; Lvovich 2005, 141–142). This universalistic orientation went along with determined efforts to strengthen trade unionism in weakly organized branches in the industrial, service, and the agricultural sectors (Doyon 1984, 205). Although the railway sector remained the best-organized sector, by the early 1940s trade unions in other sectors such as commerce, construction, and the textile industry were growing rapidly (Doyon 1984, 222). Overall trade union membership increased from about 370,000 in 1936 to about 470,000 in 1940 and 530,000 in 1945 (Doyon 1975, 154–160; Murmis and Portantiero 2004, 133).<sup>88</sup> In the late 1930s and early 1940s, tensions within the labor movement emerged between the defense of existing social protections for some occupational groups, the aims of newly organized occupational groups to attain similar protections in the short term, and the universalistic medium- and long-term goals of the CGT and the Socialist Party (Gaggero and Garro, 176; Lvovich 2005, 148). This led to the acceptance of a double strategy on the part of the CGT and the Socialist Party

<sup>88</sup> Trade union density among all wage earners was approximately 11% in 1941 and 1945 (estimates based on Doyon 1975, 154–160). Among industrial workers trade union density was estimated between 20% and 30% in 1939 (Murmis and Portantiero 2004, 134– 135).

that approved the gradual extension of social protections through the creation of new social insurance funds for certain occupational groups in the short-term without renouncing the long-term goal of a universal social security system (Gaggero and Garro, 176; Lvovich 2005, 150).

With the *coup d'état* of 1930, the military entered the political arena, in which it was going to constitute an important actor during the following 53 years. Although there existed different factions within the military, its overall political orientation was right-wing, Catholic, nationalist and authoritarian (Riekenberg 2009, 133). Towards labor and its demands for redistributive policies it took a repressive stance. During the 1930s, parts of the military were increasingly inclined to support industrialization policies, as this was seen to open the way for more national sovereignty and the construction of an independent national arms industry (Hänsch and Riekenberg 2008, 61).


Table 3: Popular class power at the beginning of the period, 1943–1946

In sum, since the early 1930s, the economic, social, and political context had undergone substantial transformations. The labor movement was significantly strengthened. Expanding union affiliation had increased its associational power resources. Low levels of unemployment and the ongoing process of industrialization provided it with additional structural power resources. The increasing predominance of the ISI paradigm, both among experts and among the wider public, generated a favorable societal climate for redistributive social policies and increased the discursive power resources of progressive actors. Arguments of employers' associations and conservative politicians about the disadvantageous economic effects of redistributive social policies were no longer credible in the eyes of the majority. Furthermore, the increasingly

class power

reformist orientation of the labor movement opened new options for strategic alliances. By the early 1940s, these developments had created favorable conditions for the formation of a new, more inclusionary governing alliance that would push forward the extension of social protections to hitherto excluded low-income sectors.

## **Labor Incorporation and Social Policy Expansion under Military Rule, 1943–1946**

The process of re-composition of the governing alliance commenced with the *coup d'état* of 1943 and the designation of Juan Domingo Perón as secretary of labor. From this position, Perón strategically pursued the construction of an alliance with sectors of the trade union movement, which later constituted the principal source of support that enabled Perón to win the presidential election in 1946.

Several concerns motivated the right-wing military coalition that carried out the *coup d'état* on the 4th of June 1943. Among them were the rejection of the civilian government's likely presidential candidate Patron Costas and worries that the ongoing electoral fraud would pave the way for political radicalization and communist subversion (Collier and Collier [1991] 2009, 331; Rock 1987, 247).

During the first months, the military government adopted a decidedly repressive stance towards labor. Shortly before, different approaches to partisan politics had led to a split of the CGT into a CGT-1 and a CGT-2. This division significantly reduced labor's capacity to contest government repression (Doyon 1984, 205). The CGT-2 comprised most of the communist-led unions and those socialist-led unions that closely cooperated with the Socialist Party. It was immediately dissolved after the *coup d'état* and many communist union leaders were arrested (Del Campo 2012, 181). The CGT-1, which comprised most of the unions that were more independent of political parties, was allowed to persist but its powerful railway workers' unions were put under control of military-nominated trustees (Murmis and Portantiero 2004, 150).

Due to internal realignments, however, the military government's repressive approach to labor would soon give way to a more cooperative stance. In October 1943, the Grupo de Oficiales Unidos (GOU) faction became the dominant force within the military government. Juan Domingo Perón, a member of the GOU, became secretary of labor and social security (Collier and Collier [1991] 2009, 331–332). Like most of the military, the GOU was authoritarian, nationalist, and fiercely anti-communist. Notwithstanding, in contrast to sectors of the military that were more closely aligned with the traditional oligarchy, the GOU promoted protectionist industrialization policies that were perceived to increase the economic independence of Argentina and to enable the strengthening of the national arms industry (Birle 1995, 91–92).

While the GOU held a dominant position within the military, the military government pursued policies congruent with the ISI paradigm, which aroused the resistance of the traditional employer sectors, which continued to promote free-market policies based on the liberal paradigm. Employer opposition meant that the military government lacked virtually any powerful civilian support base for its program (Rock 1987, 249–250). The urgent need for civilian support certainly contributed to an increased acceptance within the military of Perón's strategy towards labor. Perón replaced the purely repressive approach with a mixed approach that combined repression of communists and uncooperative unions with massive benefits for unions that were willing to support the military government and himself as secretary of labor (Collier and Collier [1991] 2009, 332; Torre 1989, 539).

Perón had good reasons to reach out for labor support. First, by 1943 labor had achieved unprecedented power resources and the structural conditions were favorable for a further increase. Second, in contrast to the employers' associations, the trade unions and also the unorganized working class were supportive of the industrialization policies based on the ISI paradigm that the military government was hoping to implement (Collier and Collier [1991] 2009, 349; James 1988, 19–20). Third, Perón saw no contradiction between anti-communism and an alliance with the trade union movement. For him, strong labor unions and social policies were the most effective barrier against communism (Birle 1995, 104; Rock 1993, 159). Fourth, the construction of an alliance with organized labor constituted the basis for Perón's personal quest for power and influence (Collier and Collier [1991] 2009, 335).

Given the relative strength of the Argentine trade union movement, Perón had to make unprecedented material and symbolic concessions in order to secure trade union and working-class support. At the same time, the successful construction of an alliance between Perón and a large part of the trade union movement also rested on the fact that by the early 1940s a significant portion of the latter was open to considering such an alliance. One factor contributing to this openness was the ideological moderation of the labor movement and the growing independence of many socialist-led trade unions from the Socialist Party during the 1930s (Murmis and Portantiero 2004, 138). A second factor lies in the traumatic individual and collective experiences that workers and their organizations had undergone during the period of the conservative regime between 1930 and 1943. Average real wages had stagnated between 1929 and 1942, most strikes had been lost and the working class felt excluded from political decisions and stigmatized with low social status (James 1988, 25–30; Murmis and Portantiero 2004, 141–145). Against this background, the unprecedented material concessions and the symbolic recognition that Perón granted to labor fell on fertile ground. Finally, although the trade union movement had gained significantly in size and strength, it lacked a powerful partisan representation. Neither the Socialist nor the Communist Party had managed to win significant electoral support (Riekenberg 2009, 135). For labor, there seemed to be no realistic means to participate in policy making other than cooperating with Perón, at least in the short run.

The first trade unions that openly declared their support for Perón were the railway workers' unions, at that moment by far the biggest and most powerful unions (Del Campo 2012, 249–256; Doyon 1984, 222). In October 1943, Lieutenant Colonel Domingo Mercante, a close ally of Perón and son of a railway unionist, was appointed new trustee of the railway workers' unions. Mercante initiated a close cooperation with the displaced former union leaders and in September 1944 returned union control to them (Del Campo 2012, 254; Murmis and Portantiero 2004, 150). Simultaneously, important social policy concessions were granted. Among the concessions were enormous subsidies to the mutual benefit society of the railway workers. These were used for the construction of polyclinics, hospitals, pharmacies and service centers. To lastingly improve the finances of the trade union-controlled mutual benefit society, an obligatory contribution from all workers in the sector and regular public subsidies were introduced (Del Campo 2012, 201–202). These measures were received with gratitude by the railway unions. On the 26th of April 1944, railway union leaders published an open letter in which they enumerated the multiple benefits that the secretary of labor had implemented in favor of the railway workers and declared their loyalty to him (Del Campo 2012, 252–253).

Soon after, Perón praised the railway workers' decision to support the military government in a way that can be understood as an appeal to the other unions to adhere as well:

The railway workers will always have the glory of having been the first who understood us and supported us. When the history will judge this new era of Argentine social policy, when many will have to be ashamed of having obstructed the course of our great social achievements, the railway union may waive its crest clean and proud because it was the precursor of the triumph of our righteousness over demagoguery, sectarianism, and abstentionism. (Perón quoted in Del Campo 2012, 253–254, author's translation)

The third largest Argentine trade union, the shop workers' union, initially took a critical stance towards the military government. But after the Labor Department had implemented the pension fund for shop workers for which the trade union had intensively campaigned since 1940, this union as well declared its support for Perón and its general secretary, Ángel Borlenghi, grew to be one of Perón's closest allies (Del Campo 2012, 276–277; Lvovich 2005, 150). Over the period 1944 to 1945, the number of trade unions supporting the military government, and in particular Perón, increased rapidly and finally comprised the vast majority of all trade unions (Murmis and Portantiero 2004, 153). The CGT and other labor organizations even organized massive demonstrations in favor of Perón, the military government, and its social and labor policies (Collier and Collier [1991] 2009, 339). Unprecedented material benefits such as favorable social and labor policies and support in collective bargaining were decisive in the decision of many of these unions to support Perón (Collier and Collier [1991] 2009, 337–338; Del Campo 2012, 256–266). In addition to important material benefits, Perón understood the importance of symbolic gestures such as the recognition of labor as a respected political actor and its essential contribution to the country's wealth and culture (James 1988, 7–40).

While cooperative trade unions could expect massive concessions, oppositional trade unions had to expect fierce repression. The communist-led construction workers union, which in 1941 was the second biggest Argentine trade union, rapidly disintegrated after about 1,200 of its leaders and activists were arrested and 130 of its local offices closed. Other influential communist-led trade unions in the steel and meatpacking sector met similar fates. At the same time, Perón supported parallel unions that had often been created beforehand by socialists (Del Campo 2012, 266–273). The capacity of the state to control trade unions was increased even more in 1945 with the issue of Decree-Law 23.852,<sup>89</sup> which provided the Labor Department with the capacity to decide which unions would be officially recognized and hence enabled to undertake collective bargaining. In many cases, union members themselves deserted the organizations that did not shift to support the government in exchange for the enormous benefits that this implied (Collier and Collier [1991] 2009, 339).

In terms of social policy, the process of construction of a new governing alliance between 1943 and 1945 led to an important expansion of coverage. As it had before 1943, the extension of social rights was largely dependent on the distribution of power resources within the popular sector. The powerful railway unions were among the first to receive social policy benefits in the form of favorable regulations and important subsidies that helped to further develop and institutionalize the trade unions' social health insurance. However, the dynamic of social policy expansion after 1943 significantly differed from the gradual expansion prior to that time. Perón was not only interested in controlling and de-radicalizing the trade union movement, but also in gaining its support. This meant that the social policy agenda of the secretary of labor soon reached out for the support of workers and trade unions in other branches. In pursuit of this aim, a variety of measures were implemented that expanded social protections to hitherto excluded groups, many of which constituted sectors with relatively low-income. With the implementation of the pension fund for the commercial sector and the incorporation of new groups of workers into preexisting funds, pension coverage increased steeply from less than 500,000 affiliated workers in 1943 to more than 1,200,000 in 1945 (Gaggero and Garro 2004, 181). In October 1943, the National Division of Public Health and Social Assistance was created in order to increase the state's role in the still largely private health care sector (Andrenacci, Falappa, and Lvovich 2004, 103–104).

<sup>89</sup> More detail on all cited laws and decrees can be found in the *Laws and Decrees* section after the literature list.

One year later, the National Social Security Institute was founded to advance a program of centralization and universalization of social protections inspired by the Beveridge Report (Golbert 2010, 83). Furthermore, several new regulations restricted rents for housing, implemented accident insurance, reinforced workplace regulations, introduced paid vacations and holidays, provided severance pay and restricted dismissal, created labor courts and an annual thirteenth month's wage (Collier and Collier [1991] 2009, 337–338; Golbert 2010, 80–81; Torre and Pastoriza 2002, 280–282)

The process of social policy expansion was, however, not devoid of contradictions. On the one hand, newly created institutions such as the National Social Security Institute promoted the centralization and universalization of social protections in line with ideas that were popular among international experts at that time. Perón, on the other hand, de facto initiated a fragmented expansion of social protections by providing specific occupational groups with pension funds or by supporting selected trade unions by extending their health insurance funds (Andrenacci, Falappa, and Lvovich 2004, 100–101; Belmartino 2005a, 109–129).

## **Democratization, Low-Income Earner Incorporation, and Social Policy Expansion, 1946–1955**

While the military government, and in particular Perón as its secretary of labor, enjoyed widespread support from the union movement and the working class, the employer associations increasingly criticized the government for its social and labor policies (Birle 1995, 94). In June 1945, a broad coalition of employers' associations sent a memorandum to the President and published a manifesto in which they fiercely opposed the program of the secretary of labor. In the following days, large segments of the middle classes, nearly all political parties, most of the media, the US ambassador and important parts of the judiciary and the military voiced support for the manifesto (Del Campo 2012, 239– 248).

Perón rejected the criticism outright, and responded with a radicalization of his discourse, speaking of the "dictatorship of capital" and the intentions of the "eternal economic oligarchy" to "violently shout down the claims of the humble"<sup>90</sup> (quoted in Del Campo 2012, 283–284). The shop workers' union quickly declared its willingness to defend the social policies of the secretary of labor and called upon the other unions to organize a joint demonstration. The vast majority of the trade unions joined the effort and on the 12th of July 300,000 workers gathered on the streets (Del Campo 2012, 288; Murmis and Portantiero 2004, 151–152). From now on, Perón repeatedly emphasized that the working class had to defend the social achievements of the Department of

<sup>90</sup> Translated by the author.

Labor and that with their support many more social improvements were going to be implemented (Del Campo 2012, 305–306).

In response the opposition intensified their own agitation. During the months of August and September, several massive demonstrations were undertaken by oppositional parties, trade unions, students' and employers' associations, some of which were met with brutal repression (Del Campo 2012, 295– 308). In the face of the escalation of the conflict, an oppositional military sector finally achieved the ouster and arrest of Perón in October 1945 (Torre 1989, 545). Most of the organized labor movement and the working class perceived Perón's removal from office as an attack on the social policies implemented under his influence, and hence, as an assault on their interests (Murmis and Portantiero 2004, 162). On the 17th of October, a massive demonstration of workers filled the Plaza de Mayo in front of the Casa Rosada, the government's headquarters. Impressed by the capacity of mobilization and internally divided over the question of Perón, the military government finally decided to release him and to announce elections for the following year (Collier and Collier [1991] 2009, 339–340).

The presidential and legislative elections held on the 24th of February 1946 were a contest between two broad, ideologically heterogeneous, multi-class alliances. Both contained progressive and conservative elements, yet to different degrees. The Democratic Union was supported by the socialist and communist parties, but the overwhelming majority of Democratic Union voters were mobilized by conservative forces such as the UCR,<sup>91</sup> the PDP, and the employers' associations. The Peronist alliance was supported by the Catholic church, the military, a splinter group of the UCR,<sup>92</sup> the conservative Partido Independiente and the progressive Partido Laborista. The Partido Laborista, which had been created by prominent trade union leaders shortly after the announcement of elections, was by far the party with the biggest capacity for electoral mobilization (Birle 1995, 95–96; Collier and Collier [1991] 2009, 345).

Thanks to the massive electoral support of the working class and the lower popular classes, the Peronist alliance won the presidential elections and broad majorities in the Chamber of Deputies and the Senate (Collier and Collier [1991] 2009, 340; Del Campo 2012, 346–348; Levitsky 2003, 27). While Perón had been able to establish a high level of control over the labor movement during his time as secretary of labor, the developments since mid-1945 had also impressively shown how much he depended on labor's powerful support. Perón himself was in many aspects right-wing and authoritarian, but the incorporation of the labor movement and the assumption of influential positions in the government and the parliament by long-standing labor leaders implied the

<sup>91</sup> The UCR cannot be classified as a typical conservative party, but in 1946 it was led by its conservative wing, to which the presidential candidate José Tamborini belonged.

<sup>92</sup> The UCR Renewal Board.

strengthening of progressive elements within the governing alliance. Most of the Partido Laborista leaders had socialist and anarcho-syndicalist backgrounds, and the party's rank and file was influenced by the membership of defecting socialists, communists, Trotskyists and progressive UCR members (Del Campo 2012, 330–332; Murmis and Portantiero 2004, 160). These ideological influences found their expression in the progressive character of the Partido Laborista's declaration of principles and its electoral platform for 1946 (Partido Laborista 1945, 1946). In these documents, the Partido Laborista explicitly rejected racism and other forms of discrimination and inequality and demanded an extension of political rights to women. Furthermore, it demanded far-ranging redistributive measures, including the extension of the social security system to all Argentine residents. As concrete steps in this direction, the electoral program explicitly called for the introduction of pension funds for industrial, rural and domestic service workers and the extension of public health care provision. Interestingly, the platform extended far beyond representing the social policy interests of the organized working class to include those of still unorganized and marginalized groups. This corresponded to both the importance that its left-leaning leaders attributed to solidarity and egalitarianism and their political aim to construct the broadest possible electoral alliance in favor of distributive policies, comprising "workers, employees, peasants, artists, and intellectuals as well as small merchants, manufacturers and agricultural producers"<sup>93</sup> (Partido Laborista 1945). The vital contribution of the Partido Laborista to the electoral victory and Perón's dependence on the support of the popular classes contributed to the fact that during the following years most of the proposed measures would be implemented.

The electoral success was moreover facilitated by the circumstance that the overwhelming majority of the Argentinians at that time were influenced by the ISI paradigm (James 1988, 19–20; Sikkink 1991, 3). This provided the Peronist alliance with additional discursive power resources, in the sense that the promise of economic growth with and through redistribution was credible in the eyes of the majority, while the anti-redistributive arguments of the employers' associations adhering to the Democratic Union had lost much of their persuasiveness (James 1988, 19–20). Redistribution was widely perceived as contributing to economic growth and social development (Sikkink 1991, 31).

Another decisive factor for understanding the following extension of social protections to hitherto unprotected low-income groups is that the return to democracy not only increased the overall power resources of the popular classes but also altered the distribution of power resources within and among these classes. Between 1943 and 1945, Perón had bid mainly for the support of organized labor, knowing that their real and potential associational and structural power resources could constitute a crucial pillar of support. Yet, with the

<sup>93</sup> Author's translation.

developments of 1945 and the return to democracy, Perón felt an urgent need to broaden his support base. Initially, Perón's objective was to construct a coalition that would comprise both labor and capital. However, the principal employers' associations took a clearly oppositional stance.94 Under these circumstances, Perón opted for reinforcing as much as possible his support base among the popular classes. While associational power resources were mainly concentrated on certain segments of the working class, the right to vote provided institutional power resources to other segments of the popular classes, including unorganized and economically marginalized urban and rural sectors. In a similar vein as the Partido Laborista, Perón reacted to the new distribution of power resources and the existing alliances with the conception of a broad popular class-based coalition that was going to require further social policy concessions to both the working class and the marginalized urban and rural populations in order to be consolidated.

Hence, from 1945 on, the governing alliance experienced another broadening through the incorporation of economically marginalized groups, such as rural workers and internal migrants. In part due to this process, the participation of people of indigenous and mixed descent was considerably higher in the Peronist alliance than in the opposition (Camarero 2004, 33; Itzigsohn and Vom Hau 2006, 204–205). Anti-Peronist discourses therefore often took advantage of the racism inherent to widespread conceptions of the Argentine nation as "white and European" and othered the Peronists as "los negros" and "los cabecitas negras" (Itzigsohn and Vom Hau 2006, 204–205).<sup>95</sup> Through the introduction of female suffrage in 1947 and the formation of the Women's Peronist Party in 1949, the alliance furthermore strengthened its appeal to women (Barry 2008; Fisher 2000, 323–324; Guy 2009, 153–157). The following elections showed how successful the incorporation of these groups was. The share of Peronist votes was significantly higher among women than among men, and among economically marginalized rural and urban sectors than among the established working class (Barry 2011, 25; Levitsky 2003, 41). While neither of these newly incorporated groups achieved political influence comparable to that of the organized labor movement, their incorporation would nevertheless be reflected in social policies that aimed at reinforcing their support. As Lewis (1993, 189) rightly formulated it, social policy "served as the cement of the Peronista, Populist alliance."

<sup>94</sup> In 1952, Perón established a pro-Peronist employer association called Confederación General Económica (CGE) which represented mostly small- and medium-sized industrial companies of the interior of the country that produced for the local market (O'Donnell 1988, 34–35). Even this pro-Peronist association was more distanced from the government than the trade union movement. While it supported the government's interventionist economic policies, it criticized what it saw as the excessive costs of the social policies (Birle 1995, 107–108).

<sup>95</sup> Literally meaning "blacks" and "black heads."


Table 4: Composition of the new governing alliance, 1946\*

\* Although the composition of the alliance underwent changes and gradually weakened between 1946 and 1955, the inclusionary-redistributive character persisted over the period. \*\* Later these parties merged to form the Peronist Party. \*\*\* Later left the alliance and joined the opposition.

Perón's agenda was inclusionary, in the sense that it promoted the political and social incorporation of long-excluded social groups. However, it was also profoundly authoritarian, in the sense that it aimed at achieving nearly complete control over the supporting forces and increasingly repressed the opposition. As will be seen, the way social policies were institutionalized reflected this dual strategy of mobilization and control (Andrenacci, Falappa, and Lvovich 2004; Belmartino 2005a, 109–129).

In terms of social policy expansion, Perón continued to deliver quickly after the elections. The often immediate character of Perón's responses to social demands significantly contributed to his popularity among the popular classes as it provided credibility to one of his frequently repeated slogans: "Doing is better than talking, implementing is better than promising" (Del Campo 2012, 200). In May 1946, the Decree-Law 13.937 created the pension fund for industrial workers and hence implemented one of the measures the Partido Laborista had proposed in order to universalize pension insurance coverage. This measure, together with the incorporation of several occupational groups into other existing pension funds, gave another decisive push to the extension of pension coverage to hitherto excluded groups. The number of covered workers increased from 1,200,000 in 1945 to 2,250,000 in 1946, which represented approximately 45% of the workforce (Gaggero and Garro 2004, 180–181; Golbert 2010, 84).

Only nine days after the creation of the pension fund for industrial workers, Perón ordered the dissolution of the relatively autonomous parties in his coalition and created a single party that he planned to keep kainly under his own control. Partido Laborista leaders that openly opposed this order were harshly repressed, some even subjected to violent assaults and imprisonment (Del Campo 2012, 350–357). In a similar way, Perón continued to increase his control over the trade unions. Through the provision of benefits for cooperative unions and the application of repression to oppositional unions, Perón was able to achieve far-ranging authority over nearly all trade unions by the early 1950s (Collier and Collier [1991] 2009, 341; Mainwaring and Seibert 1982, 515; McGuire 1997, 56–59; Torre 1989, 547). At the same time, the government actively supported the organizing efforts of the unions. In combination with a favorable social and economic context,<sup>96</sup> this allowed for an unprecedented expansion of union density from approximately 18% in 1946 to 43% in 1954 (Doyon 1975, 154–160).<sup>97</sup>



<sup>96</sup> Full employment, the rapid expansion of the industrial sector and the dominance of the ISI paradigm provided an extraordinarily favorable context for efforts to increase trade union affiliation.

<sup>97</sup> In absolute numbers this represented an increase from 880,000 union members in 1946 to 2,260,000 union members in 1954 (Doyon 1975, 154–160).


Sources: Doyon (1975, 154–160).

The fact that Perón had gained extensive control over union leaders did not stop their rank and file members from demanding further benefits (Mainwaring and Seibert 1982, 519). Given that the popular classes increasingly constituted his only solid base of support, Perón had good reasons to frequently attend to their social demands with new reform measures (Torre 1989, 548; Mainwaring and Seibert 1982, 518). Indeed, due to a variety of factors, including the introduction of female suffrage, the takeover of hitherto Catholic social assistance by the Fundación Eva Perón (FEP) and the increasingly quasi-religious traits of the Peronist leader cult, the government was gradually losing the support of the Catholic Church (McGuire 1997, 73–74; Rock 1993, 176–177). The increasing tensions with the church also contributed to the growing opposition of military leaders, who rejected Perón's redistributive agenda as well as his alliance with the trade union movement. These tensions further increased when the economic crisis that began in 1949 led to limitations on the acquisition of new military equipment (Rock 1993, 154–182). Moreover, his attempts to construct a support base among employers were hardly successful. Despite the foundation of a government-friendly umbrella organization, the Confederación General Económica (CGE), the vast majority of employers stuck to a liberal economic world view and remained in political opposition to the government (Collier and Collier [1991] 2009, 347-349; Mainwaring and Seibert 1982, 518). In the area of pension policies, the government's dependence on popular class support contributed decisively to an ongoing expansion of coverage. During the early 1950s, the CGT reinforced its demand for incorporation of rural workers into the pension system, to which the government responded with the creation of the rural workers' pension fund in 1954 (Gaggero and Garro 2004, 183; Law 14.399). That same year, Law 14.397 created obligatory pension coverage for self-employed workers and employers. The following year, a member of the parliament representing the Women's Peronist Party, Delia Degliuomi de Parodi, successfully passed a bill that aimed at providing work regulations and pension access to domestic workers (Tizziani 2013). The *coup d'etat* of 1955 impeded the implementation of the law, but soon after Decree 326/56 implemented a similar although less generous proposal. By 1956, the pension system had been extended to the whole economically active population (Golbert 2010, 84).

Figure 2: Incorporation of additional occupational groups and evolution of legal pension coverage, 1943–1955

This did not mean that in practice the whole economically active population was contributing to one of the pension funds. From early on, the CGT criticized the evasion of pension contributions by many employers (Gaggero and Garro 2004, 182). In 1950, the existing pension funds received correct contributions for 86% of the theoretically covered workers. In 1955, this number decreased to only 50%, which was mainly due to the fact that affiliation with the newly created pension funds was still low (Arza 2010, 260; Gaggero and Garro 2004, 180). Nevertheless, the expansion of pension coverage during the Peronist government was massive. By 1955 nearly 2,500,000 Argentinians were effectively contributing to the pension system, at least six times as many as in 1944 (Gaggero and Garro 2004, 180; Feldman, Golbert, and Isuani 1988, 30). In contrast to the slow extension of pension coverage before 1943, the expansion affected not only relatively small and well-organized occupational groups but benefited many low-income earners in initially weakly organized sectors. What furthermore reinforced the redistributive character of this expansion was the permissive access criteria and the progressive benefit calculation formulas. In 1954, Law 14.370 determined that only five years of contributions were necessary to achieve pension rights. Under these circumstances, even workers who had been excluded from formal work during most of their lives could relatively easily secure access to a pension.

Sources: Feldman et al. (1988, 30) and Arza (2010, 259).

Sources: Law 14.370.

The same law created uniform, progressively scaled replacement rates for all public pension funds. This meant that low-income earners could get up to 100% wage replacement. For higher incomes, the rate gradually decreased to reach only 15% for every peso earned above a certain threshold (Law 14.370).

The Peronist government also expanded the provision of health care to the popular classes. In May 1946, the Public Health Department was created (Golbert 2010, 87). Inspired by the international discourses on public health of the time, the department elaborated a universal program of public health care expansion which was also anchored in the first five-year plan of the government (Belmartino 2005d, 127–128; Golbert 2010, 88). The program was mainly aimed at guaranteeing free access to health care to an estimated 65% of the population, those who could not afford medical attention (Andrenacci, Falappa, and Lvovich 2004, 105). The ambitious project contemplated an increase of the number of hospital beds from 60,000 to over 140,000, of the number of nurses from 8,000 to 46,000, a geographically more equitable distribution of physicians, and the construction of 800 health care centers distributed all over the whole country (Andrenacci, Falappa, and Lvovich 2004, 105–106; Ross 1993, 120). However, right from the beginning, the implementation of this program met with considerable difficulties.

One of the main difficulties was the existence of contradicting political strategies for health care expansion within the Peronist alliance. In contradiction to the proposal of a centralized system, Perón had de facto already institutionalized the first trade union-controlled health insurance fund for railway workers in 1944. This contributed decisively to the fact that the powerful railway workers' unions were the first to openly declare their loyalty to Perón (Del Campo 2012, 201–202). As this strategy proved effective in mobilizing trade union support, the government continued with this strategy after 1946. Until 1950, similar health insurance funds were established for workers of the glass industry, bank and insurance employees and the employees of the Congress and of the Ministry of War (Belmartino 2005a, 115–116; Danani 2005, 170). Nevertheless, the relative importance of these health insurance funds should not be overstated. In contrast to the pension insurance funds, coverage remained relatively low and the expansion of the universal public health care system was clearly receiving far more resources during the first Peronist government (Andrenacci, Falappa, and Lvovich 2004, 105–106; Danani 2005, 130).

Compared to the establishment of a limited number of trade union-controlled health insurance funds, the expansion of the public health system was capable of mobilizing broader popular class support. Both the organized working class and the electorally important but largely unorganized rural and urban low-income sectors benefited from this expansion. Notwithstanding, conflicting interests within the Peronist alliance posed difficulties for the implementation of a unified public system. The first problem regarded the financing of the public health system. The Public Health Department originally aspired to receive a proportion of the funds of the National Social Security Institute. The institute, however, was de facto undermined by Perón's strategic decision to expand the pension system by introducing additional pension funds rather than establishing a centralized system (Golbert 2010, 83). This meant that the Public Health Department lacked its own resources. Instead, it had to rely on the provincial governments and other ministries for the expansion of public health care facilities, which decisively contributed to the institutional fragmentation of the sector (Andrenacci, Falappa, and Lvovich 2004, 100–106). This fragmentation was further exacerbated when Perón supported a more politicized expansion of public health care through the FEP, which constituted one of the mechanisms through which the Peronist alliance mobilized the political support of economically marginalized sectors (Belmartino 2005a, 116–117; Stawski 2004, 222). This fragmentation and the unwillingness of physicians to work in marginalized rural areas contributed to the persistence of important degrees of inefficiency and inequality in the public health system (Ross 1993, 118–122). In sum, however, the public health care system was massively expanded and the access of low-income earners significantly improved. Between 1946 and 1953 the number of hospital beds increased from 60,000 to 130,000 and the number of nurses from 8,000 to 18,000 (Andrenacci, Falappa, and Lvovich 2004, 105; Ross 1993, 119–120). The Public Health Department, which was upgraded to a ministry in 1949, increased its overall number of employees from 7,000 in 1946 to 35,000 in 1955 (Katz and Muñoz 1988, 104). Several successful campaigns against infectious diseases were undertaken in rural areas and 200 health centers built in small communities (McGuire 2010, 131). This contributed to a significant improvement in health indicators. Infant mortality fell from 80 per thousand in 1943 to 60 in 1954 and life expectancy increased from 62 years in 1947 to 67 in 1953 (Escudero 1981, 562; Torre and Pastoriza 2002, 293).

The incorporation of marginalized rural and urban sectors into the associational and electoral support base of the Peronist alliance was also reinforced by an expansion of social assistance policies (Golbert 2008, 27–28). In 1946, the government took control of the semi-private charity societies that had carried out most of the publicly financed social assistance until that moment. During the following two years, most of the assets and activities of these charity societies were integrated into the Public Health Department, the National Department of Social Assistance and the FEP, whereby the latter clearly became the most important institution (Andrenacci, Falappa, and Lvovich 2004, 107; Golbert 2008, 27–28). Besides the aforementioned role in the provision of health care, the FEP distributed cash benefits, scholarships, and goods to poor households. Moreover, it provided an extensive infrastructure for social assistance including emergency accommodations, retirement homes, social housing projects, holiday camps and supply stores with subsidized prices (Andrenacci, Falappa, and Lvovich 2004, 109; Golbert 2010, 95–96). In general, the activities of the FEP aimed at both mitigating poverty and generating support among the marginalized sectors of society. Frequently these activities included clientelistic practices and were accompanied by intensive pro-Peronist publicity (Andrenacci, Falappa, and Lvovich 2004, 108–109; Stawski 2004, 222). It proved useful to the Peronist government to channel the majority of the funds to the FEP instead of state agencies, because the former largely escaped parliamentary control due to its formally private status as a foundation; this facilitated the use of the funds for shoring up political support. In order to assure the political impact of the FEP's actions, only committed Peronists were hired as staff (Stawski 2004, 210–214). The mechanisms through which the foundation mobilized popular class support are well exemplified by the distribution of non-contributory pensions to needy elderly persons. Applications for these pensions had to be made to the FEP, where the politically committed staff had large discretion in deciding whose applications to accept, providing ample opportunities for clientelistic practices (Stawski 2004, 200). Once the first thousand beneficiaries had been selected, a glamorous public event was organized at the Teatro Colón in which Eva Perón personally handed out the benefits (Gaggero and Garro 2004, 187). Although the non-contributory pension scheme was restricted to only certain geographical areas and remained quite limited, in sum the resources channeled to social assistance policies through the FEP were considerable. In 1951, the FEP constituted nearly 7%<sup>98</sup> of the total public expenditure (Belmartino 2005a, 116–117). Due to the fact that the FEP was financed to a large degree by obligatory wage deductions paid by formal workers and semi-voluntary donations by the trade unions, the foundation constituted also a significant mechanism of income redistribution from the established working class to the economically marginalized segments of the popular classes (Torre and Pastoriza 2002, 295).

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

From the mid-1930s on, the Argentine labor movement saw its power resources significantly enhanced. Expanding union affiliation increased its associational power resources. At the same time, the process of industrialization and declining levels of unemployment provided it with additional structural power resources. Furthermore, the increasing dominance of the ISI paradigm in public opinion provided the labor movement with ample discursive power resources that enabled it to frame its demands for redistribution as viable and desirable for the society as a whole. Against this background, a re-composition of political alliances was initiated with the *coup d'état* in 1943, which eventually gave way to the emergence and consolidation of a new, more inclusive governing alliance. As secretary of labor of the military government, Perón originally pursued the construction of a broad multi-class alliance including both capital and labor. Considering the relative strength of the labor movement, its incorporation required significant symbolic and material concessions, among which the expansion of social policy played a prominent role. While big parts of the labor movement proved to be cooperative, the employer associations refused the government's social and economic policies and so did the currents of the military elite that adhered to a rather liberal economic paradigm. By 1945, pressures from employers and the political opposition had increased in such a way that Perón was removed from office and arrested. However, due to massive strikes and demonstrations in favor of Perón, the military government decided to release him and to announce free elections for 1946. Quickly the majority of trade union movement joined forces to establish the Partido Laborista, which would constitute the backbone of Perón's electoral success in the upcoming elections. The return to democracy furthermore brought about a reshuffling of power resources, which would be reflected in another broadening of the governing alliance and an extension of social policy benefits to economically marginalized urban and rural sectors. These largely unorganized sectors possessed few associational and structural power resources, but the

<sup>98</sup> This equaled approximately 1% of the GDP (author's calculations based on data from Belmartino 2005a, 116–117, and Martínez 1998, 17).

return to democracy provided them with the right to vote, and hence with a crucial institutional power resource. Moreover, the predominance of the ISI paradigm in public opinion made it easier to frame redistributive policies in favor of the marginalized sectors as being beneficial to society as a whole. Under these circumstances, the Partido Laborista decided to adopt a program of social policy universalization that reflected its ambitions of creating a broad, popular class-based electoral alliance ranging from the marginalized rural and urban sectors through the organized working class to segments of the middle classes. The adoption of an egalitarian social policy program was moreover facilitated by the left ideological positions of most of its leaders.

Once in power, Perón delivered quickly and extended social protections for both the organized labor movement and the largely unorganized marginalized urban and rural sectors. The gradual loss of support from the church and the military further exacerbated Perón's political dependence on popular class support, which in turn contributed to the continuity of the process of social policy expansion until 1955.

By 1956, all occupational groups were covered by the pension system and the criteria for access and the calculation of benefits were highly progressive, clearly favoring low-income earners. Even though the generalization of a contribution-based social insurance system faced in practice difficulties in incorporating informal workers as contributors, the permissive access criteria which required only five years of contributions de facto meant that the vast majority of informal workers were going to be able to receive pension benefits at retirement age. Access to health care was extended through the creation of a free and universal public health system, the strengthening of trade union-controlled health insurances and the expansion of a network of hospitals controlled by the FEP. The latter also significantly expanded the coverage of social assistance policies, including the distribution of basic supplies, cash benefits and social housing. Although certain privileges for powerful occupational groups were preserved in the pension and health care systems, most resources were concentrated on the expansion of coverage to hitherto excluded groups. All in all, social protection for low-income earners was massively expanded between 1943 and 1955.

In many senses, this period also shaped the foundations on which later social politics would take place. Important path dependency effects emanated from the particular social policy legacy left by this period. The fragmented expansion of the pension system and the institutionalization of some trade union-controlled social health insurances created significant vested interests in the preservation of particular privileges. At the same time, the politicized and hardly institutionalized character of social assistance made it easy for subsequent governing alliances to dismantle these structures. Further crucial path dependency effects resulted from Perón's authoritarian restructuring of the labor movement. On the one hand, the Peronist government fostered trade union organization and therefore contributed to the achievement of unprecedented levels of associational power resources and considerable degrees of unionization also in small companies and among low-income earners.

On the other hand, Perón established nearly complete control over the movement. This process led to the marginalization of left currents in the movement, generated a significant autonomy of the leaders from the rank and file, and instituted a far-reaching dependence on the state. Taken together, these factors contributed to the development of a broader but less progressive and less democratic trade union movement in Argentina. For a majority of the union leaders, the defense and extension of particularistic benefits would come to constitute a crucial axis of involvement in social politics.


Table 6: Change in social protection for low-income earners, 1943–1955

On the other hand, Perón established nearly complete control over the movement. This process led to the marginalization of left currents in the movement, generated a significant autonomy of the leaders from the rank and file, and instituted a far-reaching dependence on the state. Taken together, these factors contributed to the development of a broader but less progressive and less democratic trade union movement in Argentina. For a majority of the union leaders, the defense and extension of particularistic benefits would come to constitute a crucial axis of involvement in social politics.

## **4.3 Military Intervention, Exclusionary Governing Alliances, and Regressive Social Policy Reform, 1955–1973**

The 1955 military coup led to an abrupt expulsion of the popular classes from the governing alliance. During the following 18 years, recurrent military interventions, the proscription of Peronism and a reduction of the power resources of progressive popular class actors effectively blocked the formation and rise to power of inclusionary alliances. Under these circumstances, overall social policy change was clearly regressive. As during the decades preceding 1943, improvements in social protection were mainly conceived as concessions to those parts of the popular classes that had significant structural and associational power resources at their disposal. In contrast to the pre-1943 period, the union movement had become significantly broader and stronger, so that social policy concessions between 1955 and 1973 were more generous and reached significantly more people. Low-income sectors, however, lacked associational and structural power resources and were largely deprived of their electoral power resources in the face of authoritarian and semi-authoritarian regimes. Against this background, most low-income sectors experienced a significant restriction of their access to social protection.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

An economically benign international context, the continuity of the process of industrialization and a massive expansion of public employment quickly created conditions of full employment during the first years of the Peronist governing alliance (Birle 1995, 102; McGuire 1997, 58; Torrado 1992, 175). Even when Argentina slid into economic crises during 1949 and 1952, employment levels remained largely stable (Andrenacci, Falappa, and Lvovich 2004, 93– 94; Torrado 1992, 175). This situation would change with the *coup d'état* of 1955. The implementation of structural adjustment programs and the economic crises of 1959 and 1962–63 triggered a rise of open unemployment to nearly 9% in 1963, which was followed by a gradual decline to around 5% during the late 1960s (McGuire 1997, 95; Torrado 1992, 176). However, by and large, the ISI model remained intact. Industrial production as a share of GDP rose from 35% in 1955 to 41% in 1973 and industrial employment moderately increased from 32% to 33% of the labor force. At the same time, the declining importance of the agricultural sector continued, so that between 1955 and 1972 its share of GDP fell from 13% to 9% and its share of the labor force from 26% to 17% (Timmer, de Vries, and de Vries 2014). Under these circumstances, the structural power resources of the labor movement remained relatively high, although the rise in unemployment implied a moderate decline compared to the previous period.


Table 7: Economic and social context, 1955–1973

Sources: Rapoport (2013) for GDP growth and inflation; Timmer, de Vries, and de Vries (2014) for industriy share of GDP and share of industrial employment; Torrado (1992, 175) for unemployment.

The establishment of authoritarian and semi-authoritarian regimes deprived the popular classes of their electoral power resources for advancing redistributive policies. Already Perón had begun to undermine democracy through a variety of measures such as repression against non-Peronist politicians, constraints on opposition parties' access to the media and the partisan manipulation of district boundaries (Collier and Collier [1991] 2009, 342–343; McGuire 1997, 69; Rock 1987, 305). Nevertheless, the ways in which Perón gradually undermined democracy did not curb the necessity of the Peronist alliance to appeal to the popular classes, as they constituted the main electoral support base and source of legitimation. Growing opposition within the army and the church even exacerbated the dependence of the governing alliance on the popular class vote (Birle 1995, 98; Collier and Collier [1991] 2009, 334–342). Under these circumstances, the electoral process continued to provide the popular classes with power resources that could be mobilized through and for the expansion of social protections. The 1955 coup finally eliminated this mechanism. From 1955 to 1958 and from 1966 to 1973 Argentina was governed by authoritarian military governments. Between 1958 and 1966 several elections were held, but the electoral force that captured the vast majority of the popular class support, Peronism, was banned. The military furthermore vetoed any attempt of these civilian governments to adopt a pro-Peronist or redistributive agenda (Collier and Collier [1991] 2009, 486–488; Golbert 2008, 43; McGuire 1997, 82). By this means, the mobilization of the popular classes' electoral power resources through and for redistributive policies was largely blocked between 1955 and 1973. This proved particularly disadvantageous for most of the low-income sectors, as these sectors lacked structural and associational power resources (Golbert 2010, 100).

Import Substitution Industrialization remained the dominant economic paradigm during the period (Sikkink 1991, 3). This provided progressive actors with discursive power resources, which enabled them to frame their demands for redistribution as desirable for the society as a whole and supportive of the overriding goal of industrial development. Yet, in the absence of free elections, one of the most important mechanisms for transmitting discursive power resources into policy change was blocked; namely the formation of electorally successful redistributive alliances.

In terms of associational power resources, the trade union movement had grown massively during the previous period. Although the more repressive stance of the post-1955 governments showed some effects, the organizational strength of the trade union movement remained on a high level compared to other Latin American countries. Between 1954 and 1964 union density declined moderately from an estimated 43% to an estimated 36% (Doyon 1975, 154–160; Torre 1973, 909).<sup>99</sup> The structural, associational and discursive power resources of the trade union movement enabled it to exercise significant pressure on both the employers and the government (Golbert 2010, 100). However, the distribution of associational power resources was highly unequal among different segments of the popular classes. While approximately 80% of all workers in the energy sector and 66% of transport and communication workers were organized in 1964, this was the case for only 5% of the rural workers (Torre 1973, 909). In general, the economically marginalized sectors of the popular classes tended to possess low associational power resources. The solidarity of better organized sectors could have partially remedied such

<sup>99</sup> There are differing estimates of union density for both years, however all authors agree that there was a moderate decline (see e.g., Doyon 1975, 154–160; Lamadrid and Orsatti 1991, 153–156; Marshall and Perelman 2004, 36; Torre 1973, 909).

associational weakness, yet for reasons described above the emphasis of trade union participation in social politics was increasingly on the defense and improvement of existing social rights and less so on the extension of social protections to still insufficiently protected groups.


Table 8: Popular class power at the beginning of the period, 1955–1958

The employer associations continued to oppose redistributive social policies and the permanent threat of military intervention deterred most of the civilian parties from an open pursuit of redistributive agendas (Birle 1995, 110–135; Collier and Collier [1991] 2009, 496–497; Sikkink 1991, 8–9).

In sum, the context between 1955 and 1973 was unfavorable for the emergence of powerful alliances in favor of an expansion of social protections for low-income earners. Only during the last years of the period did left-leaning forces gain strength within and without the Peronist movement and the trade unions (Collier and Collier [1991] 2009, 496; McGuire 1997, 150).

## **The Expulsion of the Popular Classes from the Governing Alliance and the Regressive Turn in Social Policy Reform, 1955–1958**

By 1954 the Peronist alliance showed clear signs of erosion. The introduction of female suffrage, the takeover of Catholic charity functions by the FEP and a reported affair of Perón with a fourteen-year-old girl led the church to assume an oppositional role (McGuire 1997, 73). The increasingly authoritarian character of the government further intensified the antagonism of the opposition political parties (Birle 1995, 98). In the military the oppositional faction became increasingly dominant due to a variety of concerns with Perón's redistributive social and labor policies and his relationship to Eva Perón (Collier and Collier [1991] 2009, 334). By the mid-1950s the popular classes were the only strong support base of the Peronist alliance (Mainwaring and Seibert 1982, 518).

Against this background, the military coup of 1955 was welcomed by a coalition that included ample parts of the middle classes, the church, and nearly all employer associations and opposition parties (Birle 1995, 111; Cavarozzi 2006, 16). Consequently, the re-composition of the governing alliance following the coup saw the expulsion of the popular classes and the inclusion of a variety of forces that opposed redistributive social policies; most importantly the employer associations, conservative politicians and right-wing military sectors that adhered to a liberal economic world view (Birle 1995, 112–113; Cavarozzi 2006, 16–17). Nevertheless, conscious of the significant power resources of the largely Peronist labor movement, the military government under General Eduardo Lonardi took a reconciliatory stance during the first two months of its rule, and so most of the labor and social policies that benefited the organized working class were preserved (Collier and Collier [1991] 2009, 484; McGuire 1997, 81–82). At the same time, it was only a few days before the military government dissolved the FEP, which had been in charge of most of the social assistance policies that benefited mainly unorganized, economically marginalized sectors (Golbert 2008, 29–39). This pattern of regressive social policy reform would predominate politics over the following 18 years.


Table 9: Composition of the new governing alliance, 1956\*

\* Although the composition of the alliance underwent several changes between 1955 and 1973, the exclusionary-regressive character persisted. \*\* Between 1958 and 1966 the regime type was semi-authoritarian. The electoral stronghold of the UCR governments during these years consisted of the middle classes.

As the majority faction of the military and the economic elite favored harsher repression of the Peronist movement, Lonardi was replaced by General Pedro Aramburu only two months after the coup (Collier and Collier [1991] 2009, 484–485). Aramburu outlawed the Peronist Party, took over the CGT and removed and persecuted thousands of Peronist trade union leaders and activists (McGuire 1997, 81–82). The labor movement met these repressions with increasing resistance, which culminated in three years of massive strikes, factory occupations, sabotage, and even bombings (Collier and Collier [1991] 2009, 485). Under these circumstances, the Aramburu government adopted a double strategy towards the labor movement that combined repression with social and labor policy concessions for selected groups (Collier and Collier [1991] 2009, 485; Golbert 2010, 99–100). The most important social policy concession was the institutionalization of employer-financed family allowance funds for workers in the industrial and commercial sectors. By 1954, most trade unions had already achieved the establishment of modest family allowances through collective agreements (Gaggero and Garro 2004, 187). Nevertheless, in the 1956 collective bargaining round, the commercial sector workers' union achieved a more generous arrangement that foresaw the legal implementation of a family allowances fund that would be financed through a 4% payroll tax paid by the employers (Falappa 2014, 201–202). In the following year, Decree-Law 7913 implemented this fund with explicit reference to the 1956 collective agreement. In order to reduce the complexity of the wage negotiations in the industrial sector in a context of intense social conflict, Decree-Law 7914 established a similar fund for industrial workers, making reference to the explicit support of the UIA for the measure. Although the institutionalization of family allowance funds had progressive redistributive effects among formal workers of the industrial and commercial sector, low-income earners in other sectors were not affected.

Overall social policy change during the military government was highly detrimental to most low-income earners. Not only was social assistance drastically reduced; the universal public health sector also suffered from retrenchment. In order to reduce the fiscal deficit, the government transferred the responsibility for hospitals to the provinces without transferring adequate resources (Belmartino 2005a, 129–132). This initiated a long-term trend of deterioration of public health services. One indicator of the rapid decline during the military dictatorship was that between 1954 and 1958 the number of hospital beds was reduced by 13,000 (Katz and Muñoz 1988, 106).

#### **Civilian Governments and Military Veto, 1958–1966**

Aramburu's attempt to destroy the Peronist movement triggered massive conflicts and yielded little success. The unbroken appeal of Peronism to the popular classes was not only reflected in the inability to "de-Peronise" the trade unions but also in the fact that during the 1957 elections for a constituent assembly nearly one quarter of the voters cast blank ballots in support of the banned Peronist Party (Collier and Collier [1991] 2009, 737; Golbert 2008, 30–31). Against this background of limited political success and the erosion of the military government's political support base, general elections were scheduled for February 1958 (Rock 1987, 337). Nevertheless, due to the proscription of the Peronist Party, the force that attracted the vast majority of the popular class vote remained excluded from political competition until 1973 (McGuire 1997, 82).

In 1957, the UCR split into two because of internal differences regarding the strategy towards the Peronist movement and its policy legacy. The Unión Cívica Radical Intransigente (UCRI), which was led by Arturo Frondizi, wanted to preserve the inherited social and labor policies. The Unión Cívica Radical del Pueblo (UCRP), which was led by Ricardo Balbín, explicitly supported the military government's attempts to repeal many of these policies and to "de-Peronise" the society (McGuire 1997, 85). While the UCRP received support from the military government, the UCRI negotiated a deal with the exiled Perón who agreed to recommend that his followers vote for Frondizi in exchange for policy measures in favor of the labor movement and an eventual suspension of the ban against the Peronist Party. With the support of most Peronist votes, Frondizi triumphed with a landslide electoral victory, winning also nearly all provincial governments and a majority in the legislature (Collier and Collier [1991] 2009, 489–490; Golbert 2010, 100–101).

During the first three months of its administration, the Frondizi government maintained a good relationship with the Peronist movement. It decreed wage increases, allowed several Peronist politicians and union leaders to return to politics and restored the main provisions of the Law of Professional Associations that Perón had instituted in 1944 (McGuire 1997, 86). A second important concession was Law 14.499,<sup>100</sup> which introduced an inflation-based pension indexation mechanism. In the face of high inflation rates, this was perceived as an important improvement by the trade union movement. Because the popular classes were excluded from the governing alliance, there was nothing that would have guaranteed regular and adequate adjustment by means of presidential decrees. However, the indexation mechanism came also at a cost. Law 14.499 introduced slightly lower replacement rates and, more importantly, established that a part of the surplus of the pension system would be

<sup>100</sup> Most of the literature analyzing Law 14.499 emphasizes the relatively high replacement rate of 82% that was established by it (Feldman, Golbert, and Isuani 1988, 45; Golbert 2010, 103; Isuani 2008, 175). However, the replacement rate of 82% only affected the lowest income bracket defined in the law. The preceding Law 14.370 from 1954 had actually fixed a replacement rate of 100% for the lowest income bracket. The main innovation of Law 14.499 was therefore not the replacement rate but the automatic indexation mechanism that it implemented.

transferred to the government for the purpose of public investments. This transfer of funds in conjunction with easy access criteria, high levels of evasion and high replacement rates accelerated the maturation of the pension system so that by the early 1960s the first pension funds began to accumulate deficits (Arza 2010, 263).<sup>101</sup> In political terms, Law 14.499 fulfilled mainly three functions for the government. First, it contributed to avoiding conflicts with the powerful trade union movement. Second, it enabled the government to achieve control over significant financial resources. And third, the law responded to the constitutional amendment of 1957. Although the constituent assembly of that year had the main function of legitimizing the replacement of the Peronist constitution with the constitution of 1853, a minority faction of the assembly with good relations to the labor movement had managed to incorporate the article 14 bis, which established among others the right to adequate pension indexation ("jubilaciones y pensiones *móviles*").

While these policies were well received by the labor movement, they caused significant discontent within the military and fears of a renewed *coup d'état* arose within the government (Cavarozzi 2006, 173). When soon after the threat of military intervention was reinforced by the threat of an approaching foreign-exchange crisis, Frondizi radically changed the agenda and adopted austerity measures including privatizations and a devaluation of the currency. The consequences for the popular classes were severe: economic recession, high rates of inflation and a significant reduction of real wages (McGuire 1997, 86; Rock 1987, 338–339). Austerity policies also included retrenchment in the public health care and social assistance sectors (Golbert 2008, 32; Katz and Muñoz 1988, 108; McGuire 2010, 132). Some provinces even initiated experiments with the introduction of market mechanisms and fees for treatment in the formerly free and universal public health care system (Belmartino 2005a, 132–134). While this deterioration of the public health system reduced the access to and quality of health care for low-income earners, better-earning workers increasingly affiliated with private and trade unioncontrolled health insurances, which decisively contributed to rising inequalities in health care coverage (Danani 2005, 181–182).

The austerity program was supported by the main employer associations, but the Peronist labor movement fiercely opposed it and responded with strikes and factory occupations (Birle 1995, 116–117). These protests, in turn, were violently repressed by the military and finally ended the alliance between Frondizi's government and the Peronist labor movement (Collier and Collier [1991] 2009, 491; Golbert 2010, 102–103). At the same time, the military significantly intensified its political interference and virtually eliminated any space for pursuing pro-Peronist or redistributive policies (Birle 1995, 116; Golbert 2010, 103; Rock 1987, 340). When in the 1962 elections neo-Peronist parties won a

<sup>101</sup> In 1963 the first fund ran into financial crisis and stopped payments to approximately 200,000 pensioners for several months (Danani 2005, 155).

combined national vote of 32% and won in most provinces, the military finally undertook another *coup d'état* and annulled the electoral results (Birle 1995, 117; Collier and Collier [1991] 2009, 728). Frondizi was replaced by José María Guido, the president of the Senate, who governed for 18 months in tight collaboration with the military and whose cabinet was composed of conservatives, employers and big landowners (Birle 1995, 121; McGuire 1997, 92).<sup>102</sup>

In 1963 new elections were held, but this time both the neo-Peronist candidate Vicente Solano Lima and Arturo Frondizi were excluded from the competition. In this context, Arturo Illia, the candidate of the UCRP, assumed the presidency with only 25% of the vote (Rock 1987, 344). Compared to his predecessors, Illia pursued more interventionist economic policies that provided the ground for real wage increases and rising consumption (Birle 1995, 123). Yet, considering his weak electoral and legislative support, his margins of action were even tighter than those of Frondizi. The threat of military intervention largely obstructed the possibility of mobilizing low-income earner support with redistributive or pro-Peronist policies. Furthermore, both the Peronist labor movement and the employer associations actively opposed the government (Golbert 2010, 108; Rock 1987, 345–346).

Few progressive social policy innovations could be expected in this situation. The most significant reform was an increase in the value and the coverage of the family allowances. In 1964, these were extended to dock workers, a tiny but structurally powerful occupational group (Falappa 2014, 205). Furthermore, covered workers were entitled to receive allowances for their husbands and wives as well as for their children in secondary school (Golbert 2010, 110). The value of the allowance was fixed at 10% of the minimum wage (Falappa 2014, 213). In the area of public health care and social assistance, the government made some experiments with community-based approaches and opened approximately 250 maternal and infant health care centers (Golbert 2010, 108– 116; McGuire 2010, 133). Nonetheless, the resources that the national government devoted to social assistance programs were extremely low and never surpassed 0.2% of the GDP (Golbert 2008, 33).<sup>103</sup>

In sum, the social policies of the UCR governments between 1958 and 1966 provided some improvements for formal workers, yet informal employment increased<sup>104</sup> and social protections for low-income earners and informal workers were cut significantly.

<sup>102</sup> The short period of the Guido government was furthermore shaped by a profound economic crisis (Rock 1987, 342–343).

<sup>103</sup> Considering that during the early 1950s the FEP alone disposed of over approximately 1% of the GDP, these numbers indicate a significant reduction of social assistance.

<sup>104</sup> Between 1955 and 1960 informality increased significantly in several sectors. Among workers of the commercial sector it increased from 26% to 40% and among industrial workers from 29% to 50%. Even the already high levels of informality among selfemployed workers increased from 85% to 95% (Arza 2010, 260).

## **Military Rule and the Consolidation of Exclusionary Social Insurances at the Expense of Universality and Redistribution, 1966–1973**

The major employer associations took an increasingly hostile stance towards the Illia government because they felt that its rather interventionist economic policy was not adequately considering business interests and that its approach to labor protests was too permissive. This hostility became so intense by 1966 that the major employer associations actively campaigned for a military coup (Birle 1995, 124–125). A steep fall of economic growth from double-digit rates to virtual stagnation additionally complicated the situation. In response to these employer pressures and a looming economic crisis, the Illia government implemented austerity measures. Yet, the labor movement responded with such a massive a wave of strikes that the government eventually abandoned the wage freeze, which constituted one of the main components of the adjustment program (Rock 1987, 346). Against this background, General Onganía, who had been promoting pro-coup sentiments within the military since April 1965, took over the government by means of a *coup d'état* on the 28th of June 1966 (McGuire 1997, 145; Rock 1987, 346).

Initially, the coup was welcomed by a wide array of actors, including a majority of the more conservative union leaders, the major employer associations, and most opposition parties (Birle 1995, 126; Golbert 2010, 116; McGuire 1997, 145–150). In political terms, the Onganía government quickly advanced its authoritarian agenda. The parliament and all political parties were dissolved, judges of the supreme court were removed, and the universities were put under military surveillance Employers were furthermore granted important influence on public policy (Birle 1995, 127–129). In economic terms, the Onganía government implemented an austerity program that combined a currency devaluation with a wage freeze, price controls and increased export taxes for agricultural products. Not surprisingly, the relations between the government and the majorX employer associations were good, although the most powerful agricultural employers' association, the Sociedad Rural Argentina (SRA), criticized the government for allegedly promoting industry more vigorously than agricultural production (Birle 1995, 128–129). However, despite the initial support of the coup by many trade union leaders, they remained excluded from the governing alliance and the above-mentioned measures led to a decline of real wages, particularly for white collar workers (Birle 1995, 130; O'Donnell 1988, 265–268). Therefore, soon after the coup, the CGT launched a series of protests against declining wages and public sector worker layoffs. In February 1967, the trade unions attempted to extend their protests with a wave of strikes and factory occupations, yet the military government successfully repressed the strikes and withdrew the official recognition of several important unions (Golbert 2010, 116–117).

Until 1969, the government was highly successful in the pursuit of its agenda. The economy grew at high rates, inflation declined, the balance of payments improved, and investment increased. The dissolution of the political parties had effectively paralyzed most of their activities and the repressions against the trade union movement had led to a complete failure of the CGT's strike wave in 1967 (Cavarozzi 2006, 39). In this context of apparently stable political domination, the exclusionary governing alliance initiated a far-ranging and profoundly regressive reform of the pension system.

In August 1967, shortly after the successful suppression of the trade union movement's wave of strikes, Law 17.385 increased pension contribution requirements from 5 to 10 years. Decree 4576/67 simultaneously adjusted the rules for the calculation of the replacement rate, slightly reducing their progressive character. One and a half years later, the Laws 18.037 and 18.038 established that the 10-year contribution requirement would rise by one each year up to a maximum of 30 years. These increases in the contribution requirements severely restricted the pension access of many low-income earners as their unstable employment histories and precarious employment relations obstructed the accumulation of a high number of years of contributions. By this means, the above-mentioned high levels of informality in certain sectors increasingly translated into exclusion from access to pensions. The Laws 18.037 and 18.038 furthermore introduced new replacement rates that implied a significant redistribution of pension resources in favor of high-income earners by abolishing the progressive scaling of benefits. While the Law 14.499 of 1958 had established replacement rates that ranged from 82% for the lowest income bracket to 10% for the highest, the Laws 18.037 and 18.038 established a uniform 70% replacement rate. Workers who had contributed more than 30 years could receive one additional percentage point for each additional year of contribution and reach a maximum replacement rate of 82%. In practice, this meant that while before 1968 low-income earners received significantly higher replacement rates, after 1968 higher income groups tended to receive higher replacement rates because they were on average more likely to have accumulated over 30 years of contribution. Camila Arza (2010, 265–266) estimated that the new pension calculation formula was one of the main reasons behind the drastic rise of inequality within the group of pensioners from a Gini-coefficient of 0.29 in 1970 to 0.39 in 1980.

Figure 4: Increasingly less progressive replacement rates in 1954, 1958, 1967 and 1968

 \* For the 1968 replacement rate 30 years of contribution are assumed. Sources: Law 14.370 for 1954, Law 14.499 for 1958, Decree 4576/67 for 1967 and Laws 18.037 and 18.038 for 1968.

Besides these highly regressive alterations of the access criteria and benefit calculation formula, the 1968 reform laws also introduced profound changes in the organizational structure of the pension system. The thirteen existing pension funds were replaced by only three funds, one for public sector workers, one for private sector workers, and one for self-employed workers (Golbert 2010, 117). Law 18.037 introduced uniform contribution rates and retirement ages for salaried workers and Law 18.038 did so for self-employed workers. At the same time, the government preserved pension privileges for the military and security forces (Lo Vuolo and Barbeito 1998b, 125). Although these centralizing measures reduced administrative costs and enabled the compensation of deficits of some pension funds with the surpluses of others, the financial stabilization of the pension system did not last for long and by 1971 the system had again a negative financial balance (Arza 2010, 267). The reform furthermore eliminated the participation of trade unions in the administration of the funds (Arza 2010, 265–266; Isuani 2008, 176). This was criticized by the trade union movement, however, against the background of the failed strike effort of 1967, it did not initiate broader protest activities against the reform law (Danani 2005, 155–156).

In this context of apparent political strength, the military government also initiated regressive reforms in the health care sector that proved to be favorable to the employer sectors who supported the governing alliance. One of the first measures was the signing of an international agreement that tied Argentina to the international network of patents and improved the business opportunities of multinational drug companies in the country (Escudero 1981, 563). At the same time, the government began to analyze the fragmented health insurance sector and implemented further retrenchment and decentralization in the universalistic public health care system (Bermann and Escudero 1978, 533; Danani 2005, 152; Katz and Muñoz 1988, 108). The first proposal for a reform of the health insurance sector was elaborated by the Public Health Department. The main goal of the proposal was to reduce the inefficiency of the fragmented and weakly regulated sector and to separate the health insurances from union control (Belmartino 2005a, 165–166). This project, however, faced opposition from two sides. For the trade unions, the control over important parts of the social health insurance sector provided significant financial resources and an important tool for the recruitment of members (Danani 2005, 164). The most important confederation of physicians, the COMRA, furthermore rejected the idea that the state would regulate the business relations between professionals and health insurers. Therefore, both COMRA and the CGT opposed the proposal of the Public Health Department (Belmartino 2005a, 170–171).

While the pension reform process did not involve concessions to the trade union movement, the health insurance reform would follow a different path in the face of the initiation of a new wave of trade union and left-wing militancy. During his government, Perón had marginalized the left within the trade union movement. Yet, shortly after the coup of 1955, he started to support the left wing of the unions from exile in order to build a loyal counterweight to the more conservative but also more autonomous trade union sector led by Augusto Vandor (Collier and Collier [1991] 2009, 494–495; James 1976, 273– 275; McGuire 1997, 87–92).<sup>105</sup> In the context of a worldwide strengthening of the radical left, an ongoing deterioration of real wages and the support of Perón, the combative left-wing currents were able to gain increasing strength within the union movement during the late 1960s (Collier and Collier [1991] 2009, 496; Falappa 2014, 210). In March 1968, a left-leaning trade union sector finally won the leadership elections of the CGT, which led to a split of the trade union movement into a left-leaning and combative CGT de los Argentinos<sup>106</sup> and a more conservative CGT Azopardo, which was more inclined to negotiate with the military government (James 1976, 275; McGuire 1997,

<sup>105</sup> Ideologically Perón stood much closer to the conservative trade union faction, so that his support for the left was of a strategic nature (Collier and Collier [1991] 2009, 494– 495; McGuire 1997, 88).

<sup>106</sup> Reflecting its left-leaning ideological position, the program of the CGT de los Argentinos emphasized the necessity of universal access to health care, education, and housing. Furthermore, the program expressed solidarity with the inhabitants of the Villas de Emergencia, the poor neighborhoods that had emerged in many cities (CGT de los Argentinos [1968] 2006).

156). Although the CGT de los Argentinos faced severe repression by the government and started to disintegrate by 1969, this could not stop the initiation of a new wave of left-leaning labor and student protests from mid-1968 on (Falappa 2014, 210; Godio 2000, 1035–1040; McGuire 1997, 156). In December 1968, the military government decreed a wage increase far below inflation. In order to partially compensate for the deterioration of wages, Law 18.017 enacted simultaneously a substantial rise of family allowances for formal industrial, commercial and dock workers.<sup>107</sup> The same law also introduced an additional increase for families with three or more children, which was justified on the grounds of nationalist ideas that this would increase "the number of births in order to enrich the human potential of the country" (Law 18.017). In May 1969, the protests nevertheless began to run out of the government's control. Three students were shot by policemen in Rosario and Corrientes (McGuire 1997, 157). On the 29th of May, students and auto workers joined forces in a demonstration in Córdoba in opposition to policies that promoted price hikes and declining wages. The next day, this protest was joined by thousands of further citizens that occupied most of Córdoba's downtown and defended it for 48 hours against the police, using barricades, rocks, and Molotov cocktails. When the army finally repressed the protest, fourteen protesters had been killed and the perception of stable governmental control shattered (McGuire 1997, 157; Rock 1987, 348–349). In the aftermath, violent forms of political protest and guerrilla activity grew markedly (O'Donnell 1988, 176– 177).

The Córdoba riots and the unprecedented radicalization of student and labor protests posed a real challenge to the military government and the business community. By 1969, unemployment had decreased to 4.3%, its lowest point in many years, so that striking workers were not easily replaced. In the automobile sector, strikes imposed extraordinarily high costs on manufacturers due to the relatively high capital intensity of the sector (O'Donnell 1988, 179; Torrado 1992, 176). In short, the increasing radicalization of protests and the associational strengthening of the left coincided with the possession of significant structural power resources by participating labor sectors. Against this background, a wave of capital flight began and currency reserves declined rapidly, threatening the advent of a new balance of payments crisis (Birle 1995, 131).

These developments caused a division within the military regarding the kind of response that was perceived to be adequate (O'Donnell 1988, 172– 173). While some sectors wanted determined repression, others preferred to make concessions. The latter faction would eventually prevail (Rock 1987, 349). On this basis, social policy reform took a different path, yet not neces-

<sup>107</sup> The rise of family allowances soon created financial imbalances, so that in October 1970 the contribution rate for employers was elevated to 12%. In 1972, the value of the allowances for disabled children was furthermore doubled (Falappa 2014, 213–214).

sarily one that was more favorable for low-income earners. The minister of social welfare was replaced by Carlos Consigli, who distanced himself from the initial health reform proposal of the Public Health Department and embraced an alternative proposal of the Community Assistance and Development Department. The latter was led by intimates of the president himself and suggested a health insurance reform that would be devised in cooperation with collaborative trade union leaders (Belmartino 2005a, 166–167; O'Donnell 1988, 73). Only a few weeks after the Córdoba riots, the government implemented concessions that extended the coverage of social health insurance funds of a few selected trade unions, by then commonly referred to as Obras Sociales (Danani 2005, 185).108 These concessions were well received, particularly by the most conservative trade union leaders. In contrast to the left-leaning CGT de los Argentinos, which emphasized the need for universal access to health care in its program, the collaborationist unions concentrated on particularistic demands, among which the strengthening of the trade union-controlled Obras Sociales occupied a high priority (CGT de los Argentinos [1968] 2006; Godio 2000, 1051). In November 1969, the military government eventually announced that it was studying a law for introducing obligatory social health insurance coverage for all wage workers financed by employer and worker contributions but managed largely by the trade unions. While this announcement was heavily criticized by all employer associations, it coincided with a marked improvement of the government's relations with the more conservative union leaders. In December 1969, the government met with a group of 25 powerful conservative union leaders in order to discuss the reunification of the CGT under conservative leadership. A short time later, Law 18.610 put the obligatory health insurance affiliation into effect, which additionally increased the readiness of the group of 25 union leaders to cooperate (O'Donnell 1988, 181–182). In May 1971, Law 19.032 complemented these reforms with the implementation of a separate social health insurance fund for pensioners.

Social health insurance reform, in several ways, played an important role in stabilizing the political situation after the 1969 riots. First, it constituted a significant concession to the labor movement as a whole, which was thought to reduce the militancy of the more combative wing and to increase the readiness to cooperate of the more conservative wing. At the same time, the strengthening of the union-controlled Obras Sociales also increased the power of the union bureaucracy over the affiliates, which proved beneficial in a situation in which conservative union leaders were looking for ways to moderate or even suppress the radicalization that was taking place among some of their bases (Danani 2005, 195). And finally, the Law 18.610 also involved the

<sup>108</sup> Besides social health insurance, the Obras Sociales also provided affordable holiday packages and other benefits to their members. Nevertheless, social health insurance clearly is the most important component of their services. Therefore, the terms social health insurances and Obras Sociales will be used interchangeably in this study.

creation of the National Institute of Obras Sociales (Instituto Nacional de Obras Sociales / INOS), which was officially responsible for overseeing the system of Obra Sociales and for mitigating financial inequalities by redistributing a certain share of the revenue between the different Obras Sociales. However, there were no clear rules for resource transfers, so that in practice, the distribution of funds ended up being a political mechanism through which the government could exercise a certain level of control over trade union leaders (Mera 2010, 105; Interview Neri 2014a).

Law 18.610 mandated a certain transfer of resources from employers to formal workers, yet its overall effects were far from being progressive. First and foremost, it led to the consolidation of a highly fragmented, inefficient and inegalitarian system in detriment of the universal health care system (Belmartino 2005a, 155–175; Interview Neri 2014a). Before the reform, the universal and free public health care system provided coverage for most of the population on a relatively egalitarian basis. In 1967, only about 15%109 of the population were covered by Obras Sociales. In contrast, by 1971, 66%<sup>110</sup> of the population was covered by Obras Sociales and the public health care system was gradually downgraded to a safety net of last resort that lacked adequate infrastructure, technology, and financing (Katz and Muñoz 1988, 107; Interview Neri 2014a). While private physicians and health care companies benefited from the contracts they were able to sign with the strengthened Obras Sociales, the access to health care of low-income earners de facto deteriorated with the decline of the public health care system (Belmartino 2005a, 170; Lloyd-Sherlock 2005, 1895; Interview Neri 2014a). Furthermore, access to health care varied significantly between different occupational groups and their respective Obras Sociales, so that also among those affiliated with the system of social health insurance, considerable inequalities existed (Belmartino 2005a, 177–183).

The regressive nature of health care reform under the military government was also reflected in the poor evolution of health indicators. According to Bermann and Escudero (1978, 533), infant mortality increased by a stunning 19% between 1966 and 1970. Although the rise was exceptionally strong under the military dictatorship, the decline of the public health sector contributed to weak health indicators over the whole period between 1955 and 1973. While infant mortality had declined by 29% between 1943 and 1954, it increased by 4% between 1954 and 1971 (Escudero 1981, 562; McGuire 2010, 133; Torre and Pastoriza 2002, 293). Life expectancy rose by 8% between 1947 and 1953 but had decreased by 2% by 1970, despite important technological advances and economic growth (INDEC 2015; Torre and Pastoriza 2002, 293).

<sup>109</sup> Author's calculation on the basis of Danani (2005, 193) and CEPAL (2017b).

<sup>110</sup> Same as preceding footnote.

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

The *coup d'état* against the Peronist government in 1955 marked the rise to power of new, more exclusionary governing alliances that heavily depended on the support of anti-redistributive forces such as the majority faction of the military, the major employer associations and centrist and conservative political forces. During the following years, repression, economic adjustment and authoritarianism weakened the overall power resources of the popular classes so that the possibilities for the formation and rise to power of an inclusionary alliance were severely limited. However, the loss of power resources was unequally distributed among different segments of the popular classes. The fact that military interferences and the proscription of the Peronist Party largely blocked possibilities to mobilize electoral and discursive power resources for redistributive policies meant that economically marginalized and weakly organized segments of the popular classes were excluded from social politics. In contrast, the labor movement preserved levels of associational and structural power resources that enabled it to cause significant political and economic instability, so that the governments repeatedly perceived the need to make social policy concessions. These policies, however, only benefited formal workers or even only certain groups of formal workers. The majority of low-income earners, in contrast, faced a profound deterioration of their access to social protections due to the regressive reform of the pension system and the retrenchment of the universal public health system and social assistance policies.


Table 10: Change in social protection for low-income earners, 1955–1973

Social policy reform between 1955 and 1973 also created important path dependency effects that would go on to influence later political dynamics of social policy reform. The centralization of the pension system resulted in the elimination of most pension funds for particular occupational groups and with them the influence of trade unions in their administration. Therefore, the reforms significantly reduced the interest of the trade unions in defending the status quo. This, in turn, would facilitate later attempts to reform pensions. In the case of health insurance, the path dependency effects left by the reform, on the contrary, strongly increased the interests of the trade union movement in the preservation of the status quo. The 1970 reform gave trade union leaders control over significantly expanded Obras Sociales, which provided them with important financial resources and constituted effective tools for controlling and expanding their rank and file members. The consolidation and expansion of the system of Obras Sociales would go on to provide a strong motivation for trade unions to resist both progressive and regressive efforts to reform health care in the following decades.

## **4.4 Social Policy Expansion and Political Conflict during the Second Peronism, 1973–1976**

The discredit of the military government, the rise of left-wing youth movements and combative trade unions, the return to democracy and the lifting of the ban on Peronism in 1973 provided favorable conditions for a realignment of political forces and the formation of a more inclusionary governing alliance that comprised the Peronist Party, organized labor, unorganized low-income sectors, left-leaning youth organizations and the CGE. Quickly this new alliance developed an ambitious agenda for progressive social policy reform. Due to severe conflicts within the Peronist movement, economic difficulties, and the growing influence of right-wing Peronists, important parts of this agenda were never implemented. Nevertheless, during the three years of democracy between 1973 and 1976, the public health system was financially strengthened, social assistance policies expanded, and access to pensions strongly facilitated. In sum, social protection for low-income earners was significantly expanded.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

The Peronist alliance assumed government in an extraordinarily benign economic context. Economic growth during the preceding years had reduced unemployment to moderate levels between 5% and 6% (Torrado 1992, 176). Furthermore, a world commodity boom boosted export earnings by approximately 65% and led to a swift expansion of foreign currency reserves (Rock 1987, 361). Coupled with policies that promoted the expansion of the local market, this unexpected increase in export value initially fostered rapid economic growth. Although the export boom ended by early 1974 and economic difficulties returned (Birle 1995, 137), unemployment continued to fall, reaching a long-time low of under 4% in 1975 (Torrado 1992, 176). At the end of the Peronist government, in 1976, the industrial sector employed 32% of the workforce and produced 40% of the GDP (Timmer, de Vries, and de Vries 2014). Taken together, the process of industrialization, the related increase in capital intensity, and the virtual elimination of unemployment provided the popular classes with considerable structural power resources.


Table 11: Economic and social context, 1973–1976

Sources: Rapoport (2013) for GDP growth and inflation; Timmer, de Vries, and de Vries (2014) for industry share of GDP and share of industrial employment; Torrado (1992, 175) for unemployment in 1973; INDEC (2003) for unemployment 1974–1976; Beccaria and Groisman (2009, 110) and Jiménez (2013, 59) for informality.

In addition to these structural power resources, the recovery of democracy and the lifting of the ban on Peronism returned electoral power resources to the popular classes. This increased particularly the political importance of rural and urban low-income sectors that mostly lacked structural and associational power resources but constituted a crucial electoral support pillar of the Peronist movement (Levitsky 2003, 41).

At the same time, the national and international discursive context was extraordinarily favorable for projects of welfare state expansion. Across Europe, and indeed many countries around the globe, Keynesian thinking reached its heyday during the 1960s and 1970s and the growth of welfare states accelerated considerably (Hay and Wincott 2012, 22–23). These processes were in turn reinforced by regional and international organizations that recommended the extension of social policies. In this vein, the summit of the ministers of public health of the Americas in 1972, and a few years later the Alma-Ata conference of the WHO, recommended the extension of public primary health care in order to improve the access of the poorest to the health system (Escudero 1981, 559–560; Interview Neri 2014a; Tittor 2012, 116–117). The ILO and the Inter-American Conference on Social Security continued to promote the strengthening and extension of social security arrangements through numerous policy papers, conferences and international conventions (Deacon 1997, 74; Falappa 2014, 219). On the national level, this global trend found its counterpart in a broad public approval of state intervention on social and economic issues that extended throughout most of the political spectrum, including the two major political parties PJ<sup>111</sup> and UCR (Belmartino and Bloch 1994, 251). Progressive political forces and popular class organizations that were interested in an extension of social protections for low-income earners thus enjoyed significant discursive power resources.

The evolution of important social policy actors since the late 1960s, and the distribution of associational power resources among them, was also favorable to a renewed attempt to achieve social policy expansion for low-income earners. Orchestrated by a global wave of left-wing protest and thought, and catalyzed by the 1969 riots against the conservative Onganía dictatorship, the political left emerged as an increasingly important actor in the Argentine political arena (Collier and Collier [1991] 2009, 496; McGuire 1997, 151). Trade union density started to recover towards the early 1970s and reached an estimated 43% in 1974 (Lamadrid and Orsatti 1991, 150–156). Although some independent Marxist organizations gained considerable strength in Córdoba, Tucumán, and other cities of the interior, in large part the rise of the left took place within the Peronist movement and to a certain extent also within the Radical Party and the established trade union movement. Within the Peronist trade union movement, the rise of the left was supported by Perón himself as a strategic move to counter the increasing autonomy of the more conservative trade union leaders (Collier and Collier [1991] 2009, 494–495; James 1976, 273– 275; McGuire 1997, 87–92). In 1968, the left-leaning faction of the trade union movement founded the CGT de los Argentinos, which adopted a progressive program that explicitly emphasized the necessity of universal access to health care, education, and housing (CGT de los Argentinos [1968] 2006; Godio 2000, 1051). However, the left remained a minority faction within the trade union movement and during the early 1970s the CGT reunified under rather conservative leadership (Rock 1987, 357). More lasting was the growth and radicalization of the left-leaning youth sector of the Peronist movement that had been taking place since the late 1960s. In 1972, the various left-leaning Peronist youth organizations united and formed the Peronist Youth, which developed an impressive associational strength and capacity for mass mobilization. The Peronist Youth became one of the main organizational pillars of Peronism alongside the trade union movement and was incorporated into its national decision-making bodies and attained considerable political influence (Godio 2000, 1069; James 1976, 282–284). The Peronist left was furthermore

<sup>111</sup> In 1971, Law 19.102 prohibited party names containing references to particular political leaders, so that the Partido Peronista changed its name to Partido Justicialista (PJ).

successful in organizing residents of poor neighborhoods and in 1973 the slum dwellers' movement Movimiento Villero Peronista (MVP) was founded (Ziccardi 1984, 148, 159–160). Within a short time, the MVP became able to mobilize thousands of low-income earners and sustained base-level structures in 450 slums throughout the country (Gillespie [1983] 1998). Left-leaning think tanks such as the Technological Council of the National Peronist Movement elaborated extensive economic and social policy proposals of clearly redistributive and universal character (Rougier and Fiszbein 2006, 53; CTMNP 1973a; 1973b; 1973c). At the same time, less radical center-left forces gained strength within the UCR Party. In 1968, center-left student activists founded the Junta Coordinadora Nacional and in 1972 Raúl Alfonsín founded the social democratic Movimiento de Renovación y Cambio. Although Balbín's conservative faction maintained its majority, the center-left had gained significant weight within the UCR by the early 1970s (McGuire 1997, 162–163). As was the case with the Peronist left, the center-left of the UCR supported redistributive social policy (Belmartino and Bloch 1994, 223; Interview Neri 2014a).


Table 12: Popular class power at the beginning of the period, 1973

Progressive currents had also considerably gained in strength at the base-level of the Catholic Church since the early 1960s and intensified their cooperation with other progressive and lower popular class organizations (Ghio 2007). Simultaneously, the weakening of the traditionally liberal employer associations and the growth of the more interventionist CGE beginning in the late 1960s weakened the capacity and willingness of employer associations to openly oppose redistributive policy proposals (Birle 1995, 134–135).

Summing up, by the early 1970s, the social, economic, and political context had become considerably more favorable to the renewed formation of an inclusionary governing alliance and the pursuit of redistributive social policy reforms.

## **Social Policy Expansion during the Rise and Fall of the Short Peronist Government**

The Córdoba riots of 1969 marked the beginning of the end of the military government as it initiated a massive increase in protest and guerilla activity, contributed to capital flight and the return of economic problems and divided the military about the question of how to respond to these challenges. When General Aramburu was killed by a guerilla group in May 1970, opposition within the military grew and President Onganía was replaced by General Levingston (Rock 1987, 351–356). Levingston implemented a more interventionist economic policy but was unable to control the growing inflation and discontent among employers increased (Birle 1995, 131–132). In early 1971 a second episode of riots took place in Córdoba, which led to the replacement of Levingston by General Lanusse. Lanusse promised elections and the restoration of civilian rule but tried to convince the major political parties to agree to the formation of a national coalition that would nominate him as its presidential candidate (Rock 1987, 356–358). In order to foster popular support, his government implemented several concessions such as the creation of a public fund for the construction of housing in 1972. The fund was financed by 2.5% employer contributions and aimed at improving workers' access to credits without however reaching the lowest income groups (Golbert 2010, 120). The formation of a national coalition nevertheless failed. Instead, the major parties joined forces and formed an alliance called Hora del Pueblo to oppose the dictatorship (Godio 2000, 1068).

As social protests and guerilla activity continued with intensity, the military government finally gave in and decided to schedule presidential elections for the 11th of March 1973. Perón himself was not permitted to run, but in an attempt to counter the radicalization of his followers, the proscription of the PJ was lifted (Adamovsky 2012, 298–299). In view of the presidential elections, the Peronist movement formed a broad electoral alliance (FREJULI<sup>112</sup>) that besides the PJ included minor parties from both the conservative and socialist spectrum, the left-leaning Peronist Youth, the CGT, the CGE and left-leaning slum dwellers' organizations (Godio 2000, 1068–1069; Ziccardy 1984, 148). Although the alliance was heterogeneous, its electoral core constituency was clearly to be found among the popular classes. The selection of candidates and the organization of the campaign was accompanied with euphoria by the

<sup>112</sup> Frente Justicialista de Liberación.

Peronist left. Finally, Héctor Cámpora, who was seen with sympathy by the Peronist Youth but with reservations by the CGT, was chosen as the presidential candidate. Several sympathizers of the Peronist left—but no union leaders—were nominated for Argentina's 23 governorships. Due to the official recognition of the Peronist Youth as one of the main pillars of the Peronist Movement, it was now also receiving a considerable proportion of the candidacies for the parliament (James 1988, 242). Angered by these decisions, most union leaders virtually pulled out of the electoral campaign, so that it was the Peronist Youth that set the tone of the campaign and organized most of the electoral mobilization (James 1976, 282; McGuire 1997, 161).


Table 13: Composition of the new governing alliance, March 1973\*

\* Due to intensive internal conflicts, the composition of the alliance and the dominance of different sectors within it varied between 1973 and 1976. The general character remained nevertheless inclusionary-redistributive. \*\* The majority of the trade union movement opposed universalistic health care reform, but supported the the extension of pension coverage to lowincome earners.

The landslide victory of the FREJULI in the March 1973 elections indicated the massive mobilization capacities that the Peronist Youth, and more generally the Peronist left, had developed during the preceding years (Godio 2000, 1069; James 1988, 242). The Peronist left had become de facto one of the main support pillars of the Cámpora government and achieved significant personal presence in various state apparatuses, providing several ministers, governors, parliamentarians, and technical staff (Adamovsky 2012, 300; Belmartino and Bloch 1982, 10; Godio 2000, 1070). In combination with the traditional dependence of the Peronist alliance on the electoral support of the urban and rural low-income groups, the increased influence of the left contributed to the development of an ambitious and decidedly progressive agenda of social policy reform.

In the area of health care, the newly elected government of Héctor Cámpora proposed the most far-reaching and progressive transformation yet by creating a single universal national health system that would provide equal access and treatment to all residents of the country. The demand for a universal and egalitarian health care system was shared by a heterogeneous spectrum of center-left and left political forces, and proposals for reform in this direction had been made by progressive forces within both major political parties, the PJ and the UCR (Belmartino and Bloch 1994, 223; Golbert 2008, 41; Interview Neri 2014a). The rapid growth of the left within the Peronist movement and its decisive role in Cámpora's campaign and government also enhanced its possibilities to influence the electoral platform and the government's policy agenda. Cámpora included the proposal of a national health system in his electoral platform and inaugural speech (Belmartino and Bloch 1994, 238; Cámpora 1973, 35).<sup>113</sup> Once he had assumed the presidency, Domingo Liotta, subsecretary of health, was put in charge of elaborating a concrete reform project. For this purpose, a commission of mostly progressive health care experts with different party affiliations was formed and started to design a clearly redistributive reform proposal that foresaw a reassignment of existing resources in favor of low-income earners and an additional expansion of overall public health care spending (Interview Neri 2014a). Initially, Perón, who still lived in exile, also approved the reform project (Belmartino 2010, 119–120).

However, resistance to the project was soon mounted from different sectors. The pharmaceutical industry did not like the idea that the new system would create a monopoly of demand, most associations of health care professionals preferred a largely private system, and the companies running health care facilities feared losing all their business opportunities (Belmartino and Bloch 1982, 11; Interview Neri 2014a). Most important, however, was the resistance stemming from trade union leaders who opposed the integration of the

<sup>113</sup> The social policy demands of the Peronist left were expressed through the think tank Technological Council of the National Peronist Movement, which elaborated policy programs for various fields of social and economic policy (Belmartino and Bloch 1982, 10; CTMNP 1973a; 1973b; 1973c). In the area of health care, the Technological Council recommended the introduction of a "Single National Health System that includes all the human, material and financial resources of the sector" and provides "free and equal access" (CTMNP 1973b, 14, author's translation). As a clear sign of the think tank's influence, Cámpora literally copied a part of the text of its policy paper into his inaugural speech to the parliament (Belmartino and Bloch 1994, 238; Cámpora 1973a, 62).

trade union-controlled health insurance funds into the new system (Belmartino and Bloch 1994, 238; Interview Neri 2014a). None of these actors opted for a strategy of open confrontation. The private health care industry and the medical associations intensified their lobbying activities. But according to Aldo Neri, one of the commission's leading health experts, they had only limited influence on the final project (Interview Neri 2014a). Much more effective in influencing the final law were the trade union leaders. These formed an alliance with the minister of social welfare, López Rega, who, as a leader of the Peronist right, had little sympathy for the progressive health reform project. López Rega, demanded modifications of the reform project from the subsecretary of health, Domingo Liotta, who offered little resistance in the face of the rapidly changing power relations within the governing alliance to the detriment of the left wing and in favor of the trade union movement (Belmartino 2010, 121; Interview Neri 2014a).

At the moment the ambitious reform project was announced, the Peronist left had attained unprecedented influence within the governing alliance and could count on the sympathy of the president, Héctor Cámpora (Adamovsky 2012, 300; Belmartino and Bloch 1982, 10; Godio 2000, 1070). This situation, however, quickly changed. On the 20th of June Perón finally returned from exile and thousands of his followers waited nearby the Ezeiza airport to greet him. What was supposed to become a big welcome event ended in what was called the massacre of Ezeiza when a group of right-wing Peronists opened fire on columns of left-wing Peronists (Rock 1987, 360). Perón decided to side with the right wing of the movement and started to attack the Peronist left as infiltrators. In the following month, after repeated press reports of Perón's dissatisfaction with Cámpora for being too closely related to the Peronist left, the latter was replaced by Lastiri, son-in-law of López Rega. Lastiri replaced several ministers and high-ranking officials that were associated with the Peronist left and scheduled new elections for the 23rd of September 1973. This time it was not the Peronist Youth but the CGT who played the leading role in the electoral campaign. In August, Perón further tightened his relations with the union leaders and confirmed a series of changes in the party hierarchy that vastly strengthened the right wing. In the September elections, Perón won with over 60% of the votes (Cavarozzi 2006, 188–189; James 1976, 286–288). Only a few days later, the assassination of the CGT leader Rucci by a leftist guerrilla group marked a further escalation of political tensions and contributed to increasing marginalization of the movement's left wing (Godio 2000, 1075).

In this context, the initially ambitious reform project was increasingly watered down. Already in August a version of the project was presented that exempted the health insurance funds of the trade unions and the provincial public health systems from obligatory integration into the new system and increased the number of union representatives in its administrative body. When the project finally passed the parliament in December 1973, little was left of the originally universalist and egalitarian spirit. In the end, only four provinces voluntarily joined the program (Belmartino and Bloch 1994, 244–247).

Nevertheless, between 1973 and 1975 the new governing alliance increased public health expenditure from 1.4% to 1.7% of the GDP and several provinces initiated primary health programs that improved access to health care among urban and rural low-income groups (Escudero 1981, 564; Vargas de Flood 2006, 188). At the same time, the governing alliance also continued to strengthen the health insurance system by increasing contribution rates from 2% to 4.5% for employers and from 1.7% to 2.7% for workers (Lo Vuolo and Barbeito 1998a, 337). Coverage was furthermore extended to recipients of non-contributory pensions (Feldman, Golbert, and Isuani 1988, 54). In sum, although the fragmentation and the inequalities of the Argentine health system persisted, both low-income sectors and organized labor benefited from health care expansion, reflecting in this sense the composition of the governing alliance's popular support base. After 15 years of stagnation, child mortality declined massively from 61.2 to 43.2 per 1000 between 1970 and 1975 (McGuire 2010, 311).

Income transfers were also significantly expanded during the short Peronist government. The electoral platform of the Peronist alliance appealed to the popular classes with the promise of a progressive extension of income transfers, explicitly announcing its intentions to augment the value of the minimum pension, reduce informality through tougher controls and reintroduce a progressive scale for pension replacement-rates (Cámpora 1973, 32–34). These and other proposals were then incorporated into a social compact which was signed by the government, the CGT and the CGE during early June 1973 in order to generate a consensus for the economic and social policy agenda of the government.114 Among other things, the social compact115 promised to "eliminate social marginalization through effective public policies in the areas of housing, education, health and social assistance," to eliminate unemployment, to increase low pensions and to increase the income of low-income earners and families with many children (*El Día*, June 9, 1973). The progressive character

<sup>114</sup> One of the most important components of the social compact was to freeze both wages and prices in order to reduce inflation (Rougier and Fiszbein 2006, 60–65). According to Oraldo Britos (Interview 2014), an important trade unionist and Peronist Senator at that moment, the CGE was skeptical because of the costs that the expansion of social policies would mean for the employers, but nevertheless agreed to the package as a whole.

<sup>115</sup> The social policy measures proposed by the social compact were later specified in the Social Security Program which was signed by the government, the CGT and the CGE in late November 1973 (Decree 466/73). This program also foresaw the implementation of an unemployment benefit, which was, however, never implemented. According to Oraldo Britos (Interview 2014), the trade unions did not insist on its implementation, given the low levels of unemployment and the many other pressing problems at that time.

of the social policy guidelines established in the social compact should be understood in the wider context. The prevalence of the ISI paradigm as the dominant perspective through which to analyze the economy made it possible to frame redistributive social policy as both benefiting the poorest and the economy as a whole. In the text of the compact it was explicitly argued that income redistribution and social housing projects would foster economic growth through increasing demand (*El Día,* June 9, 1973). In addition, minister of economy José Gelbard,<sup>116</sup> who had much influence on the elaboration of the compact, was well aware of the heterogeneous support bases of the governing alliance, and hence the necessity to incorporate significant concessions to the Peronist left and the rural and urban low-income groups (Godio 2000, 1071– 1072; Rougier and Fiszbein 2006, 72; Seoane 2003, chap. 7).

In October 1973, Law 20.541 established a 23% to 28% increase for all pension benefits below 1000 pesos. This initiated a series of policies that led to a 43%<sup>117</sup> increase in the real value of the minimum pension during the short Peronist government. As higher pensions increased much less, pensioners with lower income were the principal beneficiaries of the reshuffling of resources. At the same time, the social compact initiated measures that involved the trade unions in the control of employer contributions in order to reduce informality (Falappa 2014, 218–219). By means of a sectoral agreement that stipulated the automatic deduction of social security contributions at the moment of sale of agricultural products, these efforts were particularly successful in extending pension coverage to approximately 600,000 rural workers (Feldman, Golbert and Isuani 1988, 53–54). Because the agricultural workers' union was going to play an important role in the registration of workers, the measure also helped to strengthen union organization in the agricultural sector (Interview Britos 2014). In April 1975, Decree 796/75 established an increase in workers' pension contributions to 11%. In October 1975, a moment at which the CGT dominated the government, Law 21.118 implemented more generous rules for pension calculation and virtually abolished contribution requirements, which enormously facilitated informal workers' and low-income earners' access to pension benefits **(**Feldman, Golbert, and Isuani 1988, 54–55; Rougier and Fiszbein 2006, 100–106). Taken together, pension reform during the short Peronist government was favorable for low-income earners as it extended coverage and led to a more egalitarian distribution of resources among pensioners. In contrast to the reform path in the health care system, trade union leaders had no compa-

<sup>116</sup> José Gelbard was formerly head of the CGE and maintained close relations with the Communist Party (Brennan and Rougier 2009, 177; Seoane 2003). The latter might have contributed to the fact that the social policy guidelines included in the compact tended to favor the lowest income groups.

<sup>117</sup> Based on the comparison of the real value of the minimum pension in January 1973 and January 1976 (Feldman, Golbert, and Isuani 1988, 119).

rable interests in defending the status quo and actively participated in progressive reform in favor of low-income earners.

The policies expanding family and social assistance also benefited lowincome earners. In 1973, family allowances were increased by 40% and in 1974, following recommendations of the ILO, they were extended to pensioners and pregnant women (Interview Britos 2014; Falappa 2014, 218). Expenditure for social assistance programs increased from 1% of GDP in 1973 to 1.3% in 1975 and expenditure for social housing increased from 0.4% to 0.9% (Vargas de Flood 2006, 188). In the case of social housing, however, the benefits for low-income earners were very limited. During the initial phase of the Peronist government, progressive housing construction projects had been developed. However, with the marginalization of the Peronist left after the resignation of President Cámpora, the right-wing minister of social welfare channeled most of the resources towards better-situated beneficiaries (Ziccardi 1984, 151–156, 169).

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

The economic and political context during the early 1970s was favorable for a renewed attempt to extend social protections for low-income earners. Economic growth, industrial development, and low unemployment increased the structural power resources of the popular classes. At the same time, these structural developments and the global wave of left-wing protests contributed to the recovery of the trade union movement and a marked growth of the left's associational power resources. The return to democracy and the lifting of the ban on Peronism furthermore provided electoral power resources to the popular classes and increased the political weight of low-income groups that had only weak structural and associational power resources. At the same time, the continued predominance of the ISI paradigm provided ample discursive power resources to progressive actors to frame their demands for redistribution as viable and desirable for the society as a whole. Against this background, an inclusive electoral alliance was formed and won the elections on the 11th of March 1973. During the electoral campaign, the alliance mobilized popular class support through the promise of redistributive social policy. Once in government and with the left occupying important political positions, the alliance elaborated a highly progressive and ambitious social policy agenda.

However, the ambitious agenda was only partially implemented. Soon after the 1973 elections, political tensions within the governing alliance came to the forefront. Perón decided to side with the trade unions and the right wing of the movement and the left was increasingly marginalized. In this context, the trade union leaders, who had strong interests in maintaining control over their health insurance funds, could successfully stop the implementation of a universal public health system reform. Yet, in other social policy areas, the trade unions supported progressive reform. All in all, social policy change during the three years of democratic government was highly favorable for low-income earners. Spending on public health care and social assistance was significantly increased and the coverage of family allowances and pensions extended among low-income groups. Furthermore, the minimum pensions were significantly increased so that the distribution of resources among pensioners became increasingly egalitarian.


Table 14: Change in social protection for low-income earners, 1973–1976

## **4.5 The Retrenchment of Social Policy under a Neoliberal Dictatorship, 1976–1983**

By 1976, internal conflicts had significantly weakened the governing alliance. Furthermore, economic recession, high inflation and an increasing escalation of politically motivated violence aroused widespread public discontent. Against this background, the military staged a *coup d'état* in March 1976 and formed a reactionary governing alliance that excluded the popular classes. The labor movement and the organizations of the political left were fiercely repressed. In contrast, the big, anti-interventionist employer associations were incorporated into the core of the governing alliance and members of them assumed crucial positions in ministries and other state apparatuses. On this basis, the military government implemented a series of profoundly regressive social policy reforms that significantly reduced social protections for low-income earners.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

Steeply rising oil prices, import restrictions implemented by the European Community, and a further escalation of political tensions after the death of Perón on the 1st of July 1974 initiated a deterioration of the economic situation (Birle 1995, 137; Godio 2000, 1094). By 1976 Argentina had an enormous trade deficit, inflation reached over 440% and the economy shrank by 0.7% (Rapoport 2013, Table 5; Rock 1987, 363–364). During the following years, the military government attempted to control inflation by implementing a radically neoliberal economic program. Within a very short time, export and import taxes were reduced or eliminated, cross border financial transactions were liberalized and the domestic financial market was deregulated. At the same time, collective wage bargaining was suppressed, which led to an average real wage reduction of over 30% in 1976 alone (Beccaria, Groisman, and Maurizio 2009, 13–14). While these measures failed to bring inflation under control, they led to a profound financial crisis during the early 1980s and set into motion a process of industrial decline due to the fall of national consumption and the influx of cheaper imported products (Damill and Frenkel 2006, 110; Rock 1987, 369). Industry's share of the GDP fell from 40% in 1975 to 37% in 1983 and its share of the overall employment decreased from 31% to 27% (Timmer, de Vries, and de Vries 2014). In the strongholds of Argentine unionism, in industrial companies with more than 1000 workers, the reduction of employment reached nearly 50% (Abal Medina 2012, 113). These processes also correlated with a gradual deterioration of the labor market. The latter was not reflected in a massive increase in open unemployment. Instead, as Torrado (1992, 175–178) and Marshall (1988, 12–13) argue, it found its expression in a shrinking of the economically active population, which constituted a form of hidden unemployment. In sum, the processes of de-industrialization and labor market deterioration induced a gradual decline of the popular classes' structural power resources. The imposition of military rule furthermore deprived the popular classes of electoral power resources.

Although the ISI paradigm remained influential in Argentine public opinion, the economic difficulties of the mid-1970s and the failure of the social compact to resolve them began to nurture doubts about its adequacy. At the same time, signs of discursive change appeared on the international level, such as the election of radically neoliberal governments in Great Britain and the US. Furthermore, the massive increase of Argentina's foreign debt from less than 10 billion US Dollars in 1976 to over 45 billion in 1983 contributed to a relative increase in the influence of neoliberal-oriented international organizations, such as the World Bank and the IMF, compared to that of the WHO and the ILO, which continued to promote social policy expansion (Deacon 1997; Vargas de Flood 2006, 147).


Table 15: Economic and social context, 1976–1983

Sources: Rapoport (2013) for GDP growth and inflation; Timmer, de Vries, and de Vries (2014) for industry share of GDP and share of industrial employment; INDEC (2003) for unemployment and Jiménez (2013, 59) for informality.

The constellation of social policy actors and the distribution of associational power resources between them also evolved in a way that proved to be unfavorable for progressive forces. After the replacement of president Héctor Cámpora, left-wing leaders were quickly removed from influential positions within the governing alliance and the trade unions (James 1976, 286–288; 1988, 243–244). Simultaneously, the activity of right-wing paramilitary forces massively increased with the support of the Peronist right, parts of the military and the police (Rock 1987, 355; Cavarozzi 2006, 49–55). It is estimated that by 1975 the right-wing paramilitary organization Argentine Anti-Communist Alliance, commonly referred to as Triple-A, killed around 50 left-wing activists and intellectuals per week and thereby contributed to a rapid decline of leftist movements, such as the Peronist youth, the slum dwellers' movement MVP, progressive base-level structures of the Catholic church and independent Marxist groups (Esquivel 2000, 19–24; Rock 1987, 360–363; Ziccardi 1984, 168). While the Peronist left had constituted an influential component of the governing alliance during 1973, by mid-1975 it had been marginalized and its organizations decimated (Cavarozzi 2006, 53). After the military coup in 1976, the persecution and killing of tens of thousands of activists and supposed sympathizers finally led to the complete disintegration of many organizations of the left. The center-left current within the UCR party led by Raúl Alfonsín was one of the few progressive forces that were able to grow and extend its influence (McGuire 1997, 162–163).

Although the trade union movement suffered from fierce repression as well, union density decreased only moderately from an estimated 43% in 1974 to 41% in 1982–1983 (Lamadrid and Orsatti 1991, 156). Nevertheless, many of the base-level associations of the unions were crushed, so that despite the relative stability of union density its activist structures were significantly weakened (Rock 1987, 376).

Realignments among the employer associations were also detrimental to the prospects of social policy expansion. While the more interventionist CGE suffered from numerous defections during the mid-1970s and was outlawed in 1977, the neoliberal-dominated employer associations absorbed many of these defectors and closed ranks by forming the multi-sector employers' assembly APEGE in August 1975 (Birle 1995, 141, 163; Rougier and Fiszbein 2006, 94–108). The economic transformations that were initiated by the military junta's neoliberal policies further consolidated this development by economically weakening those parts of the industrial sector that were more inclined to support, or at least tolerate, state interventionist policies (Birle 1995, 176).

In sum, from the mid-1970s on, the social, economic, and political context turned increasingly unfavorable for the extension of social protections for lowincome earners.


Table 16: Popular class power at the beginning of the period, 1976

## **Exclusionary Military Rule and Regressive Social Policy Reform**

The promising context under which the Peronist alliance had assumed government in 1973 soon evaporated. Right from the beginning, the alliance was plagued by enormous internal tensions between the left-leaning Peronist youth, the centrist trade union sector and the right wing of the PJ. Perón was able to contain some of these tensions, yet when he died in July 1974 the willingness to cooperate of the different organizations within the alliance virtually disappeared and gave way to a violent escalation of conflicts. Several of the organizations that had constituted important components of the alliance in 1973, such as the left-leaning youth organizations and the CGE, were significantly weakened (Rapaport 2006, 547–551; Cavarozzi 2006, 53–55). The trade union movement retained high levels of affiliation, but politically it was paralyzed after it had failed to halt the deepening of the political and economic crises during the second half of 1975 when the movement enjoyed unprecedented control over the government. By 1976, the governing alliance had practically fallen apart and the country was plagued by an economic recession, sky-rocketing inflation and a massive lockout organized by a majority of the employer associations that had joined forces and formed the APEGE (Birle 1995, 142; Rapoport 2013, Table 5; Rock 1987, 360–366; Rougier and Fiszbein 2006, 94– 108).

Against this background, the military seized power in March 1976. The *coup d'état* led to the complete expulsion of the popular classes from the governing alliance (Belmartino and Bloch 1982, 12). Instead, the neoliberal-oriented employer associations came to constitute the core of the new alliance (Birle 1995, 175). José Alfredo Martínez de Hoz, who belonged to a traditional family of big landowners and was president of the Argentinean Employers' Council, became the minister of economy during the first five years of the dictatorship. Members of other employers' associations were put in charge of central positions in several ministries and the central bank. In particular, the ADEBA, the association of the biggest Argentine banks, was able to exercise significant influence on the government (Birle 1995, 169–174).

Due to the nearly total collapse of Peronism, the paralysis of the trade union movement and the pulverization of the left, the military junta possessed ample freedom of maneuver and encountered only weak resistance (Birle 1995, 147; Rock 1987, 367). Against this background, the junta adopted a program of radical neoliberal reform, which reflected the dominance of neoliberal thinking within the military elite and most of the employers' associations (Belmartino 2010, 125; Birle 1995, 169–174). Taxes on foreign trade were reduced or eliminated, cross border financial transactions were facilitated and the regulations on the domestic financial market were significantly loosened (Beccaria, Groisman, and Maurizio 2009, 13–14). Progressive direct taxes were increasingly replaced by regressive indirect taxes (Marshall 1988, 143–144). At the same time, wages and payroll taxes were cut massively; free collective bargaining and the right to strike were eliminated. The CGT was outlawed and trade union activity was fiercely repressed. Political opponents, especially those that were assumed to be left-wing, were persecuted and approximately 30,000 were murdered (Godio 2000, 1103). Labor laws were reformed in a way that increased the discretionary power of employers and reduced the costs of dismissals.


Table 17: Composition of the new governing alliance, 1976

As a result of these policies, the wage share of the GDP dropped from 44% in 1975 to only 31% by 1977, which constituted its lowest level since 1935 (Cortés and Marshall 1993, 402; Golbert 2010, 129; Lindenboim, Graña, and Kennedy 2005, 35; Rock 1987, 367–370). Income inequality, measured with the Gini-coefficient, increased steeply from 0.35 in 1975 to 0.43 in 1983 (Gasparini and Cruces 2010, 5). While these developments constituted a massive loss of welfare and freedom for the popular classes, the biggest employers' associations, except the CGE, were highly pleased by the junta's policies (Birle 1995, 164).

In the area of social policy the junta's agenda also reflected the interests of the anti-redistributive forces that dominated the governing alliance. Rather than compensating for the dramatic social consequences of its economic and labor policies, social policy reform produced an additional deepening of income inequalities (Marshall 1988, 11, 139). The retrenchment of social protections under the junta was detrimental for the popular classes as a whole, yet it was most profound for low-income earners, which reflected the unequal distribution of power resources among different segments of the popular classes. While the abolition of democracy eliminated the electoral power resources of the popular classes as a whole, the union movement preserved considerable associational power resources. Their internal left being marginalized, the unions focused largely on the defense of the social rights of the formal working class. This provided the junta with incentives to offer concessions to this segment of the popular classes in order to contain strike activity or facilitate collaboration with some of the most reactionary union leaders. Most low-income earners lacked access to the formal labor market and were not represented by the union movement, and as a result of this lack of associational power resources, the military government was able to more quickly dismantle social policies that benefited low-income earners than those which benefited the formal working class.

In the area of health care policies, the junta immediately implemented a radical retrenchment of the universal public health sector. Within only two years, spending as a percentage of GDP was cut by half, falling from 1.7% in 1975 to only 0.9% in 1977 (Vargas de Flood 2006, 188). One of the mechanisms used to consolidate this spending cut was to transfer public health care responsibilities to the provincial level without an equivalent transfer of resources. At the same time, a rapidly growing part of the resources that remained in the public health system was used for military hospitals that were not open to the general public (Marshall 1988, 88). The effects of these policies were dramatic. Entire hospitals were closed and the access of the popular classes to health care as well as the quality of attention was significantly reduced (Lo Vuolo and Barbeito 1998b, 126; Marshall 1988, 60–68). In late 1978 the junta passed Law 21.902, which completely derogated the national health system law that had been introduced by the Peronist alliance. In 1981, a new migration law limited the access of migrants to the public health system and obliged health care professionals to report undocumented immigrants to the authorities (Novick 2008, 140).<sup>118</sup> Low-income earners were disproportionally affected by all these restrictions and retrenchment policies because most of them lacked health insurance coverage, and hence depended vitally on the free and universal public health system. Due to the deteriorating labor market, the share of the population that had no health insurance coverage increased during the years of the dictatorship, so that increasingly fewer resources had to provide health care for ever more people. Formal workers in occupational groups with comparatively low incomes also suffered from these spending cuts, as the health insurance of these groups tended to provide poor services so that their members were often treated in public hospitals (Marshall 1988, 129–130). The only exception to this general trend of retrenchment in public health provision was that the military government during the first two years continued the expansion of rural health programs that aimed at providing basic and cost-effective access to primary health care. Between 1976 and 1977 the number of persons attended by such programs increased from approximately 500,000 to 1,400,000 and remained on that level during the following years (McGuire 2010, 137–138; Belmartino and Bloch 1984, 269). This expansion, however, went along with the elimination of beneficiaries' participation in the organization of these programs and the persecution and expulsion of many progressive

<sup>118</sup> This reinforced the exclusionary-regressive character of the governments' health policies as a significant share of low-income earners were migrants.

local political activists and organizations that had been involved during the preceding years (Bertolotto, Fuks, and Rovere 2012, 363). Despite the limited resources devoted to these programs, their continuity nevertheless contributed significantly to a further decline in infant mortality, reaching 33.2 per 1,000 in 1980 (McGuire 2010, 311).

In the area of social health insurance, the junta pursued two principal aims. The first was to eliminate union control over health insurance funds as part of a wider strategy of weakening the labor movement. The second was to promote further privatization of the system. In line with the junta's neoliberal approach and the interests of its social support bases, these reforms deepened the regressive nature of the system at the same time that they generated new investment and profit opportunities for the financial sector. In contrast to the drastic spending cuts in the egalitarian public health sector, spending on the in-egalitarian health insurance system was markedly increased from 1.3% of GDP in 1976 to 2% in 1983 (Vargas de Flood 2006, 188). According to Gerardo Munck (1998, 236–237), the junta used the spending increase to show that social health insurance worked better under military control than under control of the trade unions, which were accused of being corrupt and inefficient. In December 1976 the junta passed the first resolution that aimed at a restriction of the trade unions' influence on the use of health insurance resources (Belmartino 2010, 125), and soon after the junta's minister of economy, Martínez de Hoz, announced his plans to combine the elimination of union control with the privatization of the system. For both components of the proposal he received explicit support from all of the big employer associations (Belmartino 2005a, 174–175; Belmartino 2010, 126–132). In November 1979, the military government passed the Law 22.105 which ordered the dissolution of the CGT and simultaneously prohibited trade unions from administering health insurance funds. In August 1980, Law 22.269 established that the administration of health insurances would be taken over by an administrative body whose staff would ultimately be designated by the government's health insurance institute (INOS). At the same time, this law allowed beneficiaries to switch to private health insurance, which benefited the best-earning affiliates at the expense of the low- and middle-income beneficiaries (Belmartino and Bloch 1982, 13; Marshall 1988, 133–134). However, in a context of resurging trade union protest and the initiation of a profound economic crisis during the early 1980s, the law was never regulated by decree and went largely unimplemented (Interview Neri 2014b).

Although the major employer associations maintained generally good relations with the military junta, the junta's failure to stabilize the economy and the negative effects of some of its measures for the industrial and agricultural sectors gradually led to the emergence of tensions within the governing alliance (Birle 1995, 153–172). At the same time, external pressures on the junta increased. While the trade union movement had been paralyzed during the first two years of the dictatorship, a general strike in April 1979 signaled its comeback as a powerful actor in Argentine politics. Among the demands of the strike were the "normalization" of the trade unions and health insurance funds that had been put under control of state-nominated trustees. In 1980 several important unions reconstituted the CGT despite legal prohibition and in June 1981 another general strike took place (Godio 2000, 1117–1122). Against this background, the military government decided to refrain from implementing the privatization of social health insurance funds and made them part of their carrot and stick tactics towards the labor movement. Hence, instead of generally depriving trade unions of their control over social insurance funds, as foreseen by Law 22.269, the junta negotiated participation in the administration of each fund with individual union leaders, which served as an effective tool to divide the union movement and to provide cooperative union leaders with benefits.<sup>119</sup>

All in all, the dictatorship cut down the universalistic public health system, which resulted in a drastic deterioration of health care for most low-income earners. Due to the authoritarian character of the regime and the violent dismantlement of most left-wing political organizations, the lowest income sectors lacked power resources to resist these policies. In contrast, the trade union movement was able to mobilize significant associational and structural power resources in a moment in which economic problems were already arousing tensions within the governing alliance. This led the junta to make considerable concessions in the area of social health insurances, which however only covered the shrinking formal working class.

The junta's pension policies were also profoundly regressive. The drastic fall of real wages during 1976 caused a parallel reduction in the real value of the average pension by nearly 50% that year. During the following years, the value of the average pension recovered slightly, but in 1983 it was still 26% lower than in 1975 (MTEyFR 2001). In addition to an overall decrease of the pension value, the military government induced a redistribution of resources among pensioners that benefited higher-income pensioners to the detriment of low-income pensioners. In March 1976, shortly before the military coup was carried out, the value of the minimum pension constituted approximately 82% of the average pension, which indicated a highly egalitarian distribution of resources among pensioners. By August 1979, the value of the minimum pension had fallen to only 35% of the average pension (Feldman, Golbert, and Isuani 1988, 119–120). Furthermore, the government implemented stricter access criteria. In June 1976, the junta repealed Law 21.118, which had virtually abolished contribution requirements eight months before (Law 21.327). In November that year, the government finally raised contribution requirements to 15 years (Law 21.451). These measures excluded many low-income earners from access to pensions, as most of these worked in informal jobs and were therefore

<sup>119</sup> According to Aldo Neri (Interview 2014b), most health insurances were de facto under control of trade unionists when he took over as the minister of health in 1983.

unable to accumulate sufficient years of contribution. In October 1980, Law 22.293 completely abolished employer contributions to the pension system and social housing fund (FONAVI), which meant an overall reduction of employer contributions by a stunning 20 percentage points and a lasting destabilization of the system's financial balance. In order to mitigate the negative financial effects, regressive value-added taxes were increased (Lo Vuolo and Barbeito 1998b, 137). This significantly shifted the financing burden from capital to the popular classes and was heavily criticized by trade unions (Arza 2010, 268– 169; Feldman, Golbert, and Isuani 1988, 58; Marshall 1988, 13). When the country entered into economic crisis in 1981–82, the government felt forced to cut pensions again. However, as was the case with health care reform, the junta was now more inclined to make concessions to the popular classes in the context of growing internal tensions and a revival of trade union protest. In contrast to the drastic pension cuts of 1976, the government decided to maintain at least the real value of the minimum pension, so that by 1983 the distribution of resources among pensioners had become again more egalitarian (Cárcamo Manna 1998, 15; MTEyFR 2001).

As the financial deficit of the pension system ran out of control, the government began to transfer resources from the family allowance funds, which, in turn, resulted in a reduction of family allowance expenditure from 1.5% of GDP in 1975 to 0.5% in 1983 and a corresponding fall of the real value of this benefit (Falappa 2014, 222–223; Marshall 1988, 118; Vargas de Flood 2006, 188). In a similar vein, the military government radically cut down social assistance expenditure from 1.3% of GDP in 1975 to 0.4% in 1983 (Vargas de Flood 2006, 188). In the area of social housing, spending cuts were combined with the application of regressive access rules that de facto excluded most lowincome earners. At the same time, the junta carried out violent evictions of unregistered settlements, which taken together with the aforementioned policies led to a massive deterioration of the housing situation of low-income earners (Belmartino 2005e, 251).

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

The military formed an exclusionary governing alliance that heavily relied on the support of neoliberal-oriented employer associations and conservative parts of the middle and upper classes. Immediately it started to implement a radical neoliberal reform program that involved a far-ranging restructuring of the system of social policies in favor of capital and high-income earners and to the detriment of the popular classes.

Figure 5: Evolution of income inequality measured with the Gini-coefficient, 1976–1983

Sources: Fishlow et al. (1993, 3) for 1976–1979 and Gasparini and Cruces (2010, 5) for 1980– 1983.

Initially the junta faced hardly any resistance, as the collapse of the Peronist alliance in conflict had either decimated and dispersed the associational power resources of popular class forces, as happened with the organizations of the Peronist left, or politically paralyzed them, as was the case with the trade union movement and the PJ party. At the same time, the establishment of an authoritarian regime deprived the popular classes of electoral power resources, so that the junta's lack of support among the sizeable but weakly organized lower popular classes appeared to be of little threat. Therefore, the dictatorship was able to advance quickly with its agenda during its first years. However, from 1979 on the trade union movement gradually recovered the capacity to act and began to mobilize their associational, structural, and discursive power resources to confront measures of the government and to extract concessions. The lowest income sectors of the popular classes, however, lacked comparable power resources. The destruction of the organizations of slum dwellers and of the political left and their deprivation of electoral power resources resulted in an extreme marginalization in the political arena. This unequal distribution of power resources within the popular classes was also reflected in the way social policies were cut. In combination with growing tensions within the governing alliance and a deterioration of the economic context towards the early 1980s, trade union protests led the government to make concessions and to refrain from the original plan of complete separation of the social health insurance funds from union control. Expenditure on social health insurance was also increased, in contrast to the original plan of strengthening the private health insurance sector. However, in those social policy areas that provided most of the social protection for the lowest income sectors, retrenchment was far-reaching. The public health system, social assistance policies, housing programs, and family allowances were massively reduced. The regressive reform of the pension system also disproportionally affected low-income earners.

The political, economic, and social transformations initiated by the military government left a complicated legacy, which would go on to shape the dynamics of later struggles over social policy in important ways. On the political level, the persecution and assassination of thousands of progressive political activists lastingly weakened the Argentine left as a political actor. The only exception to this development was the increasing strength of the center-left wing of the UCR party. The financial strengthening of the social health insurances and the weakening of the public health system increased the interests of the trade unions in preserving this fragmented and in-egalitarian system. On the economic and social level, the junta's neoliberal adjustment policies initiated a profound transformation, which had an important impact on later social politics. The passage from a process of industrialization to one of de-industrialization and financialization induced a long-term deterioration of the structural power resources of the labor movement and provided a severe challenge to its associational power resources, as trade unions were well organized in the declining industrial sector.


Table 18: Change in social protection for low-income earners, 1976–1983

Moreover, with the demise of local market oriented industrial companies, the more interventionist employer association CGE lost importance. This contributed to the overcoming of the ideological division among employers' associations in favor of a neoliberal consensus. Finally, the exponential increase of public and private foreign debts would constitute a heavy burden during the following decades and provide neoliberal IFIs with significant political influence.<sup>120</sup>

## **4.6 Return to Democracy, Stagflation, and the Failed Attempt of Inclusionary Social Policy, 1983–1989**

After having caused a severe economic crisis and provoked a war over the Falkland Islands that ended in a clear-cut defeat, resistance to the military government grew stronger during the early 1980s. Finally, democratic rule was reestablished in late 1983. The return to democracy, a temporary comeback of the ISI paradigm and the recovery of trade unions and political parties created once again a chance for the formation of a more inclusionary governing alliance. Under the leadership of its center-left faction, the UCR won the elections, making significant inroads into the traditionally Peronist popular class electorate and proposing a progressive social policy agenda that contemplated a significant expansion of social protections for low-income earners. However, in the face of economic crisis and the fierce opposition of most corporate organizations, the alliance largely failed to implement progressive social policy change.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

The democratically elected government of Raúl Alfonsín inherited a heavy burden of social and economic problems from the military dictatorship. During the crisis of 1981 and 1982, the GDP had fallen by over 11% and by 1983 inflation reached over 400% (Rapoport 2013, Table 5). Foreign debt amounted to over 45 billion dollars and foreign debt services swallowed 50% of Argentina's export earnings (Rock 1987, 391; Vargas de Flood 2006, 147). After 1983, continued capital flight, deteriorating world market prices for grains and high interest rates further exacerbated these problems (Basualdo 2011, 58–60; McGuire 1997, 185). Although the Alfonsín government was unable to lastingly reverse the tendency of economic and social deterioration during the 1980s, its heterodox economic and social policies were able to stabilize the situation during its first years in office. Through a package of economic policies referred to as Plan Austral, inflation was reduced to 82% in 1986 and economic growth temporarily returned (Rapoport 2013, Table 5). However, from that year on a renewed deterioration of the terms of trade and a bad harvest

<sup>120</sup> An analysis of the social, economic and political legacy of the military government can be found in Basualdo (2011, 41–58).

exacerbated the difficulties of servicing foreign debt and put enormous pressure on the central bank's foreign currency reserves (Birle 1995, 239). In March 1988, the government was finally forced to cease interest payments to commercial banks. By February 1989, foreign currency reserves ran out and the country entered into a deep hyperinflation crisis (McGuire 1997, 215).


Table 19: Economic and social context, 1983–1989

Sources: Rapoport (2013) for GDP growth and inflation; Timmer, de Vries, and de Vries (2014) for industry share of the GDP and share of industrial employment; INDEC (2003) for unemployment and Jiménez (2013, 59) for informality.

Despite these adverse economic developments, the heterodox policies of the Alfonsín government contributed to a relative stability of the country's industrial base. Between 1983 and 1989, industry's share of GDP remained relatively stable between 35% and 37% and its share of the employed population oscillated modestly between 26% and 29% (Timmer, de Vries, and de Vries 2014). At the same time, the labor market underwent profound changes. There was a marked increase in women's labor market participation. While women's share of the economically active population had increased by a mere seven percentage points between 1947 and 1980, during the 1980s alone, it increased by another seven percentage points.<sup>121</sup> Unemployment increased from 5.5% in May 1983 to 8.1% in May 1989 (INDEC 2003). Informality among salaried workers rose from less than 20% in 1985 to 29% in 1989 (Jiménez 2013, 59). According to McGuire (1997, 185), this expansion of unemployment and informality after 1976 generated for the first time an extensive reserve army of labor that caused a moderate but steady erosion of the popular classes' structural power resources.

On the institutional level, the re-establishment of democracy returned electoral power resources to the popular classes, which strengthened these classes

<sup>121</sup> Author's calculations based on Torrado (1992, 92) and CEPAL (2017b).

as a whole and increased the political weight of the lowest income sectors within them.

On the discursive level, the gross economic failure of the military's neoliberal economic policies and the massive human rights violations that were associated with their implementation led to a profound discrediting of the neoliberal paradigm in the eyes of the mass public and a revival of the popularity of Keynesian ideas and the ISI-paradigm (Munck 1987, 109; Portantiero 1994, 303; Svampa 2005, 25). This provided progressive political forces with discursive power resources to justify the expansion of social protections for lowincome earners as fostering overall economic growth and being desirable for the society as a whole. The change in the public mood facilitated the construction of electoral and associational alliances between the lower popular classes and other sectors of the society in order to press for progressive social policy change. The neoliberal paradigm only started to become more popular again during the late 1980s, when the economic situation deteriorated and a growing share of the public perceived the ISI-paradigm to be unable to provide adequate solutions.

These broader economic, social and political developments also influenced the evolution of the main actors involved in social politics. The return of democracy and the lifting of repression favored a process of associational recuperation of popular class actors, in particular of the trade unions and political parties (Palomino 2005, 391). The PJ was stunningly efficient in re-building its base-level organizations. Base units of the party virtually mushroomed up everywhere and within a very short time, the PJ had signed up approximately three million members (Levitsky 2003, 49). However, the PJ party of 1983 was very different from the one that had existed ten years earlier. The persecution and killing of thousands of activists and supposed sympathizers had drastically decimated the Peronist left (Cavarozzi 2006, 53). In 1983 the PJ party was dominated by centrist union leaders, which were much less enthusiastic about defending the interests of the lowest income sectors (Goldin 1997, 123–130; Palomino 2005, 390). Things were different in the UCR Party, where the social democratic wing was able to extend its influence. After the death of the conservative party leader Ricardo Balbín in 1981, a member of the social democratic faction, Raúl Alfonsín, won the internal elections and became president of the UCR. Under center-left leadership, the UCR developed concepts for universalizing social protections, in particular in the health care sector, and actively reached out to the traditionally PJ-dominated workingclass and poor neighborhoods in order to mobilize electoral support (Belmartino 2010, 138; McGuire 1997, 162–191; Interview Neri 2014a).

The associational strength of the trade union movement had been surprisingly resilient during the years of the military dictatorship. The return to democracy and constitutional rights, as well as the extensive strike activity during the 1980s, enabled the trade unions to increase union density from an estimated 41% in 1982–1983 to approximately 44% in 1989 (Lamadrid and Orsatti 1991, 156). However, under conservative leadership, trade union involvement in social politics was ambivalent regarding the interests of low-income earners. In the health care sector, interests of the union leaders in preserving control over health insurance funds led to union opposition against universalizing reforms, while in other social policy areas the unions were more supportive of progressive change.

Right-wing paramilitary and military repression had also decimated a wide array of other progressive social movements such as movements of residents of poor neighborhoods and feminist organizations (Fisher 2000, 324; Ziccardi 1984, 171). With the return to democracy, social movements of these types began to gradually re-emerge. The majority of them favored universalizing social policy reforms, as the existing fragmented system tended to disadvantage or even exclude most residents of poor neighborhoods as well as most women. Yet, the political influence of these movements remained very limited (Fisher 2000, 324; Merklen 2005, 51–59).


Table 20: Popular class power at the beginning of the period, 1983

Among the employer organizations, the 1980s brought a consolidation of the neoliberal consensus and a general rejection of redistributive social policies. Although the more interventionist employer association CGE was re-founded in 1984, it did not recover the influence that it had had during the preceding decades (Birle 1995, 225; 1996, 207–209). Moreover, employer-financed neoliberal think tanks, such as the Fundación Mediterránea, the Fundación de Investigaciones Económicas Latinoamericanas (FIEL), and the Centro de Estudios Macroeconomicos de Argentina (CEMA), became increasingly influential through lobbying and a strong presence in the media (Basualdo 2011, 56).

All in all, during the 1980s, the context was ambivalent for the extension of social protections for low-income earners. While such proposals of social policy extension seemed to be electorally viable, economic restrictions were tight and the particular constellation of interests and the distribution of power resources provided difficult conditions for the construction of a governing alliance that could sustain a progressive reform process.

#### **The Formation of a Weak Center-Left Governing Alliance and its Limited Success in Expanding Social Protections for Low-Income Earners**

By the early 1980s, pressure on the military government increased from various sides. The labor movement, which had been largely paralyzed during the first years of the dictatorship, gradually regained a capacity to take action and staged successful general strikes in April 1979 and June 1981 (Godio 2000, 1117–1122). The protests of human rights groups also grew in strength and visibility, particularly the ones of the Mothers of the Plaza de Mayo , who demanded justice for all those who were killed or kidnapped by the dictatorship (McGuire 1997, 175; Rock 1987, 384–385). At the same time, the gross failure of the junta's economic policy and the advent of crisis led to the emergence of tensions within the governing alliance. From 1979 on, internal criticisms were voiced by the industrial and agrarian employers' associations, which were discontented with the government's exchange rate policies that led to a massive overvaluation of the currency. With the deepening of the economic crisis in 1981 these internal tensions grew in intensity and even the financial sector associations, which were among the closest supporters of the regime, started to express discontent with the government for not being radical enough in implementing neoliberal policies (Birle 1995, 153–175).

Under pressure from within and without the governing alliance, General Videla was replaced by General Viola as the head of the government in March 1981. However, Viola was also unable to stabilize the economy and to control popular protests. In December 1981, he was replaced by the nationalist hardliner General Galtieri (Rock 1987, 374–375). While popular resistance continued to grow and a massive general strike paralyzed large parts of the country in March 1982, the Junta decided to invade the British-controlled Falkland Islands. In the short run, this strengthened nationalist sentiments among the mass public and helped to recover popular support for the military. However, as the war resulted in a stunning defeat after only 72 days, popular protests and tensions within the governing alliance re-emerged with even higher intensity than before. Against this background, General Galtieri was replaced by General Bignone, who finally scheduled elections for October 1983 (Birle 1995, 147– 148; Rock 1987, 375–385).

With the elections nearing, the major parties were stunningly quick in rebuilding their base level organizations and reaching out for popular class support. Within a short period, the PJ had affiliated approximately three million members and the UCR came close to 1.5 million (Levitsky 2003, 49). Despite the significantly broader membership of the PJ, Raúl Alfonsín of the UCR was able to mobilize broad-based support and won the presidential elections with 52% to 40% of the vote (McGuire 1997, 183). The UCR's victory was possible for several reasons. First, among the major candidates, Alfonsín had been the most vehement opponent of the military junta, while some of the conservative trade union leaders that dominated the PJ at that time were accused of collaboration with the regime. Alfonsín was hence perceived as the most credible candidate to restore and consolidate democracy and the rule of law. After seven years of violent dictatorship, this reputation gained him massive popular support (Birle 1995, 191; Portantiero 1994, 295–298). Second, the center-left currents122 of the UCR that supported Alfonsín had successfully constructed baselevel structures that were able and willing to reach out to working-class and poor neighborhoods that were traditionally dominated by the PJ. In contrast to the 1973 elections, in 1983 it was the UCR rather than the PJ that received impetus from left youth organizations. This provided Alfonsín's campaign with a mobilizational dimension that proved to be highly effective (McGuire 1997, 191). Third, in conjunction with the base-level activities of the party's youth in poor and working-class neighborhoods, the social-democratic platform of Alfonsín was able to generate hopes for real material improvements for lower income sectors, particularly in regard to social policy. In a context of a renewed popularity of Keynesian ideas and the ISI Paradigm, progressive social and economic policies were furthermore widely perceived as being beneficial for the society as a whole, by promoting economic recovery through revitalizing local demand. The combination of these factors made it possible for Alfonsín to construct a broad electoral support base, which involved a majority of middle-class voters, who traditionally supported the UCR, and a significant share of the poor and working-class voters, who had traditionally constituted the stronghold of the PJ (Cavarozzi 2006, 95–96; Adamovsky 2012, 340).

<sup>122</sup> Alfonsín received his main intra-party support from three center-left organizations within the UCR: the Movimiento de Renovación y Cambio, of which Alfonsín was one of the founders in 1972, and the youth and student organizations Junta Coordinadora Nacional and Franja Morada.


Table 21: Composition of the new governing alliance, 1983\*

\* In the face of failed attempts to stabilize the economy and to form a strong progressive governing alliance, towards the second half of the presidency the influence of neoliberal sectors within the UCR grew. \*\* The alliance is classified as trundcated in the sense that it attempted to incorporate the popular classes, however failed to achieve the adherence of major parts of the trade union movement. Its electoral support among the popular classes also remained limited.

The newly elected government pursued with much vigor an ambitious reform agenda that comprised economic and labor market policies inspired by Keynesian ideas, a democratization of the trade union movement, a restriction of the autonomy of the military, the legal prosecution of human rights violations that occurred during the dictatorship and an expansion of social protections for low-income earners by means of expanding social assistance policies and universalizing the health care system (Birle 1995, 198; Golbert 2010, 132– 136; McGuire 1997, 183–186; Portantiero 1994, 302). However, it soon turned out that the margins of maneuver of the government were much smaller than it had expected. An important limitation resulted from the complicated fiscal, economic and social legacy left by the dictatorship. Neither did the return to Keynesian policies result in the expected recovery of the economy nor did the IFIs make any significant concessions to the Argentine government (Birle 1995, 228–229). Even more important were the limitations resulting from the political weakness of the governing alliance. Although the UCR was able to repeat its electoral success in the 1985 legislative elections, it never reached a majority in the Senate and the 1987 elections even resulted in a loss of its majority in the Chamber of Deputies. In contrast to its initially broad electoral support, it failed right from the beginning to incorporate any significant support from corporate organizations.

Immediately after taking office, the government attempted to widen the UCR's influence within the trade union leadership by democratizing union elections and by strengthening the representation of minorities. However, instead of generating trade union support, the measure led the mostly Peronist trade union leaders to overcome their divisions and to form a united front against the government (Golbert 2010, 133; McGuire 1997, 183–193; Portantiero 1994, 302). Due to the opposition of the senators of the PJ and two minor parties, the corresponding bill failed to pass the Senate, in which the UCR held no majority on its own (Interview Britos 2014). Hence, from the beginning, relations between the trade union movement and the government were characterized by conflict and in only five years the CGT organized 13 nationwide general strikes (Palomino 2005, 390). At the same time, the government attempted to establish a supportive relationship with the industrial employers' association UIA, which however rejected the government's heterodox economic policy and took an oppositional stance, as did the other the employers' associations (Birle 1995, 210). To make things even more complicated for the government, the CGT and several employers' associations repeatedly formed alliances and jointly opposed the government's policies (Palomino 2005, 390; Portantiero 1994, 300–301).

This opposition of both employer and working-class organizations put severe limits on the government's capacity to implement its social policy agenda. The centerpiece of this agenda was a universalizing health care reform that aimed at creating a national health insurance that would equally cover all inhabitants of the country, independent of whether they were able to contribute or not (Golbert 2010, 135). The most emphatic promoter of this reform was Aldo Neri, who, as one of the country's leading experts in public health, had already played a crucial role in the elaboration of the universalizing health reform proposal of the last Peronist government. Neri belonged to the center-left wing of the UCR Party and had been in close contact with Alfonsín since the early 1970s. As was the case with the PJ, also within the UCR the more conservative sectors viewed egalitarian health care reform with skepticism while the party's left supported it. The rise of the center-left faction of the party under the leadership of Alfonsín therefore created the party internal support base for the elaboration of an egalitarian health care agenda (Interviews Isuani 2014a; Neri 2014a; 2014b). During the electoral campaign, health care universalization was discussed as part of a wider package of social democratic reforms that increased the UCR's electoral appeal to the lower income sectors (Belmartino 2010, 138). Once the UCR was in government, Neri was put in charge of the Ministry of Health and immediately began to work out concrete drafts for health care reform. By and large, these drafts were inspired by the same

universalistic aims as the national health system proposal of 1973 (Interview Neri 2014b). Initially, the government planned to advance quickly on this issue. However, against the background of the failure of the trade union reform bill to pass the Senate, the government decided to refrain from sending the health reform bill to the parliament in early 1984 in order to avoid a second failure and a further escalation of the confrontation with the trade union leaders, who had a strong interest in defending the existing health insurance system (Belmartino and Bloch 1994, 324). The first detailed reform proposal for a national health insurance system was finally presented to the public in November 1984. The core of the proposal was the creation of a unified health insurance system with universal coverage and equal quality of treatment for all residents of the country. The coverage of the lowest income groups, among them many informal and unemployed workers and their dependents who were not able to pay contributions, was going to be financed by government contributions to the insurance. The proposal foresaw the incorporation of the already existing social health insurance funds into the national insurance system as mere financing bodies. The influence of the trade union leaders on these financing bodies was furthermore going to be drastically reduced, as their administration would be carried out by a body elected by the beneficiaries (Belmartino 2010, 139; Interview Neri 2014b). At the same time, the 3% workers' contributions were going to be abolished and employer contributions raised from 4.5% to 7.5% (Golbert 2010, 135).

As Neri had actively participated in the process of failed health care reform during the 1970s, he was well aware of the capacities and willingness of the trade union leaders to block any kind of reform that would restrict their control over the health insurance funds. Against the background of the ample electoral victory and significant inroads into the traditionally Peronist working-class electorate, Neri's proposal aimed at driving a wedge between the rank and file of the unions and their predominantly Peronist leaders. By means of the elimination of workers' contributions and the introduction of democratic elections of the insurances' administrators, the proposal provided important material and political benefits for the affiliated workers while at the same time restricting the influence of the union leadership. In practice, however, the proposal reinforced the aggressively oppositional stance of the trade union leaders, who proved capable of mobilizing their rank and file to participate in a series of general strikes against the government's policies (Interview Neri 2014b).<sup>123</sup> Furthermore, the increase of employers' contributions and the expansion of the role of the state in the proposed national health insurance aroused the clear opposition of the employer associations, both within and without the health sector (Alonso 2000, 78; Belmartino and Bloch 1994, 346–348). Supported by

<sup>123</sup> The first general strike took place in September 1984. The second took place in May 1985 and 11 more followed until the end of Alfonsín's presidency in 1989 (Cavarozzi 2006, 196–202).

the World Bank, the IMF and the IADB these associations actively promoted privatization and decentralization of the health sector (Belmartino 2005c, 294, 308; Tittor 2012, 182). The majority of the associations of health care professionals also rejected the reform proposal for potentially reducing their entrepreneurial independence (Interview Neri 2014b). To make things even more complicated, delegates of the CGT and the main employer associations held several meetings during the second half of 1984 in order to join forces against the government. In late January 1985, they finally agreed on a joint manifesto of twenty demands that involved the consolidation of the existing, fragmented social insurance system and the complete restoration of trade union control over it. In exchange for support on this issue, the trade union leaders even agreed to a variety of demands of the employer associations that were detrimental to the working class, such as wage restraints and privatizations (Birle 1995, 232–233; Godio 2000, 1163–1164).

Despite the hostility of the corporate organizations, the popularity and the margins of action of the government increased during the second half of 1985 due to the initial success of its heterodox economic policy program Plan Austral in reducing inflation (Birle 1995, 234–236). In this context of apparent strength and in the hope of gaining a majority in the Senate in the approaching legislative elections, the government decided to send the health reform bill to the Congress in September 1985 (Belmartino 2010, 140). However, despite a favorable electoral result in the November 1985 elections, the governing alliance failed to win a majority in the Senate. This meant that there was a risk for the government that the health reform bill would not pass the Senate. Against this background, Alfonsín decided to withdraw the bill and to support the more hesitant faction of his government, represented by the Ministry of Labor and the Social Security Department, which favored a negotiated reform instead of risking an unfavorable decision in the parliament. In April 1986, Neri, who represented a more confrontational approach and was not willing to accept a dilution of the contents of the project, finally resigned as minister of health (Interview Neri 2014b). As the general economic situation began to deteriorate again during 1986 and the threat of a military coup re-emerged in an acute manner,124 the government finally decided to provide some of the more cooperative trade union leaders with decisive influence on the elaboration of labor and health care legislation in order to lower the tensions with the labor movement. In late March 1987, Alfonsín appointed Carlos Alderete, leader of the light and power workers' union, as minister of labor. On the 6th of May finally, an agreement regarding health care reform was signed, which was heavily influenced by the involved union leaders (Belmartino and Bloch 1994, 351; Belmartino 2010, 142). As a result of this process, the two health reform laws that passed the Congress in late 1988, Law 23.660 and Law 23.661, by and

<sup>124</sup> During Easter 1987, the government was even confronted with an open rebellion of parts of the military.

large consolidated the existing system of fragmented social health insurances under trade union control. In order to address the financial problems of the health insurance funds, employer and worker contribution rates to the pensioners' health insurance were increased by 2 percentage points in June 1988 and employer contribution rates to workers' health insurance were raised by 1.5 percentage points in early February 1989 (Lo Vuolo and Barbeito 1998a, 338). Law 23.661 mandated the extension of health insurance coverage to hitherto excluded sectors under egalitarian premises. The concrete mechanisms and modalities of incorporation were left to be established by decree, however, in the face of the following period of hyperinflation and the assumption of the neoliberal government of Carlos Ménem, these decrees were never promulgated and so Law 23.661 did not result in any relevant change regarding lowincome earners' access to health care (Interview Neri 2014b).

In sum, due to tight economic restrictions and fierce trade union and employer opposition, the government was neither able to shift resources from the inegalitarian health insurance system to the public health sector nor to assign major additional fiscal resources to the latter. Public health sector expenditure increased from 1% to 1.3% of the GDP between 1983 and 1989 (DAGPyPS 2011). However, at the same time, the reduction of health insurance coverage in the face of growing informality in the labor market meant that a growing share of the population was dependent on the public health sector.<sup>125</sup> Against this background, health indicators improved only moderately between 1983 and 1989, with child mortality falling from 30 to 25 per 1,000 live births and life expectancy increasing by less than one and a half years (World Bank 2018).

In the area of pension policies, the agenda of the Alfonsín government was far less ambitious and heavily shaped by the financial crisis of the system. Nevertheless, regarding coverage and the distribution of the available resources, it was progressive and therefore fit well with the center-left character of the governing alliance and its attempt to consolidate a support base among low-income earners.

When the government assumed office, the pension system was in a disastrous financial situation. Furthermore, from 1982 on, the military government had started to deviate from the legal indexation mechanism and raised pension benefits at a slower pace than inflation, which led to a rapid decline of the real value of pensions. The first response of the new government to these problems was to gradually reintroduce employer contributions from 1984 on (Lo Vuolo and Barbeito 1998a, 338). Nevertheless, due to the general economic crisis, growing informality and unemployment and the effects of the long-term process of population aging, this measure was far from sufficient to solve the financial problems of the system. In this context, the government continued the practice of deviating from the legally defined pension adjustment criteria so

<sup>125</sup> Informality among salaried workers rose from less than 17% in 1980 to nearly 30% in 1988 (Beccaria and Groisman 2009b, 110; CEDLAS and World Bank 2018).

that by 1988 the real value of the average pension had fallen by approximately 36% compared to 1980 (Arza 2006, 88; Madrid 2003, 103). This caused a flood of legal demands by pensioners against the state. In 1986, the government decreed a situation of pension emergency, suspended all legal processes and de facto introduced a new benefit calculation formula that consisted of a minimum pension plus a reduced replacement rate. In contrast to the replacement rates that had been introduced by the military government of Onganía, the new mechanism provided higher replacement rates for low-income earners and therefore contained a progressive element (Arza 2006, 88). At the same time, the government took measures to keep the access of low-income earners to pension benefits relatively open, despite a rapidly deteriorating labor market. Reiterated moratoriums preserved the eligibility for benefits of retiring workers who had achieved pension age but lacked sufficient years of contribution due to unemployment or informality (Lo Vuolo and Barbeito 1998b, 162–163). Furthermore, Law 23.226 from 1985 guaranteed a right to a widow pension also to unmarried life partners while the number of non-contributory pensions, which mainly benefited low-income households, was increased from 30,400 in 1981 to 137,500 in 1989 (Marshall 1988, 119; Bertranou and Grushka 2002, 41). According to Rubén Lo Vuolo and Alberto Barbeito (1998b, 161), these measures contributed to a situation in which nearly all elderly persons who had been economically active got access to pensions during the 1980s. Despite the overall progressive character of the government's pension policies, they contributed little to consolidating a political support base among low-income earners. In the context of a flood of legal claims, the declaration of pension emergency and the decline of the average pension, a perception of government failure in this area prevailed throughout the society.

Another social policy that aimed at the alleviation of poverty and the strengthening of lower-class support was the Plan Alimentario Nacional (PAN), which the government implemented in 1984 (Garay 2010, 45). The program was massive in scale and consisted of the regular provision of food boxes to up to 1.4 million poor households (Lo Vuolo and Barbeito 1998b, 180). Overall spending for social assistance increased from 0.4% of GDP in 1983 to nearly 0.9% in 1989 (DAGPyPS 2011). In 1985, the government furthermore introduced an unemployment allowance , which covered hitherto formal workers. The benefit was paid from the family allowance funds and consisted of 70% of the minimum wage plus family allowance benefits for a period of up to four months. The average number of beneficiaries ranged from 56,795 in 1985 to 153,059 in 1989, which represented 9% and 17% of the unemployed (Chebez and Salvia 2001, 6). In sum, these policies constituted a moderate improvement of low-income earners' social protection against unemployment and poverty and contributed to the government's success in competing for the traditionally Peronist popular class electorate in the 1985 legislative elections. However, this support eroded soon after. The increasingly rapid deterioration of the economic situation during the following years meant that living conditions for low-income earners were nevertheless deteriorating. Furthermore, the provincial and municipal governments, a majority of which belonged to the oppositional PJ, used their control over the distribution of food boxes associated with the PAN to apply clientelistic practices, which both limited the government's ability to use the policy to consolidate lower-class support and led to a wave of criticism in the media that damaged its approval among the middle classes (Garay 2010, 45–46).

### **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

The conditions for the formation of an inclusive governing alliance and an expansion of social protections for low-income earners were ambivalent during the 1980s. On the one hand, the popular classes regained electoral power resources with the return to democracy. Once military repression was lifted, political parties, trade unions, and social movements were quick in reconstructing their associational power resources. The temporary discrediting of neoliberalism and a resurgence of Keynesian thinking widened the discursive power resources of progressive social forces that aimed at the extension of redistributive measures. Under these conditions, the center-left candidate of the UCR Party, Raúl Alfonsín, was able to win the 1983 presidential elections based on a progressive political platform. On the other hand, the dictatorship left a disastrous economic heritage, which consisted of not only a heavy fiscal burden but a significant deterioration of the country's industrial base and labor market as well, which significantly weakened the structural power resources of the labor movement. Furthermore, the killing of thousands of leftist activists had moved both the PJ party and the union movement to more conservative positions.

Under these circumstances, the government had difficulties in constructing a sufficiently strong inclusionary alliance that would have been able to sustain its ambitious social and economic policy agenda. While it was able to mobilize broad electoral support, the corporate organizations of both the working class and capital, as well as the oppositional PJ party, successfully used their power resources to block health care reform, the centerpiece of the government's progressive social policy agenda. The fierce resistance of trade union leaders to progressive health care reform had, however, less to do with a general rejection of redistributive social policies than with the fact that this reform would have significantly limited their control over the health insurance funds. Therefore, the Alfonsín government did not face similar resistance in other areas of social policy, where it was able to pursue a moderately progressive line. In the face of a deep financial crisis of the pension system and a rapid expansion of informality, the government implemented measures that successfully maintained relatively broad access to pension benefits also among low-income earners. Furthermore, it changed the mechanism of benefit calculation in a way that favored low-income groups. The government also introduced a massive social assistance program and an allowance for unemployed workers. Family allowance and social housing spending were maintained roughly at the levels inherited from the military government (DAGPyPS 2011). However, due to the profound deterioration of the labor market and the economy in general, these measures were not enough to halt the steady increase of inequality.

Figure 6: Evolution of income inequality measured with the Gini-coefficient, 1983–1989

The 1980s once more evidenced the powerful path dependency that characterizes the Argentine health insurance system and which makes reform a politically difficult task. At the same time, the profound crisis of the pension system and the de facto application of mechanisms for benefit calculation that differed considerably from the legally established rules seemingly eroded path dependency in the pension system. In the face of a breakdown of the system, farranging reform became more acceptable.

Sources: Gasparini and Cruces (2010, 5).


Table 22: Change in social protection for low-income earners, 1983–1989

## **4.7 Neoliberalism and Regressive Social Policy Reform, 1989–2001**

The multiple crises of the 1980s not only ended in a clear-cut electoral defeat of the UCR in the 1989 presidential elections but also resulted in an alteration of the distribution of power resources to the detriment of the popular classes and progressive political actors. On this basis, the victorious candidate of the PJ Party, Carlos Menem, was able to form a solid cross-class governing alliance, in which neoliberal actors including members of the conservative economic elite, the big employer associations, and the IFIs obtained a high level of influence. In contrast, popular class actors such as the trade union movement and more progressive parts of the PJ Party were incorporated in a subordinated manner and politically marginalized. The new government pursued a genuinely neoliberal policy approach that included the pursuit of a highly regressive agenda of social policy reforms. These reforms significantly eroded social protections for low-income earners.

#### **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

The government of Carlos Menem assumed power in a situation of profound economic crisis. In 1989 the Argentine GDP fell by 4.4%, and inflation skyrocketed to over 3000% with the consequence of a sudden and massive impoverishment of large parts of the population (Rapoport 2013, Table 5; Birle 1995, 277). The government confronted these economic problems with a neoliberal agenda of privatizations, fiscal austerity, labor market deregulation, and the removal of trade and financial market regulations. However, only after the national currency was pegged to the US dollar in a fixed exchange rate in April 1991 was inflation brought under control and economic growth returned (Lo Vuolo 1998b, 189–190; Rapoport 2013, Table 5).


Table 23: Economic and social context, 1989–2001

Sources: Rapoport (2013) for GDP growth and inflation; Timmer, de Vries, and de Vries (2014) for industry share of GDP and share of industrial employment; INDEC Database for unemployment; Jiménez (2013, 59) for informality 1989–1990; and CEDLAS and World Bank (2018) for informality 1991–2001.

Nevertheless, in the medium and long run, these policies had severe negative effects. The pegging of the peso to the US dollar on a one-to-one basis led to a strong overvaluation of the local currency, which diminished the competitiveness of national manufacturing industries. In order to compensate for this, companies tried to reduce wages and imported machinery to replace labor. The high value of the peso furthermore favored the importation of consumer goods and hence contributed to a negative foreign trade balance and a high dependence on foreign capital inflows. In the face of a largely deregulated financial market, Argentina's economy became extremely vulnerable to the volatility of international capital flows. In this context, the Mexican Tequila Crisis of the mid-1990s had profound and direct consequences for the Argentine economy, which experienced a nearly 3% decline of GDP and an increase of the unemployment rate to over 18% (INDEC 2003; Rapoport 2013, Table 5). Although financial support from the IMF helped to temporarily stabilize the economy again, the next downturn came in 1998. By 2001, the Argentinian crisis had finally become so severe that the one-to-one exchange rate was no longer sustainable (Beccaria, Groisman and Maurizio 2009; Damill and Frenkel 2006).

The neoliberal reform agenda also had profound transformative effects on the structure of the economy and the labor market. Trade liberalization and an overvalued currency triggered a process of de-industrialization. Industry's share of GDP decreased from 35% in 1989 to 30% in 2001 and the share of industrial employment fell from 26% to 19% (Timmer, de Vries, and de Vries 2014). These developments went along with an increase of the unemployment rate from 7.6% in 1989 to 17.4% in 2001 (INDEC 2003). Informality among salaried workers increased from 29.0% in 1989 to 38.7% in 2001 (Jiménez 2013, 59; CEDLAS and World Bank 2018). These trends towards de-industrialization and labor market deterioration led to an erosion of labor's structural power resources, while at the same time the economic model's high dependency on capital inflows and foreign currency credits significantly increased the structural power resources of big capital and the IFIs.

Despite these adverse long-term effects, the initial success of these policies in getting inflation under control and stabilizing the economy in conjunction with the promotion of neoliberal thought by a variety of employer-financed think tanks, the IFIs, and a global shift to neoliberalism contributed to a broadbased agreement with neoliberal beliefs during the early 1990s.<sup>126</sup> At the same time, the hyperinflation crisis of the late 1980s led to an increasingly widespread perception that the ISI-Paradigm and its interventionist policies had failed and were no longer adequate to resolve the economic challenges of the time (Basualdo 2001, 51; Goldin 1997, 124; Svampa 2005, 31, 112–115). Data from the World Values Survey shows that in 1991 a stunning 65% of the respondents agreed that larger income inequality was needed in order to provide incentives, while only 21% stated that incomes should be made more equal.<sup>127</sup> Also an ample majority agreed that private ownership of business should be

<sup>126</sup> With the global rise of neoliberalism, the World Bank and the Inter-American Development Bank (IADB) increasingly replaced the less neoliberal ILO as key agents of external pension policy advice in Latin America (Kay 2000, 7).

<sup>127</sup> The question of the World Values Survey was: "Now I'd like you to tell me your views on various issues. How would you place your views on this scale? 1 means you agree completely with the statement on the left; 10 means you agree completely with the statement on the right; and if your views fall somewhere in between, you can choose any number in between. Sentences: Incomes should be made more equal vs We need larger income differences as incentives." For the calculation of the numbers, the answers 5 and 6 have been considered to be expressions of neutrality, while the answers 1–4 and 7–10 have been considered as inclination to one of the sentences (Bossert 2011).

increased to the detriment of state ownership and that individual responsibility should be increased to the detriment of public welfare provision (Inglehart et al. 2014: *World Values Survey Wave 2*). This pervasive dominance of neoliberal beliefs in the broader Argentine public weakened the discursive power resources of the popular classes and provided the political right with the ability to mobilize broad-based public support for regressive social policy reform, which was widely perceived as necessary and ultimately beneficiary for the society as a whole and even for the poorest sectors of it. Only when the dramatic social consequences of neoliberalism became increasingly visible and progressive forces managed to reorganize and to effectively articulate their criticism of the economic model did this predominance of neoliberal thinking began to erode again (Inglehart et al. 2014: *World Values Survey Waves 3 and 4*).

The prolonged crisis of the 1980s and the profound social, economic and political changes of the 1990s also coincided with important transformations of the main collective actors in Argentine welfare politics. During the late 1980s and early 1990s all major parties adopted neoliberal programs. Against the background of the failure of the heterodox economic policy package denominated Plan Austral, the UCR adopted an increasingly orthodox approach from 1987 on (Basualdo 2011, 69). In the internal elections of 1988, the conservative wing of the UCR won and Eduardo Angeloz, who proposed a consistently neoliberal electoral platform, became the presidential candidate (Birle 1995, 369). At the same time, the genuinely neoliberal party UCeDé, which had close ties to members of the Argentine economic elite, grew to become the third strongest party with nearly 10% of the vote in the 1989 elections (Birle 1995, 276; Cavarozzi 2006, 201).

After winning the 1989 presidential elections, Carlos Menem and his allies also transformed the PJ from a state interventionist party, whose policies largely constituted a populist version of the ISI-Paradigm, to a neoliberal party that pursued deregulation and free-market policies. Not surprisingly, this dramatic change of policy aroused resistance from more progressive parts and the trade union wing of the party. However, Menem and the neoliberal faction of the party were quite effective in co-opting and disciplining those leaders within the party and the CGT who were willing to cooperate and in dividing and marginalizing those who were not (Levitsky 2003, 144–185). Within five years, the share of trade union deputies in the PJ bloc in the Chamber of Deputies declined from 20% to less than 5% and the share of unionists in the executive board of the party decreased from 25% to only 12.5% (Levitsky 2003, 134– 135). Several circumstances contributed to the fact that Menem and his allies could shift the party to a completely different political orientation without destroying it. First of all, his success in stabilizing the economy and getting inflation under control provided the PJ with considerable electoral success during the first half of the 1990s. Second, the PJ had always been ideologically heterogeneous and the party's program had frequently been adapted to a changing environment or new internal relations of power (Levitsky 2003, 27). Third, the weakness of formal decision-making structures and party bureaucracy favored in practice a concentration of power in the hands of public office holders and particularly the President. Fourth, mechanisms that were established during the first Peronist government, such as the necessity of official recognition of unions and the repartition of funds among trade union-controlled health insurance funds, served effectively to co-opt and discipline most trade union leaders.128 Fifth, the loss of support from more progressive Peronists was compensated for by absorbing voters and activists from the neoliberal UCeDé, as well as some of the rather conservative voters of the UCR (Birle 1995, 286). And finally, the PJ was able to use its dense network of local units to establish a patronage-based relationship with the lower popular classes that heavily relied on the repartition of state resources, such as public jobs and social assistance benefits (Levitsky 2003, 123–124). This way, the PJ was relatively successful in mobilizing a large portion of low-income earner votes, despite the government's pursuit of an agenda that was highly detrimental to this social stratum.

The neoliberal turn of the UCR and the PJ, however, paved the way for the emergence of a new center-left party. In 1993 eight deputies of the Peronist left, who opposed Menem's neoliberal course, decided to leave the PJ and form a new party called Frente Grande. One year later the Frente Grande joined forces with PAIS, another party that had split off from the PJ, and several smaller center-left and left groups to form the FREPASO. The new center-left party was highly successful in absorbing the more progressive parts of the middle-class electorate and making some inroads into the working-class and lower popular class electorate. In the 1995 presidential election the party had developed such that it won the second biggest share of votes (Cavarozzi 2006, 165).

With the neoliberal turn, employer associations became increasingly influential actors in social politics. Their influence was strengthened not only by the employer associations' increased structural and discursive power resources, but an increase in their associational power resources due to the overcoming of internal divisions and the development and strengthening of a variety of influential think tanks. (Basualdo 2011, 56; Svampa 2005, 112–115).<sup>129</sup>

<sup>128</sup> In Argentina, trade unions are not legally entitled to conduct many of their essential functions, such as collective wage bargaining, without the official recognition of the Ministry of Labor. Neither can they maintain a social health insurance without recognition (Goldin 1997, 139).

<sup>129</sup> Three very influential employer-financed think tanks were the Mediterranean Foundation (Fundación Mediterránea), the Center for Macroeconomic Studies (Centro de Estudios Macroeconomicos) and the Foundation for Latin American Economic Research (Fundación de Investigaciones Económicas Latinoamericanas). Two of the ministers of economy came from the first two think tanks. The last developed an extensive and much

For the trade union movement, the 1990s were going to constitute a difficult decade. De-industrialization and the rise of unemployment and informality not only reduced the structural power resources of labor but also affected its associational power resources. Employment declined drastically in some of the best-organized sectors of the economy and hence led to an erosion of many unions' membership and activist structures. The metal workers' union UOM was particularly hard hit. For decades it had constituted the strongest union in Argentina, reaching a peak membership of over 500,000 in the 1970s. By 2001, the membership had fallen to only 70,000 (Abal Medina 2012, 120). At the same time, however, sectors such as food processing, private oil, and road haulage prospered during the 1990s and the respective unions did so too (Etchemendy and Collier 2007, 366–367). In total, Adriana Marshall estimates that union density shrank moderately and accounted for an estimated 40% of wage earners in 2001 (Marshall 2005, 25). As important as the decline of membership was the fact that, in the face of high unemployment, informality and the deregulation of the labor market, the fear of job loss had strong disciplining effects on workers, which decimated the activist structures of many unions and undermined their capacity to mobilize for collective action (Abal Medina 2012, 113–114; Wyczykier and Barattin 2012, 59).

The neoliberal turn of the PJ and the social and economic transformation of the Argentine society during the 1990s triggered important processes of political change and realignment within the labor movement. The majority of the unions in the CGT decided to cooperate with the neoliberal government and to try to retain members through an extension of their services and engagement in business activities instead of a focus on collective action (Alonso 2007, 166– 167).<sup>130</sup> The social policy demands of the CGT were mainly oriented towards preserving the existing system of social protections, and in particular in defending the union-controlled social health insurance funds. The extension of social protections to low-income earners as well as informal or unemployed workers was hardly on the agenda of the CGT. Nevertheless, an internal dispute over a canceled general strike in December 1993 led to the formation of a more combative faction within the CGT called Movimiento de los Trabajadores Argentinos (MTA), which was led by the increasingly powerful transport workers' unions. Although the MTA was not particularly concerned with the

cited proposal for social security reform that largely coincided with the agenda pursued by the government (Basualdo 2011, 56; CEA-FIEL 1995; Svampa 2005, 112–115)

<sup>130</sup> Most of the Argentinian unions had already developed a broad range of services since the first Peronist government, which among others often included health insurance, hospitals, hotels, sport facilities, libraries, pharmacies, and so on (Abal Medina 2012, 110). During the 1990s some unions took advantage of the privatization of state enterprises and the pension system and extended their activities to the realm of business by offering pension funds and credit cards or buying shares of privatized companies (McGuire 1997, 237; Murrillo 1997, 433).

extension of the system of social protections to low-income earners, it actively opposed the neoliberal government and cooperated frequently with more progressive political forces and unemployed workers' organizations (Interview Ardura 2014; McGuire 1997, 232–233).

A more radical approach to oppose neoliberalism was pursued by a handful of white-collar workers' unions, including the state workers' union (ATE) and the teachers' union confederation (CTERA), which left the CGT in 1992 and formed an independent confederation called Argentine Workers Congress (CTA) (Interview Fuentes 2014; Levitsky 2003, 137). The political approach of the CTA was quite different from that of the CGT. Not only did the CTA oppose the government in a combative manner, it also organized more democratically and allowed direct affiliation, so that informal and unemployed workers were able to join. In contrast to the CGT, the CTA actively demanded the universalization of social policies, which implied their extension to lowincome groups (Interview Fuentes 2014; Interview Larisgoitia 2014). On the level of party politics, many of the CTA leaders were involved in the foundation of the center-left Frente Grande, which later became part of FREPASO (Interview Fuentes 2014). The fact that the ATE and the CTERA, the largest member unions of the CTA, did not directly control their own health insurances and hence possessed greater independence from the state and weaker interest in the maintenance of the status quo than other unions, contributed to the CTA's combative nature and progressive orientation (Larisgoitia 2014). Under the motto "The neighborhood is the new factory" the CTA actively reached out to incorporate and support unemployed workers' organizations, peasants' movements, indigenous communities, migrants' organizations, poor neighborhoods and tenants' associations (Adamovsky 2012, 419; Caggiano 2011, 4–9). Furthermore, women's collectives and gender departments were formed within the CTA to fight gender inequalities (Interview Díaz 2014; Interview Ocar and Martínez 2014). Once part of the CTA, these organizations were able to participate in the confederation's decision-making structures and to influence its political agenda (Interview Larisgoitia 2014). By this means the CTA developed a strong commitment to broad-based redistributive social policies which explicitly included the extension of social protections to low-income earners. While the CGT membership gradually decreased during the 1990s, the new strategies allowed the CTA to grow throughout the decade and to reach a membership of nearly 1.5 million during the early 2000s (Interview Fuentes 2014; Interview Lozano 2014).

Furthermore, from the mid-1990s on, several new and increasingly powerful lower popular class movements emerged. Not surprisingly, one of the decisive impulses came from the CTA. In 1996, the confederation began to organize national unemployed workers' congresses and intensified its cooperation with neighborhood associations and cooperatives in poor areas. This contributed to the foundation of the Federación de Tierra y Vivienda (FTV) in 1998, which soon after became the largest unemployed and informal workers' movement in Argentina (Adamovsky 2012, 419; Svampa and Pereyra 2003, 40–46).<sup>131</sup> In the northern province of Jujuy a second important unemployed workers' movement emerged within the CTA, the Tupac Amaru (Tavano 2015, 231–232). Another impulse came from the small Maoist Partido Revolucionario Comunista (PCR). In 1994, in the midst of strike activities of municipal workers in the province of Jujuy, the PCR founded the Corriente Clasista y Combativa (CCC) as a network for combative unionists. In the following years, the CCC decided to broaden its scope of organization and developed branches for unemployed and retired workers as well as for indigenous movements (Natalucci 2010, 104–105; Interview Peñailillo 2014; Svampa and Pereyra 2003, 22). After its foundation in 1998, the unemployed workers branch rapidly developed into the by far biggest and most important branch of the CCC (Interview Ardura 2014). The significant success of these organizations in organizing unemployed and informal workers and in developing effective forms of collective action was also based on the enormous frustration and anger among large parts of the lower popular classes in the face of a dramatic deterioration of their social and economic situation. Beginning in 1996, Argentina experienced a series of spontaneous riots. What sometimes started as protests against the destruction of formal jobs quickly converted into week-long self-organized roadblocks in which marginalized sectors of the population not only participated but increasingly played a crucial role. The government reacted to these riots by conceding social assistance measures as a means of pacifying the situation (Adamovsky 2012, 422; Svampa and Pereyra 2003, 101– 149). These experiences became an important impetus and inspiration for the efforts of the FTV and the CCC. Both organizations took up the method of roadblocks and the CCC even adapted its decision-making structure to resemble the assembly-based approach that emerged during the above-mentioned protests (Interview Ardura 2014; Interview D'Elía 2014). In addition to the FTV and CCC, a large range of smaller movements emerged either as independent organizations or with the support of leftist parties, left-wing Peronist groups, or progressive sections of the church (Adamovsky 2012, 425–426; Golbert 2004, 18). By the early 2000s, the FTV had grown to approximately 90,000 and the CCC to over 70,000 members (Interview Ardura 2014; Interview D'Elía 2014). The overall membership of all unemployed workers' organizations was estimated at 300,000 at that time, which meant that the lower popular classes had for the first time achieved considerable associational

<sup>131</sup> Most movements such as the FTV and the CCC organize a variety of different groups belonging to the lower popular classes, such as peasants, cooperative workers, informal workers, neighborhood activists, indigenous activists, and so on. Unemployed workers and their demands, however, came to constitute their biggest mass base during the late 1990s. In the following I will therefore refer to these movements as unemployed workers' movements.

power resources (Alcañiz and Scheier 2007, 160). The majority of the baselevel activists were women, and a high percentage had a migratory background (Adamovsky 2012, 428).


Table 24: Popular class power at the beginning of the period, 1989–1991

At the same time that women played an important role in the emergence and growth of unemployed workers' movements, the women's movement gradually became broader. A majority of activists belonged to the popular classes and had links to progressive political organizations, such as the CTA, the unemployed workers' movements, left parties, and student organizations. From 1986 on, the Encuentro Nacional de Mujeres was established as a yearly gathering point for the women's movement. While in 1986 approximately 600 activists participated, the number rose to over 12,000 by 2001 (Adamovsky 2012, 431–432). Furthermore, from 1991 on a new law prescribed a minimum of 30% female candidates for general election party lists, which contributed to a significant increase of the women's movement's political influence (Cavarozzi 2006, 203). Although collective actions focused most often on gendered violence and the right to abortion, the large majority of the movement was also in favor of better social protection for low-income earners, as such policies were considered to contribute to a diminution of both gender and class inequalities (Interview Díaz 2014; Interview Ocar and Martínez 2014; Interview Rodriguez Enriquez 2014).

All in all, the 1990s can be described as a decade of profound social, economic, and political transformation. The early years of the decade were characterized by a distribution of power resources that was highly detrimental to progressive political forces and the popular classes. Neoliberalism had risen to be the dominant worldview not only within the employer sector but also within public opinion and the leadership of the two big parties, which were dominated by their respective right wings. Furthermore, the structural and associational power resources of the popular classes were weakened by the profound economic crisis. The lack of electoral alternatives and the effective patronagebased local structures of the PJ in poor neighborhoods additionally limited the degree to which the popular classes were able to use their electoral power resources in favor of social policy change during most of the 1990s. Only towards the end of the 1990s, after years of social deterioration and a painstaking process of realignments of left and popular class organizations, did resistance begin to strengthen, new movements emerge, and neoliberal thinking lose its dominant position.

## **The Formation of a Neoliberal Cross-Class Governing Alliance and the Initiation of Regressive Social Policy Reform, 1989–1995**

The second half of the 1980s was characterized by a deepening of economic problems and social conflict and the erosion of the support basis of the governing alliance. In 1986 alone, the government was confronted with four general strikes (Cavarozzi 2006, 199). In 1987, three more followed and the midterm parliamentary elections led to the government's loss of its majority in the Chambers od Deputies (Cavarozzi 2006, 199–200; Portantiero 1994, 306– 307). The trials against human rights violators of the dictatorship furthermore aroused three armed rebellions in 1987 and 1988, which added another dimension to an atmosphere already characterized by conflict (McGuire 1997, 186). In the economic realm, the inability of the heterodox Plan Austral to enduringly stabilize the economic situation led to a gradual shift within the governing alliance towards more neoliberal oriented policies and an increase of the influence of neoliberal technocrats and the center-right wing of the UCR (Birle 1995, 226). Initially, the neoliberal turn of the government mobilized some active support from the biggest employer associations. This support, however, began to erode as it soon became evident that the political circumstances and resistance within the party would impede the consequent application of a profound neoliberal reform program (Birle 1995, 247–249). By 1988, the economic situation had deteriorated to such an extent that the government had to declare its temporary inability to service foreign debts. The worst drought in 120 years contributed to a further escalation of the economic crisis and initiated a period of hyperinflation in 1989, which ended in a massive impoverishment of a significant proportion of the population and incidents of looting in several cities (Basualdo 2011, 61; Birle 1995, 262–263).

Not surprisingly, economic policy was the central issue of the 1989 presidential elections. Eduardo Angeloz of the UCR and Álvaro Alsogaray of the UCeDé built their electoral campaigns around genuinely neoliberal reform programs. Menem of the PJ based his campaign on a more eclectic platform, promising on the one hand massive wage hikes while on the other hand announcing privatizations and spending cuts. The elections resulted in a clear-cut victory for Menem. In the face of a dramatic deepening of the social and economic crisis, President Alfonsín decided to hand over power to Menem on the 8 th of July 1989, five months before the transfer of office was scheduled (Birle 1995, 276–77; Lupu 2014, 578; Madrid 2003, 106).

Once president, Menem shifted immediately to a genuinely neoliberal agenda, which consisted of the privatization of public enterprises, the removal of trade and financial market regulations, fiscal austerity, the deregulation of the labor market and the convertibility of the Argentinian peso with the dollar on a one to one basis (Lo Vuolo 1998b, 189–190). While this appeared like a sudden turn to some Peronist activists, others had seen it coming due to the alliances Menem had forged to win the primary elections (Interview Trotta 2014, Interview Vinocur 2014). In fact, Menem's bid for becoming the PJ Party's presidential candidate was primarily supported and financed by the Union Roundtable Menem for President (URMP), which united right-wing Peronist union leaders, many of whom had collaborated with the military dictatorship and shared the neoliberal analysis of the Argentinian economic situation. In contrast, his rival Antonio Cafiero was supported mainly by the Group of 25, which united union leaders who were mostly inclined to the left of the political spectrum (Godio 2000, 1180–1187; McGuire 1997, 204–213). Once Menem had won the internal election and become the presidential candidate of the PJ, the URMP members Jorge Triaca, leader of the plastic workers' union, and Armando Cavalieri, leader of the retail clerks' union, used their connections to the business community and Argentine high society to raise significant campaign funds from business magnates like Carlos Bulgheroni and Amalia Lacroze de Fortabat (McGuire 1997, 214). According to Godio (2000, 1184), Menem also had direct consultations from a range of big companies such as Bunge y Born, R. Handley, and Citibank during his campaign.

Immediately after he won the 1989 presidential elections, Menem had a series of meetings on economic policy with top managers of the biggest corporations and representatives of all the important employer associations. All the big employer associations thereafter declared their support for Menem's neoliberal policy program. The CEA, a circle of many of the most powerful members of the Argentine economic elite which had actively cooperated with the last military government, reappeared and financed a series of newspaper ads that promoted the government's economic platform (Birle 1995, 336–337). As minister of economy, Menem appointed a former manager of the Bunge y Born corporation, who according to the conservative union leader Luis Barrionuevo (quoted in Basualdo 2001, 58), had donated 700,000 US dollars to Menem's electoral campaign. Although the minister of economy was changed several times during the first years of the government, there was continuity in that all the ministers of economy were representatives or close allies of the neoliberal oriented economic elite and that the ministry obtained a de facto dominant role in the decision-making process within the cabinet (Basualdo 2011; Lloyd-Sherlock 2006, 358; Madrid 2003, 108). In a very short time, the new governing alliance had deliberately incorporated the big corporations, employer associations, the IFIs, and the US government into the core of its base of support and conceded to these actors strong political influence (Birle 1995, 279; Lupu 2014, 579). Menem also established a parliamentary alliance with the neoliberal UCeDé Party, which had won nearly 10% of the votes and 15 seats in the Chamber of Deputies in the 1989 parliamentary elections. The party's leader, Álvaro Alsogaray, was appointed presidential consultant (Birle 1995, 276–286). As Menem veered to the right also on moral issues, such as abortion, he was also able to establish benign relations to the Catholic church (Blofield 2006, 122).

In turn, the popular classes were incorporated in a subordinated manner with little influence on the agenda and limited power to block policies that were detrimental to their interests. Compared to Alfonsín, Meném had the advantage that most trade unions were willing to cooperate and refrained from confronting the government with strike activity due to their historical linkages to the Peronist movement. Furthermore, Menem was skillful in employing divide and rule and carrot and stick tactics to co-opt more cooperative union leaders and to marginalize those who decided to resist. Menem's main allies in disciplining the union movement were several right-wing union leaders that had already participated in the URMP and were now assigned to positions that gave them ample leverage to punish or reward other union leaders. Jorge Triaca of the plastic workers' union became minister of labor and Luis Barrionuevo of the hotel and restaurant workers' union was put in charge of ANSSAL,<sup>132</sup> an entity with ample leeway for redistributing resources between union-controlled social health insurances. The exemplary case of the metal workers' union shows quite clearly how these entities were used to co-opt union leaders and how effective they were in doing so. Originally, the long-standing leader of the metal workers' union, Lorenzo Miguel, was one of the union leaders that in October 1989 had founded the CGT-Azopardo, which united those unions of the CGT that disagreed with Menem's neoliberal agenda. However, when the CGT-Azopardo was about to declare a general strike in March 1990, Miguel suddenly opposed the initiative after the Ministry of Labor had assigned his trade union the representation of the workers of two car parts factories rather than giving it to the auto workers' union. A few months later, he even withdrew completely from the CGT-Azopardo after the ANSSAL had agreed to absorb 25 million US dollars of debt of the metal workers' union's health insurance. Another example is that in December 1991 several trade union representatives in the chamber of deputies approved the labor market deregulation

<sup>132</sup> The Administración Nacional de Seguro de la Salud (ANSSAL) replaced the INOS during the late 1980s.

law in exchange for the assumption of debts accumulated by the trade unions' social health insurance funds (Etchemendy and Palermo 1998, 567; Goldin 1997, 116–127; McGuire 1997, 220–229). By the end of 1991, the oppositional CGT-Azopardo had lost most of its supporting unions and hence the CGT was de facto reunited (Godio 2000, 1196). Combined with a deep-rooted tradition of trade union support for Peronist governments, an overall weakening of the unions' power resources, and the fact that many of the more rightist union leaders were quite sympathetic to the neoliberal worldview by the early 1990s, the above-described co-optation tactics were highly effective. Most of the unions finally aligned with the government and limited themselves to minor acts of resistance in the framework of an overall cooperative relation with the government. At the same time, Menem and his allies used their influence to dramatically reduce the overall number of union representatives within the party and the Chamber of Deputies (Levitsky 2003, 134–135).

In a similar vein, progressive party internal forces were quickly marginalized and by 1993 the mandate of the eight most resistant deputies had run out (Birle 1995, 284). Although these deputies then founded a center-left party called Frente Grande, they possessed little power in the parliament until the 1995 elections. Taken together, these developments meant that even the limited capacity of representatives of the unions or the more progressive wing of the PJ to block neoliberal economic and social policy reform initiatives in the Congress was gradually vanishing during the first years of the 1990s.

The political success in quieting resistance from within the popular classbased Peronist movement was furthermore facilitated by the fact that the introduction of a fixed exchange rate with the US dollar led to a short-term recuperation of real wages. According to Beccaria, Groisman, and Maurizio (2009, 28), real wages increased by nearly 28% between 1989 and 1993. Although they were still 13% lower than in 1987, the partial recuperation represented a significant improvement in the economic situation of the popular classes compared to the period of hyperinflation. In combination with the increasing dominance of neoliberal thinking in the public discourse, this context contributed to the widespread perception that neoliberal policies were the only possible response to the economic and social challenges of the time.

Another factor that facilitated the high velocity with which the Menem government implemented the neoliberal turn was the increasing concentration of power in the hands of very few high-ranking officeholders, especially the president, both within the PJ party and in the wider institutional setup. Within the PJ, Levitsky (2003, 123–124) observes a consolidation of machine politics during the 1990s, "which may be defined as an informal pattern of political organization in which state resources, particularly government jobs, are the primary currency of political exchange between higher- and lower-level party actors. (…) Due to their access to state resources, public office holders enjoyed a great advantage in the patronage game." Hence, very much as ANSSAL served the Menem government to discipline and co-opt union leaders by providing or denying financial support to their social health insurance funds, a variety of further state resources, such as financial and technical support for local social assistance, infrastructure, or housing programs, served to effectively bring governors, mayors, party leaders and activists in line. And as governors play a decisive role in composing the lists of candidates for the parliamentary elections, these mechanisms also provided the president with strong leverage vis-a-vis the deputies and senators of the party. Some of these mechanisms even worked to enforce cooperation of office holders that belonged to other parties. The opportunities to use patronage further increased by means of several reforms enacted during the 1990s that augmented the financial dependence of provinces and municipalities on the central government. In 2000, provinces and municipalities accounted for 47% of the public expenditure, but only for 20% of public revenue (Cetrángolo and Gatto 2002, 6).

Looking at the formal institutional setup, Menem additionally implemented measures that strengthened the power of the president vis-a-vis the parliament. When Menem and Alfonsín agreed that Menem would assume the presidency five months early, they also agreed that the UCR would vote in favor of government bills until the newly elected deputies took office in five months. Thus, during the first months, Menem had an extraordinarily broad parliamentary majority, which he used to pass the state reform law and the economic emergency law (Birle 1995, 284). These two laws provided Menem with the ability to privatize state enterprises and to implement economic policies by decree without requiring the approval of the parliament (Cavarozzi 2006, 113). Furthermore, Menem increased the supreme court from five to nine members, filled the new seats with political allies, and thereby eliminated this actor as a veto player. This cleared the way for Menem to legislate with socalled emergency decrees, which were often of questionable constitutionality. With the 1994 constitutional reform, the relatively broad use of emergency decrees was finally legalized (Lo Vuolo 1998b, 210–211).<sup>133</sup>

All in all, and in contrast to Alfonsín a few years earlier, Menem was able to form a relatively broad and solid cross-class governing alliance, in which neoliberal and conservative actors and the economic elite clearly dominated the agenda-setting process. Within the popular classes, new progressive actors were emerging, such as the CTA, MTA, and FREPASO, but until the mid-1990s these had little ability to block government policies. On this basis, the government was able to initiate a profound neoliberal restructuring of the Argentine state and economy.

<sup>133</sup> Between 1989 and 1999 Menem issued 545 emergency decrees (Ferreira Rubio and Goretti quoted in *La Nación,* April 13, 2008).


Table: 25: Composition of the new governing alliance, 1991\*

\* With the electoral victory of the UCR-FREPASO alliance in the 1999 elections the composition of the governing alliance changed significantly. However, the new alliance also had a regressive character. \*\* Over the years, the majority of more progressive forces left the governing alliance and joined the opposition.

Regarding social policy, pension privatization clearly constituted the first priority of the government. Not only was the system stricken by a profound financial crisis—it was also an area of key importance for a variety of actors within the governing alliance, especially banks and insurance companies, who were expecting enormous business opportunities (Interview Alonso 2014). One of the first measures of the government was to deal with the enormous debts that the system had accumulated with its beneficiaries by de facto deviating from the legally established replacement rate. By decree it was established that all debts of the pension system until March 1991 would be automatically converted into public bonds with a duration of ten years (Lo Vuolo 1998b, 217–218). The return to legally established replacement rates in combination with the freezing of the minimum pension at the relatively low value of 150 pesos led to an increasing differentiation of pension benefits, and hence to an increasingly unequal distribution of resources among pensioners to the detriment of the lower-income sectors. While in 1989 the minimum pension represented 86% of the value of the average pension, in 1995 it represented 50% and in 2001 only 42%. This increasing inequality was caused by both a recovery of pension benefits above the minimum and by a decline of the real value of the minimum pension due to the lack of any adjustment criteria. Between 1989 and 1995, the real value of the minimum pension fell by over 21%.134 At the same time, the government took measures to ameliorate the financial crisis of the system. In 1991, decree 2016/91 raised the contribution requirement from 15 to 20 years, which led to the exclusion of important parts of the popular classes from access to pensions, who due to a variety of reasons such as unemployment, informality, migration or child and elderly care were not able to accumulate sufficient years of contribution. In the same year, the government decided to cut the funding of the national housing fund (FONAVI) and instead direct the 5 percentage points employer contributions to the pension system (Lo Vuolo and Barbeito 1998a, 338). In consequence, expenditure on housing policies as a share of the GDP declined by nearly 40% between 1990 and 1995, despite the fact that about one third of Argentina's households were suffering from housing deficits (Belmartino 2005e, 252; DAGPSyPS 2011).

Traditionally, Argentine governments have appointed lawyers of the Ministry of Labor to head the Social Security Department that is in charge of the pension system. But in 1991, the neoliberal economist Walter Schulthess, a person who had the confidence of the minister of economy, was assigned to this office (Lo Vuolo 1998b, 215). Immediately, the Social Security Department began to work on proposals for the privatization of the pension system. In 1992, the government finally sent a bill to the Congress, which foresaw the replacement of the public pension system with that of a predominantly private system, the increase of the contribution requirements from 20 to 30 years, and an increase of the retirement age by five years. The bill was supported by the major employer associations, especially by those of the financial sector, yet was opposed by most of the trade unions and all oppositional parties. The CTA and a variety of pensioners' associations organized weekly protests in front of the Congress building. Due to the fact that in 1992 15% of the PJ bloc in the Chamber of Deputies were still representatives of trade unions, the bill could not win a majority (Alonso 1998, 601–613; MTEySS 2003, 16).

Potential investors declared that privatization via emergency decree would not give them sufficient confidence in the political sustainability of the reform. Hence, Menem was forced to negotiate with the CGT and the trade union leaders in the PJ bloc (Alonso 1998, 595–604). Initially, the CGT totally rejected privatization. But over the course of negotiations, the government made various concessions that led the CGT and most member unions to finally accept a

<sup>134</sup> All numbers have been calculated by the author on the basis of the following sources: MTEyFRH (2001) for the average pension benefit in 1989; Cárcamo Manna (1998, 15) for the minimum pension benefit in 1989; MTEySS (2010) for the average and minimum pension benefits in 1995 and in 2001; and INDEC (2008) for inflation adjustment.

modified reform proposal. The most important concession was that workers would be able to choose between a private capitalization scheme and a reformed public pay-as-you-go scheme. Both schemes contained a public component called Universal Basic Benefit.<sup>135</sup> Over the course of the negotiations, the trade unions won the concession that the value of this component be raised from 20 to 27.5% of the Argentinean average wage, which was particularly beneficial to lower income pensioners. Furthermore, the income-related component of the public scheme was strengthened and the establishment of a pension fund in the hands of a public bank, the Banco Nación, was supposed to create a safe option within the capitalization scheme. Finally, as many unions were expanding their business activities during the 1990s, the right to establish their own pension funds was seen with sympathy by a range of more conservative union leaders. Eventually, the modified reform bill passed Congress in October 1993 (Alonso 1998; Kay 2000, 12; Lo Vuolo 1998b, 222).

Although the trade unions were able to negotiate some important concessions, these hardly ameliorated the regressive character of the reform and the dramatic exclusionary effects it had on low-income earners. In theory, workers were able to choose between a public and a private scheme. However, those who did not actively choose within a certain period were automatically transferred to the private scheme, which led in practice to the transfer of most contributors (Arza 2009, 9). Furthermore, the reform meant that the legally-defined 70%–82% replacement rate of the public system was abolished in favor of new, more differentiated and generally much lower replacement rates (Kay 2000; Lo Vuolo 1998b). The most regressive component of the reform was the rise of the eligibility criteria first to 20 and then to 30 years of contributions, which de facto led to the exclusion of the vast majority of low-income earners from pension benefits. In combination with a dramatic deterioration of the labor market, this led to a decline of pension coverage among the lowest income quintile of persons in retirement age from 73% to 50% between 1991 and 2001 (Rofman and Oliveri 2011, 50). The decline in coverage, in turn, contributed to a growing need of elderly persons too seek employment, which further added to the already high unemployment rates of the 1990s (Alós et al. 2008, 34). In addition to these problematic social effects, the privatization of the system constituted an enormous financial burden for the state, because any contributor who switched to the private system was in consequence providing nothing towards the financing of the current obligations of the public system (Rofman 2006, 94; Kay 2000, 7). This already challenging fiscal deficit of the system was further exacerbated by agreements between the national government and provinces about a differentiated reduction of the employer contributions (Lo Vuolo 1998b, 221). According to the Ministry of Labor, these added

<sup>135</sup> Despite its name, the benefit was by no means universal considering that 30 years of contributions would be required to access it. Indeed, due to the tough access criteria, most low-income earners had hardly any chance to receive it upon retirement age.

up to an average national reduction of employer contributions of slightly over one third in 1996 (*La Nación*, November 6, 1996).

Health care reform was the second big area of neoliberal social policy reform for the government. Already during the 1980s, many of the actors that constituted the core of Menem's governing alliance, such as the major employer associations, the World Bank and the neoliberal UCeDé Party, promoted privatization and decentralization in the health sector. Although the specific interests and regulatory demands of different employer sectors varied, they agreed on the direction of change towards more market, more private business opportunities, and less influence from the trade unions (Alonso 2007, 171–173; Belmartino 2005a, 189; 2005c, 294).

In the public health sector, which provided health care coverage for most low-income earners, the government pursued three main aims: decentralization, cost-containment, and the introduction of market mechanisms. In July 1992, Decree 1269 transferred public hospitals from national to provincial responsibility without an adequate transfer of fiscal resources. As a result, the public health sector not only faced fiscal challenges as a whole, but regional inequalities were also further exacerbated to the detriment of poorer provinces, both of which caused a deterioration of health coverage for low-income groups (Cetrángolo and Gatto 2002, 18–19; Lo Vuolo 1998b, 231). In 1993, with the technical and financial support of the World Bank, the government furthermore increased the autonomy of hospital managers and tried to introduce competition and market mechanisms in the sector (Lloyd-Sherlock 2005, 1898).

The main reform project of the government, however, was to introduce market mechanisms in the social insurance sector and to open it to private insurers, which was repeatedly demanded by different employer associations during the early 1990s. In April 1991, the first convention of CONFECLISA,<sup>136</sup> the main employers' association of the private health sector, demanded better market access for private health insurers. In August of the same year, the employer-financed think tank FIEL presented an extensive reform proposal for the health sector that contained the complete opening of the social health insurance market to private competition.<sup>137</sup> The trade unions, not surprisingly, clearly repudiated these demands (Alonso 1996, 753–756).

The Menem government, as it did on most issues, decided to pursue the main demands of the employers' associations. In November 1991, Alberto José Mazza, at this time president of the employers' association of private health insurers, was appointed secretary of health and later ascended to minister of health (Alonso 2007, 171–173). Soon after, in January 1992, the government sent a bill to the Congress that proposed the gradual deregulation of the social health insurance sector with the final aim of full competition between trade

<sup>136</sup> Confederación Argentina de Clínicas, Sanatorios y Hospitales Privados.

<sup>137</sup> The presentation took place on a convention held under the auspices of the employers' association of Argentine private banks ADEBA.

union-controlled and private health insurance funds. Like the pension reform bill, the health insurance reform bill had been elaborated under the leadership of the Ministry of Economy (Alonso 1996, 757–758).

The proposal aroused heavy protest from trade unions, not only because they completely rejected its contents, but also because they had not been consulted in any way in its development. In this way, the bill contributed decisively to the reunification of the CGT in March 1992. In the face of this, the government decided to pull back and temporarily freeze the reform project, also because it did not want to endanger the even more important project of pension reform by over-stretching the union leaders' willingness to cooperate (Alonso 1996, 757–759; Lloyd-Sherlock 2005, 1896). In January 1993, the government finally announced that it was planning to reform the social health insurance sector by decree, however, making a significant concession to the trade unions by limiting the introduction of competition to the universe of social health insurance funds and excluding private insurers from access to the market. Although the majority of trade union leaders also rejected this proposal, in contrast to the bill of 1992 it created divisions within the CGT. Some of the more conservative union leaders indeed perceived the proposal as an opportunity to strengthen their own health insurance funds by attracting affiliates of other funds. Against this background, the government decided to issue the decree swiftly and to establish a short period during which the details were supposed to be negotiated with the trade unions. However, it soon turned out, that the implementation required the collaboration of the trade unions not only for political but also for technical reasons, given that the trade unions were administering most of social health insurance system. In practice, this led to the implementation of the decree being put on hold (Alonso 1996, 760–762).

Another aim of the government in social health insurance reform was to reduce costs for employers. Therefore, once the pension reform law had successfully passed Congress with the support of most trade union deputies, the government announced a reduction of employer contributions to the social health insurances in December 1993. As the trade unions saw their funds in danger, the CGT immediately scheduled a general strike. The action was called off as soon as the government showed a disposition to negotiate and tax authorities announced that they might undertake an investigation of the unions' finances and the personal possessions of a range of trade union leaders, however. Within the CGT this decision was controversial, and finally led to the foundation of the more combative internal faction MTA (McGuire 1997, 232– 233). All in all, due to the unified resistance of the trade union movement and the fact that the government used the threat of health insurance reform to obtain concessions on other issues, health care reform advanced at a very slow pace during Menem's first government compared to other areas, such as labor market and pension reform.

Spending on family allowances as a share of GDP slightly recovered compared to the years of hyperinflation, but it remained at a generally low level during the 1990s (DAGPyPS 2011). The main change during Menem's first term was of an administrative nature and consisted of the merging of the hitherto three existing family allowance funds (Lo Vuolo 1998b, 234).

In 1990, the Menem government eliminated the unemployment allowance, which had been established under Alfonsín in 1985. In 1991, it was replaced with the creation of the so-called National Employment Fund. The fund was financed by 1.5% employer contributions, which had hitherto been used to finance the family allowance system. The central component of the fund was the introduction of an unemployment insurance with considerably stricter eligibility criteria than the former unemployment allowance. In combination with the deterioration of the labor market, this led to a dramatic decline in coverage. While the earlier unemployment allowance covered 17% of the unemployed in 1989, the coverage of the new unemployment insurance fell to only 6% in 1995. Supported by a 350 million US dollar credit from the IADB, the fund also financed a variety of short-lived and focalized temporary employment and qualification programs (Chebez and Salvia 2001, 6; Lo Vuolo 1998b, 206, 240–241). However, even taken together, in 1995 the unemployment insurance and the temporary work programs only covered less than 8.4% of the unemployed.<sup>138</sup>

In general, contrary to the official discourse and despite the massive rise of unemployment and poverty, the first government of Menem only increased expenditure on focalized programs by a meager 0.1% of the GDP (DAGPyPS 2011). Initially, Menem even suspended most of the programs of the prior administration, including the PAN, which distributed up to 1.4 million food boxes to poor households (Garay 2010, 46–47; Lo Vuolo 1998a, 67). Only in the Province of Buenos Aires, whose governor was the Peronist Duhalde, was a food distribution program called Plan Vida implemented which reached up to 400,000 households. Instead, the government established resource transfers in order to support provincial governments' maintenance of soup kitchens and other assistance measures,<sup>139</sup> introduced a new social assistance pension for mothers of seven or more children and significantly increased the overall number of non-contributory pensions from 136,000 in 1989 to 291,000 in 1995 (Bertranou and Grushka 2002, 41; Golbert 2010, 61).<sup>140</sup> Additionally, a variety

<sup>138</sup> Author's calculations based on Golbert (2004, 22) for the number of workfare programs, Velásquez Pinto (2005, 51) for the number of beneficiaries of the unemployment insurance and DGEyFPE (2003) for the total number of unemployed persons.

<sup>139</sup> These transfers were established through the Communitarian Social Policy Program (POSOCO).

<sup>140</sup> The strongest rise can be observed for those types of non-contributory pensions that have no clear eligibility criteria and therefore serve well for clientelistic purposes (Bertranou and Grushka 2002, 41).

of over 60 mostly small and short-lived social assistance policies were introduced, which, despite the financial and technical support of several IFIs, received very little resources (Lo Vuolo 1998b, 238). For Menem and his closest allies within the governing alliance these fragmented and discretionary social assistance programs nevertheless fulfilled an important role in maintaining the lower popular classes as one of the subordinate bases of support of the alliance. Many of these policies served to consolidate and expand the PJ's clientelistic<sup>141</sup> base-level networks, which were very effective in mobilizing the lower popular class vote and controlling resistance (Garay 2010, 48; Levitsky 2003). An unemployed organizations' leader described the role of these networks and the distribution of social assistance like this:

The Peronists are always present on the territory, with their local brokers, the social plans. They have thermometers and they control the thing … In poor neighborhoods here, you can find 6 or 7 [PJ] basic units and there is no single radical comite<sup>142</sup>. [other parties] don't have thermometers … The Peronists pick up the phone and say: 'there's trouble here, things are getting nasty'… and intervene. (quoted in Garay 2010, 48)

All in all, the first government of Carlos Menem initiated a far-ranging reform process with a profoundly negative impact on the social protection for lowincome earners. Its centerpiece was the privatization of the pension system in a way that de facto excluded most of the low-income earners from pension coverage and led to an increasingly unequal distribution of income within the population of pensioners. These regressive policies thereby reinforced the negative social consequences of the neoliberal restructuring of the economy and the labor market.

## **The Deepening of Regressive Social Policy Reform in the Context of Growing Resistance, 1995–1999**

Over the course of his first term as president, Menem was able to consolidate his governing alliance and even to increase the dominance of neoliberal actors within it. During the 1991 and 1993 legislative elections, the government was not only able to get rid of dissidents within the PJ bloc, but also to increase its number of seats in the Congress. Furthermore, in 1993, the government reached an agreement with the UCR party about a constitutional reform, which, among other things, allowed Menem to run for a second term in 1995 (Cavarozzi 2006, 204). Menem focused his electoral campaign on his successes in recovering economic growth and reducing inflation, which earned him a landslide victory with nearly 50% of the votes (Birle 1995, 279; Lupu 2014, 581).

<sup>141</sup> Political clientelism in this context is conceptualised as "a relationship based on political subordination in exchange for material rewards" (Fox 1994, 153).

<sup>142</sup> Radical comité refers to the base level units of the UCR party.

While the neoliberal agenda of the government had caused the defection of many progressive middle- and popular-class voters, it had won the votes of conservative upper- and middle-class voters, which implied the virtually complete absorption of the UCeDé's electoral base (De Riz 1996, 136). The FREPASO, the new center-left party, was able to fill the vacuum on the left of the political spectrum and emerged as the second strongest party, reaching nearly 30% of the votes in the presidential and over 20% in the legislative elections. The UCR was relegated to third place (Cavarozzi 2006, 206).

Hence, the 1995 elections expressed both ongoing broad public support for the governing alliance and growing disillusionment and opposition in the face of a dramatic rise in unemployment, informality, and poverty. Data from the World Values Survey clearly shows that, although still in the majority, the approval of key neoliberal arguments had significantly declined by the mid-1990s (Inglehart et al. 2014: *World Values Survey Wave 3*). The strong results of FREPASO as well as the emergence and strengthening of the combative union factions CTA and MTA furthermore evidenced that progressive actors were successfully reorganizing.

The mid-1990s were generally characterized by growing polarization, which found its expression in an increasing intensity of protest and social conflict. The protest movement reached a first peak in 1994 when a broad alliance initiated by the CTA organized the Federal March to protest the government's neoliberal policy. On the 6th of June, the march culminated with a massive demonstration in front of the presidential palace. In 1996 and 1997, protest reached another peak with the first joint national general strike of all trade union confederations, a massive "cacerolazo"<sup>143</sup> in the capital and the eruption of week-long rebellions and roadblocks in several cities of the interior (Adamovsky 2012, 419–422; Godio 2000, 1210–1211; Svampa and Pereyra 2003, 101– 149). A short time later, the first national unemployed workers' movements emerged (Svampa and Pereyra 2003, 53–70).

While the growing protest movements held the government's neoliberal policies responsible for the dramatic deterioration of the social situation, Menem blamed undocumented immigrants, non-wage labor costs, and labor market regulations for the rise of unemployment. In other words, the government argued that the solution was not to be found in an abandonment but in a deepening of the neoliberal reform path. This approach received the decided support of the employer associations and the IFIs (McGuire 1997, 222–223). Shortly after the 1995 elections, the intersectoral employers' association CEA and the neoliberal think tank FIEL published an extensive study, in which they proposed a wide-ranging neoliberal reform of the entire system of social protection. The proposed measures included free competition in the health insurance sector, further privatization and less redistribution in the pension system

<sup>143</sup> Cacerolazo refers to a form of protest, in which the participants gather on the street and beat cooking pots.

and the introduction of work requirements for social assistance benefits (CEA-FIEL 1995). Simultaneously, the IADB supported the government with a 450 million US dollar Sector Program in Support of Fiscal Adjustment and Social Reform which explicitly aimed at smoothing the political process by providing resources and consulting for overcoming resistance (Tittor 2012, 184).

The first social policy initiative during Menem's second term was to undertake a renewed attempt at health insurance reform. Under the leadership of the Ministry of Economy, the government and the World Bank developed a new reform proposal for the complete opening of the health insurance market. A few months later, the government agreed with the IMF that private insurers would become able to enter the market. In the face of a skyrocketing of unemployment to over 18%, the government furthermore took up a recommendation of the industrial employers' association UIA and reiterated its proposal to reduce employer contributions to the social health insurances. Both measures were, once more, repudiated by the trade unions and contributed to the decision of the CGT to announce a general strike for the 6th of September 1995. However, two weeks before the strike, the leaders of the CGT and the government reached an agreement and the strike was called off. The CGT finally accepted the introduction of competition and a reduction of employer contributions by one percentage point. In return, Menem promised that private insurers would remain excluded and agreed to the creation of a commission with the participation of the CGT that would elaborate a Programa Médico Obligatorio (PMO), which defined a catalog of treatments that any social health insurance would have to cover. Furthermore, the government conceded to the conversion of several social health insurance funds with tripartite administration into purely trade union-controlled social health insurance funds.<sup>144</sup> Finally, after nearly four years of back-and-forth negotiations, Decree 1141 broke up the monopoly that the different health insurance funds had in their respective occupational groups and established that from January 1997 workers would be able to choose their social health insurance providers (Alonso 1997, 169-173; Lloyd-Sherlock 2006, 358; Lo Vuolo 1998b, 232). One year later, these rules were complemented by Decree 504/98, which allowed social health insurance funds to decline applicants and to offer differential coverage to new affiliates based on the amount of their contributions (Lloyd-Sherlock 2005, 1900).

In practice, the reform had highly regressive effects and deteriorated the quality of coverage for low-income affiliates. High-income affiliates, however, benefited from the new rules, as they massively undermined the existing mechanisms of redistribution within occupational groups. Not surprisingly, the newspaper *El Clarín* (April 24, 2003) concluded that most workers who switched social health insurance between 1998 and 2003 were disproportionally young and high earning. In order to compete for these affluent affiliates

<sup>144</sup> This applied for example to the social health insurance of banking, glass and railway workers.

with differentiated insurance packages, many health insurances started to contract out to private insurers, which led to a gradual process of privatization (Lloyd-Sherlock 2006, 359). The regressive effects of the reform were additionally intensified by the fact that the increase in unemployment and informality led to a reduction of coverage from over 63% of the population in 1991 to under 45% in 2004.<sup>145</sup>

In the area of pension reform, the government focused on cost containment. In 1995 and 1997, Law 24.463 and Emergency Decree 833/97 first restricted and then eliminated the indexation of public pensions, which led to a steady decline of their values (Kay 2000, 12–13; Lo Vuolo 1998b, 222).

In 1996, a reform of the family allowance system introduced three different benefit levels that were inversely related to the income of the recipient (Lo Vuolo 1998b, 234). Although the new benefit formula slightly increased redistribution within the group of formal wage earners, the core problem of the system persisted. It continued to exclude the rising number of informal and unemployed workers, and therefore most poor households. According to a survey from 2001 only 31.2% of all minor Argentinians were covered by the system and only 4.7% of those living in extreme poverty (Salvador 2007, 24).

With the rise of unemployment, poverty, and social conflict from the mid-1990s on, social assistance policies became increasingly important for both the exercise of political control through the government and the development of associational power resources through unemployed workers' movements. While during Menem's first term, the number of non-contributory pensions was significantly increased, during his second term workfare programs became the centerpiece of social assistance policy (Bertranou and Grushka 2002, 41; Svampa and Pereyra 2003, 87–88). As recommended by the CEA and FIEL (1995), and in contrast to the non-contributory pensions, these programs consisted of hiring unemployed workers for temporary jobs that were considered of public interest. In exchange, the beneficiaries received a small monthly nonwage payment of typically no more than 200 pesos.<sup>146</sup> The biggest programs of this type were the national Work Program (Plan Trabajar) and the provincial Buenos Aires Neighbourhoods Program (Plan Barrios Bonaerenses) (Lo Vuolo 1998b, 241; Svampa and Pereyra 2003, 88–89). The extension of these programs was closely related to the massive eruption of lower popular class protests and roadblocks that became viral after the 1996 and 1997 rebellions against the privatization of the state oil company in the provinces of Neuquén

<sup>145</sup> Calculations of the author based on the following sources: INDEC (2010) national census data for 1991 overall health insurance coverage, MSN (2010) for 2004 social health insurance coverage and Lloyd-Sherlock (2006, 358) for an estimate of persons covered by private health insurance.

<sup>146</sup> Due to the convertibility regime, 200 pesos equaled 200 US dollars at that time. Hence, these non-wage payments were far below both the wage for regular employment and the poverty line for an average household.

and Salta. By that time, the breadth and the militancy of the protests had reached such a level that the national and regional governments decided to complement repression with the repartition of workfare benefits to the protestors (Adamowsky 2012, 423). In order to intervene as quickly as possible, the Ministry of Labor even created two additional small workfare programs which had highly flexible rules and therefore served perfectly for negotiations with protestors (Garay 2010, 60). The Buenos Aires Neighborhoods Program was implemented in July 1997 as a direct reaction to the spreading of roadblocks (Svampa and Pereyra 2003, 90). Nationwide, the average number of workfare beneficiaries increased from 63,000 in 1995 to 130,000 in 1997 (Golbert 2004, 22).

In addition to the pacifying function, the distribution of workfare benefits also served the government to maintain their ties to the lower popular classes, which suffered severely from the neoliberal transformation. Most benefits were still not handed out to protestors but distributed through the dense clientelistic networks of the PJ in poor neighborhoods (Svampa and Pereyra 2003, 87). In an analysis of over one thousand municipalities in the years 1998 and 1999, Marcelo Nazareno, Stokes, and Brusco (2006, 78–80) found that the quantity of locally distributed workfare benefits had, in fact, a significant influence on the electoral results.

Through the emergence of unemployed workers' organizations and their growing associational power resources after 1998, however, lower popular class actors were increasingly able to challenge these clientelistic structures and to enforce the receipt of workfare benefits despite their opposition to the government (Garay 2010, 48). In turn, the administration of workfare benefits and related projects, such as soup kitchens or small manufactures, provided the organizations with key resources with which to consolidate their membership and attract new affiliates (Golbert 2004, 21–22; Natalucci 2010, 96).

Against this background, expenditure for social assistance and unemployment protection increased from 0.97% of GDP in 1995 to 1.44% in 1999 (DAGPyPS 2011). Although this constituted an important increase, unemployment insurance and workfare programs together still covered only about 9.6% of the unemployed in 1999.147 Therefore, it hardly compensated for the dramatic rise in unemployment and poverty. Furthermore, the regressive reform of the much more resource-intensive health insurance sector meant that overall social policy reform during Menem's second term continued to deteriorate social protections for low-income earners and shifted resources increasingly to the better off.

<sup>147</sup> Author's calculations based on DGEyFPE (2003), MTEySS (2010) and Golbert (2004, 22).

## **The Formation of a Weak Neoliberal Governing Alliance and the Vanishing Viability of Regressive Social Policy Reform, 1999–2001**

The perpetuation of high unemployment and poverty rates and the increasing intensity of social conflict during the second half of the 1990s led to a gradual erosion of the government's public support. In order to take advantage of this situation, the main opposition parties, the FREPASO and the UCR, decided to form an electoral alliance in August 1997 called Alliance for Work, Justice and Education. In public statements, the alliance committed to a vaguely defined center-left electoral platform, criticized corruption and the social consequences of Menem's neoliberal policies and promised the improvement of infrastructure, working conditions, and education. On that basis, the Alliance was able to wrench a share of the popular class vote from the PJ and to win the 1997 legislative elections (Cavarozzi 2006, 121, 208–210; Godio 2000, 1149–1151; Salvia 2015, 112).

Within the alliance, however, existed enormous contradictions. The center-left profile was supported only by a part of the coalition, namely the FREPASO and the social democratic wing of the UCR. In fact, since the crisisridden presidency of Alfonsín, the UCR had been largely dominated by its conservative-neoliberal wing. This led to the joint 1998 primary elections being contested between the center-left FREPASO candidate Graciela Fernández Meijide and the conservative-neoliberal UCR candidate Fernando De la Rúa. Mainly due to the disproportionally stronger base-level structures and party membership of the UCR, De la Rúa won the primaries by a wide margin. As a result, the neoliberal wing of the UCR took the lead of the alliance (Cavarozzi 2006, 128; Godio 2000, 1149–1151).

After arduous internal disputes, the governing PJ Party named Eduardo Duhalde its presidential candidate. However, when economic crises erupted in Southeast Asia, Russia, and Brazil, economic growth collapsed and so did the main source of legitimacy of the government. Against this background, the October 1999 general elections ended in a sweeping victory for the opposition UCR-FREPASO alliance. Once in office, De la Rúa swiftly abandoned the party's original center-left electoral platform and shifted towards a radically neoliberal agenda, which immediately aroused heavy criticism from FREPASO deputies and cabinet members but was welcomed by IFIs and employer associations. As Menem had done ten years before, De la Rúa took advantage of the high concentration of power in the office of the president provided by a range of formal and informal institutional practices to marginalize internal critics and to circumvent parliamentary resistance through the issuance of emergency decrees. In practice, the agenda-setting of the new government was dominated by a small group of technocrats and neoliberal politicians surrounding the president and the minister of economy. Although FREPASO and social democratic UCR politicians were assigned to lead the Ministry of Social Development and the Ministry of Labor, these were provided with hardly any freedom of action in a framework of fiscal adjustment. The marginalization of the left wing within the alliance went so far that, soon after becoming president, De la Rúa even ceased to consult with the FREPASO on many important policies (Interview Isuani 2014b; Lupu 2014, 577–583; Interview Vinocur 2014).

Despite certain similarities to the neoliberal turn of President Menem ten years before, De la Rúa was unable to form any comparably strong governing alliance in support of his agenda. In consequence, a large part of the government's social policy initiatives failed to overcome internal and external resistance. This situation can be credited to various factors. First, Menem undertook the neoliberal turn in a historical situation in which neoliberal thinking dominated public opinion. This provided him with ample discursive power resources to successfully frame his agenda in a way that it appeared both necessary and desirable for most of the population. By 1999 this had radically changed, as data of the World Values Survey shows (Inglehart et al. 2014: *World Values Survey Wave 4)*. In complete contrast to the 1991 survey, a large majority rejected privatizations and rising income inequality and preferred more public welfare provision over individual responsibility. Second, due to the close link between the majority of the trade unions and the PJ, Menem had an enormous advantage in taming trade union resistance, while De la Rúa's government faced three general strikes only in its first year (Cavarozzi 2006, 211–213). Third, with the emergence of unemployed workers' movements, the lower popular classes had developed their own associational power resources that De la Rúa completely underestimated (Interview Berra 2003). Fourth, while Menem's employer support was reinforced by a strong unity among different employer sectors, over the years tensions had arisen as some sectors benefited more than others from the neoliberal transformation. Furthermore, by the early 2000s employer associations were increasingly divided regarding the question of whether to continue with the convertibility regime, dollarize the economy, or float the peso, so that De la Rúa's base of support among employers was much more complex and less reliable. Fifth, both the viability of the convertibility regime and the support of certain employer sectors depended on the ongoing privatization of state enterprises, but by 1999 most of them had already been sold (Basualdo 2001, 85–101). Sixth, disciplining deputies and cabinet members and creating a public image of unity turned out to be much more difficult for De la Rúa because the majority of his internal critics belonged to the FREPASO, and hence he lacked intra-party disciplining mechanisms.

Against this background, the president started his term faced by the rising of unprecedented levels of social and political conflict. Only one month after assuming office, in January 2000, the government sent a labor market reform bill to the Congress, which, if passed, would enact even more flexible work contracts and a significant weakening of the trade unions by decentralizing collective bargaining and limiting the validity of collective agreements. Two weeks later, the IMF linked an urgently needed credit of 7.4 billion US dollars to the condition that the labor market reform bill and further measures pass the parliament. However, the labor reform was repudiated by a majority in the parliament, including many deputies of the governing alliance. The trade union confederations, in turn, joined forces and called for a national general strike for late February. After several negotiations with more conservative union leaders and a range of deputies from both the coalition and the opposition, the bill surprisingly passed parliament in April 2000. At the same time, the opposition on the streets grew continuously and at the end of May the government was confronted with a massive demonstration against its neoliberal agenda, supported by a broad alliance of the MTA, the CTA, unemployed workers' movements, a part of the FREPASO, left parties, students' associations, human rights organizations and sectors of the Catholic church. Finally, it turned out that the bill's approval had been reached through systematic bribery. This caused a massive public scandal and the resignation in protest of the vice president and several other more progressive cabinet members. It also accelerated the withdrawal of FREPASO deputies from the governing alliance, which meant a further disintegration of its already weak support base (Cavarozzi 2006, 211–213; Interview Isuani 2014b; Salvia 2015, 120).

In the middle of this tumultuous period, the government decreed that the social health insurance market would be opened to competition from private insurers from January 2001 on. Furthermore, the decree abolished the right of insurance funds to reject applications of new affiliates and changed the formula for the guaranteed minimum revenue from 40 pesos per affiliate to 20 pesos per covered person, including family members. This renewed attempt to deregulate the health insurance system was again strongly rejected by the trade unions. Several union-controlled social health insurers took legal action against the decree and it was temporarily stopped in the courts. The decree received also very hesitant support from private health insurers, as most of these had already established profitable contracts with trade union-controlled social health insurers. In March 2001, the government finally gave in and suspended the decree (Alonso 2007, 161–162).

In the area of pension policies, the government's agenda mainly consisted of trying to comply with demands of the IFIs and the employer associations to reduce both the fiscal deficits of the system and employer contributions. Encouraged by the IMF, the government reduced public pensions and federal public sector salaries by 13% in July 2001, which aroused massive strikes and demonstrations by the trade unions. Due to agreements with the opposition PJ Party, the measure nevertheless passed the parliament (*La Nación*, July 31, 2001; Salvia 2015, 125). At the same time, Decree 814/01 and Law 25.453 replaced the fragmented arrangements for the reduction of employer contributions made during Menem's government with two uniform nationwide contribution rates. For the vast majority of companies, overall social security contributions were fixed at 21%, which meant that within ten years employer contributions had been reduced by an impressive 12 percentage points (Lo Vuolo and Barbeito 1998a, 337).<sup>148</sup> In an increasingly desperate social and economic situation, the government finally also reduced worker contributions to the private pension scheme from 11% to 5%, arguing that this would contribute to stimulating much-needed consumption (Arza 2009, 8).

When De la Rúa had assumed the presidency, unemployment had already stood at nearly 14% and approximately 25% of the population had lived under the poverty line. By late 2001, these numbers had increased to over 18% and 40% respectively (Cavarozzi 2006, 214; Gasparini and Cruces 2010, 5; INDEC 2003). Despite the dramatic impoverishment of major parts of society, De la Rúa initially assigned relatively low importance to social assistance and unemployment protection programs. As did the Menem administration in 1989, his government even systematically cut the social assistance programs of its predecessor, assuming that these benefited the clientelistic networks of the oppositional PJ party (Garay 2010, 63). Although the government simultaneously implemented similar new programs, such as the Employment Emergency Program, the total amount of workfare benefits declined significantly (Golbert 2004, 22).<sup>149</sup> Not surprisingly, this led to a massive increase and radicalization of protests and roadblocks undertaken by unemployed workers' organizations. The government's response was harsh repression, which rather than quieting the protests fueled cooperation between these movements. In October 2000, FTV and CCC undertook a joint roadblock of the important highway National Route 3 in the Matanza district near the city of Buenos Aires. The massive participation of over 3,000 activists for six days finally forced the government to give in and to negotiate with the leaders of these two unemployed workers' organizations the provision of 9,500 workfare benefits (Salvia 2015, 117– 118).<sup>150</sup> The successful protest and the control over a significant number of workfare benefits not only strengthened the FTV and the CCC, it also inspired other unemployed workers' organizations throughout Argentina to reinforce their activities. Without any clearly recognizable strategy, during the following year the government's responses to these protests oscillated between repression and concessions, while unemployed workers' organizations grew rapidly (Interview Berra 2014; Svampa and Pereyra 2003, 71–72). In the context of

<sup>148</sup> Only for big service sector companies was a higher contribution rate of 25% established. The numbers also include social health insurance contributions.

<sup>149</sup> The average number of persons covered by workfare benefits declined from 106,000 in 1999 to 86,000 in 2000 (Golbert 2004, 22).

<sup>150</sup> A certain share of the negotiated benefits was directly managed by the unemployed workers' movements, which enrolled unemployed activists and created small cooperatives aimed at covering diverse needs of the local community (Adamovsky 2012, 424).

this dynamic, the number of workfare benefits and social assistance expenditure slightly recovered during 2001 (DAGPyPS 2011; Golbert 2004, 22).

Two national assemblies of nearly all unemployed workers' organizations in July and September 2001 additionally strengthened their capacity for mobilization and coordinated action. Among other demands, the assemblies insisted that workfare programs must be universalized to cover all unemployed workers over the age of 16. They also agreed on a common protest plan based on the nationwide intensification of roadblocks and the initiation of an open-ended protest action towards the end of the year (Golbert 2004, 20; Svampa and Pereyra 2003, 209–210). Parallelly, in July 2001, the CTA founded the FRENAPO (Frente Nacional contra la Pobreza), a national alliance of progressive forces that committed to the eradication of poverty in Argentina through the universalization of social policies. Within a short period, a broad range of political and social leaders and organizations joined the alliance.<sup>151</sup> Concretely, FRENAPO demanded the implementation of a non-contributory unemployment allowance of 380 pesos for all unemployed heads of household, the universal coverage of child allowances of 60 pesos per month and a universal minimum pension of 150 pesos per month for the elderly (Interview Lozano 2014). Faced with these demands for universal social protections, the government elaborated—without consulting trade unions, the FRENAPO or unemployed workers' movements—an emergency decree that replaced the existing family allowance system and workfare programs with a system of family allowances of 30 pesos per month for all children under the age of 14 and a universal minimum pension for persons over 75 without income (Decree 1382/2001). Although this decree would have significantly expanded family allowance coverage towards low-income earners, it was heavily repudiated by all popular class actors, including the FRENAPO. Unemployed workers' organizations, such as the CCC, also participated in protests against the decree. This broad-based rejection had much to do with the fact that the decree did not deviate from the general logic of permanent fiscal adjustment that characterized the government of De la Rúa. Indeed, by implementing the decree, the government estimated it could reduce spending in 2002 by approximately 500 million pesos. In practice, the decree meant that most formal workers would see their family allowances either significantly reduced or even completely eliminated.152 For unemployed workers and their organizations it was unacceptable to implement the reform on costs of formal labor as well as to lose the

<sup>151</sup> Members of the alliance included among others unemployed workers' organizations, such as the FTV, human rights organizations, progressive sectors of the church and a variety of politicians from the FREPASO, ARI and the PJ. One of them was the Peronist governor of the Province of Santa Cruz and future president Néstor Kirchner.

<sup>152</sup> The decree abolished all family allowances for all children at the age of 14 or older and generally restricted the receipt of the benefit to households with incomes below 1,000 pesos.

workfare benefits many of them had struggled for in exchange for a small benefit of 30 pesos per month (Interview Ardura 2014; Interview Lozano 2014; *El Clarín*, November 2, 2001).<sup>153</sup> Finally, the decree was successfully opposed in the courts by the CGT and never implemented (Garay 2010, 65).

All in all, the government of De la Rua attempted to deepen the regressive restructuring of social policies that had been successfully pushed forward during the preceding governments. However, due to the growth of protest against these measures and the weakness of De la Rúa's governing alliance, a major part of its agenda could not be implemented.

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

The rise of neoliberalism was rooted in the failure of heterodox policies to stabilize the economy during the 1980s. The virtually permanent economic and political crisis of the so-called lost decade had led to a shift in the distribution of power resources that was increasingly detrimental for the implementation of progressive social policy. Rising unemployment and informality reduced the structural power resources of the popular classes, while a high dependence on foreign debt increased the influence of IFIs. The increasing dominance of conservative and neoliberal currents within the main parties, the PJ and the UCR, dramatically reduced the associational power resources of the political left. A neoliberal shift in public opinion furthermore provided ample discursive power resources for neoliberal actors to frame regressive social policy reform in such a way that it appeared both necessary and desirable for the majority of the society and hence facilitated the electoral viability of a neoliberal alliance. Based on these shifts in the distribution of power resources, Menem was able to form a strong, regressive governing alliance in which neoliberal technocrats and representatives of the economic elite dominated agenda-setting. This, in turn, initiated a decade of regressive social and economic policy reform and rapidly rising income inequalities.

<sup>153</sup> Indeed, the two biggest unemployed workers' movements, the FTV and the CCC, have very close links to the trade union movement, so that it was quite common that these organizations supported demands of the trade union movement even when these demands affected only formal workers.

Figure 7: Evolution of income inequality measured with the Gini-coefficient, 1989–2001

Sources: Gasparini and Cruces (2010, 5).

In contrast to the exclusionary governing alliance of the military junta between 1976 and 1983, Menem's governing alliance incorporated popular class actors, though in a strongly subordinated way. This subordinated incorporation meant that reform was undertaken gradually and went along with nearly permanent internal discussions and concessions. Due to the historically strong interests of the trade union movement in the preservation of the system of social health insurances, the main concession to organized labor consisted of refraining from the complete deregulation of this sector. With the growth of lower popular class protests and the emergence of unemployed workers' movements since the mid-1990s, these sectors were also able to extract concessions, mainly in the form of the expansion of social assistance policies. Throughout the decade, resistance to neoliberalism grew and the distribution of power resources turned gradually more favorable to progressive actors. Nevertheless, popular struggles were defensive during the whole period and regressive social policy reform continuously outweighed concessions. Social protections for the vast majority of the popular classes deteriorated significantly and low-income earners were particularly hard hit. Through privatization, the introduction of market mechanisms, increasingly regressive benefit calculation formulas and exclusionary eligibility criteria, the health and pension systems were restructured in favor of high-income earners. Low-income earners either completely lost access or saw their benefits dramatically reduced. Simultaneously, employers benefited from massive reductions in social insurance contributions. The gradual increase in social assistance expenditure neither compensated for these massive disadvantages nor did the increase adequately reflect the dramatic rise in poverty and unemployment. Hence, despite the neoliberal rhetoric about the focalization of social expenditure on low-income earners, overall social policy reform during the 1990s took away resources from the poorest sectors and concentrated them increasingly on the better off.


Table 26: Change in social protection for low-income earners, 1989–2001

Due to the truncated nature of social health insurance reform, the strong interests of the trade unions in the system were not significantly altered and therefore the main mechanism of path dependency remained intact. In fact, the conversion of social health insurances with mixed administration into purely trade union-controlled entities provided trade unions with additional interests in maintaining the status quo. The increasingly high importance of workfare benefits for the functioning of unemployed workers' movements and the inability of De la Rúa's government to cut them on a massive scale indicate furthermore that this time path dependency effects also emerged in the system of social assistance policies.

## **4.8 Crisis, Paradigm Change, and the Expansion of Social Protections for Low-Income Earners, 2002–2015**

The inability of De la Rúa's government to create a stable base of support, to overcome the economic crisis, and to contain the ever-growing protest movements finally ended in a dramatic breakdown of the neoliberal governing alliance. The emergence and growing strength of new progressive actors and the discredit of the neoliberal model furthermore led to a significant shift in the distribution of power resources. Against this background, the center-left electoral alliance of Néstor Kirchner won the 2003 presidential elections and initiated a period of social policy expansion, which strongly extended and improved social protections for low-income earners.

## **The Political Arena: Social and Economic Context, Constellation of Actors, and Distribution of Power Resources**

When president De la Rúa resigned in late 2001 in the midst of massive protests, the Argentine economy was stuck in a profound crisis, foreign debt had become a massive burden and the one-to-one exchange rate regime between the peso and the US dollar had become unsustainable. His successors, the interim presidents Adolfo Rodríguez Saá and Eduardo Duhalde, first announced debt default in late 2001 and eventually, in early 2002, abolished the convertibility regime and devalued the peso (Golbert 2004, 13). Four years of crisis finally culminated in 2002 with an accumulated reduction of 20% in GDP, 21% unemployment and over 50% of the population living under the absolute poverty line (Beccaria, Groisman, and Maurizio 2009, 21; Vargas de Flood 2006, 201).

However, from mid-2002 on, the Argentine economy began a rapid process of recovery. The devaluation of the currency had altered the structure of relative prices in favor of domestically produced goods and services, which constituted an important comparative advantage for the crisis-ridden local industries and facilitated the expansion of production and employment. Furthermore, a strong increase in commodity prices led to an expansion of the primary sector, especially soybean production, which contributed to the generation of fiscal revenues and trade surpluses (Burchardt 2016, 60–69). The government reinforced these benign economic trends with a low-interest-rate policy and a range of measures that fostered a steady growth of domestic demand (Etchemendy and Collier 2007, 390). In effect, between 2003 and 2015 the Argentine economy expanded at an average yearly rate of nearly 4.8% (INDEC 2016, 7; Trujillo 2017, 9–16). Industry's share of GDP recovered slightly from 29% to 33% between 2002 and 2011 and industrial employment rose from 17% to 21% of overall employment (Timmer, de Vries, and de Vries 2014). Between 2002 and 2007 alone, the absolute number of industrial workers grew by an estimated 36% (Wyczykier and Barattin 2012, 68). Unemployment fell from 20% in late 2002 to 10% by the end of 2005 and 6.6% in mid-2015 (INDEC 2003; MPyT 2019). Informality among wage earners decreased from over 48% in 2004 to under 33% in 2015 (CEDLAS and World Bank 2018).


Table 27: Economic and social context, 2002–2015

Sources: Rapoport (2013) for GDP growth and inflation 2002–2006; Trujillo (2017, 9–16) for GDP growth 2007–2014; INDEC Database for GDP growth 2015; IPC del Congreso (2018) for inflation 2007–2015; Timmer, de Vries, and de Vries (2014) for industry share of GDP and share of industrial employment; INDEC Database for unemployment; and CEDLAS and World Bank (2018) for informality.

These developments meant that the disciplining effects of unemployment and informality weakened, that workers became more difficult to replace and that employment in capital intensive sectors grew. Taken together, this led to a significant recovery of the structural power resources of the popular classes.

At the same time, the break with the convertibility exchange regime, the default on foreign debts and the return of fiscal and trade balance surpluses drastically reduced Argentina's dependence on capital inflows, and hence, reduced the structural power resources of the IFIs and foreign investors.

Moreover, growing resistance to the neoliberal model and its failure to comply with its promises of progress, stability, and wealth for all led to a widespread discrediting of neoliberalism and the return of a more state-interventionist consensus in public opinion. The rapid initiation of economic and social recovery with the implementation of policies inspired by and justified with the logic of the Keynesian and ISI paradigms further contributed to the consolidation of this consensus. In contrast to data from 1991 and 1995, World Values Survey data from 1999, 2006 and 2013 show stable majorities in support of an active role for the state in reducing income inequalities, increasing public welfare provision, and expanding national ownership of companies (see Figures 8–10; Inglehart et al. 2014: *World Values Survey Waves 2–6*). As part of this general discursive shift in favor of state intervention and redistributive policies, proposals for universal social policies that would extend coverage to low-income earners became increasingly popular in both academic and political circles (Danani and Beccaria 2011; Hintze and Costa 2011, 164). On the international level, universal and redistributive social policies were promoted by the ILO through the Social Protection Floor program. These developments provided progressive actors with much more favorable conditions in which to frame their demands in a positive manner and to mobilize mass support, and hence, significantly increased their discursive power resources.<sup>154</sup>

The profound social, economic, and political crisis of the late 1990s and early 2000s also significantly affected the constellation of actors participating in Argentine social politics. A range of corruption scandals and the incapacity of governments of different colors to reverse the process of economic and social deterioration led to a widespread and profound loss of confidence in the Argentine political elite (Adamovsky 2012, 436). Not surprisingly, the governing parties were particularly hard hit. The center-left FREPASO, whose participation in the neoliberal government was widely perceived as a complete betrayal of its own program, completely disintegrated.<sup>155</sup> The UCR became a shadow of its former self, winning only 2% in the 2003 presidential elections and 14% in the same year's legislative elections. Instead, former UCR politicians left the party to form the neoliberal Recrear party and the center-left ARI party. The PJ party did not suffer from similar splits. However, internal differences led to the decision that three Peronist candidates were allowed to compete separately in the 2003 presidential elections.

<sup>154</sup> Figures 8, 9 and 10 show changes in the Argentine public opinion that reflect the shift from neoliberal predominance during the early 1990s to the prevalence of a more interventionist post-neoliberal paradigm during the 2000s. In four waves of the World Values Survey respondents were asked to position themselves on a scale from 1 to 10 in order to express their inclination to either of two given options. For the elaborations of the figures the positions 1–4 have been summarized and represent those who prefer the first option. The positions 7–10 have been summarized and represent those who prefer the second option. The middle-positions 5 and 6 are considered expressions of neutrality.

<sup>155</sup> One of the two founding parties of the FREPASO, the Frente Grande, continues to exist but plays only a minor political role compared to the 1990s.

Figure 8: Public opinion with regard to government and individual responsibility in welfare provision, 1991–2013

Figure 9: Public opinion with regard to income inequalities, 1991–2013

Sources: World Values Surveys (1991, 1995, 2006, 2013).

Sources: World Values Surveys (1991, 1995, 2006, 2013).

Figure 10: Public opinion with regard to public and private ownership of business companies, 1991–2013

Carlos Menem was the candidate uniting most of the right and neoliberal factions of the party. Adolfo Rodríguez Saá was the candidate of the center. Néstor Kirchner adopted a center-left profile and united support from both the center and the left wing of the PJ. To sustain his campaign, he formed the centerleft electoral alliance Frente para la Victoria (FpV), which additionally incorporated several smaller center-left and left political parties and social movements. Kirchner finally won the 2003 presidential elections (Zelaznik 2011, 96–99). The swift economic recovery and a range of redistributive policies soon contributed to a strong rise of Kirchner's popularity. By the mid-2000s, the FpV became the dominant electoral force in Argentina (Cavarozzi 2006, 143). Although it received considerable middle-class support, its electoral stronghold was composed of the popular classes and its support was strongest among those groups with low incomes and low levels of formal education (Moncagatta and Safranoff 2013, 32–43). Overall, these developments in the Argentine party system were highly favorable for an extension of social protections low-income earners.

While during the 1990s employers' associations, employer-financed think tanks and IFIs were united by a strong neoliberal consensus, during the 2000s these actors suffered from fragmentation. In the context of profound economic crises, tensions between different sectoral interests regarding the exchange regime came increasingly to the forefront. On the one hand, external creditors,

Sources: World Values Surveys (1991, 1995, 2006, 2013).

service sector associations and enterprises that had accumulated foreign debt favored the replacement of the peso with the US dollar as official currency. This way they hoped to avoid a fall of the dollar value of their investments in Argentina and a respective relative increase of the value of their foreign debts. On the other hand, the agricultural sector and export-oriented industries, as well as companies that had invested larger amounts outside Argentina, supported a currency devaluation, because this would increase their international competitiveness and the local value of their foreign investments (Basualdo 2001, 58–101). The Kirchnerist governments furthermore used state interventionist policies, such as import regulations and subsidies, to selectively favor certain sectors and companies and thereby to additionally divide the capital side (Bonvecchi 2011, 143–146). The involvement of employer-financed think tanks, employer associations and the IFIs in the agenda-setting of the 1990s furthermore led to the strong discrediting of these actors in the opinion of the public. Taken together, these developments implied a significantly weakened influence of these actors. While during the 1990s they had a decisive influence on the agenda-setting of social policy reform, during the 2000s their struggles were rather defensive and consisted of criticizing and resisting the government's agenda.

The profound economic and social crisis also affected the trade union movement. Although union density fell only slightly to an estimated 40%, the absolute number of affiliates declined significantly due to the rise of unemployment (Marshall 2005, 25). The threat of unemployment and the defensive strategy towards neoliberal policies pursued by the majority faction of the union movement led to a significant weakening of most unions' activist structures and mobilization capacities (Etchemendy and Collier 2007, 369–370). Furthermore, after the MTA left the CGT due to a dispute over labor market reform in 2000, the labor movement was politically divided into three confederations (Campione and Rajland 2006, 317–318). Hence, the associational power resources of the trade union movement were significantly weakened by the early 2000s. However, with the economic recovery and a range of government measures that strengthened collective bargaining and employment growth, a process of union revitalization began after 2003 (Weinmann, Bossert and Hecker 2016, 195–189). Between 2003 and 2005 the number of strike actions and signed collective agreements tripled (Etchemendy and Collier 2007, 369– 370; Etchemendy 2011, 59–60). According to Carlos Gdansky, secretary general of the metal workers' union in the La Matanza district, his union grew from 70,000 members in 2001 to 300,000 in 2011 while union specialist Héctor Palomino estimated a 24% growth of overall union membership between 2003 and 2009 (Abal Medina 2012, 120). Besides this important growth, the capacity to influence policy was strengthened by the reunification of the CGT in 2004. Hugo Moyano, the former leader of the combative MTA, became its president and later also assumed the vice-presidency of the PJ party (Campione and Rajland 2006, 317–318; Svampa 2008a, 88). Although both the CGT and the CTA split again into various factions after 2008, all in all, an important recovery of associational power resources of the Argentine union movement can be observed (Weinmann, Bossert, and Hecker 2016).

In contrast to the union movement, unemployed workers' organizations were able to massively increase their associational power resources already during the years of the crisis. According to the national coordinator of the CCC, the organization grew from approximately 12,000 members in late 2001, to 50,000 in early 2002 and 70,000 in 2003 (quoted in Golbert 2004, 21). The FTV and other organizations also grew strongly during this period so that the overall membership of unemployed workers' movements reached approximately 300,000 at that time (Interview D'Elía 2014; Alcañiz and Scheier 2007, 160). Due to various factors, the general growth of their associational power resources during the crisis was succeeded by a period of more ambivalent developments after 2003. One reason was that the number of unemployed workers and the public acceptance of roadblocks declined with the recovery of the economy. Another reason was that the unity among unemployed workers' movements during 2001 did not last for long and differences regarding political orientation and strategy came increasingly to the forefront (Natalucci 2010, 109–110). During the Kirchnerist governments, the movements furthermore split into two factions that either supported or opposed the government. The alliance with the government enabled the FTV to maintain a membership of roughly 90,000 during the following years (Interview D'Elía 2014). Other allied social movements, such as the Tupac Amaru, the Movimiento Evita, or the Frente Transversal were even able to grow significantly and to convert into mass movements (Natalucci 2012a; 2012b). Government support and privileged access to subsidies and social assistance programs made an important contribution to this growth. At the same time, opposing unemployed workers' movements, which received fewer subsidies and faced frequent judicial persecution for their protest actions and roadblocks, mostly suffered from declining membership (Svampa 2008b, 9–13). Nevertheless, many of these movements also retained significant associational power resources. The CCC, for example, was able to maintain a membership of approximately 35,000 and a high level of protest activity (Interview Ardura 2014). Summing up, despite a certain decline in public visibility compared to the crisis period, the unemployed workers' movements were able to consolidate their role as important actors in the national political arena after 2003.


Table 28: Popular class power at the beginning of the period, 2002-2005


At the same time, the women's movement became a real popular mass movement (Di Marco 2010, 203). By the mid-2010s it proved capable of mobilizing several hundred thousand demonstrators throughout the country against gendered violence and for legal, safe, and free abortion. Participation in the Encuentro Nacional de Mujeres, the most important annual gathering of the Argentine women's movement, rose from around 12,000 in 2001 to over 60,000 in 2015 (Tarducci 2005, 397; *Página 12*, October 11, 2015). The massive growth of the women's movement was to an important part related to the growing participation of women from progressive parties, trade unions, and unemployed workers organizations. The annual national gatherings served as an opportunity to participate in issue workshops and build networks, which in turn constituted an important support for strengthening the influence of women in their respective organizations. By 2002 the CTA and several unemployed workers' movements had incorporated the demand for legal abortion into their programs (Daich and Tarducci 2012, 7; Sutton and Borland 2013, 213). For the vast majority of activists, women's interests also included broader redistributive and universalizing social policy reform, because such policies were considered to contribute to a diminution of both gender and class inequalities (Interview Díaz 2014; Interview Ocar and Martínez 2014).

In summary, the distribution of power resources and the constellation of actors that emerged out of the multiple economic, social and political crises of the 1990s and early 2000s provided a favorable context for the formation and consolidation of an inclusionary-progressive governing alliance. The re-emergence of Keynesian, demand-oriented thinking in public opinion led to a renewed strengthening of the discursive power resources of popular class actors. Progressive forces, such as the CTA, unemployed workers' and women's movements, and the center-left FpV electoral alliance, had emerged and grown in strength. This meant an important increase of the associational power resources of the popular classes. The economic boom after 2002 and a slight trend towards re-industrialization furthermore contributed to a recovery of labor's structural power resources. In contrast, key drivers of the regressive social policy agenda of the 1990s, such as employer-financed think tanks, the IFIs and employer associations, suffered from discrediting before the public and increasing internal fragmentation.

#### **Struggles over Social Policy in a Period of Profound Economic, Social and Political Crisis, 2002–2003**

By late 2001, poverty had risen to unprecedented levels and affected approximately 40% of the population (Cavarozzi 2006, 214). This not only fostered organized protests by unemployed workers' movements but also led to the proliferation of somewhat loosely organized lootings. In the face of an acceleration of capital flight, the government announced on the 1st of December a restriction on the withdrawal of money from bank accounts, which led to a wave of middle-class protests as well as a joint general strike of the CTA and the CGT on the 13th of December. In response to the proliferation of lootings and the increasing intensity of protests, President de la Rúa declared a state of emergency on the 19th of December and ordered the repression of protests, which resulted in severe street battles and numerous deaths. Instead of calming the situation, however, the measure led to a further growth and radicalization of the protests. On the 20th of December, De la Rúa saw no other choice than to resign (Adamovsky 2012, 436–440; Cavarozzi 2006, 213–214).

On the 22nd of December, the parliament elected the Peronist governor Rodríguez Sáa as his successor (Cavarozzi 2006, 215). The new president was immediately confronted with social demands. During the preceding days, the FRENAPO had mobilized over 1.7 million supporters in an informal referendum in favor of its demand for universal social protection. On the day of Rodríguez Sáa's election, the FRENAPO furthermore announced that it was going to present a bill with its demands in the parliament. Simultaneously, several social movements met in the city of Rosario and demanded an immediate halt of foreign debt payments (*La Nación*, December 22, 2001). In response to the growing strength of these demands and in consideration of the delicate political context, Rodríguez Sáa announced in his inaugural speech a default on foreign debt and the creation of one million state-sponsored jobs. He immediately invited representatives of unemployed workers' movements to discuss the creation and distribution of these jobs (Interview Ardura 2014; Interview Berra 2014). While these measures were well received by most of the social movements, Rodriguez Sáa lacked support from his own party, which led him to resign on the 30th of December (Cavarozzi 2006, 215–216).

On the 2nd of January 2002, the Peronist senator Eduardo Duhalde took office as his successor (Cavarozzi 2006, 216). One of his first announcements was to end the fixed one-to-one exchange rate between the US dollar and the Argentine peso. This caused a sharp devaluation of the local currency and thereby a further loss of income for most Argentinians (Golbert 2010, 147). The unemployed workers' movements immediately returned to the streets to demand that Duhalde keep his predecessor's promise to create one million jobs and FRENAPO leaders reiterated their demand for a universal unemployment allowance (*La Nación*, January 14, 2002; *La Nación*, January 15, 2002). Aware of the vulnerability of his government and the power resources of the social movements and the FRENAPO, Duhalde met with social movement leaders and shortly thereafter announced the implementation of a non-contributory workfare program called Programa de Jefes y Jefas de Hogar Desocupados (PJJHD), which would provide a monthly allowance of 150 pesos in exchange for participation in communal work (Interview Ardura 2014, Interview D'Elía 2014). In order to calm down political conflicts, he also invited these movements as well as the trade union centrals, various employers' associations, and representatives of the Catholic church and other religions to participate in a multisectoral roundtable to agree on policies to confront the crisis (Interview López 2014). One of the policies discussed in this context was the PJJHD. The government's original proposal limited the number of beneficiaries to 500,000 and the duration of the benefit to only three months (Golbert 2004, 27; Decree 165/2002). These restrictions were rejected by the CTA, the CCC and the FTV, which demanded broader social policies that would cover all those in need and have stronger redistributive impact (Interview Ardura 2014; Interview D'Elía 2014; Interview Lozano 2014). Finally, the multisectoral roundtable agreed to significantly extend the program. The introductory paragraph of the Decree 565/2002 that introduced the PJJHD stated that the multisectoral roundtable agreed on an "urgent necessity to universalize the Plan Jefes y Jefas de Hogar in order to secure a minimum monthly income to all Argentine families" and that "it is important to promote a redistribution of resources from those with high incomes towards the social sectors with low income."<sup>156</sup> This allowed an unprecedented expansion of coverage and within a few months the PJJHD

<sup>156</sup> Author's translation.

reached nearly two million households (Golbert 2004, 23–27).<sup>157</sup> Furthermore, the universalistic discourse and the proposition of rights to "social inclusion" found in the text of the Decree 565/2002 distinguished the PJJHD strongly from preceding social assistance programs. Parallel to the PJJHD, the government introduced the Plan Nacional de Seguridad Alimentaria (PNSA), which consisted mainly in the provision of food to low-income households, and the program Ingreso para el Desarrollo Humano (IDH), which was similar to the PJJHD but did not require participation in communal work and reached about 200,000 households (Cruces, Epele, and Guardia 2008, 22; Vargas de Flood 2006, 167–168).

All in all, social assistance and workfare policies that benefited the lowest income sectors were massively expanded during the interim presidency of Eduardo Duhalde. Between 2001 and 2003 the overall expenditure on social assistance and unemployment protection increased from 1.5% to 2.5% of GDP (DAGPyPS 2011). The political process behind this expansion reflects the shift in the distribution of power resources in favor of progressive popular class forces, such as the unemployed workers' movements and the FRENAPO, and the subsequent redefinition of the relationship between the national government and these forces. While the government of President De la Rúa was reluctant to receive these actors as negotiation partners, responded to their protests with repression and finally fell over a wave of protest and social conflict, Duhalde acknowledged the need to receive these actors and to negotiate the terms of social policy with them. This disposition was related less to any genuinely progressive political orientation of his government than to the fact that social conflict, roadblocks, and mass protests had reached such an intensity that they were perceived to be a real threat to the ability to govern. Most employers' associations also perceived this situation as being highly explosive, considering that it included an increasing number of factory occupations, and so refrained from more vehement resistance to the expansion of social assistance (Adamovsky 2012, 421–422; Golbert 2004, 17–18; Interview López 2014).158 159

<sup>157</sup> This meant that coverage was approximately 15 times higher than that of the biggest preceding workfare program, the Plan Trabajar (Golbert 2004, 23–27).

<sup>158</sup> Between 2002 and 2003 approximately 160 factories were occupied and resumed production under worker control (Adamovsky 2012, 421–422).

<sup>159</sup> Stronger objections came only from the associations of the big agricultural producers SRA and CRA, which protested the implementation of export taxes on agricultural products and processed foods to finance social assistance measures (*La Nación*, March 5, 2002). However, despite their public refusal, no major protests were organized. Roughly two thirds of the PJJHD was financed through export taxes (Golbert 2004, 30).

Figure 11: Roadblocks and social assistance, 1997–2005

Although policy change was most profound in the area of social assistance and unemployment protection, the new distribution of power resources also influenced social policy making in other areas. In early 2002 health workers, associations of patients with chronic conditions, and unemployed workers' movements organized several protests to demand free access to medical care including pharmaceuticals.<sup>160</sup> Hospital workers furthermore staged several strikes for improvements in supply of pharmaceuticals and working conditions (Garay 2010, 68). In response to these protests and the severe crisis of the health system the government decreed a "sanitary emergency" in early April 2002 and announced a range of measures to ensure the financial viability of the system, to increase the redistribution of resources among health insurance funds and to "universalize free access to pharmaceuticals" for low-income earners (Decree 486/2002). Soon after, in June 2002, the government finally announced the implementation of the Plan Remediar, which gradually extended the free provision of basic pharmaceuticals nationwide to over 6500 primary health care

Sources: Massetti (2006, 30) for the number of roadblocks; DAGPyPS (2011) for social assistance and unemployment protection expenditure.

<sup>160</sup> The unemployed workers' movements alone organized over 30 roadblocks and other protests to demand free access to pharmaceuticals (Garay 2010, 68).

centers and approximately 15 million Argentinians living in low-income households (*La Nación*, June 19, 2002; Garay 2010, 81–83; Tobar 2004). In order to address the financial deficit of the health system, employer contributions to the health insurance system were raised from 5% to 6% and to the pensioners' health insurance from 1.5% to 2% (*La Nación,* November 17, 2002). Furthermore, the Obligatory Medical Plan, which detailed the treatments all health insurance funds had to cover, was replaced by the Emergency Obligatory Medical Plan, which contained a revised and reduced catalog of treatments and pharmaceuticals. Among other things, it was mandated that physicians had to prescribe generic drugs wherever possible (*La Nación*, February 15, 2002).

In July 2002, the government decided to raise the minimum pension from 150 to 200 pesos (*La Nación*, July 17, 2002). Compared with the inflation rate of slightly over 25% in 2002, this meant that the pensioners with the lowest incomes experienced a significant increase in the real value of their benefit (MTEySS 2017, 37; Rapoport 2013, Table 5). This measure, however, did not alter the problem that, due to the high contribution requirements instituted in the early 1990s, most low-income earners completely lacked access to any kind of pension benefit. Against this background, the biggest unemployed workers' movements started to organize protest activities in late 2002 to demand the expansion of pension coverage to low-income groups without access. Duhalde received the leaders of these movements again and finally promised the creation of one million additional pension benefits. However, as no concrete measures followed, the unemployed workers' movements together with allied trade unions staged massive protests for an expansion of pension coverage in January and February 2003. In response to these, the government finally implemented the Plan Mayores in March 2003, which provided a monthly benefit of 150 pesos to persons over 70 who live in the poor provinces in the north of Argentina and have no other source of income. Yet, due to the restrictive access criteria, the program had only a very minor impact (Garay 2010, 70–84; Pautassi, Rossi, and Campos 2004, 11).

All in all, the ability of Duhalde's government to restore a certain level of social peace and governability was very much related to his recognition of the strength and mobilization capacity of a range of progressive popular class actors that had emerged and grown during the 1990s, such as the unemployed workers' movements and combative union confederations. He acknowledged them as legitimate political actors, initiated a dialogue with them, and responded to several of their demands. This led to a significant extension of social protections for low-income earners. Notwithstanding these concessions, the redefinition of the relationship between the government and these progressive social forces was more a result of perceived political necessity than a genuinely pursued aim of President Duhalde. As soon as the political situation appeared to be stabilizing, the government returned to a more conservative stance. Shortly after implementing the supposedly universal PJJHD program, an internal resolution of the Ministry of Labor decided that no new enrollments to the program would be permitted after the 17th of May 2002. This contributed to the fact that merely 62% of the households that met the eligibility criteria ended up receiving benefits from the program (Neffa 2008 et al., 102ff).

In general, social policy responses to unemployed worker and trade union protests became much more limited from mid-2002 on and the government returned increasingly to a more repressive stance. Repression finally culminated with two young protestors being killed by the police in a roadblock on the 26th of June 2002 (Pereyra, Pérez, and Schuster 2008, 20). This escalation of violence, in turn, severely damaged the image of the government, led to a further growth of protests, and aroused the opposition of significant parts of the lower and middle classes and the media. Eventually, this contributed to Duhalde's decision to schedule elections for early 2003 and to refrain from running for president (Svampa 2008b, 11; Norden 2011, 102).

#### **The Formation of a Strong Inclusionary Governing Alliance and the Initiation of Progressive Social Policy Expansion, 2003–2007**

A more solid transformation of the relationship between the government and the new popular class actors took place under the following presidency of Néstor Kirchner, who deliberately pursued their integration into his governing alliance. As governor of Santa Cruz, Kirchner had initially supported the government of Carlos Menem; from the mid-1990s, however he had become a fierce critic of Menem and the neoliberal model (Galasso 2016, 28–35). During the crisis he supported the FRENAPO and its demand for universal social protection (Interview Fuentes 2014). And when he ran for president as one of the three Peronist candidates in the 2003 elections, he based his campaign on a center-left and strongly anti-neoliberal rhetoric, promising the implementation of a more state interventionist economic model focused on social inclusion, redistribution, and national sovereignty as well as a strengthening of the public health and pension systems (*La Nación*, April 26, 2003; Zelaznik 2011, 96). This platform was well received by many of the unemployed workers' movements, the more combative factions of the union movement, many human rights organizations, the more progressive sections of the PJ party, and several smaller center-left and left parties. During the campaign, Kirchner was able to forge an alliance with the biggest unemployed workers' movement, the FTV, the center-left party Frente Grande and some leaders of the CTA (Interview D'Elía 2014; *La Nación*, February 19, 2003). Because Kirchner was relatively unknown to the public compared to most of the other candidates, forging political alliances within the powerful PJ party apparatus was also crucial for the success of his candidacy. While most of the influential Peronist politicians supported one of the other two Peronist candidates, Meném or Rodríguez Sáa, Kirchner eventually won the backing of interim President Eduardo Duhalde, which added the support of significant base-level party structures to his campaign, especially in the populous province of Buenos Aires.<sup>161</sup> Furthermore, Duhalde was able to gather the support of 71 mostly smaller CGT unions for Kirchner (*La Nación*, February 21, 2003). With this support, Kirchner finally obtained 22% of the votes in the first round of the presidential elections, which enabled him to enter the final ballot against Ménem, who had reached 24% of the votes. In the face of unfavorable public opinion polls Ménem finally withdrew his candidacy and Kirchner became president (Cavarozzi 2006, 136– 138).

Once in office, Kirchner actively pursued the consolidation and expansion of his progressive governing alliance. In order to incorporate a broader range of progressive and popular class-based organizations, the government implemented a range of both symbolic and material measures (Etchemendy and Garay 2011, 284; Zelaznik 2011, 97). Kirchnerism shifted economic policy from neoliberalism to Keynesian state interventionism, started to renationalize formerly privatized companies, expanded social expenditure, established price and trade controls, and promoted collective bargaining and wage hikes as a way to foster local demand as a driver of the economy (Bonvecchi 2011). In public speeches the government frequently emphasized that its redistributive policies were not only improving the living conditions of the beneficiaries, but that they were also crucially contributing to the growth of the economy and the progress of the society as a whole, thereby effectively taking advantage of the post-neoliberal development paradigm that dominated public opinion (Díaz Rosaenz 2017, 101–106). In addition to these broad-based policies in favor of both formal and informal labor, the government used the distribution of selective benefits, such as political posts and resources of social assistance and infrastructure programs, to favor loyal supporters and to negotiate political support on different levels of the political arena. On the one hand, basic units of the PJ party in poor neighborhoods continued to function as clientelistic networks using social programs to establish direct linkages to low-income earners. On the other hand, on a more aggregate level, the selective distribution of resources through political posts, social assistance, and infrastructure programs served as leverage in negotiations with social movements and lower rank leaders of political parties (Etchemendy and Garay 2011, 283–285; Mauro and Rossi 2011, 17). Given that in Argentina provinces and municipalities are highly dependent on fiscal transfers from the national government, this helped

<sup>161</sup> Duhalde's support of Kirchner was of a strategic nature rather than based on genuine political proximity. Indeed, Duhalde first tried to convince two political allies, the governors of Santa Fé and Córdoba, to run for the presidency, but failed in both cases. As he nevertheless wanted to avoid a victory of his long-term rival Carlos Ménem, he finally forged an agreement with Néstor Kirchner in early 2003 (Cavarozzi 2006, 136– 138).

Kirchner to rapidly build a network of supporters from both major parties, the PJ and the UCR (Cavarozzi 2006, 142).

With the aim of extending his alliance with the unemployed workers' movements, Kirchner received leaders of the FTV and the CCC immediately after assuming office. A short time later he also invited leaders of smaller social movements and established a fluid channel of communication through the General Secretariat of the Presidency. Kirchner himself participated several times in meetings, emphasizing that he was willing to defend lower popular class interests, but that in order to do so the social movements would have to support the governing alliance (Boyanowski Bazán 2010, 114–116; Galasso 2016, 62–63). Besides this general promise to defend low-income earners' interests, the government conceded participation in housing construction programs and other public works to several organizations. This strategy turned out to be relatively successful, so that soon many of the unemployed workers' organizations either joined the alliance or at least loosely cooperated with it (Interview Ardura 2014; Mauro and Rossi 2011, 169; Tavano 2015, 232). In the following years, new Kirchnerist social movements such as the Movimiento Evita and the Frente Transversal were founded and grew rapidly. The collaboration with the government was advantageous for these movements, and in particular for their leaders. The distribution of the resources of social assistance and housing programs was strongly biased in favor of loyal organizations, which ultimately increased the capacity of these groups to attract new members. Furthermore, the receipt of state resources strengthened the control of the leaders over the rank and file, because the leaders were able to grant more benefits to particularly committed and loyal members (Interview Falcón 2014; Golbert 2004, 21–22; Massetti 2009, 60–88). In return, these organizations frequently mobilized their members in support of the government both electorally and on the streets (Etchemendy and Garay 2011, 284).

At the same time, Kirchner actively sought the incorporation of the trade union movement into his government. While the center-left orientation of his government was in many ways closer to that of the progressive CTA, Kirchner believed that a broad-based inclusionary governing alliance required also the support of the two CGTs, which still represented the majority of the union movement (Etchemendy and Collier 2007, 390). To achieve this support, Kirchner established a close alliance with the more combative unions that were organized in the CGT Disidente<sup>162</sup> and also invited the more moderate sector (*La Nación*, December 4, 2003). In parallel, the government began to implement decidedly pro-union policies that helped unions to revitalize collective bargaining, reinstituted tripartite concertation, reinforced workers' rights, and fostered employment creation (Weinmann, Bossert, Hecker 185–189). These measures provided the ground for significant wage increases and were well

<sup>162</sup> After the MTA had left the CGT due to political differences, the member unions of the MTA founded the CGT Disidente.

received by the union movement and the majority of the workers.<sup>163</sup> In 2004, the two CGTs finally reunited and in 2005 Hugo Moyano, leader of the combative teamsters' union and close ally of the government, ascended to become its sole secretary general (*La Nación*, July 7, 2005). In 2006, the CTA elected as secretary general a supporter of the government, Hugo Yasky of the teachers' union confederation (*El Clárin*, November 11, 2006). Hence, by the mid-2000s, the trade union movement had become a solid pillar of support of the governing alliance (Wyczykier and Barattin 2012, 55).

A majority of the human rights organizations also joined the governing alliance, which reopened trials against human rights violators of the last military dictatorship and provided several of these organizations with funding to carry out education and social assistance projects (Mauro and Rossi 2011, 171; Zelaznik 2011, 97). All in all, within two years a strong progressive governing alliance with broad-based popular class support had emerged which was both willing and able to implement a redistributive social policy agenda.

The most profound reform process was undertaken in the pension system, which had constituted an important focus of unemployed workers' movement protests during early 2003. To countervail the inegalitarian distribution of resources within the system, the government implemented a combination of general benefit increases and disproportionally higher minimum pension increases. Between 2003 and 2004 the real value of the minimum pension was raised by a stunning 32% while the average pension benefit grew by 13%.<sup>164</sup> Until the end of Kirchner's presidency in 2007, the relative value of the minimum pension increased from 50% to 80% of the average pension, so that the distribution of resources within the group of pensioners had become highly egalitarian (MTEySSS 2017, 37).

Ten months before the 2005 mid-term elections, the government then addressed one of the main demands of the unemployed workers' movements and the CTA by implementing the Pension Inclusion Program, which resulted in a massive expansion of pension coverage among the elderly and thereby strongly benefited unemployed and informal workers.<sup>165</sup> The program introduced the possibility of early retirement for unemployed workers at the age of 60 or later, as well as providing access to pension benefits to persons at retirement age who did not meet the requirement of 30 years of contribution. (Law 25.994). In order to account for the missing years of contribution, the benefit was then

<sup>163</sup> During Néstor Kirchner's presidency both the minimum wage and the wages accorded in the main collective agreements nearly doubled in real terms (Etchemendy 2011, 57, 76).

<sup>164</sup> Estimates of the INDEC Database have been used for inflation adjustment.

<sup>165</sup> Although the CGT did not explicitly demand such measures, as part of the inclusionary governing alliance the CGT clearly supported the extension of pension coverage beyond the margins of formal labor (Puricelli 2019, 118–133).

reduced by a certain proportion during the first five years; afterward it converted into a full regular pension benefit (Law 25.994; Decree 1.454/05).<sup>166</sup>


Table 29: Composition of the new governing alliance, 2005\*

\* The composition of the alliance experienced minor changes over the years, however, both its composition and its inclusionary-progressive orientation remained largely stable between 2003 and 2015.

This resulted in an enormous increase of pension coverage among the lower popular classes. Pension coverage among elderly persons living in households of the lowest income quintile increased from 42% in 2004 to 85% in 2010, and for the second lowest income quintile the respective increase was from 66% to 95% (Rofman and Olivieri 2011, 49). Furthermore, the program had a profoundly egalitarian impact regarding gender and migratory background, because both women and migrants were much more often unable to reach 30 years of contribution and were thus disproportionally beneficiaries of the program (Rofman and Olivieri 2011, 45; Sala 2014, 9). The major limitation of the program consisted of the fact that it was designed as a temporary measure and not as permanent change of the eligibility criteria. Nevertheless, it resulted in an enormous improvement of social protection for low-income earners. In

<sup>166</sup> The discounted amount depended on the number of missing years of contribution and averaged approximately 40% (Arza 2009, 12).

political terms, the measure contributed significantly to the consolidation of majoritarian support from the lower popular classes, the unemployed workers' movements, and the CTA (Interview D'Elía 2014; Interview Fuentes 2014; Interview Larisgoitia 2014).

By 2007, the pension inclusion program made up about 30% of the total pension expenditure (Arza 2009, 12). To finance these costs and to strengthen the public pension system, the government undertook further highly popular measures, which received explicit support from most trade unions and political parties (Arza 2012a, 49; Garay 2010, 84). Law 26.222 allowed affiliates of the private pension system to switch to the public system, established the rule that new contributors would be automatically affiliated with the public instead of the private system, and significantly increased the replacement rate.167 Affiliates of the private system within 10 years of retirement age who had accumulated only low amounts on their individual pension capitalization accounts were automatically transferred to the public system. Decree 313/07 furthermore increased the contribution rate for the private pension system from 5% to 11% and thereby eliminated an important financial incentive to opt for the private system. This led not only to a significant number of additional contributors in the public system but also provided access to funds that hitherto privately insured contributors had accumulated on their pension accounts.

The massive increase in pension coverage also had positive effects regarding health care coverage, as approximately two and a half million new pensioners were now also covered by the pensioners' health insurance (Arza 2012b, 36). Furthermore, in May 2004 the government introduced the Federal Health Program. A central aim was to strengthen the universal public health sector by improving low-income earners' access to preventive treatments. In order to do so the number and quality of primary health care centers and the distribution of free pharmaceuticals were extended (Cohen 2009, 30; Golbert 2012, 145). Another measure in this framework was the implementation of the Plan Nacer, which aimed at reducing child and mother mortality by providing health insurance coverage to uninsured pregnant women and children up to the age of six in nine northern provinces (Golbert 2010, 151). By 2007, the coverage of Plan Nacer had reached 450,000 and child mortality had fallen by one quarter in the provinces where it was implemented (McGuire 2010, 125–142).

All in all, the coverage and quality of health care for low-income earners improved significantly during Kirchner's presidency. The expansion of pension coverage and formal employment led to an important increase in health insurance coverage, and the government also increased the resources available for the universal public health system (MSN 2010). In real terms, the expenditure per person covered only by the public sector approximately doubled

<sup>167</sup> For an average wage earner with 30 years of contributions, who retired in 2010, Law 26.222 meant an increase of over 10 percentage points of the replacement rate (author's calculations).

between 2004 and 2007.<sup>168</sup> However, in contrast to previous progressive health care reform efforts of the 1940s, 1970s and 1980s, the government did not attempt to overcome the systemic fragmentation and inefficiency of the system, which continued to be divided between a universal public sector, approximately 300 social health insurance funds and a wide range of private insurers (Maceira 2009). This has much to do with the fact that all these previous attempts to enact integral reform failed due to significant resistance. In interviews, health experts from within the governing alliance assured that they still preferred a unified public system. However, for political and strategic reasons they refrained from proposing integral reform in order to avoid conflicts with the trade union wing of the governing alliance (Interview Gago 2014; Interview Trotta 2014).

Social assistance policies continued to be crucial for mobilizing lower popular class support during Kirchner's presidency. However, while workfare programs had been the main tool during the previous governments, Kirchner shifted the focus towards housing and infrastructure programs and social assistance pensions.<sup>169</sup> Although the protest wave by unemployed workers' movements calmed down significantly with the assumption of the new government, pressure from below nevertheless remained significant with over 1,200 roadblocks in 2003 (Massetti 2006, 30). Furthermore, several unemployed workers' movements made publicly clear that they would return to the streets if it turned out that the government would not implement policies in the interest of the lower popular classes (*Página 12*, August 3, 2003). In this context and with the aim of winning the unemployed workers' movements for the governing alliance, the government introduced the program Manos a la Obra in August 2003, which provided subsidies to worker cooperatives of municipalities and 17 different unemployed workers' organizations (Natalucci 2012a, 143). Although not all of these organizations eventually joined the governing alliance, several of them did. Those movements furthermore benefited particularly well from the following expansion of housing and infrastructure programs, which created both better living conditions and jobs for many residents of poor neighborhoods (*Página 12*, August 3, 2003; Svampa 2008c, 22–25; Interview D'Elía 2014). Between 2004 and 2007 expenditure as a percentage of GDP on housing, running water, and sewage nearly doubled from 0.77% to 1.32% (SPM 2016). This provided the basis for the construction of over 400,000

<sup>168</sup> Aurthor's calculations based on DAGPyPS (2011) for expenditure data and MSN (2010) for the number of covered persons.

<sup>169</sup> In contrast to the extension of pension coverage among the elderly through the Pension Inclusion Program, social assistance pensions were usually provided to beneficiaries in working age, so that in practice their social function was to provide protection against risks such as unemployment, poverty and invalidity. In political terms, the distribution of social assistance pensions was often clientelistic and served to mobilize lower popular class support.

houses for low-income families and a significant extension of basic services and public schools in poor areas (Garay 2010, 73).

The government also increased the number of conditional cash transfers without work requirements. The main measure to do so was the expansion of non-contributory invalidity and motherhood pensions. As a result, the overall number of non-contributory pensions nearly doubled from 345,000 in 2003 to 625,000 in 2007 (MTEySS 2017, 49). Compared to other social assistance programs, these benefits were relatively high, constituting between 70% and 100% of the minimum pension and increasing with each adjustment of the latter. In 2004, the government furthermore implemented the Program for the Social Inclusion of Families,<sup>170</sup> which was mainly thought of as a replacement for the PJJHD without work requirements for women with children under the age of 19 (Zaga Szenker 2009, 4–11). In 2007 the program reached 542,000 beneficiaries.171 Along the same lines, a conditional cash transfer linked to a training program called Employment and Training Insurance was established in 2006, which enrolled only very few beneficiaries, however (Rodríguez Enríquez and Reyes 2006, 28–29). In sum, expenditure as a percentage of the GDP on social assistance policies including housing and non-contributory pensions increased between 2003 and 2007, which represented a significant improvement in social protection for low-income earners if we take into consideration that the number of unemployed and poor persons fell by nearly one half during the same period (Gasparini and Cruces 2010, 5; MPyT 2019; SPM 2016).

The rapid expansion of social protection for low-income earners was furthermore reinforced by two additional factors. First, the development of significant associational power resources of the lower popular classes not only increased their capacity to influence the agenda-setting of the policy process but also their capacity to ensure the effective implementation of these policies. In many cases unemployed workers' movements informed the residents of poor neighborhoods about new social programs, helped with the enrollment process, and provided legal assistance, thereby contributing to the fact that even quite complicated policies such as the Pension Inclusion Plan rapidly expanded among low-income earners (Almonacid 2013, 245). And second, since the late 1990s, an alliance of the CTA with several migrants' organizations and progressive politicians from different parties had promoted a new migratory law. With the support of the government, this initiative finally led to a new law in 2003, which explicitly stated that the government is responsible to ensure the same social rights to migrants as to Argentine citizens and that under no circumstances could health care can be denied (Caggiano 2011, 8–9; Domenech 2013, 126–127; Law 25.871). The new migratory law furthermore

<sup>170</sup> Programa Familias por la Inclusión Social.

<sup>171</sup> http://www.desarrollosocial.gov.ar/Planes/PF/resena.asp. Accessed on November 15, 2010.

permitted the legalization of many migrants' migratory status, which facilitated the formalization of working relations and the enrollment in social programs. Hence, in several ways, the new legislation contributed to an important expansion of social protections for migrants, who were disproportionally often low-income earners (Bertranou 2014, 105; Sala 2014, 18).

## **Social Policy Expansion, Popular Class Mobilization, and the Growing Intensity of Political Conflicts after the Economic Boom, 2007–2011**

After four years of massive social policy expansion, high real wage increases, and an average economic growth rate of nearly 9%, the governing alliance enjoyed broad-based public approval (Levitsky and Murillo 2008, 16–19). Its activist base among the popular classes had been successfully consolidated through the incorporation of the majority of the trade union movement and several of the strongest unemployed workers' and human rights movements. In view of the 2007 presidential and legislative elections, the governing alliance therefore aimed at increasing its support among the middle classes. In order to do so, Cristina Fernández de Kirchner, the President's wife and now presidential candidate, forged an alliance with several leaders of the center-left wing of the oppositional UCR party, which traditionally captured an important part of the middle class vote (Zelaznik 2011, 99). Furthermore, two small center-left parties, Libres del Sur and the Partido Intransigente (PI), joined the alliance. On this basis, the governing alliance captured 45% of the votes and emerged as the dominant political force from the election, controlling absolute majorities in the Chamber of Deputies and the Senate and three-quarters of Argentina's 23 governorships (Levitsky and Murillo 2008, 16). Also, for the first time, five unemployed workers' leaders were elected to Chamber of Deputies (Mauro and Rossi 2011, 170).

The picture soon began to change, however. A first fissure in the unity of the governing alliance appeared when the ex-President Néstor Kirchner revealed his intention to preside over the PJ party. Several unemployed workers' movements in the governing alliance had emerged during the 1990s as a leftwing opposition to the PJ and now feared that the party would assume a dominant role within the alliance. Although the two biggest unemployed workers' movements, the FTV and the Movimiento Evita, finally accepted Kirchner's intention to preside the PJ, another big organization, Barrios de Pie, decided to leave the governing alliance in 2008 (Boyanovsky Bazán 2010, 32–34; Mauro and Rossi 2011, 176). Tensions arose moreover in relations between the government and a part of the trade union movement. In July 2008, the hotel and restaurant workers' union under the lead of the right-wing Peronist Luis Barrionuevo and 56 smaller unions split off from the CGT to form the CGT Azul y Blanca, which assumed a distanced relationship with the government (*El*  *Clarín*, July 8, 2008).<sup>172</sup> Two years later, after divisive internal elections, the CTA also split into two, the CTA de los Trabajadores (CTA-T) and the CTA Autónoma (CTA-A), the latter adopting a combative oppositional stance (Interview Fuentes 2014; Retamozo and Morris 2015, 76).<sup>173</sup> Although the two majoritarian trade union confederations remained in the governing alliance, the emergence of two opposition confederations clearly weakened its support base among formal labor.

Another challenge for the government emerged from a deterioration of the economic context. From 2007 on, inflation returned to values over 20% and became a constant source of criticism from the media and the opposition.<sup>174</sup> Furthermore, due to the global financial crisis, economic growth slowed down significantly in 2008 and Argentina entered into recession in 2009. The fiscal surplus gave way to a growing deficit and the government decided to address this situation with an increase in export taxes on soybeans and other agricultural products (Bonvecchi 2011, 144–150). The measure was strongly objected to by the four major agricultural employers' associations, which immediately formed an alliance to organize resistance. This unity put the employer sector in a powerful position given that, taken together, they not only united enormous economic resources but also several hundred thousand members with high lobbying capacities and political influence in nearly all political parties and major media companies. On this basis, the agricultural employers' associations developed a broad-based campaign that involved extensive presence in the media with the aim of winning public support, close cooperation with the political opposition, intensive lobbying with the aim of promoting divisions in the governing alliance, and massive protest actions and roadblocks aiming at disturbing the supply of certain types of food to the cities (Bonvecchi 2011, 149; Mauro and Rossi 2011, 175). The government, in turn, communicated that the revenue from the tax increase was going to be used for the construction of public hospitals, housing for low-income earners, and better infrastructure in impoverished rural areas (Lodola 2011, 220). At the same time, it mobilized allied trade unions and unemployed workers' movements to engage in mass

<sup>172</sup> During the 1990s Barrionuevo had been one of the strongest supporters of Carlos Menem's neoliberal government. Besides political differences with the leadership of the CGT and the government, the principal motivation for the split was that Barrionuevo aspired becoming secretary general of the CGT yet lacked the necessary support for being elected (*El Clarín*, July 8, 2008).

<sup>173</sup> An important reason for this split was that a part of the CTA was strongly disappointed with the government's decision to not officially recognize the CTA in order to avoid conflicts with the CGT (Interview Fuentes 2014). Furthermore, several of the leaders of the newly created CTA-A, such as Claudio Lozano and Víctor De Gennaro, were at the same time leaders of a minor left-wing opposition party that electorally competed with the governing alliance.

<sup>174</sup> Additionally, the government pressured the national statistics institute INDEC to lower the official numbers on inflation, which became another important focus of criticism.

protests in favor of the measure. The powerful teamsters' union helped to ensure food supply to the cities. After four months of intensive conflict, the government finally decided to try to legitimize the resolution that raised the export taxes by converting it into a law bill and approving it in the parliament. The bill passed the Chamber of Deputies, but the fact that 19 deputies of the governing alliance voted against it, showed that the agricultural employer associations had managed to drive a dividing wedge into the alliance's unity. One day before the treatment of the bill in the Senate on the 16th of July 2008, the agricultural employers' associations, as well as the government, mobilized again several hundred thousand protestors against and for the passing of the bill. Finally, the vote in the Senate ended in a 36-to-36 standoff, so that Julio Cobos, the Vice-President of the government and president of the Senate, was enabled to vote. Cobos voted against the bill, which meant that it failed to pass the parliament. The government finally gave in and revoked the export tax increase (Garay 2010, 74; Mauro and Rossi 2011, 175; Serrafero 2011, 30–32).

This defeat significantly damaged the governing alliance's standing in public opinion and led to the desertion of numerous deputies, senators, governors, and mayors. Most of these officials represented centrist and conservative political currents and had joined the alliance in order to improve their access to funds distributed by the national government and because it had appeared politically opportune during the years in which the government enjoyed high popularity. Furthermore, the fight over the export tax increase inaugurated a much more hostile relationship with the biggest media companies and the employers' associations of the agricultural sector, which had initially benefited from Néstor Kirchner's exchange rate policy (Varesi 2011, 44; Zelaznik 2011, 102–103).

The government responded to these developments with the adoption of a more confrontational course towards conservative political and economic sectors and a reinforcement of its efforts to mobilize popular class support (Varesi 2011, 44; Zelaznik 2011, 102–103). Pension policies played again a crucial role in this strategy. In late July 2008, President Cristina Kirchner officially announced that the government was preparing a bill to introduce an automatic mechanism for pension benefit indexation (*La Nación*, July 28, 2008). Thereby she took up a long-standing demand of all trade union confederations, many pensioners' associations, and several opposition parties (*Página 12*, August 13, 2006; *Página 12,* November 15, 2006; Smulovitz 2008, 297).<sup>175</sup> By the 1st of October the bill had successfully passed both chambers with large majorities. On the 21st of October, the government sent a second bill to the Congress, which aimed at the complete renationalization of the pension system under favorable terms for those previously affiliated to the private system (Mesa-Lago

<sup>175</sup> In simplified terms, the new indexation mechanism was based on two main variables: the evolution of the system's fiscal revenue per beneficiary and the evolution of wages (Arza 2009, 16–17).

2009, 18–19). This measure was highly welcomed by the trade union confederations and unemployed workers' movements. The major employer associations and IFIs rejected the measure but refrained from further confrontation in the face of approximately 90% public approval for the measure. Finally, the bill passed the parliament with ample support even from a wide range of opposition deputies and senators (Arza 2012a, 49–56; Bonvecchi 2011, 150).

The nationalization of the pension system was not only highly popular it also provided the government with abundant additional fiscal resources. The incorporation of millions of contributors helped to finance the massive expansion of coverage among low-income earners and women as well as the new pension indexation mechanism.176 Furthermore, the nationalization of the assets accumulated in the individual pension capitalization accounts valuing approximately 9% of GDP) could be used by the government to confront the economic crisis with countercyclical measures and further social policy expansion in other areas (Arza 2009, 4; *El Clarín*, November 30, 2008).<sup>177</sup>

Among those countercyclical measures was massive government investment in the universal public health sector, the construction of social housing, and the extension of running water and sewage in poor neighborhoods. In each of these areas public spending as a percentage of the GDP increased by approximately 20% between 2008 and 2009 (SPM 2016). The Plan Nacer, which provided basic health insurance to uninsured low-income mothers and children up to the age of six, was extended to all Argentine provinces and by 2010 reached 1,500,000 beneficiaries (Golbert 2012, 146).<sup>178</sup> The government also increased expenditures on social assistance policies. In May 2008, it created a new program called Jóvenes con Más y Mejor Trabajo which provided a monthly allowance to support unemployed persons between 18 and 24 who lacked diplomas to attend school or participate in professional training. In 2009, the program already had over 72,000 beneficiaries and by 2011 this number had risen to approximately 150,000 (Brown 2012, 21–24). In early 2009, the government introduced another new social assistance program in collaboration with one of the biggest unemployed workers' movements, the Movimiento Evita. In order to create jobs and to improve public infrastructure in poor neighborhoods in greater Buenos Aires, the Programa de Inversión Social (PRIS) provided subsidies for the creation of up to 1,000 cooperatives with approximately 15,000 workers (Natalucci 2012, 143). Furthermore, the

<sup>176</sup> In large part due to these measures, pension expenditure skyrocketed from 5.6% to 11.4% of GDP between 2005 and 2015.

<sup>177</sup> The government furthermore used the nationalized shares of companies to obtain company internal information in order to reduce tax evasion and to influence company decisions. This, in turn, led to intense conflicts with several companies, including the country's biggest industrial company, Techint (Bonvecchi 2011, 150).

<sup>178</sup> In 2007 the Plan Nacer had covered nine provinces and approximately 450,000 beneficiaries (McGuire 2010: 125–142).

expansion of social assistance pensions was significantly accelerated. Between 2008 and 2009 alone the number of non-contributory pensions increased from 720,000 to over 923,000. In 2011, it reached 1,195,000 (MTEySS 2017, 49).

Despite these efforts to mobilize popular class support, the governing alliance was not able to recover sufficiently quickly from the costly dispute over export taxes in 2008. In the context of economic crisis and increasing criticism from the mainstream media, the legislative elections of June 2009 resulted in a clear-cut defeat and the governing alliance lost its majorities in both chambers (Jones and Micozzi 2011, 59–61). Again, the response of the government was to increase confrontation with the opposition and to deepen the redistributive agenda. Two months after the elections, the government sent a bill to the Congress, which aimed at replacing the media law implemented by the last military dictatorship. Among others, the bill foresaw limitations regarding the concentration of media in the hands of a few private companies. This, not surprisingly, further exacerbated the already existing conflict between the most powerful media companies and the government. With the support of several center-left and left deputies and senators, the bill successfully passed both chambers (Kitzberger 2011, 186–187). At the same time, the government implemented a new social assistance policy called Ingreso Social con Trabajo,<sup>179</sup> which massively extended the creation of subsidized worker cooperatives in poor neighborhoods. In 2011, the program had nearly 160,000 beneficiaries with monthly wage subsidies of 1,200 pesos and health insurance coverage (*La Nación,* November 30, 2014). However, because most of the subsidies were channeled to mayors and unemployed workers' movements within the governing alliance, the program quickly gave rise to repeated and massive waves of roadblocks by opposition social movements and heavy criticism from opposition parties (Guimenez and Hopp 2011, 5–6). These protests, in turn, provoked widespread negative media coverage for the government, which was accused of using the program for clientelistic purposes.<sup>180</sup> Furthermore, the selective distribution of subsidies led to considerable tensions among social movements supporting the government. The two social movements most favored in the disbursement of subsidies were the Movimiento Evita, which was able to register approximately 15,000 members in the program, and the Frente Transversal, which was able to register over 5,000 members (*La Nación*, February 14, 2010). The more autonomous FTV, for its part, received considerably fewer subsidies and therefore started to publicly threaten with roadblocks. In order to avoid further desertions from the governing alliance, additional subsidies were finally granted to cooperatives led by the FTV (Interview D'Elía 2014). All in all, the program had ambivalent results for the government. On the one

<sup>179</sup> Popularly known as Programa Argentina Trabaja.

<sup>180</sup> The conflict was virtually omnipresent in the media, the topic of hundreds of newspaper articles. Here only three examples shall be mentioned: *La Nación*, December 30, 2009; *Página 12*, November 4, 2009; and *El Clarín*, November 9, 2009.

hand, it provided significant resources that helped allies to strengthen their base-level networks, to create jobs, and to improve infrastructure in poor neighborhoods.<sup>181</sup> On the other hand, the political conflicts associated with the clientelistic distribution of benefits damaged the image of the government and indicated that the policy was hardly able to generate broad-based electoral support.

Once again, the government decided to address the situation by taking up an existing and increasingly popular social policy demand: the introduction of a universal child allowance (Bossert 2016, 120–121). Already during the economic crisis in the early 2000s a wide range of progressive trade unions, social movements, and politicians had demanded a similar measure through the FRENAPO alliance. From 2008 on, the demand gained new momentum within both the opposition and the governing alliance (Interview Linares 2014; Interview Lozano 2014). In 2008 and 2009 alone, five bills were introduced in the Congress that proposed an extension of the child allowance system to children of informal or unemployed workers. One of them was even promoted by a deputy of the governing alliance (Hintze and Costa 2011, 181). Against this background, Cristina Kirchner issued a presidential decree in October 2009 which implemented the Asignación Universal por Hijo (AUH) and extended the child allowance system to approximately 3.5 million informal and unemployed workers. It provided a monthly payment of 180 pesos per child on the condition that school-aged children attend school and comply with a plan of vaccinations. Although the measure replaced both the PJJHD and the Programa Familias por la Inclusión Social, it represented a massive improvement regarding both the coverage and the value of the cash transfers for low-income households. Furthermore, in contrast to the partly clientelistic programs it replaced, it constituted a genuine social right (Gasparini and Cruces 2010, 15–16). In March 2011, only a few months before the next national elections, the government extended the benefit to pregnant women, which increased coverage by approximately another 180,000 beneficiaries (*Página 12*, March 2, 2011).

## **Resurgence and Decline of the Inclusionary Governing Alliance, 2011–2015**

After the clear-cut defeat in the 2009 mid-term elections, the governing alliance found itself in a severe crisis. Soon after, however, the alliance experienced a stunning resurgence. One important factor was that the economic crisis was quickly overcome. Anti-cyclical policies contributed to maintaining low unemployment levels and to a significant increase in real wages during the

<sup>181</sup> In a qualitative study, Zaragoza (2015, 348) revealed that most of these jobs in practice were a precarious form of (public) employment that did not really comply with the idea of autonomous cooperatives. Interviews with former cooperative workers point in the same direction (Interview Córdoba 2014; Interview Falcón 2014).

short period of recession. In 2010 and 2011, fostered by demand-oriented policies on the national level and a combination of low interest rates and rising commodity prices on the international level, Argentina returned to extraordinarily high growth rates (Gerchunoff and Kacef 2016, 21; Trujillo 2017, 15).<sup>182</sup> Thus, in contrast to the 2009 mid-term elections, the 2011 elections took part in a highly favorable economic context.

The public perception of progress was furthermore reinforced by the deepening of the government's redistributive social policy agenda. Between 2009 and 2011 alone, the lowest income decile's share of the national income rose by approximately one third (CEDLAS and World Bank 2018).<sup>183</sup> In particular, the universalization of the child allowance system was extraordinarily popular and became a crucial component of the electoral campaign. This and other policies, such as the free public broadcasting of football, became effective symbols of the government's commitment to both economic growth and social inclusion (Cherny 2011, 137; Zarazaga 2015, 353–354).

Despite the massive conflicts associated with the introduction of the cooperative program Ingreso Social con Trabajo, this measure also fulfilled an important function in the political resurgence. Rather than being suitable for mass communication, it contributed to a broadening and disciplining of the fragmented activist base of the governing alliance. As the distribution of subsidies was largely discretionary and controlled by the national government, it provided important leverage for carrot-and-stick tactics directed at mayors and governors. Those who decided to support the governing alliance were granted with numerous subsidies, nurturing thereby electorally important clientelistic networks in poor neighborhoods. Those mayors and governors who did not cooperate received very few subsidies. In their districts, the distribution of subsidies and the creation of cooperatives was then often carried out by social and political movements of the governing alliance. This contributed decisively to the alignment of local politicians and to a rapid expansion of the base-level structures of completely loyal organizations, such as the Movimiento Evita, the Frente Transversal, La Cámpora, and Kolina. All of these organizations had been created in close cooperation with the government and Kolina was even led by the sister-in-law of the President (*La Nación*, March 10, 2013; *La Nación*, November 30, 2014; Zarazaga 2015, 347–348).

These factors together with the fragmentation of the opposition and a wave of compassion and public attention aroused by the sudden death of ex-President Néstor Kirchner in October 2010 led to a clear-cut electoral victory for the governing alliance, which won the presidency with 54% of the vote and recovered majorities in both chambers (Cherny 2011, 137).

<sup>182</sup> In 2010 9.3% and in 2011 6.5% (Trujillo 2017, 15).

<sup>183</sup> Between 2009 and 2011 the Gini coefficient fell from 0.444 to 0.419 and the lowest income decile's share of the national income rose from 1.5% to 2.0% (CEDLAS and World Bank 2018).

However, the resurging unity and strength of the governing alliance was only temporary. Since at least 2011, there had been internal discussions between Hugo Moyano, the general secretary of the CGT, and Cristina Kirchner about the representation of the trade union movement on electoral lists and in the cabinet. While Moyano considered the status quo to be insufficient, Cristina Kirchner was much more inclined to provide influential positions to more progressive and less autonomous organizations, such as La Cámpora, than to the CGT. Furthermore, Moyano demanded the elimination of income taxes for wage earners, which the government rejected. In turn, the government supported the candidacy of Antonio Caló, leader of the metal workers' union, for general secretary of the CGT. In June 2012, the dispute eventually escalated and Hugo Moyano publicly broke with the governing alliance, which led to a further split of the trade union movement. While the majority of the trade unions remained in the governing alliance and formed the CGT-Oficialista under the lead of Antonio Caló, a minority faction under the lead of Hugo Moyano created the CGT-Opositora (*Perfil*, October 3, 2012; Schmall 2012, 93). This meant that by 2012 the trade union movement was divided into five confederations, of which only the two biggest adhered to the governing alliance.

In addition to these internal disputes, the economic context deteriorated sharply. In 2012 the GDP fell by 0.5% and in 2013 growth remained moderate at 2.3% (Trujillo 2017, 15–16). Due to an increase in capital flight and the relative appreciation of the Argentine peso during the preceding years, the foreign currency reserves of the central bank were decreasing rapidly. To deal with this situation the government undertook several measures: tight restrictions on currency exchange, import restrictions on a broad range of products, and the nationalization of the YPF oil company with the aim of increasing oil production and therefore reducing dependency on energy imports (Gerchunoff and Kacef 2016, 21–27; Vilas 2017, 51). Restrictions on the capacity to buy US dollars and certain imported goods caused significant discontent among important sections of the middle and upper classes. In combination with high inflation, a growing perception of insecurity, and various allegations of corruption, this led to the emergence of middle- and upper-class demonstrations throughout the country in 2012 and 2013 (Gómez 2014, 96–98). At the same time, the opposition trade union confederations organized several protests calling for the derogation of wage taxes and an increase in the minimum pension to 82% of the minimum wage, so that taken together the government was faced with a significant rise in protest activity (Trujillo 2017, 14).

As in 2009, the government responded to the economic cool down with demand-oriented measures and to the political conflict with redistributive policies that aimed at reinforcing its alliance with the lower popular classes and loyal organizations. In June 2012, the government implemented a new housing program called Pro.Cre.Ar, which provided highly subsidized credits to lowand middle-income earners (Decree 902/2012). Subsequently, expenditure on housing policy increased steeply from 0.75% of GDP in 2012 to 1.19% in 2015. Thereby, the government injected money into the economy to prevent a crisis in the construction sector, reinforced its alliance with the powerful construction workers' union and redistributed resources to low- and middle-income families (Trujillo 2017, 15). In order to avoid negative income effects of the economic cool down, it furthermore extended social assistance policies for low-income earners of working age. In March 2013, the government implemented a program called Ellas Hacen. The policy was designed in collaboration with the national women's council<sup>184</sup> and addressed issues of gendered violence and economic emancipation, which the increasingly strong Argentine women's movement had brought to broad public attention (Hauría and Mendoza 2017, 4). Through the program approximately 88,000 jobs in subsidized cooperatives were created for women in difficult economic situations. Women who had experienced gendered violence were given priority (*La Nación*, November 30, 2014). In addition, the government increased the number of social assistance invalidity pensions by over 250.000 between 2011 and 2013. The overall number of non-contributory pensions reached 1,450,000 in 2013 and 1,515,000 in 2015 (MTEySS 2017, 49).

In the area of health care policies, the government continued the expansion of the Plan Nacer. The number of mothers and children receiving basic health insurance coverage through the program increased from 1.5 million in 2010 to 4.7 million in 2012 (*Página 12*, March 25, 2014). In August 2012, the policy was renamed Programa Sumar and coverage was extended to uninsured women up to the age of 64 and adolescents up to the age of 19, so that by late 2015 the number of beneficiaries rose to nearly 13 million (MSN 2016, 4; *Página 12*; August 7, 2012).

Another measure that improved social protections for low-income earners was the sanctioning of the Law 26.844 in March 2013, which improved labor regulations for domestic workers by reducing the working day from 12 to 8 hours, promoting the formalization of working relations, and including domestic workers in the family allowance and work accident insurance systems. The sector remained predominantly informal, yet formality increased from 15% to 25% within one year (Pereyra 2017, 29–32).

These efforts to extend social protections and to mobilize popular class support could be considered by and large successful. Notwithstanding, the desertion of an important part of the labor movement and the emergence of massive middle- and upper-class protests meant that the governing alliance faced a complicated environment in the 2013 mid-term elections. Contributing to the challenging conditions, Sergio Massa, one of the conservative Peronist politicians who had left the governing alliance during the preceding years, had

<sup>184</sup> The national women's council is a public entity in Argentina in charge of promoting gender equality. The council traditionally cooperates closely with organizations of the Argentine women's movement.

successfully constructed an oppositional Peronist alliance. This alliance received support from a part of the base-level structures of the PJ party and the opposition trade unions and benefitted from the discontent of important parts of the middle and upper classes. Against this background, the governing alliance won only 34% of the votes, though it was able to retain majorities in both chambers (Pagni 2016, 111).

Once the elections were over, the government decided to tackle the alarming reduction of foreign currency reserves by devaluating the Argentine peso and raising the interest rate. While the measure had little success in augmenting the reserves, it fostered inflation and threw the economy back into recession. In order to avoid any increase of inequality and poverty in consequence of the devaluation, the government immediately enacted several measures to increase social protection for low-income earners (Gerchunoff and Kacef 2016, 30; Vilas 2017, 53). In January 2014 the government announced a new social assistance program called PROGRESAR, which provided an allowance of 600 pesos per month to unemployed or informal workers between 18 and 24 years of age, who participate in a professional training program, finish their school education, or study at a university (Di Giovambattista, Gallo, and Panigo 2014, 32–41). One year later, the allowance was increased to 900 pesos and by late 2015 the program reached over 900,000 beneficiaries (*Página 12*, July 21, 2017; Vilas 2017, 53). In June 2014, Cristina Kirchner furthermore announced a renovation of the Pension Inclusion Plan, which until 2017 allowed over 800,000185 additional low-income earners at retirement age to receive a contributory pension benefit. An important difference to the Pension Inclusion Plan of 2005 was that this time enrollment in the program was explicitly restricted to low-income households (*La Nación*, June 4, 2014; *La Nación*, August 14, 2014). At the same time, the government introduced tighter controls and higher penalties for companies evading social insurance contributions and increased expenditure on the universal public health system, passing from 2.67% of GDP in 2013 to over 3% in 2015 (Law 26.940; SPM 2016). And finally, four months before the presidential elections of 2015, the government announced the extension of the automatic indexation mechanism for pensions to the family allowance system (*Página 12*, June 16, 2015).

All in all, the expansion of social protections to low-income earners continued between 2011 and 2015, despite the economic recession, falling commodity prices, and increasing political conflicts.

#### **The End of the Inclusionary Governing Alliance**

The general elections of 2015 took place in a complicated economic context: declining commodity prices, rising fiscal deficit, slow economic growth, and

<sup>185</sup> Calculations of the author based on MTEySS (2017, 47).

high inflation (Gerchunoff and Kacef 2016, 31). Although the government's demand-oriented economic and social policies were relatively effective in protecting the purchasing power of most low- and middle-income households, discontent over the general state of the economy was growing, especially in the middle and upper classes. This discontent was additionally enhanced by a range of allegations of corruption against high ranking government officials (Manzetti 2014, 185–192). Another scandal erupted when in early 2015 federal prosecutor Alberto Nisman was found dead in his apartment right before he was to speak before the Congress where he wanted to make several accusations against President Cristina Kirchner. Although the circumstances of his death remained unclear, the case evolved into an omnipresent issue in the media with negative impacts on the governing alliance's electoral outlook (Murillo, Rubio, and Mangonnet 2016, 5).

At the same time, the opposition converged in two main electoral alliances, UNA and Cambiemos, and thereby overcame the fragmentation that had decisively contributed to their electoral defeat in 2011. The first was led by the conservative Peronist mayor Sergio Massa and the second by the neoliberal businessman and mayor of Buenos Aires Mauricio Macri. Due to the employer-friendly platform of Mauricio Macri's alliance and his extensive network within the economic elite, Cambiemos was able to raise approximately 70% more campaign funds than the governing alliance (Tagina 2016, 238). While the opposition was increasingly united, the governing alliance engaged in long internal arguments over the question of who would run as its candidate (Pagni 2016, 110). Finally, it was decided that Daniel Scioli, governor of the province of Buenos Aires, would be the candidate. However, the decision remained controversial and not all sectors of the alliance were willing to support him energetically (Grimson 2016, 18). The electoral campaign of the governing alliance hence experienced several difficulties.

At the same time, Macri was able to reach out to new groups of voters. By assuring that his government would maintain the existing redistributive social programs, he deliberately addressed a broad group of voters who recognized the social progress made under the Kirchner governments but disliked their confrontational style (Tagina 2016, 235).<sup>186</sup>

Against this background, the presidential elections of 2015 finally ended in a narrow electoral victory for Cambiemos, and Mauricio Macri became

<sup>186</sup> This promise to maintain the existing social protections for low-income earners was clearly a recognition of the societal distribution of power resources and the acknowledgement that a confrontation with the beneficiaries of these programs and the social movements of the lower popular classes would result in a significant loss of votes. While the actual preference of Macri's party was to cut most of the social assistance programs and to replace them with "incentives" for employers to create jobs (Interview Schmidt-Liermann 2014), the 2015 presidential campaign promised to complement one with another.

president. Right from the beginning, a range of influential employers was included in the core of the new governing alliance. Ten businessmen became ministers in Macri's cabinet, the big private media companies immediately changed their relationship with the government from confrontation to thinly concealed support and on the global level the international financial institutions welcomed both his election and his political agenda (Lijalad 2016, 8; Pagni 2016, 113). On this basis, the new governing alliance initiated a period of regressive social policy reform. As a consequence, inequality, unemployment, and informal salaried employment began to rise again (CEDLAS and World Bank 2018).

## **Political Dynamics, Patterns of Social Policy Change, and Path Dependency Effects**

By the early 2000s, the distribution of power resources had shifted in favor of the popular classes. Progressive actors had increasingly gained associational power resources, intensified the coordination with each other and forged alliances with middle-class and small business sectors. The discrediting of neoliberalism and the resurging popularity of a state interventionist economic paradigm had furthermore shifted significant discursive power resources towards more progressive actors. Unemployed workers' movements and the CTA effectively mobilized these power resources for the extension of social protections for low-income earners. Against this background, the economic crisis and the fall of the neoliberal government of President De la Rúa in late 2001 initiated a profound shift in social policy. The following interim President Eduardo Duhalde acknowledged the increasing power of these new actors, received them as legitimate negotiation partners and responded to their demands and protest actions with a massive expansion of social assistance policies. In 2003 the center-left electoral alliance of Néstor Kirchner won the presidential elections and deliberately pursued the integration of the new progressive actors into the core of its support base. Backed by a series of redistributive labor and social policies, Kirchner was eventually able to form a solid and inclusionary governing alliance that incorporated trade unions, unemployed workers organizations, the PJ party, smaller center-left parties and even some business actors. Labor and the lower popular classes soon became the electoral backbone of the alliance.

On this basis, the new government massively expanded public expenditure from 29% of GDP in 2003 to 47% in 2015, pursued demand-oriented economic policies, reintroduced labor market and trade regulations and renationalized several privatized companies (DAGPyPS 2011; SPM 2016). This general intensification of state intervention involved at its core a massive expansion and reform of social policies, which strongly improved social protections for lowincome earners regarding coverage, quality, and value of the benefits. Overall social spending increased from slightly over 19% of GDP in 2003 to nearly 31% in 2015. Access to contributory old-age pensions was nearly universalized and the real value of the minimum pension was massively increased. Family allowances were extended to children of informal, unemployed, and domestic workers and their real value was raised repeatedly. After a tenfold increase in coverage, non-contributory invalidity and motherhood pensions provided basic incomes to approximately 1.4 million poor persons of working age, while further programs provided allowances to workers in cooperatives and students living in low-income households.


Table 30: Change in social protection for low-income earners, 2002–2015

The universal public health sector, which served mostly low-income groups, was strengthened and expenditure increased from less than 1.9% of GDP in 2002 to over 3% in 2015 (SPM 2016). In addition, the Plan Sumar provided basic health insurance to nearly 13 million uninsured women, children, and adolescents.

Taken together these measures contributed to a significant improvement of social indicators. Between 2003 and 2015, life expectancy rose by over two years and infant mortality fell by 37%, reaching a value of 10.3 per 1000 (World Bank 2018).

Income inequality measured with the Gini coefficient decreased from 0.53 in 2003 to 0.41 in 2015 (CEDLAS and World Bank 2018). In the same period, the lowest income decile's share of the national income increased by 73%, while the richest decile's share fell by 30%. According to calculations of Lustig and Pessino (2014, 321) and Rofman and Apella (2016, 103), the expansion of social protections accounted for a major share of this improvement. Unemployment fell from 16% in 2003 to under 7% in 2015 and informal salaried employment from 49% to 33% (CEDLAS and World Bank 2018).

Figure 12: Evolution of income inequality measured with the Gini-coefficient, 2002–2015

Sources: CEDLAS and World Bank (2018).

While pension and family allowance policies underwent important structural changes, health care policy was shaped by path dependency. Progressive health experts of the governing alliance consciously refrained from proposing structural health care reform in order to avoid internal conflicts with the union movement. Nonetheless, by increasing expenditure on the universal public health sector and by expanding primary health care programs, the government used the room to maneuver within the existing structure to improve the protection of low-income groups. The period also left an important legacy out of which new path dependencies might emerge. The highly popular universalization of family allowances and the introduction of an automatic indexation mechanism for both family allowances and pensions will make it politically costly for future governments to significantly cut these policies. In the face of high inflation rates, former governments carried out retrenchment by simply not adequately adjusting the benefits. Legally binding indexation rules restrict this option. In addition to that, the enormous extension of coverage means that a significantly enlarged share of the electorate benefits from these policies. Another type of path dependency is related to workfare and cooperative programs, which are of crucial interest for unemployed workers' movements and have been fiercely defended in the past. The active support of the government has furthermore contributed to the emergence and growth of additional Kirchnerist social movements, who have a vital interest in defending their control over cooperatives and workfare benefits.

## 5 Analyzing the Structural Bases, Actors and Political Mechanisms of Change

The last chapter has reconstructed the evolution of social protection for lowincome earners in Argentina between 1943 and 2015. To identify and analyze the underlying political and structural factors, the period has been divided into seven sub-periods. The analysis of each subperiod started with an overview of the economic, social, and political context. On this basis, the constellation of actors and the distribution of power resources were identified. This, in turn, served as a basis for the next step, in which the processes of alliance formation and social policy reform were traced during each sub-period.

This structured analysis facilitated the systematic application of two methodological approaches to identify causal factors and mechanisms underlying the extension and retrenchment of social protections for low-income earners. The first methodological approach consisted of a longitudinal comparison of different sub-periods, which revealed correlations between structural and political characteristics of different sub-periods and social policy outcomes. In the following, the main results will be summarized and discussed. In accordance with the theoretical framework, the focus will be on the constellations of actors, the distributions of power resources, the characteristics of governing alliances, and the direction of social policy change.

The second methodological approach applied in this study was process tracing. By this means, I was able to reveal not only whether but also how underlying configurations—particularly the distribution of power resources influenced the formation of governing alliances and subsequent social policy reform processes. Furthermore, process tracing enabled me to reveal the importance of different actors' political strategies for the durability of governing alliances and social policy outcomes. In the following the results from the process tracing analysis will be summarized and merged with those of the subperiod comparison.

## **5.1 Social Policy towards Low-Income Earners during Different Periods: Between Truncated and Inclusionary Social Protection**

Since the mid-2000s, it has become a broad consensus among social policy researchers that Latin American social policy regimes have historically developed in a truncated manner, meaning that they cover only one part of the population, namely formal workers and their families, while they largely exclude unemployed and informal workers and their dependents (Barrientos 2004; De Ferranti et al. 2004; Lindert, Skoufias and Shapiro 2006).187 Barrientos and Santibáñiez (2009) explained the emergence and consolidation of truncated social policy regimes with the ISI development model, which required the political support of state employees and the increasingly powerful formal working class and hence led to the introduction and expansion of social insurance policies that benefited these sectors but excluded the lower popular classes. In this line, it has been argued that during the 2000s an important expansion of social assistance policies and the introduction of universalistic cash transfer policies in several Latin American countries constituted for the first time a significant break with the long-standing history of truncated social policy regimes on the continent (De Ferranti et al. 2004; Lustig, Pessino, and Scott 2014).

In the face of insufficient statistical data to undertake quantitative analyses of the redistributive impacts of Latin American welfare states during the ISI period, the thesis of a historically stable evolution of truncated welfare states in Latin America depended essentially on two assumptions. The first was that social policy expansion during the whole ISI period was based on social insurance and the second was that these social insurance programs consistently excluded informal workers and other low-income groups. Although these assumptions are highly plausible from a theoretical point of view, empirical evidence from the Argentine case suggests that the picture might be much more complex. As can be seen in Table 31, social protection for low-income earners and informal workers indeed varied significantly during and after the ISI model. During some periods, universal and targeted social policies were expanded significantly and access criteria to social insurance programs were loosened to such an extent that informal and unemployed workers were able to receive generous benefits despite lacking formal work contracts and regular social insurance contributions.

<sup>187</sup> Socialist Cuba constitutes an exception in this regard.


Table 31: Patterns of change regarding social protection for low-income earners, 1943–2015

\* During that period the party was named Partido Peronista. Later it was renamed Partido Justicialista (PJ).

The period between 1943 and 1955 was characterized by the rise of Peronism and a massive expansion of social protections for low-income earners. Legal pension coverage was gradually broadened to finally cover all occupational groups. Progressive replacement rates in combination with permissive access criteria furthermore enabled most informal and unemployed workers to receive a generous pension benefit upon reaching retirement age. Health care expansion was also rapid and despite the consolidation of social health insurance programs for certain occupational groups, the vast majority of resources was concentrated on the creation and strengthening of a universal and free public health care system. At the same time, up to 1% of the GDP was devoted to social assistance policies of the Eva Perón Foundation (FEP).

The following period between 1955 and 1973 was characterized by military dictatorships and civilian governments led by the UCR. The latter were de facto heavily limited by military interferences. During this period many of the most progressive policies were significantly cut and most social policies attained a more truncated character. The FEP was dismantled and social assistance policies massively reduced. Social health insurance coverage was extended to all formal workers and contribution rates increased, but the universal and free public health care system was retrenched and therefore most low-income earners and informal workers suffered a significant deterioration of health care coverage. The progressive replacement rate in the pension system was replaced with a uniform rate and the access criteria were restricted in such a way that a large portion of low-income earners were de facto excluded from pension benefits. Although family allowances were introduced and expanded during this period, they were only provided to formal workers so that coverage among low-income earners remained limited.

Between 1973 and 1976, Peronism returned to the government and social policy moved again in a more inclusionary direction. New legislation for the agricultural sector first led to a considerable increase in pension coverage among rural workers. In 1975 a reform of the access criteria finally completely abolished contribution requirements, so that access to pension benefits was de facto universalized among low-income earners and informal workers. Furthermore, disproportional increases in the minimum pension benefit meant that low-income earners could expect relatively generous benefits. The public health care sector was strengthened again and by the mid-1970s it administered 50% of overall public health care spending (Vargas de Flood 2006, 188). In addition, expenditure on social assistance, housing, and family allowances was significantly increased. Once again, the Argentine social policy regime attained an inclusionary rather than truncated character.

In contrast, the military dictatorship between 1976 and 1983 pushed social policy back into a decidedly truncated direction. The value of the minimum pension was drastically reduced and a requirement of 15 years of contributions was introduced. Expenditure on the public health system was cut by 42%, on social assistance by 72%, and on social housing by 25%.<sup>188</sup> Hence, within a few years, social protections for low-income earners and informal workers were reduced to a minimum. At the same time, social insurance expenditure, which was increasingly limited to formal workers due to stricter access criteria, remained largely stable (DAGPyPS 2011; Vargas de Flood 2006, 188).

Between 1983 and 1989 a UCR government attempted to return to a more inclusionary social policy regime but largely failed to do so in the face of a difficult economic and political context. Only social assistance policy was significantly expanded. In contrast, a universalizing health care reform was blocked by political and corporate opposition and the pension system suffered from a severe financial crisis.

Between 1989 and 2001, the Peronist government of Carlos Ménem and the UCR/FREPASO government of Fernando De la Rúa implemented a pro-

<sup>188</sup> Measured as a share of GDP.

foundly neoliberal social policy reform agenda, which deepened the exclusionary and truncated character of the social policy regime. The increase of the pension contribution requirements to 30 years and the partial privatization of the system strongly favored high-income earners and reinforced the exclusion of the vast majority of low-income earners from access to pension benefits. The decentralization of the public health system furthermore exacerbated regional disparities and the deregulation of the social health insurance system led to an increasing concentration of resources on the better-off affiliates. Social assistance policy constituted the only social policy area in which social protections for poor residents were moderately expanded.

Between 2002 and 2015, social policy again moved in a much more progressive direction under the Peronist presidents Eduardo Duhalde, Néstor Kirchner, and Cristina Kirchner. Social assistance and housing policies were massively expanded. The Pension Inclusion Plan nearly universalized access to pension benefits while disproportional increases of the minimum pension benefit led to a distribution of resources within the system that increasingly favored low-income earners. Expenditure on the universal and free public health sector was strongly increased and the provision of free pharmaceuticals expanded. Furthermore, child allowances were extended to cover children of informal and unemployed workers. Taken together, these changes meant that social policy became much more inclusionary.

The comparison of the patterns of change regarding social protections for low-income earners during different periods reveals that there have been indeed surprisingly stark contrasts in the direction of social policy. Each change in direction was followed by high reform intensity; that is, the introduction of a significant number of profound changes within a short period. These findings question the assumption that the development of truncated social policy regimes in Latin America was a continuous process and directly linked to the ISI model.

In order to shed light on the societal configurations that were related to the often abrupt and profound social policy changes between the different periods, the following sections extend the comparison and focus on underlying political, economic and social factors.

## **5.2 Constellations of Actors and Interests in the Argentine Politics of Social Protection for Low-Income Earners**

A crucial part of the empirical study was dedicated to analyzing the role of different actors in the Argentine politics of social protection for low-income earners. The constellations of actors and interests not only proved to be considerably different from that of most of the European experiences that inspired the vast majority of welfare state theories but also varied over time and with regard to different areas of social policy. Many of these differences were related to the specific formation of the Argentine popular classes as political actors, which was heavily shaped by the emergence of Peronism in the mid-20th century. However, a range of other factors also profoundly affected the constellation of actors in Argentine social politics, such as the high level of informal employment and the creation of vested interests through the introduction of union-controlled health insurances.

The findings largely support my initial assumptions. Actors were much more likely to actively demand and support an extension of social protections for low-income earners when their ideological orientation was left-of-center. Although the formation of political organizations and coalitions was strongly shaped by the divide between Peronism and anti-Peronism during the whole period of study, within these political camps the left-right divide was crucial to understanding why certain organizations and internal currents supported an extension of social protection for low-income earners and others opposed it. This applied to political parties as well as to trade unions, social movements, the church, and the military.

Actors were also much more likely to propose or support a progressive extension of social protections when their social support bases contained parts of the lower popular classes. Social policy proved to be an effective means of consolidating and enlarging this support. However, as there were also other means to mobilize lower popular class support, actors were most likely to pursue progressive social policy agendas when dependence on lower popular class support converged with a left-of-center political orientation or at least the strong influence of left internal currents. The experience of the trade union confederation CTA furthermore indicates that democratic organizational structures can reinforce the positive effect of lower popular class support by enabling participation in the agenda-setting process. Nevertheless, most political coalitions, parties, and organizations were highly hierarchical throughout the period and lacked working, institutionalized channels for base-level participation in the agenda-setting process. The elaboration of progressive social policy agendas was most often carried out in top-down processes, in which leaders proposed or implemented such policies in order to mobilize and reinforce lower popular class support. This also contributed to the fact that social policies were often designed in a way that allowed leaders to enhance their control over the beneficiaries and the rank and file. In the long run, these control- and cooptation-mechanisms were detrimental to the evolution of social protections for low-income earners, as they also served right-wing governments' efforts to influence, control, and weaken popular class mobilization.

As assumed in the theoretical framework, the structural context of Argentina's labor market with its strong formal-informal divide had an important impact on the way different segments of the popular classes were able to organize and to participate in the politics of social protection. The dispersion of informal workers among small-enterprises, the high proportion of self-employed and agricultural workers, the short average duration of employment, the lack of social protections, and the precarious nature of employment all made it more difficult for informal workers to organize in unions or to generate alternative forms of workplace organizations. This contributed decisively to the fact that union density in Argentina throughout the period of study was high among formal workers but low among informal workers. However, in the context of the neoliberal economic transformation during the last quarter of the 20th century, political ties between formal and informal workers strengthened as more and more unionists lost their jobs and contributed to the emergence and growth of unemployed and informal workers' movements.

State intervention and path dependency also shaped the role of actors in social politics. The incorporation of the trade unions into the first Peronist governing alliance was not only accompanied by a massive growth in membership, but also by a large-scale removal of communist, socialist, and anarcho-syndicalist leaders and their replacement by often much more conservative Peronists. This move towards a more centrist political orientation contributed, together with their social bases in the formal sector, to an increasing focus on the interests of the well-organized formal working class to the detriment of informal workers' and other marginalized sectors' interests. The introduction of a system of fragmented union-controlled health insurance funds furthermore created strong interests on the part of union leaders in defending the status quo in the health sector against both privatizing-regressive and universalizing-progressive reforms. Although the Peronist government attempted to restructure the employer associations as well, it was not able to have a comparable lasting effect.

The empirical findings furthermore support the hypothesis that regime type influences the likelihood of actors to pursue progressive social policy agendas. Centrist currents of the PJ, the UCR, and the trade union movement recognized the need to construct broad-based alliances and to mobilize at least some lower popular class support in order to be able to gain control over the government during democratic periods and therefore were more open to support progressive social policies under democratic than under authoritarian or semi-authoritarian rule.

In the following, the role of the most important actors and the factors that shaped their evolution will be analyzed in more detail.

#### **Political Parties**

In general terms, the Argentine case supports the assumption of the power resources approach that left and center-left parties tend to support the extension of social protections for low-income earners while center-right and right parties either prefer stratified or market-oriented social policies that benefit the better off and exclude or disadvantage low-income earners. However, the difficulty of the traditional power resources approach in dealing with the Argentine party system resulted from the fact that during the past 70 years both left and right parties were comparatively small, had limited influence in the parliament, and did not win any presidential election until 2015. Instead, two centrist parties, the PJ and the UCR, dominated electoral competition during most of the period. Both parties have been highly flexible with regard to ideology and party program, meaning that their platforms varied significantly over time and were much less tied to long-standing policy traditions than is the case of most communist, social democratic, Christian democratic, or conservative parties. As can be observed in Table 31, this flexibility also found its expression in the fact that governments led by both the PJ and the UCR expanded social protections for low-income earners during some periods and implemented regressive reform during others.

Ideological position nevertheless is a crucial category to understand these parties' social policies. While as a whole they can be described as centrist, their activist bases were indeed highly heterogeneous and contained both left and right internal currents. During all PJ and UCR led governments that expanded social protections, it was the parties' progressive currents that emphatically promoted egalitarian social policies and that pushed the agenda-setting process in this direction. The influence of left currents was furthermore reinforced by the fact that these governments either depended on a social support base that was composed to an important extent of low-income earners or tried to extend their electoral appeal to them.

During the first Peronist government (1946–1955) the creation of the Partido Peronista involved the absorption of the left-leaning Partido Laborista as its strongest organizational pillar. As most Partido Laborista leaders and members had socialist, anarcho-syndicalist, or communist backgrounds, this meant the constitution of a significant left wing within the Partido Peronista. Perón himself was furthermore aware that his electoral victory depended as much on massive support from the lower popular classes as it did on the support of organized labor. This combination of a strong left wing that promoted the extension of social protections for both political and ideological reasons with the pragmatism of Perón himself, who aimed at mobilizing and consolidating the support of the lower popular classes, led to the fact that the Peronist Party took a decidedly progressive course in social policy.

A similar constellation can be observed during the second Peronist government (1973–1976). This time the left wing had gained significant influence within the PJ party due to the tactical support it had received from Perón during his time in exile and due to the significant growth of leftist Peronist youth organizations. Especially during the early phase of the PJ government under President Héctor Cámpora, the party's left was able to influence the social policy agenda. The project of a universal public health sector was even promoted jointly by left sectors of the PJ and the UCR against the interests of the more conservative factions of both parties. Although this project failed due to the resistance of the trade union wing of the PJ, the left was nevertheless able to push the party's social policy agenda in an egalitarian direction. The centrist wing of the party, including Perón, by and large, supported this direction as it effectively served the goal of consolidating the party's broad base of support among the lower popular classes.

The moderate expansion of social protections for low-income earners under the UCR led government between 1983 and 1989 also took place in a moment when the social democratic wing of the party dominated the agenda setting and explicitly aimed at broadening its traditionally middle-class support base by incorporating additional popular class support. President Alfonsín himself belonged to the center-left wing of the party and supported the extension of social assistance policies and the universalization of health insurances.

In a similar way the Peronist government between 2003 and 2015, which initiated a massive expansion of social protections for low-income earners, was led by the left wing of the PJ party in coalition with a range of progressive non-Peronist politicians and organizations and depended crucially on the lower popular class vote.

In contrast, both the PJ led government of Carlos Menem (1989–1999) and the Alianza government led by Fernando de la Rúa (1999–2001) were dominated by the respective right wings of their parties and pursued a highly regressive social policy agenda. In both cases, the respective left wings of the parties criticized this direction and elaborated alternative, more egalitarian social policy proposals, yet were unable to prevail in the agenda-setting process. The PJ government of Carlos Menem had the peculiarity that it was led by the party's right and promoted a highly regressive social policy agenda at the same time that it continued to depend on the lower popular class vote, which it increasingly tried to retain through a dense network of clientelistic base-level structures and arbitrary social assistance programs.

During the UCRI-led government of Arturo Frondizi (1958–1962) and the UCRP-led government of Arturo Illia (1963–1966) the left wings of their parties were de facto impeded from having major influence on the social policy agenda as the military regularly intervened to block any attempt of significantly redistributive or pro-Peronist policies.

In sum, the left-right divide and the composition of their bases of support decisively shaped the positions of the main Argentine parties in the politics of social protection. As the two biggest parties can be described as centrist it depended on whether left or right internal currents dominated the agenda-setting process concerning social policy.

#### **Trade Unions**

Seen from a global perspective, the Argentine trade union movement was an important political force in enabling the extension of social protections to lowincome earners. Any government that implemented such change had the trade union movement as one of its most important alliance partners. This supports the assumption of the power resources approach, that strong trade union movements favor the development of egalitarian welfare states. Nevertheless, the role of the Argentine trade union movement has been ambivalent. While sometimes it supported or even promoted progressive reform, at other times it focused on improving or protecting existing benefits that covered only formal workers. In the health care sector, this protection of vested interests even led to the blocking of universalizing health care reforms during the 1970s and 1980s.

As was the case with the two major Argentine parties, ideological position mattered decisively. It was left factions within the union movement, such as the CGT de los Argentinos during the late 1960s and the CTA since the early 1990s, which explicitly promoted universal social policies, defined the working class in broad terms to include informal, unemployed, and self-employed workers, and pushed organizing strategies to reach out to those with the lowest incomes.

It should be remembered that the centrist factions of the union movement, which comprised the majority of the union leaders since the first Peronist government, also played an overall supportive role in the processes of social policy expansion towards low-income groups. This behavior was linked less to the ideological positions of these factions than to strategic and tactical choices. One important incentive for these choices was the composition of the social bases of the Argentine union movement. The massive state support provided to union organization during the first Peronist government led not only to a membership growth of more than 300% within only ten years, but also drastically changed the social composition of their rank and file membership by increasing the number of low-income affiliates working in small companies and hitherto unorganized sectors, including also a steep increase in female and migrant worker affiliation. To this day Argentine union membership is characterized by the peculiarity that affiliation is widespread in small enterprises with fewer than 50 employees and among workers with low levels of education (Delfini 2013, 108–111). Therefore, important sections of the formal workers organized in unions benefit as much from universal and egalitarian social policies as they do from social insurance coverage. Although worker-centered social insurance policies remained the first preference for most centrist union leaders, this composition of the rank and file contributed to the fact that universal and egalitarian policies indeed often constituted an acceptable second preference. Another important reason is related to the category of regime type. Under authoritarian and semi-authoritarian regimes, the centrist faction of the union movement mainly concentrated on defending and improving social insurance policies which benefited primarily the formal working class. The rightwing of the union movement even repeatedly collaborated with military governments on this issue. However, during democratic periods the union movement formed a component of broader political alliances that attempted to reach out to informal, unemployed, and self-employed workers in order to win electoral majorities. In this context, the centrist wing of the union movement supported the expansion of social protections for low-income earners through the PJ. Many trade union leaders were at the same time Peronist politicians and had concrete electoral ambitions. In most cases, this support was rather passive and indirect. However, in some situations, it was also quite active. For example, when the CGT had attained hegemony over the Peronist government in October 1975, it implemented the de facto universalization of old-age pensions by abolishing contribution requirements.

A particular pattern of union strategy can be observed in the area of health care policies. Some unions had started already in the late 19th century to organize mutual benefit societies that provided health services and holiday facilities to their affiliates. During the first half of the 20th century, some of the strongest unions were furthermore able to achieve the implementation of state subsidies and obligatory employer contributions to finance these services. Between 1943 and 1955, Perón extended and improved arrangements of this type in order to foster union support and finally the military government of Juan Carlos Onganía extended the system of sector-specific health insurance to the whole formal workforce in an attempt to defuse labor militancy and to co-opt cooperative union leaders. While the resulting system was highly inegalitarian and comparatively expensive due to high administrative costs, for the unions it constituted an important source of income and an effective tool for the recruitment of new members. In contrast to the pension and family allowance funds, most of these health insurance funds were directly controlled by the union of the respective sector. Not surprisingly, union leaders had therefore an extraordinarily strong interest in defending and strengthening these insurance funds. In practice, this led to the fact that the trade union movement resisted any attempt to significantly reform the sector, no matter whether the reform proposal was progressive (1970s and 1980s) or regressive (1990s) nor whether it was proposed by the PJ (1970s and 1990s) or the UCR (1980s). This, however, did not prevent the union movement from supporting expenditure increases on the universal public health sector as long as these did not negatively affect the unioncontrolled health insurance system.

#### **Social Movements**

Most Argentine social movements were politically left-wing and actively demanded or at least supported the extension of social protections for low-income earners. In this sense, the empirical analysis confirms the initial hypotheses, that social movements that organize disadvantaged or even marginalized groups, such as peasants, indigenous communities, women, or unemployed and informal workers, tend to support the extension of social protections for low-income earners as their struggles are aiming for more equality and their bases of support tend to benefit from such policies. Compared to the main political parties and trade unions, the strength and political influence of social movements was generally weaker and varied much more strongly over time. Notwithstanding, during some periods, social movements played a crucial role in the extension of social protections.

Until the early 1970s, the political power and influence of autonomous lower popular class organizations were relatively limited. The PJ party and some leftist trade unions, such as the sugar cane workers' union, were the main entities that organized populations in these social sectors. During the early 1970s, residents of poor neighborhoods, left-wing Peronist groups and the Movement of Priests for the Third World successfully constructed a slum dwellers movement, the Movimiento Villero Peronista. By 1973 the movement had become capable of mobilizing thousands of low-income earners and sustained base-level structures in 450 slums throughout the country. Together with other groups, it constituted part of the progressive wing that pushed for redistributive policies within the broader alliance of the Peronist government. However, subsequently, the organization declined in the face of brutal right-wing repression first through the paramilitary Triple A and then through the military dictatorship. With the return to democracy, social movements of this type reemerged on a smaller scale and carried out land occupations and founded soup kitchens and worker cooperatives. The outreach of these movements remained very limited, however, until the mid-1990s when new lower popular class movements emerged and within a short time converted into mass movements. The rise of these movements was facilitated through several factors. First, years of neoliberal policies and social regress had created enormous frustration and anger among large parts of the lower popular classes. Second, the massive rise of informality and unemployment led to the situation that significant numbers of hitherto formal, unionized workers joined the ranks of the lower popular classes and contributed with their previous experiences to processes of base-level organization. Third, the progressive union confederation CTA and smaller leftist parties supported these processes of organization. Fourth, by setting up roadblocks they were highly effective in putting pressure on the government and extracting social policy concessions in favor of their members. By the early 2000s, the overall membership of these social movements had reached approximately 300,000 and the ideological position of virtually all of them was clearly left-wing. Although, at least initially, they were most successful in organizing around the issue of unemployment, the majority of them also contained branches that organized other segments of the lower popular classes, such as peasants, cooperative workers, informal workers, neighborhood activists, and indigenous communities. Against this background, groups that were suffering from several intersecting discriminations constituted the majority at the base-level: most of the activists were women, lived in impoverished and stigmatized neighborhoods and a high percentage had migratory or indigenous backgrounds. Not surprisingly, these movements carried out extensive activism to demand social protections for low-income earners. Most of the concrete actions focused on specific benefits for their members, but on the more aggregate level, they demanded the universalization of unemployment protection schemes and pressured from within and without the Kirchnerist government for an extension of redistributive social policies.

During the 1970s and 1980s, leftist youth and student movements also played a significant role in promoting lower popular class organization and social policy expansion. With the global wave of student protests in 1968 and the Cordoba-riots in the following year, left-leaning student organizations began to grow quickly in Argentina. Although different types of Marxist movements emerged throughout the universities of the country, the movements that were most effectively able to influence social policy were those with links to the Peronist movement or the UCR party. In 1972, various left-leaning Peronist youth organizations united and formed the Peronist Youth, which soon converted into one of the main organizational pillars of Peronism, was incorporated into its national decision-making bodies, heavily influenced the 1973 electoral campaign, and attained a considerable portion of parliamentary seats and public offices. Members and sympathizers of the Peronist Youth subsequently used their political positions to promote egalitarian social policies such as a unified universal health system and large-scale housing programs. However, due to brutal repression through the paramilitary Triple A and later the military Junta, the Peronist Youth and most of the Marxist student movements increasingly weakened and finally disintegrated.

The less radical center-left youth organizations of the UCR, the Franja Morada and the Junta Coordinadora Nacional, suffered considerably less repression and were able to extend their influence within the UCR throughout the 1970s and early 1980s. During the electoral campaign of 1983, these youth groups actively reached out to mobilize support in working-class and low-income neighborhoods and supported a redistributive social policy agenda that had among its aims the universalization of health insurance coverage in favor of low-income groups.

In response to the murder of approximately 30,000 mostly left-wing activists during the military dictatorship between 1976 and 1983, Argentina developed comparatively strong human rights organizations, such as the Madres de la Plaza de Mayo, Abuelas de Plaza de Mayo and the Centro de Estudios Legales y Sociales (CELS). All these organizations were politically left-of-center and promoted the implementation of social human rights, which included explicit support for expanding social protections for low-income earners. Nevertheless, their focus remained on the human rights violations of the military government and on finding the stolen children of murdered activists, so their influence on social policy was rather limited.

The Argentine women's movement became increasingly broad and influential after 1986 when the Encuentro Nacional de Mujeres was established as a yearly gathering point for activists. By the mid-2010s participation in these gatherings reached over 60,000 and the movement was capable of mobilizing several hundred thousand demonstrators throughout the country. The massive growth of the women's movement was to an important extent related to the growing participation of women from progressive parties, trade unions, and unemployed workers' organizations so that the movement increasingly attained a popular class and left-of-center base. For most activists, women's interests also included broader egalitarian social policy reform, because such policies were considered to contribute to a diminution of both gender and class inequalities. The influence of the movement on broader social policy issues was nevertheless rather indirect, operating through the egalitarian perspectives that activists of the women's movement promoted in the broader discourse and the trade unions, parties, and social movements they were affiliated with, while the major campaigns and mass demonstrations focused on the issues of gendered violence and the right to abortion.

#### **Employers' Associations**

The main employers' associations were opposed to all major initiatives to extend social protections to low-income earners throughout the period of study. This supports the initial hypothesis that capital tends to oppose such policies. However, whether this resulted in open and aggressive opposition or in mildly and implicitly formulated critique depended strongly on the broader political context and on the kind of employer association.

Structural context proved to be an important category for understanding employer opposition to social policy expansion. In a context of mostly abundant labor supply, due to an ample reserve army of unemployed or informal self-employed workers, most employers felt little need to support processes of skill formation and reproduction through extending social protections to lowincome sectors. In contrast, the negative aspects of social policy expansion dominated in the view of most employers. The three crucial arguments that employers' associations repeatedly brought forward were that such policies would impose considerable fiscal costs and potentially higher tax burdens, that they would foster laziness in the beneficiaries, and that they would inhibit the operation of the benign forces of the market and reduce competition among wage earners. In this way, the arguments brought forward by the major employer associations resembled quite strongly the assumptions of the power resources approach, that due to structural reasons, employers had important incentives to oppose egalitarian social policy.

In order to understand the different intensities of resistance to redistributive policies of different associations, it also proved useful to take into account their ideological positions, social bases, and organizational structures. The most powerful employer associations in the agricultural (SRA, CRA), financial (ABRA, ADEBA), commercial (CAC, BOLSA) and construction sectors (CACON) were throughout the period committed to economic liberalism, which provided a cognitive and analytical framework from which an expansion of social protections appeared harmful for their business activities as well as the economy as a whole. Within the strongest industrial sector association (UIA), economic liberalism was the dominating perspective during most of the period; however, between the 1940s and 1970s significant internal discussions took place over the question of whether state-sponsored import substitution or liberalism would be the economic policy most favorable for the sector. While most of the big companies clearly adhered to the liberal paradigm, many smalland medium-sized and national market-oriented enterprises favored ISI policies that would foster local demand and protect them from foreign competition. Although the latter group grew significantly between the 1940s and 1970s, the organizational structure of the association allowed a comparatively small group of big and rather liberal oriented companies to wield a disproportionate and often dominant influence over the political course of the UIA. After economic deregulation was initiated by the military junta between 1976 and 1983, the liberal economic position became eventually hegemonic within the association. Thirty of the most influential and powerful business leaders from different sectors furthermore formed the CEA, which represented the top of the Argentine economic elite and also adhered to a neoliberal economic worldview. Although with different intensities, all these associations actively and openly opposed redistributive social policies. This involved lobbying, dissemination of studies undertaken by employer-financed think tanks, and public criticism of or support for certain policies. During military dictatorships and right-wing governments, representatives of these organizations regularly occupied crucial positions in the cabinet and were able to exercise strong direct influence on the direction of economic and social policy. Under progressive governments, several of these employer associations pursued aggressive oppositional strategies, going at times as far as to promote and support military coups.

A second group of employer associations was less aggressively anti-redistributive, though much less powerful. In order to unite those employers that favored the ISI model, Perón created alternative employer associations for

different economic sectors during the early 1950s. The industrial sector association CGI was however the only one of these organizations that was able to attract a significant level of employer support. In contrast to the above-mentioned associations, the CGI was dominated by small- and medium-size national market-dependent companies and maintained good relations with the Peronist governments. Rather than openly confronting the government on social policy issues, the CGI preferred to reach compromises in tripartite negotiations. Although the association clearly supported the ISI model with its focus on protecting national industries and fostering local demand, it favored a version of this model that created demand through investment rather than through increasing consumption with redistributive policies. Due to the related fiscal costs, the CGI also opposed an expansion of social protections for low-income earners. However, in contrast to the associations adhering to economic liberalism, this opposition found its expression mainly in lobbying and negotiation activities rather than in public confrontation. With the process of economic liberalization during the military junta between 1976 and 1983, the CGI eventually declined and lost most of its influence.

## **The Military**

The Argentine military constituted an important actor in social politics during the first 40 years (1943–1983) of the period of study. Although there existed different internal factions, the hegemonic political orientation in the military leadership was right-wing authoritarian and its political influence on social policy was profoundly regressive, and thus highly detrimental to low-income earners.

In two ways the developments within the military during the 1920s and 1930s were important to understanding its later involvement in politics. First, the *coup d'état* of 1930 reflected the growing strength of factions within the military leadership that were convinced that the armed forces had the duty to participate as a major force for order in the political arena. Second, against the background of the international economic crises and the rise of Keynesian thinking, nationalist currents that favored state-interventionist industrialization policies became increasingly influential and started to challenge the dominance of economic liberalism within the military elite (Hänsch and Riekenberg 2008; Riekenberg 2009; Rock 1993). Although these currents differed from the economically liberal currents in the sense that they were more strongly influenced by Catholic Social Teaching and assumed that the state should play a certain role in welfare provision, their agendas were far from being progressive. Rather than seeing social policy as a tool for reducing social inequalities, they saw it as an effective instrument of cementing existing social hierarchies, controlling and dividing popular class organization, and constructing effective barriers against the influence of communist ideas. In economic terms, these currents favored a variant of the ISI paradigm, that, like the developmentalist model, put emphasis on investment rather than on redistribution and national demand. In contrast, the economically liberal currents within the military favored a considerably smaller role for the state in the economy as well as in welfare provision (Rock 1993; Sikkink 1991). In concrete terms, these different orientations found their expression in the fact that state-interventionist currents tended to prefer centralized, state-led social insurance policies as the main social policy instrument, while economically liberal currents preferred market-oriented private insurance. However, both currents had in common that their social policies were largely regressive and detrimental to low-income earners. Both statist and economically liberal military governments radically reduced spending on social assistance and universal policies and promoted social insurance reforms that limited coverage to formal workers and favored higher income workers over lower income workers. In the face of a broad informal sector in Argentina, these policies significantly reduced social protections for low-income earners and thus contributed pivotally to the formation of a truncated social policy regime. In political terms, the military elite's capacity to carry out its regressive agenda rested on alliances with other center-right and right political forces, such as the conservative leadership of the Catholic Church, the majority of the employers' associations, and right-wing politicians.

The only major exception to this regressive role of the military was the dictatorship between 1943 and 1946. However, even in that case, the extension of social insurance policies to hitherto excluded groups was related to Juán Domingo Perón's personal political strategy and his influence through the Department of Labor, not to a general preference for redistributive policies among the military elite in general. Perón formed part of the GOU, a group of rightwing nationalist military leaders that promoted state interventionist industrialization policies and managed to become the dominant force within the military government soon after the coup in 1943. From his position as head of the Department of Labor, Perón implemented a range of carrot-and-stick policies which included repression and intervention as well as social and labor policies to create an increasingly strong political support base within the trade union movement. Although his policies mainly covered formal workers, they had nevertheless significantly progressive redistributive effects due to the fact that the majority of formal workers were still not covered by social insurances and that previous labor market regulations had been largely unfavorable to workers in general. In order to foster military support for his agenda, Perón argued that his policies were necessary in order to eliminate communist influence within the labor movement, to establish state control over the working class, and to reach a situation of stable class harmony. Notwithstanding these arguments, his redistributive agenda soon aroused significant unrest among military leaders and even led to his temporary arrest in 1945. After his election as president in 1946, Perón was initially able to reinforce military loyalty through strong increases in military spending and arms production; however, as soon as these expenditure increases became difficult to sustain in the face of the economic crisis of the late 1940s and early 1950s, the military became an increasingly oppositional actor. By 1955 military opposition had become openly aggressive, with both rather statist and economically liberal military leaders attacking Perón for his redistributive agenda and his alliance with popular class forces (Rock 1993). In September 1955, the military eventually staged a *coup d'état* against Perón's government.

Despite the varying influence of economically liberal and more statist internal factions, all three later military dictatorships (1955–58, 1966–73, 1976– 83) were fiercely right-wing and pursued highly regressive social policy agendas that significantly deteriorated social protections for low-income earners. Furthermore, between 1958 and 1966 the military intervened any time government policy went significantly in a pro-Peronist or redistributive direction.

Only after the return to democracy in 1983 did military involvement in politics decreased significantly. Due to the debacle of the Falklands war, the massive human rights violations, and the economic disaster generated by the last military junta, the rejection of military involvement in politics had become a broad-based societal consensus, which enabled the democratic government of Raúl Alfonsín to restrict the autonomy of the armed forces, to reduce military expenditure, and to initiate trials against human rights violators. Consequently, the military ceased to be an influential actor in social politics after 1983.

#### **The Catholic Church**

The empirical findings largely support our initial hypothesis that due to the influence of Catholic Social Teaching, the leadership of the Argentine Catholic Church tended to support stratified, status conserving social policies. However, during most of the period of study, the involvement of the Catholic Church in social politics was rather indirect. Leaders of the church preferred to criticize growing individualism, materialism, and inequality in general terms rather than to participate in concrete processes of social policy reform. In some cases, such as the introduction of the workfare program PJJHD in 2002, the church took the role of a mediating force between different contending interests.

Although the church itself did not actively pursue a regressive social policy agenda, it nevertheless contributed to the implementation of such policies by legitimizing or even openly supporting military and right-of-center governments. The collaboration with right-wing military dictatorships was related to an historically close relationship between right-wing nationalist military leaders and the formal leadership of the Argentine church. Furthermore, both military and democratic right-wing governments shared the church's positions on issues such as divorce, religious education, and abortion. All these issues were highly important to the leadership of the church and shaped their patterns of political alliance making more than social policy issues. In contrast, the relationship of the leadership of the church with center-left governments that promoted egalitarian social policies was more conflictive, as these governments also promoted policies that strengthened the rights of women and homosexuals, liberalized divorce, promoted sexual education, or allowed for discussions on the issue of abortion. A peculiar case was the first government of Perón, which implemented a progressive social policy agenda while initially enjoying strong support from the hierarchy of Catholic Church. This particular constellation had much to do with the fact that Perón was a former military leader and that his government reaffirmed the system of religious education that was considered to be of extraordinary importance by the church's leadership. As soon as the government began to take measures that contradicted the interests of the church, such as the introduction of female suffrage and the takeover of Catholic social assistance by the Eva Perón Foundation, however, the relationship between the church and the government increasingly deteriorated. By 1955 the church had finally become one of the driving forces behind the *coup d'état* that overthrew the Peronist government.

Like other organizations, the Argentine Catholic Church was not politically homogeneous. Again, the left-right divide played an important role in understanding the involvement of different internal currents in social politics. While conservative currents remained the dominant force at the formal leadership level throughout the whole period of study, progressive base-level movements grew significantly during the 1960s and 1970s. While the former tended to support regressive governing alliances, the latter often forged alliances with left-of-center organizations and contributed to processes of lower popular class organization. This, however, also led to progressive activists of the church becoming the targets of brutal paramilitary and state repression during the 1970s and early 1980s, which contributed to a restoration of the hegemony of the conservative leadership. More progressive currents nevertheless recovered some strength during the 1990s with the emergence of new social movements and growing resistance to the neoliberal model.

#### **International Organizations**

The empirical evidence on the role of international organizations in the Argentine politics of social protection for low-income earners largely supports the initially formulated hypotheses as well as the findings of Tittor (2012) and Deacon (1997, 2005). UN agencies and programs such as the ILO, the WHO, and UNRISD generally promoted the expansion of social policy and offered technical support for the development and implementation of policies. International Financial Institutions, such as the IMF, World Bank and IADB, in

contrast promoted the privatization and retrenchment of social policies following the Chilean model in conjunction with the introduction of residual targeted social assistance programs. Although these targeted social policies explicitly favored low-income earners, their coverage and financing were so limited that they were not able to compensate for the loss of social protections for lowincome earners related to the regressive restructuring of social insurance policies. Hence, despite the promotion of targeted social assistance policies, the overall impact of the IFIs on Argentine social policy was highly detrimental to low-income earners.

Over the period of study, the influence of different international organizations varied as well as the mechanisms through which they participated in the policy-making process. Compared to the above described political and economic actors on the national level, the impact of international organizations was at most times rather indirect and operated mainly through the promotion of ideas that shaped the views of both the wider public and the Argentine community of social policy experts in science and public office. In the 1930s the ILO promoted the state-led expansion of social insurance arrangements and became increasingly influential in the Argentine social policy discourse. During the 1940s, inspired by, among other things, the Beveridge report, it gradually shifted towards broader concepts of social security that encompassed nonsalaried parts of the population. The newly founded WHO also took up the idea of broad-based universal social security, so that from the 1940s on, both the ILO and WHO favored the extension of social protections for low-income earners by promoting a favorable discursive environment and by providing technical advice and assistance for the development and implementation of policies. Ideas promoted by these organizations were important inspirations for the initiatives to build a universal health sector and broad-based pension system during the 1940s, 1950s, and 1970s in Argentina. However, with the dramatic increase of Argentine foreign debts and the rise of neoliberalism between the mid-1970s and early 2000s, the influence of the state-centered ILO and WHO gradually declined and the influence of IFIs on social policy intensified. In contrast to the mostly indirect influence of the UN agencies, the influence of the IFIs increasingly rested on more direct mechanisms such as the financing of reform programs through the World Bank or the conditionalities imposed by the IMF in exchange for the approval of loans. For example, the World Bank actively financed regressive and privatizing reforms in the health and pension sectors during the 1990s at the same time that the IMF linked financial support to the government to the effective implementation of these programs. In order to assure the political viability of these reforms and to mitigate some of its negative social impacts the World Bank furthermore provided credits to finance residual social assistance policies for short periods of time. Finally, with the economic collapse of the neoliberal model during the early 2000s, the increasing discrediting of neoliberal ideas, and the government's default on foreign debts, the influence of the IFIs on social policy-making in Argentina decreased again.

## **5.3 The Distribution of Power Resources and Social Protection for Low-Income Earners**

One of the main hypotheses associated with my theoretical framework is that the direction of social policy is crucially dependent on the distribution of power in the society. I assumed that the extension of social protections for low-income earners is much more likely when the popular classes are powerful. To assess the level of popular class power I distinguished four types of power resources: associational, structural, institutional, and discursive power resources. In the following I will summarize and analyze the long-term variation of popular class power resources and its impact on social protection for low-income earners.

## **The Long-Term Evolution of Popular Class Power Resources**

As can be seen in Figure 13, the overall level of popular class power resources has varied significantly over time and thereby provided widely differing frameworks for social policy reform processes.

In order to analyze this variation, it is helpful to look at the different types of power resources separately. Figure 14 shows the evolution of the levels of structural and discursive popular class power resources. As can be seen, both remained relatively stable on a medium-high to high level between the mid-1940s and the mid-1970s. The high level of discursive power resources was related to the predominance of the ISI paradigm among the wider public as well as expert communities and politicians in Argentina during these years. Due to the interventionist and rather demand-oriented logic of the paradigm, its popularity provided progressive and popular class actors with a favorable discursive environment in which to frame redistributive social policies as being benign for the economy and society as a whole. Hence, it facilitated the generation of mass support for redistributive policies.

Figure 13: The evolution of the overall level of popular class power resources

\* The assessment of the level of popular class power resources refers always to the approximate timeframe within which a new governing alliance was formed at the beginning of a period. This timeframe varied with respect to the different alliances.

Figure 14: The evolution of structural and discursive popular class power resources

\* See figure 13.

In turn, ISI policies and a favorable international context contributed to the growth of the industrial sector and to low levels of unemployment, which strengthened labor's capacity to cause significant economic damage through strike activities and therefore contributed to high levels of structural power resources.<sup>189</sup>

<sup>189</sup> Between the late 1940s and the mid-1970s, the industrial sector constituted between 33% and 41% of the GDP and employed between 30% and 35% of the workforce. Unemployment remained most of the time below 6%.

Nevertheless, as can be seen in Figure 14, the discursive and structural popular class power resources declined steadily afterward. Political and economic crises during the 1970s and 1980s contributed to a growing perception that the ISI-paradigm was exhausted and unable to deliver solid solutions for the problems of the time. Furthermore, the proliferation of neoliberal employer-financed think tanks during the military dictatorship (1976–83) in conjunction with a worldwide trend towards supply-side-oriented economic perspectives led to an increasing influence of neoliberalism, which finally became the dominant economic paradigm towards the 1990s and significantly reduced the level of discursive popular class power resources. Simultaneously, the frequent eruption of profound economic crises since the 1970s as well as the implementation of neoliberal policies led to a steady decline of the industrial sector, rising unemployment and informality.<sup>190</sup> This implied a gradual reduction of the popular classes' structural power resources.

These crises, however, also led to growing resistance and increasingly widespread doubts about the adequacy of the paradigm to serve the interests of the majority of the population. Thereby, they provided the ground for the return of demand-oriented economic perspectives and a recovery of the popular classes' discursive power resources. From the mid-1990s on, this led to a divergence between rising levels of discursive and falling levels of structural popular class power resources. However, rather than showing a dissociation of the two categories, the temporary divergence was in large part a result of their close interaction and interdependence. Indeed, the discrediting of the neoliberal paradigm and the gradual recovery of the popular classes' discursive power resources was decisively fueled by the dramatic deterioration of the labor market, and hence, by a development that at the same time eroded the popular classes' structural power resources. Subsequently, the recovery of discursive power resources in combination with a strengthening of the associational power resources of the popular classes enabled the formation of a progressive governing alliance and a deep shift in economic policy that provided the bases for a gradual recovery of the labor market and thereby of the popular classes' structural power resources.<sup>191</sup> Hence, it is precisely their interdependence and interaction that explains the delay between the moments when they each began to recover.

All in all, the evolution of structural and discursive popular class power resources was marked by relatively long timeframes and gradual changes. Both remained on a high level between the mid-1940s and 1970s, then gradually

<sup>190</sup> Between 1976 and 2002, the industrial share of GDP declined from 41% to 29% and the share of industrial employment fell from 32% to 17%. In the same timeframe, unemployment increased from less than 5% to nearly 20% and informal salaried employment from 18% to 44%.

<sup>191</sup> Between 2002 and 2015 unemployment decreased from nearly 20% to less than 7% and informal salaried employment from more than 44% to less than 33%.

declined and finally began to recover during the mid-1990s, in the case of discursive power resources, and the mid-2000s, in the case of structural power resources.

Figure 15: The evolution of associational and institutional popular class power resources

In contrast, the evolution of associational and institutional power resources followed quite different patterns, as can be observed in Figure 15. During the first half of the 20th century, the Argentine labor movement experienced a gradual, yet not linear, process of growth and unification. This process was significantly accelerated after 1943 when state policies favored union organization and the emergence of a strong popular class-based party.<sup>192</sup> By this means, during the second half of the 1940s, the associational power resources of the popular classes had increased to a medium level. Afterward, this level proved to be resilient to changing economic and political environments. After the overthrow of the Peronist government in 1955, the Peronist Party, the main popular class-based party, was prohibited. However, the labor movement had attained unprecedented levels of organization, which to a certain degree compensated for the decrease of associational power resources generated by the prohibition of the Peronist Party.<sup>193</sup> Furthermore, the activist structures of the unions provided an important space for the continuity of Peronist political activities despite the official ban. With the lifting of the ban against the Peronist Party and the emergence of significant leftist student, slum dwellers, and base-level union movements, the associational power resources of the popular classes reached their peak with a medium-high level during the early 1970s. During the following

<sup>\*</sup> See figure 13.

<sup>192</sup> The first popular class-based mass party that emerged was the short-lived Partido Laborista, which was soon replaced by the Partido Peronista.

<sup>193</sup> Trade union density increased from approximately 11% in 1941 to 18% in 1946 and 43% in 1954. Afterwards it remained stunningly stable at values between 36 and 44% until the end of the period of study in 2015.

years, severe conflicts within the Peronist movement significantly weakened its left wing and led to a temporary paralysis of the union movement. Under these conditions, the military seized power in 1976 and carried out brutal repression against all leftist organizations, combative union factions as well as progressive intellectuals. This led to a temporary decline of the popular classes' associational power resources to a medium-low level during the second half of the 1970s. However, the union movement was able to maintain a high level of union density and towards the early 1980s also recovered important capacities of political action. With the fall of the military dictatorship in 1983 political parties were able to rebuild their organizational structures, which led to a recovery of medium levels of popular class associational power resources. While during the following two decades economic crises and neoliberal policies weakened the union movement, new center-left parties and lower popular class organizations emerged, so that the overall level of associational power resources remained medium. All in all, the associational power resources remained relatively stable throughout the period of study, despite significant qualitative differences regarding the concrete composition and orientation of popular class actors over time.

In stark contrast, the level of institutional popular class power resources oscillated strongly between different periods. This was largely due to the fact that the institutional power resources of the numerically strong popular classes depended most importantly on their capacity to exercise democratic rights and to influence the course of governments and parliaments by means of electoral participation. Repeated *coups d'état* therefore led to abrupt and dramatic declines of this type of popular class power resources. As can be seen in Figure 15, the institutional power resources of the popular classes reached a medium level during the second half of the 1940s, when democracy was first reinstalled and extended to women but then gradually restricted by the first Peronist government. The *coup d'état* of 1955 and the following authoritarian and semiauthoritarian regimes in combination with the prohibition of the main popular class-based party eventually led to a dramatic reduction of these power resources to a low level. In 1973, the return to democracy and the lifting of the ban against the Peronist Party initiated a short phase of high levels institutional power resources, which ended again in 1976 when the military erected a brutal authoritarian regime. Only in 1983 was democracy lastingly restored and since then there have been medium-high to high levels of institutional popular class power resources.

In sum, the evolution of the different types of popular class power resources showed considerably different patterns. While the levels of associational power resources remained surprisingly stable, the levels of discursive and structural power resources underwent gradual change over longer periods of time and the levels of institutional power resources were characterized by frequent and dramatic oscillations. Against this background, it becomes evident that a focus only on the evolution of the associational power resources, as in the traditional power resources approach, would have been insufficient to understand the significantly varying directions of social policy reform. This underlines the empirical importance of my theoretical extension of the classical power resources approach in welfare state research by distinguishing four types of power resources.

## **The Overall Level of Popular Class Power Resources and its Impact on Social Policy Change**

As can be observed in Table 32, the overall level of popular class power resources at the beginning of each period correlates coherently with the subsequent direction of social policy change.


Table 32: Popular class power and patterns of change regarding social protection for lowincome earners, 1943–2015

\* At the beginning of each period.

At the beginning of the periods 1943–55, 1973–76, and 2002–15, the overall level of popular class power resources was medium-high or high and social protection for low-income earners expanded strongly during all these periods. In contrast, at the beginning of the periods 1976–83 and 1989–2001, the overall level of popular class power was medium-low and social protection for lowincome earners was reduced. These coincidences support the main hypothesis of the power resources approach that progressive social policy becomes more likely when popular class power resources are high. More variable were the social policy outcomes during periods at the beginning of which the overall level of popular class power was medium. During the period 1955–73, social protection for low-income earners was strongly reduced, while during the period 1983–89 it was moderately expanded. Again, this fits well with the theoretical framework. The level of popular class power resources was not assumed to translate linearly into social policy change but to be mediated by politically contingent processes. Furthermore, as will be shown in the following, the specific combinations of different types of popular class power resources during the two periods had an impact on the respective dynamics of social policy change.

There were two main political mechanisms through which the overall level of popular class power resources was causally related to social policy outcomes. First, the overall level of popular class power determined to an important degree the capacity of popular class actors to exercise external pressure on governments, which hence led governments to use social policy concessions as a means to contain protest and strike activity, to divide and control popular class forces, or to co-opt union and social movement leaders. Second, the overall level of popular class power decisively shaped the formation of different types of governing alliances. High levels of popular class power resources at the beginning of a period made the formation of progressive and inclusionary governing alliances politically viable, while low levels of popular class power resources provided a favorable framework for the formation of exclusionary and regressive governing alliances. In turn, the evolution of power resources within each period was one of the crucial factors determining the sustainability of the respective alliance.

In the Argentine case, the second mechanism, that is the formation of different governing alliances, was empirically much more important in shaping the long-term evolution of social protections for low-income earners. Such social policies were consistently extended by progressive and inclusionary governing alliances and cut by regressive and exclusionary governing alliances. External popular class pressures on exclusionary governing alliances were mainly exercised by the trade union movement, with the result that regressive governing alliances responded with social policies that benefited the organized formal working class but usually excluded the lower popular classes. The capacities of the lower popular classes to exercise external pressure on regressive governments were much more limited during most of the period of study. The only major exception constituted the proliferation of lower popular class protests and movements after the mid-1990s, which led to reiterated social policy concessions, particularly in the area of social assistance policies.

## **Different Types of Power Resources and the Dynamics of Social Politics**

Table 33 provides a more detailed overview of the distribution of power resources and social policy change by distinguishing different types of power resources. Combined with findings from the process tracing analysis, these details reveal further patterns. As we can observe, not only the overall level of popular class power resources differed between periods, but also the combination of different types of power resources. This variation, in turn, coincided with a varying distribution of power resources within the popular classes. Due to their high union density and capacity to disturb processes of production, the formal working class' structural and associational power resources were significantly stronger than those of the lower popular classes throughout the whole period of study. In contrast, the right to vote and the dominance of economic paradigms that were favorable for redistributive policies strengthened the institutional and discursive power resources of the popular classes as a whole, and hence, increased the relative political weight of the lower popular classes. During the periods 1943–1955, 1973–1976, 1983–1989, and 2002– 2015, the levels of both institutional and discursive power resources of the popular classes were either medium or high. This contributed not only to overall levels of popular class power resources that were medium to high but also to a distribution of power resources among and within the popular classes that gave more weight to low-income groups and left-of-center currents than during other periods. In the theoretical framework, it has been assumed that this combination provided a particularly favorable political framework for progressive social policy reform. If we now look at the social policy outcomes of these periods, the findings provide strong support for this hypothesis. During all four periods, social protections for low-income earners were expanded.

The historical reconstruction of the processes of social policy reform during these periods not only provided further empirical evidence in support of this hypothesis but also revealed that there were two main political mechanisms that linked this particular distribution of power resources with egalitarian social policy outcomes. First, as argued earlier, the overall medium to high levels of popular class power resources provided a context under which the formation of inclusionary and progressive governing alliances was politically viable and sustainable. And second, the distribution of institutional and discursive power resources provided governments and political parties with important incentives to try to mobilize lower popular class support through inclusionary social policy agendas.<sup>194</sup> Indeed, the mobilization of the lower popular class vote was a crucial prerequisite for the electoral success of any of these alliances. Simultaneously, the distribution of discursive power resources provided actors with capacities to successfully frame policies in favor of the lower popular classes as beneficial and desirable for the society as a whole. This, in turn, provided a discursive environment under which centrist currents within the PJ and UCR as well as the trade union movement were inclined to forge alliances with left-of-center and lower popular class forces.

<sup>194</sup> Such incentives also operated during the regressive governing alliance of President Carlos Menem in the 1990s. This led to the fact that, despite the overall regressive nature of the alliance's policy agenda, it extended clientelistic social assistance policies in order to slow down the erosion of lower popular class support.


Table 33: Four types of popular class power resources and patterns of change regarding social protection for low-income earners, 1943–2015

\* At the beginning of each period.

A closer look at the processes of social policy change shows how these mechanisms operated in practice. The first period was initiated with the military coup of 1943 and the assumption of Juan Domingo Perón as head of the Department of Labor shortly later. The coup was preceded by decades of gradual industrialization and the emergence of the working class as an increasingly powerful political actor. The military government responded to these developments by implementing and expanding redistributive social policies in order to foster working-class support and to enhance state control over the union movement. Perón explicitly used the implementation of new pension funds and subsidies for union-controlled health insurance funds to establish alliances with some of the most powerful trade unions. At the same time, these policies also served as mechanisms through which labor militancy could be contained and defused. The increasing predominance of the ISI paradigm in public opinion furthermore provided discursive power resources that enabled Perón and the unions to frame these policies as being favorable to the goal of economic development, and hence, as being beneficial for the society as a whole. This discursive context could have provided an environment in which social policies in favor of unorganized low-income earners were justified in a similar way, but in the face of the low structural, institutional, and associational power resources of the lower popular classes, the political benefits of doing so appeared very limited. As a result, until the return to democracy in 1946, social policy expansion mainly benefited sectors of the formal working class that had attained significant associational and structural power resources. With the return to democracy in 1946 this constellation changed significantly, as the right to vote provided the lower popular classes with a significant level of institutional power resources. Indeed, neither of the two major electoral alliances during the 1946 elections was able to win without mobilizing significant low-income earner support. At the same time, the particular combination of democracy with a high level of discursive popular class power resources enabled political actors to mobilize lower popular class support through inclusionary social policy agendas without having to renounce formal working-class and middle-class support, as the ISI paradigm allowed them to frame these policies as ultimately benefiting all social classes. On these grounds, both electoral alliances decided to compete with progressive social policy agendas. The Peronist alliance was finally more successful in mobilizing popular class support and won the elections. To consolidate and broaden this support, the Peronist government then implemented an inclusionary social policy agenda that benefitted both the formal working class as well as the lower popular classes. While it continued to expand social insurance funds that favored formal workers, it combined these policies with a massive expansion of social assistance policies, the creation of a universal public health sector, and the introduction of highly permissive pension access criteria. Taken together these policies amounted to a strong expansion of social protections for low-income earners.

Similar processes can be observed during the Peronist governments in the 1973–1976 and 2002–2015 periods. Both periods were preceded by several years of increasing popular class protests as well as an increase of the overall level of popular class power resources, which finally contributed to the decline of the respective regressive governing alliances and provided viable options for the formation of more inclusionary governments. In both cases, preceding developments also led to an increase in the relative political weight of the lower popular classes. In 1973, the return to democracy provided the scarcely organized but numerous lower popular classes with significant electoral weight. During the early 2000s, the discrediting of neoliberalism and the increasing popularity of a more demand-oriented economic paradigm as well as the successful construction of social movements strengthened their discursive and associational power resources. Against these backgrounds, Peronist presidents during both periods undertook significant efforts to mobilize lower popular class electoral support through the expansion of social protections for low-income earners. In order to justify these policies in the eyes of the wider public, these governments took up the prevalent economic paradigms and argued that increasing social protections would foster national demand and hence contribute to creating jobs and economic wealth for everyone. Centrist currents of the PJ and the union movement agreed with this claim and took a supportive stance towards these policies. At the same time, these measures were combined with an improvement of existing social protections for formal workers. As they had between 1946 and 1955, formal labor and the lower popular classes constituted the main electoral and associational strongholds of these Peronist governments.

The operation of the above mentioned three political mechanisms can also be observed during the 1983–1989 period, yet the political dynamics were different. Protest activity and popular class power resources increased during the last years of the preceding military dictatorship, finally contributing to the return to democracy and providing options for the formation of more inclusionary governing alliances. Nevertheless, the brutal repression of the military government and the initiation of a process of de-industrialization had lasting negative effects on the associational and structural power resources of the popular classes. Against this background, the overall level of popular class power resources was medium and therefore lower than during the previously discussed periods. At the same time, the ongoing popularity of the ISI paradigm and the return to democracy in 1983 provided institutional and discursive power resources that favored the lower popular classes in a similar way as in the above described periods. In the face of this, both major electoral alliances in the 1983 elections made efforts to compete for the vote of low-income earners as well as formal workers. Although the principal issues during the campaign were centered on the restoration of democratic institutions, the dramatic economic and social crisis caused by the military junta made it inevitable that social policy would also be a recurrent issue. While the PJ reached out to the popular classes by emphasizing its historic commitment to social justice, the UCR, under the center-left leadership of Raúl Alfonsín, undertook strong efforts to mobilize popular class electoral support through a social democratic agenda that included the promise of strengthening the universal public health system and extending social assistance policies. Together with a highly professional campaign organization, this enabled the UCR to make significant inroads into the traditionally Peronist working-class and lower popular class electorates and contributed to Alfonsín's electoral victory. Through the quick implementation of a massive social assistance program, the PAN, the government was able to consolidate some of the lower popular class support during the 1985 elections. However, in contrast to the previously discussed Peronist governments, Alfonsín was never able to win significant trade union support. The enduring and deep-seated Peronist identity of the union movement as well as tactical errors led to a highly contentious relationship with organized labor. In combination with the deterioration of the economic and social situation after 1985, the PJ was soon able to reaffirm its dominant position among the popular classes and the governing alliance increasingly lacked the power to implement its original policy agenda.

At the beginning of the periods 1955–1973, 1976–1983 and 1989–2001 the overall level of popular class power was medium-low or medium, which enabled the formation of regressive governing alliances and the deterioration of social protections for low-income earners. However, the combination of different types of popular class power resources also differed between these periods and led to significantly different dynamics and outcomes.

The gradual disintegration of the Peronist governing alliance and growing military, church, and employer opposition during the mid-1950s enabled the military to seize power in 1955 and to initiate a period of authoritarian and semi-authoritarian exclusionary governing alliances that lasted until 1973. Nevertheless, throughout the whole period, the formal working class remained capable of mobilizing significant associational and structural power resources at the same time that the popularity of ISI paradigm provided discursive power resources that enabled popular class and progressive actors to frame social policy expansion in economically favorable terms. In contrast, the largely unorganized lower popular classes lacked associational and structural power resources and, due to the authoritarian and semi-authoritarian regime types during this period, they were also largely deprived of institutional power resources. This meant that while the formal working class preserved significant capacities to exert pressure, the lower popular classes suffered from extreme political marginalization. Against this background, social protections were severely reduced for the mostly informal and scarcely organized low-income earners. In contrast, social insurance programs that covered formal workers were strengthened and extended each time the governments felt the need to diffuse labor militancy or to co-opt collaborative union leaders. Instead of improving social protections for low-income earners, these concessions to the formal working class turned out to be detrimental for the lower popular classes, as they were financed in part by reducing expenditure on the universal and free public health care system and social assistance and housing policies.

During the 1976–1983 dictatorship, social policy took an even more regressive direction. During the mid-1970s, violent conflicts within the inclusionary Peronist governing alliance had significantly weakened the political left and largely paralyzed the trade union movement. Against the background of political and economic crises and the related decline of popular class power resources, the military was able to seize power and to form an exclusionary governing alliance in 1976. The abolition of democracy and hence the reduction of the popular classes' institutional power resources led again to a nearly complete political marginalization of the lower popular classes. This time, however, the associational power resources of the trade union movement were also significantly weakened, so that the military junta was not only able to radically cut social protections for low-income earners, such as social assistance and public health, but also to reduce social insurance benefits and employer contributions, which negatively affected the formal working class. Only towards the late 1970s did the union movement recover parts of its associational strength and capacity of action, which caused the dictatorship to refrain from its original plan of social health insurance privatization and even led to a significant expenditure increase in this area.

The dynamics of regressive social policy reform during the 1989–2001 period were again different. The formation of a regressive governing alliance was preceded by a decline of the overall level of popular class power resources during the second half of the 1980s. This decline was mainly explained through the rise of informal employment and unemployment, which led to a reduction of the popular classes' structural power resources, and the increasing influence of neoliberal thinking, which led to the decline of discursive popular class power resources. At the same time, union density remained relatively high and democratic rights provided the popular classes as a whole with significant electoral importance. Against this background, and in contrast to the above described authoritarian periods, the preservation of medium-high levels of institutional popular class power resources in addition to associational power resources motivated the governing alliances to make social policy concessions not only to the formal working class but also to the electorally important lower popular classes. In practice this meant that labor protests and union representation in the parliament forced the government to refrain from the complete privatization of social health and pension insurance programs and that the need to retain at least some level of lower popular class electoral support motivated the government to increase social assistance expenditure in order to extend the resources available to its clientelistic networks. Nevertheless, these concessions did not compensate for the overall regressive character of social policy reform during the period. The fact that regressive governing alliances were electorally sustainable with only relatively limited concessions to the numerically strong popular class electorate can be best understood through the interaction of democracy with low levels of discursive popular class power as a consequence of the popularity of the neoliberal paradigm. As survey data clearly shows, during the early 1990s a sizable proportion of the popular class electorate shared the neoliberal worldview and believed that regressive social policy, although initially painful, would be in their long-term interest, as it was seen to foster economic growth and employment. When, however, during the second half of the 1990s, the ongoing deterioration of the social situation more and more evidently contradicted this hope for long-term improvements, and hence contributed to the decline of neoliberal thinking in the wider public, the Menem government had to significantly increase social assistance measures to retain the lower popular class vote.

## **5.4 Governing Alliances and Social Protection for Low-Income Earners**

The formation of different types of governing alliances constituted the most important political mechanism that mediated between the distribution of power resources and the constellation of actors and interests, on the one hand, and the direction of change regarding social protections for low-income earners, on the other. Nevertheless, the processes of alliance formation and the achievement and preservation of control over the government were historically contingent processes. That is, the distribution of power resources and the constellation of actors and interests set margins within which these processes took place, but different actors' decisions regarding whether and how to participate in alliances were also subject to their strategic decisions and their pre-existing affinities or animosities, as well as period-specific events, such as economic crises and political scandals. Therefore, the concrete compositions and internal dynamics of different alliances varied significantly, which in turn also affected the social policy outcomes towards low-income earners.

## **Distribution of Power Resources, Types of Governing Alliances and Policy Outcomes**

Table 34 summarizes the empirical findings regarding four pivotal research categories for each period: the overall level of popular class power resources, the formation of different types of governing alliances, the strength and unity of these alliances, and the direction and intensity of change regarding social protection for low-income earners. If we focus our attention first on the relation between the overall level of popular class power, the type of governing alliance and the resulting social policies, the observed pattern consistently supports our initial hypothesis that higher levels of overall popular class power resources favor the formation of inclusionary and progressive alliances that extend social protection for low-income earners, while lower levels of popular class power resources favor the formation of exclusionary and regressive alliances that carry out retrenchment.


Table 34: Popular class power, governing alliances, and patterns of change regarding social protection for low-income earners, 1943–2015

\* At the beginning of each period. \*\* At the moment of the formation of the alliance. Strength refers to the extent to which the alliance effectively incorporated actors with significant power resources. Unity refers to the degree of cooperation and consensus within the alliance.

In practice, changing distributions of power resources interacted with realignments of important political actors in the rise and fall of governing alliances. Industrialization, the growth of the union movement, the increasing popularity of the ISI paradigm and the return to democracy towards the mid-20th century constituted the bases on which the popular classes became a powerful political force in Argentina. This provoked resistance on the part of the landed oligarchy, most business representatives , and the political elite. However, at the same time, it provided important incentives for centrist and progressive actors to consider alliances with popular class forces in order reach and sustain governmental power. The first successful attempt to form an inclusionary governing alliance was led by Juan Domingo Perón in 1946, who was able to win support from sectors of the military, the church, and the UCR party as well as from the labor movement and the lower popular classes. The inclusionary and progressive governing alliances of the periods 1973–1976, 1983–1989 and 2002–2015 were characterized by both a preceding increase of the overall level of popular class power and an alignment of centrist political actors and currents with the redistributive project. In turn, the demise of these alliances began with internal conflicts related to their social, ideological, and cultural heterogeneity. These conflicts finally led to the weakening of the alliances' bases of support and the realignment of some of its actors with oppositional forces. Together with a decrease in the overall level of popular class power resources linked to the decline of progressive organizations, changes in the labor market, shifts towards neoliberal thinking, or the abolition of democracy, these realignments eventually provided the ground from which exclusionary and regressive governing alliances emerged. Initially, these were sustained by a broad variety of political forces, such as representatives of different business sectors, rightwing military factions, the conservative leadership of the Catholic Church, and politicians from different parties. However, over the years, internal tensions also tended to grow within these alliances and subsequently led to their gradual disintegration. Some business sectors left the alliances because they felt that other sectors were benefiting more from their economic policies, the church slowly started to criticize the rise of social tensions, statist and neoliberal military factions disagreed on how to confront economic crises, and so on. Together with a recuperation of popular class power resources through successful processes of organization, industrial expansion, tightening labor markets, a shift towards more demand-oriented economic thinking, or the return to democracy, these realignments provided the grounds for the renewed formation of inclusionary or progressive governing alliances.

#### **Contingent Dynamics of Governing Alliances and Intra-Alliance Mechanisms of Social Policy Expansion**

A closer look at Table 34 shows that governing alliances differed significantly with regard to their composition and unity, their strength to implement social policy change, and their durability in years. This meant that the intra-alliance mechanisms of agenda setting regarding social protection for low-income earners varied.

Two types of governing alliances that implemented a strong expansion of social protections for low-income earners can be distinguished: inclusionaryredistributive governing alliances during the years 1946–55 and 1973–76 and an inclusionary-progressive governing alliance during 2003–15. The major difference between these two types of alliances consisted in the varying degree of influence of left-of-center forces within the respective alliances. While in the inclusionary-progressive alliance, left-of-center forces formed the leading circle, the inclusionary-redistributive alliances included strong left-of-center actors but were led by a politically centrist circle. Although both types of alliances implemented strong social policy expansion, the distinction helps to grasp the underlying political dynamics in each case.

The inclusionary-redistributive alliances were led by Juan Domingo Perón and his close circle, who were politically centrist and pragmatic rather than left-of-center in their worldview. However, Perón and his close circle were also aware that the main political mass basis of his alliance rested on the support of the popular classes, including both the strongly organized formal working class and the largely unorganized lower popular classes. The mobilization of these sectors meant also the incorporation of significant left-of-center political forces such as the majority of the trade union movement and the labor party in 1946 or the Peronist Youth, the combative minority faction of the union movement, and a left-leaning slum dwellers' movement in 1973. In contrast, Perón's attempts to mobilize mass support from conservative and employer sectors were much less successful. On this basis, it was mainly three political mechanisms within the governing alliance that favored the setting of a redistributive social policy agenda in favor of low-income earners. First, leftof-center forces within the alliance exerted pressure for inclusionary social policies that, in accordance with their political ideology, aimed at reducing inequalities and injustices. This mechanism can be observed especially in the early phase of the inclusionary-redistributive governing alliances when the left within these alliances was relatively strong and autonomous. Second, the fact that the sustainability of the government crucially depended on the mass support of the lower popular classes motivated the pragmatic circle at the top of the governing alliance to use inclusionary social policies as a means of gaining support and mobilizing voters. Third, social policies were often deliberately implemented in such a way that they provided state officials with considerable power over the beneficiaries. Hence, they served to enhance control over their own base of support. Taken together, these mechanisms led to the setting of a political agenda that included a strong expansion of social protections for lowincome earners.

Perón's political centrism and multi-class appeal furthermore enabled the formation of initially strong governing alliances that brought together a wide array of actors. The inclusionary-redistributive alliances won ample majorities in both chambers of the parliament, nearly all governorships and most local governments, which provided favorable conditions for the sanctioning of a significant number of profound social policy changes within a short period. At the same time, the breadth of the alliances led to significant internal tensions and struggles over the direction of policy between actors with widely differing ideological orientations. During the years 1946–55 the internal political polarization was, however, significantly less intense than during the 1973–76 period, which explains to an important degree why the latter governing alliance disintegrated much more quickly. What was common to the two governments and distinguished their legacy from that of the inclusionary-progressive alliance was that while taking up demands of left actors within the alliance, the leading circle of the alliance also took measures to marginalize the left in political terms. In the 1943–55 period, Perón and his close circle dissolved the left-ofcenter labor party and replaced many leftist union leaders. In their place he promoted the creation of the centrist Partido Peronista and fostered the emergence of rather conservative union bureaucracies. In the 1973–76 period, the leading circle of the alliance politically marginalized the leftist Peronist Youth and removed left-leaning Peronists from public offices. In addition, the paramilitary Triple-A carried out violent repression and persecution of left-wing activists. In the long run, this decisively weakened the Argentine left and provided therefore a political legacy that was unfavorable to the defense of social protections for low-income earners during the following exclusionary governments.

The inclusionary-progressive governing alliance between 2003 and 2015 was led by Néstor and Cristina Kirchner and their close circle, which in contrast to Perón, belonged to the center-left wing of the Peronist movement. The Kirchners deliberately sought to incorporate a broad range of progressive organizations into the core of the governing alliance and to build its electoral stronghold among the popular classes as well as progressive parts of the middle classes. In contrast to the 1940s and 1970s, the array of actors this time included several strong leftist unemployed workers' movements. These new actors had emerged during the 1990s and early 2000s in opposition to the PJ and UCR/FREPASO governments so that most of them were hesitant to join a governing alliance which included major parts of the PJ. The government had to show considerable commitment to the interests of these movements to win their adherence, and despite efforts to attract them, several decided to continue their activities as left oppositional forces. Against this background, within the inclusionary-progressive governing alliance, it was mainly four political mechanisms that contributed to the setting of a redistributive social policy agenda in favor of low-income earners. First, this time a center-left political current of the PJ was heading the government and was able to shape the social policy agenda. The weight of the political left within the alliance was furthermore reinforced by the incorporation of several left-leaning parties, unemployed workers' movements, and human rights organizations, which thus became able to press their demands in the internal agenda-setting process. Second, the expansion of social protections for low-income earners successfully served to create the government's most reliable electoral stronghold among the lower popular classes and to convince unemployed workers' movements and the left union wing to join the governing alliance. Third, the Kirchner governments also designed social policies, and in particular social assistance and housing programs, in such a way that they served to co-opt and control social movement leaders, lower rank politicians, and beneficiaries. Fourth, the fact that several unemployed workers' movements remained in the opposition and frequently undertook roadblocks motivated the government to use social policies as a tool to contain and defuse protest activity.

Due to the medium-high level of the overall popular class power resources and the fact that the governing alliance was able to incorporate a majority of the labor, the unemployed workers' and human rights movements as well as the PJ party, the alliance was quite broad and strong, which facilitated the sanctioning of profound social policy reforms within a short time. Although there were certain tensions between left and centrist actors as well as between Peronist and non-Peronist actors, these tensions were much lower than during the 1973-76 period, so that they neither hampered social policy change nor led to an early disintegration of the alliance. In contrast to the inclusionary-redistributive alliances, the inclusionary-progressive alliance provided significant resources and public offices to center-left and left political actors and thereby fostered the growth of these forces within the governing alliance.

A moderate expansion of social protections for low-income earners was undertaken by the truncated-progressive governing alliance during the 1983– 89 period. As in the case of the inclusionary-progressive alliance, the government was led by center-left political actors, in this case Raúl Alfonsín and his close circle, who represented the social democratic wing of the UCR. The two main mechanisms that led to the incorporation of inclusionary social policies into the government's agenda were that the left wing of the UCR attained significant influence and that proposals for an extension of social protections were part of an attempt to broaden electoral support among the lower popular classes. This attempt was, however, not very successful, as most of the workingclass and lower popular class electorate maintained their traditional adherence to the PJ. In this sense, the alliance remained truncated.

In contrast to the inclusionary-redistributive and the inclusionary-progressive alliances, the truncated-progressive alliance was weak in political terms. Its power rested nearly completely on the UCR party and its predominantly middle-class base of support. Furthermore, the alliance never achieved a majority in the Senate, so that taken together, the possibilities to impose profound social policy change were much more restricted. Against this background, the main inclusionary reform project, a unified universal health system, failed in the face of the opposition of trade unions, the PJ, and employers. The only significant expansion of social protections for low-income earners was reached through the introduction of new social assistance and unemployment protection measures.

During the 1989–2001 period, a moderate reduction of social protections for low-income earners was undertaken. The fact that the social situation of the lower popular classes deteriorated more profoundly than during other periods of retrenchment was related to the confluence of the reduction of social protections with a dramatic deterioration of the labor market. Except during the last two years of the period, the government was sustained by a broad-regressive alliance. The alliance can be described as regressive because the agendasetting process was clearly dominated by the conservative-neoliberal wing of the PJ, business representatives, and neoliberal think tanks. Nevertheless, the alliance was also broad in the sense that, due to historical affinities and the proliferation of clientelistic networks, it also included large parts of the trade union movement and captured the majority of the working-class and lower popular class vote. Left-leaning sectors of the PJ and the trade union movement, however, occupied a clearly subordinate position in the alliance and had little influence on the agenda-setting process. As the social situation deteriorated, the dominance of the neoliberal paradigm in the public opinion declined and new oppositional popular class actors emerged and grew in strength. Against this background, lower popular class support for the governing alliance increasingly eroded. Due to the dependence of the government on at least a major share of the lower popular class vote, it responded to the gradual erosion of support and the rise of competing political actors with the extension of social assistance policies. It was mainly two mechanisms that characterized this response to lower popular class power. One the one hand, most of the social assistance expenditure was channeled to clientelistic Peronist networks in poor neighborhoods. Thereby the government aimed at enhancing its direct control over its lower popular class base and at countervailing the gradual emigration of low-income earners towards other political forces. On the other hand, a smaller share of resources was used to contain and defuse unemployed workers' movements' protests through directly negotiated concessions. However, all in all, the increase of social assistance expenditure far from compensated for the dramatic reduction of social protections for low-income earners in the fields of pension and health care policies.

Due to the variety of actors included in the governing alliance and majorities in both chambers of the parliament, the alliance was strong and able to implement profound social policy changes. Ménem was furthermore highly effective in dealing with internal tensions by applying stick-and-carrot tactics that led to the alignment of the more cooperative union leaders and the marginalization of the more combative union factions as well as the minority of progressive Peronist politicians. During the second half of the 1990s, the deterioration of the social situation because of the government's neoliberal policies led nevertheless to a gradual erosion of the government's popular class support. Against this background, the 1999 elections resulted in the replacement of the Peronist government by an alliance of the UCR and the FREPASO. The new government was dominated by the neoliberal faction of the UCR and largely continued the social policy agenda of the preceding government. Due to a considerably smaller base of support, an economic downturn, and growing resistance from trade unions and social movements, the alliance was however considerably weaker and largely disintegrated within only two years.

The periods 1955–73 and 1976–83 were characterized by exclusionaryregressive governing alliances. During both periods, popular class actors were de facto excluded from the alliances while right-wing military and business leaders gained far-reaching influence on social policy.195 Against this background, social protections for low-income earners were significantly reduced.

<sup>195</sup> The first few months of the Frondizi government in 1958 might be considered the only significant exception.

Social policy change was particularly regressive during the military dictatorships in the years 1955–58, 1966–73, and 1976–83. During the civilian UCR governments of 1958–66, the deterioration of social protections for low-income earners was more limited, yet the prohibition of Peronism and the repeated military interferences also inhibited significant moves towards inclusionary social policy. While these governments from time to time responded to trade union protests with concessions in the area of social insurance that benefited formal workers, there were no significant political mechanisms that would have led to an expansion of social protections for low-income earners.

The governing alliances during these periods were medium-strong and beside military, employer, and church support also enjoyed some level of support from the middle classes and non-Peronist parties. Due to the severe restriction of institutional checks and balances, they were able to carry out significant social policy change within a short time. However, internal tensions led to repeated internal coups and finally the decline of these alliances.

Table 35 summarizes the main political mechanisms related to the expansion of social protections for low-income earners. As can be observed, the intra-alliance processes of agenda-setting in this respect were consistently related to the underlying distributions of power resources within and without the alliance. On the one hand, the distribution of power resources permitted progressive and lower popular class actors a certain level of direct influence on the agenda-setting during the progressive-inclusionary, the progressive-truncated and to a lesser degree during the inclusionary-distributive governing alliances. On the other hand, the leading circles of these alliances responded to the power resources of the lower popular classes and progressive actors by proposing and implementing social policies that either mobilized and consolidated their support for the alliance or aimed at controlling or containing their political activities. The extension of targeted social assistance programs during the broadregressive alliance of the 1990s also has to be understood as a response to the lower popular classes' institutional and associational power resources, as they aimed at controlling beneficiaries through clientelistic practices and at containing protests.

The surprising coherence in the direction of social policies of each governing alliance despite their often very heterogeneous bases of support was related to the generally highly hierarchical structure of governments, organizations, and alliances in Argentina. In practice, presidents and their close circles possessed sufficient power to control state bureaucracies and lower-level political office holders as well as their own parties.<sup>196</sup> As a result, top-down mechanisms of agenda-setting prevailed during virtually all periods, enabling

<sup>196</sup> A variety of factors contributed to this concentration of power, such as the strong constitutional powers of the president, the capacity to exchange bureaucratic personnel, a high dependency of lower levels of government on resource transfers from the central government, and the de facto weakness of party internal institutions.

leading circles to impose their agendas over the objections of dissenting minority factions within their alliances.

Table 35: Intra-alliance mechanisms related to the expansion of social protections for lowincome earners, 1943–2015


\* Mechanisms in bold were of particular importance within the respective alliance.

## 6 Conclusions

The empirical historical analysis of more than seven decades has provided an encompassing account of *when* and *why* social protections for low-income earners have been extended or cut in Argentina. The findings furthermore provide significant empirical support for the initially developed theoretical framework, as they show both *that* and *how* the popular classes' power resources and historically specific constellations of actors and interests shaped the dynamics of governing alliance formation and social policy change. The following chapter will offer an overview of the main findings and reflect on their implications for welfare state theory.

## **6.1 Main Findings: Power, Alliances, and Redistribution**

In congruence with the theoretical framework, the empirical analysis shows that high levels of popular class power resources favor the extension of social protections for low-income earners while low levels of popular class power favor retrenchment. This relationship, however, was not linear and direct but depended crucially on historically contingent processes of governing alliance formation as the most important mechanism that causally linked the underlying distribution of power resources to specific outcomes regarding social protections for low-income earners. The dynamics of these processes of governing alliance formation, in turn, were heavily influenced by historically grown constellations of actors and interests as well as the current social, politico-institutional and economic contexts. In order to provide a more detailed overview of the main findings, I will first make a compact assessment of the initial hypotheses, then work out the crucial political mechanisms that shaped processes of expansion and retrenchment of social protections for low-income earners and finally illustrate in a strongly summarized way how these mechanisms operated over the whole period of study in Argentina.

## **Assessment of the Initial Hypotheses**

*Popular class power and redistribution:* In accordance with the traditional power resources approach (Esping-Andersen [1990] 1998; Huber and Stephens 2001; Korpi 1983) and one strand of Marxist state theory (Poulantzas [1978] 2000), I have assumed that the redistributive character of public policy can be best understood as the outcome of relations of power. The empirical analysis of the Argentinian case provided strong support for this hypothesis. As has been shown in Chapter 5, high levels of popular class power resources consistently corresponded with an extension of social protections for low-income earners, while low levels were consistently associated with retrenchment. This consistency, however, does not mean that there was an automatic functional fit between the two categories. Rather, the distribution of power resources constituted a framework that enhanced or decreased the chances of different actors of imposing their social policy interests at the same time that it provided incentives to actors to pursue certain alliance strategies and to make social policy concessions to certain sectors and political forces. Within this framework, the concrete political dynamics and outcomes remained always contingent on the strategies pursued by different actors and a range of events specific to the historical moment, such as political scandals, economic crises, and so on.

*Distribution of power resources among different popular classes:* I furthermore assumed that the expansion of social protections for low-income earners becomes more likely when this subsector of the popular classes achieves considerable power resources, as this might, (1) enable them to exert pressures on policy-makers, (2) enable them to become themselves participants in the design and implementation of social policy, or (3) at least provide incentives to political elites to mobilize their support through social policies. This hypothesis is strongly supported by the empirical findings. However, in practice, the three mentioned mechanisms were of different importance. Until the mid-1990s, low-income groups, such as peasants or unemployed and informal workers, were not well-organized and so their associational power resources were comparatively low. Due to their marginalized position in the system of production and their high vulnerability on the labor market, their structural power resources have also been low throughout the period. This meant that (1) their capacity to exercise external pressures and to extract significant social policy concessions was severely limited. The weakness of political organization meant that (2) a direct influence of lower popular class actors on social policy agendas through the active participation in governing alliances was the exception. After the mid-1990s, the emergence of lower popular class mass movements changed this situation to a certain degree. External pressures through these movements and later the participation of some of them in the Kirchnerist governing alliance significantly contributed to the extension of social assistance policies. Nevertheless, seen over the period of study as a whole, the most important mechanism that related the power resources of the lower popular classes to an expansion of social protections was (3) that institutional and discursive power resources provided political elites with significant incentives to mobilize their electoral support and to promote their affiliation to parties and movements associated with the government.

*The necessity of broader alliances:* Due to the lack of economic resources and the spatial dispersion of low-income earners, I have assumed that the construction of strong lower popular class movements is challenging and hence that the extension of social protections usually requires the formation of broader alliances with other political forces. I have furthermore assumed that governing alliances can influence social policy much more strongly than can oppositional alliances. Therefore, relatively limited differences in the distribution of power resources can be associated with significant differences in social policy when they enable the assumption of a different kind of governing alliance. The empirical analysis provides strong support for this hypothesis. Social protections for low-income earners were only significantly extended by inclusionary governing alliances that actively sought to incorporate the lower popular classes as an important electoral and associational support base. External pressures by unemployed workers' and slum dwellers' movements as well as by broad-based oppositional alliances, such as the FRENAPO, had comparatively limited direct effects on the level of social protections for low-income earners. Although these pressures led to concessions, such as the extension of certain social assistance programs, their most significant impact was indirect insofar that they contributed to the decline of the respective regressive governing alliance and to the subsequent formation of an inclusionary alliance, which eventually initiated deeper social policy change.

*Actors and social protections for low-income earners:* I have assumed that the most likely allies for redistributive alliances are left and popular classbased political parties, trade unions, and social movements. Right political parties, right-wing military leaders, neoliberal IFIs, and employers were assumed to pursue retrenchment or inegalitarian types of social policy. As can be seen in Table 36, the empirical analysis supports these initial hypotheses, yet also shows that the constellations of actors and interests were even more complex than expected.


Table 36: Overview of the constellation of actors and interests regarding the extension of social protections for low-income earners

Due to the politically heterogeneous bases of many political actors, in practice different internal currents took distinct positions towards social protections for low-income earners. This finding is nevertheless congruent with the power resources approach, as it was the left currents within these actors that consistently supported increases in social protection for low-income earners and right internal currents that opposed them. Furthermore, major factions of the union movement and the two biggest parties can be described as politically centrist. Over the years these centrist factions oscillated between supporting and opposing social protections for low-income earners. Whether they inclined to one side or another was in practice importantly shaped by the levels of institutional and discursive power resources of the popular classes. When these were high, the centrist factions inclined towards supporting social protections for low-income earners; when they were low, they tended to oppose them. High levels of institutional power resources meant that the numerically strong lower popular classes attained considerable electoral importance, which provided strong incentives for centrist political actors to attempt to mobilize their support. At the same time, high levels of discursive power resources provided them with capacities to successfully frame inclusionary social policy as being beneficial for the society as a whole and thus enabled them to engage in alliances with left and lower popular class forces without necessarily renouncing their betteroff and at times more conservative middle-class and formal labor support. In practice, this mechanism was of high importance for the politics of social protection for low-income earners, as during no period were left and lower popular class forces able to form a sustainable governing alliance without the support of centrist currents and organizations. This fact was reinforced by the crosscutting divides that characterize the Argentine political system. On the one hand, positions towards redistributive policies were strongly shaped by the leftright divide. On the other hand, major political actors organized around the divide between Peronism and anti-Peronism. This often meant that due to historically grown identities and affinities, political alliances between left and non-left Peronist organizations and currents were more likely to emerge than between left currents that belonged to Peronist and anti-Peronist political traditions.

*The development of a truncated welfare state:* I have furthermore assumed that truncated systems of social policy that provide social protections mainly to formal workers and the middle and upper classes result from exclusionary and regressive governing alliances that aim to foster middle-class support and control trade union agitation. This hypothesis is empirically supported; however, the findings showed that the development of truncated social policies was at times also favored by more progressive governing alliances and actors. There were several reasons for why exclusionary and regressive governing alliance were particularly prone to promote truncated policies. First, based on their rightist ideological orientation, these exclusionary alliances either preferred highly stratified or market-oriented social policies. Both types of social policy were regressive and provided social protection mainly to the better-off segments of the society. Second, these exclusionary governments in each case inherited a range of social policies that benefitted both the formal working class and the lower popular classes. As the latter had lower structural power resources and were scarcely organized, and therefore had weaker capacities of resistance, social protections for low-income earners were cut quicker and more profoundly. The trade union movement, in contrast, was able to actively resist the retrenchment of social insurance protections for formal workers and at times even able to extract positive concessions. Rightist governing alliances furthermore used truncated social insurance policies to contain trade union agitation or to co-opt and divide the popular classes and their organizations. In sum, the interplay of these factors led to the outcome that exclusionary governing alliances pushed the Argentine social policy regime strongly in a truncated direction. Although inclusionary and progressive governing alliances importantly expanded social protections for low-income earners and hence pushed the social policy regime in a less truncated and more egalitarian direction, their policies were not completely coherent in this sense because their leading circles also used fragmented and truncated social policies to co-opt union leaders and thereby to attain more control over their own social bases.

*Path dependency effects:* Finally, I have assumed that inegalitarian social policies, especially exclusionary social insurance schemes, tend to create vested interests and hence path dependency effects that pose additional difficulties for progressive social policy reform, as the resistance from employers and conservatives might be reinforced by popular class actors who benefit from the respective policies. The empirical findings support this hypothesis but at the same time also indicate that the strength of path dependency effects highly depends on the types of vested interests created by a particular policy. For example, the system of social pension insurance has been restructured several times, while similarly far-reaching change in the area of social health insurance was successfully blocked by the trade union movement. This has very much to do with the fact that, in contrast to the pension system, most social health insurance funds are controlled by trade unions and this creates strong vested interests in these organizations as well as among their members. In practice, this meant that the defense and strengthening of the social health insurance funds constituted one of the highest political priorities of organized labor. Therefore, resistance to reform was more intense in this policy area and governments took up the demands of the union movement when they perceived it necessary to contain labor militancy or to negotiate union support for reform in other policy areas. Nevertheless, even in the case of health insurance, when it comes to the possibilities of reform, much depended on the strategy of the government. For example, both regressive and progressive governments were able to gradually alter the relative amount of resources provided to the universal public health system and the social health insurance funds in one or another direction. In addition to this, governments were able to gradually transform the regulations regarding affiliation and benefits, which was in practice of high importance to the question of how many low-income earners were able to access social health insurance and of what quality the coverage was.

## **Main Factors and Mechanisms behind Policy Change**

The empirical analysis revealed that the level of social protection for low-income earners underwent strong variation during the 72 years under study. Rather than being linear, this change was characterized by alternating periods of expansion and retrenchment. While the length and reform intensity of the different periods varied, the direction of policy change within each period was very coherent. Against this background, it has been possible to undertake a twofold methodological approach to identifying the main factors and mechanisms associated with the extension and retrenchment of social protections for low-income earners. In a first step, the study traced the processes of policy change and thereby reconstructed which actors expanded or retrenched social protections for low-income earners under what circumstances and with what kind of goal and strategy. In a second step, the different periods of expansion and retrenchment were compared in order to achieve a clearer picture of the political patterns that were associated with progressive and regressive social policy reform cycles. In combination, these two approaches allowed for a quite concise identification of the main factors and mechanisms that shaped the processes of social policy change towards low-income earners in Argentina. Figure 16 provides a summarized illustration of the results.

The main driving forces behind changes in the direction of social policy were alterations in the distribution of power resources and in the constellations of actors and interests. Any period of expansion of social protections for lowincome earners in Argentina was preceded by national and international structural changes and political developments that led to an increase of the overall level of power resources of the popular classes and a realignment of political forces. These realignments were characterized by rising tensions among antiredistributive forces, on the one hand, and a growing disposition of centrist political forces to forge alliances with left and lower popular class forces, on the other. Taken together, these developments generated favorable conditions for the formation of inclusionary and progressive governing alliances. In practice, the concrete processes of alliance formation were different in each period, and so were the alliances' compositions, political orientations, and capacities to impose far-reaching policy change. Notwithstanding, in all cases, these inclusionary and progressive governing alliances implemented an expansion of social protections for low-income earners. Within these alliances it was mainly four political mechanisms that shaped the respective agenda-setting processes.

Figure 16: Summarized illustration of the main factors and causal mechanisms shaping the evolution of social protections for lowincome earners in Argentina

First, the leading circles of these alliances used progressive social policies as measures to mobilize the electoral and associational support of the lower popular classes, and hence, to strengthen or consolidate their bases of support. Second, left and lower popular class actors and currents within these alliances used their power to influence the agenda-setting process in favor of low-income earners. Third, the implementation of inclusionary social policies was used to co-opt and control beneficiaries and political leaders, and hence, to increase the control of the leading circle over the bases of support of the alliance. Fourth, governing alliances implemented redistributive social policies as measures to contain and diffuse lower popular class protests.

As the Argentine popular classes have been organized in multiple movements and currents with different ideological orientations, inclusionary alliances tended to be strongly heterogeneous. This characteristic was additionally reinforced by the fact that the political left was never strong enough to form a governing alliance on its own and hence always depended on the collaboration of centrist political actors. This heterogeneity, in turn, favored the emergence of internal tensions as soon as political or economic difficulties arose and the promotion of concrete policies revealed conflicts of interest between some of the allied forces. Internal tensions were of varying intensity during the different inclusionary governing alliances, but in all cases, they sooner or later led to a weakening of the alliance. In parallel, it could be observed that the implementation of redistributive policies aroused growing resistance from employers and right-wing political forces, which gradually contributed to processes of reconstruction of anti-redistributive alliances. Political and economic crises and a subsequent decrease of popular class power resources eventually provided political opportunities for these alliances to replace the weakened inclusionary governments either by military coup or democratic elections. As soon as the exclusionary alliances took control of the government, they initiated a series of reforms that drastically cut social protections for low-income earners. After some time, however, internal heterogeneity and economic crises also produced increasing tensions within the exclusionary alliances. Recurring phenomena were, among others, that certain sectors of the economic elite started to complain that other sectors would benefit more from state policy, that different military factions disagreed on how to deal with popular resistance, and that the church started to criticize the rise of social tensions. In combination with a recovery of popular class power resources, this allowed for renewed attempts to form more inclusionary governing alliances.

In order to provide a concise overview of the above described pendular movements of social politics in Argentina, Figure 17 summarizes the operation of the main causal factors and mechanisms over the period of seven decades.

In addition to the formation of inclusionary and exclusionary alliances, the extraction of social policy concessions by means of exercising oppositional pressures constituted another mechanism through which different distributions of power resources and constellations of actors were causally linked to changing levels of social protection.

Notably, during approximately five of the seven decades under study, this mechanism was mainly related to trade union pressures and resulted in concessions regarding social insurance for formal workers. As most low-income earners lacked social insurance coverage, they hardly benefitted from these concessions. Only after the emergence of strong unemployed workers' movements during the mid-1990s did oppositional pressures lead to significant concessions in the area of social assistance policies and hence become more important in influencing the evolution of social protections for low-income earners.

Considering that social policy reform is commonly characterized by path dependencies and gradualism, the extraordinarily stark contrast between the strong expansion of social protections for low-income earners during some periods, and their dramatic reduction during others, appears at first glance surprising. However, the empirical analysis revealed several aspects that explain these stark contrasts, some of which were quite particular to the Argentine political system and culture. First, the period of study was shaped by an extraordinarily strong polarization between inclusionary and exclusionary political projects, neither of which was able to lastingly alter the balance of power in its own favor. Juan Carlos Portantiero (1973) even referred to this situation as a hegemonic tie. The strength of the polarization resulted in large part from the aggressively anti-redistributive stances of most of the employers and other right-wing forces as well as the populist and confrontational style of Peronist politics. Related to this, Argentine politics were characterized by a relative weakness of tripartite and other institutional structures that could have promoted compromises and middle ways. Second, most Argentine political organizations, alliances and governments functioned as highly personalistic and hierarchical apparatuses. This, in turn, enabled the leading circles of the different governing alliances to impose quite coherent and deep social policy reform agendas despite their highly heterogeneous bases of support. So, mechanisms that might have promoted the generation of political compromises tended to be weak not only outside but also within governing alliances, which, as a side product, accelerated the realignment of former allies as opponents on various occasions. Third, within all governments, the direction of social policy was subordinated to the overriding goals of economic policy and hence depended highly on whether a demand-oriented or a supply side-oriented approach was adopted. In this regard the inclusionary and exclusionary alliances differed consistently. Fourth, as social protections for low-income earners often resulted from top-down strategies to mobilize and control lower popular class support, they were often not institutionalized as social rights but implemented as loosely regulated assistance measures that provided political leaders with opportunities for the application of clientelistic strategies. The resulting lack of solid legal structures and their perception as gifts rather than conquered social rights made their later retrenchment politically much easier. Fifth, in contrast to social protections for formal workers, social protections for low-income earners lacked a strong autonomous support basis, as the lower popular classes did not develop comparable associational and structural power resources. This enabled exclusionary governments to advance much quicker with the dismantling of social protections for marginalized groups than with the retrenchment of social insurance.

## **The Argentine Case in Regional Perspective**

While the extraordinary radicality of changes might be an Argentine particularity, the expansion of social protections for low-income earners often resembled regional trends. When in Argentina the first Peronist governing alliance during the 1940s and 1950s massively expanded health care, pension, and social assistance provision to low-income earners, several other Latin American countries incorporated hitherto excluded groups into their social security systems (Haggard and Kaufman 2008, 79–113; Segura-Ubiergo 2007, 24–76). Although not all reforms reached the poor, several countries undertook significant efforts to expand social protections to the poor. Costa Rica, among other countries, approved social insurance legislation during the 1940s that prioritized low-income earners and was universal in its aspiration (Huber and Stephens 2012, 91; Rosenberg 1983, 45–104). Uruguay implemented and universalized family allowances and included rural workers and other hitherto excluded low-income groups into its social security system (Filgueira 1995, 17– 23). In a similar vein, when during the 1970s the second Peronist government strengthened the universal public health system, extended low-income earners' access to pensions and raised social assistance expenditures, governments in several other Latin American countries pursued progressive social policy agendas. Costa Rica advanced towards the de facto universalization of health care coverage, expanded access to pensions among the poor, and set into motion massive social assistance programs (Martinez Franzoni and Sánchez-Ancochea 2013, 20–23; McGuire 2010, 86–87). In Chile, the governments of Eduardo Frei and Salvador Allende strongly strengthened the provision of basic health services, education, and housing for peasants and marginalized urban sectors (Arellano 1985, 413–414). And finally, when the Kirchnerist government universalized pensions and child allowances and expanded housing policies during the 2000s and early 2010s, the regional political landscape experienced an unprecedented left turn and social protections for marginalized groups were extended in most Latin American countries (Garay 2010; Levitsky and Roberts 2011; Lustig 2015; Pribble 2013). Among other countries, Brazil massively expanded targeted cash benefits through the Bolsa Familia program (Hunter 2011, 316–318). Chile strongly extended pension coverage among low-income earners and Uruguay significantly improved family allowances and health care provision for the poor (Pribble 2013, 39–119).

Interestingly, all mentioned examples have in common that they took place after a variety of developments led to a considerable strengthening of the overall level of popular class power resources and the formation of inclusionary governing alliances. They took place during periods of democracy and in contexts where at least one important political force made significant efforts to mobilize low-income voters, which increased the institutional power resources of the popular classes in general and of the lower popular classes in particular. Furthermore, in all examples, social policy expansion took place in contexts where pro-redistributive development and social policy paradigms were influential in the wider public as well as in expert communities, which implied a comparatively high level of discursive popular class power resources. Between the 1940s and 1970s, the ISI development paradigm became increasingly popular throughout the region and provided progressive forces with the ability to credibly frame social policies in favor of low-income earners as being beneficial for economic development and hence ultimately benefitting all social classes. At the same time, the universalistic ideas of the Beveridge-model and progressive "sanitarista" public health expert circles became increasingly influential in the regional social policy debate. In a similar way, the discrediting of neoliberalism and the emergence of a rather more demand-oriented and stateinterventionist economic paradigm during the 2000s, combined with a revival of universalism as a guiding social policy idea in regional expert communities and international organizations, provided a favorable discursive context during the recent phase of social policy expansion. In addition to this, most of the mentioned examples took place during periods of comparatively fast economic growth and high commodity export prices, which led to a tightening of labor markets, and hence, increased the structural power resources of the popular classes.<sup>197</sup> And finally, all examples were characterized by an increase of the associational power resources of the popular classes through a strengthening of left parties and currents, social movements, or trade unions. In each case, social policy expansion for low-income earners was to an important degree a response to these increases of the popular classes' power resources. In Costa

<sup>197</sup> For an overview of economic developments in Latin America during the 20th century, see Hofman (2001). For an overview of long-term commodity price developments, see Jacks and Stuermer (2016). Costa Rica during the 1940s might be considered an exception, due to the still very low level of industrialization and a variety of economic difficulties during that decade. Nevertheless, also in this case the considerable expansion of banana plantations increased the structural power resources of the popular classes, which found its expression in the relative effectiveness of plantation workers' strikes (Hytrek 1999).

Rica during the 1940s, the government of Rafael Ángel Calderón Guardia responded with the incorporation of the communist party and the trade union movement into his governing alliance. The legislation of an inclusionary social insurance system constituted in this context the outcome of the increased influence of popular class forces and progressive Catholic currents within the alliance and the alliance's attempt to mobilize lower class support (Huber and Stephens 2012, 91–93; Hytrek 1999; Yashar 1997, 101–151). During the 1970s, social policy expansion under the Presidents José Figueres Ferrer and Daniel Oduber was motivated by attempts to deal with an increasingly strong left wing in the PLN, intense electoral competition for low-income earners' votes among the PLN, Social Christianism and the left as well as growing challenges to the PLN's hegemony among peasants through the growth and militancy of leftist rural movements (Martinez Franzoni and Sánchez-Ancochea 2013, 20–23; McGuire 2010, 86–87; Yashar 1995, 85–87). The expansions of social protections for low-income earners in Uruguay during the 1940s, 1950s, and 2000s constituted responses to tight electoral competition for the votes of the lower popular classes, pressures from an autonomous and increasingly strong labor movement, and the influence of progressive political currents and parties. While progressive politicians mainly operated as currents within the two major centrist parties during the 1940s and 1950s, the Partido Colorado and the Partido Nacional, during the 2000s they were able to push social policy expansion through the governing center-left Frente Amplio party (Filgueira 1995, 17–23; Pribble 2013, 39–119). In Chile, during the 1960s, 1970s, and 2000s the strengthening of the popular classes' power resources found its expression in a strengthening of left parties, particularly the Partido Socialista de Chile, and an increased electoral competition for the votes of labor and the lower popular classes. During the 1960s and 1970s, these developments were reinforced by growing labor and social movement activism and a strengthening of the left wing of the Partido Demócrata Cristiano. This constellation of power resources motivated the Christian Democratic government of Eduardo Frei after 1964 to implement progressive social policies and to incorporate popular class forces into his electoral and organizational base of support. Between 1970 and 1973, social policy expansion was continued by the left government of Salvador Allende (Arellano 1985; Huber and Stephens 2012, 90–91). During the 2000s, social policy expansion for low-income earners was promoted by governing alliances under the leadership of the Partido Socialista de Chile (Pribble 2013, 39–119). Similarly, in Brazil, the gradual increase of the popular classes' power resources found its expression in a strengthening of peasant and labor activism and the formation of governing alliances under the leadership of the center-left Partido dos Trabalhadores, which expanded cash transfers and social services covering the poor (Hunter 2011). In sum, the contextual factors, political constellations, mechanisms, and subsequent social policy outcomes were in all examples highly consistent with the theoretical framework developed in the present study.

Although this brief overview does not constitute anything close to a comprehensive regional analysis, it nevertheless strongly suggests that the theoretical approach developed in this study has explanatory power beyond the Argentinian case and constitutes a fertile perspective to explore the evolution of social protections for low-income earners in other Latin American countries.

## **6.2 Theoretical Implications**

So far, long-term studies have been predominantly interested in elaborating theoretical explications for social spending variation and general welfare regime development in Latin America (Filgueira 2005; Haggard and Kaufman 2008; Huber and Stephens 2012; Huber, Mustillo, and Stephens 2008; Mesa-Lago 1978; 1989; Segura-Ubiergo 2007). However, Latin American social policy regimes and spending have expanded without generating social protections for low-income earners at any comparable pace. While this paradox has dramatic social consequences, severely limits the life quality and prospects of hundreds of millions of Latin Americans, and increases social tension and inequality, few studies have focused explicitly on the politics of social protection for low-income earners. Those that have focused on this issue cover a relatively short timeframe, which inhibits evaluation of the role of major economic, social, and politico-institutional factors whose impacts become visible only over longer periods of time (Garay 2010; Pribble 2013). The most important theoretical contribution of the present study was, therefore, the development of the first comprehensive, coherent, and specific approach to analyzing the longterm politics of social protection for low-income earners in Latin America. The approach is comprehensive as it goes far beyond single-factor explications and instead merges ideas and concepts from different theories and approaches and adapts them to the Latin American context and to the issue of social protection for low-income earners. The approach is coherent as it does not simply add explanatory factors and categories but systematically connects them. It explains how different economic-structural, institutional, discursive, and politico-associational factors shape the constellations of power and interest in society, and through this, how they provide at times more and at times less favorable conditions for the implementation or expansion of social protections. Finally, the approach is specific as it focuses explicitly on the issue of social protections for low-income earners and describes the political mechanisms that relate underlying constellations of power and interest to social policy outcomes.

The strongest theoretical influence stemmed from power resources theory and hence the empirical analysis produced first and foremost implications for that approach. However, due to the discussion and incorporation of ideas and concepts from several other approaches and theories, a range of suggestions and implications for these can be derived from the analysis as well.

## **Implications for Power Resources Theory**

In general, the empirical findings provide strong support for the central theoretical claim of the power resources approach that left and working-class forces promote broad and redistributive social policies while right and capitalist class forces promote retrenchment and regressive reform. Regarding Latin America, power resources theorists have furthermore argued that the length of democracy decisively influences the redistributive character of social policy in the sense that longer democratic records are related to more redistributive effects of social spending (Huber et al. 2006; Huber and Stephens 2012). Again, this argument is strongly supported by the empirical findings. All periods of major expansion of social protections for low-income earners in Argentina have been characterized by democratic rule and the tracing of the processes of alliance formation and agenda-setting also showed how and through which mechanisms democracy affected political dynamics and different actors' strategic decisions.

At the same time, the findings strongly supported the modifications and specifications of the approach made in the theory chapter. These proved able to provide empirically sustained explications for the development of social policy in complex contexts that were difficult to analyze with the original, more parsimonious version of the approach. First, the modified approach uses the broader concept of popular class power resources instead of working-class power resources. This choice for the broader concept turned out to be empirically important. The lower popular classes such as the peasantry, informal and unemployed workers are much larger in Argentina and other Latin American countries than in the European countries that constituted the main empirical bases for the development of the original power resources approach. In this context, a focus on the working class alone would have been too narrow to determine when the distribution of power resources was favorable or not for the implementation of progressive social policies. Peasants, informal and unemployed workers were of high electoral importance in practice and their social movements influenced social policy making through the organization of oppositional protests and roadblocks as well as participation in inclusionary governing alliances. Indeed, whether social policy reform mainly favored the formal working class or also the lowest income groups depended decisively on the relative distribution of power resources among the popular classes. Social policy was particularly progressive when a strong overall level of popular class power resources coincided with a relative strength of the lower popular classes.

Second, in order to come to a meaningful assessment of the relative strength of the political left it turned out to be insufficient to look at political organizations as ideologically homogeneous units. The modified approach therefore also considers the political divides and struggles that take place within political organizations. In Argentina, neither of the two major parties during the period under study, the PJ and the UCR, can be categorized as left or right. In order to understand their role in social politics, it was pivotal to look inside these parties and to assess the relative strength of left and right internal currents and the conditions under which the centrist factions of these parties tended to ally with either of the two sides. Based on this modification, the study was able to show that the left-right divide was of crucial importance even in a party system that was organized mainly around another divide, namely between Peronism and anti-Peronism. It was consistently left internal currents that promoted and defended social protections for low-income earners and right internal currents that favored regressive social policy reform. Similar observations were made regarding the trade union movement. While conservative union factions tended to focus their political demands on social insurance benefits that favored exclusively the formal working class, left-leaning union factions promoted universalistic and redistributive social policies based on a broad definition of the working class that explicitly included informal and unemployed workers. It seems promising to apply this theoretical modification also to the study of regressive social policy reforms undertaken by left and social democratic parties during the 1990s and 2000s in several Latin American and industrialized countries. It is likely that in these cases as well the policy shift was related to the emergence and increasing influence of right internal currents within these parties.

Third, the empirical findings confirm that considering four instead of only one type of power resource can significantly enhance the analytical capacity of the power resources approach. Not only the associational strength of the popular classes, but also their institutional, structural, and discursive power resources were pivotal to understanding the redistributive character of social policy reform. Indeed, as has been shown, the level of associational power resources of the popular classes remained relatively stable over the period of study while in contrast social protections for low-income earners underwent periods of significant expansion and retrenchment. Hence, a focus on associational power resources alone, as in the traditional power resources approach, would clearly have been insufficient to explain social policy change. Yet, if we take the four different types of power resources into account, the levels of social protection for low-income earners and the combined levels of popular class power resources developed in a congruent manner. Moreover, the process tracing analysis was able to show why and how variations in the different types of power resources affected processes of governing alliance formation and social policy reform. Thus did the rise and fall of democracy, and hence varying levels of institutional power resources, deeply affect the capacities of the popular classes and their organizations to take advantage of their numerical strength and to influence the formation of governments and their political agendas. In practice, the level of institutional power resources interacted closely with the level of discursive power resources. When import substitution and post-neoliberalism constituted the dominant development paradigms in public opinion, left and popular class forces were successfully able to frame redistributive social policy as being beneficial for the overriding goal of economic development, and hence, as being beneficial for the society as a whole. This, in turn, enhanced the capacities of pro-redistributive forces to mobilize mass support and facilitated the formation of inclusionary governing alliances that united a range of different actors with heterogenous social bases. In contrast, when neoliberalism dominated public opinion, mass support for social protections for low-income earners was considerably more difficult to attain and centrist actors were more hesitant to support such policies, as through the neoliberal perspective these appeared both economically and morally harmful. Finally, processes of industrialization and de-industrialization and varying levels of unemployment and informality affected the capacity of labor to exercise economic pressure through the interruption of the process of production, and hence, led to varying levels of structural power resources of the popular classes. In practice, structural power resources interacted closely with associational power resources and constituted an important basis for organized labor to pursue the interests of the formal working class.

In addition to a more adequate assessment of the overall level of popular class power resources, the distinction of four types of power resources also proved useful to understanding the distribution of power resources between different segments of the popular classes and its effects on the dynamics of alliance formation and social policy-making. In the Argentine case, during most of the period covered in this study, structural and associational power resources were highly concentrated on the formal working class while institutional and discursive power resources were more equally distributed among the different segments of the popular classes. This meant that the relative political weight of the lower popular classes, and hence the incorporation of their interests into the social policy agenda, was in practice highly dependent on the levels of institutional and discursive power resources. Furthermore, organized labor was able to defend most of the social protections for formal workers even under the conditions of authoritarian regimes and neoliberal public opinion. In contrast, under such conditions, social protections for the scarcely organized and structurally weak lower popular classes were radically cut.

And finally, the distinction of different types of power resources also increases the coherence of the power resources approach in welfare state research and therefore constitutes a qualitative contribution. So far, power resources studies in the field of welfare state research have considered democracy, neoliberal influence, and globalization as factors that affected social policy outcomes in addition to the distribution power resources (Esping-Andersen [1990] 1998; Huber and Stephens 2001; Huber and Stephens 2012). Inspired by the debate on labor's power resources in the discipline of industrial relations<sup>198</sup> and by feminist policy analysis,199 the theoretical framework of this study has advanced a more integrated approach. Instead of considering institutional, structural, and discursive contexts as additional explanatory factors aside from the distribution of power resources, my approach looks at these contexts as directly influencing the distribution of power resources. The empirical analysis supported this integration, as it turned out that the main mechanism through which these contextual factors influenced the dynamics of governing alliance formation and social policy-making was through altering the capacities of different social and political forces to exercise influence. This does not, however, mean that the importance of these contextual factors was entirely restricted to their impact on the distribution of power resources. They also, for example, influence the historically contingent development of the constellations of actors and interests in the Argentinian political arena.

#### **Pluralist, Marxist, and Feminist Perspectives on the Role of Actors in Social Politics**

The literature review in chapter two has worked out different theoretical perspectives on the role of actors in the politics of social protection for low-income earners, distinguishing among pluralist, Marxist, and feminist approaches.

According to the pluralist perspective on Latin American welfare states, most importantly represented by Carmelo Mesa-Lago (Mesa-Lago 1978; 1992; Mesa-Lago, Cruz-Saco, and Zamalloa 1993), the unequal access to social protection has been the outcome of the influence of different pressure groups, such as military men, civil servants, white-collar workers, blue-collar workers, and peasants. The weakness of social protections for the lowest income sectors is thereby considered the result of a lack of power of low-income groups, such as peasants, unemployed and informal workers. The empirical findings of our study support this argument of the pluralist perspective to a certain degree. Associational and structural power resources were concentrated on formal labor during most of the period, which provided it with more capacities to demand and defend social protections than the lower popular classes. This

<sup>198</sup> See e.g., Brinkmann et al. 2008; Dörre 2011; McGuire 2013; Schmalz and Dörre 2013; Silver 2003; and Wright 2000.

<sup>199</sup> See e.g., Hobson and Lindholm 1997.

constituted indeed one important component of the explication. However, there were also other important components, such as the opposition of employers and their influence on exclusionary governing alliances. Moreover, the pluralist perspective cannot explain why social protections for low-income earners were massively expanded during several periods.

Different approaches based on Marxist premises relate the evolution of social protections to the wider and often complex dynamics of class struggle (Block 1977; Esping-Andersen [1990] 1998; Huber and Stephens 2001; 2012; Offe 1984; Piven and Cloward [1971] 1972; Poulantzas [1978] 2000). The empirical findings largely support the core argument of these approaches that the expansion of social protections is an outcome of popular class power and mobilization. In contrast to the pluralist perspective, they assume that capital and its political allies play an important anti-redistributive role in social politics, which is congruent with the empirical finding that employers repeatedly pushed for retrenchment, privatization, and regressive re-regulation. However, different approaches based on Marxist premises have emphasized distinct mechanisms that relate popular class power to redistributive social policies. Piven and Cloward ([1971] 1972) and Poulantzas ([1978] 2000) argued that redistributive social policy is primarily the result of popular class mobilization and unrest, and hence, the outcome of bottom-up processes. In contrast, Offe (1972; 1984) and power resources theorists (Esping-Andersen [1990] 1998; Huber and Stephens 2001; 2012) more strongly emphasized top-down mechanisms. Offe assumed that redistributive social policy is in large part a consequence of the necessity of democratic governments to achieve a certain level of legitimacy in the eyes of the popular masses. Power resources theorists, in turn, emphasized that redistributive social policy is expanded most significantly when popular class and left political forces achieve influence in or control over governments, and hence, are able to introduce policy change from above. In practice, all three mechanisms could be observed, however, in the case of social protections for low-income earners in Argentina, the strongest positive effects were indeed observed when popular class and left forces took part in governing alliances, which coincides largely with the emphasis of the power resources approach.

Moreover, the empirical findings support the arguments of Esping-Andersen ([1990] 1998, 67) and Przeworski (1985, 74–75) that working-class organizations are not necessarily focused solely on attaining benefits for specific occupational groups, as suggested by the pluralist approach of Carmelo Mesa-Lago (1978). The findings show that they can also adopt a broad definition of the working class that emphasizes common interests with other popular classes and promotes inclusionary social protections. In contrast to the pluralist perspective, these accounts can hence explain the formation of inclusionary governing alliances during certain periods that led to the expansion of social protections for low-income earners despite significant disparities in the distribution of power resources between formal labor and the lower popular classes.

The empirical findings moreover support several arguments of the reviewed contributions from feminist social policy research. They support the hypothesis that large parts of the women's movement support the struggle for social protections for low-income earners, as such policies tend to reduce gender inequalities (Arza 2012c; Huber 2006; Sainsbury 1996).<sup>200</sup> The findings also support the argument that the concrete forms and interests of popular class actors cannot be derived neatly from their class position, as internal disputes over the ideological orientation and processes of non-class boundary drawing, that might exclude low-income groups along lines such as gender, race, skill or informality, importantly shape their role in politics (Hite and Viterna 2005; O'Connor 1998; Orloff 2009; Rose 1997). In the Argentine case, informality and unemployment turned out to be important lines along which low-income earners have been excluded from trade union organization. Indirectly, this also led to the exclusion of many female and migrant workers, as these were disproportionally often affected by informality and unemployment. In terms of social policy, this contributed to the fact that social protections for low-income earners had low priority in the political agenda of the biggest Argentinian trade union confederation, CGT. In contrast, the smaller confederation CTA explicitly aspired to organize informal and unemployed workers as well as other oppressed groups and was a strong supporter of an expansion of such policies.

#### **The Role of Social and Economic Structures**

In general, the empirical findings show that economic and social transformations had important impacts on the political dynamics of social policy change. At times, the impact was relatively unambiguous and direct. For example, the eruption of social and economic crises consistently led to increasing political tensions and a weakening of the incumbent governing alliances. However, more often, the political effects were rather indirect and contingent on a range of further factors. Industrialization enabled the working class to become a powerful political actor in Argentina, but the concrete form, political orientation, and strength that the labor movement attained was the outcome of relatively autonomous processes that entailed, among other things, the consequences of strategic choices of the trade unions and their opponents as well as the intervention of the state. In a similar vein, the increase of informality and unemployment in combination with processes of residential segregation after the mid-1970s provided possibilities for the organization of low-income

<sup>200</sup> Furthermore, in the Argentine case, the pro-redistributive orientation of the women's movement is also related to its left-of-center political orientation and its strong support base among the popular classes.

earners based on the territorial dimension. Yet, the how and the when of the emergence of unemployed and informal workers' movements depended on a range of further factors, such as the decision of left parties and trade union factions to undertake deliberate attempts to organize these sectors as well as the particular way the government responded to lower popular class protests by handing out social assistance benefits. In line with the arguments of Collier and Collier ([1991] 2009), the empirical findings show that major transformations of the political arena, such as the political incorporation of the labor movement during the 1940s and of lower popular class movements during the 2000s, were enabled by particular social and economic contexts, but became then increasingly institutionalized and developed path dependency effects that assured their continuity over longer periods of time with considerable autonomy from subsequent social and economic transformations.

Hence, in sum, the empirical findings suggest that for the understanding of the Argentine politics of social protection for low-income earners between 1943 and 2015, social and economic factors were important, yet their impact on political dynamics was always shaped by their interaction with other factors such as the respective historical, institutional, and discursive contexts. Furthermore, the impact of social and economic factors was mostly indirect, meaning that they shaped political processes of alliance formation and social policy reform by enabling the emergence and strengthening of certain actors and interests, which then however interacted with relative autonomy from the social and economic structure.

This indicates that the empirical findings do not support the hypotheses which have been derived from structuralist approaches that suppose a direct determination of social policy outcomes by structural factors. Neither did social policy mechanically respond to social and economic transformations, as assumed by the Logic of Industrialization approach (Wilenksy 1975), nor did they functionally adapt to the needs of capital accumulation as suggested by more structuralist Marxist approaches (e.g., Offe 1972; O'Connor [1973] 2002) and the Varieties of Capitalism approach (Iversen and Soskice 2001; Mares 2001; Soskice and Schneider 2009). The hypothesis derived from PREALC/ILO research, which suggests that the level of social protection for low-income earners depends directly on the dual character of Latin American economies and the related high levels of informality, is not supported by the empirical findings either (Portés and Schauffler 1993, Tokman 1987). Informality did heavily condition the access of workers to social insurance coverage, however, whether government policies attempted to formalize workers or to extend social policy coverage also to informal labor cannot solely be explained by the dual character of the economy. For example, if we observe the evolution of pension coverage in Argentina, we can clearly see that inclusion or exclusion depended in practice much more strongly on legal changes regarding the access criteria undertaken by different governments than on the percentage of the workforce that was informal. Indeed, the pension coverage among the elderly became nearly universal due to the Pension Inclusion Plan during the mid-2000s despite high levels of informality both at that moment and during the preceding decades.

The empirical findings furthermore contradict the hypothesis that globalization as such renders redistributive welfare states unsustainable (see e.g.; Rhodes 2000; Scharpf 2000; Segura-Ubiergo 2007; Strange 2003). In fact, one of the strongest phases of expansion of social protections occurred after 2002 and was paralleled by a strong expansion of foreign trade. Instead, the findings show that more than the quantitative aspect of world trade, it was the qualitative nature of the Argentine insertion into the international division of labor as primary goods exporter that had important impacts on the political dynamics of social policy reform. The findings thereby support several hypotheses that can be derived from dependency (e.g., Senghaas 1974; Cardoso and Faletto [1969] 1977) and extractivism theory (e.g., Burchardt 2016; Gudynas 2012; Peters 2016). As argued by both approaches, the dependence on primary goods exports increased the relative social, political, and economic power of the landed oligarchy and big agricultural companies and weakened the structural power resources of the popular classes by hindering sustainable industrialization and favoring the persistence of dual labor markets. Hence, the way the Argentine economy participated in the international division of labor provided a challenging framework for the sustainability of redistributive governing alliances and contributed to the capacity of anti-redistributive elite circles to repeatedly install repressive inegalitarian regimes. Moreover, the findings empirically confirm the operation of several of the mechanisms described by authors of the (neo-) extractivism approach that link the dependency on primary goods exports to the political dynamics of social policy change. First, due to the dominance of primary export products, the oscillation of international commodity prices led to a significant external vulnerability of the Argentine economy and contributed decisively to the frequent eruption of crises. These crises usually led to a reshuffling of power resources, disruptive changes in public opinion, and realignments of political actors, which in turn provided the ground for the rise and decline of strongly differing governing alliances and radical shifts in the direction of social policy. Second, in line with the concept of neoextractivism, during all three phases of strong expansion of social protections for low-income earners (1943–1955, 1973–1976, 2002–2015), inclusionary governing alliances initially benefitted from a rise in commodity prices, which enabled them to implement legal instruments to capture a part of the primary sector profits and to use these resources for redistributive policies that consolidated or even strengthened their heterogenous popular class support bases. Third, at the same time that the appropriation of primary sector profits enabled redistributive policies, it led to a structural dependence of the inclusionary governing alliances on both national agricultural production and international commodity prices. This meant that any drought or fall in commodity prices constituted a significant challenge to the sustainability of the inclusionary project. Interestingly, the empirical findings show that in the Argentine case, the respective governing alliances did not respond to these challenges by cutting or dismantling redistributive social policies. Indeed, in certain social policy areas they even tended to deepen the redistributive agenda in favor of lowincome earners. However, at the same time, they tackled fiscal and foreign trade balance deficits through measures that induced inflation and a devaluation of the currency and hence decimated the purchasing power of wages, which turned out to be catalyzers for alliance internal tensions. These tensions, in turn, caused a significant political weakening of the governments' bases of support and hence contributed to their subsequent loss of power. By this means, as argued by theorists of (neo-) extractivism, the dependency of inclusionary governments on primary goods exports not only challenged the fiscal but also the political sustainability of social protections for low-income earners.

#### **The Role of Institutions**

The empirical findings show that institutions had important influence on the dynamics of the politics of social protection for low-income earners in Argentina. From a global perspective, the findings support Bob Jessop's (1990; 1999) state theoretical argument that institutions are strategically selective. The historical analysis reveals that this selectivity affected social policy change mainly through three mechanisms. First, institutions were selective in the sense that they provided some forces with more capacities to influence the policy process than others and hence directly affected the distribution of power resources among contending actors. Second, institutions influenced the historical formation and evolution of the subjective political interests of different actors in social policy. Third, they provided a framework which limited the array of potentially successful actions and alliances for the pursuit of their respective interests and hence affected the development of strategies of different actors.

Besides supporting this rather abstract state theoretical argument, the empirical findings have a range of implications for more fine-grained theoretical arguments regarding the role of specific institutional structures for social policy change. In agreement with Segura-Ubiergo (2007), Haggard and Kaufman (2008), and Huber and Stephens (2012), among others, the empirical findings show that regime type is a decisive institutional factor. Democratic regimes provided a much more favorable institutional framework for the extension of social protections for low-income earners than authoritarian and semi-authoritarian regimes. In practice, three main underlying mechanisms could be observed. First, democracy provided the popular classes with significant electoral weight and with civil liberties that facilitated processes of organization, which taken together had a positive effect on the level of power resources of the popular classes and therefore increased their ability to attain governmental power and to influence the social policy agenda. Second, the electoral weight of the popular classes provided significant incentives for politicians and governments to attempt to mobilize low-income earner support by proposing or implementing an extension of social protections. Third, in line with the arguments of Przeworski (1985) and Esping-Andersen ([1990] 1998), democracy motivated organized labor to forge political alliances with unorganized low-income groups and social movements in order to reach electoral majorities. At the same time, the findings also show that the operation of these mechanisms depended not only on the prevalent regime type but on further factors. For example, the dominance of neoliberal thinking during the early 1990s undermined these mechanisms to an important degree, as the dominant world view enabled the political right to generate mass support for regressive social policy reform even among low-income earners.

The empirical findings moreover support the hypothesis derived from Immergut (1990; 2010) that the number and strength of institutional veto points are inversely related to the velocity of social policy change. In the Argentine case, institutional veto points were relatively weak during most of the period of study, which contributed to a strong concentration of power in the hands of the leading circles of the respective governing alliances and enabled them to undertake radical shifts in the direction of social policy and to implement profound change within relatively short timeframes. In line with Levitsky (2003), the findings furthermore showed that a meaningful assessment of veto points cannot rely solely on formally existing institutional structures but must also take into consideration informal structures and political practices that often have an important role in shaping policy processes.

The tracing of successful and failed social policy reform processes also provided clear support for the hypotheses that once implemented, inegalitarian social policies tend to deepen social and political divides within the popular classes, as argued by Abel and Lewis (1993), and to create vested interests that make future reform politically more costly, as argued by Huber and Stephens (2001) and Pierson (2001).

The Argentine case does not support the hypothesis derived from Rosenberg (1979; 1983) and Martínez-Franzoni and Sánchez-Anconchea (2012) that bureaucracies of state apparatuses, and in particular of social insurance institutions, tend to attain significant influence on the direction of social policy change. Due to the highly vertical organization of governments in Argentina and the ability of presidents to exchange bureaucratic personnel, bureaucrats had little ability to develop significant long-term influence on the direction of social policy. Interviews with a former minister and several state secretaries showed that even when bureaucrats were able to develop social policy proposals, these were commonly ignored as soon as they diverged from the general line of economic and social policy pursued by the leading circle of the governing alliance.

## **The Role of Ideas**

Two ideational approaches have been distinguished in the literature review. Authors of the first approach stress the importance of policy learning and the proliferation of specific ideas about the regulation of social programs through national, regional, and global social policy expert communities (e.g., Danani and Beccaria 2012; A. Deacon 2000; Martínez Franzoni, Sánchez-Ancochea, and Solano 2012; Weyland 2005). The second approach emphasizes the importance of overarching economic paradigms and worldviews, such as Keynesianism or neoliberalism, that provide discursive frameworks through which both the wider public and political decision-makers assess different social policy options (e.g., Blyth 2002; Hall 1993; Hay 2001; Wincott 2011).

The empirical findings provide support for the arguments of both approaches—to different degrees, however. The diffusion of specific social policy recommendations and ideas through international organizations and academic circles often influenced the way specific social policies were introduced, extended, cut, or reformulated. However, the empirical findings suggest that the influence of such ideas was strongly limited as soon as these contradicted the dominant economic discourse. For example, universalistic social policy proposals elaborated by Argentinian academics, the WHO, and several think tanks hardly had any impact on social policy outcomes during the 1990s, when neoliberal thinking guided policy-making in nearly all areas. In contrast, shifts in the prevalence of overarching economic paradigms influenced the general direction of social policy towards low-income earners much more profoundly. This was mainly due to the fact that broad economic paradigms developed their impact on the policy process through a variety of direct and indirect mechanisms. First, as did specific social policy recommendations and ideas, they influenced the cognitive and analytical perspectives of social policy-makers. Second, they also shaped the general political orientation of governments, which usually provided clear margins for the development of policies within the different ministries and government departments. Third, they had a significant influence on the perceptions and opinions of the wider public, with the result that they shaped the societal distribution of discursive power resources and facilitated or hampered the formation of inclusionary governing alliances.

The empirical findings however also showed that ideational factors do not lead to certain policy outcomes in a direct or mechanic way. Instead, they always interact with other contextual factors and their concrete impact is furthermore dependent on historically specific political processes. For example, the prevalence of the ISI paradigm as an overarching economic logic and the specific policy ideas proliferated by the ILO, the WHO, and academic communities during the 1960s and 1970s provided a benign ideational context for the extension of social protections to low-income earners. However, the fact that long periods of these decades were characterized by exclusionary rightwing dictatorships in Argentina successfully blocked the implementation of such policies.

## **6.3 Outlook**

With the victory of the center-right neoliberal Cambiemos alliance in the 2015 presidential elections, a cycle of 13 years of progressive reform ended in Argentina. Rather than being an isolated phenomenon, this development coincided with a broader swing of the political pendulum to the right and the decline of several inclusionary governing alliances in the Latin American region.

On the one hand, economic difficulties decisively contributed to the resurgence of the right. Against the background of a failure to diversify production and export structures, the decline of commodity prices heavily affected growth rates in most Latin American countries, generated fiscal pressures and, most importantly, intensified distributive conflicts within and without the governing alliances. As in Argentina, most progressive governments did not respond to sluggish economic growth and declining fiscal revenues by cutting previously implemented redistributive social policies. In fact, in many cases, they even deepened or extended these programs in order to compensate, with considerable success, the social consequences of the economic downturn. However, in a context where it was no longer possible to finance progressive policies by distributing the gains from economic growth and commodity export, inclusionary governing alliances had increasingly no other choice than to implement measures that seriously affected the interests of one or another societal group. In practice, the array of such policies was quite heterogeneous and the emphases varied across different countries, ranging from tax and tariff increases to the shifting of funds between different policy areas to indirect financing through inflationary policies. Despite considerable differences in the kinds of policies implemented, all of them produced winners and losers and hence increased tensions among different groups and interests within the governing alliances as well as intensifying oppositional right-wing, upper- and middle-class resistance. Taken together, economic difficulties and subsequent political tensions tended to reduce the popularity of the progressive governments, to weaken some of their organizations, and to generate doubts about the appropriateness of the demand-oriented and state-centered post-neoliberal paradigm to deal with the current economic challenges. Eventually, these developments provided opportunities for the re-emergence of center-right and right governing alliances.

On the other hand, a more fine-grained look at the political developments in the region also indicates that the rise of the political right did not constitute an automatic or inevitable outcome of the end of the commodity export boom. As the victory of the left MAS-Party in the 2020 presidential elections in Bolivia shows, progressive governments cannot only countervail the negative social impact of the economic downturn but also steer successfully through the subsequently arising political tensions. As argued by Anria and Roberts (2019), in the Bolivian case, internal alliance mechanisms for participation and consensus generation have been of key importance for preserving the cohesion and strength of the governing alliance even in times when tensions between different factions and interests tended to increase. In a similar way, the analysis of the Argentine case suggests that the decline of the Kirchnerist governing alliance was as much caused by political errors, corruption scandals, verticalism, and a lack of mechanisms for the generation of effective internal consensus as by the economic downturn. Additionally, the current economic difficulties not only challenge progressive governments but also their conservative and neoliberal counterparts. In México, the 2018 elections resulted in a victory of the center-left alliance of Andrés Manuel López Obrador. In Argentina, the neoliberal policies of the right Cambiemos alliances have pushed the economy into an even deeper recession which contributed to the land-slide victory of the more inclusionary alliance of Alberto Fernández in the 2019 presidential and parliamentary elections.

The four years of Macri's presidency furthermore showed that there were significant restrictions on cutting social protections for low-income earners. In contrast to past experiences of retrenchment, the Cambiemos alliance was neither able to build on the repressive means that were available to the authoritarian right-wing regimes between 1955 and 1983 nor on the deep-rooted neoliberal consensus that enabled the Menem government to mobilize broad-based popular support for regressive social policy reform during the 1990s. Together with the fact that the Argentine working class and lower popular classes had conserved significant associational and structural power resources and capacities of resistance, this context led the Cambiemos government to embark on a rather slow and gradual process of retrenchment, which has left many of the popular social policies of the previous government intact. As argued by Levitsky and Roberts (2011, 423–424), the Chilean experience of the right Piñera government points in a similar direction.

What are, then, the future political prospects for sustaining or even extending measures of social protections for low-income earners in Latin America? While conservative and neoliberal governments are certainly increasing pressure on such policies, path dependency effects as well as the still relatively strong power resources of the popular classes, and hence the capacities of left and popular class forces to resist retrenchment or even to regain governmental power, constitute considerable counter-pressures. A first superficial analysis suggests that the recent formation of right-wing governing alliances in several countries of the region was less due to a significant reduction of popular class power resources than to divisions among progressive and popular class actors, strategic mistakes, and corruption scandals, on the one hand, and successful political strategies as well as an increased unity of the political right, on the other. This raises important questions that should be addressed by future studies: What were the strategies of the political right to achieve governmental power and what made them successful in some countries but not in others? What structures, constellations of actors, and strategies enabled inclusionary governing alliances in some countries to sustain unity even in the face of difficult economic environments and increasingly fierce oppositional pressures? What strategies were most effective in impeding a reduction of social protections, or even to achieve a further expansion, under conservative governments? The modified power resources approach can provide a fruitful theoretical framework to explore these questions.

However, not only the future but also the past of social protection for lowincome earners and state redistribution in Latin America is still less clear than often suggested. Influential studies on Latin American welfare states by Barrientos (2004); De Ferranti et al. (2004); Lindert, Skoufias, and Shapiro (2006); Mesa-Lago (1978); and others argued that social protection started to expand relatively early on the continent but since then excluded or disadvantaged lowincome groups and therefore attained a regressive truncated character. This picture of rather constantly and consistently exclusionary social policy regimes appears surprising for a region whose newer history is characterized by severe social conflicts, the emergence of powerful popular class movements, guerrilla wars, uprisings, and even revolutions. In fact, the present study shows of the Argentinian case that social policy was anything but constantly exclusionary. In practice, phases of strong expansion of social protections repeatedly interchanged with phases of radical retrenchment. The ability of the Argentinian elite to push the social policy regime in a truncated direction thereby depended less on their capacity to pre-empt or avoid the implementation of redistributive policies than on their capacity to reverse these progressive changes at a later point in time, which not seldom occurred with considerable support from repressive military forces, western governments, international financial organizations, and transnational corporations. As social policy change is strongly influenced by path dependencies and historically grown constellations of actors and interests, understanding the past is crucial for an appropriate understanding of present and future processes of political change. Therefore, the application of the modified power resources approach to the history of other Latin American countries could contribute significantly to our understanding of redistributive processes on the continent and enhance our capacities to make sense of future change.

Finally, one of the crucial strengths of the modified power resources approach is the fact that it has been specifically adapted to fit the analysis of redistributive social policy change in Latin America. The other side of this coin is, of course, that the approach in its current form has limitations and cannot be neatly applied to explain political outcomes in other policy areas or in other world regions. However, these limitations are not due to a general inappropriateness of the approach for such tasks but related to the particular operationalization developed in the framework of this study. Associational, structural, institutional, and discursive power resources as well as historically contingent constellations of actors and processes of (governing) alliance formation are all likely to constitute pivotal categories for understanding political change in other policy areas and geographical regions. However, Import Substitution Industrialization and Neoliberalism, for example, might not necessarily constitute the decisive ideational structures which account for the distribution of discursive power resources among contending actors when it comes to explaining policy outcomes regarding for example gendered violence. Furthermore, the array of contending actors is likely to differ significantly as well as the kind of institutional and structural power resources that shape their capacities to make their voices heard in the political process. Hence, the application of the approach beyond Latin America and beyond redistributive social policy change will require some adaptation. Yet, if such adaptation is carefully undertaken, its application might yield significant new insights into political processes across different regions and policy areas.

## References


Bertranou, Juan M. Palacio and Gerardo M. Serrano, 69–82. Buenos Aires: Prometeo Libros.


## **Laws and Decrees**

Decree 165/02. 2002. "Emergencia Ocupacional Nacional. Programas Jefes de Hogar - Creación." Boletín Oficial, January 23, 2002. Buenos Aires.


## **Interviews**


<sup>201</sup> The political and academic positions of the interviewees are those held as of early 2014 when the interviews were conducted.


<sup>202</sup> Name changed because interviewee preferred anonymity.

expert commission that designed the universal health reform project of the Peronist government in 1973. Interview conducted by the author in Buenos Aires. February 26.

Neri, Aldo. 2014b. Interview conducted by the author in Buenos Aires. March 10.


## Index

ABRA (Asociación de Bancos de la República Argentina) 266 Actor-centered approaches 32– 36, 56 ADEBA (Asociación de Bancos Privados de Capital Argentino) 161, 200, 266 Administración Nacional de Seguro de la Salud *see ANSSAL* Alderete, Carlos 178 Alfonsín, Raúl 149, 159, 169ff., 174, 176–179, 181, 193f., 196, 202, 208, 260, 269, 282f., 290 Alianza Popular Revolucionaria Americana *see APRA* Allende, Salvador 22, 306, 308 Alsogaray, Álvaro 192, 194 Anarchism 101f., 104 Anarcho-syndicalism 104, 109, 117, 258f. Angeloz, Eduardo 186, 192 ANSSAL (Administración Nacional de Seguro de la Salud) 194f. Anti-communism 111f. Anti-Peronism 23, 118, 297 APRA (Alianza Popular Revolucionaria Americana) 28 Aramburu, Pedro 134, 150 Argentine Anti-Communist Alliance (Triple-A) 159, 289 Arms production 106, 110f. 269 Asociación de Bancos de la República Argentina, *see ABRA*  Asociación de Bancos Privados de Capital Argentino, *see ADEBA*  AUH (Asignación Universal por Hijo) 242 Australia 19 Authoritarianism 21, 24, 29, 43f. 61, 68f., 77, 87, 93, 95, 110f., 116, 119, 127, 129–133, 138, 145, 165, 167, 258, 262, 267, 276, 283f., 312, 318, 322, *see also Military dictatorship* Auto workers 142, 194 Balance of payments *see Trade balance* Balbín, Ricardo 135, 149, 171 Barrionuevo, Luis 193f., 237f. Barrios de Pie 237 Bignone, Reynaldo 174 Blank ballots 135 Bolivia 28, 65, 322 Bolsa de Comercio de Buenos Aires (BOLSA) 266 Brazil 13, 18, 65f., 208, 306, 308 Buenos Aires, City of 211 Buenos Aires, Province of 99, 202, 230, 247 Bulgheroni, Carlos 193 Bunge y Born 193 Business associations *see Employers' associations* Business leaders 266, 286, 290f. Business representatives *see Business leaders* CAC (Cámara Argentina de Comercio) 266

CACON (Cámara Argentina de Construcción) 266 Cafiero, Antonio 193 Cambiemos 247, 321f. Cámpora, Héctor 151–154, 156, 159, 259 Cámara Argentina de Comercio, *see CAC* Cámara Argentina de Construcción, *see CACON* Canada 19 Capital flight 142, 150, 169, 224, 244 Capitalism 34, 38, 40, 50, 52, 62, 73ff., 91 Catholic Church 54, 71f., 74, 116, 119, 121, 127, 130, 132f., 149, 159, 162, 190, 194, 210, 212, 225, 257, 268ff., 283, 286f., 292, 296, 301 Cavalieri, Armando 193 CCC (Corriente Clasista y Combativa) 190, 211ff., 222, 225, 231 CEA (Consejo Empresario Argentino) 188, 193, 204ff., 266 Central Bank 161, 170, 244 Central de Trabajadores Argentinos, *see CTA* CGE (Confederación General Económica) 118, 121, 146, 149f., 154f., 160ff., 168, 172 CGI (Confederación General de la Industria) 267 CGT (Confederación General del Trabajo) 105, 109, 111, 113, 119, 121f., 134, 138f., 141, 143, 148, 150f., 153ff., 161, 164f., 176, 178, 186, 188f., 194f., 198, 201, 205, 213,

221f., 224, 230ff., 237f., 244, 262, 315 CGT Azul y Blanca 237 CGT de los Argentinos 141ff., 148, 261 CGT Disidente 231 CGT-1 111 CGT-2 111 CGT-Azopardo 141, 194f. CGT-Oficilista 244 CGT-Opositora 244 Chamber of Deputies 116, 176, 186, 194f., 198, 237, 239 Child allowances *see Family allowances* Chile 22, 66, 271, 306, 308, 322 Christian Democrats 57, 59, 259, 308 Church *see Catholic Church* Civil society 26 Class struggle 34, 39, 314 Clientelism 67, 125, 128, 181, 187, 191f., 195f., 202f., 207, 211, 230, 235, 241ff., 260, 279, 284, 290ff., 306 Collective action 63, 76f., 188, 190f. Collective agreement 134, 210, 221, 232 Collective bargaining 114, 134, 161, 210, 221, 230f. Colombia 66 Commodity prices 216, 243, 246, 317f., 321 Communism 51, 111f., 114, 117, 258f., 267f. Communist Party 69, 112, 116, 133, 155, 308 Comparative method 85–97 Confederación General Económico, *see CGE*

Confederación General del Trabajo *see CGT* Confederaciones Rurales Argentinas, *see CRA* Congreso de Trabajadores Argentinos, *see CTA* Congress *see Chamber of Deputies; Parliament; Senate* Consejo Empresario Argentino, *see CEA* Consigli, Carlos 143 Constitution 66ff., 105, 136, 171, 196, 203, 259, 292 Construction workers' union 114, 245 Convertibility 193, 206, 209, 216f. Corriente Clasista y Combativa *see CCC* Corruption 105, 208, 218, 244, 247, 322f. Costa Rica 17ff., 22, 46, 306ff. Coup 86, 104, 110f., 121, 126, 129, 131, 133f., 136ff., 141, 145, 157, 159, 161, 165, 178, 267–270, 276, 280, 301 Coup d'etat *see Coup* CRA (Confederaciones Rurales Argentinas) 226, 266 CTA (Central de Trabajadores Argentinos) 189ff., 196, 198, 204, 210, 212, 222–225, 229, 231f., 234, 236, 238, 248, 257, 261, 263, 315 CTA-A (CTA-Autónoma) 238 CTA-T (CTA de los trabajadores) 238 Currency convertibility *see Convertibility*

devaluation 136, 138, 216, 221, 225, 246, 318 overvaluation 173, 184f. De Gennaro, Victor 238 De la Rúa, Fernando 96, 208f., 211ff., 215f., 224, 226, 248, 255 Debt, foreign *see Foreign debt* De-industrialization 158, 168, 185, 188, 282, 312 Democracy, definition of 92 Dependency theory 42, 317 Devaluation *see Currency* Dock workers 137, 142, 145 Duarte, Eva, *see Perón, Eva* Duhalde, Eduardo 202, 208, 216, 225f., 228ff., 248, 256 Economic growth 18, 35, 37, 40, 80f., 100, 104f., 117, 138, 144, 146f., 155f., 169, 171, 184, 203, 208, 237f., 243, 246, 284, 307, 321 Economically active population 63, 100, 103, 121f., 158, 170, *see also Labor force* Election 13, 18, 86, 92f., 104, 111, 116, 118f., 126, 131, 135ff., 141, 150f., 153, 156, 158, 169, 171, 174, 176ff., 181, 183, 186f., 191–197, 203f., 208, 216, 218, 220, 224, 229f., 232, 237f., 241ff., 245–248, 259, 268, 281f., 291, 301, 321f. Employers' associations 48, 56, 59, 66, 70, 73, 108ff., 112, 115–118, 138, 161f., 168, 173, 176, 200, 204f., 220, 225f., 238f., 265, 268 Eva Perón Foundation *see FEP (Fundación Eva Perón)*

Factory occupations 134, 136, 138, 226 Falklands War 269 Family allowances 17, 20, 98, 128, 134, 137, 142, 145, 156f., 168, 182f., 202, 212, 215, 249f., 255f., 306f. FEP (Fundación Eva Perón) 121, 124–127, 132f., 137, 145, 254f., 270 FORA (Federación Obrera Regional Argentina) 102 Foreign debt 158, 168ff., 192, 213, 216f., 221, 224f., 271f. Foreign trade 41, 47, 108, 161, 184, 317f. Formal employment 63f., 234 Formal labor 65, 163, 212, 232, 238, 282, 297, 313, 315 Formal workers 20f., 28, 59, 63ff., 68, 70, 126, 134, 137, 144f., 180, 212f., 253, 255, 258, 261, 268, 281ff., 292, 297f., 305f., 312 FpV (Frente para la Victoria) 220 FREJULI (Frente Justicialista de Liberación) 150f. Frente Grande 187, 189, 195, 218, 229, 233 Frente Transversal 222, 231, 241, 243 FREPASO (Frente del Pais Solidario) 187, 189, 196f., 204, 208ff., 212, 218, 254f., 289, 291 Frondizi, Arturo 135ff. 260, 291 FTV (Federación de Tierra, Vivienda y Hábitat) 189f., 211ff., 222, 225, 229, 231, 237, 241

Fundación Mediterránea 172, 187 Galtieri, Leopoldo 173 General secretary 113, 221, 232, 238, 244 General strike 102, 165, 173, 176f., 188, 192, 194, 201, 204f., 209f., 224 Gini coefficient 18f., 139, 162, 167, 182, 214, 243, 250 GOU 111f., 268 Governing alliances definition of governing alliances 55 exclusionary governing alliance 109, 129, 139, 145, 166, 214, 278, 283, 298, 314 inclusionary governing alliance 106, 111, 146, 150, 169, 231f., 248, 278, 282, 286, 296, 301, 307, 310, 312, 314, 317, 320f., 323 progressive governing alliance 175, 224, 230, 232, 274, 279, 286f., 289f., 297ff. regressive governing alliance 213, 270, 278f., 281, 283f., 287, 291, 296f. Group of 25 143, 193 Guerrilla 142, 153, 323 Guido, José María 137 Health care Health care reform 143f., 151, 153, 166, 176–179, 181, 201, 235, 250, 255 Public health system 22, 124, 127, 136, 145f., 157, 163, 165, 167f., 215, 234, 246, 255f., 282, 299, 306, *see also Social health insurance* Hora del Pueblo 150

Housing 18, 22, 46, 98, 106, 115, 125, 127f., 141, 148, 150, 154– 157, 166ff., 182, 196, 198, 215, 231, 235f., 238, 240, 244f., 249, 255f., 264, 283, 289, 306 Human rights violations 171, 175, 265, 269 Hyperinflation 170, 179, 185, 192, 195, 202 IADB (Inter-American Development Bank) 71, 178, 185, 202, 205, 270 Ideational approaches 32, 47–49, 56, 78f., 320 Illia, Arturo 137f., 260 ILO (International Labor Organization) 20, 38, 71, 95, 108, 147, 156, 158, 185, 218, 270f., 296, 316, 320 IMF (International Monetary Fund) 71, 158, 178, 185, 205, 210, 270f. Immigrants 100f., 104, 118, 163, 168, 189, 204, 233, 236f., 261, 315 Import Substitution Industrialization *see ISI* Income inequality 19, 167, 182, 185, 209, 214, 250, *see also Gini coefficient* Indigenous communities and movements 52, 54f., 65, 70, 74, 96, 118, 189f., 263f. Industrialization 24f., 29, 33, 37f., 42, 62, 76, 84, 91, 101, 104, 108, 110ff., 126, 129, 132, 147, 149, 160, 168, 172, 191, 223f., 267f., 280, 307, 312, 317 Inflation 107, 130, 135f., 139, 142, 147, 150, 154, 157ff., 161, 169f., 178f., 183–186, 198,

203, 217, 228, 232, 238, 244, 246f., 250, 318 Informality Informal employment 29, 63f., 137, 257, 284 Informal labor 22, 230, 316 Informal sector 38, 87, 268 Informal workers 20, 22, 24, 28, 51, 59ff., 63ff., 67, 69f., 74, 127, 137, 145, 155, 168, 183, 190, 215, 232, 246, 253, 255, 258, 263f., 295, 313, 316 Informality 38, 55, 60, 64, 91, 137, 139, 147, 154f., 159f., 170, 172, 179f., 182, 184, 188, 191, 198, 204, 206, 213, 217, 223, 263, 274, 312, 315ff. INOS (Instituto Nacional de Obras Sociales) 144, 164, 194 Institutionalist approaches 32, 42–47, 52 Inter-American Development Bank, *see IADB* Interim president 216, 226, 230, 248 International Labour Organization, *see ILO* International Monetary Fund, *see IMF*  Junta Coordinadora Nacional 149, 174f., 264 Justicialist Party, *see PJ (Partido Justicialista)*  Kirchner, Cristina 239, 242, 244, 246f., 256, 289 Kirchner, Néstor 212, 216, 220, 229–232, 234f., 237, 239, 243, 247f., 256, 289

Kirchnerism 221f., 230f., 251, 264, 295, 306, 322 Kolina 243 La Cámpora 243f. Labor Department 113f. Labor force 33, 42, 65, 76, 100f., 129 Labor laws 63, 161, 204, 268 Labor market regulations *see Labor laws* Labor movement 21f., 24, 35f., 44, 51, 60, 62, 64ff., 68, 100f., 104f., 109–112, 116, 118, 126f., 129, 133–138, 143, 145, 157, 164f., 168, 173, 178, 181, 188, 221, 245, 268, 275, 286, 308, 315f., *see also Trade union movement* Landowners 33, 70, 102, 137, 161 Lanusse, Alejandro 150 Lastiri, Raúl 153 Legislature *see Parliament* Levingston, Roberto 150 Light and power workers' union 178 Liotta, Domingo 152f. Lobbying 35, 61, 74, 153, 173, 238, 266f. Lonardi, Eduardo 133f. López Rega, José 153 Macri, Mauricio 247f., 322 Madres de la Plaza de Mayo 173, 265 Malvinas War, *see Falklands War* Martínez de Hoz, José 161, 164 Marxist approaches 34, 39f., 50, 54, 62, 70, 294, 314, 316 Marxist groups 148, 159, 264, *see also Communist Party*

Media 34, 48, 74, 115, 130, 173, 181, 229, 238f., 241, 247f. Menem, Carlos 183, 186f., 192– 198, 200–211, 213f., 220, 229, 238, 260, 279, 285, 322 Mercante, Domingo 113 Metal workers' union (UOM) 188, 194, 221, 244 Mexico 23, 28, 65f. Middle class 22, 52, 55, 100, 103ff., 115, 119, 127, 133, 151, 174, 181, 187, 204, 220, 224, 229, 237, 248, 260, 281, 289f., 292, 297, 321 Migrants *see Immigrants* Miguel, Lorenzo 194 Military coup *see Coup* Military dictatorship 29, 55, 111ff., 115f., 126, 131, 133ff., 138, 140–144, 146, 150, 157f., 160f., 163–166, 168f., 171, 173f., 179f., 182, 193, 214, 232, 241, 254f., 262–269, 274, 276, 280, 282, 284, 292 Military government *see Military dictatorship* Military junta *see Military dictatorship* Military rule *see Military dictatorship* Minister of social welfare 143, 153, 156 Ministry and Minister of Economy 95, 155, 161, 164, 193, 198, 201, 205, 208 Ministry and Minister of Health 165, 177f., 200 Ministry and Minister of Labor 178, 187, 194, 198f., 207, 209, 229

Ministry of Social Development 209 Ministry of War 124 Mothers of the Plaza de Mayo *see Madres de la Plaza de Mayo* Movimiento de los Trabajadores Argentinos *see MTA* Movimiento Evita 222, 231, 237, 240f., 243 Movimiento Nacionalista Revolucionario (MNR) 28 Movimiento Villero Peronista (MVP) 149, 159, 263 Moyano, Hugo 221, 232, 244 MTA (Movimiento de Trabajadores Argentinos) 188, 196, 201, 204, 210, 221, 231 Nationalist 110f., 142, 173, 267ff. Neo-Peronist 136f. Neri, Aldo 14, 153, 165, 176ff. Obras Sociales 143f., 146, *see also Social health insurance* Oligarchy 25, 99, 102, 104f., 111, 115, 286, 317, *see also Landowners* Onganía, Juan Carlos 138, 148, 150, 180, 262 PAMI, *see Pensioners' health insurance* Parliament 26, 46, 52, 73, 91ff., 96, 105, 116, 121, 135, 138, 151ff., 177f., 195f., 210, 224, 239f., 259, 276, 284, 288, 291, *see also Chamber of Deputies; Senate* Partido Colorado 308 Partido Communista, *see Communist Party*  Partido Demócrata Cristiano 308

Partido Demócrata Progresista (PDP) 116 Partido dos Trabalhadores 308 Partido Independiente 116, 119 Partido Intransigente 237 Partido Justicialista (PJ) 23, 28, 51, 69, 118f., 121, 134f., 145f., 148–152, 160, 167, 171f., 174, 176, 181, 183, 186ff., 191ff., 195, 197f., 203, 207–213, 218, 220f., 223, 229ff., 233, 237, 246, 248, 254, 258ff., 262f., 275f., 279, 282f., 288–291, 296, 311 Partido Laborista 116–120, 126, 259, 275 Partido Liberación Nacional (PLN) 22, 308 Partido Nacional 308 Partido Peronista *see Partido Justicialista* Partido Revolucionario Comunista (PCR) 190 Partido Revolucionario Institucional (PRI) 23, 28 Partido Socialista de Chile 308 Patronage *see Clientelism* Peasants 20, 22ff., 28, 33, 44, 52, 55, 61, 63, 65, 69f., 72, 74, 117, 189f., 263f., 295, 306, 308, 313 Pensioners' associations 198, 239 Pensioners' health insurance 179, 228, 234 Pensions Average pension 165, 180, 197f., 232 Minimum pension 154f., 157, 165f., 180, 197f., 212, 215, 228, 232, 236, 244, 249, 255f.

Non-contributory pensions 125, 154, 180, 183, 202, 206, 215, 236, 241, 245 Pension coverage 22, 103, 114, 119, 121f., 151, 155, 199, 203, 228, 232–235, 249, 254f., 307, 316f. Pension Inclusion Plan 236, 246, 256, 317 Pension reform 48, 141, 155, 201, 206 Replacement rate 123, 128, 135f., 139f., 145, 180, 183, 197, 199, 215, 234, 249, 254f. Social assistance pensions 202, 235, 241, 249 Perón, Eva 125, 132 Perón, Juan 66, 111–121, 123f., 126ff., 130, 132, 135, 141, 145, 148, 150, 152f., 156, 158, 160, 254, 259f., 262, 266, 268ff., 280f., 286–289 Peronist Party *see Partido Justicialista (PJ)* Peronist Youth 148, 150f., 153, 159f., 259, 264, 288f. Peru 28 PJ, *see Partido Justicialista (PJ)* Plan Alimentario Nacional (PAN) 180f., 202, 282 Plan Austral 169, 178, 186, 192 Plan Jefes y Jefas de Hogar (PJJHD) 225f., 229, 236, 242, 249, 269 Plan Mayores 228 Plan Nacer 234, 240, 245, 249 Plan Nacional de Seguridad Alimentaria 226 Plan Remediar 227 Plan Sumar 249

Plan Trabajar 206, 226 Plan Vida 202 Plastic workers' union 193f. Pluralist approaches 33f., 37, 50, 313f. Polarization 204, 288, 305 Poverty 82, 99, 106, 125, 180, 202, 204, 206ff., 211f., 214, 216, 224, 235, 246 Power resources approach 23, 28, 34ff., 40, 42, 49–54, 57ff., 62, 69–73, 75f., 78, 83, 258f., 261, 266, 277, 294, 297, 310ff., 314, 323f. Power, definition of 72 Press, *see Media* Price controls 138, 154, 230 Privatization 17, 23, 71, 81, 136, 164f., 178, 184, 188, 193, 196– 200, 203f., 206, 209, 214f., 230, 248, 256, 271, 284, 314 Process tracing 85–97 Productivity 38f. Proscription 77, 129, 135, 145, 150 Public employment 101, 129 Public enterprises 188, 193, 196, 209 Railway workers' unions 111, 113, 124 Rodríguez Sáa, Adolfo 224f., 229 Rucci, José 153 Secretary of health 200 Secretary of labor 111–116, 126 Semi-authoritarian *see Authoritarianism* Senate 77, 116, 137, 176ff., 237, 239, 290 Social assistance 18, 22, 48, 71, 86, 98, 121, 125, 127f., 133f., 136f., 145f., 154, 156f., 166ff.,

175, 180, 182f., 187, 190, 196, 202f., 205ff., 211f., 214f., 222, 226f., 230ff., 235f., 240f., 245– 249, 253ff., 260, 268, 270f., 278f., 281–285, 289–293, 295f., 305f., 316 Social compact 154f., 158 Social Democracy 51, 69, 149, 171, 176, 208, 259f., 282, 290, 311 Social health insurance 114, 127, 141, 143ff., 157, 164f., 167f., 177, 179, 187f., 194ff., 200f., 205f., 210f., 214f., 235, 249, 254, 256, 284, 298f. Social housing *see Housing* Social movements 26, 28, 44, 54f., 58–62, 64f., 72, 74, 91, 93, 172, 181, 220, 222, 224f., 230f., 241f., 247, 251, 257, 263, 265, 270, 281, 291, 296, 307, 310, 319 Social security 17, 33, 37ff., 41, 46, 48, 57, 63, 69, 71, 76, 80f., 96, 108, 110f., 117, 148, 155, 188, 211, 271, 306 Social Security Department 178, 198 Socialist Party 105, 109, 111f. Sociedad Rural Argentina (SRA) *see SRA* Soup kitchens 202, 207, 263 SRA (Sociedad Rural Argentina) 138, 226, 266 State enterprises *see Public enterprises* State of emergency 224 State workers' union 189 Strikes 75f., 102f., 112, 126, 134, 136, 138ff., 142, 161f., 165, 171, 190, 194, 205, 210, 221,

227, 273, 278, 307, *see also General strike* Structuralist approaches 37–42, 56, 316 Students 116, 142, 210, 249 Subsecretary of health 152f. Subsidies 46, 113f., 221f., 235, 240f., 243, 249, 262, 280 Suffrage 92, 110, 118, 121, 132, 270 Sugar cane workers' union 263 Supreme court 138, 196 Tamborini, José 116 Tariffs 83 Teachers' unions 189, 232 Teamsters' union 232, 239 Techint 240 Technocrats 48, 192, 208, 213 Trade balance 83, 106, 139, 142, 217 Trade union movement 23f., 28, 35, 51, 59, 88, 111f., 114f., 118, 121, 126, 128f., 131, 135f., 139ff., 146, 148, 151, 153, 156, 159, 161–165, 167, 171, 175f., 181, 183, 188, 194, 201, 213f., 221ff., 229, 231ff., 237, 244, 250, 258, 261f., 268, 275f., 278ff., 282ff., 286, 288, 290f., 296ff., 308, 311 Triaca, Jorge 193f. Trotskyists 117 UCeDé (Unión del Centro Democrático) 186f., 192, 194, 197, 200, 204 UCR (Unión Cívica Radical) 52, 69, 103f., 116f., 119, 133, 135, 137, 148f., 152, 159, 168f., 171f., 174ff., 181, 183, 186f., 191f., 196f., 203f., 208, 213, 218, 231, 237, 254f., 258ff.,

262, 264, 279, 282, 286, 289– 292, 296, 311 UCRI (Unión Cívica Radical Intransigente) 135, 260 UCRP (Unión Cívica Radical del Pueblo) 135, 137, 260 UIA (Unión Industrial Argentina) 109, 134, 176, 205, 266 Unemployed workers' movements 54, 96, 190f., 204, 206, 209–215, 222–229, 231f., 234–238, 240f., 251, 289, 291, 305 Unemployment 20, 48, 64, 75f., 84, 91, 98, 101, 104, 106f., 109f., 126, 128ff., 132, 142, 145ff., 149, 154, 156–160, 168, 170, 172, 179f., 183ff., 188, 191, 198f., 202, 204–208, 211– 217, 221, 223, 225ff., 235, 242, 248f., 263f., 273f., 284, 290, 312, 315 Unemployment allowance 20, 180, 183, 202, 212, 215, 225, *see also Unemployment insurance* Unemployment insurance 98, 202, 207, 215, *see also Unemployment allowance* Unión Cívica Radical *see UCR* Unión Cívica Radical del Pueblo *see UCRP* Unión Cívica Radical Intransigente, *see UCRI* Unión del Centro Democrático, *see UCeDé* Union density 60, 73, 109f., 120, 131f., 148, 159f., 171, 188, 191, 221, 223, 258, 275f., 278, 284

Unión Ferroviaria, *see Railway workers' unions*  Union leaders 96, 111, 113, 116, 121, 128, 134f., 138, 143f., 146, 148, 151ff., 155f., 163, 165, 171f., 174, 176ff., 181, 187, 193–199, 201, 210, 258, 261, 262, 283, 288, 291, 298 Unión Obrera Metalúrgica, *see UOM*  United States 17, 67 Universities 138, 264 UOM (Unión Obrera Metalúrgica) 188 Urbanization 100f., 105f. Uruguay 13, 17, 19, 88, 306, 308 Vandor, Augusto 141 Varieties of Capitalism *see VoC* Verticalism 319, 322 Veto points 42f., 319 Videla, Jorge 173 Viola, Roberto 173 Violence 157, 191, 223, 229, 245, 265, 324 VoC (Varieties of Capitalism) 37–40, 316 Voting *see Election* Wages 20, 50, 63f., 70, 75f., 83, 109, 112, 115, 123, 126, 134– 138, 141ff., 154, 158, 161f., 165, 178, 180, 184, 187f., 193, 195, 199, 204, 206, 217, 230ff., 234, 237, 239, 241f., 244, 266, 318 Weber, Max 72 WHO (World Health Organization) 71, 147, 158, 270f., 296, 320 World Bank 71, 95, 158, 178f., 184f., 200, 205, 217, 243, 248, 250, 270f.

World Health Organization, *see WHO* World Values Survey 95, 185, 204, 209, 218

World War I 104 World War II 17, 55, 106 Yasky, Hugo 232 YPF 244

Pamela M. Barnes | Ian G. Barnes

## **The Politics of Nucelar Energy in the European Union**

## **Framing the Discourse: Actors, Positions and Dynamics**

For the foreseeable future the overall use of nuclear electricity in the European Union is unlikely to change significantly despite the controversies surrounding its use amongst the EU's nation states. The author questions the role that nuclear electricity plays in meeting the challenges of providing secure, competitive and sustainable energy to support the development of the low carbon economy in the EU.

*2018* • *292 pp.* • *Pb. 58,00 € (D)* • *59,70 € (A) ISBN 978-3-8474-0687-7 eISBN 978-3-8474-0831-4*

Maria Marczewska-Rytko (ed.)

## **Handbook of Direct Democracy in Central and Eastern Europe after 1989**

Since the collapse of the Soviet Union the political history of Central and Eastern Europe has been mainly the story of arise, consolidation, transformation and struggles of new democratic regimes and societies. This handbook offers an instructive approach to that history focusing on the relevance of practices and institutions of direct democracy.

*2018* **•** *351 pp.* **•** *Pb. 64,90 € (D)* **•** *66,80 € (A) ISBN 978-3-8474-2122-1 eISBN 978-3-8474-1110-9*

www.barbara-budrich.net

*2021***.** *328 pp.* **•** *Pb.* **•** *39,90 € (D)* **•** *41,10 € (A) ISBN 978-3-8474-2377-5* **•** *eISBN 978-3-8474-1663-0*

Who decides to initiate change processes in organizations? Who sets the goals? What does it mean for employees to participate in change processes? The book examines organizational change processes based on collaboration between employers, employees and action researchers in Europe and the U.S. in the later part of the 20th century. The authors offer important insights into participation and change in organizations for researchers and practitioners by identifying dilemmas and paradoxes, conflicting interests and exercising of power.

## www.shop.budrich.de

Jerzy J. Wiatr (ed.)

# **New Authoritarianism**

Challenges to Democracy in the 21st century

2019 **•** 184 pp **•** Hc. **•** 36,00 € (D) **•** 37,10 € (A) **•** GBP32.95, US\$52.00 ISBN 978-3-8474-2221-1 **•** eISBN 978-3-8474-1249-6

The authors deal with comparative aspects of contemporary authoritarianism. Authoritarian tendencies have appeared in several "old democracies" but their main successes take place in several states which departed from dictatorial regimes recently. The book contains case-studies of contemporary Hungarian, Kenyan, Polish, Russian and Turkish regimes.

**Gender in Focus: Identities, Codes, Stereotypes and Politics**

Andreea Zamfira Christian de Montlibert *2018 • 370 pp. • Pb. • 49,90 € (D) • 51,30 € (A) ISBN 978-3-8474-2183-2 • eISBN 978-3-8474-1211-3*

This book deals with the interplay between identities, codes, stereo types and politics governing the various constructions and decon structions of gender in several Western and non-Western societies (Germany, Italy, Serbia, Romania, Cameroon, Indonesia, Vietnam, and others). Readers are invited to discover the realm of gender studies and to reflect upon the transformative potentialities of globalisation and interculturality. Gender in Focus: Identities, Codes, Daniela Radu (eds.)

## **The Editors:**

Gender in Focus: Identities, Codes, Stereotypes and Politics

is book deals with the interplay between identities, codes, stereotypes d politics governing the various constructions and deconstructions gender in several Western and non-Western societies (Germany, Italy, rbia, Romania, Cameroon, Indonesia, Vietnam, and others). Readers e invited to discover the realm of gender studies and to reflect upon e transformative potentialities of globalisation and interculturality.

entities, Codes, Stereotypes and Politics

*We are witnessing an incredible development in gender studies due to the volvement of many dedicated people with good knowledge and deter-*

*nation to study the relevant issues, using a complex set of social science truments. This book is an excellent example of how this research area grows ganically, as authors from different countries and cultural contexts are le to have a common vision on how gender studies shall evolve. In order see how gender issues are connected with political science and human hts this book is an instant classic."*

Andreea Zamfira, Assistant Professor, "Lucian Blaga"

Christian de Montlibert, Emeritus Professor, "Marc Bloch"

www.barbara-budrich.net

tors:

iversity of Sibiu, Romania

re credits: photo: www.lehfeldtmalerei.de ender in Focus:

N 978-3-8474-2183-2

783847 421832

iversity of Strasbourg, France

Daniela Radu, Independent Researcher

*Radu Carp, Professor of Political Science, University of Bucharest*

A. Zamfira, Ch. de Montlibert, D. Radu (eds.)

Dr. Andreea Zamfira, Assistant Professor, Lucian Blaga University of Sibiu, Romania | Dr. Christian de Montlibert, Emeritus Professor, Uni versity Marc Bloch, Strasbourg, France | Dr. Daniela Radu, Independent Researcher Stereotypes and Politi

www.barbara-budrich.net Barbara Budrich Publishers