The Autocratic Middle Class: How State Dependency Reduces the Demand for Democracy 9780691192192, 9780691192185, 9780691209777

Citation preview

t h e au t o c r at i c m i d d l e c l a s s

p r i n c e to n s t u d i e s i n p o l i t i c a l b e h av i o r Tali Mendelberg, Series Editor The Autocratic Middle Class: How State Dependency Reduces the Demand for Democracy, Bryn Rosenfeld After Repression: How Polarization Derails Democratic Transition, Elizabeth R. Nugent The Loud Minority: Why Protests Matter in American Democracy, Daniel Q. Gillion Steadfast Democrats: How Social Forces Shape Black Political Behavior, Ismail K. White & Chryl N. Laird The Cash Ceiling: Why Only the Rich Run for Office—and What We Can Do about It, Nicholas Carnes Deep Roots: How Slavery Still Shapes Southern Politics, Avidit Acharya, Matthew Blackwell & Maya Sen Envy in Politics, Gwyneth H. McClendon Communism’s Shadow: Historical Legacies and Contemporary Political Attitudes, Grigore Pop-Eleches & Joshua A. Tucker Democracy for Realists: Why Elections Do Not Produce Responsive Government, Christopher H. Achen & Larry M. Bartels Resolve in International Politics, Joshua D. Kertzer

The Autocratic Middle Class h o w s tat e d e p e n d e n c y r e d u c e s the demand for democracy

bryn rosenfeld

p r i n c e to n u n i v e r s i t y p r e s s p r i n c e to n & ox f o r d

c 2021 by Princeton University Press Copyright  Requests for permission to reproduce material from this work should be sent to [email protected] Published by Princeton University Press 41 William Street, Princeton, New Jersey 08540 6 Oxford Street, Woodstock, Oxfordshire OX20 1TR press.princeton.edu All Rights Reserved Library of Congress Cataloging-in-Publication Data Names: Rosenfeld, Bryn, 1981–author. Title: The autocratic middle class : how state dependency reduces the demand for democracy / Bryn Rosenfeld. Description: Princeton : Princeton University Press, [2021] | Series: Princeton studies in political behavior | Includes bibliographical references and index. Identifiers: LCCN 2020014695 (print) | LCCN 2020014696 (ebook) | ISBN 9780691192192 (hardback) | ISBN 9780691192185 (paperback) | ISBN 9780691209777 (ebook) Subjects: LCSH: Civil service—Political activity—Former communist countries—Case studies. | Middle class—Political activity—Former communist countries—Case studies. | Authoritarianism—Former communist countries. | Democracy—Former communist countries. | Democratization—Former communist countries. | Former communist countries—Economic conditions. | Former communist countries—Politics and government. Classification: LCC JN96.A69 P64 2021 (print) | LCC JN96.A69 (ebook) | DDC 306.20947–dc23 LC record available at https://lccn.loc.gov/2020014695 LC ebook record available at https://lccn.loc.gov/2020014696 British Library Cataloging-in-Publication Data is available Editorial: Bridget Flannery-McCoy and Alena Chekhanov Production Editorial: Kathleen Cioffi Cover Design: Layla Mac Rory Production: Erin Suydam Publicity: Kate Hensley and Kathryn Stevens Cover image: Rally for Russian President Vladimir Putin at the Luzhniki Stadium, Moscow, March 3, 2018. Maxim Shemetov / REUTERS This book has been composed in Arno Printed on acid-free paper. ∞ Printed in the United States of America 10

9

8

7

6

5

4

3

2

1

contents

Acknowledgments 1

vii

The Autocratic Middle Class

1

2 State Dependency and Middle-Class Demand for Democracy

37

3 The Post-Communist Middle Classes, the State, and Democratization

66

4 Rethinking the Middle-Class Protest Paradigm

100

5 Choosing to Work for the State

134

6 Revolution, Democratic Retrenchment, and the Middle Class

155

7 Aligning the Middle Class with Autocracy: Rhetoric and Practice

193

8 Conclusion

222 Appendix I: Regression Results Bibliography

245

251

Index 269 A link to the online appendix can be found at https://press.princeton.edu/ books/paperback/9780691192185/the-autocratic-middle-class.

v

acknowledgments

in the writing of this book, I have incurred many debts. The project began as a doctoral dissertation at Princeton University, where I was fortunate to have the guidance of several scholars whom I greatly admire. Mark Beissinger advised this project from its earliest beginnings. He asked difficult questions, offered patient encouragement, and has been an invaluable source of guidance and support. Chris Achen was also an exemplary mentor. His wit and wisdom made our conversations one of the singular pleasures of graduate training at Princeton. I am also deeply indebted to Grigo Pop-Eleches for his insightful comments on every draft. His feedback pushed me to sharpen the project’s empirics and helped me to wrestle with alternative arguments. At Princeton, I also wish to thank Carles Boix, Rafaela Dancygier, Amaney Jamal, Steven Kotkin, Markus Prior, and Rory Truex. Nuffield College, where I was a postdoctoral fellow during the 2015– 2016 academic year, provided the perfect setting for revising this manuscript. Ben Ansell gave valuable guidance on framing and engaged generously with the project’s arguments. My wonderful writing group in Oxford—Ezequiel González-Ocantos, Jody LaPorte, Luis Schiumerini, Jazmín Sierra, and Maya Tudor—helped me negotiate the difficult transition from dissertation to book. Their friendship and thoughtful suggestions kept me on track. This project also benefitted greatly from conversations at Nuffield with Nancy Bermeo, Ray Duch, Geoff Evans, Olga Onuch, Henry Thompson, and Laurence Whitehead. At the University of Southern California, my colleagues generously engaged with this project in seminars and workshops, in conversations on campus and on the hiking trail. Gerry Munck, whose own work on democratization has taught me so much, has been a wonderful mentor. His detailed feedback at various junctures and many helpful insights significantly strengthened the manuscript. Ben Graham has been a terrific friend and incredibly generous colleague. He supported this project through careful comments, vii

viii

acknowledgments

transcription of my book workshop, and encouragement when I needed it most. I could not ask for better junior faculty friends than Erin Baggott Carter, Brett Carter, Andrew Coe, Morris Levy, James Lo, and Jon Markowitz. I also thank Jeb Barnes, Dennis Chong, Ann Crigler, Pat James, Jane Junn, Saori Katada, Stan Rosen, and Jeff Sellers for their constructive engagement and advice along the way. Individual chapters of the manuscript were greatly improved by feedback from participants in the CIS Working Paper Series, and especially from comments by Evgeniia Iakhnis, Xinru Ma, and Kelly Zvbogo. I am also extremely grateful for a terrific discussion of the manuscript at the book workshop held at USC’s Center for International Studies in the winter of 2016. I thank especially my external discussants Stephen Haggard and Cynthia Kaplan for their comments on each and every chapter and for providing valuable advice on the project’s overall framing. I owe a special debt of gratitude to Dave Kang for supporting the workshop as director of the Center for International Studies, to the many faculty and graduate students who participated, and to the wonderful CIS staff who helped to organize it, Cort Brinkerhoff and Madeline Brown. Before completing this manuscript in the fall of 2019, I joined the faculty at Cornell University and relocated with my family to Ithaca. I wish to thank my wonderful colleagues in the Department of Government and especially Begüm Adalet, David Bateman, Val Bunce, Ali Cirone, Gustavo Flores-Macías, Jill Frank, Sabrina Karim, Tom Pepinsky, Rachel Riedl, Ken Roberts, Sid Tarrow, Jeremy Wallace, Nic van de Walle, and Jessica Weiss for advice and encouragement as I made final revisions. Previous versions of several chapters were presented at annual meetings of the American Political Science Association and Midwest Political Science Association. I was fortunate to benefit from the insightful comments of Scott Gehlbach, Henry Hale, Tom Remington, and Rory Truex, my discussants at those conferences. I also wish to thank seminar participants at Columbia, Essex, George Washington, the Higher School of Economics in Moscow, the London School of Economics, NYU, Oxford, Princeton, UCLA, the University of Southern California, Vanderbilt, and Washington University St. Louis for excellent feedback. For helpful discussion and comments along the way, I also thank Graeme Blair, Noah Buckley, Volha Charnysh, Michael Donnelly, Sarah El-Kazaz, Jordan Gans-Morse, Yanilda Gonzalez, Aram Hur, Adam Krawitz, Tomila Lankina, Noam Lupu, Israel Marquez, Gwyneth McClendon, Michael Miller,

a c k n o w l e d g m e n t s ix

John Reuter, Joan Ricart-Huguet, Graeme Robertson, Arturas Rozenas, Rudy Sil, David Szakonyi, Tariq Thachil, Josh Tucker, Carlos Velasco Rivera, and Seth Wessler. Tim Frye kindly read drafts of several of this book’s chapters and offered valuable advice at various stages of the project. My field research would not have been possible without the support of many different people and institutions. In Russia, I discussed ideas from this project with Anna Andreenkova, Elena Avraamova, Vladimir Gimpelson, Yevgeny Gontmakher, Alexey Grazhdankin, Lev Gudkov, Svetlana Mareeva, Lilia Ovcharova, Ovsey Shkaratan, Irina Soboleva, Natalia Tikhonova, Mark Urnov, and Denis Volkov. The project, in turn, has benefitted from their valuable insights and suggestions. The Higher School of Economics was a terrific place to pursue my research. Mark Urnov opened his personal archive to me and was particularly generous with his contacts in Russian politics and academia. His support and good humor greatly enriched my time in Moscow. I am also grateful to the Levada Center and the Foundation for Public Opinion (FOM) for generously sharing their data. It has been a pleasure to collaborate with the Levada Center and the Institute for Comparative Social Research (CESSI) for a number of years. In Ukraine, I wish to thank Anatoly Arseenko, Elena Besedina, Liudmila Cherenko, Oleksandr Demianchuk, Evgeny Golovakha, Oleksey Haran, Olga Kupets, Olga Kutsenko, Mikhailo Mishchenko, Denys Nizalov, Olena Nizalova, Svetlana Oksamitna, Vladimir Paniotto, Tatiana Petrushina, Liudmila Shanghina, Elena Simonchuk, and Ganna Vakhitova for helpful discussions. Vladimir Sayenko and his mother Lydia were wonderful hosts, and my time in Kiev was improved immeasurably by their hospitality. Being met by a familiar face when I got off the train and the singular pleasure of seeing a friend I hadn’t seen in years added to the joy of fieldwork. I am grateful also to have shared many cups of tea and conversations with Sasha, with whom I lived in Kiev. I also owe a debt of gratitude to the Institute of Sociology of the National Academy of Sciences of Ukraine and to Victor Stepanenko for facilitating my research in Kyiv. The Institute of Sociology, the Razumkov Centre, and the Institute for the Study of Labor (IZA) very generously allowed me to use their data for this project. In Kazakhstan, Sara Alpysbaeva, Alima Bissenova, Sholpan Esimova, Gulmira Ileuova, Lyaila Ivatova, Elena Maltseva, Daulet Mynzhasarov, Bakhytnur Otarbaeva, Botagoz Rakisheva, Nurbek Sayasat, Caress Schenk, Nazym Shedenova, Alexei Trochev, Aigul Zabirova, Ernar Zharkeshov, and Aiman

x

acknowledgments

Zhusupova made important contributions to this research. I also wish to thank the Academy of Public Administration under the President of the Republic of Kazakhstan in Nur-Sultan (formerly Astana) for welcoming me as a visiting researcher and my colleagues at the Academy for their camaraderie over many cafeteria lunches and cups of tea. Professor Lyaila Ivatova opened her home to me and was an incredibly gracious host. I am grateful for her company, conversation, and warm manty as temperatures in Astana dipped and the wind whipped in winter months. The research for this book was made possible by the generous financial support of several institutions. I gratefully acknowledge funding for my field research from the U.S. Department of Education’s Fulbright-Hays program, the Princeton Institute for International and Regional Studies, and the Mamdouha S. Bobst Center for Peace and Justice. Dave Kang provided the funding for a workshop on this manuscript at USC’s Center for International Studies. At Princeton University Press, I wish to thank Eric Crahan for his early interest in my manuscript and Bridget Flannery-McCoy for seeing it through to publication. I am also grateful to Alena Chekanov and Kathleen Cioffi for their help along the way, and to Wendy Washburn for meticulously editing every page. Tali Mendelberg, the series editor, and three anonymous readers offered excellent feedback, along with praise for the project that helped motivate me to finish it. My husband, Ian Kysel, has shared this project’s evolution and much more since we met at Swarthmore College. He read every draft and talked through every chapter—on bikes along the Hudson, on the towpath along the Thames, on the hiking trails in Griffith Park, and climbing the gorges in Ithaca. His love and patience made the writing possible. Littlest but not least, Ezra joined our family between rounds of revisions. I was at my desk working on chapter 3 when I went into labor. This project has grown up with him. Finally, my parents, Raymond Rosenfeld and Janelle McCammon, met as doctoral students in political science at Emory in 1970. When I took my first tenure-track job, my father sent me fifty years of the American Political Science Review to make my office feel more like home. Their many gifts inspire me as a parent and scholar. This book is dedicated to them.

t h e au t o c r at i c m i d d l e c l a s s

1 The Autocratic Middle Class Thus it is manifest that the best political community is formed by citizens of the middle class, and that those states are likely to be well-administered in which the middle class is large, and stronger if possible than both the other classes, or at any rate than either singly; for the addition of the middle class turns the scale, and prevents either of the extremes from being dominant. — a r i stot l e , t h e p o l i t i c s1

much has been expected of the middle class. In two thousand years of scholarship on democratization—from Aristotle in the fourth century to Acemoglu and Robinson today—two views of the middle class have dominated. On the one hand, the middle class has been seen as a source of social stability. In this view, the middle class favors durable economic institutions and maintenance of the political status quo. Politically and socially, it acts as a conservative force. By moderating the conflicting redistributive demands of rich and poor, it stabilizes existing regimes. On the other hand, the middle class has been cast as an agent of political change and democratization. In scholarship on the the first wave of democratization in the nineteenth century, the second wave of democratizations following World War II, and the third wave of democratic transitions from the mid-1970s through the collapse of communism, the middle classes have been ascribed a critical role (see, e.g., Lipset 1960; Huntington 1991; Acemoglu & Robinson 2006; Ekiert 2010). To date, no comparative scholarship has sought to reconcile these contrasting views.

1. Aristotle (1999, 96). 1

2 chapter 1

This is a matter of some urgency. Civic revolution has increasingly become the norm in recent episodes of democratization. The very fact of mass mobilization during transition, as well as its character, appears to have lasting impact on the subsequent quality of democracy (Haggard & Kaufman 2016). Although these popular insurrections are often labeled as middle-class revolutions, our understanding of the constituencies involved as well as their motivations remains limited. Moreover, while the role of the middle classes in mass uprisings against authoritarianism has varied across time and space, no clear consensus has emerged to explain these patterns (e.g., Koo 1991; Jones 1998; Shin 1999; Rueschemeyer, Stephens & Stephens 1992, 272). Recent protests around the world demonstrate the critical importance of understanding how citizens form democratic preferences under autocracy, and thereby the sources of bottom-up pressure for political change. In Eastern Europe and the former Soviet Union, the collapse of communism inaugurated a period of market reforms, promising social transformation, prosperity, and democracy. The growth of a new, market-based middle class was widely expected to buttress these twin economic and political transitions. The transition from state socialism ended the institution of universal state employment and broadened the scope of economic activities neither initiated nor directly controlled by the state. While these processes reduced the size and influence of the state, they did so unevenly and in unexpected ways, leaving the state as an influential patron of the region’s educated, professional classes. At the same time, the early post-communist years saw a staggering rise in unemployment, inflation, and poverty across the region. Initially, at least, these conditions produced a sharp differentiation of incomes and the immiseration of the communist-era middle class (Ekiert 2010, 112). During the 1990s, inequality rose most rapidly in the former Soviet states, where market reforms proceeded more slowly than in Eastern Europe, and were least complete (Hellman 1998). Economic recovery took off only a decade later, around 2000, when the region’s economies joined the high growth trajectory of other emerging markets. During this recovery, the post-Soviet middle classes have grown markedly, yet not only, or even primarily, because small business, entrepreneurship, and private enterprise have flourished. Instead, the rise of the middle class has been largely tied to the state. Today, it is clear that market reforms have failed to replace the old state-dependent middle classes of the communist era with an independent, entrepreneurial middle class. Moreover,

t h e au t o c r at i c m i d d l e c l a s s

3

a new state-dependent middle class has emerged, only deepening existing cleavages. This book seeks to expand our grasp of authoritarian resilience and bottom-up pressures for democratization in states where economic growth is increasing the size of the middle class. Contrary to the conventional expectation, I ask why and under what conditions growth of the middle class may not increase popular pressure on regimes to democratize. To answer this question, I turn to a wide array of survey data on the political preferences and behaviors of the middle classes in the post-communist countries as well as qualitative evidence collected during nine months of fieldwork in the region. In contrast to existing scholarship, this book emphasizes that a variety of development strategies can drive an expansion of the middle class. These strategies differ in their effect on the formation of democratic constituencies. There are multiple pathways to the middle class, and not all of them, I argue, lead to greater support for democracy. In this book, I tell the tale of two middle classes, with differing degrees of state dependency, that are nonetheless treated as one homogenous group by existing theories. The story focuses on contexts that combine autocracy with an economically interventionist state. These conditions are not rare. Pervasive public sectors are a feature of politics in the Middle East, Africa, Asia, Latin America, as well as the former Soviet Union (e.g., Hertog 2013; Diamond 1987; Weyland 1996; Gervasoni 2010; Oliveros 2016; Darden 2008). These conditions are widespread in resource states, but they are also present in post-socialist countries both with (e.g., Azerbaijan) and without oil (e.g., Belarus). The post-Soviet middle classes typify the type of divided middle class that develops under autocratic state institutions and extensive state economic engagement. Following the collapse of communism, an important new cleavage appeared between the old, state-dependent middle class and the nascent middle class of the new market-based economy (see also, e.g., Ekiert 2010). Thanks to statist economic policies and a decade of economic recovery, the state-dependent sectors of the middle class in the former Soviet Union are growing. Civil servants and other government budget-sector employees are also increasingly well paid. Where formal wages remain low, the state has turned a blind eye to bribe-taking and graft by those on its payroll, ensuring that corruption helps to compensate. The upshot is that lucrative white-collar positions in public-sector enterprises, banks, and financial service firms as

4

chapter 1

well as the state administration have become among the clearest pathways to the middle class across the former Soviet states. As society has grown wealthier, even state teachers and doctors find it increasingly lucrative to exploit their official positions (INDEM 2005). This is the new state-dependent middle class of post-communist quasi-capitalism. The process of middle class formation now transpiring in the postSoviet states recalls what Barrington Moore called the “feudalization” of the bourgeoisie. While a number of influential approaches to democratization— among them modernization theory, its values-based variants, and redistributive theories—expect the middle class to be a force for democracy,2 an earlier macro-sociological literature understood that when the middle class maintains close ties with the state it can be a deeply conservative political force. This book problematizes recent theories of democratization, which largely ignore the implications of how growth of the middle class is achieved. Indeed, this book’s central argument is that failure of the middle class to gain economic autonomy from the state stymies support for political change and contributes to authoritarian resilience. This study fills an important gap in the literature. While the middle classes have often been seen as a linchpin in successful democratic coalitions, very little work to date has provided a detailed empirical examination of this group’s political preferences. As Ansell & Samuels (2014, 46) write, “Scholars have paid insufficient attention to the concrete interests of the ... middle classes in the study of regime change.” The distinctive contribution of this book is to show concretely, at the individual level and for a broad set of cases, how state-led development produces middle classes that are beholden to autocratic regimes. The findings thus contribute to our understanding of why some countries are democratic, while others are not. They also speak to long-standing debates in comparative politics about the role of class actors in democratic transition and participation in contentious politics, as well as a newer literature on the rise of state capitalism in contemporary autocracies. In addition, this research advances our understanding of the consequences of state-directed development. While this subject has thus far been studied primarily at the institutional and macrosociological level (Gerschenkron 1962; Bellin 2002; Kohli 2004; Wengle 2015), this book focuses on how statist strategies of development structure 2. See, e.g., Lipset (1960), Inglehart & Baker (2000), Boix (2003), Acemoglu & Robinson (2006), Welzel & Inglehart (2008).

t h e au t o c r at i c m i d d l e c l a s s

5

the incentives of individual actors to side with autocracy. It also provides new insights into the persistent puzzle of development without democratization. This introductory chapter is organized as follows. I first lay out my argument in brief. I then situate it within existing scholarly debates. My focus is at the micro-level, on how state dependency shapes the incentives faced by the middle classes to oppose the extension of democratic institutions. I next provide an overview of my empirical strategy. I then conclude with an outline of the book’s remaining chapters.

Argument in Brief: The Autocratic Middle Class The question of how society’s class structure affects the prospects for democracy has long been a core preoccupation in the study of politics. One of the most enduring arguments in this literature holds that growth of the middle class gives rise to democratization. This view is echoed in many canonical approaches to democratization, including political economy’s redistributive theories, modernization theory, and its values-based variants. These theories expect the middle class—defined variously in terms of income, education, or occupation—to play a key role in democratic transition. Yet remarkably little systematic research examines middle-class attitudes toward democracy in autocratic settings. As the developing world’s middle classes have grown rapidly over the past two decades, largely in nondemocracies, this gap has become even more significant. Will these growing middle classes “turn the scale,” enhancing their countries’ prospects for democracy? Or standing the conventional logic on its head: Might it be the case that certain modes of state-supported middle-class growth, in fact, delay democratization? This book argues that a middle class whose status depends on public employment for an authoritarian state is often antithetical to democracy. Drawing lessons from the post-communist countries, it sheds light on how the economic institutions of state employment benefit autocrats, helping them to secure the support of key middle-class constituencies. While existing theories expect that growing middle classes will confront old networks of patronage and privilege, they ignore crucial variation in the extent to which the middle classes are stakeholders in existing autocratic systems. Indeed, in many contemporary autocracies, expanding opportunities within the state sector—within the state bureaucracy, public institutions, and state-owned

6

chapter 1

enterprises—supply the principal avenues of mobility into the middle class and drive its numerical expansion. Under autocracy, an expansive public sector can dampen democratic demands. How a person gains and maintains their place in the middle class affects their expected benefit from democracy and willingness to challenge the political status quo. The distinction between the middle classes of the state and private sectors is thus real, politically significant, and becomes the locus for divergent economic and political interests in autocratic systems. This distinction forms the basis for access to privileges, benefits, and at times even differing status before the law. Where the state supplies the principal avenues of social mobility into the middle class, a self-conscious, democratizing middle class fails to form. Material incentives based on employment status and workplace mobilization all militate against the expression of democratic demands. Rather, selective incentives are used to mobilize growing middle classes in support of autocracy. I connect the middle classes’ reticence about democracy to the power of authoritarians to bestow or withhold benefits that are contingent and could be diminished or disrupted by change in the political control of the state. Some of these benefits are formal, like jobs, employment protections, shorter working hours, access to healthcare, and loans on advantageous terms from state banks. Other benefits are informal, though rooted in an official position, like opportunities to solicit bribes and kickbacks and preferential treatment by other state institutions. In turn, the power of state selective incentives varies with an individual’s exit options. These options are limited when an extensive public sector employs a preponderance of the middle class, crowding out private-sector alternatives. Limited exit options deepen dependence on the state. By tying future benefits to regime continuity, state economic engagement that concentrates rents in the public sector creates its own middleclass constituencies, encouraging opposition to democratization and dividing potential democratic coalitions. If we survey the developing world today, it is possible to tell a tale of two middle classes: one whose economic opportunities and life chances are owed directly to the state, and the other whose economic livelihood is less directly dependent. These categories roughly correspond to two pathways into the middle class: one through the state, the other through the private sector. While the latter group may fear predation by a large authoritarian state, the former group has the state to thank for its upward mobility and middleclass status. White-collar workers at state and quasi-state enterprises, teachers, doctors, and other public-sector professionals, as well as those employed in

t h e au t o c r at i c m i d d l e c l a s s

7

the state administration, all typify what I call the state middle class. These groups comprise a middle class that, I argue, helps to stabilize autocracy and has been largely ignored by existing theories. While reticence about democracy often goes hand in hand with greater satisfaction with the status quo and support for the incumbent autocrat, this need not and, indeed, is not always the case. Even where the middle class expresses grievances with the status quo or is not generally supportive of the current regime, it will remain unsupportive of democratization if it associates transition with diminished life chances. The latter case describes Ukraine after the Orange Revolution. The former case describes Russia during the post-election protests preceding Putin’s third term as President. Each of these cases as well as cross-national evidence shows that reticence about democracy is not merely a function of support for a particular autocrat or satisfaction with the status quo. This study adopts a sociological definition of middle class emphasizing human and social capital. The definition, described in detail in the next chapter, captures the distinction between highly educated white-collar and professional strata versus less educated routine and manual laborers. This approach is notably distinct from a normative view of the middle class as a carrier of democracy, as synonymous with the capital-owning bourgeoisie, or as an exclusively income-based category. The middle class thus defined most closely resembles the middle class of educated professionals in modernization theory and its values-based variants (e.g., Lipset 1959; Welzel & Inglehart 2008). It also shares with Collier & Collier’s (2002) “middle sector” an emphasis on members of a broad range of occupational groups between the working class and economic elite. At the same time, it is broader than Moore’s (1966) bourgeoisie and empirically has above-median income like the middle class of elite competition theories (Ansell & Samuels 2008). This approach facilitates careful comparison with these other theories and a sound basis for empirical investigation. Education and occupation are theoretically well-grounded determinates of class and socioeconomic status (SES), readily available in most cross-national surveys, well correlated with each other empirically, and much more reliably measured than income—for which they are often used as proxies.3 3. Indeed, education and occupation are the social stratification measures that scholars of class have considered most relevant in shaping life chances. See, e.g., the review of Weberian class analysis in Breen (2005) or the work of Weeden & Grusky (2005) on the importance of occupations. The approach I take represents a theoretically informed and context-appropriate

8 chapter 1

Scope of the Argument The argument just summarized is not specific to particular types of autocracy, with the potential exception of rightist dictatorships.4 The cross-national sample I employ in chapter 3 covers virtually the full spectrum of authoritarian regime types from highly repressive (Uzbekistan) to wavering on the cusp of democracy (Georgia). Moreover, the cases I examine in greater depth in subsequent empirical chapters range from hegemonic single party (Kazakhstan) to relatively pluralistic competitive authoritarianism5 (Ukraine). Although there are no right-wing dictatorships among the cases considered in this book, such regimes are probably beyond the scope of my theory. If faced with the choice between substantial neoliberal economic reforms under a rightest military junta and a populist democrat advocating leftist policies, the intertwining of economic and political considerations regarding transition would likely lead the public-sector middle class to side with democracy. This choice set describes several transitions to democracy in Latin America, including in Chile, Brazil, and Argentina. That said, the argument’s intended scope is not limited to the postcommunist region. Rather than hinge on historical legacies of communism, in line with Bellin (2000), I expect the argument to travel to other latedeveloping autocracies with high levels of state economic engagement and to apply whether the logic of the state is developmental or patrimonial. Over the past two decades, according to World Bank criteria, the middle class expanded rapidly across the developing world. At the same time, a wave of market backlash and nationalization has returned to the fore the notion of state control over strategic sectors and ideas implicit in the developmental state tradition. Although many studies, including this one, show that economic and political reforms are intertwined in the minds of post-communist citizens, the book’s central argument does not hinge primarily on the association of democratization with neoliberalism. Even where neoliberal reforms are not expected to accompany the triumph of democratic forces, middle-class

variant of neo-Weberian class typologies, such as the widely used measure associated with Erikson & Goldthorpe (1992). 4. Under democracy, civil service protections, greater transparency, and the rule of law significantly change the dimensions of state dependency examined in this study. 5. Or, in some estimates, unconsolidated democracy.

t h e au t o c r at i c m i d d l e c l a s s

9

table 1.1. Post-Communist Countries by Regime Type Nondemocracies

Democracies

Armenia Azerbaijan Belarus Bosnia Kazakhstan Kyrgyzstan Russia Tajikistan Turkmenistan† Uzbekistan

Albania Bulgaria Croatia Czech Republic Estonia Georgia∗ Hungary Latvia Lithuania Macedonia Moldova Montenegro Poland Romania Serbia Slovak Republic Slovenia Ukraine∗

Note: ∗ Wavered on the cusp of democracy during the period under study. † Excluded from survey samples.

groups dependent on the state may be reticent about democracy for at least two additional reasons: (1) they may fear that being associated with the preceding regime will bode poorly for them following transition, and (2) they may expect that rising attention to the rule of law under democracy will reduce opportunities to earn informal rents.

Descriptive Data on the Post-Communist Region Table 1.1 lists the post-communist countries that I examine in this book and breaks them down by regime type. Throughout the text, I use the term “post-communist” to refer collectively to all of the countries in table 1.1 and “post-Soviet” or “formerly Soviet” to refer to the fifteen republics that once comprised the Union of Soviet Socialist Republics (USSR). Of these, my primary interest is in the nondemocracies, plus states that wavered on the cusp of democracy during this period (e.g., Georgia and Ukraine). The three Baltic States (Latvia, Lithuania, and Estonia) and Moldova, which democratized earlier and maintained stable democracies over this period, are thus excluded

10 c h a p t e r 1

Total employment

Percent in public sector

80

Middle−class employment 71 58

60 46 40

35

20

0

Autocracies

Democracies

Autocracies

Democracies

figure 1.1. Public employment in 27 post-communist countries by regime type. Data source: European Bank for Reconstruction and Development (EBRD) Life in Transition Survey (LiTS), 2006 (all post-communist countries).6

when I refer to the post-Soviet autocracies. Due to the absence of reliable survey data, Turkmenistan is not included in this book’s empirical analyses. Figure 1.1 provides descriptive statistics on public employment in the nondemocracies of the former Soviet Union and the democracies of Central and Eastern Europe. Recalling that these countries once shared the institution of universal state employment, it is not difficult to see why state employees in the autocracies might be worried about democratic transition.7 As the first bar on the far left of the plot shows, the public sector is a major employer in the post-Soviet autocracies. Even after fifteen years of privatization, public employment continued to comprise just under half of total employment 6. Mongolia and Turkey are excluded from this otherwise East European and formerly Soviet sample. 7. These figures combine all public employees, though certainly the implications of democratization may vary across parts of the public sector. While employees of state-owned enterprises lost their jobs as firms privatized and new owners sought to enhance profitability, in some cases the size of the state bureaucracy actually grew following transition. For example, in the decade after the transition, O’Dwyer (2004) finds that the state administration grew dramatically in Poland and Slovakia, as weak parties entrenched themselves through patronagebased recruitment. By contrast, no such growth occurred in the Czech Republic or Hungary, where the party system was stronger. O’Dwyer’s theory implies, however, that these effects are likely to be short-lived, since institutionalized party systems will, over time, constrain patronage-based recruitment.

t h e au t o c r at i c m i d d l e c l a s s

11

in the nondemocracies at the time these data were collected in 2006. In the democracies, meanwhile, it comprised only about a third of total employment. Turning to the right side of the figure, we see that the post-Soviet middle classes remain more heavily dependent on the state than other groups. Indeed, the public sector employs roughly seven in ten middle-class individuals (71%) across the post-Soviet countries (vs. 58% in the democracies). The share of working middle class employed in the state sector is highest in Azerbaijan (84%), Belarus (81%), and Uzbekistan (79%), and considerably lower (though still a clear majority) in Russia (59%). These figures make a strong case for the importance of public sector institutions in shaping the life chances and career prospects of the post-Soviet middle class. Finally, it is worth noting that in terms of household expenditures, state employees in the former Soviet countries are better off than others. In fact, median household expenditures in the autocracies are 17 percent higher for state employees than for those employed outside the state sector. Given that the official wage gap favors the private sector in many former Soviet countries, these expenditure data strongly suggest that state employees in the autocracies benefit from informal rents (e.g., Sharunina 2013; Gorodnichenko & Peter 2007). By contrast, equivalent data from the democracies suggest virtually identical median household expenditures by state and non-state employees.8 As long as the public-sector middle class remains large and enjoys access to illicit income, it will have fewer incentives to favor democratization and associated political transparency.

The Middle Class and Democratization This argument, though intuitive, is at odds with much of what has been written on the politics of the middle class. As the middle classes expand across the developing world, observers have frequently assumed that they will demand better governance and support democratization. What do existing theories tell us about the mechanism through which development, by enlarging the middle class, will lead to democratization? This section surveys existing explanations for the political preferences of the middle class 8. Consistent with these statistics, Róna-Tas (1997) reports that by the mid-1990s in Hungary, salaries and benefits had become mostly comparable across the public and private sectors. However, as labor discipline became stricter, state employees were no longer able to make use of company resources and time for informal business activities that they once conducted on the side (203, 205).

12 c h a p t e r 1

and relates the book’s theoretical framework to several newer approaches and recent literature on authoritarian resilience. Conventional wisdom holds that one of the signal developments en route to liberal democracy in Western Europe was the rise of the middle class. That reading of history continues to shape observers’ expectations about the likely political influence of growing middle classes in the developing world today. Fukuyama (2014) provides one account: In nineteenth-century Europe, wages began to rise, improving workers’ standards of living. At the same time, demand increased for skilled professionals like engineers, accountants, and lawyers. As a result, workers became more affluent, came to own property, and became more highly educated. They also became more politically conservative: “more likely to vote for political parties that could protect their privileges rather than ones pushing to overturn the status quo” (415).9 By the second half of the twentieth century, the median voter had become “a middle class individual with a stake in the existing system” (417). This process is said to have shored up democracy in Europe and is thought to be happening across many parts of the developing world today. However, the conditions greeting the rise of the middle class were notably different in the early democratizers, and especially in England (Landes 1998). First, the state apparatus, as van de Walle (2014) has observed, was much smaller historically. Government expenditure as a share of national income was a mere fraction of what it is in the modern era. Second, while autocracy is the dominant regime type among developing countries today, England had been experimenting with the institutions of democracy since the first half of the 19th century. Accordingly, the status quo bias exhibited by Europe’s middle classes as they became more politically influential favored increasingly democratic political systems. The post-war expansion of the public sector only reinforced western Europe’s democratic middle-class constituency. Moreover, growth of the middle class has not always enhanced the prospects for liberal democracy, even in Europe. Although the English case is often taken as paradigmatic of the first-wave democracies, other first-wave cases suggest a more complex view. As an earlier sociological literature highlights, monarchic France and absolutist Germany exemplify cases where the 9. Of course, there are also critiques of class-based accounts of European democratization. Capoccia & Ziblatt (2010), for example, argue that competition, conflict, and alliance among political parties (which did not merely replicate class divisions) better account for democracy’s emergence in Europe.

t h e au t o c r at i c m i d d l e c l a s s

13

rising middle classes actually reinforced authoritarianism. Moore (1966), for instance, argues that the French bourgeoisie’s dependence on the crown delayed the emergence of full democracy. And Dahrendorf (1967) observes that Germany’s state-directed industrialization resulted in the co-optation of the upwardly mobile middle class. In both cases, the absence of an independent economic base apart from the state limited rising groups’ interest in making political claims favoring democratization. Why then do existing theories expect the prospects for democracy to improve with growth of the middle class, which occurs under autocracy? This book’s framework can be distinguished from influential arguments about the political role of the middle class in two broad strands of literature: modernization theory and its values-based variants and political economy’s redistributive theories of democratization.

The Middle Class in Modernization Theory Classical modernization theory posits a pivotal role for the middle class in economic development and democratization. As articulated by Lipset (1960), and echoed by other influential scholars (e.g., Dahl 1971; Huntington 1991), economic development changes the class structure of society, improving the chances for democratization and democratic stability. At the individual level, scholars have long linked rising educational levels and greater occupational specialization with democratic values (Lipset 1960) and the resources to participate effectively in democratic politics (Almond & Verba 1963; Inkeles 1969). Education has been associated with rising political efficacy, greater political trust, and a growing commitment to tolerance, freedom, and democracy.10 At the same time, changes in the nature of work are thought to have political consequences. Democracy’s chances are said to improve as the share of jobs that place a premium on independent thinking and endow individuals with the skills for political activism grows (Lipset 1960).11 10. See, for example, Prothro & Grigg (1960); Gibson, Duch & Tedin (1992); Barro (1999); and Przeworski, Alvarez, Cheibub & Limongi (2000), but also Weil (1985) for an example of the perspective that education’s effect on liberal attitudes is conditional; Hakhverdian & Mayne (2012) for the view that the relationship between education and trust is context specific; and Acemoglu et al. (2005) for evidence that the link between education and democracy crossnationally is spurious. 11. O’Donnell’s (1973) theory of bureaucratic authoritarianism famously challenges this view. The perspective in this book is consistent with his claim that under certain conditions

14 c h a p t e r 1

More recent modernization arguments build explicitly on this earlier preoccupation with values (e.g., Inglehart 1977; Inglehart 1990; Abramson & Inglehart 1995; Inglehart & Baker 2000; Welzel & Inglehart 2008). “Effective” democracy, Welzel & Inglehart (2008) argue, requires a social base of support to ensure that formal democratic institutions are honored in practice. Economic development makes democracy a stable political outcome, because it is accompanied by a deepening of self-expression values and human empowerment (see also Welzel 2013). In short, modernization theory proposes that human capital formation paves the way for democracy. Like these theories, I also emphasize education and occupation. Indeed, the post-communist middle classes are the product of rising educational levels, first under communism and then with the expansion of higher education following independence. They are likewise the product of rising managerial and professional opportunities—again, first in the communist era and then with the expansion of the white-collar workforce in the post-independence period. Importantly, however, modernization theory and its values-based variants cannot easily explain why individuals, who are similar in terms of income, education, and occupation hold different views of democracy depending on their relation to the state. In contrast to my own framework, modernization theory ignores the intervening role of the state through public enterprises and organizations. By ignoring how state economic institutions shape citizens’ incentives, these theories over-predict homogeneity in the political preferences of the middle class. My approach foregrounds precisely these interests and occupational settings, revealing that the rise of a state-supported middle class may actually reinforce authoritarianism and delay democratization. Moreover, while modernization theory expects education to increase support for democracy, I show in chapter 3 that higher levels of education do not invariably make the middle class more democratic.12

development is inimical to democracy. Indeed, the technocratic actors within the state bureaucracy that play an important role in O’Donnell’s story are this book’s state middle class. I, however, extend the scope of the argument to cases beyond Latin America where politics are not driven primarily by the deficits of early industrialization and the legacies of populism associated with import substitution industrialization (ISI). 12. Welzel (2013) argues that, under communism, modernization led to a “build-up of emancipative values for quite some time, until the downfall of Soviet imperialism open[ed] the gate for democratization” (163). If such a process is underway today in the modernizing

t h e au t o c r at i c m i d d l e c l a s s

15

This book thus also speaks to debates on the relationship between education and democratic values, suggesting that it depends crucially on political and institutional context. As Campbell (2013, 37) writes, of the various ways to measure SES, education is typically the most predictive and robust; other factors “must ‘survive’ being in the statistical ring with [this] 800-pound gorilla.” Indeed, my findings suggest that the additional explanatory power of income is quite limited after accounting for the effect on democratic preferences of human and social capital. I also find, however, that education alone, and in particular higher education under autocracy, does not consistently give rise to greater support for democracy (see also Wang 2018). Post-Soviet education—during a period when many governments experimented with democracy and institutions of higher education took on the task of preparing democratic citizens—has a weak positive effect on democracy support. By contrast, Soviet education—with its ideological and explicitly antidemocratic character—has if anything a negative effect. Most importantly, education is closely tied to occupation, which this book shows decisively shapes preferences toward the political economic system. In the statistical ring, I find that state employment not only survives but triumphs over the 800-pound gorilla of education in analyses of democratic attitudes, pro-democracy protest participation, and voting. Thus, whether educated in the Soviet period or the product of post-Soviet higher education, the state middle class shares the same low level of support for democracy as the working class. These findings imply that to understand whether the educated will prefer democracy, we must also consider how state economic institutions affect their life chances. They may also be interpreted as implying that as autocratic countries promote tertiary education in pursuit of development, they will likely need to balance these policies with good jobs, good benefits, and other perks that keep educated groups satisfied.

The Middle Class in Redistributive Theories of Democratization A second broad strand of literature focuses on the economic position of the middle class. Whereas modernization theories highlight the importance of a cluster of socio-structural variables for predicting political preferences and post-communist autocracies, pent-up demand should be evident in reliable surveys, particularly among groups that have benefitted the most from economic development. The empirical findings that follow do not support that interpretation.

16 c h a p t e r 1

values (e.g., resources, education, and occupational status), the defining feature of the middle class in redistributivist theories is individual income. From the perspective of these theories, growth of the middle class is favorable to attaining and maintaining democracy, because it lessens the redistributive consequences of democratization (Boix 2003; Acemoglu & Robinson 2006). Indeed, the middle class is pivotal: whether it sides with the rich or the poor determines the fate of democracy. As long as the middle class is small relative to the mass of poor below it, fear of losing resources to the poor will keep the middle classes from favoring political liberalization. Thus until the middle class is large, these theories expect its representatives to act like elites and oppose democratization. Although redistributivist theories are concerned primarily with macrolevel relationships, they are built on microfoundations that are open to critique. First, the underlying individual-level mechanics of these theories follow from the Meltzer & Richard (1981) model, and assume that one’s place in the income distribution drives political preferences. Whether a person is rich or poor relative to others in society determines his/her expected benefit from democracy. Classic redistributivist theories assume that democracy makes society more equal, meaning that the poorer one is relative to others the more one stands to gain by democratization. Overlooked by these theories is whether an individual’s income and expected future benefits are tied to the state and thus likely to be diminished or disrupted by a transition to democracy. In the post-communist countries, democracy did not in fact make society more equal. Democracy made society less equal. In response, some scholars have proposed that rising inequality with the victory of democratic forces leads the rich to be more supportive of democratization than the poor (Karakoc 2018). Yet all approaches to democratization that focus on inequality are limited in their ability to explain variation in democratic preferences among individuals within the same society who share similar levels of income.13 This is true of arguments that anticipate rising and declining inequality following democratization. These theories are thus silent on what I argue is a key source of intra-class conflict: whether individuals associate their life chances with benefits and rents that are tied to autocracy. Do autocratic state institutions provide the economic opportunities by which individuals 13. More specifically, the only mechanism for doing so in redistributive theories is asset mobility, an intervening variable that helps to predict how successful elite owners of capital are likely to be in avoiding redistribution of their assets to the poor.

t h e au t o c r at i c m i d d l e c l a s s

17

secure their middle-class status? Is downward mobility out of the middle class or loss of rents a feared consequence of democratization? These theories make no allowance for statist strategies of development that use a large public sector to generate particularistic rents for the regime and its allies. Further, redistributive theories’ prediction that the poor will always prefer democracy produces certain empirical anomalies. Again, based on MeltzerRichard these theories predict that, all else equal, income will be inversely related to democracy support—since the more one has, the more one has to lose from the redistribution of assets following democratization. Although my focus is on developing states, where unskilled and semi-skilled workers still outnumber professionals and semi-professionals with university degrees, this framework does not provide a compelling alternative to my argument. In particular, it is unclear from these theories why the private-sector middle class would be more likely than the working class to favor democracy, especially where the middle class still constitutes a minority.14 Redistributive theories also fail to predict that the demand for democracy will be similar among state-dependent middle-class groups (with above-median income) and the working class. In short, my argument, that it is not principally the relative income of the middle class, but its source that determines the demand for democracy, better accords with the evidence in subsequent chapters. It follows from my framework that economic development does not encourage democratization by reducing inequality and the redistributive fears of the rich. Nor does development matter for democratization principally by increasing educational endowments. In fact, I propose that not all forms of economic development actually increase democracy’s prospects (see also Tang & Woods 2014). Instead, my argument implies that development matters insofar as it reduces the economic role of the state and increases the availability of alternative sources of livelihood beyond the autocrat’s reach.

Other Approaches and New Directions This study also builds on and contributes to a newer literature on co-optation in autocracies. In studying the co-optation of elite actors, this literature has focused primarily on the role of formal political institutions—regime parties, 14. See van de Walle (2014) for a comparison of the size of contemporary Africa’s middle classes and that of Europe’s in the age of democratization. Defined in terms of income and consumption, van de Walle argues that Europe’s middle class was probably narrower.

18

chapter 1

legislatures, and elections (Gerschewski 2013; Svolik 2012; Brownlee 2007; Gandhi & Przeworski 2007; Magaloni 2006). Co-optation of the mass public has been examined principally through the lens of electoral patronage— targeted benefits to individual, usually poor, voters by the ruling party and their brokers in the context of elections. My approach differs from these other studies in terms of both actors and settings. First, I focus on potential opposition from a different set of societal actors: the middle class, a group that has often been neglected in accounts of authoritarian resilience. Second, my account points beyond representative political institutions to the importance of public sector enterprises and organizations for co-opting potential opposition. State employment remains a central institution for ensuring authoritarian stability and a way of life for a substantial share of the middle class across many nondemocratic settings. In addition, this book speaks to recent theoretical perspectives on democratization that emphasize the avenues of social mobility available to rising middle classes and intra-class competition. For example, Leventoglu’s (2014) framework predicts that either higher downward mobility or lower upward mobility for the middle class under autocracy will increase its incentives to support democratic transition (835). Ansell & Samuels’s (2014) elitecompetition approach, meanwhile, predicts that democracy will be more likely to emerge when rising middle-class groups find their upward mobility blocked. Rather than see democracy as the product of inter-class conflict, Ansell & Samuels view it as the product of grievances among the relatively well-off but politically disenfranchised, who fear state expropriation (172). Indeed, they find that the preference for state ownership, which they use to proxy fear of expropriation by the state, is negatively associated with democracy support (195). Moreover, in contrast with the predictions of redistributivist theories, Ansell & Samuels show that only among citizens who oppose state ownership does income have any discernible effect (198). These findings are consonant with my own empirical evidence. By examining how socioeconomic status interacts with an individual’s relationship to the state, I show explicitly that “having more to lose” has the greatest effect on those who are formally excluded from state patronage. This book’s findings thus complement elite-competition theory by showing how direct state intervention in the economy leads some rising groups to turn toward the state for economic opportunity and social mobility, rather than away from it fearing expropriation. At the same time, Ansell & Samuels (2014) situate their theory in the context of land-based economies and focus empirically

t h e au t o c r at i c m i d d l e c l a s s 19

on nineteenth-century Europe’s democratization. In this regard, my work is different. In the late-developing countries I study, the economy is not agrarian but industrial; land reform and the rise of independent freeholders are not the key variables they were in earlier waves of democratization. Unlike nineteenth-century England, the state’s extensive role in the modern industrial and financial sectors supports and promotes a middle class that expects better economic conditions under autocracy. Again, this story is about a middle class that thinks and acts differently than the middle class of canonical theories. Finally, my approach adds to a line of work which shows that classes are often divided and questions the notion that shared material circumstances are sufficient to produce common political preferences (Angrist 2005; Tsai 2005; Yang 2007; Slater 2009). As Tsai (2005) points out, an authoritarian state’s growing middle classes may not share the same social identity or interests. Given the diverse occupational backgrounds of China’s capitalists and the absence of common grievances, Tsai concludes that class formation has not yet occurred and therefore that the mechanisms of middle-class influence on democratization are absent. The importance of diverse occupational backgrounds is also echoed in recent work on the postindustrial democracies by Kitschelt & Rehm (2014). Writing on occupations as sites of political preference formation, Kitschelt & Rehm suggest that as the number of professionals and semi-professionals has expanded, divisions among these groups have become more politically consequential. The implication, these scholars conclude, is that “the conventional notion of the middle class” is inadequate “for many theoretical problems in political science” (1691). This study concurs with that perspective. Across the post-Soviet states, my findings demonstrate that divisions within the middle class are already perceptible—and politically consequential. Further, I show that these cleavages are the result of extensive state economic engagement and the distinctive variety of middle-class formation that statist economies engender. The evidence in this book thus suggests that rather than treat the middle class as a coherent actor, we urgently need to distinguish between middle-class groups that are and are not state dependent. Finally, as the preceding discussion makes clear, the literature on democratization has focused overwhelmingly on macro-level variation. To date, that is, most studies have investigated the relationship between a country’s characteristics (such as inequality, per capita GDP, etc.) and its political system, whether democratic or authoritarian. This theoretical emphasis has

20

MACRO

chapter 1

MICRO

MACRO

Individual preferences

Regime change

Value change

Regime change

Redistributive theories

?

Level of inequality

MY ARGUMENT

Modernization theories Economic development

? MY ARGUMENT

figure 1.2. Causal logic of two theories of regime change. The figure shows where this book’s argument lies in the causal chain described by redistributive and values-based modernization theories.

come at the expense of understanding how such macro-level variation affects micro-level processes or how micro-level processes in turn affect macro-level outcomes. In other words, what has been missing from the discussion is clear and empirically verified mechanisms. Ideally, we would have theories at each step of this causal chain. The contribution of the present study is in theorizing and empirically testing an argument that bridges macro and micro. This argument specifies why and how economic and political structures that create an autocratic middle class affect individual attitudes and behavior. Only at the individual level can I test the expectation that the middle classes’ relationship to the state explains important variation in preferences over democracy. At the same time, individual-level survey data allow me to address crucial questions of selection bias: that is, whether state employment has any meaningful effect on preferences or, alternately, whether the types of middle-class individuals who choose state employment merely differ from those who do not. This subject has been largely ignored in existing work on the middle class. Figure 1.2 shows where this argument lies in the causal chain described by redistributive and values-based modernization theories. The observable implications of my theory for both attitudes and behavior are distinct from those of canonical approaches. In both cases, my theory of the autocratic middle class helps to explain why the expected macro-micro link in the logic of existing theories is sometimes broken. I next expand on a central theme of this study: that there are a variety of ways in which middle-class formation can take place, not all of which create a constituency for democracy. Indeed, the ex-communist cases that this book

t h e au t o c r at i c m i d d l e c l a s s

21

examines do not conform to the expectation that as development proceeds, the state sector will invariably contract. The discussion moves from historical cases of state-led late development to the large, patronage-based public sectors typical of today’s developing economies.

Economic Statism and Democratization The argument in this book revives and extends an old insight from Moore’s (1966) Social Origins of Dictatorship and Democracy. My point of departure is Moore’s analysis of France. In the French case, Moore writes that modernization took place through the crown, resulting in the “partial feudalization” of the bourgeoisie and delayed emergence of full democracy (109). “Without the Revolution,” he concludes, “this fusion of nobility and bourgeoisie might have continued and carried France forward into a form of conservative modernization from above” (109). It is this insight, I propose, rather than Moore’s more widely remembered contention, “No bourgeoisie. No democracy,” that serves as a sound basis for a contemporary theory about statist economies and an autocratic middle class. Similarly, in contemporary autocracies, when modernization takes place through the state, the interests of elites and rising middle classes in autocracy are often fused, perpetuating conservative modernization from above and delaying democratization. Dahrendorf ’s (1967) seminal work on Germany makes a similar point. In Society and Democracy in Germany, Dahrendorf traces Germany’s failure to embrace liberal democracy until after the Second World War to the process of state-led industrialization that took place under Bismarck. While England’s industrialization favored smaller units with lower capitalization, Germany’s turned decisively toward large industrial units backed by powerful state banks (35). Quoting Veblen, Dahrendorf writes that “in Germany industrialization has formed a fruitful misalliance ... with the ‘dynastic state’ and has therefore taken a course entirely different from that of England” (33). He goes on to observe: “At what should, in terms of the English model, have been the heyday of private enterprise and liberal social and political patterns, the state was, in Germany, the largest single entrepreneur” (37). The result was a middle class that remained relatively small and dependent and, moreover, lacking in its own political aspirations (44). In sum, these older macro-historical analyses recognized that, under autocracy, statist modes of development have historically favored the co-optation of rising middle-class groups, with implications for democratization.

22

chapter 1

While similar in spirit to the argument advanced here, this earlier macrosociological literature could only speculate about how lack of autonomy from the state affected actors’ political behavior and preferences. The individuallevel empirical evidence that I marshal in this study simply was not present. This study is among the first to employ micro-level evidence to test key implications of these seminal studies using a strong quantitative design. Moreover, the historical literature speaks only indirectly to contemporary conditions of middle-class growth and democratization. Again, these conditions differ from those of nineteenth-century Europe in important ways. The size of the public sector boomed in response to the exigencies of late industrialization in the twentieth century. Statist models of development held much greater sway by the post-war period than they had in prior eras of democratization. My research extends and validates in the context of contemporary state capitalism the insights of this earlier literature on how certain modernization trajectories stunt the formation of a democratizing middle class. By contributing a very different type of evidence, this book provides a strong complement to these existing studies. Meanwhile, at least two of comparative politics’s most canonical theories of democratization largely ignore the economic role of the state. Modernization theory, and its values-based variants, expect autonomous thinking and a desire for greater liberty to grow with the middle class’s material, intellectual, and social resources. Redistributive theories focus on the distribution of wealth in society and individual income, disregarding its source for all but capital-owning elites. These theories must be contorted a great deal to accommodate the systematic differences in political preferences that I show arise from working for the state. Further, more recent literature on authoritarian resilience has focused heavily on institutions, but has privileged the study of political over economic institutions. Institutions help autocratic systems survive by constraining behavior. Yet the incentives and constraints generated by large public-sector institutions have been underappreciated in recent accounts of autocratic persistence. This leaves a significant gap in our understanding of contemporary processes of democratization and authoritarian resilience, since in many latedeveloping countries, economic modernization is accompanied by, even led by direct state intervention. Beginning with Gerschenkron’s (1962) seminal study, a large literature has described the state’s pivotal role in late

t h e au t o c r at i c m i d d l e c l a s s

23

development (e.g., Bellin 2000; Bellin 2002; Kohli 2004). Over the past two centuries, the economic role of the state has steadily expanded—a process that Gerschenkron argues accelerated in response to the demand for rapid, capital-intensive development in the modern global economy. The result is that in many late-developing countries, state-owned enterprises dominate in key economic sectors, and expanding economic opportunities, including rising wages, in the state sector drive the growth of the middle class. This study builds on an older literature on the political economy of authoritarianism to shed light on how statist economies can benefit autocrats and impede democratization. Work by Kornai (1992) powerfully illustrates how public-sector institutions and enterprises fulfilled key political as well as economic functions under communism. With the explicitly ideological basis of these institutions now gone, work on their political function is urgently in need of updating. Of more recent literature, Greene (2010) demonstrates cross-nationally that the size of the public economy strengthens dominant party regimes. His study of Mexico further illustrates an important mechanism linking a large public sector to stable autocracy: the ease with which state employees are recruited as electoral mobilization agents. This is a role the state middle classes also frequently play in the post-Soviet region. Indeed, as Frye, Reuter & Szakonyi (2014) show, large enterprises and institutions that depend on state support provide economies of scale in rallying workers to subvert democracy. What is more, a growing literature on state capitalism recognizes that privatizations of the past two decades reduced state control less than previously imagined (Musacchio & Lazzarini 2012; Musacchio & Lazzarini 2014; Naughton & Tsai 2015; Kurlantzick 2015). Alive and well today, state-led development privileges large state banks and public investments in transportation and communication. It often relies on monopolies or near monopolies in production that are directed or controlled by the state. Increasingly extensive public services in health and education typically accompany this economic expansion. These public service sectors and expanding government bureaucracies are the engine of an enlarged middle class in many developing countries. Existing comparative research has, however, mostly overlooked whether the character of the middle class changes fundamentally under conditions of late development and autocratic state economic engagement. And much of the evidence that middle-class formation can take place in a variety of

24

chapter 1

ways, with distinctive political implications, has remained sequestered in the area studies literatures or embedded in case-specific explanations (e.g., Brown & Jones 1995; Jones 1998; Tsai 2005; Yang 2007; Maleva & Ovcharova 2008; Remington 2011; Chen & Lu 2011; Chen 2013; Cheeseman 2015; Gontmakher & Ross 2015; Handley 2015; Resnick 2015). These insights thus remain detached from broader debates about democratization and authoritarian resilience. Handley (2015, 615), for instance, finds “that the processes of class formation and transformation in Africa unfolded in ways that are distinct from the classical models developed to describe and understand events in Western Europe.” In particular, Africa’s middle classes have not historically emerged from the private sector, “but instead out of a relationship, first and foremost with ... the state.” Indeed, since Sklar (1965) at least, scholars of African politics have argued that bloated states are a major impediment to democratization (see also Diamond 1987). Diamond (1992) writes: As a result of this state expansion in the quest for rapid development, control of the state itself has become the principal means of personal accumulation and hence the principal determinant of class formation. ... This distorted relationship between state and society has been one of the most fundamental causes of democratic breakdown in Africa and Asia.... By crowding out economic competition from the private sector, it prevented the emergence of an autonomous, productive (rather than parasitic) bourgeoisie. By subjecting virtually all developmental activity to state mediation and control, it made community as well as individual advancement dependent on control of the state (483–484). Still other scholars have seen similar processes at work in China and Southeast Asia. In his book on China’s growing middle class, Chen (2013) cites the state’s far-reaching impact on social stratification: more than half of the middle class in the three cities he studies are employed in the state administration (59). Chen finds that middle-class Chinese citizens, and particularly state workers, are more supportive of the current regime and less supportive of political rights than other Chinese citizens (77, 101). He explains these attitudes as the product of extensive state influence over the middle class’s career chances. Along similar lines, Jones (1998) notes that Indonesia’s middle class grew with the bureaucracy’s expansion under Suharto (155), while in Malaysia, he observes that government policies created the indigenous Malay

t h e au t o c r at i c m i d d l e c l a s s 25

middle class (154). As in China, Jones concludes that these state-created middle classes were loyal regime clients. A variety of single-country studies provide a rich portrait of the composition of the middle classes at particular times and places. However, they also raise important questions. How has state expansion in the quest for rapid development affected class formation beyond Africa and Asia? Is the dependency culture Jones describes peculiar to Southeast Asia? Does the Chinese party-state’s paternalism make that case unique? Also unclear in virtually all of this work is whether the middle-class professionals who choose state employment are somehow different, or less democratic ex ante. Is it dependence that matters, or is a risk-averse middle class simply wary of both democracy and the vagaries of the private sector? Despite the important contributions of this literature, most work on the middle class, including the studies just mentioned, cannot answer this question. Limitations of the data and research strategy employed prevent them from doing so. The present book takes seriously such alternative explanations. Using panel data with pre-employment measures of democratic attitudes, I can be more certain that growing state dependency does indeed reduce the demand for democracy. Thus, in this book, I integrate the richness of fine-grained case study work, strong research design for identifying causal effects, and rigorous cross-national comparative research. The argument in this book sees the post-Soviet cases as part of a broader pattern, in which encouraging the growth of the middle classes has been a policy priority of some of the world’s largest nondemocracies (Johnson 1985; Sundhaussen 1991; Brown & Jones 1995; Jones 1998; Shin 1999). Along with Indonesia and Malaysia, in the past two decades alone, the governments of both China and Russia have actively pursued policies to increase the size of the middle class (Chen 2002; Remington 2011). Through the expansion of their state-dependent middle classes, these regimes aim to weaken potential middle-class coalitions in support of democratization. Indeed, recent cross-national studies suggest that such strategies pay dividends. Tang & Woods (2014), for example, find that state economic engagement, measured as the share of government-run businesses and public investment, mitigates the positive effect of development on the probability of democratic transition. In fact, these authors show that economic development benefits democracy only when the state’s level of economic engagement is low. This is an intriguing finding, which may help to account for development’s diminished effect on democratization after World War II, the era of the

26

chapter 1

expanding public sector (Przeworski et al. 2000; Boix 2003; Boix 2011). The argument in this book provides one plausible microfoundational explanation: state economic engagement weakens the incentives of key societal actors to prefer and pursue democratization. Consistent with that proposition, this study demonstrates that dependence on the state for economic opportunity attenuates individuals’ support for democracy and weakens pro-democracy coalitions. Finally, the arguments in this book speak to theories that link privatization and democratization. These theories variously posit that privatization provides a financial base for opposition, enhances the class of small property holders, and has a psychologically liberating effect.15 In the Eastern European context, Róna-Tas (1997) argues that universal state employment was the central institution of social order under communism, and that once it gave way to the emergence of the private market, communism’s demise was inevitable (5). Indeed, the first regimes to collapse, Poland and Hungary, were the first to develop a limited private sector and to give members of the nomenklatura the opportunity to participate in legal private enterprise (181). This study shows concretely how privatization that creates alternatives for the public sector middle class can enhance the prospects for democracy by changing the incentives of a key group in the autocrat’s coalition. Beyond the post-communist region, Greene (2010) similarly links the decline of dominant parties and democratic transition to the privatization of state-owned enterprises (see also Greene 2007). Greene finds that in Mexico under the PRI and cross-nationally, autocrats “virtually transform public agencies into campaign headquarters,” using “public employees themselves to inform and mobilize voters” and doling out employment for political gain (812). With privatization, authoritarian incumbents lose these sources of leverage. This study finds micro-level evidence that state patronage effectively shapes the regime preferences of public employees and shows that public-sector institutions influence the political economy of a wide variety of nondemocratic, and not only dominant-party, regimes. As Fish & Choudhry (2007) note, one factor that has limited our understanding of the impact of privatization on political democracy is the absence of research about a key intervening variable, the political preferences of the

15. See Fish & Choudhry (2007) for a review of these theories.

t h e au t o c r at i c m i d d l e c l a s s 27

middle classes (272). Chapter 3 of this study answers the call for clear crossnational measures of these preferences. Its findings have several implications for our understanding of the political consequences of economic restructuring and privatization. First, the micro-level evidence in each of the subsequent empirical chapters implies that statist models of development attenuate the positive role that social mobility can play in broadening democratic coalitions. Second, it follows that privatization that reduces state employment is likely to improve long-run prospects for democratization, all else equal. However, this book’s findings also offer a note of caution: privatization policies will be least effective at aiding democratization if they create downward mobility for the public-sector middle class. After all, as chapter 3 clearly shows, reducing the public-sector middle class’s employment prospects to routine and manual labor yields no clear benefits for democracy. I find that whether employed in the public or private sector, the working classes share the state middle class’s weak support for democracy. In sum, contrary to the conventional view that growing middle classes pose a threat to authoritarian stability, I build the case that a middle class dependent on the state may in fact be a source of autocratic resilience. This contingency arises from both structural and strategic factors. Virtually all authoritarian systems target some groups over others for advancement. This book takes the view that nondemocratic states pursue policies that expand middle-class employment in state-dominated sectors as a means of ensuring loyalty and regime survival. As high levels of human capital have become increasingly important to sustain economic growth, developing nondemocracies have sought to increase the size of the middle class, while remaining conscious of the potential political threat of independent economic activity. As a result, under autocracy, the state often plays a direct role in forging and reinforcing the middle class, with state and quasi-state-owned enterprises and an expansive public sector providing a key avenue of upward mobility. My framework suggests that democratization is likely to occur as economic co-optation becomes more costly, in line with Magaloni’s (2006, 20) work on hegemonic parties. However, I depart from Magaloni (2006) in proposing that these costs depend not only or even primarily on a country’s level of economic development. Rather, the argument I advance in this book implies that economic diversification, when it reduces the state’s role in the economy, is a much more important factor.

28

chapter 1

Overview of the Empirical Strategy To date, no systematic cross-national research has sought to explain the conditions under which the middle class contributes to authoritarian resilience. In order to address this gap, I draw on new opportunities to study public opinion and political participation under various forms of nondemocracy, from hegemonic single party to relatively pluralistic competitive authoritarian. The past two decades have witnessed a dramatic expansion of social science research investigating attitudes toward democracy, including in places where reliable public opinion data used not to exist (Diamond & Plattner 2008, ix). The empirical work in this study focuses on Eastern Europe and the former Soviet Union, where the status of democracy and efforts to consolidate it remain uncertain. No other world region offers such rich sources of micro-level evidence over a period marked by multiple mass mobilizations in the name of democracy. The region is also particularly well suited for studying the middle class. By the World Bank’s calculations, the percentage of the population that joined the middle class from approximately 2000 to 2010 in Eastern Europe and Central Asia was the highest in the world.16 Theories that emphasize the political consequences of a growing middle class would thus expect to find an increasingly confident and assertive democratic constituency. Yet empirically the record is mixed. Countries like Russia, where the middle class has grown fastest, have failed to consolidate democracy and even slipped back toward greater authoritarianism.

Research Design and Methods The empirical chapters of this book are designed to test the main hypotheses derived from my theory. To provide the strongest possible test of my argument, I endeavor to present convincing evidence of a causal relationship, increase confidence in the explanation by differentiating among mechanisms, and demonstrate similar effects across different contexts and samples. The first two points establish internal validity while the third reflects on external 16. Calculations by Augusto de la Torre and Jamele Rigolini on the basis of the World Bank’s PovCal and household surveys. The sample is limited to low- and middle-income countries. Results presented at the MIC Forum: The Rise of the Middle Class. Accessed June 23, 2015, www.worldbank.org/content/dam/Worldbank/document/MIC-Forum-Rise -of-the-Middle-Class-SM13.pdf.

t h e au t o c r at i c m i d d l e c l a s s

29

validity. In keeping with these principles, this book’s research design consists of several interrelated types of analyses. I first probed whether the argument’s predictions held across a variety of settings and different national samples. To show that the book’s central argument explains patterns of democracy support across the former Soviet states, I carefully analyzed the broadest available survey of post-communist public opinion with detailed data on regime preferences and labor market position. Together with additional evidence from different survey samples in subsequent chapters, this large cross-national sample gives us greater confidence in the generality of the arguments illustrated by the analysis. I then turned to the question of whether the state dependency of the middle classes shapes only attitudes or also action. With the aim of reflecting on whether such dependence is a shared driver of attitudes and several different types of politically consequential behavior, I analyze post-election protest, participation in a civic revolution, and voting in studies of Russia and Ukraine. Subsequent chapters thus attempt to answer the question of whether state employees merely appear to support autocracy, but would be equally likely to take part in a democratic uprising. I also take up the challenge in subsequent chapters of exploring who chooses a career in the state sector. Answering this question is crucial for understanding whether state careers have an independent impact on the political preferences of the middle class, or merely attract more authoritarian personalities. To convincingly answer this question, I fielded an original survey of students who had not yet entered the labor market. I also employ a strong research design to provide evidence of causal relationships that are difficult to establish in cross-sectional data. To implement this design, I identified a panel survey, consisting of reinterviews with the same respondents, that measured new labor market entrants’ regime preferences both before they entered the workforce and after. The findings of this analysis and that of the original survey provide additional evidence that career experiences matter independently of the factors that affect career choices, with state dependency reducing middle-class demands for democracy. Finally, much of the book deals with the impact of public-sector policies on the political attitudes and behavior of the middle class without addressing the regime intentions. In the book’s final chapter, however, I consider the creation of an autocratic middle class as an explicit regime strategy. Before concluding, I also probe the importance of newer forms of state economic engagement and indirect influence that cross-national data do not easily capture.

30 c h a p t e r 1

I test my theory using both quantitative analysis of surveys and qualitative evidence gathered in the field. To complement cross-national survey data, I collected a variety of additional empirical evidence, facilitated by close collaboration with local scholars, during nine months of research in Russia, Ukraine, and Kazakhstan. This material includes approximately fifty expert interviews with academics and government officials (the vast majority of which I conducted in Russian), survey datasets that have yet to appear in English-language scholarship, library and unpublished manuscript sources, as well as statistical and archival data on the structure of each country’s labor market. Among the datasets I either collected or obtained though my fieldwork are: two massive nationally and regionally representative surveys in Russia (N = 34,000), specialized surveys of Russia’s urban middle class, a highly unusual series of four protest surveys from the mobilizational cycle following Russia’s 2011 parliamentary elections, an original survey of Russian university students, a unique three-wave panel survey (with reinterviews of the same respondents) spanning the Orange Revolution in Ukraine, and a brief original survey of senior civil servants in Kazakhstan. In short, this study is informed by a wide array of both quantitative and qualitative data from the field that has not previously been brought to bear on our understanding of the interests and regime preferences of the middle classes. Four advantages of my mixed methods research design merit mention. First, one concern with existing case studies of the middle class is their potentially limited external validity. As discussed above, including an analysis of multicountry survey data helps to address this issue of generalizability. My research design further advances existing case studies of the middle class in at least two other ways. One, I do not limit my scope to the middle class, as do, for example, studies of business owners, urban professionals, and the self-identified middle class. Focusing only on the middle class prevents such research from reliably concluding that middle-class attitudes are distinctive or determining why (because they lack the necessary baseline). Two, I do not limit my sample to a single regime type. By focusing exclusively on either democracies or nondemocracies, previous studies cannot reach reliable conclusions about the interaction of regime type and public-sector institutions on middle-class attitudes toward democracy. The second important methodological feature of this study is that it combines evidence on political preferences with evidence on actual political behavior. At the individual level, most studies are limited to data on attitudes or reported political behavior. This is a significant limitation since activities

t h e au t o c r at i c m i d d l e c l a s s

31

like protest are more proximate to regime change and one of the ways that even a numerically small middle class can influence democratization. To address this gap, I draw on an unusual series of protest surveys conducted in Russia, showing that my theory also has implications for politically consequential real-world behavior. I employ a novel research strategy—a variant of the case-control method of causal inference used widely in epidemiology— for the first time applying it to the study of protest participation. This design helps to address the challenges of collecting and analyzing individual-level protest data, and represents a significant improvement over studies that rely on respondents’ own retrospective reports about their protest attendance. No existing study of the middle class presents opinion data alongside data on party support, vote intention, and actual protest participation. Throughout the book, I carefully address alternative causal logics. This point brings me to a third advantage of the research design: In addition to various forms of cross-sectional evidence, I also test my argument dynamically using panel data, a particularly robust source of evidence that is very rarely available outside of consolidated democracies. Because panel data— consisting of repeated interviews with the same respondents over several years—have an explicitly time-ordered dimension, they provide additional leverage in identifying causal relationships (Miller 1999; Finkel 1995). Unlike other observational data, panel data offer unique opportunities to eliminate threats to inference arising from endogeneity, like reverse causation, selection based on permanent characteristics, and omitted variable bias. Specifically, panel data from Ukraine provide a “pretreatment” measure of regime preferences for those whose labor market position changes between subsequent survey waves. In other words, they allow me to say with confidence that an increase in state dependency negatively affects democracy support among those with previously similar regime preferences. I demonstrate these effects using a design that resembles difference-in-differences, but with the added benefit of following the exact same individuals over multiple survey waves. In addition, panel data allow me to rule out reverse causation by focusing on new labor market entrants with no prior employment history—that is, young people who enter the labor market during the period under study. The results of that analysis once more confirm that dependence on the state affects support for democracy, and not the other way around. Yet even when limited to cross-sectional data, I work hard to improve identification by exploiting natural variation. Thus, in the cross-national analysis

32 c h a p t e r 1

that follows, I take advantage of variation across different occupations within the state sector to show that the effects of state employment hold not only among those most likely to be influenced in their choice of career by political considerations: professionals employed in the state administration. I also exploit variation in career trajectories to show that democrats are no more likely to leave state employment sooner (i.e., differential attrition). Each of these analyses helps to build the case, within the natural limitations of observational data, that selection, or the self-sorting of individuals into sectors and careers, is not a plausible alternative to my argument. The fourth important feature of this book’s research design is that it speaks to the self-conscious and strategic ways that nondemocratic regimes deploy the concept of middle class. In the final empirical chapter on Kazakhstan, I trace the emergence of this concept in official rhetoric and policy through field interviews, the observations of senior civil servants, primary source documents, and in-country library research. I also exploit the fact that Kazakhstan’s state involvement is particularly extensive in certain industries. This involvement occurs both directly, through outright ownership, and indirectly, through minority stakes and other investments. Focusing on industries where state involvement is extensive but indirect allows me to probe whether the logic of the argument extends beyond the formal state sector.

Plan of the Book The remainder of the book is organized as follows. Chapter 2 expands on the argument I have laid out in this chapter. It outlines my explanation for variation in middle-class regime preferences and details the individual-level logic of state dependency. It also discusses the book’s key concepts in greater depth, alongside descriptive data on the middle classes and state economic engagement in the countries under study. Chapter 3 begins the empirical investigation of the book’s central premise. I construct individual-level measures of class, career trajectories, and democracy support, using a large cross-national survey dataset. Using detailed evidence on individual employment histories from twentyseven post-communist countries selected for their common history of state employment and communism, I show that, under autocracy, state-sector careers weaken support for democratic institutions, especially among the middle class. In the democratic post-communist countries with less statedependent economies, I do not expect state employment to weaken support

t h e au t o c r at i c m i d d l e c l a s s

33

for democracy, and indeed, no such relationship is evident. Broadening the sample to include Central and Eastern Europe also helps to clarify why prior research, which has overwhelmingly focused on democracies, has failed to find the contingent relationship between middle-class growth and democratic preference formation that this book highlights. By contrast, a sample that includes multiple regime types reveals that, under authoritarianism, the state has a distinctive impact on the middle classes’ political orientations. By exploiting natural variation across state-sector professions and over time, as individuals move into and out of the public sector, I demonstrate that state employment has meaningful, independent effects on regime preferences. I also consider but find lacking several alternative explanations, including communist-era socialization, generational effects, social desirability bias in survey responses, and self-sorting into career paths based on prior characteristics (i.e., selection). These results are robust to changes in the measurement of middle class, a broadening of the definition of democracy support, and a reassignment of countries with unstable regimes. The findings imply that authoritarian states benefit by creating economic opportunities and employment security in the state sector, which limits middle-class demands for democracy. The chapter provides the first rigorous crossnational examination of the regime preferences of the post-communist middle classes. Chapter 4 raises micro-foundational questions about the expectation that a rising middle class will lead a democratic civic revolution. In this chapter, I shift focus from political attitudes to actual political behavior and test my argument against observed patterns of mobilized contention during Russia’s 2011–2012 electoral cycle. I do so by nesting a unique series of protest surveys within detailed data on the population from which protesters were recruited. By combining these two sources of data and using statistical methods inspired by work on rare disease to study protest participation (an outcome that is also rare), I show that how one enters the middle class and what alternatives one possesses affect participation in risky collective action. Amid heightened middle-class participation in anti-regime protests, I find that professionals in the state sector were significantly less likely to mobilize against electoral fraud, even after controlling for ideology. If this group had participated at the same rate as middle-class professionals from the private sector, I estimate that up to ninety thousand additional protesters would have taken to the streets. While I find that middle-class protesters from the private sector were much more likely than the working class to join the protests’

34 c h a p t e r 1

democratic coalition, middle-class protesters from the public sector were not. I trace these patterns of participation to lower support for democracy among state-sector professionals and career incentives. I also show how the power of these incentives varies with available alternatives to state employment. Finally, I contrast my explanation with its leading alternative, the self-selection of people who are less democratic into state employment, and show that my framework better accounts for variation in the propensity to join opposition protests. The chapter’s findings highlight how state-dependent economies undermine the formation of broad democratic coalitions of the middle class. I next take a step back to further probe the question of selection raised in preceding chapters: namely, in a country such as Russia, what kinds of people select state-sector careers in the first place? Are less democratic types more likely to gravitate toward employment in the public sector? Is the publicsector middle class systematically different because it attracts individuals with systematically different characteristics? Chapter 5 uses original survey data to investigate the political orientations and career aspirations of students who intend to join Russia’s public sector. This analysis uses the same measures of democracy support as in chapter 3, but in contrast to preceding analyses captures these attitudes prior to state employment. While I examine a range of factors highlighted by existing cross-national research on public employment in established democracies and developing states, my principal focus is on whether the political views of people who will become public-sector workers differ from others. Rather than make career choices on the basis of political orientations, I find that the single most important factor shaping public-sector career preferences is preexisting networks—having a parent who works in the public sector or attending a university with a strong alumni network in state institutions. The availability of these informal networks, rather than views on the importance of political freedom, order, national security, or strong economic performance, shapes who is most likely to join Russia’s public sector. By focusing on students’ political preferences before they enter the labor market, this chapter’s research design helps to show that democratic attitudes have virtually no bearing on career choices, suggesting that differences arise only later. I build on this research design, using a panel study of new labor market entrants, in the subsequent chapter. Chapter 6 provides a dynamic test of my theory over a period when Ukraine was struggling to consolidate democracy. Though existing theories expect human capital formation and a growing middle class to enhance

t h e au t o c r at i c m i d d l e c l a s s

35

prospects for democratization, a paucity of longitudinal evidence from countries in political transition has impeded research on these questions. Focusing on the case of Ukraine, I examine whether dependence on the state for economic opportunities and life chances moderates middle-class demands for democratic institutions, using a panel survey spanning the Orange Revolution. I go on to assess how the distinctive political orientations of different groups within the middle class affected the nature of protest coalitions during Ukraine’s 2004 democratic breakthrough. I then show that these same factors subsequently shaped how popular support for the Orange forces unraveled. More specifically, I use a difference-in-differences design to demonstrate that reliance on the state for economic opportunity caused the political preferences of otherwise-alike new labor market entrants to diverge. This analysis shows with striking clarity that state dependency significantly reduced support for democracy at a time when Ukraine’s political future was contested. I then shift from political attitudes to political actions and broaden the analysis to include all Ukrainians. Similar results for protest and voting behavior are found in the second part of the chapter. Aided by reinterviews with the same respondents, the analysis in this chapter helps to clarify the causal effect of state dependence on democracy support, a relationship that is difficult to assess with cross-sectional data. Further, the findings in this chapter help to illuminate the micro-level mechanisms by which public-sector retrenchment and privatization might improve prospects for democratization. Chapter 7 studies the vocabulary and practice of state-led middle-class growth in Kazakhstan, a Central Asian country whose post-independence development has been shaped by both Soviet legacies and the experience of its East Asian neighbors. This case thus suggests how the argument might travel beyond the post-communist region and speak to newer modes of state capitalism. In it, I use field interviews, an original survey of senior civil servants and other public opinion data, archival material, and statistical sources to illustrate the book’s argument. The chapter begins by examining the discursive context in which fostering middle-class growth has become a priority of post-Soviet states. Though symbolically important today, the middle class was conspicuously absent from the official class constellation of the communist era. After the 1917 October Revolution, there were peasant and proletariat classes, plus the intelligentsia, but formally no middle class. The discussion demonstrates that Kazakhstan’s regime has self-consciously adopted a view of the symbolic and strategic importance of the middle class as a guarantor of stability and

36 c h a p t e r 1

source of societal mobilization. I next turn to the size and composition of the middle class that has begun to emerge in post-independence Kazakhstan. I show that lack of autonomy from the state is associated with less middleclass support for democracy and political competition, consistent with the book’s central argument. In fact, my analysis suggests that there is virtually no support for democracy among Kazakhstan’s state middle class. Support for democracy is also lower in parts of the private sector where state engagement is extensive (i.e., in the natural resource and defense sectors). The chapter ends by showing how the state deploys its middle-class clients to mobilize society in support of autocratic stability. The final chapter concludes with implications of the evidence presented for the literatures on development and democratization, resource states, and protest mobilization. It demonstrates that the logic of middle-class state dependency holds not only in states where the economy is based on extractive industries. It also places the book’s argument in broader comparative perspective. Paying greater attention to the autocratic middle class supported by state economic engagement helps make sense of delayed democratization and democratic failures in the Middle East, Africa, and Latin America. Development, which increases the size of the middle class, need not lead to democratization and may even stabilize authoritarianism. Much depends on the state-supported middle class as an important swing group in the autocrat’s coalition.

2 State Dependency and Middle-Class Demand for Democracy

the dominant image of modernization in developing nondemocracies is of social transformation that undermines regime stability. Modernization unleashes new social forces. It forces autocrats to adapt, to find new ways to neutralize development’s beneficiaries, or else risk regime collapse. This, however, is only a partial account of modernization. Modernization can occur in both top-down and bottom-up ways. When top-down, modernization’s principal beneficiaries are often within the state itself. By concentrating rents in the state sector, such modernization from above can be a source of regime resilience rather than a threat to autocratic rule. Consider the case of Russia. Russia’s modernization under Vladimir Putin has been a key source of his regime’s resilience. Why? The Russian state plays a larger role in the economy today than it did when Putin took power (e.g., Kudrin & Gurvich 2015). The state’s renewed economic engagement has helped the Kremlin to shore up a loyal base of support. Between 2005 and 2015, during a period of fast-paced growth, the Russian State Statistics Service reported that the number employed in public administration, defense, and social security grew in both absolute and relative terms (Rosstat 2017, 93). While the number of public-sector employees in education declined somewhat over this period, the size of the public-sector payroll in health and social services rose. Employment also increased in the largest state-owned enterprises (Abramov, Radygin & Chernova 2017), bringing the state’s share of formal-sector employment to 50%. While estimates vary, a widely cited International Monetary Fund (IMF) paper estimated the state’s share of the Russian economy at 70 percent

37

38 c h a p t e r 2

in 2014.1 Other, more conservative estimates place the figure around 50 percent.2 What is clear, however, is that the number of enterprises and organizations owned by the state rose each time Putin was reelected, even as the overall trend has been toward fewer, more consolidated state entities (Rosstat 2016, 193). Moreover, these figures do not take account of enterprises with mixed ownership types, which allow the Kremlin to exert influence without direct state control. By all accounts, these forms of ownership are on the rise. While calculating the size of indirect state participation is not straightforward, one study finds that, in 2014, state-owned enterprises and organizations with mixed ownership employed approximately one-third of all individuals with jobs across the Russian economy (Abramov, Radygin & Chernova 2017). Even as world oil prices plummeted and the West stepped up sanctions against the Russian regime for its annexation of Crimea in late 2014, Putin preserved employment in the state sector. Between 2014 and 2015, the share of total employment in the state sector actually grew slightly, as did the absolute size of public sector employment (Rosstat 2016, 92). Such trends say something about how the regime survives even in an inauspicious climate of austerity. To protect his base, Putin has worked to preserve, and when possible strengthen, his alliance with the state sector, and the middle-class state salariat in particular. In chapter 1, I discussed existing theories that expect the middle class with its high educational and professional qualifications to fare better under democracy, all else equal. However, all else often is not equal and context matters (Franzese 2007, 55). I account for this variation explicitly by directly considering the role of the state. In this book, I argue that people with high educational and professional qualifications often support authoritarianism, because authoritarian regimes play an important role in providing resources (like jobs, salaries, benefits, and opportunities for rent-seeking) to these same groups. Authoritarian regimes monopolize opportunities to access resources and reward people who act as their brokers. Patronage, corruption, and other phenomena typically tied to partisanship function much more broadly 1. See the IMF’s May 2014 report, “Russian Federation: Fiscal Transparency Evaluation.” https://www.imf.org/external/pubs/ft/scr/2014/cr14134.pdf. 2. See, e.g., the Center for Strategic Research, January 2018 report, “Effektivnoe upravlenie: gosudarstvennoj sobstevennost’iu v 2018-2024 i do 2035.” https://www.csr.ru /wp-content/uploads/2018/02/Doklad_effektivnoe_upravlenie_gossobstvennostyu_Web .pdf. Lower-bound estimates range to about 33%, though note that many in this range include only general government spending and revenue and exclude state-owned enterprises.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 39

than mere partisan tools—they function to promote autocratic regime stability by creating greater uncertainty around the consequences of democratization and stifling dissent. As described in chapter 1, public employment is widespread among well-educated professionals and managers, and more common among the middle than working class, again in contrast to what scholars have typically assumed.3 So the educated middle classes become the linchpin of autocracy instead of its achilles heel. This framework highlights the importance of education and occupation—and more specifically how a person’s occupation defines their place in the political economic system—as opposed to income or wealth in determining regime preferences. I find support for this argument using the most comprehensive data yet assembled on the democratic preferences of the people with high educational and professional qualifications that existing theories expect to champion democracy. In sum, I expect that those whose middle-class status is sustained by the state benefit more from autocracy and will be less likely to support democratization. Dependence hinges on discretionary benefits wielded by the state, on the state’s ability to use selective sanctioning, and on a particular political economy in which the middle classes have few exit options beyond the public sector. I expect that those with greater alternatives to state employment will be less dependent and more democratic than those with fewer alternatives. The story this book tells thus focuses on the public-sector institutions and labor market conditions that inhibit bottom-up demand for democracy. The evidence shows why—especially where state economic engagement is extensive—growth of the middle class in the state sector is frequently not accompanied by democratization. It remains now to more fully lay out why this pattern of preferences arises. This leads me to a discussion of the theoretical mechanisms underpinning the state’s influence over the regime preferences of the middle class and then onto an explanation of the book’s key concepts. These arguments are then tested using a range of empirical strategies in subsequent chapters.

The Political Economy of Authoritarian Rule A central argument of this book is that an autocrat’s control over public-sector employment serves as an important tool for managing the economic selfinterest and loyalty of citizens, and the middle classes in particular. The broad 3. See figure 1.1.

40

chapter 2

control over employment that autocrats across the post-communist region enjoy today is in part a structural inheritance from the communist period, but also an evolving and deliberate survival strategy with novel features. Under state socialism, politics—and, in particular, official ideology—shaped all aspects of the economy. Public sector institutions, state-owned firms, and state cooperatives became more than merely the scene of work (Kornai 1992, 222). They became, in Kornai’s (1992) phrase, “cells of totalitarian power,” fulfilling key social and political, as well as economic, functions. Public-sector entities retained discretion to set wages and bonuses to favor and discriminate on political bases. They were the site of ongoing political agitation. They functioned as ladders of career mobility, including for political positions. They monitored and persecuted would-be troublemakers among their ranks. They controlled the provision of myriad state services, access to scarce goods, and in some places the right to residency. These functions gave public-sector managers enormous power over their employees and employees strong incentives to remain in the good graces of their superiors (Kornai 1992, 220). With these levers, universal state employment ensured the communist regimes a high degree of social control and relative longevity. This book shares with Kornai’s (1992) seminal study of state socialism the view that control of employment is a crucial lever of political control over a population. However, this book inverts the logic of his argument—which ran from politics to economics. Here I am interested in how a large public sector, once established, influences a country’s political institutions and molds the value system and choices of ordinary citizens. The thrust of the argument is that location within the economic structure of state capitalism drives preferences with respect to political outcomes, among no group more forcefully than the middle classes. As in the classical communist system, the regime elicits loyalty through particularistic financial and career incentives, backed by workplace-based mobilization in the context of an ongoing relationship. But unlike classical communism, the official ideology is extinct, private property is permitted, and formal restraints on the mobility of labor have been lifted. The political economy of authoritarianism described by this book does not rest on an official ideology. In its latest incarnation, state capitalism is given a non-ideological justification. For example, in Russia, greater state economic engagement is said to ensure proper financing for high-risk economic transformation. Nor does it require that state employment be universal, only that the state’s engagement in particular sectors trends monopsonistic. Control

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 41

in sectors such as banking, transportation, energy, construction, machinebuilding, chemicals, and defense gives states a dominant stake in so-called strategic industries. While calling for the state’s role in the economy to be reduced, contemporary autocrats strategically maintain more discreet forms of engagement. They sell shares in state corporations but maintain controlling stakes, create new state holding companies, and increase the size of the state administration. State enterprises, in turn, enjoy the benefits of cheap credit from statecontrolled financial institutions and a favorable regulatory environment. These corporations succeed and even thrive because they are sheltered from competitive pressures in industries that the state dominates. Though the model varies across countries and time periods, typical state-dominated sectors include defense, banking, transportation, and extractive industries. This is, of course, in addition to government-provided service sectors like health and education. The health systems of most post-Soviet countries, for example, are almost entirely state dominated. Private practice remains uncommon and hospital privatization has been limited (Rechel, Richardson & McKee 2014, 33). As of 2016, there were no private hospitals in Belarus and, in Russia, there were only 357, or 6.5 percent of the total. The state’s role in education across the region is similarly encompassing. For educators at all levels, an independent livelihood in the private sector remains rare. A growing economy needs professionals and managers with higher educational qualifications. Public entities within the sectors just mentioned absorb this demand, creating a middle class whose formation is state dependent. Thus, in many developing autocracies the state provides a significant share of permanent employment, which provides for steady earnings that are status enhancing. These formal-sector jobs contrast with lower-status and less secure informal employment, abundant in developing countries’ large gray economies. In the post-communist countries, as elsewhere, state employment comes with both formal and informal benefits. The formal benefits of state employment may include a formal labor contract, shorter working hours, paid vacation and medical leave, medical insurance and care at state medical clinics, transportation benefits, housing benefits, and access to credit on preferential terms. Data collected by the European Bank for Reconstruction and Development show that formal employment is much more common within the state sector than outside it: across the nondemocratic post-communist countries, those working in middle-class occupations in the state sector are virtually

42

chapter 2

assured the protections of a contract, whereas in the private sector one in five lack them.4 A separate study by the Russian Academy of Sciences confirms that the guarantee of social benefits under a formal labor contract is much more widespread among professionals in the public than private sector (Russian Academy of Sciences 2014, 50). Sharunina (2013) notes that one in three public-sector employees, compared with only one in ten privatesector employees, works less than forty hours per week. Such benefits help to offset lower formal wages in the state sector. The point is that jobs in the public sector foster political alignment with the regime when government policies provide superior benefits and working conditions. Uncertainty on three fronts—about how the formal benefits of state employment will compare under a democratic regime and whether one’s own position will remain secure after a regime transition, along with the risk of losing one’s job for political dissent—creates reticence about democratization. Public-sector professionals are also ideally situated to extract informal benefits (rents) from their formal positions. Indeed, autocrats often compensate for low formal salaries by permitting middle-class clients to solicit bribes and engage in other corrupt behavior. One study sums up the divisions within Russia’s middle class as between bribe-takers and their victims (Tikhonova 2008, 25). While corruption among state bureaucrats receives the greatest attention, corruption is also common in other government budget sectors, like health and education. A 2010 report by the corruption watchdog INDEM found that the everyday corruption market in state clinics and hospitals exceeded any other at more than 35 billion rubles annually. Bribes in the education sector—when applying for state kindergarten, secondary school, or university; to receive a diploma; or as a “required” form of gratitude—were nearly as common.5 Because these informal benefits of public employment 4. By contrast, in the democracies, informal employment is less common and these discrepancies between state and private are much smaller. 5. The share of citizens forced to give a bribe to receive a free medical service increased from 23.5% to 37.7% between 2000 and 2005 as the Russian economy began to take off. The share who paid a bribe to enter a required secondary school and to finish successfully also rose (from 13.2% to 41%) over the same period, as did the share who paid a bribe to enter an institution of higher education (from 36% to 52.1%) (INDEM 2005). While the everyday corruption market in higher education is the largest within Russia’s educational sector (21 billion rub), corruption in early childhood (14 billion rub) and secondary education (2 billion rub) is also substantial. Individual bribes in education, like those in healthcare, also tend to be larger than many other types of everyday corruption (Ministry of Economic Development of the Russian Federation 2011).

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 43

are tied to political control of the state and threatened by rule of law and political competition, I expect them to affect more than just support for a particular party or candidate, but also an individual’s preference for democratic institutions. Public-sector positions also offer fertile ground for particularistic relationships to develop. This is true whether state employment is initially granted as a political favor by the current regime or is a legacy of a previous economic system. In both cases, employment relations give an autocrat unique leverage over individual incentives. Job loss, promotions, and other benefits can all be made contingent on the right political views. Moreover, as scholars have widely observed, state patronage thrives when “organizational devices and social networks of supervision” provide the basis for repeated interactions and make it easier to monitor defection (Kitschelt & Wilkinson 2007, 8). In public-sector workplaces, schools, and hospitals, these conditions are already present (Hale 2007; Frye, Reuter & Szakonyi 2014). More broadly, it is not unusual across a variety of contexts for employers to act like political machines, effectively carrying out political recruitment on behalf of parties and candidates (e.g., Mares, Muntean & Petrova 2018; HertelFernandez 2017; Mares & Young 2016; Wolfinger 1972; Scott 1969).6 Such workplace-based efforts are frequently assisted by unions, and a number of studies highlight specifically the role of public-sector unions in delivering support for autocratic patrons.7 To summarize, then, employment settings allow an autocrat to efficiently confer privileges in exchange for support, and withdraw benefits from those who organize for democracy or participate in popular insurrections. Repeated interactions and straightforward levers of economic influence ease the recruitment of state-sector professionals— both as supporters and as political brokers, who in turn mobilize others—as suggested by existing studies of clientelism. At the same time, this book’s framework contrasts in at least three ways with standard accounts in the patronage literature. First, patronage is usually conceived in terms of the linkage between an excludable benefit and the success of a particular party or candidate (e.g., Robinson & Verdier 2013). Instead,

6. Interestingly, Mares, Muntean & Petrova (2018) find that such strategies are most effective under conditions of high economic concentration, a perspective that dovetails nicely with this book’s argument. 7. See, for example, Forrat (2018) on Russian teachers or Larreguy, Montiel Olea & Querubin (2017) on the support of Mexico’s largest teacher’s union for the country’s long-time dominant party, the PRI.

44

chapter 2

I consider how an excludable benefit, such as a public-sector job (as well as the privileges and side payments that go with it), may be linked to the continuation not only of a particular party or patron, but also a political regime. In other words, I expect that patronage also has the power to diminish support for democracy by tying future benefits to stable autocracy. Second, the existing literature has overwhelmingly focused on patronage as a tool for winning elections. This study broadens that focus to examine how patronage produces incentives to support not only candidates and parties at election time, but also regime stability at other times. More specifically, I argue that state patronage induces support for stable authoritarianism during periods of protest when the autocrat is most vulnerable. By holding out employment and associated benefits as politically contingent, regimes are able to demobilize key constituencies within the middle class from participating in anti-regime demonstrations and mobilize them in defense of existing autocratic institutions. Chapters 4 and 6 demonstrate this dynamic, using data on mass mobilizations in Russia and Ukraine. Third, this book helps to explain why autocrats would choose to target patronage to the middle classes, rather than specialize in relatively cheaper, low-skill clients. Studies of patronage typically focus on low-wage employment, while the middle classes are thought capable of “ideological investments.” Such an arrangement is assumed to be compatible with the incentives of both clients and patrons. From the client’s perspective, because the poor gain greater marginal utility from employment, scholars have assumed that they will be more likely to enter into clientelistic relationships (Dixit & Londregan 1996; Chubb 1982; Calvo & Murillo 2004; Stokes 2005; Robinson & Verdier 2013). However, in situations where low-skilled jobs are relatively abundant in the private-sector economy and high-skilled jobs are not, this logic begins to fray. Where high-skilled professions are monopolized by the state and private-sector opportunities are lacking, labor market alternatives can actually be worse for the middle class. We observe precisely this pattern in many post-Soviet states. Taking Russia as an example, these factors help to explain why the middle class is more, not less, likely to want to work for the state (Russian Academy of Sciences 2014, 51). Similar patterns hold in other developing economies. Conversely, from the patron’s perspective, scholars have assumed that specializing in low-skill clients nets more votes for the same outlay of resources. So why would autocrats buy off the middle classes, who are expensive, rather than the poor, who are cheaper? In part, this focus on efficiency in the

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 45

allocation of patronage is a consequence of the literature’s focus on elections— and patronage as a tool to win them. Each voter (each potential client) casts a vote, which is equal to and substitutable for any other. This argument, however, neglects the fact that regimes often press middle-class clients to engage in more sophisticated activities that are worth many votes, receiving a higher return on their investment. Throughout the electoral cycle, state-sector professionals can be mobilized to provide politically contingent assistance to voters in accessing a range of services.8 During elections, the state middle class can be mobilized for activities that include campaigning, organizing and attending rallies, and serving on election commissions. In fact, the recruitment of state employees as electoral mobilization agents (or brokers) is quite widespread (e.g., Bratton & van de Walle 1997; GrzymalaBusse 2007; Greene 2007; White 2011; Mares 2015; Larreguy, Montiel Olea & Querubin 2017; Forrat 2018). And, as noted already, the incentives embedded in public sector institutions facilitate these transactions. Further, the incentives facing a regime differ if remaining in power depends not only on winning elections, but also on keeping those who are most likely to mobilize off the streets. Public displays of defiance entail great risk for authoritarian incumbents (Lee & Zhang 2013; Bunce & Wolchik 2011; Robertson 2011; Bayat 2009). By demobilizing a large segment of the urban middle class, patronizing the public-sector middle class can contribute to a regime’s resilience. Chapters 4 and 6 draw on evidence from Russia and Ukraine to show that state dependency not only reduces the middle classes’ support for democracy in the abstract, but also their participation in protests and civic revolution. While Weitz-Shapiro (2012) rightly notes that patronage targeted to the poor can alienate the middle class, regimes may also target patronage to directly benefit middle-class constituencies. In focusing on the state as patron not only of the poor but also of the middle classes, this study contributes to our understanding of the conditions under which modernization may lead to an expansion of patronage, rather than its contraction (Kitschelt & Wilkinson 2007). The argument in this book also advances scholarship over the past decade on the pernicious consequences of state dependency for democratization. Arriola (2013) relates the emergence of electoral competition in Africa to the rise of private ownership in the banking sector. Bellin (2000, 180) argues 8. For an excellent discussion of the types of assistance employees of the state can provide to mobilize voters, see Mares, Muntean & Petrova (2018).

46 c h a p t e r 2

that state-sponsored development makes organized labor and capital-owning elites “diffident about democratization,” because these actors realize that their economic prospects “hinge on state discretion.” While I share Bellin’s contention that state dependency moderates demands for democracy, I move beyond her focus on workers and capitalists, returning the middle classes to the story of democratization. And whereas Bellin (2002) wrestled with how to measure diffidence in the absence of opinion data, I am able to measure diffidence directly, in terms of both attitudes and behaviors, using a broad array of individual-level surveys. Also closely related to the micro-level mechanisms I investigate is McMann’s (2006) work, which links the uneven development of democracy in Russia and Kyrgyzstan to subnational differences in economic autonomy.9 McMann argues that lack of autonomy, or the ability to earn a living independent of the state, is “critical to the practice of democracy,” since “otherwise, citizens will avoid activism for fear of economic reprisals” (4). Fear of negative sanctions is undoubtedly an important part of the story. However, I also argue that positive inducements and opportunities for career advancement shape perceptions of the performance of different political systems and help to generate support for autocracy, a mechanism I refer to below as satisfaction. This argument leads to distinct empirical implications for those who are past beneficiaries but no longer directly subject to state sanctioning. I test these implications by differentiating between current and former state employees in chapter 3. I also depart from McMann in my empirical approach. To test her argument, McMann uses detailed case studies of subnational variation in the regional economies of two post-Soviet countries; but this macro-variation leaves individual-level mechanisms unexplored. While McMann’s evidence speaks to aggregate patterns of engagement, it does not distinguish the middle class. At the micro-level, McMann relies on interviews with political activists (and a helpful comparison group of non-activists). By contrast, I evaluate my argument on the middle class using representative surveys of citizens, across the ex-communist countries and a variety of institutional contexts. I also demonstrate that state dependency not only lowers levels of engagement, but reduces support for democracy. In sum, my analysis shows that economic 9. See also Gervasoni (2010, 304), who cites economic autonomy as a key mediating mechanism between subnational rentierism, rooted in fiscal federalism, and regime type in the Argentine provinces.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 47

autonomy has the strongest and most consistent effects on attitudes toward democracy among the middle class, but that other groups within society are vulnerable to the same set of mechanisms. Why would state dependency matter more for middle- than working-class support for democratization? The framework in this book suggests that state dependency may have a more pronounced effect on the middle class, because the middle class has both more to lose and fewer exit options, under certain labor market conditions. However, several other explanations are also possible. First, where state involvement in the economy is extensive, the working classes—both state and private—are likely to see opportunities for upward mobility into the middle class as tied to the regime. The working classes may thus be uniformly unwilling to embrace democracy, as they observe the value of loyalty and the price of opposition for career advancement. Second, in keeping with elite-competition theory, the working class may be unlikely to challenge the existing regime until it has something to lose. Third, specifically in the post-communist region, the working class’s uniform lack of support for democracy may be a legacy of the communist period, during which it was the principal target of ideological indoctrination. Fourth, modernization theory suggests that the working class lacks the prerequisites for democratic value formation, like education and occupational specialization. As such, it expects demands for democracy to emerge only as society becomes more middle class. Each of these mechanisms, or several in tandem, could explain why state employment has a less pronounced effect on the democratic attitudes of the working class. This study’s principal contribution, however, lies in demonstrating the effect of state dependence on middle-class support for democracy and in parsing the mechanisms that account for observed patterns of democratic preferences among middle-class groups.

The Logic of State Dependency In this section, I summarize the logic of state dependency at the heart of my argument. The analytical framework tested in this book emphasizes that democratization is, first, fraught with uncertainty and, second, accompanied by change in control of the state. In this sense, democratization carries many of the same risks for state-dependent groups as other forms of regime change. All regime change threatens status quo rules about whose interests are represented in policy. However, state-dependent groups can also expect democratization, by broadening representation and accountability, to

48

chapter 2

increase transparency and reduce opportunities to earn informal rents. As democracy provides a more meritocratic basis for hiring, promotion, and retention in the government sector, employees whose primary asset has been acceptance of authoritarian rule may see their prospects diminished. Further, most successful transitions involve both political and economic reform (e.g., Nelson 1993; Amin et al. 2012). In many contexts, including the one examined here, democratization raises the specter of state retrenchment, though certainly the neoliberal project has at times been more closely associated with authoritarian rule. Still, even in Latin America, young democracies enacted painful market reforms beginning in the 1980s, dismissing many government employees and privatizing public enterprises (Weyland 2004). Not only in the post-communist countries but around the world, the downsizing of public payrolls has frequently accompanied, or soon followed, democratization. As such, fears of state retrenchment, particularly in less competitive state enterprises, may shape the state middle class’s attitude toward democratization. Such fears are not, however, the only reason to expect state middle-class opposition to democracy. Whether democratization brings privatizations, lustration, or a growing state bureaucracy, there is little guarantee that the middle class that prospers under one regime will not be sacked or sidelined by its successor. Structural conditions of corruption and a weakly professionalized public sector are especially unlikely to produce a democracy capable of delivering on programmatic policies, as pointed out by the parties’ literature (Kitschelt et al. 1999). Hence, the state middle class in a country with a corrupt public sector and state apparatus that is only weakly guided by law should have little expectation that democracy will bring strong public goods provision. Instead, weak parties are more likely to form on clientelistic bases, packing the public sector with their own loyalists—and plausibly disadvantaging those associated with the preceding regime. Meanwhile, for those who retain their jobs, rising attention to the rule of law under democracy promises to staunch the flow of informal rents. For all of these reasons, the cost of supporting democratization is higher for the middle class in state-dependent sectors than for those whose economic opportunity, upward mobility, and life chances are independent of the state. As insiders to existing networks of patronage and political loyalty, those with careers in the state, or who are paid from the state budget, will be more likely to view democracy as threatening to the status, benefits, and rents that their public employment provides.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 49

The task of the book’s remaining chapters is to demonstrate the generality of the central argument that middle-class dependence on the state stands in the way of democracy by affecting middle-class attitudes and political behavior, then pin down how and why state dependency reduces middle-class demands for democracy. The latter emphasis on mechanisms is especially important, since even those studies that describe sectoral cleavages within the middle classes (e.g., Chen 2013; Gontmakher & Ross 2015) have failed to probe whether working for the state has any independent effect, or whether those who choose public employment are simply different. For social scientists, this search for mechanisms increases an explanation’s credibility (Elster 2015) by helping to “distinguish between genuine causality and coincidental association” (Hedstrom & Swedberg 1998). Each of the book’s empirical chapters thus probes the mechanisms outlined below. The book’s final empirical chapter additionally underscores how autocrats work actively to align growing middle classes with autocracy and seek to promote authoritarian stability by shaping middle-class formation. Before moving on to define key concepts, I summarize the arguments above by laying out several mechanisms, and alternatives, which could account for the link between state employment and middle-class reticence about democracy. Material incentives. It is well known that autocrats build coalitions of support using instrumental payoffs to secure loyalty. Public sector positions are an especially effective means of co-opting the middle class, because the benefits of a public-sector job and the threat of employment loss can generate strong incentives to support autocracy under certain labor market conditions. First, when the public sector is extensive, it crowds out private sector alternatives (Behar & Mok 2015). Indeed, across the post-communist region, state control in sectors like education, health, energy, transport, oil, and gas trends monopsonistic. Consequently, the state monopolizes employment opportunities in a number of occupations requiring specific training and credentials. This constrains the state middle class’s prospects of finding alternative employment, as discussed further below. Second, and relatedly, many public-sector professions are highly specialized and require skills, networks, and know-how whose value is highly specific. Whereas individualized, project-based work and flexible forms of employment are more abundant in the private sector, public-sector organizations and enterprises tend to be larger with idiosyncratic products, procedures, tasks, and technologies, which “bind work experience to a specific

50 c h a p t e r 2

context” (Marx 2011, 8). Public administration is a case in point, but other occupations in the public sector require skills that are similarly specific. Third, public-sector positions have certain material advantages. Though for some state-sector experience may result in skills and personal relationships that are valued in the private sector, forgoing a public-sector position can still have drawbacks. As discussed previously, the benefits of public-sector employment in an autocracy can be substantial: a formal labor contract and employment stability, on-time payment of wages, paid vacation and medical leave, access to preferential state medical treatment, transportation benefits, housing, and even differing status before the law. Under autocracy, government jobs often also provide opportunities to earn informal rents. Change in political control of the state and rising attention to the rule of law under democracy threaten these benefits. Not all of these considerations apply equally to every type of public employment. But together I argue that they generate a consistent set of incentives that makes the public-sector middle class more diffident about democratization than the private-sector middle class. As such, I expect that the middle class employed in public-sector jobs will be more likely than the middle class that is not, all else equal, to prefer stable authoritarianism. This proposition is tested in each of the book’s subsequent empirical chapters. The final empirical chapter also tests an extension of this logic to individuals employed in sectors of the economy that have a very high degree of indirect state involvement, despite structures of ownership that are formally private. Two concerns related to the loss of material benefits may drive state-sector managers and professionals’ reticence about democracy. One, state employees are likely to fear any political transition in which they might lose their jobs (as well as other particularistic benefits and rents afforded by their official position). This could occur either (a) because the entity that employs them ceases to exist or is transformed or (b) because they are perceived as “damaged goods” by the new regime, which may prefer to pack the public sector with its own loyalists. The adoption of lustration laws and compulsory employment vetting are examples of the risks to public-sector job stability that transition can entail.10 Two, publicsector employees may fear being fired for articulating democratic values that threaten the current system. 10. The use of these practices in a number of East European countries following the transition from communism may make such fears especially pronounced in this region.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 51

If fear of being fired in retaliation for holding the wrong political views or losing one’s job in the context of transition is the primary mechanism by which state employment affects support for democracy, we would expect the effect of state employment to be strongest for current employees.11 Likewise, we would expect those who are not current state employees, but receive potentially excludable benefits from the state (e.g., because they are injured or disabled), as well as those who are temporarily not working, but intend to return to state employment, to more closely resemble state than non-state employees in their degree of dependence and attitudes toward democracy. I examine these propositions in chapter 3 using survey data that allow me to distinguish between current state employees, recipients of state benefits that are not universal, those temporarily on leave from the public sector, and those who have left public-sector employment. Lower support for democracy also among the latter group suggests that fear of job loss and mobilization in state workplaces are not the only reasons why a large public sector aids authoritarian resilience. Satisfaction. When state employment improves individuals’ economic opportunities and life chances, it also generates satisfaction that is based in performance. Material benefits are used not only to coerce, but also to persuade. If personal satisfaction with “autocracy that delivers” shapes the state middle class’s regime preferences, we would expect to find that state managers and professionals are more satisfied with their lives and evaluate the performance of autocratic institutions more highly than others. In chapter 3, I investigate whether the state middle class expresses greater satisfaction with life overall and with the opportunities life has afforded them relative to their peers. In chapter 4, I show that Russia’s state middle class feels more secure and less vulnerable than their private-sector counterparts. This mechanism thus implies that working for an authoritarian state can have the hypothesized effect not only for immediate material reasons, and by inducing support for the current regime, but because it alters long-term 11. Empirically, it is important to distinguish those who sincerely report preferences consistent with the current regime from those who do so insincerely (i.e., because they fear retaliation). While it is very difficult to isolate this type of bias in survey data once collected, a variety of techniques allow researchers to indirectly ask politically and socially sensitive questions in such a way as to assure respondents’ confidentiality and minimize their incentives to lie (see, e.g., Rosenfeld, Imai & Shapiro 2016). While this type of response bias would be concentrated among current employees facing retaliation, I find that both current and former public-sector employees share similar views on democracy.

52 c h a p t e r 2

perceptions of the efficacy of autocratic institutions, calculated based on all available retrospective information. As in models of Bayesian learning, which emphasize the weight of past experience, this mechanism predicts that state employment shapes an individual’s history and experience of performance under an autocratic regime (or in Bayesian terms, contributes to an individual’s “prior” on autocracy).12 Rather than constituting a specific exchange between patron and client, public-sector jobs may foster a preference for authoritarianism by providing benefits that cumulatively enhance satisfaction with what autocratic institutions can deliver. This framework implies that state employment will be relevant not only for current employees but past employees, as well. It follows that, even when current wages are no longer at stake, those who have personally experienced greater benefits from autocracy over their lifetimes will be less supportive of democracy. To test these propositions, I turn in chapter 3 to survey data that include detailed employment histories for a large number of both current and former state employees. This mechanism also implies that when economic conditions deteriorate, and autocrats cut wages and benefits or allow the gap between public and private-sector compensation to grow, satisfaction with performance will diminish. Chapter 4 presents evidence consistent with this proposition: protest by the state middle class in 2011–2012 was motivated more by economic than ideological grievances, whereas for others in the middle class ideological grievances were paramount. The Kremlin’s promise, in response, to raise public-sector compensation for professionals in the government budget sectors further accords with the strategic logic of performance-based satisfaction. Over time, if unable to address such grievances, an autocrat will be forced to rely more heavily on immediate inducements like the threat of dismissal. The absence of alternatives, which I take up next, helps to sharpen these incentives. In short, then, the state middle class’s reticence about democracy is often associated with greater satisfaction with the status quo and support for the incumbent autocrat; however, a belief that autocracy has delivered over the long term (even if the current regime has not) and the perception of limited exit options also shape demands for democracy. The Ukrainian case, described in chapter 6, shows that even where dissatisfaction with the status quo is widespread, there is no guarantee that the state middle class will embrace democracy. 12. See, e.g., Magaloni (2006) following Achen (1992).

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 53

Long-run alternatives. Finally, an individual’s interest in ensuring the uninterrupted continuation of public-sector benefits is a function of his/her alternatives in the private sector. Where alternatives are plentiful, individuals can afford to make ideological investments. Lack of private-sector alternatives, by contrast, fosters long-term dependence on the state, which I hypothesize will negatively impact democracy support. The availability of private-sector alternatives depends on a variety of factors, chief among them individual endowments of human and social capital, experience, age, and gender. This implies that state workers who have spent some of their careers outside the state sector will be less dependent on the state than those who have spent their entire career in public employment. While the former have demonstrated their ability to access private-sector alternatives, the latter have no demonstrated exit options. To proxy for an individual’s available alternatives, chapter 3 uses detailed employment histories and chapter 4 exploits life cycle variation, while chapter 6 relies on a direct question asking respondents’ about the likelihood of finding equivalent employment in case of job loss. Finally, chapter 7 explores patterns of democracy support in economic sectors that are controlled overwhelmingly by the state. Regardless of the measure used, the evidence suggests that limited labor market alternatives constrain support for democratization. There are two principal alternatives to the logic based in material incentives and individual interest that I have just described. Socialization. Socialization or social conditioning implies that the experience of working in the state sector under authoritarianism inscribes state employees with enduring orientations such that they acquire a taken-forgranted quality and persist even as incentives change. A hallmark of socialization effects, as typically conceived, is that they cumulate or deepen over time (e.g., Pop-Eleches & Tucker 2014). While scholars have shown that socialization to antidemocratic attitudes occurred as a result of living under communism (Pop-Eleches & Tucker 2017), the effects of living under a post-communist authoritarian regime have received less attention. If socialization effects exist among state workers in the post-communist period, then we would expect that the longer one has worked for an authoritarian state after independence, the more likely one is to have been imprinted with antidemocratic attitudes. In other words, a significant negative effect for additional years of state employment would be consistent with socialization, understood as a cumulative deepening of antidemocratic attitudes.

54 c h a p t e r 2

In chapter 3, I examine both the possibility that socialization effects increase with each additional year of public-sector employment as well as the alternative that the first few years of a public-sector career matter more, with the effect of additional years diminishing over time. My findings suggest that cleavages within the post-Soviet middle class are not primarily a function of long-run socialization or ongoing indoctrination. Nor are they an artifact of socialization that occurred during the communist era. Instead, as the analyses in chapter 6 reinforce, opinion change occurs relatively quickly—a pattern that is inconsistent with long-term socialization and more in keeping with changed material incentives.13 This suggests that a long-standing and ideological authoritarian regime, such as existed under communism, is not necessary to produce the pattern of preferences I identify. This pattern of preferences can also arise as a result of economic incentives in the context of a political economy based on extensive state economic engagement, and not only ongoing ideological indoctrination. Selection. Throughout the analysis, I consider selection to be the main rival hypothesis. An account based on selection implies that individuals sort themselves into different career paths, electing careers in either the state or private sector for political reasons (or reasons correlated with politics), while the careers themselves have no independent impact on democracy support. I consider three variants of the selection hypothesis. First, I consider initial self-sorting of types in the analyses presented in chapters 3, 5, and 6.14 This would be a problem if democrats were less likely than nondemocrats to choose state employment or if, for example, risk-averse individuals were both more likely to seek a public-sector career and less likely to want democratization. In short, this alternative implies that observed differences between the state and private-sector middle classes are due primarily to the influence of prior characteristics on the choice of career path. Chapter 6’s research design, based on a panel survey of the same individuals over time, provides the strongest evidence against this alternative; that conclusion is further bolstered by analyses of original survey data with measures of pre-employment regime preferences in chapter 5. Chapters 6 and 7 additionally show that selection into state employment on ethnic or linguistic

13. Or social conditioning that occurs very quickly, as distinct from socialization that takes place over years. 14. See also the online appendix to chapter 4.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 55

bases is inadequate to explain the state middle class’s weaker support for democracy. Second, I address the concern that those who are more democratic leave state employment sooner than those who are more authoritarian (i.e., differential attrition). In other words, an alternative account based on differential attrition suggests that state employment’s apparent effect is due primarily to the fact that democrats are less likely than nondemocrats to “survive” in state employment. I probe the plausibility of this alternative in chapter 4, again making use of detailed data on long-term employment histories. Third, I investigate whether recruitment after regime transition could be responsible for state employment’s apparent dampening effect on support for democracy. Clearly, each of these forms of selection plausibly occurs in some cases. However, in subsequent empirical chapters, I rule each of these out in turn as viable alternative explanations for the book’s main findings.

Key Concepts Before moving onto the empirical investigation, it is necessary to give greater precision to several key concepts and definitions. I describe in this section how I conceptualize the study’s three main variables: democracy support, the middle class, and state employment. Because each of these concepts is the subject of vigorous debate, clarity over terms is essential.15

Democracy Support Democracy support can be expressed through both attitudes and actions. This book considers both. While chapter 3 focuses on preferences, subsequent chapters turn to behavior. Chapter 4 examines the intersection of support for democratic political actors and protest behavior in Russia, while chapter 6 investigates the relationship between beliefs about democracy, protest participation, and voting in Ukraine. To be sure, the preference for democracy and behavioral renderings like participation in a pro-democracy protest or voting for a democratic party/candidate are not the same thing. Some people who would prefer to live in a democracy may be fearful of revealing their preferences through political action. Conversely, some who take part 15. On the challenges of working with “essentially contested” concepts, see the discussion in Coppedge (2012, 11–14).

56 c h a p t e r 2

in pro-democracy protests or vote for democratic forces in an election may do so for other reasons. This study takes the view that both latent and revealed preferences are politically consequential. While revealed preferences drive democratic openings, studying latent preferences supplies information about how likely those openings are to succeed and which groups they are likely to resonate with. Whenever possible, I combine attitudinal and behavioral data, triangulating democracy support and the actions that advance it politically to shed light on these processes. The definition of democracy I adopt is conventional, though more encompassing than the minimalist requirement of regularly scheduled elections. Focusing on the institutions and procedures of democratic governance, I define democracy as: free and fair elections, held regularly on the basis of universal suffrage and contested by multiple parties or candidates; an independent judiciary; freedom of expression; and freedom of association (Schumpeter 1942; Huntington 1991; Collier 1999; Huber & Stephens 1999, 761). In the remainder of this section, I focus on conceptualization. Separate discussions of measurement appear in each empirical chapter. All together, the treatment of these issues follows Munck & Verkuilen’s (2002) related work on the methodological challenges of defining indicators of democracy. The preference for democracy as I conceptualize it in this study captures an individual’s support for the central features of a democracy, the liberal institutions that safeguard democracy in practice. These include an individual’s beliefs about competitive elections, independent courts, freedom of speech, and minority protections. Closely related is the belief that regime type matters, or “salience.” Whether individuals believe that the choice between democracy and autocracy affects them personally will shape the character of their democracy support. So too is whether their regime preference is strongly held or conditional on economic and political circumstances. This approach is more expansive than minimalist definitions of democracy support that focus exclusively on the preference for free and fair elections. It also rejects such elements of maximalist definitions as the belief in individual liberty, tolerance, and distrust of political authority (Gibson 1995). With this approach, the relationship between preferences over the institutions and procedures of democratic governance and these beliefs is left to empirical investigation. While the survey items I use to measure the preference for democracy vary across the analyses, I strive throughout to address three criticisms of citizens’ understanding of the concept. First, one concern is that most citizens lack

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 57

a clear appreciation of the meaning of democracy. Second, previous studies have suggested that the meaning of democracy is culturally bound and may be “lost in translation” (e.g., Shaffer 1998). A common version of this argument is that citizens, in the developing world especially, associate democracy with economic rather than political standards (Camp 2001; McIntosh et al. 1993). According to this view, support for democracy signals a desire for greater wellbeing rather than a democratic political system. Third, scholars have argued that citizens’ endorsements of democracy are hollow in an age when even autocrats embrace democratic rhetoric. Understanding what citizens mean when they endorse democracy is critical. To avoid this guesswork, I employ measures that are as concrete and explicit as possible. Still, it is worth noting that while some post-communist citizens have had personal experience living under democracy, others have had no such experience. For the latter group especially, interpreting survey responses about democracy poses certain challenges.16 Within the limits of available survey data, these challenges are best addressed by triangulating responses to several different items on democracy support, including items that inquire about democratic attitudes without employing the word “democracy” itself, as advised by Coppedge (2012, 323). Thus rather than rely only on generic endorsements of democracy, I opt for indicators that inquire about the preference for specific aspects of democratic governance, when they are available (see also Pop-Eleches & Tucker 2017). When such measures are unavailable, I chose alternatives that offer respondents a continuum of political regime choices that is appropriate to the context. At the same time, it is worth noting that critical claims about citizens’ understanding of democracy may be somewhat exaggerated. Dalton, Shin & Jou (2008, 6), for example, find that when asked to define democracy in their own words, respondents in fortynine recent national surveys across Africa, Asia, Eastern Europe, and Latin America identify democracy primarily with freedom and civil liberties, results that differ from what many skeptics have assumed. 16. Unfortunately, the cross-national survey analyzed in chapter 3 does not inquire about respondents’ conception of democracy, only about the importance of democracy in the abstract and that of various democratic institutions. Turning to other sources, Dalton, Shin & Jou (2008) find that level of democracy and cumulative democratic experience are weakly associated with greater emphasis on freedom/liberty than political processes such as majority rule and free and fair elections. In looking within a given country, however, the authors find no clear pattern of change over time in citizens’ propensity to define democracy in terms of freedom/liberty.

58

chapter 2

The Middle Class The principal actors in this story are the middle classes. Conceptually, I use the term “middle class” to describe individuals who are distinguished from both elites and the working class by their resources, education, and occupational position. The concept of class that I employ is based on social relations of status and power rather than economic relations of production, like other studies in the neo-Weberian tradition (see, e.g., Erikson & Goldthorpe 1992; Goldthorpe & Marshall 1992; Yang 2007; Tikhonova & Mareeva 2009; Maleva & Ovcharova 2009). This approach is also notably distinct from approaches that define the middle class (i) according to its ownership of capital (e.g., entrepreneurial bourgeoisie); (ii) in normative terms as a carrier of innovation, democratization, or modernization; (iii) as a unified class actor; or (iv) as a purely stipulative and statistical construct derived from median or absolute income. In addition to debates about the definition of middle class, there are also wide-ranging perspectives on how to best measure the concept.17 Operationally, I define the middle class in terms of education and professional status, emphasizing the theoretically relevant attributes of human and social capital and labor market position. These are also characteristics that sociologists have emphasized as shaping individuals’ unequal life chances—an outcome I argue is further influenced by individuals’ relationship to the state. While an extensive empirical literature in economics and sociology has described the composition and size of the middle class, scholars have reached little agreement on the concept. Both maximalist and minimalist approaches in the existing literature have shortcomings. On one end of the spectrum are economistic approaches based on cross-sectional income. Economists most frequently employ one of two types of measures. Relative approaches define as “middle class” individuals whose income falls within a symmetric interval around the median (e.g., Thurow 1987; Birdsall, Graham & Pettinato 2000; Pressman 2007). A key problem with this definition, however, is that it travels poorly to developing countries where the income distribution has a strong right skew with many poor. In such cases, taking a symmetric interval 17. See, for example, Weeden & Grusky (2005) for a useful summary of the various points of view on measuring class and an empirical comparison of results for U.S. data or Remington (2011) on the debate among Russian sociologists. A practical discussion of various class measurement indices and their application in the European context can be found in Leiulfsrud, Bison & Jensberg (2005).

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 59

around the median makes little theoretical sense. Such measures also suffer from problems of cross-national comparability. Absolute approaches address this problem by defining the middle class according to a lower and upper threshold of income based on purchasing power parity (e.g., Milanovic & Yitzhaki 2002; Banerjee & Duflo 2008; Ravallion 2010). However, several issues with income-based measures remain. One issue, which is particularly acute in developing economies, is that current income (and, moreover, consumption, its most frequent proxy) is a weak indicator of lifetime wealth and expected variation in returns on that wealth in the future (Sørenson 2000, 1538). What economists call permanent income may be better captured by education and occupational status, especially in developing countries. Another problem is practical but has significant implications: reported income is frequently unreliable.18 For one, asking about income is often sensitive and subject to reporting bias. Typical problems of measurement error associated with using a single indicator are exacerbated where there are systematic incentives to misreport. The large informal economies and enduring corruption of many developing countries heighten these incentives. Since misreporting of income is almost certainly nonrandom and probably related to outcomes of interest, focusing exclusively on income is likely to bias results.19 Nonetheless, such minimalist definitions are not without advantages. Income measures are widely available, making it easy to cross-check indicators across multiple sources of data. Parsimonious measures also facilitate the cumulation of knowledge by enhancing scholars’ ability to compare findings and validate results. Although my main measure of middle class is sociological, where possible, I also present results using the best available income/consumption-based measures. On the other end of the spectrum are maximalist definitions and the multifactor integrated criteria frequently found in sociological studies of the middle class. As recent work notes, such approaches are especially problematic when multiple attributes are weakly related and poorly correlated indicators are aggregated into measurement scales with little or no guidance 18. See Pop-Eleches & Donnelly (2013) for a careful examination of the problems of reliability and validity in the World Values Survey income measure. 19. If, for example, misreporting of income is related to either education (i.e., if those who are more educated give more socially desirable responses) or sector of employment (i.e., if state workers withhold information on informal sources of income/corruption) and these variables are also related to democracy support, estimates will be biased.

60

chapter 2

table 2.1. Operational Definition of Middle Class Middle Class

=

Education Incomplete higher education University degree Postgraduate degree

+

Occupation Technicians & associate professionals Professionals Managers

Note: The occupations included correspond to the ILO’s ISCO-08 categories 1–3.

from theory (Weeden & Grusky 2005). Further, maximalist definitions define away many of the most interesting issues for empirical research (Munck & Verkuilen 2002). This tendency is evident in much of the single-country case literature on the middle class, where attributes of economic and political behavior are often incorporated into the very concept of middle class. Rather than ask whether or under what conditions individuals with middle-class professions, incomes, or education save, invest, express democratic preferences, or participate politically, many studies dispatch with these issues by definitional fiat. Like other political scientists, I argue that the connection between class, political preferences, and behavior is best left to empirical investigation (Brooks 1994; Manza & Brooks 1996). Table 2.1 gives my operational definition of middle class. Middle-class status is attained as individuals gain the social and professional status markers of higher education and a nonmanual, white-collar occupation. Rather than emphasizing current income, my operational measure follows my conceptual definition in emphasizing endowments of human and social capital.20 Where educational and occupational criteria overlap, we can be more confident in the reliability of our measurement instrument. I call those individuals who possess both the educational and occupational characteristics in table 2.1 “middle class,” and they are one of the two key independent variables in my analysis, together with state employment. Elites, meanwhile, comprise the small minority who hold economic or political power. Per Mills (1956), they are political leaders, senior government officials, major corporate owners and directors, and high-ranking military officers. While most elites are highly educated, most of the highly educated are not, by this definition, elites. 20. Note that economists incorporate this information in their definition of “permanent income.”

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 61

Not in survey samples

Elites

Population

Middle class In survey samples Working class

figure 2.1. Correspondence of class structure with survey samples. The population can be stratified into elites, the middle class, and the working class. This study assumes, however, that only the middle class and working class are actually represented in survey samples. Elites comprise only a very small share of the population in developing countries and are least likely to be accessible to survey researchers.

I define as non-middle class all respondents who have lower educational and/or occupational attainment. I use the terms “non-middle class” and “working class” interchangeably, employing the latter to stipulate the group below the middle rather than an organized or self-conscious actor. Though the population can be stratified into elites, the middle class, and the working class (as shown in figure 2.1), this study assumes that only the middle and working classes are actually represented in survey samples. Given that elites comprise no more than a fraction of the population in developing countries and are least likely to be accessible to survey researchers, this is a reasonable and defensible empirical assumption. Though owners and highlevel managers of enterprises in the commercial, financial, and industrial sectors are captured by my measure, the bourgeoisie, as typically defined (e.g., Moore 1966; Collier & Collier 2002), makes up only a small part of the middle class examined in this book. I define as middle-class occupations those that are nonmanual and not routine,21 focusing on white-collar workers and professionals, the group sometimes called the “new middle class” in the sociological literature. This definition thus recalls the middle class of modernization theory and its valuesbased variants (e.g., Lipset 1959; Welzel & Inglehart 2008). Like Collier & 21. And not focused on rote tasks as are clerks or traders.

62

chapter 2

Collier’s (2002) “middle sector,” its members comprise a broad range of occupational groups between the working class and economic elite. The middle class studied here is thus much broader then Moore’s (1966) bourgeoisie, though these categories are overlapping. Similar to the middle class of elite competition theories, empirically it has above-median income (Ansell & Samuels 2008). Three broad groups corresponding to the first three categories of the International Labor Organization’s (ILO) International Standard Classification of Occupations (ISCO-08) comprise this book’s middle class:22 managers, professionals, and technicians/associate professionals. In practice, this includes those who “direct, coordinate and evaluate the overall activities of enterprises, governments, and other organizations, or of organizational units within them”; doctors, teachers, lawyers, engineers, mathematicians, architects, programmers, and accountants; those with specialized technical skills and semiprofessionals in science, engineering, health, business, the law, culture, and communications.23 According to one influential perspective in contemporary sociology, occupations are the soundest basis for class analysis, because they shape lifestyle, resources, and behavior and are inscribed in the institutional structure of society (Weeden & Grusky 2005). My approach accords with this view. Finally, together with the occupational criteria just described, I define as middle class individuals who have at least some university education. This group is made up primarily of individuals with higher education, but also includes a small group with post-graduate degrees. To be sure that my choice of threshold is not unduly influential, in various analyses, I extend my definition to include those with a post-secondary technical or vocational degree. The specifics are discussed in each chapter.

22. See the ILO’s definitions at http://www.ilo.org/public/english/bureau/stat/isco /docs/gdstruct08.doc. 23. The working class thus includes individuals with less than a university education; clerical support workers, including most categories of secretaries and administrative professionals; service and sales workers (though, note that clerical and service employees in supervisory or managerial jobs are classified as managers); skilled agriculture, forestry and fishery workers; plant and machine operators and assemblers; elementary occupations; and armed forces occupations (i.e., ISCO-08 major groups 0 and 4-9). The four- or five-year specialist (spetsialist) degree conferred under the Soviet system and the bachelor’s degree, more recently adopted across most of the region, are considered higher education, while vocational and professional postsecondary degrees (professional’no-tekhnicheskoe (PTU, tekhnikum)) are not.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 63

Across the post-communist region, educational levels are high. In part an inheritance of the Soviet era, educational levels in the region resemble those of the established democracies in the OECD and are higher than in Latin America, the Middle East, or much of the developing world. Since 1989, however, countries across the region have also experienced an educational boom. According to Shleifer & Treisman’s (2014) calculations based on World Bank data, access to tertiary education has grown by an average of 33 percentage points across the post-communist states (7). Higher education enrollments in the average post-communist country outpaced Switzerland by 2012. In chapter 3, I investigate whether Soviet education makes the middle class educated under communism somehow different from the middle class educated after independence. While education comes to have a weakly liberalizing effect after independence that it did not have under communism, when the system was highly ideological and antidemocratic, state dependency similarly suppresses democratic demands among the middle classes educated in both periods, consistent with this book’s argument.

State Employment A central claim of this study is that the state shapes socioeconomic stratification in ways that affect the social bases of support for democracy. Publicsector employment distinguishes those whose economic opportunities and life chances depend most strongly on the state. This includes several broad categories of state employment: civil servants and others employed directly in the state administration, employees of state-owned enterprises, and professionals paid out of the state budget, such as medical professionals and teachers. This latter group typically forms the largest share of the state salariat. Though each of these public-sector groups is in some ways distinctive, they also share important similarities. As discussed in chapter 1, autocrats (more so than democrats) treat public-sector jobs as an excludable benefit (though democrats do too at times). They also use employment to distribute additional perks and particularistic benefits. Not only bureaucrats in the state administration, but also teachers and doctors frequently earn informal rents by exploiting their official position. And though budget-sector institutions and the state administration do differ in many respects, all provide for repeated interactions that help to generate loyalty and facilitate supervision. It is no accident that polling places are often located in government buildings, schools, and hospitals, or that pro-regime demonstrators and rally organizers are recruited from these same places.

64

chapter 2

The coding of public-sector employees in this study follows the ILO definition.24 First, this group comprises civil servants and others employed in the state administration at all levels of government (central, state, regional, and local), including ministries, agencies, and departments. Second, this group includes employees of all other institutions that are controlled and mainly financed by the state, and thus budget-sector workers. I also adopt the standard ILO definition of public corporations (or state-owned enterprises, a term I use equivalently) as legal units mainly owned or controlled by the government that produce goods and services for sale in the market. Some examples include post offices, railways, steel smelters, and mines, though there is considerable variation in the activities of state enterprises over time and across space. For the reasons just mentioned, this study often considers different types of public sector employers together. However, this approach is further justified in analyses that distinguish the state administration from health, education, and state enterprises in subsequent empirical chapters. Before concluding, it is worth noting that the prevailing form of state economic engagement has changed a great deal over the past two decades. However, rather than wither away as neoliberal reformers intended—and Marx and Engels once predicted—the state has continued to thrive. Various forms of state-led development and state capitalism are resurgent. The data in table 2.2 give some sense of the importance of state ownership for select nondemocratic countries where data are available (Musacchio & Lazzarini 2012, 7). The table shows that, after two decades of privatization, state-owned enterprises still represent a significant portion of total stock market capitalization and contribute a large share of GDP in some of the world’s largest nondemocracies. At the same time, the reemergence of state capitalism has come with many new categories of government control. These strategies as well as their effects on politics are still poorly understood. Even measuring state influence poses significant challenges in an era when states increasingly combine state ownership of enterprises with other forms of control—investment in partially privatized firms, minority stakes held by state-owned holding companies, and

24. See the ILO’s definition of public sector employment at http://laborsta.ilo.org/ applv8/data/sectore.html and http://www.ilo.org/public/english/bureau/stat/download/ wp_pse_e.pdf.

m i d d l e - c l a s s d e m a n d f o r d e m o c r a c y 65 table 2.2. State Ownership in Selected Nondemocracies

China Egypt Malaysia Russia Singapore Thailand Vietnam

SOE output to GDP (%)

Number of listed firms with some state stake

SOEs as a % of market capitalization

Number of SOEs with majority state control

30%

942

70%

15 12 12 6

36% 40% 20% 21%

167,000 57 52 8214 20 60 3364

20% 12% 26% 34%

Number of firms with minority state ownership∗ 59 28 1418

Note: This table shows patterns of state ownership for selected nondemocracies circa 2010. SOE = state-operated enterprises. ∗ Includes firms only in which the federal government has minority ownership. Source: Musacchio & Lazzarini (2012, 7).

loans by state banks or other state-controlled investment funds.25 Many of these arrangements are beyond the scope of this study.26 However, if other forms of state economic engagement affect political preferences in ways that resemble the more direct forms of control examined here, this book’s main findings may be interpreted as conservative.

25. For example, Musacchio & Lazzarini (2012) note that in some privatizations, governments take minority positions but, nonetheless, maintain veto power over corporate decisions (fn. 4). See also Bortolotti & Faccio (2009) on the difficulties of discerning a government’s true share of equity and influence in the OECD countries. These problems are only magnified in the context of developing states. On this subject, see also McGregor (2010) and Bremmer (2010). 26. Of course, dependence on the state may also be considerable among some categories of the private sector, for example, among private sector employees in industries dominated by state contracts or where private firms receive extensive government assistance. In the cross-national analysis, where case-by-case judgments based on industry-level information are impractical, I follow the convention in the literature on privatization and adopt a definition of state ownership as majority control (Musacchio & Lazzarini 2012, 3). In the book’s final chapter, I relax these sharp distinctions in an analysis of Kazakhstan, a country that exemplifies many of state capitalism’s new tactics and indirect forms of control.

3 The Post-Communist Middle Classes, the State, and Democratization

when communism collapsed in Eastern Europe and the Soviet Union, there was widespread euphoria about the region’s democratic prospects. Its middle classes—both those emerging from communism and those its new markets promised to create—were accorded a central role in the unfolding drama of democratization. As the story went, communist modernization had created a large, politically dissatisfied middle class of educated whitecollar workers and intellectuals, who had buried communism. This group was now prepared to build democracy, aided by the democratic forces now being unleashed in a new market-based middle class. To quote just one exuberant formulation from the 1990s: “As the most influential social force propelled by the modernization process, the educated middle classes are bringing electoral democracy to the center stage of history” (Vassilev 1999, 599). These processes were considered deterministic, not contingent. Yet in hindsight, democracy receded across much of the region even as the middle class expanded. The present chapter addresses this apparent paradox. In it, I provide the first rigorous cross-national examination of the regime preferences of the post-communist middle classes. I furnish clear evidence that middle-class demands for democracy are more contingent than frequently assumed. Such demands will remain limited, I have argued, where middle-class groups fail to gain economic autonomy from an authoritarian state. To show that the argument laid out in chapter 2 applies across different settings, I use detailed information about labor market participation and political attitudes for a 66

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 67

large cross-national sample of post-communist citizens in Central and Eastern Europe and the former Soviet Union. Broadening the sample to include both democracies and nondemocracies helps to clarify why prior research, which has overwhelmingly focused on democratic contexts, has failed to find the sectoral differences in middle-class support for democracy that this book highlights. By contrast, a sample that includes multiple regime types reveals that, under authoritarianism, the state can limit middle class demands for democracy through the provision of public sector employment. This chapter demonstrates the utility of disaggregating the middle classes. I show a variety of empirical evidence that the state middle class holds economic and political preferences that are distinct from those of the privatesector middle class, despite similar incomes, educations, and occupations. These differences, as I have argued, are key to understanding the political implications of an expanding middle class. The results in this chapter confirm that the state middle class is less democratic than the private-sector middle class across the nondemocratic postcommunist countries. By contrast, in post-communist democracies, where I do not expect state employment to weaken support for democracy, indeed, no such relationship is evident. There, public employment has a weakly positive effect on democracy support. These results are robust to changes in the coding of middle class and a broadening of the definition of democracy support. I find that as the share of an individual’s career spent in the state sector since the collapse of communism increases, support for democracy decreases. My analysis makes clear that this finding is not an artifact of age or Soviet socialization. Further, consistent with the argument that public-sector jobs generate satisfaction with the performance of autocracy, and do so by improving individuals’ economic opportunities and life chances, state employees are more likely than others across the post-communist nondemocracies to be satisfied with their lives. They are also more likely to report having done better than both their high school classmates and colleagues around 1989. At the same time, I show that regime preferences, support for the status quo, and support for a particular incumbent, though often intertwined, are not the same thing. A state middle class that is satisfied with the status quo and loyal to the regime will naturally be circumspect about democracy. However, these attitudes do not always go together. Even where the state middle class expresses grievances with the status quo (as in Russia during the 2011–2012 protests) or is generally unsupportive of the regime (as in Ukraine following the Orange Revolution), it may still reject democracy for the reasons

68

chapter 3

outlined in chapter 2. In a similar vein, this chapter also demonstrates that attitudes toward democracy and the free market are related, but distinct. Greater antipathy toward free markets alone does not explain the state middle class’s reservations about democracy. I consider, in addition, several alternative mechanisms by which state employment might matter for the middle classes’ democratic preferences (or appear to), but find weak support for each of them. These include: fear of being fired, blaming democracy for the economic dislocation of the 1990s, different preferences over redistribution, the persistence of communist-era socialization, and generational effects through communist party membership or repression. My analysis also rules out the possibility that the ideological and explicitly antidemocratic character of the communist educational system is responsible for low support for democracy among the state middle class. The pattern of preferences that I identify is not merely a historical inheritance based on unique features of Soviet education. Finally, the analysis addresses social desirability bias in survey responses as well as bias resulting from selection. Across a variety of tests, my results consistently suggest that employment experiences affect the political values of state workers and that observed differences between state and non-state workers are not due solely to sorting into career paths based on prior characteristics. The upshot is that control over public employment benefits autocrats through the creation of a middle class that is economically self-interested in the maintenance of an autocratic system. In the next section, I describe the gap in cross-national research that this chapter seeks to fill. I then introduce the survey data that I use to test my hypotheses, define key concepts, and explain my choice of measures. The following section reports the cross-national empirical results and examines alternative hypotheses. The final section summarizes the implications of these findings for the argument in this book and concludes.

Socioeconomic Status, State Employment, and Democratic Attitudes Despite the centrality of the middle class to prominent theories of democratization, existing cross-national research on the subject has been surprisingly devoid of politics. Indeed, such studies, primarily from the field of development economics, can be divided broadly into two categories. The first investigates the impact of the middle class on economic growth. The second

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 69

examines its size, savings rates, and consumption patterns. Ravallion (2010) studies changes in the size of the developing world’s middle class as well as cross-country differences in the pace at which the middle class has grown. Using an income-based definition, Ravallion finds that 1.2 billion people in the developing world joined the global middle class from 1990 to 2005. He attributes cross-national variation in the expansion of the middle class to initial differences in poverty and rates of economic growth. However, his view of this process is entirely apolitical and ignores the fact that regimes in China and Russia, where growth of the middle classes has been greatest, have explicitly invested in middle-class formation (Chen 2013; Remington 2011). Banerjee & Duflo (2008, 19), meanwhile, find that steady, salaried employment rather than an entrepreneurial spirit distinguishes the developing world’s middle classes from its poor. This too points to politics, since the main source of stable employment in developing nondemocracies is often the public sector. Among quantitative cross-national studies of the middle class, only Easterly (2001) deals explicitly with politics. Easterly finds that a larger share of income for the middle class is associated with more democracy and greater political stability. However, by treating the middle class as an aggregate, Easterly cannot establish whether it is middle-class demand for democracy that explains this relationship, or some other factor related both to levels of democracy and middle-class size. In short, the ability to make inferences about micro-level preferences and mechanisms based on these macro-level relationships is limited. Instead, using appropriate data, we should look for evidence that the implied relationships exist also at the individual level. Micro-level evidence of middle-class demand for democracy would: 1) increase our confidence that the apparent cross-national relationship between middle-class size and democracy is not spurious and 2) help to clarify otherwise vague causal mechanisms. Yet there are currently no individual-level analyses of these questions across multiple nondemocracies, despite new research opportunities and data sources that have emerged over the past twenty years. Since the collapse of communism, a number of scholars have emphasized the importance of new economic activity in the private sector to sustaining democratic coalitions in the region (Mach & Jackson 2006; Frye 2006; Jackson, Klich & Poznanska 2005; Frye 2003; Przeworski 1991).1 Still others point to the post-Soviet middle classes’ continuing reliance on state support 1. Frye (2003) and Jackson, Klich & Poznanska (2005), for example, deal directly with the link between private-sector development and democratic practices. Other research has focused

70 c h a p t e r 3

(Remington 2011; Shevtsova 2012) and have shown that their participation in pro-democratic protest is limited by state dependency (Rosenfeld 2017). The present chapter builds on these insights. It asks: Across the countries of eastern Europe and the former Soviet Union, is the middle class more likely to hold liberal democratic attitudes? In the autocracies, does dependence on the state diminish middle-class support for democracy? While a body of existing literature expects that the middle class will be more likely to advocate for institutionalized political responsiveness, this book makes the argument that support for democracy will depend crucially on whether the middle class’s economic opportunities and life chances are tied to the existing regime. In this chapter, I test that argument using survey data from twenty-nine post-communist countries. This approach ensures greater generalizability across different institutional contexts (e.g., resource and non-resource states, more and less repressive regimes, unitary and federal systems). At the same time, because the cases are drawn from a single region, the design controls for many unobservable aspects of geography, culture, and shared communist history. I return to the question of how well these results are likely to travel to other regions in the Conclusion.

Data, Concepts, and Measures Cross-national Data The European Bank for Reconstruction and Development’s (EBRD) Life in Transition Survey (LiTS) covers twenty-seven countries in Central and Eastern Europe and the former Soviet Union and approximately 27,000 individuals. My analysis focuses on Wave I, which was conducted in 2006.2 An advantage of the LiTS dataset is that it includes a detailed employment history for each respondent over the entire post-communist period. For each year in which respondents received a wage since 1989, they are asked to recall their occupation, industry, and the ownership of their employer—whether state, private, or foreign. on support for the liberal pro-market parties, which in practice have often also advanced democratic reforms. 2. Unfortunately, the EBRD did not include the same political values measures in the second wave.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 71

In addition, the survey includes common demographic questions as well as measures of household consumption. It also contains items about economic satisfaction, concern over income inequality, perceived social mobility over the post-communist period, and subjective position in the country’s income distribution. These allow me to compare the effect of state employment on individual regime preferences to a variety of competing explanations. Most importantly, the LiTS is unique insofar as it combines panel-like data on an individual’s employment history with measures of political preferences. The fact that the survey asks identical questions across a large number of democracies (N = 18) and nondemocracies (N = 9) that share important developmental legacies of communism provides additional analytical leverage (Pop-Eleches & Tucker 2017; Beissinger & Kotkin 2014; Pop-Eleches 2007).3 In particular, these countries share a common history of universal state employment, giving us greater confidence that when we compare across the democratic and nondemocratic cases, the results are not confounded by inheritances from the communist era. A sample that includes both democracies and nondemocracies also helps to clarify why prior research has missed the contingent relationship between an expanding middle class and the emergence of democratic values that this chapter demonstrates. Building on the discussion in chapters 1 and 2, the following sections explain how the main independent and dependent variables are measured.

Democracy Support This chapter focuses on explaining variation in patterns of democracy support cross-nationally. To measure democracy support, I use eight survey items. The first asks: “With which of the following statements do you agree most: (1) Democracy is preferable to any other form of political system; (2) For people like me, it does not matter whether a government is democratic or 3. The country-level variable democracy (democ) is coded based on the dichotomous Boix-Rosato democracy variable. The subset of democracies includes 18,002 observations (or about n = 1,000 per nationally representative country sample) in: Albania, Bulgaria, Croatia, the Czech Republic, Estonia, Georgia, Hungary, Latvia, Lithuania, Macedonia, Moldova, Montenegro, Poland, Romania, Serbia, the Slovak Republic, Slovenia, and Ukraine. The nondemocracies (n = 9,000) are: Armenia, Azerbaijan, Belarus, Bosnia, Kazakhstan, Kyrgyzstan, Russia, Tajikistan, and Uzbekistan. LiTS surveys conducted in Mongolia and Turkey are excluded from the otherwise East European and former Soviet post-communist sample.

72

chapter 3

authoritarian; or (3) Under some circumstances, an authoritarian government may be preferable to a democratic one.” This item captures both the salience of regime type and the preference for democracy. It probes whether respondents view the type of government that they live under as relevant and whether they prefer democracy, or see authoritarianism as sometimes preferable. The remaining items ask to what extent respondents think various democratic institutions are important for their country: (1) free and fair elections, (2) freedom of speech, (3) an independent press, (4) courts that defend individual rights against abuse by the state, (5) equality before the law, (6) minority rights, and (7) a strong political opposition. Responses to these items are coded on a five-point Likert scale from strongly disagree to strongly agree. Respondents’ answers to these questions are highly correlated. The Cronbach’s α for standardized scores of the eight items just mentioned is 0.86, well above the conventional threshold of 0.7, indicating strong internal consistency (Nunnally 1978). Exploratory factor analysis supports the decision to treat these indicators as constitutive of a single underlying concept.4 While imperfect like all survey questions, together these items have several advantages. In particular, the second set of items provides a concrete and multifaceted assessment of respondents’ attitudes toward democracy, which does not rely on the meaning respondents assign to the word “democracy.” Meanwhile, the first item allows me to distinguish respondents who offer interviewers their views on various democratic institutions, but feel that for people like themselves the system of government is largely irrelevant. Since democracy support requires both the belief that regime type matters and a preference for democratic institutions, my primary dependent variable is a product of respondents’ scores on each of these two attributes. I also disaggregate these items and use several alternative measures described below (Munck & Verkuilen 2002, 24). Based on the first question, I create an indicator for the belief that democracy is preferable. I code as zero those who believe authoritarianism is sometimes necessary and those who believe regime type does not matter for people like themselves. For the second component, I code an indicator for those who “agree strongly” that the seven democratic institutions mentioned above are important. In the analysis that follows, I also relax this measure to include those 4. See online appendix section A.1.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 73

who “agree somewhat” with the importance of democratic institutions and show that this choice makes little difference. It is worth noting, however, that the more lenient measure of democracy support is less highly correlated with democratic outcomes. This suggests that accounting for preference intensity as I do here results in a measure with greater predictive validity.5 The analysis in this chapter focuses on the group I call “liberal democrats,” who express strong, consistent support for liberal institutions as well as a preference for democracy over other political systems. This group of respondents has three important characteristics. First, the intensity of their democracy support is greater than that of other respondents. Second, their responses indicate that they consider regime characteristics salient; that is, they matter to the respondent personally. Third, their answers demonstrate ideological consistency and coherence across dimensions.6 In addition to my primary measure of democracy support, I also test several alternative conceptions. I examine separately the items on institutions and salience. I also investigate more and less aggregated measures of the preference for democratic institutions. Indeed, the pattern is very similar whether democracy support is measured using a single question on the importance of a political opposition or a standard index of the democratic institutions items. These results are further discussed in the section on robustness checks.

The Middle Class The measure of middle class used in this chapter follows directly from table 2.1, presented in chapter 2. Together with state employment, this measure of the middle class is one of two key independent variables in my analysis. Again, this measure supposes that middle-class status is attained as individuals gain social and professional status markers: higher education 5. See discussion below. The overall correlation between the relaxed definition of democracy support and a country’s Polity score is ρ = .384, compared to ρ = .577 for the more stringent definition, suggesting that a narrower definition is preferable. 6. Before proceeding with the analysis, I also confirmed by comparison with national democracy scores that my measure of democracy support represents more than casual attitudes or random responses, but is rather closely associated with democratic outcomes. This correlation, of course, tells us nothing about the direction of causation; it simply implies that the political attitudes I measure are anchored to political realities. Further details regarding the relationship between democracy support and democratic outcomes are given in online appendix A.1.1.

74

chapter 3

and nonmanual, white-collar occupations.7 For those who are retired or temporarily not working, I code the respondent’s most recent occupation. My occupational definition of the middle class thus includes a modestly sized cohort who are not economically active but last possessed a middle-class job. To be sure that this choice is justified, I test whether my results are sensitive to the inclusion of retirees. I find that they are not. Finally, it is worth noting that while education has long been considered an important variable in its own right, alone it is insufficient to explain middleclass regime preferences in the post-communist nondemocracies. In fact, contrary to received wisdom but consistent with recent work reconsidering the causal effect of education on democratic values (e.g., Wang 2018; Acemoglu et al. 2005), I find that higher education is not consistently related to regime preferences in this sample. Last, to be sure that my results do not hinge on defining the middle class in terms of education and occupation, I also code the middle class according to Ravallion’s (2010) income-based criterion. The results remain unchanged.

State Employment The primary state employment variable that I use in the analysis takes a value of one if an individual has been employed in the state sector since the collapse of communism and zero if not. Across the post-communist countries in my sample, the share of respondents who have been state employees at any time since 1991 is 42 percent. This figure is even higher, 59 percent, among the middle class, reflecting the important role that public employment has played 7. The empirical correlation between a university education and a middle-class occupation for the full sample is .535. Of those with higher education, about 50% are professionals in various fields, 15% are technicians and associated professionals, and 10% are managers, business owners, or senior officials. Those who have higher education but are not considered middle class in this study are employed by and large either as clerks (another 9%) or service, shop, or market sales workers (7%). Less than 10% are employed in other occupations, including unskilled labor, craftsmen, and machine operators. Across the 27 countries in my dataset, the correlation between the educational and occupational criteria for middle-class status range from .409 in Armenia to .668 in Macedonia (see online appendix figure A.5). The average correlation across the more economically advanced democracies in the sample is only slightly higher (.536) than the nondemocracies (.530). One factor affecting these correlations is that, across many postSoviet countries, the employment share of routine and manual labor has remained high even as access to education has grown. In these data, as in the real world, a low share of white-collar, professional jobs limits the overall size of the middle class.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 75

for this group. In the nondemocracies with less reformed economies, the figure is higher still: 67 percent. These figures capture the extent to which state employment has remained a way of life for a large segment of the post-Soviet middle classes since the collapse of communism. Taking advantage of the pseudo-panel nature of my employment data, I also code three additional variables in order to probe why and how state employment affects political preferences. Since only current employees are vulnerable to being fired for expressing “politically incorrect” views, including a dummy for current state employment allows me to test whether the apparent effect of a state career could be due to biased survey responses. In 2006, at the time of the survey, the overall share of state employees was about a sixth in the democracies and a quarter in the nondemocracies.8 If reporting bias is a problem, we would expect that current employment would absorb much of the effect of the other state employment variables and that its coefficient would be negative. In addition to the binary indicators for state employment, I also code continuous variables for years of employment and career share in the state sector during the post-communist period. Years of post-independence state employment vary from zero to fifteen, while post-independence state career share ranges between zero and one. Public-sector career share captures well an individual’s demonstrated access to private-sector alternatives. It also reflects the extent to which an individual’s professional networks and know-how are tied up with the regime. I interpret this variable, therefore, as long-term dependence on the state. I expect state career share to be inversely related to democracy support. After controlling for state career share, also including years of state employment allows me to investigate differential attrition as a source of selection bias and socialization effects. An account based on differential attrition implies that nondemocrats last longer in state employment under a nondemocratic regime. In a regression with both state career share and years of state employment, we would expect years in the state sector to be negative and significant if differential attrition is driving the results. Using years of state employment, I can also test for evidence of workplace-based socialization. A significant negative coefficient on years of state employment would be consistent with socialization, understood as a cumulative deepening of antidemocratic attitudes. 8. Among the employed, these figures were considerably higher: 35% and 46%, respectively.

76

chapter 3

Of those in my dataset who are middle class and employed by the state, about half are budget-sector workers. Some 42 percent are in education, while 11 percent are in public health.9 Another 12 percent are in public administration, including the military and security apparatuses. In total, approximately two-thirds of the state-employed post-communist middle class works in these three industries. The remainder works in various state-owned enterprises. Among those who have worked in the public sector since 1991, the average length of post-communist employment is 8.3 years. Across the full sample, approximately 10 percent have worked in the state sector continuously since the collapse of communism, from 1991 until 200610 In the analysis that follows, I use variation in the type and timing of state careers as well as their duration to gain leverage on the question of whether state employment itself or merely selection (self-sorting) accounts for the apparent effect of a public-sector career. In particular, I exploit the fact that medical professionals are plausibly less motivated by political considerations than those working directly for the state administration. To preview, I find consistently that the sorting of less democratic types into state employment accounts poorly for patterns in the data that are better explained by state employment subsequently shaping regime preferences. These results are thus similar to Gans-Morse et al. (2017), which concludes that corruption in Russia results from the transformation of bureaucrats’ behavior and attitudes after entering the civil service, rather than through a process of corrupt self-selection.

Controls I also control for a variety of alternative explanations. First, I include age and age squared to capture generational differences in attitudes toward democracy and differences in communist and post-communist socialization. Since state employees tend to be older, controlling for age ensures that state employment is not confounded by generational effects. Second, I include gender with males coded as one. Like age, gender potentially confounds the effect of state employment, since the state-sector workforce in many countries is 9. I code as budget-sector employees respondents who both work for the state and are employed in either health or education. 10. Online appendix table A.1 gives detailed information about the composition by sector by industry of the post-communist middle class in the EBRD sample.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 77

disproportionately female. If women’s regime preferences differ systematically from those of men, failure to include gender as a control could bias the results. Third, I use a logged measure of household expenditures that has been equivalized per adult as a proxy for income. In regions where wages are often paid informally and not reported to tax authorities, many scholars prefer this measure to current income, which is often underreported.11 I have also run the regressions with an additive scale of consumer durables; but since these two measures are correlated (0.607), I report the results with only the household expenditures variable included in the model. These models may be interpreted two ways: 1) as testing the independent effect of current income above and beyond the effect of middle-class status that runs through human and social capital, and 2) as controlling for any wage gap between the public and private sector. I expect this proxy for income to be positively related to support for democracy in line with elite-competition and modernization theories and contrary to redistributive theories. If most of the effect of being middle class runs through occupation and education and the additional explanatory power of income is negligible, its effect should be substantively small. Fourth, I include respondents’ subjective evaluation of their location in the country’s income distribution (subjective income decile). Because this measure is correlated with respondents’ evaluations of their own social mobility over the post-communist period (0.517), I report only subjective income decile and drop mobility in the regressions that follow.12 Fifth, I control for attitudes toward the free market, since studies of democracy support in the region have often found a close connection between economic and political attitudes. Sixth, I use a measure of trust in the president and two items 11. Of course, systematic measurement error in independent variables, such as underreporting of income, is only a problem if it is related to the explanatory variables (in this case, middle-class status and state employment), to other variables that are correlated with the treatment, or to the outcome (democracy support). If, for example, underreporting of income is related to either education (i.e., if those who are more educated give more socially desirable responses) or sector of employment (i.e., if those in the state sector who have access to many informal rents are more likely to withhold information on these sources of income), estimates will be biased. These scenarios are highly plausible; therefore I avoid current income. 12. The results are similar for mobility. Note also that respondents’ subjective evaluation of their location in the income distribution is only modestly correlated with both ownership of consumer durables and equalized household expenditures. For this reason, I include both measures of subjective and objective well-being.

78

chapter 3

on economic and overall life satisfaction to capture support for the autocrat and the status quo, respectively. Controlling separately for unemployment has no effect on the results, so I interpret the findings for non-state employees as indicative of preferences in the private sector. The full specifications include all controls as well as country fixed effects.13 In a final section on other alternative mechanisms, I also control for economic experiences since the collapse of communism, preferences over redistribution, and communist socialization.

Empirical Results I begin by examining whether being middle class is associated with greater support for liberal democracy. I use figures to intuitively display predicted probabilities of democracy support, separately for the democratic and nondemocratic post-communist countries.14 The first set of empirical results, for the democracies, is summarized in figure 3.1.15 The results conform strongly to the predictions of modernization theory, reproducing at the individual level what the classical literature on the middle class leads us to expect: that the middle class is more supportive of democracy than the working class. Whether the middle class is employed in the state or private sector is unrelated to democracy support. Both groups are equally democratic. Tentatively transposing these relationships to the macrolevel, we would conclude with Easterly (2001) that a large middle class is good for democracy. The second set of results in figure 3.2, however, implies 13. I do not include country fixed effects in all models, because doing so creates an incidental parameters problem. While a logistic model with fixed effects can be estimated using the conditional likelihood, the incidental parameters problem makes it impossible to recover the intercepts or therefore compute substantive quantities of interest. Because the results do not differ substantively, I therefore rely on the model without fixed effects when calculating predicted probabilities of democracy support. 14. Appendix table a1 reports the logistic regression coefficients. 15. The results reported in figure 3.1 correspond with table a1 column 2. All results reported in this and subsequent figures are based on the observed variables approach described in Hanmer & Kalkan (2013). Briefly, the probability of success of the dependent variable is estimated based on the independent variables of interest, while all other variables are held at their empirical values for each sample case. The predicted probabilities are then calculated for each observation and the average is taken over all observations under the assumption that the sample is a good representation of the population.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 79 Middle class

Probability of being a ‘democrat’

0.30

Non-middle class

Middle class

0.25

0.20

0.15

Non-middle class

Private State Private State

0.10

0.05

0.00

figure 3.1. Middle class more supportive of democracy regardless of sector in post-communist democracies. The left panel of this figure shows the predicted probability of being a ‘democrat’ given that a respondent is middle class (squares) or non-middle class (circles). The second panel shows the predicted probability of being a ‘democrat’ for middle-class and non-middle-class respondents by sector of employment. The vertical bars (hash marks) are 95% (90%) confidence intervals. Data source: EBRD LiTS, 2006 (democracies only).

clearly that the democratic virtues of the middle classes under autocracy are conditional. The left panel of figure 3.2 shows that although being middle class is associated with more liberal regime preferences, the average effect of being middle class is small.16 To quantify the size of the effect, I calculate the predicted probability of democracy support for the middle class (squares) versus others (circles). While the difference in the predicted probability of democracy support for the middle and non-middle class is significant at the 0.001 level, it is only about 4 percentage points (with a 95 percent confidence interval of [0.020, 0.063]). 16. Although none of the EBRD surveys offer direct evidence on whether respondents’ conceptions of democracy differ with democratic experience, for such differences to be driving these results, they would have to be concentrated among the state middle class. It is unclear why democratic learning would disproportionately affect the state middle class and not other groups.

80 c h a p t e r 3 Non-middle class

Probability of being a ‘democrat’

0.30

Private

0.25 Middle class

0.20

0.15

Middle class

Non-middle class

Private

State

State

0.10

0.05

0.00

figure 3.2. Middle class in post-communist nondemocracies more likely to support democracy, only if employed outside the public sector. The left panel of this figure shows the predicted probability of being a ‘democrat’ given that a respondent is middle class (squares) or non-middle class (circles). The second panel shows the predicted probability of being a ‘democrat’ for middle-class and non-middle-class respondents by sector of employment. The vertical bars (hash marks) are 95% (90%) confidence intervals. Data source: EBRD LiTS, 2006 (nondemocracies only).

The right panel of figure 3.2 tests the proposition that support for democracy will be attenuated among a middle class whose economic opportunities and life chances are tied closely to a nondemocratic state. It shows that, under authoritarianism, the association between the middle class and liberal regime preferences is virtually zero for the state-employed middle class (solid square), whereas it is sizable and significant for those members of the middle class who are employed in the private sector (open square).17 In fact, even at 17. One potential concern is that state employees are more likely to lie about their regime preferences than are private sector workers and, hence, that social desirability bias explains these results. While several survey methods have been shown to elicit more honest responses to sensitive items than direct questioning (see, e.g., Rosenfeld, Imai & Shapiro 2016), none has been applied to the question of regime preferences in the countries in this study. Analyzing Kalinin’s (2013) list experiment on turnout in Russia by sector of employment, I find that state employees were no more likely to dissemble about their politically sensitive turnout decision than others. I further explore whether social desirability bias can account for these results by looking at current state employees below.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 81

the 0.1 level, we cannot reject the null hypothesis that the actual difference in democracy support between the public-sector middle class and non-middle class is zero. Non-middle-class respondents and state workers have about the same (low) probability of being democrats, just above 10 percent. At the same time, the association between being middle class and support for a liberal democratic regime is considerably stronger for those who are not employed by the state. Members of the middle class who work outside the public sector are about 8 percentage points more likely than the non-middle class to prefer liberal democracy. This difference is both statistically distinguishable from zero at the 0.001 level and substantively meaningful, as it represents more than a 70 percent increase in the probability of democracy support.18 These results imply that, under autocracy, the effect of socioeconomic status on democracy support is conditioned by an individual’s relationship to the state.

Robustness Checks These findings hold up to a number of changes in the scope of the middle class and the definition of democracy support. In this section, I also demonstrate that the state middle class’s reticence about democracy is not simply reticence about the free market nor is it synonymous with support for the status quo or a particular autocrat. I take up each of these tests in turn.19 First, as noted already, educational levels across the post-communist countries are high, higher than in many other developing regions. To probe both the robustness and generalizability of my findings, given that I define the middle class in part according to its educational credentials, I considered several alternative models. To begin, I relaxed the educational criteria I used to define the middle class by including those with specialized post-secondary training in addition to those with university and post-graduate degrees. This change makes very little difference. Second, I probed potential differences in 18. These results hold whether we focus on those who are not employed by the state or specifically on those who are employed in the private or foreign sectors. The predicted probabilities reported in the right panel of figure 3.2 look substantively the same: at the p < 0.05 level we can still distinguish the private-sector middle class’s greater likelihood of supporting democracy from the state-employed middle class, though the standard errors are less precisely estimated for this smaller group. 19. These tests are reported in detail in online appendix A.

82

chapter 3

the effect of education on the democratic attitudes of the middle class, distinguishing between the middle class educated in the Soviet period and the middle class that gained higher education after the collapse of communism. Given the ideological and explicitly antidemocratic character of the Soviet educational system and that the content and quality of education changed in the post-Soviet period, it makes sense to further investigate whether communist education is primarily responsible for the antidemocratic attitudes of the middle class. In short, the results imply that low support for democracy among the state middle class is not merely a product of Soviet education. Whether educated under communism or the product of post-Soviet higher education, the state middle class’s attitudes toward democracy are indistinguishable from those of the working class, in keeping with the main results. I next conducted a series of tests to further explore the effect of income, given that my primary measure of the middle class is sociological. While the main results reflect the net effect of middle-class status without accounting for income, adding a control for logged equivalized household expenditures per adult makes little difference.20 The coefficient on expenditures is positive and significant. However, its substantive effect is small. Moving from the 25th to the 75th percentile in expenditures increases the likelihood of democracy support by only 2.5 percentage points.21 Further, the inclusion of expenditures has virtually no effect on the coefficient for middle class, defined using educational and occupational criteria. These results imply that the independent effect of current income, above and beyond the effect of middle-class status that runs through human and social capital, is small. Moreover, if we substitute Ravallion’s (2010) income-based definition of the middle class for our own and interact it with state employment in a model with all of the same controls, the pattern is very similar. The results remain substantively the same with the inclusion of all controls and country fixed effects.22 20. Again, this measure is the preferred proxy in a region where income is widely underreported and underreporting is plausibly related to both class and sector as well as the outcome of interest. 21. One concern is that household expenditures may not be a good proxy for income among state workers if the non-monetary benefits of public employment reduce expenditures. To investigate, I subset the sample and looked only at the effect of household expenditures on democracy support among non-state employees. The effect size is very similar. All else equal, going from the 25th to 75th percentile of household expenditures increases democracy support only by about 1 percentage point. 22. Excluding retirees from the middle class, controlling for unemployment, and accounting for subjective class also do not change the results.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 83

Next, I tested the robustness of my dependent variable. To be sure that the findings are not contingent on my coding of democracy support, I first relax the definition to also include respondents who “agree somewhat” that liberal institutions are important for their country. Though the coefficient on state employment is close to zero, the interaction between state employment and middle class remains large, statistically significant, and negative.23 The results from this model, displayed in figure a1 of the appendix, are very similar to those reported above and imply that the state middle class is 25 percent less supportive of democracy than the private-sector middle class (p < 0.05). I also examined several other alternative measures of democracy support. The results of linear models in which democracy support is measured alternately using a conventional averaged index of the seven democratic institutions items and the belief that a political opposition is essential imply that the state middle class is about 5 to 10 percent less supportive of democracy than the private-sector middle class, though these differences are imprecisely estimated. I further find that public-sector employees are both more likely to prefer authoritarianism and to view the type of political regime governing their country as personally salient. Substantively, these results suggest that the state middle class is about 25 percent more likely to support authoritarianism than the private-sector middle class. The one departure in these findings is that state employees’ stronger preference for authoritarianism and conviction that regime type matters does not appear to be conditional on class.24 In each case the results of these additional tests lean in the expected direction, though they are not all statistically significant at conventional levels. Comparing across them, it is interesting to note that the more demanding the definition of democracy support, the clearer are the differences between the state and private-sector middle classes. To summarize, then, the contrast between state and private-sector middle class largely conforms to the theorized pattern across all of the alternative specifications of the dependent variable, though the results should be interpreted with some caution. Last, I show that my measure of democratic attitudes is not simply picking up support for a market economy, satisfaction with the status quo, or loyalty

23. One potential explanation for state employment’s diminished effect is that state employees may be more likely to express support for democracy than to fully embrace democratic institutions. This would be consistent with contemporary autocrats’ democratic rhetoric but lack of substantive democratic commitments. 24. In other words, all state employees are more likely than their private-sector counterparts to believe in the occasional necessity of authoritarianism.

84

chapter 3

to a particular autocrat. First, I added a variable that measures the preference for free markets, as much of the literature on democratic attitudes after the collapse of communism has found a connection between attitudes toward political democracy and a market economy (e.g., Duch 1993; Miller, Hesli & Reisinger 1994; Gibson 1996; Frye 2003).25 The main results are unchanged. While attitudes toward the market are positively and significantly associated with support for democracy, all other coefficients remain very similar.26 These results suggest that democracy and a market economy are intertwined in the minds of post-Soviet citizens but are not synonymous.27 Were views toward democracy and economic liberalism basically the same, the preference for free markets would account for more of the variation in the model.28 Second, I show that reticence about democracy is not merely a function of support for the incumbent autocrat or satisfaction with the status quo. Most importantly, the inclusion of these additional controls does not change the main results. In short, then, each of these tests again confirms that state employment has a substantively important independent effect on middle-class support for democracy.

Alternative Explanations: Selection Having shown a robust association between state careers and regime preferences, the following sections endeavor within the natural limitations of observational data to establish that it is causal. A key challenge is to untangle 25. Though see also Finifter & Mickiewicz (1992), Hough (1994), and McIntosh et al. (1994). 26. See the online appendix. 27. Both sets of attitudes could be driven by resistance to social change or individual risk tolerance. If the willingness to accept risk is also associated with public-sector employment, then the patterns of democracy support among the state middle class found above might not be due to the causal effect of state employment but to the underlying characteristics of state employees. The results in chapter 6 and in the following section on differential attrition suggest this is unlikely. If risk aversion were driving the results, we would expect risk-averse individuals to be both less democratic and remain in state employment longer than individuals with a higher tolerance for risk, all else equal. Figure 3.5 does not support this interpretation. Firmer evidence based on pretreatment measures, albeit limited to Ukraine due to data availability, is presented in chapter 6. 28. The Spearman correlation between the preference for free markets and democracy shows a modest interrelationship of ρs = .372. Meanwhile, the correlations between the preference for free markets and support for specific democratic institutions are tiny (ρs < .077 in all cases).

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 85

the effect of a state career itself from the factors that lead individuals to sort themselves into different career paths. In econometric terms, correctly identifying the effect of career choices on support for democracy requires a strategy to tackle reverse causation and selection bias (that is, the influence of an unmeasured variable on both career choice and the outcome of interest). I next describe these problems substantively in the context of my analysis and outline the strategy I use to address them. First, if people with authoritarian values seek out state employment, the two variables will be related, but not for the reasons I cite. The case in which people who value freedom and liberal democracy gravitate to the private sector poses a similar problem. In both cases, democratic values exert a simultaneous effect on employment choices. This effect would bias our estimates. Similarly, our estimates would be biased if some selection criteria affect both employment choices and attitudes toward democracy. For example, if being unsuccessful in the new market economy leads some to both dislike democracy and seek the security of public employment, again our estimates will be biased. This latent character type (being a post-Soviet “loser”) would be an omitted variable. Aversion to risk, which might lead to a preference for public employment and a reluctance to embrace democratization, would be another example. Insofar as these characteristics are a function of age, gender, or are proxied by low subjective well-being (or low social mobility), this problem is minimized by the controls. Nevertheless, I consider this issue in greater depth below. The usual caveats regarding cross-sectional survey data apply. In chapter 6, I actually measure pre-employment regime preferences and can investigate directly whether regime preferences or their correlates are drivers of a career in the state sector. However, comparable panel data in which respondents are reinterviewed over a period of years simply do not exist for such a large number of countries. The pseudo-panel nature of the LiTS employment data allows me to address these issues more fully than is typically possible with cross-sectional data, making the EBRD survey a very good testing ground for my argument and its alternatives. Not only does the LiTS survey tell us whether respondents work for the state in their current occupation, it tells us how their career has progressed since 1989, just before the collapse of communism. As such, I can identify those respondents who have spent more and less time in the state sector and those who have moved into and out of it. The survey also further divides state employees by type.

86

chapter 3

I exploit this variation both within the state sector and across time to test the plausibility of competing arguments and speak to the question of causal mechanisms. Subsequent sections show that neither selection nor other conceivable alternatives explain my results. Instead, the analysis implies that, where private sector alternatives are lacking, authoritarians benefit from the presence of a large public sector. u s i n g o ccupat i o na l va r i at i o n to test for selection I begin this analysis of mechanisms by examining differences across occupational groupings within the state sector. Some serve the state directly, working within the state administration at various levels. Others work in education and public health and are paid out of the state budget. In this section, I exploit the fact that some public employees are more likely than others to chose their careers on the basis of preexisting political values. Specifically, I expect that civil servants, persons in the military, and those employed in various security apparatuses are more likely to be politically motivated in their choice of career than are medical professionals. The fact that very few medical professionals in the post-Soviet countries make a career exclusively in the private sector further supports my identifying assumption, since those who wish to practice medicine are virtually required to do so within the state system.29 If the negative effect of state employment shown in figure 3.2 holds among medical professionals, this would suggest that state employment does, in fact, have an effect on regime preferences that is independent of selection. Figure 3.3 shows the results of the within-state-sector analysis. The figure displays the predicted probabilities of support for liberal democracy across all four major occupational groupings within the state sector: civil service, military, and security apparatuses; education; health; and those who work for the state in some other capacity (the largest group being manufacturing in state-owned enterprises). For comparison, the first bar gives the predicted probability of democracy support among the non-state middle class (open circle). The probability of democracy support among the non-state middle class is higher than the probability of democracy support among all occupational groups within the state sector. For education and other state sector occupations, these differences are significant at the 0.01 level, while for those 29. See the figures cited on page 41.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 87

0.30

Non-state middle class

State middle class

Probability of being a ‘democrat’

0.25

0.20

State admin, police, military

Health workers

Educators 0.15

Other

0.10 (n=446) 0.05

(n=350) (n=115) (n=74)

0.00

figure 3.3. State employment has a negative effect on support for democracy across all occupational categories. This figure compares the predicted probability of support for liberal democracy across four occupational groupings within the state sector: civil servants, military, and security apparatuses; educators; health workers; and those who work for the state in some other capacity (the largest group being manufacturing in state-owned enterprises). The predicted probability of democracy support among the non-state middle class is significantly higher and shown for comparison (open circle). The vertical bars (hash marks) are 95% (90%) confidence intervals. Data source: EBRD LiTS, 2006 (nondemocracies only).

who serve the state and for health workers these difference are less precisely estimated and significant only at the 0.1 level.30 Most importantly, no group within the state sector is significantly more likely than any other to prefer a liberal democratic regime. For selection to explain these results, medical professionals and teachers would have to be just as motivated by political considerations as civil servants and police in choosing state careers. This is not to say that selection plays no role in the choice 30. Note that these latter groups are numerically smaller in the sample.

88

chapter 3

of careers; however, these results suggest that state employment has an independent effect, and that this effect, rather than selection alone, better accords with empirical patterns. In subsequent sections, I present additional evidence that neither selection nor reverse causation can convincingly account for my results. u s i n g ca r e e r t r a j e ctor i e s to t e st alternative mechanisms I next test several different mechanisms by which state employment could lower the probability of democracy support, exploiting variation over time in respondents’ careers. Whereas typical cross-sectional data only provide evidence on current employment, here I am able to juxtapose different career trajectories in order to better understand the relative importance of selection effects, immediate material incentives, long-term dependence on the regime, and socialization which deepens over time. This analysis also helps to address the concern that social desirability bias could be driving the results. To sort out the mechanisms that link state-sector employment with political preferences, I regress democracy support on a dummy for current state employment (at the time of the survey in 2006), state career share, years in the state sector, the variable for middle class, and the controls.31 The results presented in figures 3.4 and 3.5 correspond with column 4 of appendix table a1. I first consider the effect of current state employment. More than half of those who had spent their entire careers in the state sector were still employed by the state at the time of the survey. As such, this variable should absorb much of the unique effect of being a public-sector “lifer,” if such an effect exists. Including the variable for current state employment also allows me to test whether response bias can plausibly explain the effect of a state career. If fear of being fired for expressing liberal views is driving stated democratic preferences, then we would expect current employment to be of principal importance in these regressions.32 31. Besides the controls in all other models, specifications with current state employment also include a dummy for Armenia. While Armenia contributes an unusually large number of observations to the group of non-current state employees, Armenian respondents are much more negative about democracy than others in the region, making it an outlier in these data. Together these two patterns induce correlation between negative views of democracy and nonstate employment that is otherwise absent in the data. 32. Online appendix table A.9 shows that this analysis extends to other excludable categories of dependence on the state.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 89

In fact, the coefficient on current state employment is negative as expected, but does not always reach conventional levels of significance (table a1, column 4). This implies that the association between state employment and the preference for an illiberal regime is not principally an artifact of social desirability bias. If fear of being fired were of primary importance, we would expect current employment to have greater substantive and statistical significance. Instead, current state employment explains little of the remaining variation after controlling for long-run dependence on the state.33 The results in figure 3.4 suggest instead that long-term dependence on the state has the most significant effect on the political values of state workers in the post-communist nondemocracies. To proxy long-term dependence on the state, I use the share of an individual’s post-Soviet career spent in the public sector. This variable captures well an individual’s demonstrated access to private sector alternatives. While those with a state career share of less than one but greater than zero have also been employed outside the state, those who have spent their entire career in the state sector have no demonstrated ability to access private sector alternatives. I expect those with larger career shares in the public sector, who likely possess the least access to private-sector alternatives, to be the least democratic, all else equal. Consistent with this hypothesis, figure 3.4 shows that as the share of one’s career spent in the state sector increases, support for democracy decreases. Substantively, the probability of preferring a liberal democratic regime if you have never been employed by the state is about 7 percentage points greater than if you have only had state employment. This difference is significant at the 0.05 level. These results further help to rule out an alternative account based on selection. Specifically, the downward sloping line in figure 3.4 is inconsistent with a selection story that emphasizes initial sorting by democratic preference type. If that were the case, we would not expect to see any additional change after the initial sorting took place. Online appendix A.4.1 offers further evidence from Ukraine and Georgia that political self-selection following regime transitions cannot account for the main results.

33. The first four columns of online appendix table A.8 report coefficients from models with one state employment variable at a time. If fear of being fired or social desirability bias were most important, moving from column 2 in table A.8 to column 4 of table A.2, we would expect the coefficient on state career share to get smaller and current employment to be large and significant. This is not the case.

90 c h a p t e r 3

Probability of being a ‘democrat’

0.20

0.15

0.10

0.05 0.0

0.2

0.4

0.6

0.8

1.0

Share of career in state sector figure 3.4. As post-Soviet career share in state employment increases, support for democracy decreases. This figure plots the predicted probability of being a ‘democrat’ given the share of one’s career spent in the state sector since 1991, controlling for class, gender, age, and age squared. The dashed lines are 95% confidence intervals. Data source: EBRD LiTS, 2006 (nondemocracies only).

Yet we might still be concerned that those who are more democratic leave state employment sooner than those who are more authoritarian (i.e., that differential attrition could explain these results). For example, imagine two individuals who have both spent their entire careers in the state sector. Now imagine that one has spent fifteen years employed in the state sector, while the other has spent only one year. An account based on differential attrition would predict that the person who has has been employed longer by the state would be less likely to be a democrat than the person who has been employed for only one year. Evidence to the contrary is presented in figure 3.5.34 In fact, figure 3.5 shows that additional years of employment are not significantly related to sorting out of the public sector. 34. Another approach would be to obtain a pretreatment measure of salient political values—that is, a measure of regime preferences before an individual embarks on his/her career. This is the approach I take in chapters 5 and 6. As noted already, no such measure is available for a cross-national sample.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 91

Probability of being a ‘democrat’

0.20

0.15

0.10

0.05 0

5

10

15

Years in state sector figure 3.5. The effect of additional years of state employment is insignificant. This figure plots the effect of additional years of state employment on the predicted probability of democracy support based on a logistic regression that also includes state career share, class, gender, age, and age squared. The effect of additional years of state employment is insignificant, a finding that is inconsistent with cumulative, long-term socialization. Data source: EBRD LiTS, 2006 (nondemocracies only).

Finally, the insignificant effect of additional years of state employment, evident in figure 3.5, casts doubt on alternative explanations that focus on cumulative, long-term socialization or indoctrination through repeated exposure (e.g., Pop-Eleches & Tucker 2017). Only socialization that occurs very quickly, does not deepen appreciably over time, and is relatively sticky would be consistent with this evidence.35 In sum, the findings suggest that neither pure selection of less democratic types into state employment nor certain versions of long-term socialization likely account for these findings. That said, separating employment trajectories from factors like age, skills, and education is a difficult task. Although we cannot entirely rule out these alternatives given the nature of the data, the fact 35. I also tried the log(1 + number of years in the state sector) to allow additional years of state employment at the beginning of a state career to matter more than additional years later in the career, but it too was insignificant in the model.

92

chapter 3

that several pieces of evidence all point to a similar conclusion lends credence to the argument that a large public sector limits support for democracy, especially among the middle class that has long been expected to champion democratization.

Placebo Test Next I investigate whether my core result holds up in a “placebo test.” If the argument I have laid out is correct, the negative effect of state employment on democracy support should be present in autocracies but not in democracies. Dependence on a democratic regime should not make public-sector employees less democratic. If anything, we would expect the opposite. All the same forces—long-run dependence on the state and fear of being fired—plus an interest in maintaining a nonpolitical basis for promotion and retention36 suggest that employees of a democratic regime should be at least as likely, if not more likely than others, to prefer democracy. Fortunately, the LiTS sample includes eighteen democracies in central and Eastern Europe that share the post-Soviet states’ history of nearly universal state employment. A comparison of the two sets of cases allows us to investigate whether shared communist-era legacies of nearly universal state employment are a plausible source of confounding effects. Figure 3.6 shows the results of this test.37 Each point represents the average marginal effect of a change in the independent variable from no state employment to a career spent in the state sector (that is, from zero to one with all other variables held at their observed values). Such a shift from low to high is substantively interesting since the distribution of state employment is bimodal: a majority in the post-communist countries have spent their entire career in either the state or private sector. The top bar (nondemocracies) essentially replicates the finding in figure 3.4 using the full twenty-seven-country dataset, interacting the key independent variables with an indicator for democratic countries. Here again the estimated effect of state employment in the nondemocracies is negative and significant at the 0.05 level. In the democracies (bottom bar), by contrast, as we observed at the outset, sector of employment has a negligible 36. Greater job security and a nonpolitical basis for hiring, promotion, and retention tend, of course, to accompany greater emphasis on the rule of law and public sector development under democracy. 37. The logistic regression results appear in column 5 of table a1.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 93

Nondemocracies

Democracies

−0.15

−0.10

−0.05

0.00

0.05

0.10

0.15

Difference in the probability of democracy support for the state and non-state middle classes

figure 3.6. The effect of state employment on regime preferences is negative in the nondemocracies, weakly positive in the democracies. This figure shows the predicted change in the probability of being a ‘democrat’ as career share in state employment rises from 0 to 1, conditional on the respondent being middle class. The hash marks and horizontal bars denote the 90% and 95% confidence intervals. In democracies, the difference is slightly positive but insignificant. In the nondemocracies, the difference is negative and significant at the 0.05 level. Data source: EBRD LiTS, 2006 (all post-communist countries).

effect on support for democracy.38 In sum, consistent with expectations, the placebo test distinguishes between the effect of state employment in the nondemocracies and democracies (this difference is significant at the 0.01 level) and turns up no evidence that state employment under a democratic regime is associated with weaker support for democracy.

Other Alternative Mechanisms e co n o m i c p e r f o r m a n ce Before concluding, I consider three substantive alternatives, but find little evidence to support them. The first alternative explanation is that state workers have less favorable views of democracy because they hold negative perceptions of their country’s economic performance since independence 38. Of course, not all of the Central and East European democracies have been democracies over the entire post-Soviet period (contrast Serbia, Croatia, and Romania with Slovenia, for example). Even if we exclude the East European countries that transitioned to democracy only in the late 1990s or early 2000s, state employment is still weakly related to support for liberal democracy.

94 c h a p t e r 3

and are more likely to associate economic insecurity with the transition to democracy. Indeed, state workers, along with women and pensioners, are often characterized as losers in the post-communist transition, while privatesector workers, along with the young and well-educated, are often viewed as winners.39 Public-sector wage arrears were widespread in the former Soviet Union during the 1990s. Official salaries remain low for many teachers, doctors, and state-sector professionals to this day (see, e.g., Richter 2006; Sharunina 2013). If the state middle class blames democracy for downward social mobility and loss of the social status that was associated with their occupations during the Soviet period, this, rather than state dependency, could explain their attitudes toward democracy. Additional analyses, however, suggest that the state middle class’s lagging support for democracy is not due to negative economic experiences since the collapse of communism.40 While post-communist citizens may have once blamed democracy for declining standards of living, today they increasingly fault their nondemocratic governments for the failure to improve conditions. Indeed, I find that negative assessments of economic performance and the perception of downward social mobility are actually associated with less support for the status quo—that is, with more support for democracy.41 The estimated effect of public employment for the middle class, meanwhile, remains surprisingly stable with the addition of these economic controls, suggesting that negative economic experience cannot account for the state middle class’s lower support for democracy. In fact, the evidence implies that, on average, state employment actually improves individuals’ perceptions of their own economic opportunities and life chances. Although perceived social mobility is exactly equal for the state and non-state middle classes,42 state employees are more satisfied with their lives than others (53.2% vs. 47.1%, p < 0.001). State employees are also more likely than others to say that they have done better than both their high school

39. As Tucker, Pacek & Berinsky (2002, 559) note, this tendency to ascribe winner/loser status on the basis of demographic characteristics may not be consistent with individual perceptions. 40. These analyses are reported in detail in online appendix A.4.2. 41. See the online appendix. 42. The median score for both groups is −1, meaning a slight worsening of relative social position. The median score for non-middle-class respondents, regardless of sector of employment, was somewhat lower: −2.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 95

classmates (41.2% vs. 35.5%, p < 0.001) and the colleagues they had around 1989 (33.5% vs. 27.6%, p < 0.001).43 r e d i st r i bu t i v e p r e f e r e n ce s A second set of alternative explanations emphasizes the redistributive consequences of democratization. Two redistributive accounts are plausible. First, political economy models of democratization imply that the farther an individual’s income is below the median, the greater is his/her demand for redistribution. These theories further expect redistribution to be higher under democracies. It thus follows that those with below-median incomes should be more likely to support it. Perhaps the state middle class has incomes that are higher than those of the private-sector middle class, making them less favorable toward redistribution and democratization? In fact, both the state and private-sector middle classes in my sample have median incomes slightly above the overall median of the distribution.44 Moreover, the median income of the state middle class is slightly below the median income of the private-sector middle class, making the state middle class’s comparatively lower support for democracy puzzling in light of redistributive theories. Given the striking rise in inequality as the region democratized, another possibility is that post-Soviet citizens expect the triumph of democratic forces to be accompanied by less rather than more redistribution, contrary to the assumptions of canonical redistributive models (Karakoc 2018). Whereas this argument would imply that state workers are less favorable toward democracy because they are more sensitive than others to inequality, redistributive theories would expect the state middle class to hold less favorable views of democracy for precisely the opposite reason: because they are less sensitive to inequality. The survey asks respondents whether they agree or disagree that the gap between the rich and poor in their country should be reduced. There is no relationship between views on income inequality and sector of employment. Concern about inequality is shared widely across lines of class and employment sector. Empirically, middle-class respondents are no less likely to 43. It is possible that state workers may be more risk-averse and therefore more satisfied with their own lives. 44. Using per-adult equalized household expenditures as the best available proxy for income, the median for the state-employed middle class is approximately the 65th percentile of the overall distribution, while the median for the private-sector middle class is just above the 70th percentile.

96 c h a p t e r 3

favor redistribution than those below them. Employees in the state sector are no more likely to favor redistribution than those outside of it. In sum, it seems unlikely that preferences over redistribution and attitudes toward inequality could explain the pattern of democracy support found in this chapter. co m mun i st s o ci a l i z at ion A third potential concern is that the apparent effect of state employment actually reflects Soviet-era socialization. In most surveys it is impossible to distinguish the impact of Soviet-era workforce socialization from that of the post-Soviet era. The LiTS survey, however, offers a unique opportunity to disentangle the effect of post-Soviet careers from those of the Soviet era. Recent scholarship suggests that communist socialization has enduring effects (e.g., Pop-Eleches & Tucker 2017; Beissinger & Kotkin 2014; Pop-Eleches 2007). If post-Soviet state workers were more imprinted by the communist system than non-state workers, and those effects persist, state workers may be less favorable to democracy in the present period. Following existing literature, exposure to communism can be thought of as a cumulative experience (e.g., Pop-Eleches & Tucker 2017). Hence, one strategy for dealing with the potentially confounding effects of communist socialization is to control for age, as I have done throughout. Indeed, age is almost perfectly collinear in this sample of post-Soviet nondemocracies with communist socialization, when the latter is measured as years under a communist regime, so separating the two with a single survey for each country is not possible. Across virtually all of the models previously reported, age is highly statistically significant. The predicted probability of democracy support rises somewhat until about age 40, then steadily declines.45 These results imply slightly more democratic attitudes among those who were young adults at the time of perestroika, less support among those who grew up under declining pluralism after the mid-1990s, and the least support among those who lived longest under communism. Further analyses show that the effect of state employment on the middle class also cannot be explained by coming of age politically under communism, living through Stalinism, experiencing perestroika, or other communist-era experiences like party membership or a family history of repression.46 In sum, these results imply that being a 45. The results are plotted in online appendix figure A.8. 46. These analyses are reported in detail in online appendix A.4.3.

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 97

state employee in a post-Soviet autocracy has an effect on attitudes toward democracy that is not driven by differential exposure to communist socialization. Communist experience matters too. But state employment after the fall of communism matters independently.47

Conclusion The cross-national findings in this chapter raise questions about the longstanding assumption that a large middle class aids democratization. Instead, they suggest that autocrats benefit from control over public employment and the economic incentives it creates for the middle classes to remain reticent about democracy. Using multicountry survey data from the European Bank for Reconstruction and Development’s (EBRD) Life in Transition Survey (LiTS), I find that across the post-communist nondemocracies the state middle class is generally unsupportive of democracy and no more supportive of democracy than the working class. I also find that as the share of one’s post-Soviet career spent in the state sector increases and private-sector alternatives decrease, support for democracy declines—a pattern I attribute to the importance of long-run dependence on the state. My results are robust to changes in the coding of middle-class and alternative definitions of democracy support. They do not hinge on using higher education (which is especially widespread across the region) to define the middle class, nor are they unique to the middle class educated under communism. In fact, they hold even when the middle class is defined in terms of accepted income-based criteria (Ravallion 2010). I also assess a number of competing explanations, showing that neither age nor generation, views of the free market, communist era experience, or place in the income distribution can explain the effect of state dependency. Moreover, within the natural limitations of cross-national survey data, a variety of tests imply that the effect of state employment is not primarily due to processes that affect the types of individuals found in the state sector (that is, to selection), but to the factors I identify. I control for known, observable selection criteria, show that initial sorting is insufficient to explain the effect of state careers on democracy support, and demonstrate that differential 47. In one final test to be sure that no omitted factor from the communist era could account for my findings, I consider communist party membership and family history of repression by the Soviet regime in online appendix section A.4.3.

98

chapter 3

attrition is inconsistent with observed patterns of democratic attitudes. I also compare across different parts of the state sector where political motivations are more and less likely to affect career choice, yielding further evidence against selection. In the absence of true panel data, the ability to analyze variation in post-communist careers and career trajectories increases our confidence that state employment influences political preferences and not the other way round. Finally, in the democratic post-communist countries, where we would not expect state employment to weaken support for democracy, indeed, no such relationship is evident. This gives us greater confidence that the results reflect the effect of post-Soviet careers under post-independence authoritarian regimes—rather than some other shared legacy of universal state employment under communism. The region’s high educational levels are in part another shared legacy of communism. Could education’s role in defining the middle class make these findings specific to post-communism? The fact that the state middle class that attended higher education after independence looks very much like the state middle class educated under communism suggests that these findings are probably not unique to the nature of the Soviet educational system. That the results also hold for younger respondents with less exposure to communist education similarly implies the argument’s portability across different educational contexts. Of course, some caution is warranted. Educational systems in the former Soviet countries did not change overnight after independence. Legacies of communist education persist. Still, I found similar patterns for those educated under a highly ideological system and at a time when many of the region’s institutions of higher education sought to instill democratic values—suggesting that the findings may well generalize to other settings. What is clear from this chapter is that the book’s central argument holds explanatory power across a wide range of post-communist cases and different autocratic institutions. At the same time, this chapter’s findings suggest several areas for further investigation, which I pursue in subsequent chapters. First, do attitudes translate into action? Might it be the case that state employees appear to be more loyal, but would be just as likely as others to take part in a democratic uprising? In chapter 4, I examine the link between attitudes and politically consequential behaviors, drawing on surveys of protesters in Russia to assess the impact of state selective incentives on pro-democracy coalitions. Chapter 6 further probes the relationship between preferences over democracy and two types of political action: participation in a civic revolution and

t h e p o s t- c o m m u n i s t m i d d l e c l a s s e s 99

voting. Together these analyses suggest that the pattern of preferences this chapter identifies also line up with political behavior. A second question for further investigation is who chooses a career in the state sector? Do state careers have an independent impact on political preferences, as I have tried to show here, or does the state middle class in fact attract more authoritarian personalities? Chapters 5 and 6 take up this challenge. Chapter 5 shows that networks and nepotism better explain the choice of career than political preferences. Chapter 6 provides more definitive evidence on causal relationships that are difficult to establish in crosssectional data. Using a panel survey consisting of reinterviews with the same respondents, I measure the regime preferences of new labor market entrants before they enter the workforce and after. The findings provide additional evidence that career experiences matter independently of the factors that affect career choices. Finally, this chapter has dealt with the impact of public-sector policies on the loyalty of the middle class without addressing their intentionality. Chapter 7 considers the creation of an autocratic middle class as an explicit regime strategy. It also probes the importance of newer forms of state economic engagement and indirect influence that cross-national data do not capture. In conclusion, the evidence in this chapter suggests that there is no necessary connection between high levels of human and social capital and the preference for a liberal regime. I have shown that constituencies within the middle class whose livelihoods and status depend on public employment are less supportive of democratic institutions than others with similar educational and occupational profiles. This evidence illustrates clearly how a large public sector contributes to authoritarian survival. It also speaks directly, at the micro-level, to one of political science’s abiding puzzles: development without democratization. Economic opportunities and employment security in the public sector may in fact delay democratization by helping to sustain a middle class that is divided on democracy.

4 Rethinking the Middle-Class Protest Paradigm

amid accusations of fraud in Russia’s December 2011 parliamentary elections, a series of demonstrations, the largest since the collapse of the Soviet Union, took place in Moscow. The first protest on the day after the election was small. But by the end of the month, the demonstrators numbered more than 100,000. The regime responded by mobilizing its supporters. The largest pro-regime rallies were held in the lead-up to the presidential election in March and around the national-patriotic holiday on May 9, celebrating victory in World War II. These events, too, were massive, with many participants bussed to the capital from the regions. The opposition movement ebbed over the spring, holding a series of smaller protests around the presidential election, with participants in the tens of thousands. Then, again, in June, an opposition protest on Prospekt Sakharov, named for the Soviet dissident, attracted crowds as large as those that had gathered in December. Russia’s post-election protests were quickly characterized as a middle-class phenomenon. Consistent with a vast political science literature, Moscow’s middle class was cast as a force for democratization. Urban civic protests and revolutions, in particular, have been seen as a middle-class phenomenon, belonging “to the repertoire of the educated, the informed, the connected, and the relatively well-off ” (Welzel 2013, 217). Thus, Acemoglu & Robinson (2006, 39) assert that “almost all revolutionary movements were led by middle class actors,” while Huntington (1991, 67) argues that, “In virtually every county [of the third wave] the most active supporters of democratization came from the urban middle class.” In the days following the initial demonstrations, the Financial Times ran a story under the headline, “Russia’s Middle Class Finds its Feet” (December 100

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 101

12, 2011). The New York Times declared, “Boosted by Putin, Russia’s Middle Class Turns on Him” (December 11, 2011). Articles highlighting the protesters’ middle-class credentials also appeared in the Russian press.1 Seeking to capitalize on popular resentment of Moscow’s newly affluent, the Kremlin reinforced this image. Not only were journalists drawn to the narrative of middle-class revolt, this view also found resonance with scholars of the region. Krastev & Holmes (2012, 41), for example, write that Putin’s decision to return to the presidency sparked protest by “offending Russia’s middle class.” As I endeavor to show in this chapter, these characterizations of the 2011–2012 protests distract from crucial heterogeneity within Russia’s middle class. By failing to account for cleavages within the middle class, we miss an important dynamic in bottomup processes of democratization. This dynamic has been largely overlooked in both popular accounts and the existing scholarly literature. The Russian case typifies the type of middle class that develops under autocratic state institutions and extensive state economic engagement. The majority of new entrants to Russia’s middle class over recent years are professionals paid out of the state budget and civil servants (Russian Academy of Sciences 2014; Maleva & Ovcharova 2009; Avraamova 2008). The proportion of state workers in Russia’s middle class is approximately one-half to two-thirds, while the overall share employed in the public-sector is about four in ten (Russian Academy of Sciences 2014, 30). In this regard, Russia is not unique. In the Middle East, Africa, Latin America, and other states of the former Soviet Union large public sectors have shored up nondemocratic politics (e.g., Hertog 2013; Diamond 1987; Weyland 1996; Gervasoni 2010; Oliveros 2016; Darden 2008). Resource states often have large public sectors, but such conditions also exist in states that do not depend on extractive industries. While the previous chapter focused on democratic attitudes, this chapter examines the effect of middle-class state dependence on mobilizational potential. In it, I investigate the most consequential action that individuals can undertake to influence democratization: protest. The wave of demonstrations beginning in late 2011 after Russia’s parliamentary election offers an ideal opportunity in which to consider the constituencies and coalitions that are likely to challenge illiberal regimes. Further, since preferences cannot be 1. See, e.g., Makarkin (2011) or Levinson’s (2012) “It’s not the middle class—it’s everyone” for a contrarian view.

102

chapter 4

inferred directly from protest behavior, I also analyze the democratic commitments of middle-class protest participants. As recent protests around the world demonstrate, better understanding citizen preferences under autocracy is critically important for grasping the sources of popular pressure for political change (Teorell 2010; Ulfelder 2005). Yet existing scholarship, including canonical works, offers virtually no concrete micro-level evidence on the preferences and actions of the middle class vis-à-vis democratization.2 Protests afford a uniquely tractable and politically salient opportunity for filling this gap. With that aim, I leverage detailed data on protest participants and a novel empirical strategy. This approach is widely applicable to the study of contentious collective action across a variety of settings. More specifically, I employ choice-based sampling methods, which allow me to combine protest survey data with representative data on the population from which protesters were recruited. The case-control method I present—a variation on designs commonly used in epidemiological studies of rare disease but ignored to date in political science research—allows me to estimate the population prevalence of protest for groups with given characteristics. This approach improves on a range of studies that lack individual data on protesters, rely on reported rather than actual participation, are unable to compare protesters to nonparticipants, or are prevented by their small sample size from reaching reliable conclusions about protest subgroups. An additional advantage is that I need not, and do not, assume that all protesters were democrats or that participation in opposition demonstrations had uniform ideological content for all who turned out. The preceding cross-national analysis showed that when the middle class emerges from the public-sector of an authoritarian state, it is less likely to value a democratic regime. Here I extend the argument and empirical analysis to a concrete case study of protest behavior in Russia, using an unusual series of individual-level surveys from four demonstrations, detailed population data, and a novel research design. This design overcomes a variety of inferential problems in the study of protest, allowing me to test the the project’s central contentions against actual political behavior. Making this connection to 2. Exceptions include studies by Tsai (2005), Chen (2013), and Gontmakher & Ross (2015), though these studies do not directly examine the formation of pro-democracy protest coalitions.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 103

actual political behavior is important: while the survey-measured attitudes of the state middle class are less democratic, it is possible that when tested, during moments of real contention, the state middle class would be as likely as others to join a democratic uprising. The evidence in this chapter suggests otherwise. To preview the results, I find that the prospects for mobilized democratic transition hinge considerably on the middle classes’ degree of state dependency. First, I show that the fastest growing segments of Russia’s middle class, professionals employed in the public sector, were significantly less likely than the private-sector middle class to mobilize. Controlling for a variety of other factors including ideology, I estimate that state dependency reduced the likelihood of protest by more than 25 percent among the middle class and 50 percent among the non-middle class. Further, I find that state employees who did protest were less likely to do so as part of the pro-democracy coalition. State-sector protesters were less likely to use the politicized language of rights and freedoms than their private-sector counterparts. They were also more likely to value stable development and standard of living over political liberty and regime change. For state-dependent middle-class groups, protest was not about democratic transformation, but securing a better deal from the regime. I trace these patterns of participation to the interaction of individual resources and state selective incentives. In particular, I find support for two mechanisms to explain patterns of participation: negative incentives such as the threat of dismissal and positive inducements grounded in differing economic and social protections. I also show that this explanation better accounts for variation in the propensity to join opposition protests than its leading alternative: selection, or the sorting of individuals into different careers based on preexisting political orientations. These findings highlight the importance for regime survival of co-optation that takes place in public-sector workplaces, beyond formal political institutions like parties, parliaments, and elections. Whereas most studies focus on co-optation of either elites or the poor (e.g., Gandhi & Przeworski 2007; Magaloni 2006; Medina & Stokes 2007; Calvo & Murillo 2004), capitalists or labor (Bellin 2000), this study shows the significance of co-opting the middle class in terms of protest capacity when autocratic regimes are challenged. Its findings also hold broader lessons for our understanding of the middle class and democratization. The fact that middle-class state employees

104

chapter 4

protested in smaller numbers, and that those who did were less likely to do so under the banner of democracy, implies that how growth of the middle class is achieved will affect bottom-up processes of democratization, through both attitudes and actions. Finally, these findings reveal that ignoring key sources of preference heterogeneity within the middle classes obscures an important micro-level mechanism behind development without democratization.

The Middle Class and Mobilized Contention The existing literature suggests a pivotal role for the middle class in mass protests against illiberal regimes. Indeed, middle-class groups have been credited with supplying crucial support for the popular mobilizations that precipitated the overthrow of Suchinda in Thailand, the fall of Chun Doo-hwan in South Korea, and the removal of Marcos in the Philippines. Yet the middle classes’ role in pro-democracy movements has varied across time and space (e.g., Rueschemeyer, Stephens & Stephens 1992; Koo 1991; Jones 1998; Shin 1999), and there has been little agreement on the sources of this variation.3 Conceptual, theoretical, and empirical issues have all impeded progress on this agenda. First, as discussed in chapter 2 the term “middle class” carries many meanings. For some, middle class is a normative category embodying the democratic values and participatory ethic that this study aims to explain. For others, the middle class is defined by its place in the income distribution (e.g., Boix 2003; Acemoglu & Robinson 2006). In this chapter, I continue to define the middle class as university-educated white-collar strata.4 In keeping with neo-Weberian approaches I focus on educational and occupational resources and qualifications, those that the literature has long identified with more democratic values and greater mobilizational potential (Lipset 1959; Dahl 1971). Following these arguments and recent theories of democratization 3. Beyond these cases of democratic revolution, this book’s argument also sheds light on other instances of civic protest, for example in Latin America, where neither the Kirchner loyalist middle class in Argentina nor the PT loyalist middle class in Brazil lent their support to recent anti-government demonstrations against abuse of power and official corruption. 4. Although income is not directly measured in this chapter, the occupational categories represented among the middle class generally have incomes above the median. As such, redistributive theories would expect this group to act like elites and oppose democratization.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 105

(e.g., Ansell & Samuels 2014), I employ a sociological, rather than exclusively income-based, definition of the middle class. Second, few studies have had at their disposal extensive empirical evidence on the demographic characteristics and political preferences of participants in popular uprisings against authoritarian regimes. Without such evidence, the existing literature tends to homogenize the middle class—finding either that it stood with the state or against it (Huntington 1991; Acemoglu & Robinson 2006). It also tends to oversimplify middle-class actors’ prodemocratic orientation. In short, the literature on democratization has not paid sufficient attention to the fact that autocracies can co-opt citizens by providing them economic opportunities and avenues of upward mobility in the public-sector, potentially undermining the link between the middle class and democratization. Indeed, we have surprisingly little systematic behavioral evidence to support our many assumptions and assertions about the role of the middle class in mobilized contention. I address this gap using detailed individual-level data to examine how state-led development undermines potential middle-class coalitions in support of democratization. I also consider why these splits in the middle class arise. The explanation I develop focuses on regimes’ use of selective incentives to secure a middle class that supports autocracy from among employees in state-dependent sectors.5

Empirical Expectations Authoritarian regimes face two fundamental challenges: maintaining elite unity and deterring mobilization from below (Svolik 2012). This book focuses on the latter task and on how authoritarian regimes maintain social order through the management of economic self-interest. Scholars have long argued that autocrats provide targeted rents to cultivate political loyalty and employ negative sanctioning to stem dissent. Public-sector jobs are themselves a benefit; but they also provide access to networks, information, and resources that can be leveraged for private gain. This chapter provides evidence for the argument in chapter 2 as it applies to mobilizational potential. That argument, succinctly stated, is that regimes use public-sector employment to 5. As Hicken’s (2011, 304) review notes, the size of the public-sector is the most commonly used proxy for clientelism.

106

chapter 4

neutralize potential middle-class opposition by offering perks to encourage loyalty or denying benefits to those who participate in pro-democracy demonstrations.6 In general, this argument suggests more middle- than non-middle-class participation in anti-regime mobilization, since all else equal, buying off the middle class is more expensive. It also implies that the state will be most effective at demobilizing opposition and countermobilizing support among social groups that depend on it for employment, status, and life chances.7 If the argument is correct, we would expect to find evidence consistent with the following propositions. First, the argument implies that middle-class mobilization will exhibit significant sectoral differentiation, with state workers less likely to participate in anti-regime demonstrations. The first hypothesis below implies that a state’s ability to co-opt the middle class will constrain the size of anti-regime protest by limiting participation by the state-dependent middle class. Three factors potentially account for this hypothesized pattern of protest participation: positive and negative inducements, grievances, and differences in social capital. I briefly discuss each in turn together with several additional observable implications. First, fear of being fired is perhaps the most straightforward explanation for state employees’ lower protest participation. The desire to avoid retaliation should be especially acute where exit options are limited. At the same time, if public-sector careers provide excludable economic and social protections along with access to networks, information, and resources that can be used to generate additional rents, we would expect these perks and privileges to increase satisfaction. If such carrots matter, and not only sticks, then public-sector employees should perceive themselves as more secure and less vulnerable than their private-sector counterparts. Moreover, we would expect the power of these selective incentives to vary with the opportunity costs of forgoing state benefits. These costs are determined not only by an individual’s endowments, but also by available alternatives. All else equal, workers who have fewer exit options and are more vulnerable will be less likely to join opposition protests. Empirically, 6. According to Olson’s (1965, 51) classic definition, selective incentives include both positive and negative inducements that are by their nature excludable. I use selective incentives to refer to both carrots and sticks wielded by the regime. 7. For a similar perspective, see Lee & Zhang (2013) on how Chinese civil servants are mobilized against protest and in support of the regime.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 107

I expect private-sector alternatives to covary with experience, proxied by age/years of employment. Alternatives should increase with experience at the beginning of one’s career, before declining through the remaining years of economic activity. In the post-Soviet context, where generational differences in both skills and mentality are substantial, alternatives should peak earlier than in non-transitional economies. At the end of one’s career, and to a lesser degree at the very beginning, alternatives are likely to be limited, while in the mid-ranges of experience, alternatives are likely to peak. Retirement, in turn, marks a sharp decline in the relevance of exit options and the power of selective incentives. Where a state pension is a universal and non-excludable benefit, the availability of alternatives loses salience upon reaching pension age. Taken together, then, I expect the power of selective incentives wielded by the state to be positively related to age, with a sharp drop-off at retirement, as both the risk of losing one’s job and the availability of alternatives become irrelevant. These patterns would constitute further evidence for the importance of selective incentives. Second, public-sector privileges may simultaneously foster resentment among those who are excluded from special treatment. Insofar as grievances motivate protest (Gurr 1970), private-sector workers may be more likely to demonstrate. A third possibility is that private-sector professionals compensate for their exclusion from state-sponsored privileges through dense networks of friends and acquaintances. These networks could reflect differences in social capital that would increase private-sector workers’ propensity to be recruited as protest participants. I next turn from participation to protesters’ goals and motivations. If public-sector workers see future benefits as tied to regime continuity, I expect that they will be not only less likely to participate but also more likely to favor the status quo and reject the risky prospect of democratization. Accordingly, I anticipate that this group will be less likely to join pro-democracy coalitions of non-system regime opponents than others who are more insulated from government incentives. This yields the following proposition: Protest coalitions will exhibit significant sectoral differentiation, with state workers less likely to join democratic forces. In sum, the second hypothesis implies that a state’s ability to co-opt the middle class will also constrain the size of any prodemocracy coalition. Testing this proposition requires additional information about protesters’ political orientations. Besides being less likely to self-identify as democrats, if the argument I have laid out is correct and protesters in state-dependent sectors do indeed have

108

chapter 4

table 4.1. Summary of Hypotheses and Observable Implications H1

H2

Middle class mobilization will exhibit significant sectoral differentiation, with statesector workers less likely to participate in anti-regime protests. A Availability of Alternatives • Those state workers who have the fewest (most) exit options will be least (most) likely to participate in opposition protests. B Negative inducements • State employees will feel vulnerable to being fired. C Positive inducements • State employees will be more satisfied with their economic and social protections and less likely to participate in opposition protests than their private-sector counterparts. Protest coalitions will exhibit significant sectoral differentiation, with state-sector workers less likely to join democratic forces. D Values • State employees will be less likely to advocate regime transition and more likely to support the “pocket” opposition. • State employees will be less likely to express grievances in the politicized language of rights and freedoms. • State employees will be more likely to value stable development and standard of living over political freedoms and regime change.

weaker democratic commitments, two additional patterns should be evident. First, they should be less likely to advocate regime transition and more likely to support the regime’s so-called “pocket” or “loyal” opposition (i.e., parties that collude openly with the regime). Second, they should be less likely than their private-sector counterparts to use the politicized language of rights and freedoms, which might endanger their benefit streams, and more likely to value stable development and economic well-being. Finally, examining both attitudes and actions is important for one very simple reason. Popular uprisings have often resulted not in democracy but elite rotation. While groups with distinct political preferences may partner to pressure the current government, such divisions spell trouble for the formation of a lasting and coherent pro-democracy coalition. Distinguishing democratic attitudes from oppositional action thus deepens our understanding of the prospects for stable democratic transition.

Protest and the Russian Middle Class Economic modernization became the explicit preoccupation of the Russian regime during Vladimir Putin’s first term in office, a formula for recovery from Russia’s turbulent free market transition during the 1990s. Buoyed by global

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 109

commodity prices, between 2000 and 2007, the Russian economy grew at a rate unprecedented in the post-Soviet period. Real wages increased 2.5 times, while the Russian economy added 3.3 million jobs. At the same time, Russia experienced an “educational boom.” The share of the workforce with higher education rose from 21.6 percent in 2000 to 25.6 percent by 2007 (Maleva & Ovcharova 2009). In line with these changes, the Russian state statistics service (Rosstat) reports that more than a third of Russian jobs are now white collar. Over the same period, state careers became one of the clearest pathways to the middle class. Starting in 2001, the already small share of income from entrepreneurial activities in the Russian economy began to shrink (Ovcharova 2012, 30). Meanwhile, the fastest-growing segments of Russia’s middle class became state officials, regional civil servants, so-called siloviki (law enforcement, military, and intelligence), as well as public-sector managers and professionals (Russian Academy of Sciences 2014; Remington 2011; Avraamova 2008; Tikhonova 2008).8 This is the group Pavolvskii (2011) calls the “Putin majority”—those who were most negatively affected by the reforms of the 1990s and gained the most from Putin’s tenure. There are several reasons to believe that coalition formation among a middle-class group with such heterogenous interests is unlikely. In addition to their formal wages, employees in Russia’s public-sector are more likely than their private-sector counterparts to enjoy certain benefits: a formal labor contract, shorter working hours, paid vacation and medical leave, medical insurance, transportation benefits, housing benefits, and other subsidies (Remington et al. 2012; Russian Academy of Sciences 2014). While those outside the public-sector are vulnerable to economic volatility and corruption, bureaucrats and professionals paid out of the state budget are insulated from economic risk and benefit from informal rents. Tikhonova (2008) puts it starkly in terms of the inherent conflict of interest between bribe-takers and their victims (25). Besides having distinct economic and political interests, Russia’s public and private-sector middle classes are subject to different 8. Among those employed in Russia’s state-sector are civil servants; the Procuracy, military, police, and others involved in internal and external security; and professionals paid out of the government budget, including medical professionals and teachers (the so-called budzhetniki). In Russia, the so-called “force structures” or agencies engaged in internal and external security include: the Ministry of Foreign Affairs, the Ministry of Defense, the Procuracy and courts, the migration service, the tax inspectorate, and various intelligence agencies. According to estimates by Novaya Gazeta, the total number employed in these agencies exceeds 4 million. See http://www.ng.ru/news/464537.html?print=Y.

110 c h a p t e r 4

mobilizational pressures. Frye, Reuter & Szakonyi (2014, 196), for example, find that Russia’s state-owned enterprises, government ministries, and public-sector unions are easy targets for the ruling party to mobilize citizens in support of the regime (see also Hale 2007). This chapter’s findings add to recent work on protest in Russia, clarifying how patterns of collective action are shaped by extensive state economic engagement. Shevtsova (2012, 23), for example, questions “how far the Russian middle class wants to go in changing the system ...[given that] a sizable swath of [it] lives off its role as a service provider to the state bureaucracy or big state run corporations.” I link this argument to protest behavior, formulating testable hypotheses, and situating them in comparative theoretical perspective. Lankina & Voznaya (2015) find that regions with a high share of state employees experienced fewer protests with fewer overall participants. This chapter offers individual-level evidence on the political behavior of state employees to explain this aggregate association. My emphasis on the demobilization of potential regime opponents also complements Koesel & Bunce’s (2012) observation that widening anti-regime protests resulted in Kremlin attempts to co-opt and fragment the opposition, including by organizing proregime counter-demonstrations, and Smyth, Sobolev & Soboleva’s (2013a, 2013b) finding that the public-sector middle class played an important organizing role in these rallies. Finally, Chaisty & Whitefield (2013) show that few Russians who supported these protests embraced democracy—underscoring that the movement’s ideological platform was vague and its character not exclusively democratic. There was plenty of room to protest without wanting full-fledged democracy. Nondemocratic alternatives like the communists and nationalists were an established part of Russia’s political landscape. Given protesters’ heterogenous motivations, we need to disaggregate the “mass” in mass mobilization, moving beyond the study of protest events to examine individual-level participation in democratic coalitions.

Data and Empirical Strategy A major challenge of studying protest at the individual level is access to suitable data. A review by Walgrave & Verhulst (2011) finds that only a handful of studies involved surveys of protesters before the mid-1990s. Though common now in Western Europe,9 protest surveys remain relatively rare elsewhere 9. See, e.g., Walgrave & Verhulst (2011) and van Stekelenburg et al. (2012).

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 111

(for exceptions, see Volkov 2012; Smyth, Sobolev & Soboleva 2013b; Onuch 2014). Yet alone, even high-quality protest surveys can tell us little about the causes or correlates of protest participation, because they lack information about nonparticipants. Without a baseline of nonparticipants, it is impossible to reliably assess how protesters differ from the population. To address this problem, most studies use existing, usually nationally representative, surveys of public opinion that ask about past protest participation (Welzel 2013, 223). Because these studies capture both protesters and nonparticipants, they allow researchers to better identify the causal mechanisms underlying activism. However, for the study of protest, these surveys also have drawbacks. First, the number of sampled demonstrators is typically small. This results in low statistical power and difficulty detecting effects, especially among protest subgroups. Second, representative surveys rarely ask about participation in specific protest events. Instead, most survey-based research on social movements to date relies on broad measures of participation (e.g., at “peaceful demonstrations”), ignoring the particulars of movements’ activities. These studies thus lack direct evidence on protesters’ motivations and recruitment. Third, even when nationally representative surveys do ask about participation in a particular event, they do so after the fact. This means that most individuallevel studies of protest rely on measures of reported behavior, which are subject to cognitive biases and respondent recall. Especially for revolutionary uprisings and other historic events, reported participation may be biased by social pressure to have taken part or to hide one’s participation in a movement that failed. Fourth, relative to protest participation, nationally representative surveys provide only post hoc measures of other covariates, potentially introducing posttreatment bias. The research design in this chapter offers a new approach to these challenges. To address the inferential problems just discussed, I employ both a random sample of protesters and a random sample of the population from which protesters were recruited. I begin by comparing these two samples descriptively. Then, for the main results, I use a multivariate modeling strategy first proposed by Lancaster & Imbens (1996) to estimate the probability of protest as a function of individual-level characteristics. This design is a variant of the standard case-control research design used in individual-level rare events studies in epidemiology. It has the advantages of scale, specificity, and verifiability that are lacking in existing methods for studying protest. By selecting on the dependent variable, the protest sample ensures an adequate

112

chapter 4

supply of protest cases. Protest participation is observed at a specific protest event, with the interviewer, in effect, verifying the respondent’s participation. At the same time, the population sample supplies the necessary baseline. The survey data I use in my analysis were collected at protest events in Moscow between December 2011 and January 2013, following parliamentary and presidential elections. These data arguably provide the most complete individual-level record of a mobilizational cycle outside the established democracies. A respected independent polling organization, the Levada Center, used international best-practices in protest survey methodology to ensure a representative sample. Stationary protests in Russia are generally cordoned off by special forces and riot police, creating a defined perimeter for the protest event. During the post-election demonstrations, police established entry points and required protesters to pass through metal detectors. Interviewers took advantage of these procedures to randomly sample protesters at a fixed skip interval. When this was not possible, for example, at moving demonstrations and marches, interviewers moved systematically though the columns of participants, selecting every n-th respondent at an interval set by fieldwork supervisors.10 In this way, interviewers worked their way from one end of the crowd to the other. Because protesters generally assembled by ideology, political party, or identity, this procedure ensured that all groups were represented. To complement the protest data, I identified an unusual source of population data that facilitates detailed comparison. With a total sample size of 33,997 respondents, the Foundation for Public Opinion’s (FOM) GeoRating surveys are both nationally and regionally representative of Russia’s population. A minimum of 500 interviews are conducted in each region, including both the city of Moscow and Moscow region. The surveys were conducted face-to-face in respondents’ homes. When possible, these data have been cross-checked with the most recent Russian census data from 2010, though the latter is less extensive on certain variables.11 The main explanatory variables in the analysis are middle-class and state employment. I measure both as in previous chapters.12 In Russia, this includes 10. For details, see section B.1 of the online appendix. 11. The FOM sample closely mirrors the census data on gender, age, and education. The census data do not, however, include information about occupation, only broader measures of employment status (i.e., economically active vs. various categories of economic inactivity). 12. A limitation of the question asked in the present protest survey is that it does not differentiate among different categories of public employment. I thus use the terms “public-sector

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 113

civil servants, bureaucrats, educators, medical professionals, and other socalled budget employees (budzhetniki). The same definition of each of these two variables is applied to both the protest and population data. Having outlined the empirical strategy, the next section analyzes the data descriptively. A more detailed explanation of the case-control research design and estimation strategy precedes the main empirical results. The results are organized into three sections. First, I investigate anti-regime protest participation. Second, I provide evidence on the mechanisms that explain these patterns of contention. Third, I disaggregate participation to examine the democratic protest coalition.

Descriptive Analysis The Protesters Table 4.2 summarizes the protest surveys. A quick examination of the data suggests that participants were overwhelmingly college-educated managers and professionals. This is particularly true among the first movers, participants in the earlier protests. Besides professionals and small-business owners, students, retirees, and housewives made up the next largest contingent. In terms of political views, a plurality (29%–38%) in each survey identified as democrats, while an absolute majority called themselves “democrats” or “liberals.” While the terms “democrat” and “liberal” are at times used interchangeably, the “liberal” camp in Russia encompasses a range of ideologies from conservative pro-market, even libertarian, to “left-of-center,” along with a range of political views. Although some Russian liberals advocate democracy, not all do. Indeed, the Putin administration’s pursuit of liberal economic policies has ensured that there is also a sizable group of pro-government or “system” liberals (Shevtsova 2012, 29). In the main analysis, I examine these groups separately, though I show in the online appendix that my findings are unaffected by this decision. Another important pattern to emerge in table 4.2 is that communists, socialists, and nationalists comprised a substantial minority at each protest (from approximately 25%–40%). Though these events were often framed as “pro-democracy,” this label obscures participants’ significant ideological diversity (Volkov 2012). Lastly, for two protests in September 2012 and January 2013, survey questions distinguish between managers and professionals of the public and worker” and “state employee” interchangeably to refer to those employed both directly by the state and those paid out of the government budget, such as teachers and doctors.

114 c h a p t e r 4 table 4.2. Descriptive Statistics for Protest Surveys Wave 1

Wave 2

Wave 3

wave 4

December 2011 February 2012 September 2012 January 2013 % % % % Education Higher Unfinished higher Secondary/vocational Incomplete secondary

69.8 12.9 16.7 0.6

69.4 11.7 18.1 0.8

58.6 13.5 26.5 1.4

71.7 10.5 17.2 0.6

Occupation Professional/office worker Managers Entrepreneur/business owner Student Blue-collar worker Housewife Other

54.0 16.5 7.7 12.3 8.0 1.6 3.3

43.7 14.3 8.0 10.9 7.4 1.9 13.9

37.3 11.8 6.6 14.5 9.7 2.6 17.5

40.0 12.6 5.5 8.3 3.7 4.0 25.8

—

—

17.0

15.4

Ideology Democrats Liberals Communists Socialists/social-democrats Nationalists

37.8 31.2 12.6 9.5 6.1

31.4 27.5 18.2 10.4 12.7

28.9 23.1 17.1 9.9 11.7

37.2 25.2 7.4 8.9 7.4

Participation Past protest participant

63.1

79.8

—

94.8

N (Total N = 3,265)

791

1344

805

325

Sector of employment State

Notes: This table describes key variables from four protest surveys conducted in Moscow between December 2011 and January 2013 by the Levada Center, an independent sociological research organization. Proportions may not sum to 1 due to rounding.

private sectors.13 These results show a striking participatory gap. In January 2013, protesters were more than two-and-a-half times as likely to be employed 13. The first two waves did not make this distinction. It is thus possible that these findings better characterize “protest stalwarts” than protesters as a whole. In the absence of more complete data, I cannot rule out the possibility that participation among state employees was higher for the earlier protests, when a larger number of people participated.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 115

in the private-sector as by the state. I next turn to a more informative comparison of protesters and the population.

Protesters and the Population Figure 4.1 compares the demographic composition of the protesters with that of Moscow’s population.14 Estimates of the share of protesters who were middle class (overall and, separately, for the state and private-sector) are plotted on the y-axis as a function of the share of middle class in the population from which protesters were recruited (x-axis). The 45-degree line through the plot denotes proportional representation. Points above the line thus indicate groups that were overrepresented among the protesters. Because both data sources are surveys, there is uncertainty in both x and y values. This uncertainty is captured by the horizontal and vertical 95 percent confidence bars. Several patterns emerge clearly from this plot. First, the middle class was vastly overrepresented among the protesters relative to its share of the population. The magnitude of this overrepresentation (or percentage point differential) is indicated by each point’s vertical distance from the 45-degree line through the plot. Whereas close to 60 percent of all protest participants were middle class, the size of Moscow’s middle class (based on identical criteria) is just 30 percent. These were clearly protests of the “want-mores” rather than protests of the “have-nots.”15 At the same time, however, figure 4.1 highlights just how sharply Russia’s middle class is divided by sector of employment and how consequential this cleavage is for political behavior. Whereas the private-sector middle class was vastly overrepresented among the protesters, an individual from the public-sector middle class was about as likely to participate as had she been drawn randomly from Moscow’s population.16 Though purely descriptive, these figures raise questions about 14. Online appendix section B.2 presents additional descriptive plots of the data. These figures provide further evidence on the protests’ socio-demographic composition and show that patterns in the pooled data are similar across individual survey waves. 15. Following Kerbo’s (1982) typology, these demonstrations more closely resembled a movement of affluence than a movement of crisis. 16. Online appendix figure B.2 shows that state employees were significantly underrepresented among the protesters as a share of the population. These differences are so large that they are very unlikely to be artifacts either of underreporting or survey nonparticipation by state employees.

116 c h a p t e r 4

Moscow 0.6

Middle class (all)

0.5 Non-state middle class

Protesters

0.4

0.3

0.2 State middle class

0.1

0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

Population figure 4.1. Demographic comparison of protesters and the population. The vertical position of each plotted point gives the share of protesters within a given group, while the horizontal position gives that group’s share of Moscow’s population. The 45-degree line indicates proportional representation. The vertical (horizontal) bars are 95% confidence intervals based on the sample size of the pooled protest (population) data. Data sources: The Levada Center and FOM (Moscow only).

the conventional view that growth of the middle class necessarily contributes to greater societal mobilization.

Case-Control Sampling with Contaminated Controls The principal limitation of the preceding analysis is that it ignores potential confounders. Russia’s state employees are, on average, older and predominately female. Only a multivariate strategy can address the extent to which state workers’ low rates of protest participation are explained by such factors as

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 117

gender and age. Another limitation is that the preceding analysis tells us nothing about the prevalence of protest. Ideally, we want to know the probability of protest among particular groups and the population as a whole in addition to the increase (decrease) in the relative risk of protest for individuals with different covariate profiles. The case-control approach I detail next allows us to estimate these quantities. The method I use is a variant of case-control (Keogh & Cox 2014) or choice-based sampling, as it is known in the econometrics literature. For rare outcomes, like protest, choice-based sampling saves significant data collection resources (King & Zeng 2001). The standard case-control design involves sampling observations (randomly or collecting all those available) for which the choice/behavior is observed (i.e., where the outcome, Y, is equal to 1, known as “the cases”) as well as a random sample of the population for which the choice/behavior is not observed (i.e., Y = 0, “the controls”). Under this basic design, the outcome is measured for all observations and selection depends on Y. The probability of the outcome can then be estimated easily as a function of covariates using, for example, the procedure described in King & Zeng (2001).17 In practice, however, representative random samples seldom include measures of rare events. When such a population sample—that is, one consisting of an unknown mixture of cases and controls—is paired with a random sample of “cases” (with Y = 1), the design becomes one of “contaminated controls” (Lancaster & Imbens 1996, 146). The modified case-control design with contaminated controls consists of two independent random samples, one sample selected entirely on the dependent variable, the other drawn from the whole population, with only the covariates observed. If the prevalence of the outcome of interest is known (i.e., the marginal probability of Y = 1 in the population), then the problem is simplified. However, because the population prevalence of many outcomes of interest remain unknown, this approach is often impractical. One solution is to estimate prevalence endogenously as first proposed by Lancaster & Imbens (1996). The model and estimation strategy are described in detail in section B.3 of the online appendix.

17. As a robustness check, online appendix table B.7 gives an alternative set of results for a range of protest prevalence estimates using the prior-correction model for rare events in King & Zeng (2001).

118 c h a p t e r 4 table 4.3. Multivariate Model Predicting Protest Participation Using a Case-Control with Contaminated Controls Research Design Dependent Variable: Protest Participation Moscow Estimate (median) Middle class State employment State × Middle class Male Age Age2 Democrats Communists Nationalists Constant π

1.21 −1.40 0.92 0.60 −0.06 −0.02 9.16 1.71 1.09 −2.83 0.102

CI [2.5%, 97.5%] [0.91, 1.55] [−1.92, −0.90] [0.29, 1.58] [0.36, 0.84] [−0.48, 0.34] [−0.06, 0.03] [3.89, 23.62] [1.30, 2.16] [0.64, 1.60] [−3.84, −1.87] [0.064, 0.147]

Note: The table reports point estimates for the model parameters and 95% Bayesian credibility intervals. Data source: The Levada Center and FOM.

The stacked dataset used in the analysis comprises the random sample of protesters plus a random sample of the population in Moscow and Moscow region. Since protest is only observed in Moscow and protest participants came primarily from Moscow and the surrounding area, these data provide a very clean test of the hypotheses. The model regresses protest on categorical covariates for middle class, state employment, their interaction, political ideology (democrat, communist, or nationalist), and male gender, as well as continuous covariates for age and age squared.

Empirical Results Table 4.3 reports coefficient estimates and 95 percent Bayesian credibility intervals from the case-control model predicting protest participation.18 Because the model is nonlinear, and includes a multiplicative interaction of 18. Online appendix table B.3 reports additional results from a second model in which the population sample is drawn from across Russia. Problems of misclassification are obviously greater in this design, since protest is not observed outside of Moscow. I therefore focus on the first set of results in the main text.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 119

the key independent variables, I forgo a discussion of the coefficients on the explanatory variables in table 4.3 and, instead, present the key quantities of interest graphically. The estimate of π in the last line of table 4.3 (10% with a 95% CI of [6.4, 14.7]) is a sensible estimate of the fraction of Moscow’s population engaged in protest given the city’s size and estimates from other sources, though its 95 percent confidence interval lies on the upper end of conceivable participation.19 Figure 4.2 compares the probability of protest participation for several groups in terms of relative risk: the fractional increase in the probability of protest given chosen values of the explanatory variables relative to the probability of protest given some baseline values of those same variables (King & Zeng 2001, 141). Intuitively, relative risk is greater (less) than one when the probability of protest is higher (lower) for the first group than it is for the second. The median estimate of predicted relative risk is reported together with its 95 percent confidence interval.20 First, I find that state workers were only about half as likely as privatesector workers to join the protests, even after accounting for the confounding effects of age, gender, and ideology. Second, I find that the strong positive relationship between being middle class and protest participation is attenuated substantially by employment in a state-dependent sector. Specifically, I find that the state middle class was only about 75 percent as likely as the nonstate middle class to take to the streets, holding constant class status along with the other controls.21

19. Robustness checks in which π is given reasonable bounds based on all available information are discussed in online appendix section B.4. The basic strategy is to take organizer reports of protest size as the upper bound for calculating the population prevalence of protest, and police (Russian MVD) reports of protest size as the lower bound. These bounds can then be used to determine the conceivable range of parameter estimates. The results reported below remain substantively the same whether π is estimated endogenously or bounded between 0.3% and 6% of Moscow’s population. 20. I average over the empirical distribution of all other covariates in the population sample, using the observed-value approach in Hanmer & Kalkan (2013). This approach ensures that the estimates obtained are of average effects in the population and are not due to differences in the distribution of the other covariates. 21. Given that ideology likely captures some of the effect of being state employed, these are conservative estimates. Online appendix section B.5 shows that these findings are not due to differences in the age composition of the two middle classes or their maturing politically during different periods.

120

chapter 4 Risk ratio 95% CI

State vs. non-state State middle class vs. non-state middle class Middle class vs. non-middle class Non-state middle class vs. non-middle class State middle class vs. non-middle class

0.5

1.0

2.0

3.0

4.0

figure 4.2. Relative risk of protest participation. This figure compares the fractional increase in the risk of protest participation for each group relative to the given baseline. The results are based on a case control with contaminated controls design, which combines a random sample of protesters with a random sample of the population of Moscow and Moscow region. Like the preceding results, these findings indicate that state workers, in general, and the state middle class, in particular, were less likely than their private-sector counterparts to protest. Horizontal bars indicate 95% confidence intervals. Data sources: The Levada Center and FOM (Moscow only).

What do these results tell us about how middle-class growth is likely to affect protest potential? The last two bars in figure 4.2 show that while state employment does not entirely vitiate the boost in protest participation associated with a growing middle class, it does dramatically diminish it.22 This implies that mobilization rates will be lower, and anti-regime protests less likely to obtain a critical mass, when a sizable segment of the middle class is state dependent. Accounting for cleavages within the middle class thus helps to explain why threats to regime stability in Russia have not been more significant and protests more successful. According to the preceding analysis, had the state middle class participated at the same rate as the private-sector middle class, up to 90,000 additional protesters would have taken to the streets. Moreover, had state workers protested at the same rate as others, the number of protesters would have 22. See also appendix table A2.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 121

grown by up to 200,000.23 These findings point to an under-appreciated aspect of mobilizational potential in developing nondemocracies: not only is it important whether the middle classes are growing, it matters too whether that mobility is supplied by the state. Finally, although the attitudinal controls in table 4.3 confirm a key role for democrats, they also reveal that possessing clear political views— whether democratic or not—increased the probability of protest.24 Contrary to the pro-democracy frame ascribed by Western journalists, these protests attracted participants from across the ideological spectrum. Middleclass communists and nationalists, like democrats, were systematically more likely to take part than both their working-class counterparts and those who did not subscribe to any particular ideology. This underscores the fact that a growing middle class may increase social mobilization behind political demands that both are and are not compatible with democracy. Understanding the sources of democratic protest potential requires closer examination of protest participants’ ideology. First, however, I briefly explore the mechanisms by which state-dependent development shapes mobilizational potential.

Selective Incentives, Grievances, and Social Capital Incentives, grievances, and differences in social capital all potentially help to explain the variation in middle-class protest participation found in the preceding analysis. Though descriptive in nature, the evidence in this section suggests two tentative conclusions: first, that both positive and negative inducements, like fear of being fired, demobilize the state middle class, and, second, that the grievances engendered among those excluded from special treatment contribute to mobilizing the private-sector. First, the most straightforward explanation for state workers’ low protest participation is that they fear being fired from their jobs. Reports that state employers threatened dismissal to mobilize workers in support of the regime circulated widely at the time of these demonstrations. How likely is it that 23. These estimates are based on the entire mobilization cycle. For calculations, see section B.8 of the online appendix. 24. The large positive coefficient on democrats in table 4.3 implies that if more state employees were democrats, protest numbers would have been higher. However, state-sector democrats were still less likely to mobilize than private-sector democrats who were better insulated from government pressure.

122

chapter 4

table 4.4. Public-Sector Employees Cite Fewer Alternatives February 2009 April 2009 May 2009 % of each group who believe they would be unable to find alternative employment Public-sector workers Private-sector workers

45 39

44 38

48 40

Source: FOM nationally representative surveys, N = 2,000 in each wave. Notes: This table displays the share of respondents who think it is unlikely they would find employment, in response to the question “In the event you are dismissed, do you think it will or will not be possible, given your qualifications and experience, to find employment in your field in the coming 2–3 months?”

fear of being fired could account for overall sectoral differences in patterns of contention? If state workers are comparatively less confident in their ability to find alternative employment, we would expect them to be less likely than private-sector workers to join anti-regime demonstrations. While the protest surveys did not measure perceived alternatives, we can get some sense of them from other data sources. According to a series of surveys conducted in 2009 at the height of the financial crisis and shown in table 4.4, roughly half of all state employees expressed concern about finding alternative employment if dismissed. Given limited alternatives, these data suggest that state workers would indeed be more concerned about retaliation. Besides negative sanctions like threats of dismissal, could patterns of participation have been influenced by systematic sectoral differences in formal and informal benefits? A variety of evidence suggests that positive inducements were also salient. Given Russia’s large gray economy, a formal labor contract, being paid on the books, vacation, and medical leave are far from universal practices. While these benefits are widespread in public employment, just 30 percent of workers at newly established private entities receive these same protections. Even among Russian professionals, fully a quarter lack these basic social benefits and sources of job stability (Russian Academy of Sciences 2014, 49–50). The same series of surveys captures public-sector employees’ greater sense of job security.25 Surveyed three times during 2009 as the crisis progressed, as table 4.5 shows, public-sector workers were as much as 17 percentage points 25. See FOM publications Dominanty No. 23 (06-11-2009), No. 38 (09-24-2009), and No. 41 (10-15-2009).

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 123 table 4.5. Stability of Public-Sector Employment June 2009

September 2009 October 2009 % of each group saying dismissal due to layoffs very unlikely

Public-sector workers Private-sector workers

59 50

64 51

72 55

Source: FOM nationally representative surveys, N = 2,000 in each wave. Notes: This table displays responses to the question “What do you think, how likely is it that you will be dismissed in the coming 2–3 months due to layoffs?”

more likely than private-sector workers to think it unlikely that they would be laid off, even under conditions of profound economic uncertainty. Whereas nearly three-quarters of public-sector workers felt their jobs were safe, only about half of private-sector workers felt the same about theirs.26 Public-sector job security thus plausibly also encouraged regime loyalty and lower protest participation among the state middle class. Beyond the formal benefits and greater sense of job security that come with public-sector positions, government jobs also frequently provide access to networks, information, and resources that can be used for private enrichment. While the opportunity to earn informal rents binds public-sector workers to the regime, these privileges and side-payments simultaneously produce private-sector grievances. Indeed, the private-sector middle class expressed greater concern about official corruption (42% vs. 35%) and was less likely to believe the courts would protect its interests (49% vs. 57%).27 These grievances thus reinforce the rift between the two middle classes. Insofar as grievances motivate protest (Gurr 1970) and the desire for protection from a predatory state drives demands for democratic transition, these differences plausibly help to explain why the private-sector middle class was more likely to demonstrate while the state middle class was reticent. 26. Note that the question concerns layoffs in the context of an economic downturn, not dismissal for political reasons. 27. By contrast, measures of consumption offer no evidence that the private-sector middle class protested due to objective economic deprivation. Both the protest survey and the survey of the middle class described in online appendix section B.1 show that the average level of consumption of the state and private-sector middle classes is very similar. This evidence also leans against the notion that the state middle class’s lower propensity to protest could be due to its possessing fewer resources.

124

chapter 4

A related possibility is that private-sector professionals compensate for their exclusion from state-sponsored privileges through dense networks of friends and acquaintances. These networks could in turn reflect differences in social capital that increase the likelihood of private-sector protest participation. Turning to evidence from the protest survey, I find little support for this mechanism. While 10.1 percent of the state middle class reported that friends or acquaintances brought them to the protest, 5.9 percent of the private-sector middle class said the same (a difference that is statistically insignificant, but leans against the hypothesis). Clearly these data do not support causal claims. However, they do suggest that both negative and positive selective incentives could explain variation in middle-class mobilization—by demobilizing the state middle class and contributing to grievances among private-sector professionals excluded from special privileges. Before concluding this analysis of protest participation, I briefly consider one further set of observable implications: If selective incentives wielded by the state are indeed important for explaining patterns of mobilization, then we would expect state workers’ protest participation to vary as a function of their exit options (or opportunity costs of jeopardizing state benefits). In particular, those state workers who have the fewest (most) exit options should be least (most) likely to participate in opposition protests. To test this hypothesis, I examine protest participation propensities for public employees at various stages of their careers. Figure 4.3 presents the results visually by sector of employment for the middle class.28 By focusing only on Moscow’s middle class, these results hold constant factors like education, occupational type, and characteristics of the local labor market that might also affect an individual’s exit options—thus isolating life cycle variations and the effect of experience and highlighting differences across sectors. Though merely descriptive, the pattern of protest participation for state employees shown in figure 4.3 appears to be consistent with life cycle variation in the strength of the state’s selective incentives. Despite generally high levels of participation among young people, state employees in their twenties—with little career experience and fewer exit options—were somewhat less likely to protest than state employees in their thirties, who, while still 28. The distribution of age in the population has been cross-referenced with the 2010 Russian census data for Moscow, archived here: http://moscow.gks.ru/wps/wcm/connect/ rosstat_ts/moscow/ru/census_and_researching/census/national_census_2010/score_2010 /score_2010_default.

Index of representation

(proportion of protesters)/(proportion of population)

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 125 Non-state middle class

2.5 2.0 1.5

State middle class

1.0 0.5 0.0 Under 30

30’s

40’s

50’s

Over 60

figure 4.3. Exit options and protest participation. This figure compares middle-class mobilization by sector of employment at different points in the life cycle. The y-axis plots an index of representation, with higher values indicating more protest mobilization. On the x-axis are age cohorts, as a proxy for exit options. Note that by focusing on Moscow’s middle-class protest participants, this analysis holds constant educational and occupational endowments, as well as characteristics of the local labor market that may also be associated with available alternatives. The results are consistent with the claim that life cycle variation in available alternatives to state employment affects the power of state selective incentives and, thus, rates of mobilization. Data sources: The Levada Center and FOM (Moscow middle class only).

young, had more experience and therefore more options. Mobilization then declined among state employees in their forties and fifties, as they developed more specialized social capital in the public-sector and their private-sector alternatives waned.29 The same was not true of state employees over sixty, most of whom would have begun receiving a pension—a universal and nonexcludable benefit. As the right side of figure 4.3 shows clearly, middle-class state employees of pension age, no longer in need of alternatives (thus lowering the stakes of protest participation), were much more likely to participate. In sum, for those who were still young but had acquired work experience, exit options peaked, while, for those at or near retirement, they declined sharply in

29. This pattern is largely consistent with Russians’ subjective appraisals of the availability of alternatives to their current employment. See, for example, FOM (2009).

126

chapter 4

relevance. Selective incentives, thus, operated most weakly for these groups, while for those who lacked alternatives, they operated most strongly. It is also worth noting that the pattern of participation by state employees differs from that of non-state employees. If the pattern of participation across age groups was similar for public- and private-sector workers, it would cast doubt on the importance of selective incentives—which differ across sectors—and imply that these patterns are better explained by some other factor correlated with age. Instead, the patterns are systematically different: lower participation among the youngest state workers (vs. high participation among non-state twenty-somethings), participation peaking with career mobility in the thirties (vs. earlier in the private-sector), and a distinctive uptick in mobilization among those reaching pension age in their sixties (versus monotonically declining30 participation among non-state employees over forty).31 To reiterate, because private-sector workers did not face the same selective incentives as public-sector workers, these differences give us greater confidence that selective incentives plausibly explain patterns of participation.

The Democratic Protest Coalition I next disaggregate participant ideology in order to test the second hypothesis regarding sectoral differentiation in the democratic protest coalition. As Robertson (2011) observes, “Not all protesters demonstrating under (or even against) authoritarian rule are democrats pushing for liberal revolutions” (13). Like other post-election protest movements, the demonstrations in Moscow attracted groups from across the political spectrum under a vague ideological platform and agenda. This diversity was clearly visible in images of the crowds, with activists of different political parties, movements, and identity groups congregating and marching together. While organizers united in opposition to electoral fraud, far from all participants were democrats. 30. The one exception to monotonically declining rates of participation among non-state employees is for individuals in their thirties. A possible explanation is that this cohort is most likely to have small children. 31. An alternative is that the pattern of protest participation for public-sector workers just described could also be due to selection—that is to the self-sorting of individuals into different career paths. If, for example, those who joined the public sector under Yeltsin are systematically different than those who joined under Putin, then those differences—rather than selective incentives—may explain observed patterns of mobilization among state workers. Online appendix table B.15 presents the results of a test of this selection hypothesis.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 127

Again, I expect that protest coalitions will exhibit significant sectoral differentiation, with the state middle class less likely to identify as democrats. I test this proposition using responses to the survey item: “Which of these groups represent ideas that are closest to your own?” The dependent variable is “democrats” and takes a value of either zero or one.32 The main explanatory variables are, again, middle class and state employment, as well as their interaction.33 The same basic controls are included and, as a robustness check, I also add a common proxy for income based on household consumption to the analysis.34 Consumption is coded as a categorical variable with six levels from low (1) to high (6). Including this measure ensures that observed differences between the state and private-sector middle classes are not confounded by a sectoral wage gap. I use logistic regression with heteroskedastic-robust standard errors, clustered by survey wave.35 Figure 4.4 shows that the effects of both state employment and middleclass status are substantively large and in the hypothesized direction.36 Whereas a middle-class protester working in the private sector had about a 39 percent chance of identifying with the democrats, a middle class protester from the public sector had only about a 30 percent chance. This difference is significant at the α = 0.05 level. Moreover, a middle-class protester from the public sector was no more likely than a working-class protester to identify as a democrat (p > 0.20). Thus the classic result on the left side of the plot 32. Again, liberals are not included in the main analysis to minimize possible confounding by “system liberals”—those who prefer liberal policies in the economic sphere, but see them as compatible with nondemocratic rule. As a robustness check, I repeat the analysis in online appendix table B.13 using a combined dependent variable. The results are qualitatively similar, though the effect of state employment is slightly diminished. 33. Online appendix section B.6 confirms that all of the educational and occupational categories that make up the middle class affect the probability of democracy support given protest participation similarly, justifying their joint analysis. 34. See section B.1 of the online appendix for the exact wording of this and all subsequent items. 35. The regression results, for both the pooled data and individual surveys, are reported in online appendix table B.12. The results are substantively similar with and without the control for household consumption. I also reran the analysis in table B.12, without the state employment variable, on all four waves of survey data. These additional results in online appendix table B.14 highlight that middle class and sector of employment have important interactive effects which models that include only class status or occupational and educational criteria fail to capture. 36. The results displayed are from the model in column 3 of online appendix table B.12.

128

Non-middle class

0.6

Probability of being a ‘democrat’

chapter 4

0.5

Middle class

State

Private State

0.4

0.3

Middle class

Non-middle class

Private

0.2

0.1

0.0

figure 4.4. The democratic protest coalition by class and sector. This figure shows the predicted probability of identifying as a democrat. State middle-class protesters were significantly less likely (p < 0.05) to identify as democrats than the private-sector middle class and no more likely to identify as democrats than the working class. Data source: The Levada Center (all protesters).

obscures significant causal heterogeneity. While these findings suggest that changes in the class structure of society and growth of the middle class shape political preferences and democratic coalitions, they imply that the state’s role in these processes is also critical. Even those state employees who did protest were significantly less likely to join the pro-democracy coalition than would be expected given their other characteristics and class position. va lue s, in ce n t i v e s, a n d t h e f o r m at i o n of p ro -d e m o cr ac y coa l i t i o n s Though united by common opposition to electoral fraud, the groups protesting were not equally threatening to regime stability. While the democratic forces consisted overwhelmingly of regime opponents (or what in Russia is called the “non-systemic opposition”), the communists are an official opposition party with parliamentary representation. Although they do occasionally challenge the party of power, they more often act in concert with the regime. Thus, identifying with the Communist Party represents an acceptable level of dissent within the system, one that does not threaten regime stability.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 129

Indeed, I find that middle-class state employees were more likely to identify with the communists than predicted by their other characteristics and class position. These differences in political orientation were also evident in respondents’ motivations for protest attendance. While private-sector employees were more likely to have the goal of forcing Putin’s resignation and regime change (33% vs. 26%, X 2 (1) = 3.44, n = 1130, p = 0.06), public-sector workers were more likely to have economic motivations (30% vs. 22%, X 2 (1) = 6.34, n = 1, 130, p = 0.01). Both of these relationships hold after controlling for level of consumption and class. Compared to others, state-sector protesters were more motivated by economic and less by ideological grievances.37 This implies that the public-sector middle class primarily sought a better bargain with the regime. It also suggests that Russia’s leaders could ensure their loyalty more easily through economic inducements, which is exactly what they did. Following the protests, the Kremlin raised wages across many categories of public employment (especially for professionals of the budget sector) and instituted more robust measures to monitor regions’ compliance with the directives on public-sector wages issued by the federal center. Finally, these differences were also apparent in the degree of radicalism protesters from the state and private sectors believed necessary to resolve Russia’s most pressing problems. While radical protesters were a minority in both sectors, private-sector workers were significantly more likely than public-sector workers to support solving Russia’s problems through radical acts of protest with the goal of regime change (14% vs. 8%, X 2 (1) = 4.64, n = 1130, p = 0.03). In sum, these findings suggest that many publicsector employees joined the protests because they believed the state had failed to uphold its commitments. At the same time, they were less likely than private-sector protesters to frame their grievances in the politicized language of rights and repression and less likely to support radical opposition and regime change. By redoubling its efforts to support the state middle class, following the protests, the Kremlin clearly sought to renew its contract with those most invested in regime stability. 37. Indeed, only about a third of the state middle class mentioned changing the policies of those in power and beginning reforms as the reason for their participation (contrast 44% of the private-sector middle class). Nearly as many mentioned the desire for a higher standard of living.

130 c h a p t e r 4

Conclusion Popular challenges and urban insurrections against authoritarian regimes are increasingly important for contemporary democratization (e.g., Ulfelder 2005; Bunce & Wolchik 2011; Beissinger, Jamal & Mazur 2015). Yet a lack of detailed data has impeded our understanding of the constituencies involved in anti-regime protest. This leaves important questions unanswered, especially concerning the post-election demonstrations that have proved so effective at dislodging authoritarian incumbents (Tucker 2007; Magaloni 2010). Drawing on choice-based sampling methods, I take a novel empirical approach to the individual-level study of protest and apply it to the case of Russia’s recent mobilizational cycle. This approach broadens the tools available to scholars for studying the microfoundations and mechanisms of protest. It can be used across a variety of research contexts, wherever surveys of both protesters and the population from which they are recruited can be conducted. With the benefit of these new tools, this chapter nuances the notion that Russia’s middle class mobilized for democracy. Within the middle class, I found striking participatory gaps. In fact, mobilization occurred at very different rates among public- and private-sector employees with similar occupational and educational profiles. While the opposition movement drew support from the private-sector middle class, the public-sector middle class— the fastest growing segment of Russia’s middle class—was more likely to stay home and less likely to identify as democrats. One implication is that cognitive mobilization (e.g., Welzel & Inglehart 2008; Inglehart 1990) is not uniformly the consequence of rising affluence, education, and occupational specialization, but depends in part on selective incentives, like other forms of mobilization. In particular, these findings suggest that, under autocracy, both the value placed on democracy and the will and resources to intervene effectively in politics vary with an individual’s relationship to the state. This is especially true of the middle class, the group both modernization and values-based theories expect to be most active in gaining and maintaining democracy. Other implications follow for the literature on patronage. As these results demonstrate, state patronage can provide powerful incentives for political loyalty, even among the middle class for whom these incentives are usually assumed to matter less (Medina & Stokes 2007; Calvo & Murillo 2004). The analysis further shows that by demobilizing a large segment of the

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 131

urban middle classes during anti-regime protests, co-optation of the middle class contributes to authoritarian regime resilience. Based on my analysis, I estimate that if the state middle class had participated at the same rate as the private-sector middle class, up to 90,000 additional protesters would have taken to the streets. What is more, had state workers protested at the same rate as everyone else, the ranks of protesters would have risen by up to 200,000. As work on informational cascades by Kuran (1995) and Lohmann (1994) demonstrates, whether people decide to take part in protests depends on their expectations about others’ behavior. This implies that public-sector workers’ low rates of participation likely also discouraged others and helped to prevent the protests from achieving a critical mass (Marwell & Oliver 1993). Thus, both directly and in terms of their spillover effect, state employees’ poor turnout and comparatively weaker support for the pro-democracy coalition helped reinforce regime stability. While sheer numbers are effective at destabilizing autocrats, weak democratic coalitions make democratic transition, not to mention consolidation, a more uncertain proposition. As described at the outset, the mobilizational cycle around Russia’s 2011– 2012 parliamentary and presidential elections featured a series of protests by the opposition as well as counter-mobilization by the state. This book’s framework also suggests several reasons why the regime could efficiently mobilize its public-sector middle class. First, public-sector workers were more vulnerable to selective incentives. Second, they were more invested in the continuation of the regime and required less ideological mobilization. While the Kremlin could cheaply purchase the rally attendance of low-skill, low-wage workers, it required the support of the public-sector middle class to organize a credible alternative to the opposition. Smyth, Sobolev & Soboleva’s (2013a) study of these pro-regime rallies finds that participants were better educated, more affluent, and more likely to work in the public-sector than nonparticipants—while rally stalwarts were more likely than casual participants to hold supervisory roles at state jobs (31). Rally stalwarts were also more likely to report that their material well-being increased since 2000, in keeping with the strategy I have described of enlarging the state-sector middle class and cultivating its loyalty through targeted benefits. Alongside this, selective incentives like transport and housing, promises of days off, and threats of dismissal helped to persuade state workers to attend. So why are these incentives apparently so effective in the Russian case and how far is my framework likely to travel? I return to this question in the

132 c h a p t e r 4

Conclusion, but to preview that discussion beyond Russia, the argument’s scope rests on two basic conditions: illiberal state institutions and extensive state economic engagement. More specifically, three institutional factors are conducive to a high degree of dependence among public-sector professionals: 1) the absence of merit-based recruitment, 2) lack of employment protections, and 3) tolerance of official corruption (i.e., the ability to earn informal rents by exploiting one’s professional position).38 While these conditions are widespread in other autocratic settings, in cases where alternative institutional configurations characterize the state sector, the middle class may be more politically independent. Additionally, as discussed in chapter 2 the power of selective incentives varies with the opportunity costs of forgoing state benefits.39 These costs are determined not only by individual endowments, but also by available alternatives. When workers have few exit options and many excludable benefits, they will be least likely to join opposition protests. The fact that Russia’s public sector is extensive in certain industries (like health and education) means that outside options for professionals in those fields are limited. This is also true in industries deemed strategic—such as energy, defense, transportation, and more recently technology and finance—as well as the natural monopolies in electricity, gas, water, and rail. In short, the overall size of the public sector matters as does the generosity of available benefits.40 For instance, public-sector employment is more extensive in Belarus and Azerbaijan than

38. These factors are virtually universal in the post-Soviet cases. Even where dismissal on the basis of political opinion is illegal (e.g., Russia and Azerbaijan), actual practice diverges from these formal protections. 39. And, indeed, selective incentives may be less important relative to other forms of leverage, such as workplace-based socialization and indoctrination, when a regime is long-standing and ideological. By contrast, in more circumstantial and temporary regimes, working for the state poses less of a problem in terms of signaling loyalty to the opposition should it win democratizing elections. 40. As the resource curse literature notes, non-tax revenues help to sustain high public spending and bloated public sectors, inhibiting the formation of an autonomous workforce. This book dovetails nicely with that perspective and provides clear micro-level evidence of how resource states’ large public sectors limit middle-class protest potential. Importantly, however, while the mechanism that this book highlights is often present in rentier states, it is by no means limited to states with oil, as discussed further in the Conclusion. While Azerbaijan has oil, Belarus does not. Both have large public sectors and less successful protest movements than other post-Soviet neighbors.

r e t h i n k i n g m i d d l e c l a s s p r o t e s t 133

in either Ukraine or Georgia.41 Consistent with this book’s argument, challenges to the regime have been less successful in the former cases than in the latter. In terms of policy, this suggests that both state retrenchment and shifts in development strategy that favor the creation of a more economically autonomous middle class are likely to undermine regime stability. Where the middle class remains divided, as it does in Russia, we are likely to see periodic but underpowered opposition protests continue. In closing, this chapter’s findings imply that the state middle class constitutes an important swing group in the social coalition supporting a nondemocratic regime. Yet, several questions remain. Are those who gravitate toward employment in the state sector systematically different? Perhaps those who pursue public-sector careers hold different political preferences, which better explain their subsequent behavior? In all of the analyses thus far, I have attempted to grapple with this question, despite limitations of the data. The next two chapters use an original survey and strong research design to more reliably answer these questions. Chapter 6 then shows further behavioral evidence on the effect of middle-class state dependency from a country that, unlike Russia, does not have a natural resource-based economy. While survey answers are relatively costless to give, protest participation is not. The evidence in this chapter begins to show how state-led development weakens the middle classes’ commitment to democracy, with behavioral consequences for how democracy is contested in the street.

41. According to the EBRD, the private-sector’s share in employment is 10%–15% higher in the former cases than in the latter.

5 Choosing to Work for the State

the preceding chapters show that, with respect to democracy, the political attitudes and behavior of the publicly employed middle classes are different. But, in a country such as Russia, what kinds of people elect statesector careers in the first place? Are less democratic types more likely to gravitate toward employment in the public sector? Is the public-sector middle class systematically different, because it attracts individuals with systematically different characteristics? These questions are critical: If their answer is yes, then we should look earlier in the causal chain for the drivers of democratic attitudes and behavior. If their answer is no, then observed differences in state bureaucrats’ attitudes and behavior plausibly arise later, as a result of employment. The analysis in this chapter thus sheds light on whether Russia’s autocratic institutions are sustained, in part, by the self-sorting of autocratic types into public service or whether these institutions themselves promote political preferences that are conducive to regime stability, as I have argued. To address these questions, this chapter investigates the political orientations and career aspirations of students who intend to join Russia’s public sector, using original survey data. The analysis focuses on whether and how Russia’s future public servants differ from others in terms of their motivations, networks, values, and background. While I examine a range of factors highlighted by existing cross-national research on public employment in established democracies and developing states, my principal focus is on whether prospective government workers differ in their political attitudes, a set of variables less frequently studied in the extant public administration literature. Clearly, the choice of career is not random. Some set of factors dictates the preference for a government career over a career in the private sector. This chapter revolves around the question of whether those factors include regime preferences, or some other attribute or attitude correlated with democracy 134

c h o o s i n g t o w o r k f o r t h e s tat e 135

support. By examining students’ political preferences before they enter the labor market, this chapter’s research design provides further evidence for the arguments presented in previous chapters. In what follows, I study prospective Russian public servants’ views on the importance of political freedom, order, national security, and strong economic performance. I also probe whether democratic values, both in the abstract and in terms of support for specific democratic institutions, are tied to the preference for public employment. I find that students who choose public-sector careers begin with very similar political attitudes, but somewhat different backgrounds and motivations. The single most important determinant of the preference for a public-sector career is preexisting networks. I show that having a parent who works in the public sector or attending a university with a strong alumni network in government has much greater influence on career choices than do political orientations. It is the availability of these informal networks, rather than ideology, that shapes who is most likely to join Russia’s public sector.

Working for the State in Comparative Perspective More than twenty-five years after the collapse of communism and the demise of central planning, the share of state employment in Russia’s economy remains substantial. While the 1990s are often remembered as a decade of state downsizing, between 1990 and 1998, the share of total employment in the public sector actually increased. The number of teachers in state schools, academic staff in state higher education, and employees in heath care, social protection, and sports all grew (Gimpelson & Treisman 2002).1 Employment in public administration ballooned by 45 percent or approximately 900,000 employees. By 2005, roughly half (48%) of all full-time employees in Russia worked in the government sector, according to data from the International Social Survey Programme (Houston 2014, 851). OECD figures from 2011 give the rate of public employment in Russia’s labor force as 31 percent. Since the labor force includes both the unemployed and part-time workers, this figure significantly understates the real weight of government employment in the Russian economy, especially among salaried full-time workers. As noted in chapter 1, the 1. Only employment in science and research declined dramatically as the Soviet Union’s military industrial complex was dismantled.

136 c h a p t e r 5

rate of state employment among middle-class professionals is higher than for the labor force as a whole. Based on the EBRD data from chapter 3, just under 55 percent of the Russian middle class is employed in the state sector. Much as the institution of universal state employment served as an important instrument of social control in the communist period, so too does an extensive state sector help to mobilize regime support and ensure political stability in Russia today. Despite the fact that autocracy often goes hand in hand with a large public sector (Greene 2010), few studies have directly examined what makes public sector positions desirable in these contexts and how those who take government jobs differ, if at all, from their peers. As a starting point, scholars studying these questions in democratic settings often focus on extrinsic motivations like salary or, in the high-corruption context of new democracies, opportunities to earn informal rents. Sharunina (2013) notes that public-sector wage premiums are generally higher in less-developed countries, especially where there is a large informal economy. However, in Russia, she documents a sizable public sector wage gap: over the period 2000–2010, public-employees were paid 30 percent less than private-sector employees for similar jobs, even after accounting for workers’ observed and unobserved characteristics. Although there is some regional variation—the gap is smaller where there are fewer public-sector jobs and where federal subsidies are higher—in the best case, public-sector salaries remain 20 percent lower. Still, formal wages are only part of the picture. While low official salaries may drive qualified candidates out of the public sector, state jobs also frequently provide lucrative opportunities to earn informal rents. In high-corruption contexts, government workers can offset low wages by exploiting their official positions. Gorodnichenko & Peter (2007), for example, show that state employees in Ukraine hide a substantial proportion of their total income, which the authors attribute to bribes and other informal payments. Gerber & Schaefer (2004, 46), meanwhile, find that a degree in state administration yields among the highest returns (together with economics and business) of any academic major in Russia today. Thus, if workers are drawn to Russia’s public sector by the promise of financial gain, they are probably motivated by informal benefits rather than high official salaries. That said, existing cross-national research has reached mixed conclusions on the importance of financial incentives for public-sector recruitment (e.g., Dal Bó, Finan & Rossí 2013; Lewis & Frank 2002; Karl & Sutton 1998;

c h o o s i n g t o w o r k f o r t h e s tat e 137

Krueger 1988). Some studies find that those who prioritize high income are more likely to seek government employment (Lewis & Frank 2002), while others find that private-sector workers place a higher value on good wages (Karl & Sutton 1998). Still others find that there is little difference (Gabris & Simo 1995). Alongside the literature on extrinsic financial motivations, a large body of research in the field of public administration focuses on intrinsic motivations for public service. This literature hypothesizes that individuals with a high level of public service motivation are drawn to, and will self-select into, jobs that meet these needs (Houston 2014, 856). Indeed, research in the developed democracies consistently finds that aspirants to public service display higher levels of altruism and a stronger desire to be useful to society than those who seek private-sector positions (e.g., Lewis & Frank 2002; Barfort et al. 2015). Using experiments embedded in a Russian online survey, Gans-Morse et al. (2017) show that students who prefer public employment behave more altruistically than their peers. Attitudinal evidence from the same survey confirms this pattern. A third body of scholarship focuses on the role of networks in shaping the preference for public employment. On the supply side, scholars have long studied how social networks and influential relationships affect the allocation of valuable commodities, like jobs, in communist systems. Considerable evidence documents the importance of guanxi in communist China and blat in Soviet-era Russia—that is, personal connections and informal networks of exchange—for obtaining opportunities within the state (e.g., Bian 1994). More specifically, in post-communist Russia, the importance of family ties and personal connections for establishing one’s career has been linked to low intergenerational social mobility (Yastrebov 2011). Access to networks impacts the demand for government jobs by shaping individuals’ strategic considerations.2 Along these lines, Liu & Wang (2017) offer evidence from a natural experiment in China that students who perceive their universities as having strong alumni networks in government are more likely to pursue government careers. Comparing students just above and just below the threshold for admission on the nationwide university entrance exam, they find that attending one of China’s three most selective universities increases the probability of joining the civil service by 25 percent. Having

2. See, e.g., Fox & Lawless (2005) for a similar argument.

138

chapter 5

a network, whether through school or family ties, plausibly heightens the expectation of a successful public-sector career. Drawing on these bodies of literature, this chapter asks what motivates aspirants to public-sector careers in an autocratic country, such as Russia. In terms of character and ambition, how do future public servants compare with their peers? In what ways do their aspirations differ? Does career stability, greed, prestige, or a strong public service motivation animate those who intend to take up public-sector professions? A second important set of questions, which extends the analysis in previous chapters, concerns whether those who elect to work for the state hold different political values from their peers. Do they differ in how much they value freedom, political order, global influence, and economic growth? Do they differ in their views of the type of political system that is best able to achieve these goals? In short, how different are they from their peers in terms of what they want from their government and what type of government is likely to deliver it? Finally, a third important set of questions regards the background of public servants. Are their political and economic experiences similar to or different from their peers? Were their families more likely to be winners or losers in the transition from communism? What kind of educational and occupational profile do the parents have? These are critical questions given the lasting impact of adolescent socialization, the importance of networks, and the role of familial ties for enhancing perceived public-sector career prospects.

Who Goes Public? Motivation, Networks, Values, and Background Though there is limited empirical evidence beyond the developed democracies on what makes public-sector positions desirable and how public officials differ from their peers, extant studies suggest several relevant variables. First, there is a large body of work on patronage and clientelism, which implies that the primary motivation for seeking state employment in a developing nondemocracy is personal enrichment. Even if formal salaries are low, official positions are often a lucrative source of informal rents. Working for the state may provide the chance to solicit bribes, gain access to state resources, or obtain preferential treatment. Indeed, an emerging line of research considers whether corrupt types self-select into the government sector where corruption is widespread (Gans-Morse et al. 2017; Hanna & Wang 2017; Banerjee,

c h o o s i n g t o w o r k f o r t h e s tat e 139

Baul & Rosenblat 2015). If material incentives drive the decision to enter the public sector, we would expect students who prioritize high earnings or generous benefits to express a stronger preference for state employment. Second, students may make occupational choices in response to perceived career prospects, choosing to go into jobs and organizations where they are likely to benefit from preexisting networks (Liu & Wang 2017). Such networks may be essential to accessing public-sector opportunities when recruitment is not strictly meritocratic and jobs are doled out as a form of patronage. Even where recruitment is meritocratic, preexisting networks may facilitate access to preferable posts and career advancement. Networks can come from several sources. As Liu & Wang (2017) show, attending a school with a strong alumni network in government may encourage students to pursue a government career. Elite universities often supply such networks. Another source of network-based recruitment is family ties. Children who follow their parents into the same profession may benefit from their parents’ professional contacts and connections. A parent who works for the state can make introductions, provide know-how, or influence colleagues and acquaintances, enhancing their child’s career prospects in the government sector.3 In sum, preexisting networks open opportunities that plausibly enhance the motivation for choosing a public-sector career. If networks are an important factor in the choice of state employment, we would expect that both attending a school with a strong alumni network in government and having a parent who works in the public sector will increase the likelihood of a public-sector career. Third, a variety of studies have shown that political attitudes affect economic behavior (e.g., Gerber & Huber 2009; McConnell et al. 2018). Yet much of this literature focuses on partisanship, an especially strong form of political identity. For example, recent experimental evidence that reservation wages are lower when a prospective employer shares a subject’s partisan identity implies a preference to work for co-partisans (McConnell et al. 2018). We have less evidence on whether other political attitudes and orientations influence employment choices.

3. It is beyond the present study to definitively parse the effect of family ties that runs through networks from the effect of family ties that runs through human capital transfers, or transfers of knowledge that heighten the probability of success in securing a government job. However, after controlling for parental education and occupation, any additional effect of having a public-sector parent can plausibly be attributed to nepotism and networks. See, e.g., Lentz & Leband (1989).

140

chapter 5

One possibility is that political values shape the choice of career path. More concretely, students who elect public-sector professions in an autocratic system such as Russia’s may have systematically different views on the importance of political freedom, economic growth, order, and their country’s global influence than others. Another possibility is that they differ from their peers on what type of political system they believe is best suited to achieving these objectives. In recent years, Russia’s political system has become increasing autocratic. Key democratic institutions have been eroded and political competition has been limited. As political freedom has declined, the Kremlin has aimed to secure legitimacy through economic growth, political stability, and rising global influence. Are Russians’ political values, and especially their views on democracy, as compelling a factor in the choice of employment as is partisanship in the American context? While partisan affiliations are a key component of social identity in the United States, there are reasons to doubt that democratic values are as central to identity or as powerful in determining economic behavior in Russia. For one, apoliticism, and even political apathy, have long been a mode of survival under Russia’s autocratic leadership. In the late communist period, opportunism rather than ideology more often motivated the choice to join the Communist Party. In short, Russian students may prefer to work for an employer who shares their political values, or they may give political values and ideology relatively low priority next to personal, pragmatic considerations. With these points in mind, if political preferences are an important criterion in the choice of government employment, then we would expect prospective public employees to hold more authoritarian values. Fourth, the case-specific literature on Russia has long emphasized differences between those who see themselves as winners and losers in the transition from communism (Tucker, Pacek & Berinsky 2002; Tucker 2006). The collapse of communism brought radical changes in the social status of various groups. While some groups gained, other groups lost in Russia’s dramatic economic and political transformation. For those whose families benefitted from free market reforms, private sector opportunities may be more attractive than for those whose families fared poorly. Conversely, losers from the transition may be more likely to seek the security of state employment. If transition experiences are determinative, we would expect that the preference for public employment will be stronger among students whose families lost out in the post-communist transition.

c h o o s i n g t o w o r k f o r t h e s tat e 141

Another potentially important factor to consider is that the structure of the labor market constrains prospects in certain professions. For certain careers, there is no opportunity (or virtually none) to pursue employment in the private sector. This is almost universally true of those seeking a career in the military, police, or other occupation associated with maintaining public order. In the Russian economy, however, this is also true of occupations in the fields of public health, research, and education. In addition, it is true in industries that the state dominates like rail, aviation, banking, oil, and gas. Younger workers may be especially affected by a lack of private-sector opportunities if they have a strong desire for a particular occupational profile or interest in a specific industry upon entering the labor market. If these structural constraints are significant, we would expect that the preference for public employment will be stronger among students who prioritize their chosen profession as compared with students whose employment preferences are less closely tied to their majors. Finally, several other factors have been cited as drivers of the preference for public employment across other contexts. In many places, public employment is prestigious. In China, for example, a government job adds to personal and family status (Liu & Wang 2017; Lee & Zhang 2013). Public-sector jobs also frequently provide more stability than private-sector jobs, with a lower probability of being laid off. This argument implies that public employment will appeal most to those who are risk-averse and place a high value on job security (Lewis & Frank 2002; Dong 2014; Houston 2014; though see also Gabris & Simo 1995). In addition, the preference for flexibility is an important job value for many labor market participants (Karl & Sutton 1998). If parttime and flexible work arrangements are perceived as more prevalent in the private sector, it may drive some out of public employment. Finally, we would expect a strong public service motivation—defined as placing a high value on serving one’s country—to be associated straightforwardly with a preference for the public sector (Lewis & Frank 2002; Houston 2014). I control for each of the factors just mentioned in subsequent analyses.

Data and Measures To test these hypotheses, I draw on a survey I designed in collaboration with researchers at the Higher School of Economics in Moscow. The survey was conducted in late 2014 among 1,399 undergraduate students at three elite Russian universities: the Higher School of Economics (HSE, N = 540), Moscow

142

chapter 5

State University (MGU, N = 496), and the Moscow State Institute for International Relations (MGIMO, N = 363). The survey included detailed information about students’ career plans, political values, family backgrounds, and personal characteristics. Although students at these three institutions are not representative of students pursuing higher education in Russia as a whole, they are indicative of the group that already is, or stands the best chance to join, its middle class.4 These three institutions are consistently ranked among Russia’s top universities and are widely considered its most prestigious. Accordingly, their alumni are well represented in elite government circles. In the year following the survey, the schools’ alumni held five of eight deputy prime minister positions in the Russian government. MGIMO is run by the Russian Ministry of Foreign Affairs and has traditionally trained Russia’s diplomatic corps. Approximately 45 percent of the HSE’s alumni in the social sciences begin their professional careers in the state administration. Moscow State University is one of the country’s oldest universities and is much larger, with a broader educational mission in the natural sciences, than the other two. It remains at the top of several rankings of Russian universities and sends many of its graduates to public-sector institutions. The primary dependent variable in subsequent analyses is the preference for state employment. However, in view of the differences between working in public administration and other government structures, I also analyze separately the group that expects to directly serve the state. In total, approximately one-third of the full sample say that upon finishing university they expect to work in the public sector. Of these, 23 percent anticipate working in public administration. Another 23 percent expect to work in a state-owned enterprise, and 55 percent plan to work in a public medical, educational, or research institution. The full distribution of responses to the survey’s post-graduation employment question is presented in table 5.1. Next, I measure job values and motivations with a battery of fourteen items. For each item, respondents are asked to evaluate the importance to them personally of various job characteristics on a four-point scale ranging from not important (1) to very important (4). Figure 5.1 shows the average response for each of these items. The questions on salary and benefits 4. As elsewhere in this book, I mean Russia’s middle class sociologically defined, as many of these students have above-median incomes. Insofar as this group also includes members of Russia’s economic and political elite, we would expect to overestimate their ability to make ideological investments, rather than choose a career for pragmatic reasons.

c h o o s i n g t o w o r k f o r t h e s tat e 143 table 5.1. Distribution of Anticipated Careers Public-sector medical, educational, research Public administration State-owned enterprise Military, police, law enforcement Private sector/corporate Self-employment Journalism Other Difficult to say

0.17 0.07 0.07 0.01 0.42 0.08 0.08 0.02 0.08

Note: Share of respondents that expects to have each of the given careers.

tap extrinsic financial motivations, while the question on the importance of finding work that is directly related to one’s major taps a respondent’s commitment to pursuing a career in a particular field over other considerations. The item on prestige captures the importance of social esteem, while the items on risk of being fired and the desire to do something novel/risky capture the extent to which respondents prioritize job security. I also measure the importance of flexible hours, and assess public service motivation with a question on the importance of serving one’s country. Additional items measure the importance of opportunities for career advancement, work-life balance, the chance to make contacts, specialization vs. broad professional skills, and the ability to work independently. I include the latter items, which are standard in the literature, as controls. To capture political preferences and orientations, I use several items that measure these attitudes in different ways. First, I use a question which asks whether democracy is always preferable, treating democracy as an abstract concept. Second, I use an index consisting of the average of seven items measuring respondents’ views more concretely on the importance of particular democratic institutions: free and fair elections, independent courts, equality before the law, freedom of speech, an independent press, minority rights, and a strong political opposition.5 These measures are identical to those used in the cross-national analysis in chapter 3. Given Putin’s increasing authoritarianism, if political values affect employment choices, I expect that the 5. The resulting measure ranges from −2 to 2, with −2 indicating “not at all important” and 2 indicating “very important.” An index consisting of the average of these responses was chosen over an additive index in order to preserve observations for which at least one item was missing. The results are similar using either approach.

144

chapter 5

Career advancement Work-life balance High salary Specialize in clearly defined field Chance to make contacts Gain diverse professional skills Work independently Low risk of being fired Good benefits Prestigious Serve country Directly related to major Flexible work schedule Do something novel/risky 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

figure 5.1. Job values and motivations. This figure presents the average response and 95% confidence intervals for survey questions measuring the importance of different job characteristics. The response options were not important (1), rather unimportant (2), rather important (3), and very important (4). Data source: Author’s survey of students at three Moscow universities.

preference for public employment will be weaker among students who give greater importance to democracy and democratic institutions. Third, I measure authoritarian attitudes by examining the trade-offs that respondents are willing to accept. Specifically, the question asks which of the following respondents consider most important: freedom, stability, economic growth, the absence of corruption, a fair chance to succeed, or redistribution from rich to poor. I then use this item to investigate whether the preference for public service is stronger among students who prioritize things like order and economic growth over political freedom. In addition to ranking national priorities, survey respondents were asked whether democracy or autocracy is better at providing freedom, stability, economic growth, low levels of corruption, equality of opportunity, and redistribution. Based on a comparison of these two items, I coded an indicator for whether the respondent believes that an autocratic system is best suited to delivering on his/her top national priority. I use this measure to test whether the preference for public employment is stronger among students who believe that autocracies are better at providing for the things they consider national priorities.

c h o o s i n g t o w o r k f o r t h e s tat e 145

Besides looking at these explicit measures of the preference for autocracy and the willingness to trade off democracy for political stability and growth, I also leverage several indirect measures of authoritarianism. A first set of items taps chauvinistic political values such as the belief that Russia will be respected only if it is feared and the conviction that one’s own country is better than all others. A second set of items taps the willingness to accept hierarchy at home and at work as well as the belief that those in power should be feared. Again, if authoritarian dispositions affect employment choices, I expect the preference for public employment will be stronger among students who express more xenophobic values and a greater acceptance of hierarchy in different facets of their lives. Next, as indicators of the role that networks play in the decision to enter government service, I focus on two items. The first captures whether having a parent who works in the public sector increases the likelihood of a government job. For the second, I include a dummy variable for each university. If alumni networks matter, I expect that the preference for government service will be stronger among students of MGIMO, which is connected to the Ministry of Foreign Affairs and has long served a central role in training Russia’s diplomatic core. Of the three institutions, the HSE has the strongest network in the private sector. I would thus expect the link with public employment to be weakest there, though still stronger than at other Russian universities. Finally, I measure transition experience using an item that asks whether the respondent’s family is better off now or before the collapse of communism. I code as winners those who say that their families live better today and as losers those who say that their families lived better in Soviet times. I also control for several background characteristics: gender, orthodox religion, the type of settlement where the respondent’s family lives, a proxy for family income, parents’ education and occupation, as well as the respondent’s field (social sciences and humanities vs. natural sciences) and class year. I control for gender, since previous research has found that men are more likely to take up careers in the state administration, military, and law enforcement, while women are more likely to major in education, medicine, and science (Gerber & Schaefer 2004). I control for religion given the state’s close association with the Russian Orthodox Church in recent years (Anderson 2007). I include settlement type as well as various proxies for income and social status, since state employment often provides upward mobility (or “social lift”) for those who are less well off or from the provinces.

146 c h a p t e r 5 table 5.2. Willingness to Accept Authoritarianism

Authoritarianism sometimes preferable

Private Sector

Public Sector

0.60

0.61

Note: Share of respondents in each category agreeing with the statement.

Empirical Results Before presenting the main regression results, it is worth noting that even a simple comparison casts doubt on the prevalence of political self-selection. Table 5.2 presents respondents’ views on whether authoritarianism is sometimes preferable. Clear majorities agree with this proposition among those who desire to work in both the public (61%) and private (60%) sectors. In terms of their tolerance for autocracy, these figures suggest that aspirants to Russia’s state and private sectors are initially identical. I next systematically test the propositions just outlined for public-sector employment as a whole and then separately for the state administration. Across all analyses, I estimate logistic regression models predicting employment choice as a dichotomous variable. Because the hypothesized correlates of this choice—background, networks, job motivations, and values—lie at different points in the causal chain, I estimate my main models in four steps. In each step, I include factors that lie at the same or prior stages of the causal chain. Thus, in the first column of table 5.3, I include only background characteristics: gender, religion, settlement type, family income, transition experience, parents’ education and occupation, whether a parent is employed in the public sector, university, class year, and field. Model 1 suggests that proximity to public-sector networks and transition experiences have a significant impact on the likelihood of entering the public sector. Students at MGIMO and MGU are much more likely than students at HSE to plan public-sector careers. This is consistent with MGIMO’s strong ties to the state administration and MGU’s influential network in public education and research. By contrast, HSE has the strongest private sector alumni network of the three institutions, and more of its students appear to gravitate in that direction. Likewise, students who have a parent working in the public sector are more likely than others to make the same career choice. This is true controlling for both parental education and occupation, suggesting that parents’ informal networks matter more than intergenerational transmission of human capital. Finally, those whose families lost during the post-communist

table 5.3. Models Predicting Public-Sector Employment Dependent Variable: State Employment

Background (1) Prioritizes stability Prioritizes economy Prioritizes corruption Prioritizes equality Prioritizes redistribution Autocracy delivers Democracy preferable Democratic institutions index Salary Benefits Serve my country Prestige Stability Independence Related to major Broad skills Specialization Career growth Risky/novel Flexible Work/life balance Make contacts Public sector parent MGIMO MGU Transition loser Male Urbanicity Orthodox Parents’ education Parents’ occupation Income Constant Observations Log likelihood Akaike inf. crit.

Motivation (2)

−0.158 (0.148) 0.037 (0.115) 0.317∗∗ (0.098) −0.113 (0.115) 0.095 (0.109) −0.069 (0.118) 0.162 (0.096) 0.008 (0.113) 0.025 (0.123) −0.386∗ (0.159) 0.186 (0.107) −0.236∗ (0.106) −0.055 (0.124) 0.125 (0.118) 0.372∗ (0.179) 0.307∗ (0.149) 0.609∗∗ (0.224) 0.418∗ (0.189) ∗∗∗ (0.172) 0.910∗∗∗ (0.204) 0.997 ∗∗∗ (0.198) 0.705∗∗ (0.240) 0.688 −0.254 (0.148) −0.387∗ (0.179) −0.051 (0.037) −0.063 (0.044) −0.286 (0.149) −0.442∗ (0.184) 0.126 (0.146) 0.144 (0.171) −0.194 (0.184) −0.308 (0.221) 0.036 (0.053) 0.031 (0.065) −0.969 (0.562) −0.170 (1.006) 938 −576.007 1,178.015

717 −422.298 898.595

Political Values (3)

Political Values & Priorities (4)

0.098 (0.206) −0.080 (0.142)

0.215 (0.365) −0.147 (0.312) 0.096 (0.316) −0.242 (0.334) −1.111 (0.611) 0.097 (0.208) 0.111 (0.212) −0.051 (0.147)

−0.166 (0.151) 0.033 (0.116) 0.321∗∗ (0.099) −0.106 (0.115) 0.097 (0.109) −0.036 (0.120) 0.150 (0.096) 0.006 (0.115) 0.027 (0.124) −0.334∗ (0.161) 0.181 (0.108) −0.246∗ (0.107) −0.035 (0.127) 0.122 (0.120) 0.367∗ (0.180) 0.605∗∗ (0.227) 0.916∗∗∗ (0.204) 0.727∗∗ (0.242) −0.383∗ (0.181) −0.063 (0.044) −0.442∗ (0.186) 0.155 (0.172) −0.325 (0.221) 0.025 (0.066) −0.359 (1.017)

−0.173 (0.153) 0.034 (0.118) 0.321∗∗ (0.101) −0.117 (0.118) 0.073 (0.111) 0.002 (0.123) 0.168 (0.098) −0.003 (0.116) 0.043 (0.126) −0.351∗ (0.164) 0.183 (0.109) −0.248∗ (0.108) −0.034 (0.128) 0.122 (0.122) 0.366∗ (0.182) 0.611∗∗ (0.230) 0.931∗∗∗ (0.208) 0.779∗∗ (0.246) −0.436∗ (0.184) −0.075 (0.045) −0.507∗∗ (0.190) 0.171 (0.173) −0.363 (0.225) 0.022 (0.066) −0.285 (1.042)

711 −417.849 893.698

707 −411.247 892.495

Note: ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. Logistic regressions predicting public-sector employment. All models include dummies for class year and field. The reference category is “freedom” for the question on priorities and the HSE for university.

148 c h a p t e r 5

transition are more likely to choose the public sector, perhaps because they feel their prospects in the private sector are limited or because they perceive government employment as more secure. Model 2 adds the items on career motivations. Each of the findings from the first model hold, and the coefficients remain remarkably stable. At the same time, the results imply that a strong public service motivation also drives the choice of government employment, consistent with Gans-Morse et al. (2017). These findings also lend modest support to the notion that a lack of private-sector options in certain professions pushes otherwise similar students into the public sector. The desire for a career that is closely related to one’s major is positive, as expected, though only marginally statistically significant (p < 0.1). There is also some evidence that future public employees are less ambitious, or at least anticipate less career growth (perhaps because professional jobs in, for example, education and medicine imply a less hierarchical career ladder than positions in large private sector organizations). Notably, neither the value placed on salary and benefits nor job stability clearly separates future public- from private-sector workers. Finally, the results suggest that public-sector jobs are more appealing to those who are willing to sacrifice flexibility. This makes sense given that parttime and informal employment arrangements are concentrated in the private sector. In Model 3, I consider the impact of political values on career choice. If people with more authoritarian values self-select into the state sector, it would cast doubt on the notion that public employment independently shapes the political preferences of the state middle class. The results in column 3, however, confirm that neither the belief that democracy is preferable to other political systems nor the preference for democratic institutions is significantly related to the likelihood of a public-sector career. Both coefficients are close to zero and statistically insignificant. This implies that people choose state employment for other reasons, and thus that the state must work to secure their loyalty through socialization and incentives. Model 4 probes this null relationship further. The results suggest that the preference for government employment is not significantly related to prioritizing stability, economic growth, transparency, equality, or redistribution over political freedom. In fact, as figure 5.2 shows descriptively, the same priorities are widely held among students who intend to pursue careers both inside and outside of government. Additionally, I find no evidence that students who believe that autocracies are better able to deliver on the things they prioritize

c h o o s i n g t o w o r k f o r t h e s tat e 149

Public sector

Private sector

Redistribution Fair chance to succeed Lack of corruption Economic growth Stability Freedom 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.00

0.05

0.10

0.15

0.20

0.25

0.30

figure 5.2. National priorities. The figure shows that students planning to enter the public sector from elite universities share similar views on national priorities. Economic growth and combatting corruption are viewed as top priorities among both future public servants and others. The bars are 95% confidence intervals. Data source: Author’s survey of students at three Moscow universities.

are more likely to want a public-sector career. Again, the coefficient is close to zero and statistically insignificant. I next repeat these tests only for the group that is most likely to pursue a career in the state administration, a subset of all public employees. The results in column 1 of table 5.4 are broadly similar to the results in column 1 of table 5.3. Access to both parental and university networks remains a strong driver of employment in the state administration. I also find that the relative size of the coefficients changes on the variables for the universities, consistent with MGIMO’s stronger alumni network in public administration. As before, public service motivation matters too. Model 2 again adds the additional items on career motivations. While the coefficients on the variables from the preceding model remain very similar, several new patterns emerge when comparing the job motivations of those entering the public sector as a whole to the motivations of those entering public administration. In particular, I find that those entering public administration are less likely to value working independently than those entering either other government budget-sector professions or the private sector. Additionally, the choice of major less clearly dictates a career in the state administration, consistent with the skills necessary to succeed in this area being more easily transferable to other white-collar positions. There is also no evidence that students who plan to work for the state administration are less

table 5.4. Models Predicting Employment in State Administration Dependent Variable: Public Administration

Background (1) Prioritizes stability Prioritizes economy Prioritizes corruption Prioritizes equality Autocracy delivers Democracy preferable Democratic institutions index Salary Benefits Serve my country Prestige Stability Independence Related to major Broad skills Specialization Career growth Risky/novel Flexible Work/life balance Make contacts Public sector parent MGIMO MGU Transition loser Male Urbanicity Orthodox Parents’ education Parents’ occupation Income Constant Observations Log likelihood Akaike inf. crit.

Motivation (2)

−0.305 (0.295) 0.075 (0.252) 0.420∗ (0.210) 0.324 (0.239) 0.177 (0.217) −0.431∗ (0.219) 0.223 (0.201) −0.006 (0.239) 0.100 (0.260) −0.246 (0.334) 0.381 (0.219) −0.607∗∗ (0.214) −0.105 (0.252) 0.144 (0.244) 1.023∗∗∗ (0.294) 1.039∗∗ (0.355) 1.327∗∗∗ (0.346) 1.554∗∗∗ (0.431) 1.374∗∗ (0.468) 1.048∗∗ (0.388) ∗ 0.667 (0.470) 0.784 (0.373) 0.397 (0.346) 0.580∗ (0.280) −0.019 (0.074) 0.036 (0.088) 0.197 (0.279) −0.120 (0.357) −0.163 (0.249) −0.207 (0.308) −0.469 (0.320) −0.528 (0.419) 0.146 (0.101) 0.068 (0.124) −5.634∗∗∗ (1.232) −6.104∗∗ (2.196) 661 −186.190 398.381

504 −130.712 315.424

Political Values (3)

Political Values and Priorities (4)

−0.025 (0.432) −0.313 (0.265)

0.507 (0.417) 0.929 (0.804) 0.251 (0.739) 0.249 (0.768) 0.350 (0.775) 0.173 (0.466) −0.238 (0.289)

−0.323 (0.299) 0.096 (0.255) 0.356 (0.212) 0.314 (0.240) 0.192 (0.216) −0.369 (0.223) 0.223 (0.204) −0.011 (0.241) 0.116 (0.262) −0.162 (0.343) 0.390 (0.222) −0.610∗∗ (0.214) −0.074 (0.257) 0.147 (0.244) 1.010∗∗ (0.355) 1.563∗∗∗ (0.433) 1.418∗∗ (0.474) 0.641 (0.484) 0.386 (0.347) 0.030 (0.088) −0.087 (0.359) −0.217 (0.319) −0.490 (0.426) 0.075 (0.128) −6.232∗∗ (2.222)

−0.318 (0.310) 0.090 (0.264) 0.345 (0.222) 0.274 (0.248) 0.091 (0.224) −0.241 (0.237) 0.291 (0.214) −0.029 (0.252) 0.172 (0.278) −0.191 (0.362) 0.409 (0.231) −0.655∗∗ (0.221) −0.118 (0.266) 0.148 (0.250) 0.973∗∗ (0.367) 1.599∗∗∗ (0.447) 1.498∗∗ (0.485) 0.782 (0.497) 0.272 (0.362) 0.011 (0.089) −0.243 (0.376) −0.222 (0.322) −0.550 (0.436) 0.072 (0.130) −6.659∗∗ (2.297)

501 −129.430 316.859

483 −123.078 314.156

Note: ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001. Logistic regressions predicting employment in public administration. Other prospective public-sector employees have been excluded from the sample, such that private-sector employment is the reference category. All models include dummies for class year and field. The reference category is “freedom” for the question on priorities and the HSE for university.

c h o o s i n g t o w o r k f o r t h e s tat e 151

ambitious than their peers, as seemed to be the case for other public-sector professions. Next, if the state administration is more highly politicized than educational or medical institutions, political values may play a more prominent role in the choice to serve the government directly than in the preference for a budget-sector position. In Model 3, I explore this possibility, but find only weak evidence in its favor. The notion that democracy is preferable to other systems is, as before, unrelated to the choice of a career in the state administration. While the preference for democratic institutions leans negative and the coefficient is somewhat larger than before, it still does not approach conventional levels of significance. Incorporating the additional items on national priorities in column 4 further confirms that political values are, at best, weakly related to the choice to join the state administration. By contrast, attending a university with a strong alumni network or having a parent in the public sector are far stronger predictors. I also probed the robustness of these results.6 As a first step, I disaggregated the democratic institutions index into its component parts: views on the importance of elections, free speech, a free press, independent courts, equality before the law, minority rights, and a strong opposition. Also, because one might be concerned about the causal order of political values and job motivations or the overall number of parameters estimated in the preceding models, I estimated these models with only the controls for background characteristics. The results predicting employment in any part of the public sector are broadly consistent with the preceding models. Only one of the democratic institutions items is significant, minority rights, and it is signed positive— meaning that those who favor minority rights are more likely to seek state employment than others. Meanwhile, as before, networks and losing out in the post-communist transition remain clear drivers of the preference for public employment, though attending MGIMO (p = 0.07) and having a public sector parent (p = 0.08) are just shy of the conventional 0.05 threshold for statistical significance. Next, I investigated whether more autocratic types show a preference for state employment, using the implicit measures discussed earlier. Rather than measure the preference for democracy directly, I exploit items that tap chauvinism and deference to authority as indirect measures of antidemocratic tendencies. None of these five items is significantly related to state employment. Indeed, the coefficients are small and the signs inconsistent. As a whole, these 6. All alternative specifications are reported in table C.1 of the online appendix.

152

chapter 5

tests roundly confirm the importance of networks and transition experience over political self-sorting. Finally, I perform the same tests in models that predict employment in the state administration. Again, there is little evidence of political self-selection. The signs on the individual democratic institutions items are mixed and none of the coefficients is statistically significant. Similarly, there is no clear relationship between implicit authoritarianism and the likelihood of entering public administration. These models also reinforce the earlier findings regarding networks. Once more, strong networks that enhance career prospects— having a public-sector parent or attending a university with many alumni in government—are the clearest predictors of a civil service career.

Discussion The desire to serve one’s country does draw some to public-sector careers, especially in health and education. However, opportunism rather than altruism or ideology is a more important factor in the career choices of Russia’s elite university students. Although I did not find evidence that the preference for public employment is stronger among students who prioritize financial gain, I found clear, consistent support for the importance of networks— institutional and familial.7 Both attending a school with a strong alumni network in government and having a parent who works in the public sector significantly increased the likelihood of a public-sector career. Being a student at MGIMO was a strong predictor of joining the state administration, all else equal. Similarly, having a public-sector parent significantly increased the likelihood that a student believed she would obtain a potentially lucrative spot in the bureaucracy. The latter finding echoes existing work, which demonstrates strong intergenerational effects in Russia’s labor market (Borisov & Pissarides 2016; Yastrebov 2011). It underscores the importance of informal networks for securing positions in the public sector among those with 7. Measuring financial incentives in terms of salary and benefits has obvious drawbacks. Students entering the government in Russia can expect to make lower salaries than their counterparts in the private sector. Still, opportunities to exploit one’s official position—informal benefits that the available survey questions may not fully capture—could alternately explain the preference for a public-sector profession. In a closely related study also among students at the HSE, however, Gans-Morse et al. (2017) find that those who are willing to cheat and bribe self-select out of state careers.

c h o o s i n g t o w o r k f o r t h e s tat e 153

similar backgrounds and elite educations. Indeed, the importance of having a public-sector parent held even after controlling for parental education and occupation (which were not significant predictors of a public-sector career). This suggests that the effect of having a public-sector parent can more likely be attributed to nepotism and networks than the intergenerational transfer of human capital. By contrast, I find very little evidence that the preference for public employment will be stronger among students who hold more authoritarian values. Across all models, I fail to find that the preference for democracy, attitudes toward specific democratic institutions, the relative importance of political freedom versus other priorities like economic growth and stability, or the view that autocracies are better at delivering on these priorities significantly influences the choice of a public-sector career. What is more, I find little evidence to suggest that students who, even implicitly, hold more authoritarian attitudes self-select into government employment. Clearly some do gravitate toward government employment for ideological reasons. However, such considerations appear rarely to be paramount, including for those entering the state administration. The results provide mixed evidence on whether transition losers are more likely to prefer state careers. In short, the answer appears to vary across different parts of the public sector. While the relationship between postcommunist transition experiences and employment choice is weak for the state administration, it is robust for the state sector as a whole. This suggests that transition losers are more likely to take budget sector positions that are less prestigious and less well remunerated, while winners and losers alike choose positions in the state administration. Finally, there is some evidence that strongly desiring a particular occupational profile related to one’s major increased the likelihood of a budget-sector career. It was not, however, significantly associated with the likelihood of employment in the state administration. The results thus imply that at least one reason why students with certain majors expect to have government jobs is that their alternatives are limited. While the skill set for positions in research, medical, and educational institutions is highly specialized, the skills required for a position in the state administration are more readily employed in other white-collar and managerial positions in the public or private sector. Thus, skill specificity and the structure of the labor market compel students in certain fields to take up jobs in the public sector.

154 c h a p t e r 5

In terms of altruistic motives for a state career, these findings largely echo Houston (2014, 851), who concludes that the public service motivation remains weak in six former communist bloc countries—but stronger in the budget sector than in public administration.8 Like Houston (2014), I also find that the similarities between public and private-sector workers in Russia are more pronounced than the differences, specifically with regard to extrinsic motivations such as salary and benefits. Both sets of findings cast doubt on the notion that the public and private sectors attract very different types of people.

Conclusion The evidence in this chapter thus suggests that insofar as public and private sector workers hold different attitudes and engage in different political behaviors, these differences emerge as a consequence of career choice. They are not primarily the result of self-sorting. Rather, they are incentivized and socialized through the workplace and arise in the context of employment.9 With regard to bribe-taking and other illicit behavior, Gans-Morse et al. (2017) conclude similarly that “corruption in Russia results more from the transformation of bureaucrats’ behavior and attitudes after joining the civil service, rather than from a process of corrupt self-selection” (5). In sum, then, this evidence supports the view that Russia’s large public sector still serves—as did universal state employment under communism (Kornai 1992; Róna-Tas 1997)—as an important tool for managing the economic self-interest and loyalty of citizens. The next chapter provides further evidence on the importance of the state sector for securing the loyalty of an expanding middle class, as well as a strong research design to address self-selection in yet another context. Focusing on the case of Ukraine, I analyze change over time, showing how changes in labor market position and upward mobility into the middle class affect attitudes and actions that alter democracy’s chances.

8. Gans-Morse et al. (2017) are more optimistic about the role of altruistic motivations in Russia, though they again appear concentrated in the budget sector. 9. This conclusion is consistent also with Frye, Reuter & Szakonyi (2014 and 2018).

6 Revolution, Democratic Retrenchment, and the Middle Class

during the orange revolution, Ukrainians took part in mass protests that overturned fraudulent election results and produced a peaceful transition of power. Ukraine’s successful civic revolution, which occurred amid prolonged economic growth, seemed at least initially to vindicate the link between development and democratization. Following a profound economic contraction in the 1990s, Ukraine’s economy had begun to rebound around 2000 and continued to expand each year until the global financial crisis. Over the period preceding the Orange Revolution, poverty in Ukraine registered one of the sharpest declines of any transition economy.1 GDP growth peaked at 12.1 percent in the year prior. Real wages rose by 24 percent, then increased another 20 percent the following year (World Bank 2007). Declining unemployment further contributed to rising standards of living. Western observers of Ukrainian politics initially greeted the Orange Revolution as a democratic breakthrough. However, Ukraine’s democratic performance improved only haltingly in the months after. Popular discontent intensified and reforms quickly stalled. By 2006, the country’s parliamentary election campaign focused largely on the revolution’s failures. Those who had opposed it united with those who had become skeptical about its prospects to support the political forces that the Orange coalition had defeated (Copsey 2006). The Party of Regions’ comeback victory, which

1. According to the World Bank’s World Development Indicators. 155

156 c h a p t e r 6

paved the way for Viktor Yanukovych’s return as prime minister later that year, set Ukraine on a path of democratic retrenchment.2 I began the empirical portion of this book with a wide lens, documenting the existence of an autocratic middle class across the broad post-communist experience. I next showed that middle-class state dependency helps to explain not only attitudes but consequential political actions, like protest participation. Then, in the preceding chapter, I paused to consider whether the political orientations of public-sector professionals are plausibly a function of state efforts at co-optation or are instead initially distinctive. The present chapter complements these cross-sectional analyses by testing the book’s argument dynamically, over a period of political transition. To gain leverage on how the state’s extensive involvement in the economy limited popular support for democratization in Ukraine and contributed to democratic retrenchment, this chapter exploits an unusual panel survey. The survey includes detailed employment, educational, and income data as well as key political variables for a nationally representative sample of Ukraine’s working-age population. Crucially for the design of this research, respondents were interviewed three times—twice before and once after the Orange Revolution. Longitudinal data such as these are rarely available from countries in the midst of transition and provide a unique opportunity to investigate how state economic engagement shapes the formation of pro-democracy coalitions at a time of real contention. My empirical strategy consists of three parts. First, I provide a strong causal test of the state dependency argument for the cohort entering the labor market during the panel. By following the same individuals over time and using a design resembling difference-in-differences, I am able to directly address critical questions about causal order and endogeneity. Next, I broaden the analysis to all Ukrainians and turn to two behavioral measures of regime preferences: protest participation during the Orange Revolution and voting in the country’s first post-revolutionary parliamentary election. Specifically, I examine who protested for the Orange forces and who took to the streets to defend the incumbent regime. Then I investigate how popular support for the Orange coalition unraveled through an analysis of vote switching.

2. The Party of Regions was the party of the preceding regime under Leonid Kuchma. Viktor Yanukovych, the defeated candidate in the 2004 election, was Kuchma’s handpicked successor.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 157

The results again demonstrate that dependence on the state for economic opportunity and life chances limits support for democracy. I find that new labor market entrants, and especially those with white-collar professions, became less supportive of democracy after joining the state sector. This change occurred quickly, consistent with a shift in material interests that affected how individuals viewed their future prospects under democracy. Since private political preferences may differ from the actions individuals take at pivotal political moments, I then connect these attitudes to consequential political behavior. I find that educated, white-collar professionals were more likely to participate in pro-democracy protests only if they were employed outside the state sector. Those with careers tied to the state were more likely to protest in defense of nondemocratic incumbents.3 Finally, controlling for other determinates of party support, I find that, as the revolution faltered, state employees whose jobs and rents were threatened by the prospect of democratic transition were more likely to defect to the Party of Regions than private-sector employees with similar characteristics. These results imply that public-sector institutions structure incentives that shape both attitudes and action. Together this chapter’s empirical analyses show how a lack of autonomy from the state affects actors’ attitudes and choices to participate in mobilized democratic contention.

The Ukrainian Case Background Ukraine has long been a crossroads of empire. Parts of Ukrainian territory were once under Austrian, Polish, Lithuanian, Mongol, and Russian rule. These divergent political histories continue to influence Ukraine’s political geography. As compared with the other post-Soviet republics, Ukraine’s strong regionalism has produced greater political contestation, but also significant obstacles to democratic consolidation.4 Indeed, the identities shaped by imperial-era institutions still define to an important extent Ukraine’s electoral map and durable divisions between the country’s political parties. As recent work shows, Ukrainians’ contemporary foreign policy orientations are 3. This latter finding is consistent with Beissinger’s (2013) evidence on counter revolutionaries. 4. Weak and incompetent leadership over the post-communist period has further contributed to what Lucan Way calls Ukraine’s “pluralism by default.”

158 c h a p t e r 6

crucially a function of these historical legacies (e.g., Peisakhin 2018; Rozenas & Zhukov 2019). At the same time, however, historically determined regional and ethnolinguistic identities are not the only, or even the most important, drivers of Ukrainians’ attitudes toward democracy. Peisakhin (2018), for example, notes that the impact of these variables on democratic attitudes is limited, and concludes that support for democracy must be explained by other factors. Consistent with this book’s broader argument, in this chapter I show that socioeconomic position and state dependency are important determinates of regime preferences in Ukraine. In addition to having different historical identities, Ukraine’s regions also have divergent economic interests. Overall, about 14 percent of Ukraine’s economy is in agriculture, 28 percent in industry, and 59 percent in services. While Western Ukraine’s economic base is largely agricultural and increasingly tied to Europe, Eastern Ukraine’s is industrial and remains closely intertwined with Russia. The analysis in this chapter shows that economic reliance on the state has a meaningful independent effect on middle-class support for democracy, even in a setting where regional and identity-based factors are often considered determinative. Moreover, given the size of its state sector and slow pace of privatization, the political effects of state dependency are substantively important in the Ukrainian context. Despite pressure from international financial institutions to reform, the state remains heavily invested in chemicals, mining, machine-building, energy, infrastructure, banking, and finance. As of 2017, the state still operated more than 1,800 SOEs5 and owned more than half of all banking assets (International Trade Administration 2017). The state is also Ukraine’s largest employer, providing jobs for about one million people in state enterprises, administration, and other public organizations. Stateowned enterprises are controlled directly by the Ministries, making political meddling in their operations pervasive. Most accounts of Ukraine’s poor track record on privatization focus on how Ukraine’s leaders have used the public sector to co-opt the country’s oligarchs. Indeed, this small group of economic elites effectively controls Ukraine’s largest state-owned enterprises and has opposed their sale. This chapter shows that besides securing the support of elites, extensive state involvement in the economy also limits mass-based support for 5. Some 3,350 were registered.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 159

democratization, including among the country’s middle class. Thus, a large public sector serves incumbents’ interests in other ways as well—not only helping them to maintain the backing of key elites but deterring mobilization in the face of popular revolt.

Public Sector Dynamics and the Orange Revolution Under Kuchma, who was president of Ukraine until early 2005, the state sector was a principal source of patronage. For more than a decade, rewards and punishment based on access to graft as well as workplace-based mobilization kept the state sector politically pliant (D’Anieri 2007; Darden 2008). Corruption was widespread. The state turned a blind eye to bribe-taking, allowing officials as well as teachers and doctors to supplement their official income. Gorodnichenko & Peter estimate that Ukrainian public-sector employees accepted bribes worth around 1% of Ukraine’s GDP in 2003.6 Threats of job loss and selective prosecution were commonplace throughout the state apparatus, down to ordinary state workers (Darden 2008, 51). At the time of the Orange Revolution, state enterprises still made up a substantial share of Ukraine’s economy and were also large employers. State ownership remained pervasive in the natural gas, mining, transportation, energy, machine building, and banking sectors. State enterprises received government subsidies in exchange for political support, despite being inefficient or even (as in the case of Ukrainian coal production) loss-making. The future prospects of these industries were clearly tied to political control of the state. The regime also held vast patronage power through the public bureaucracy and public organizations such as schools and hospitals. Every teacher and administrator in virtually every school in the country was an employee of the Ministry of Education (D’Anieri 2007). More than 90 percent of hospitals were public and every doctor in them was a employee of the Ministry of Health. With weak de facto employment protections, the jobs, promotions, and access to informal benefits of those working in education and health as well as state enterprises and public administration depended also on political considerations.

6. Gorodnichenko & Peter (2007) also find that employees in public administration, managers, and health professionals benefitted most.

160

chapter 6

Even after his departure, Kuchma left behind a corrupt, criminal legacy. Coercion of government employees continued following the Orange Revolution (D’Anieri 2007, 63 and 253). There are credible accusations that government employees were pressured to vote “correctly” in the 2006 parliamentary election (D’Anieri 2007, 214). Because civil service protections were effectively lacking, the state’s power over patronage remained great (D’Anieri 2007, 20). Around the time of the Orange Revolution, according to the ILO, approximately 27 percent of Ukraine’s total working-age population was employed in the public sector. With an overall labor force participation rate of 58 percent,7 about half of those who were employed held jobs in the public sector. Roughly 20 percent were employed in the state administration and public organizations, while just over 30 percent were employed in state-owned enterprises. Throughout this period, Ukrainian leaders used distributive policies aimed at improving the well-being of public sector professionals to secure middle-class loyalty. On the eve of elections in 2004, Ukraine’s government, under Prime Minister Viktor Yanukovych, raised public-sector wages. These policies dramatically impacted the dynamics of upward mobility over this period. Following the Orange Revolution, decisions about privatization and the public sector were among the new governing coalition’s most contentious. Victor Yushchenko lined up quickly in favor of fiscal restraint, coming into conflict with his Orange ally, Yulia Tymoshenko, who pushed further wage hikes for state workers through the Rada. These policies created substantial real wage growth for public-sector workers in 2005 and 2006. The number of public employees also increased modestly over this period. Within a year, however, Yushchenko had replaced Tymoshenko and Yanukovych’s Party of Regions (POR) was experiencing a resurgence of popularity. After the POR’s victory in parliamentary elections in 2006, Yanukovych returned to the office of prime minister. This study shows that the POR, and to a lesser extent Tymoshenko’s bloc, 7. Labor force participation rates are calculated by the World Bank on the basis of ILO data as the share employed out of a country’s working-age population (ages 15+). For 2003– 2004, this figure was 57%, while from 2005–2007, it was 58% in Ukraine. Neither labor force participation rates nor the share of public-sector employees has changed dramatically since 2006.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 161

benefitted from the defection of state workers who opposed western-style democratization. While the Orange Revolution was initially seen as a “democratic breakthrough,” it produced little momentum for democratic reform. Before the end of Yushchenko’s first 100 days in office, an editorial in the Kyiv Post opined: “The phrase ‘Orange Revolution’ is now synonymous in the Western imagination with democracy. But in terms of domestic politics, it seems things are much the same” (Sunden 2005). After 2005, growing support for Yushchenko’s rival, Viktor Yanukovych, and his Party of Regions, in effect ratified Ukraine’s retreat from democracy and return to the status quo ante. This process was well underway by 2006 when the POR won a comeback victory in parliamentary elections, and was all but complete when Yanukovych resumed the presidency in 2010. While Yanukovych did return to the presidency through elections rather than by other means, in crucial respects Ukraine maintained the corrupt, semi-autocratic political system of previous periods.8 In short, compared with the Central and East European democracies in chapter 3, Ukraine in this period continued to lack effective civil service protections, transparency, and strong rule of law.

Case Selection Focusing on the period of the Orange Revolution affords an unusual opportunity to examine the preferences and political behavior of key social groups at a moment when democracy was actually contested. Although revolutions are by definition exceptional, understanding how state dependency affects regime preferences is substantively most interesting and politically most crucial during moments of contention. This book’s framework sheds light on why democratic consolidation in Ukraine was not more successful. Again, the framework that this book tests rests on two contextual conditions: a large public sector and illiberal state institutions. In these respects, Ukraine shares important similarities with other late-developers in the former Soviet Union and beyond. Like many other late-developing countries, Ukraine has been plagued by pervasive corruption and insecure property rights. Direct state intervention in the economy is widespread and has had 8. Freedom House’s Nations in Transit governance measures attest to Ukraine’s contradictory democratic performance during these years (Freedom House 2013; Pop-Eleches & Robertson 2014).

162

chapter 6

a profound effect on the country’s economic and social structure. Compared to its post-Soviet neighbors, the size of Ukraine’s public sector around the Orange Revolution was not unusual. According to the ILO, the share of public-sector workers among Ukraine’s economically active population was similar to Georgia’s, and somewhat smaller than in Azerbaijan, Belarus, Moldova, or Russia. While Ukraine’s public sector is large and encompassing in certain industries, Ukraine still offers a “tough test” of the state dependency argument in at least three other respects: First, Ukraine lacks the resource rents that often underpin a large public sector, so it helps to show that the argument is relevant beyond a few countries with significant natural resource wealth. Second, and relatedly, Ukraine has had difficulty financing its public sector for much of the post-communist period. Wage arrears have been extensive and publicsector salaries have often been well below subsistence levels. The Ukrainian case thus helps to show that the argument holds even in places where formal incentives are small, as long as there are few alternative sources of economic opportunity or extensive informal rents to be gained by exploiting ones’ official position. Third, part of the period I examine is one of nominal democratic progress, or at least heightened democratic rhetoric. Ukraine is arguably both less authoritarian during these years than it was in other periods and less authoritarian than other countries in the region (though clearly also not a consolidated democracy). This helps to show that the argument does not hinge on strong repressive capacity. Finally, regional and linguistic identities strongly shape Ukrainians’ attitudes and political behavior. Where these identities crosscut economic interests, they potentially overwhelm the effect of labor market position. Where these factors overlap, regional and linguistic identity may take credit for part of the impact of economic factors. For all of these reasons, we would expect the observed effect of economic dependence on the state to be weaker here.9 Thus if the argument holds in this context, we might expect that it would also apply under “easier” contextual conditions: that is, where resource rents support a large and well-financed state sector, wage arrears are uncommon and state salaries are higher, the state’s repressive capacity is greater, and noneconomic lines of cleavage are less pronounced. 9. Though uncertainty may be heightened during a transition period, leading to greater risk aversion that could conceivably strengthen the effect of state dependency, this is precisely the setting in which understanding how state dependency affects regime preferences is substantively most interesting.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 163

Data and Empirical Approach One limitation of the analyses presented in preceding chapters is that they focused on cross-sectional relationships, measured at a single point in time. Rustow (1970) was among the first to note the weakness of cross-sectional designs for testing the causal relationships implied by modernization theory. His work rightly raised the possibility of alternative causal pathways. Indeed, much of the early literature adducing support for modernization theory was based on cross-sectional correlation rather than within case variation. This chapter exploits within-case and within-subject variation. By focusing on a single country, the analysis also removes remaining sources of cross-national heterogeneity that may have posed a threat to inference in chapter 3’s crossnational study. This chapter relies on a unique source of Ukrainian public opinion data, the Ukrainian Longitudinal Monitoring Survey (ULMS). The ULMS is a panel survey consisting of three waves. Panel data—which involve reinterviews with the same respondents over time—have signal advantages over cross-sectional data, yet are very rarely available for the democratizing countries in which we would ideally like to test our theories. The data that I employ in this chapter provide a unique research opportunity. Crucially for the design of this research, respondents were interviewed both before and after the Orange Revolution. The first wave took place in 2003, about a year and a half prior to the contested election that sparked massive protests on Kyiv’s central square. The second wave was conducted in the summer of 2004, just before the Orange Revolution. The third wave was conducted in 2007, approximately three years later, after Yanukovych’s return to the prime ministership. The initial sample was representative of the working-age population of Ukraine, aged 15 to 72, and comprised about 4,000 households and 8,500 individual respondents. The response rate in the first wave was 66 percent. In subsequent waves, panel respondents continued to be interviewed even if they moved to a new household. In total, just over 5,000 individuals participated in all three waves.

Research Design In subsequent sections, I present three mutually reinforcing types of analysis. First, I offer a rigorous causal test of the central hypothesis. Next I show that the argument also matters for political behavior. In addition to asking

164

chapter 6

respondents what type of political system they prefer in the abstract, the ULMS includes information about individuals’ demonstrated regime preferences. The second and third empirical sections examine regime preferences as expressed on the streets and at the ballot box. In particular, I exploit data from the survey’s two pre-revolutionary waves to investigate how state dependency conditions the effect of socioeconomic position on the probability of democratic protest participation during the Orange Revolution. Then I investigate whether the same factors that affect attitudes toward democracy and the probability of pro-democracy protest also translated into changing party support following the Orange Revolution. The ULMS allows us to observe who already supported the Party of Regions on the eve of the Orange Revolution and who came to support the party only later. As mentioned already, growing support for the Party of Regions after 2005 marked the return of regime incumbents and a retreat from democratization. The key empirical challenge in testing this book’s theory is the potential for endogeneity between labor market position and political preferences. In addition to the causal pathway from employment to regime preferences, the relationship between these two variables may be reciprocal, with political preferences also (or even exclusively) influencing the choice of career. It is impossible therefore to construct a strong causal test of the argument using cross-sectional data, where these variables are measured contemporaneously. The chief advantage of the panel design is that it explicitly incorporates change over time through repeated measures of the same variables (Miller 1999). Specifically, I provide a strong test of the independent causal effect of state dependence on regime preferences through an investigation of new labor market entrants. The difference-in-differences design that I employ exploits the temporal dimension of the panel data to better identify causal relationships.10 The principal advantage of this design for causal inference is that it allows me to directly address threats from selection bias and reverse causality. It also eliminates potential confounding by individual characteristics that affect political attitudes and behavior but do not vary over time. These include virtually all of the control variables one might include in a crosssectional regression predicting regime preferences, as well as other unmeasured traits like ability, work ethic, and communist-era experiences that could affect regime preferences and socioeconomic position along with selection into state employment. In short, the difference-in-differences design solves 10. The design I employ is also similar to a “cross-lagged” approach (Finkel 1995).

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 165

the problem of omitted variable bias from unobserved but time-invariant sources—a central concern in cross-sectional research. Thus, focusing on new entrants to the labor market provides a particularly clean causal test of the state dependency argument. Given Ukraine’s low rate of intersectoral mobility, it also makes substantive sense.11 Since state employment tends to be highly stable, an individual’s initial choice of employment is an especially consequential career decision. Because the survey represents the population aged 15 and older, the number of initially non-working students in the sample (N = 558 in 2003 and N = 416 in 2004) is sufficiently large for such an analysis.12

Measurement of the Independent Variables The first key independent variable captures direct dependence on the state for economic opportunities and life chances. Because the ULMS was designed by economists to track changes in the Ukrainian labor force during the country’s market transition, it is particularly extensive on variables related to individual labor market participation. In contrast to many other sources of public opinion data, for example, enterprise type is carefully measured. Each wave of the ULMS includes the same question on the ownership structure/type of enterprise or organization in which the respondent was employed. Using this question, I create an indicator for public employment that includes state enterprises (about 29% of total employment), budgetary/public organizations (about 18%),13 local municipal enterprises (2%), and state farms (6%). Public-sector employment thus continued to comprise slightly more than half of total employment in Ukraine around the time of the Orange Revolution, making it a very significant contributor to the life chances of Ukrainian citizens. My second independent variable captures the distinction between highly educated white-collar and professional strata versus less educated routine and manual laborers. My primary interest lies in the political orientations of the former group, which I refer to as the middle class. Each survey included key questions on education and occupation, the variables that I have used 11. More than half of the ULMS sample either worked or had worked in the same sector for their entire career as of 2003: 48% in the state sector and 10% in the private sector. Among those who were still employed in 2007, the corresponding figures are 39% and 14%, respectively. 12. Though note that a small sample makes it harder to find significant effects. 13. These include various levels of the state administration as well as organizations such as universities, schools, hospitals, and clinics that are funded from the state budget.

166

chapter 6

throughout to define the middle class. Though I also code income per household member in constant local currency (2003 Ukrainian hryvnias), I do not use income as my primary measure of class for both the practical and theoretical reasons laid out in chapter 2. Again, income is subject to widespread underreporting. Gorodnichenko & Peter show that Ukrainian state employees hide a substantial share of their income, which the authors attribute to bribes. Focusing on income alone would thus distort the results. I instead use the measure of current income to control for the formal public/private-sector wage gap, though Gorodnichenko & Peter (2007) find that the combined total of reported and unreported compensation is similar across the two sectors. Among those who are employed, 19 percent are middle class according to my definition. Just under two-thirds (63%) of this middle-class group is employed by the state. Moreover, of those who joined the middle class between 2003 and 2007, 58 percent did so through the state sector. The share that is middle class is somewhat higher in the center, south, and west (21%) and somewhat lower in the east and Crimea (16%). State-sector employment, however, plays a consistently important role for the middle class across all of Ukraine’s regions, and it is strongest in the central region that includes Kyiv. Finally, it is important to note that Ukraine’s middle class, thus defined, is nowhere near the middle in terms of income. Approximately 75%–80% of both the state and private-sector middle classes have incomes above the median.14 Redistributive theories would thus expect these groups to be less supportive of democracy than working-class Ukrainians with below-median income.

Measurement of the Dependent Variables To capture regime preferences, I employ both attitudinal and behavioral measures. In each wave, respondents were asked about their attitude toward various political systems and party support. The 2007 wave of the survey additionally included a battery of items about the Orange Revolution. These items allow me to directly assess the relationship between abstract regime preferences and more concrete voting and protest behavior, without assuming that one is a reliable proxy for the other. For the attitudinal measure of regime preferences, I use an item that asks: “What kind of political system, in your opinion, is most suitable for Ukraine?” 14. See online appendix figure D.1.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 167

Respondents were offered four options: (1) The Soviet system that was in our county until perestroika; (2) The Soviet system, but in a different, more democratic form; (3) The political system that exists today; (4) Western-type democracy.15 A strength of this question is that it measures regime preferences in terms of a continuum that would be familiar to respondents, given the context (Gibson, Duch & Tedin 1992; Whitefield & Evans 2001). Because the communist countries used the rhetoric of democracy, respondents’ understanding of the term may differ from its conventional usage in political science. It is thus helpful that the question distinguishes between Soviet and Western conceptions of democracy. At the same time, we might worry that the question does not clearly define what is meant by the term “democracy” or what would constitute a “more democratic form” of the Soviet system. Ideally, the question would be followed by additional items probing respondents’ views on specific aspects of democracy (majority rule, minority rights, individual liberties, etc.) as in chapter 3. The fact that this chapter’s findings are consistent with those of chapter 3 at least partially allays these concerns. In addition, the response option, “the political system that exists today” raises certain questions. During the period in which these surveys were conducted, Ukraine’s political system was in flux. While interpretations may have varied, no more than 10 percent of respondents express a preference for the current system in any of the survey’s three waves. This helps to clarify that state employees are not merely more supportive of the regime that hired them. Further analyses show that respondents’ views on regime type go together with their views on related issues in sensible ways and vary meaningfully across panel waves, even after accounting for measurement error.16 In all subsequent analyses, I transform the political regime variable to vary over the interval zero to one, with higher scores representing more democratic preferences.17 Next, as a behavioral measure of regime preferences, I code respondents’ participation in the Orange Revolution. The 2007 survey asked respondents 15. A question about respondents’ preferred economic system followed. Responses on the two questions are closely related (ρ = .704). Given the order in which they appeared on the questionnaire, if spillover effects exist, they would run from the item on politics to the item on the economy. 16. See online appendix D.2. 17. See, e.g., Abelson & Tukey (1970) on the use of linear techniques with ordinal data. I also demonstrate robustness using multinomial logistic regression for a categorial dependent variable.

168

chapter 6

both whether they personally participated in any protests and which side they supported. Together these questions allow me to identify three groups: (1) those who participated on the Orange side, the side calling for democratic reform, (2) those who participated on the Blue side, the side backing the incumbent regime and its chosen successor, and (3) those who did not participate. I focus on the first two groups—those who expressed a regime preference through costly action, recognizing that some who share that preference chose not to participate. Using both attitudinal and behavioral measures allows that some nonparticipants may prefer to live in a more democratic society. And some people may just like protesting. Joining a protest, even on the side calling for democratic reform, need not proxy for democracy support. Rather, the combination of attitudinal and behavior measures sheds light on how public employment shapes democracy’s base of support and participation in the civic revolutions that often drive democratic openings. My second behavioral measure of regime preferences focuses on change in Ukrainians’ voting behavior. Using the question regarding vote intention included in each survey wave, I code an indicator for respondents who switched to support the Party of Regions between 2004 and 2007. During that time, millions of Ukrainians withdrew their support from the Orange coalition. As shown next, these coalitions were strongly associated with regime preferences. What is more, in practical terms, this shift in party support was important as it effectively ratified Ukraine’s democratic retrenchment. va r i at ion in t he de p e n de n t va r i a bl e In this section, I use inter-item correlations to show that respondents’ views on regime type go together with their views on other, related issues in sensible ways. This approach follows Converse (1964). I then look at variation in respondents’ regime preferences over time. If regime preferences are politically salient and well-measured, we would expect a respondent’s attitudes toward regime type in 2003 and 2004, before the Orange Revolution, to be good predictors of the side a respondent took on those events, as reported in 2007. In table 6.1, I regress the side a respondent supported during the Orange Revolution on lagged regime attitudes, using multinomial logistic regression.18 Respondents were presented with the following response options: 18. The question on which side a respondent supported during the Orange Revolution asked: “What was your personal attitude during the political events that surrounded the election process?”

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 169 table 6.1. Multinomial Logistic Regression Results Showing the Relationship Between Regime Preferences and Side Supported During the Orange Revolution (1) Orange Side vs. Blue and White Side

(2) Had a Different Third View vs. Blue and White Side

(3) Had No Particular Opinion vs. Blue and White Side

(4) Had a Different Third View vs. Orange Side

(5) Had No Particular Opinion vs. Orange Side

Regime preferences (wave 2)

2.125∗∗∗ (6.63)

1.309 (1.22)

1.255 (1.16)

0.616∗ (−2.22)

0.590∗∗ (−2.74)

Regime preferences (wave 1)

1.783∗∗∗ (5.29)

1.295 (2.22)

1.328 (1.52)

0.726 (−1.53)

0.744 (−1.60)

Note: The table reports risk ratios from a multinomial logit model predicting the side a respondent supported in the Orange Revolution with lagged regime preferences. Coefficients represent the increase in the odds of being in the first category over the second that is associated with a one-unit increase in regime preferences. The associated t statistics are in parentheses. N = 3, 291. ∗ p < 0.05, ∗∗ p < 0.01, ∗∗∗ p < 0.001.

“I more or less agreed with the ‘orange’ side,” “I more or less agreed with the ‘blue and white’ side,” “I had a different, third view,” and “I had no particular opinion.” Table 6.1 shows the strong association between respondents’ answers to this question and the regime preference item in terms of relative-risk ratios for several indicative outcomes. The results in table 6.1 confirm that the ULMS’s measure of regime preferences elicits responses that are well correlated with other consequential political attitudes.19 In particular, the bivariate relationships in table 6.1 demonstrate that respondents’ regime preferences in 2003 and 2004 are good predictors of the side a respondent later took in the Orange Revolution. Moving from the bottom to the top of the regime preferences scale is associated with a 29-percentage-point increase in the probability of siding with the Orange coalition and a 28-percentage-point decrease in the probability of siding with incumbents. In other words, the closer one was on the regime preference scale to preferring the Soviet system, the more likely one was during the Orange Revolution to support Yanukovych and the Party of Regions. For a 19. Strong inter-item association is taken as evidence that these attitudes form part of a broad conceptual framework or ideological orientation, while lack of coherence among attitudes (low inter-item correlations) may imply that respondents lack meaningful preferences on the regime questions elicited by the survey.

170

chapter 6

table 6.2. Aggregate Regime Preferences, 2003–2007 2003 (%)

2004 (%)

2007 (%)

Soviet system Soviet system but more democratic Current political system Western democracy

37.2 21.1 5.5 36.2

34.8 22.0 8.9 34.3

29.2 26.6 10.5 33.8

Mean regime preference score [0,1]

0.469 [.457, .480]

0.476 [.463, .488]

0.496 [.484, .509]

Note: The table reports the survey-weighted percent of respondents in each category for each of the three survey waves. Columns may not sum to 1.0 due to rounding. The bottom line gives the survey-weighted mean numeric regime preference score (recoded to vary 0–1), and its 95% confidence interval.

question involving abstract political concepts, the opinions elicited appear well anchored to political realities. Next, I look at how Ukrainians’ regime preferences changed over the period 2003–2007. I investigate two types of systematic variation, showing that most of the variation in regime preferences occurred at the individual rather than aggregate level. This suggests that events played a more circumscribed role in shaping Ukrainians’ preferences over democracy than sometimes thought. In particular, scholars have speculated about two aggregate shifts in Ukrainian public opinion: first, increased pro-democratic sentiment around the Orange Revolution, then subsequent decline in support for democracy after it, as disillusionment mounted with the Orange coalition’s performance in office. In fact, aggregate regime preferences changed only modestly over these years, according to the ULMS. As we can see from table 6.2, support for western-style democracy was roughly stable. Table D.2 of the online appendix reports additional descriptive statistics. Ukrainians became somewhat less likely to favor the Soviet political system that existed in their country until perestroika and somewhat more likely to favor either the current political system or the Soviet system in a new, more democratic form. Though this small and statistically significant shift toward more democratic attitudes is evident when we compare respondents’ mean score on the political regime question in 2003 and 2007 (0.469 vs. 0.496, p < 0.05), the substantive size of this difference is quite small.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 171

Looking at variation in individuals’ regime preferences over time, however, it is clear that aggregate opinion stability belies greater micro-level fluctuations.20 Exploiting the ULMS’s panel design, I next consider how large real individual opinion change is relative to random response instability caused by measurement error.21 The results suggest that after accounting for measurement error, respondents’ true regime preferences are fairly stable, though not immovable, across waves.22 I estimate a stability coefficient of .911 (σ = .036) between waves 1 and 2. The stability coefficient between waves 2 and 3 is .778 (σ = .028), which implies yearly stabilities of .920 (.7781/3 ). For comparison, these findings imply that Ukrainians’ regime preferences over this period were significantly less stable than partisanship in established democracies (Green, Palmquist & Schickler 2002), though more stable than Americans’ opinions on many issues of domestic and foreign policy (Achen 1975; Feldman 1989). Ukrainians’ true attitudes toward democracy fluctuated about as much as did East Germans’ interest in politics around the collapse of communism (Prior 2010). Together these results suggest that Ukrainians’ expressed regime preferences represent more than doorstep opinions offered to satisfy inquisitive interviewers—if they were this, we would expect much greater volatility (Converse 1964; Zaller 1992). Subsequent empirical analyses show how changes in socioeconomic status and state dependency help to explain this variation in Ukrainians’ democracy support.23 20. A naive approach to gauging these changes is to calculate the correlation between respondents’ answers for each pair of waves. The share of panel respondents who gave the same answer on the 4-category regime preferences scale in waves 1 and 2 of the survey is 58%, despite the fact that those interviews were conducted a year apart. The fact that responses fluctuated similarly among respondents with more education (50%) and white-collar occupations (55%) suggests that the reliability of these questions has little to do with respondents’ political sophistication. The share who gave the same response between waves 2 and 3—a longer period of 3 years spanning the Orange Revolution—falls predictably to 49%. Similarly, the share giving the same response in the first (2003) and third (2007) waves is 48%. While at first glance these patterns may appear to imply that many Ukrainians reassessed their views of democracy around the Orange Revolution, a more plausible explanation is that this volatility is due largely to measurement error. 21. For a detailed description of the measurement model used in this analysis see online appendix section D.2. 22. The full results are reported in online appendix table D.1. 23. A standard econometric result holds that measurement error in the dependent variable does not bias coefficient estimates of the explanatory variables (Wooldridge 2003, 303).

172

chapter 6

Empirical Results The remaining empirical sections of the chapter proceed as follows. I next detail the main results from an investigation of new labor market entrants. This analysis takes advantage of the explicit time-ordered dimension of the panel data to provide clear causal evidence for the state dependency argument. Two subsequent sections examine regime preferences as revealed by protest participation and vote choice.

Changes in State Dependency and Changes in Regime Preferences A causal relationship between state dependency and support for democracy implies not only that these variables be correlated cross-sectionally, but also that changes in one be related to changes in the other over time.24 While regime preferences are fairly stable over the course of an individual’s adult life, the analysis in this section uses a difference-in-differences design to show that they are shaped crucially as individuals enter the labor market. Focusing on respondents who were students during the survey’s first waves but joined the workforce in subsequent years provides a very clean causal test of the argument.25 If state employees have more to lose from democratization, we would expect diffidence about democratization to rise among those who enter the labor force in public-sector occupations. This pattern should also be evident among the middle class: among those who secure a middle-class occupation through the state, I expect support for democracy to fall, while among those More specifically, estimates from linear regressions will be unbiased and consistent so long as measurement error is assumed to be uncorrelated with each of the explanatory variables. Evidence substantiating this assumption is cited in footnote 20 above. However, measurement error in the dependent variable does result in larger error variance—an intuitive result since the variable measured with error contains less information. The upshot is that because regime preferences are measured with error, subsequent tests lean against finding significant effects. 24. Section D.3 of the online appendix presents results from a pooled cross-sectional analysis, which captures the long-run relationships between class, sector, and democracy support in Ukraine among the full sample of respondents. The results conform to the pattern identified cross-nationally in chapter 3, which uses a similar empirical setup but different data. 25. The cross-sectional results in table D.3 suggest that the pattern of lower democracy support among state workers holds not only for new labor market entrants but for the sample as a whole.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 173

with otherwise similar characteristics, I expect no such change. An advantage of narrowing the analysis to the middle class is that we hold constant level of education and the type of occupation attained as individuals enter the labor force. In sum, two pieces of evidence would constitute support for the hypothesis. First, I expect joining the state sector to weaken support for democracy. Second, I expect that, once employed, public sector professionals will be less supportive of democracy than private-sector employees. Further, it is essential to show that these differences cannot be explained by individuals’ preexisting political preferences or characteristics. The analysis that follows directly tackles the possibility that individuals sort themselves into different types of employment on the basis of their existing attitudes or attributes correlated with regime preferences. For example, we might be concerned that democrats are less likely than nondemocrats to seek public employment under a nondemocratic state. Or we might worry that risk-averse individuals are both more likely to choose state careers and less likely to favor democracy. If selection drives the association between political preferences and the choice of state employment, then we would expect individual i’s political preferences at time t − 1 to be associated with his/her choice of employment at time t. By examining lagged values of the dependent variable, and showing that they are uncorrelated with the treatment, we can rule out selection and reverse causation. The key identifying assumption of the difference-in-differences design is common trends (Angrist & Pischke 2009). Substantively, we assume that the trend in regime preferences would be the same among all new labor market entrants had their employment choices, and thus their degree of state dependency, not diverged. The common trends (or “parallel slopes”) assumption cannot be tested directly, but three-wave panel data do allow us to examine regime preferences over multiple pretreatment waves for apparent violations. Together the evidence of pretreatment equivalence and no trending in the dependent variable shown below supports the identifying assumptions of the analysis.26 Figure 6.1 presents the findings, which are based on following the same individuals over time and showing how the preferences of those same individuals change. The top panel displays the results for all new labor market 26. As an additional robustness check, I also examined balance on a number of pretreatment covariates. I find no significant pretreatment differences in gender, age, education, income, household size, or views on economic liberalism.

174

chapter 6

All new labor market entrants 1.0

Mean difference in democracy support (State − Private)

Democracy support

0.4 Private sector in wave 3

0.8

0.6

State sector in wave 3

0.2

Pre-employment Postemployment

0.0 −0.2 −0.4

0.4 Wave 1 (2003)

Wave 2 (2004)

Wave 1 (2003)

Wave 3 (2007)

Wave 2 (2004)

Wave 3 (2007)

(a) Middle class 1.0 0.4

Mean difference in democracy support (State − Private)

Democracy support

Private sector in wave 3

0.8 State sector in wave 3 0.6

Pre-employment Postemployment

0.2 0.0 −0.2 −0.4

0.4 Wave 1 (2003)

Wave 2 (2004)

Wave 3 (2007)

Wave 1 (2003)

Wave 2 (2004)

Wave 3 (2007)

(b) figure 6.1. Sectoral effects on political preferences. The dark vertical bars are 90% confidence intervals. The light vertical bars indicate 95% confidence intervals. Treatment occurs between waves 2 and 3. Data source: ULMS.

entrants. The results for the middle-class group with higher education that gained white-collar or professional employment over this period are in the bottom panel. I first focus on the group that entered the labor market between the second and third survey waves. The advantage of focusing on this group is that we can plot lagged values of the dependent variable for two pretreatment

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 175

All new labor market entrants 1.0

Mean difference in democracy support (State − Private)

Democracy support

0.4 Private sector in wave 2

0.8

0.6 State sector in wave 2

0.2

Pre-employment Postemployment

0.0 −0.2 −0.4

0.4 Wave 1 (2003)

Wave 2 (2004)

Wave 1 (2003)

Wave 2 (2004)

figure 6.2. Robustness test of sectoral effects on political preferences. The dark (light) vertical bars are 90% (95%) confidence intervals. Treatment occurs between waves 1 and 2. Data source: ULMS.

waves. We can then examine these values for evidence of trending, which would suggest a violation of the parallel slopes assumption. As a robustness check, I also repeat the analysis for the cohort entering the labor market between the first and second survey waves. The findings in figure 6.2 are identical. Turning to the results in figure 6.1, the open circles (dots) represent the mean regime preferences of those who were employed in the private (public) sector by wave 3. Treatment occurs between waves 2 and 3. Accordingly, the leftmost points in each figure show the pretreatment estimates of democracy support from wave 1, the middle set of points shows the pretreatment estimates from wave 2, and the rightmost points plot the posttreatment estimates from wave 3. The first thing to note is that the pretreatment regime preferences of the two groups are indistinguishable and there is no evidence of trending. This supports interpreting the diverging lines between waves 2 and 3 as indicating that labor market choices have a causal effect on regime preferences. Among those who joined the state sector, mean support for democracy fell from .761 to .574 (p = 0.006), while among those who joined the private sector it changed very little from .775 to .733 (p = 0.400). Further disaggregating state-sector employment, the ULMS data suggest that the vast majority of new public-sector workers joined either state enterprises (46.1%) or budget-sector organizations (37.1%). Both groups exhibit the same pattern of declining democracy support, suggesting that fears of privatization in the

176

chapter 6

state enterprise sector are not solely responsible for these results. Meanwhile, among those who secured a middle-class occupation through the state, support for democracy decreased from .905 to .639 (p = 0.007), while among those who secured a middle-class occupation in the private sector, support for democracy declined insignificantly (from .902 to .822, p = 0.416). In terms of timing, most (62.9%) of those who entered the state sector between waves 2 and 3 of the survey did so in 2006 and 2007 after the Orange coalition had already begun to unravel. Another 9.0% joined while Yanukovych was still prime minister in late 2004. While the pretreatment regime preferences of those hired in 2005 are indistinguishable from those who were hired later, the negative effect on democratic attitudes of gaining economic opportunities through the public sector is stronger for those who were hired as Yanukovych and the POR were again ascendant. It is important to note that these results are not driven by a growing preference for the current political system among new state employees (8.3% before vs. 17.7% after). Rather, the share of new state employees who preferred western-style democracy fell sharply (from 65.0% to 35.5%), while the share who preferred some version of the old Soviet political system rose (from 26.7% to 45.2%). The full matrix of transition probabilities is given in online appendix table D.4. The right panel of each figure presents the difference in means between the two groups—both before entering the labor market (waves 1 and 2) and after (wave 3). In both cases, the political preferences of the two groups are initially the same: the difference-in-means estimates sit squarely on the horizontal line at zero. This implies that existing political preferences cannot explain subsequent labor market choices. It also suggests that any pretreatment differences in the preferences or characteristics of those entering public and private employment are orthogonal to democracy support.27 Meanwhile, the post-employment difference-in-means estimates are −.201 (p = 0.001) for all new labor market entrants and −.183 (p = 0.111) for the middle class. Though the former is more precisely estimated than the latter, both estimates are sizable and negative. 27. In other words, figure 6.1 shows that neither risk aversion nor any other characteristic that might differ between the two groups is correlated with the preference for democracy prior to state employment. In econometric terms, the expression of risk aversion could be seen as a mechanism through which state employment affects political preferences, but should not (given pretreatment equivalence of economic and political preferences) be seen as a confounder.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 177

Again, this approach holds constant many individual characteristics that might confound state employment in other types of analysis. Gender and age, for example, are often key drivers of political preferences in the post-Soviet context and are empirically correlated with public-sector employment.28 To properly identify the effect of state employment in a cross-sectional analysis, it is necessary to control for these variables. In a within-subjects design, by contrast, gender is a constant and age changes uniformly for all. As such, neither can explain the results. Similarly, a variety of difficult-to-observe characteristics like individual aptitude, work ethic, or risk aversion might affect the likelihood of choosing a state versus private sector career. By following the same individuals over time, however, these unobserved characteristics do not pose a threat to inference as long as they remain constant across waves.29 Only omitted factors that changed between 2004 and 2007, affecting those who chose state employment but not those who chose private employment, could account for the results in figure 6.1. The fact that the exact same pattern holds in figure 6.2 for the cohort entering the labor market between 2003 and 2004—a much shorter period—adds to our confidence in these results.30 While the approach used in figures 6.1 and 6.2 eliminates most threats to inference, we might still be concerned about two other alternatives. First, if state employees feel pressure to misrepresent their true attitudes, then biased survey responses may cause observed democracy support to diverge even as latent regime preferences remain unchanged. This alternative seems unlikely given that the share citing a preference for the current system increased by only about 10 percentage points and never exceeded about 17 percent. If state workers feared retaliation, we would expect stronger support for the current system.31

28. The average age of state employees in wave 3 was 43.4 versus 39.0 for non-state employees, while the share of women in the public sector was 59.0% versus 45.5% in the private sector. 29. It is worth underscoring that the composition of the sample is exactly the same in each time period in figure 6.1. As such, these changes in democracy support cannot result from changes in the aggregate composition of the state or private sector. 30. This test also addresses the concern that some time-varying factor associated with the Orange Revolution could confound the results in figure 6.1. 31. Two items provide further evidence against concerns about preference falsification. At the end of the interview following the battery of political questions, respondents were asked “How comfortable did you feel talking about these political events and your political activities

178

chapter 6

Second, having eliminated threats to inference from reverse causation and time-invariant characteristics, any remaining threats to inference must emanate from time-varying confounders. Thus, only some intervening change between panel waves, which affects individuals’ political attitudes and correlates with their choice to pursue state employment, could confound the observed relationship. While it is difficult to think of such a factor, one possibility is the change in income associated with taking a new job. In order to rule out the possibility that posttreatment regime preferences are due to differences in relative position within the income distribution, rather than state dependency per se, I compare posttreatment earnings. I find, however, that the difference in earnings between the two groups is insignificant and, thus, cannot explain the results (2,507 vs. 2,675 Ukrainian hryvnias, a difference of about $34 USD; two-sample t-test p = 0.455). Given their similar place in the income distribution, it also seems unlikely that either expectations about a rise in inequality following democratization or redistribution that would lower it explain why state- and private-sector attitudes toward democracy diverge.32 To summarize, then, these findings robustly demonstrate that state dependency weakens support for democracy. They also provide strong evidence against alternative arguments. Figure 6.1 shows that changes in labor market position have a causal effect on regime preferences, and not the other way around. Prior to entering the workforce, future state- and private-sector workers shared virtually identical political preferences. Proponents of westernstyle democracy were no more likely to choose careers in the private sector. Similarly, those who preferred a Soviet-style political system were no more likely to take up positions in the public sector. Yet, once career paths were chosen, the political preferences of these two groups diverged. What is more, this

in this interview?” The responses of new state employees were indistinguishable from others. Indeed, the average answer among both groups was just above the midpoint of a five-point scale, in which higher responses represent greater comfort talking about politics (3.28 vs. 3.48; two-sample t-test p = 0.292). We can also examine interviewers’ assessments of respondents’ sincerity and openness, rather than rely on respondents’ self-reports. On this measure, too, there is no evidence of biased responses by new state employees. Interviewers rated the two groups as equally candid (2.08 vs. 2.08 on a three-point scale, p = 0.931). 32. In other words, an alternative mechanism—whereby people are expressing their feelings about inequality and differences among them stem from their differing locations within the income distribution—is not supported by the evidence.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 179

divergence occurred quickly—a pattern that is inconsistent with long-term socialization and more in keeping with changed material incentives.33 This pattern held among students completing higher education and entering white-collar occupations. Those who secured middle-class jobs through the state became significantly less favorable toward democracy, implying that education alone does not dictate more democratic values. Further, for the subset of new labor market entrants who came from working-class backgrounds,34 securing a middle-class job in the state sector had an even more pronounced effect on political attitudes.35 Among those for whom state employment provided a pathway into the middle class, support for democracy declined sharply (−.256, p = 0.039), suggesting that there is no necessary connection between intergenerational upward mobility and greater demand for democracy. Finally, it is worth underscoring that these tests focused on the cohort many consider Ukraine’s most democratic. All else equal, young people plausibly have better labor market alternatives than older workers, whose more limited prospects outside the state sector are likely compounded by additional years of state employment. Yet even among the survey’s youngest respondents, reliance on the state for economic opportunities and life chances weakened support for democracy. Additionally, we might like to know whether a mid-career move into (or out of) the public sector has a similar (opposite) effect. While the research 33. One further concern might be that the dependent variable conflates joining the West with support for political democracy. If joining the West and closer trade ties with the European Union are beneficial for the private sector, this could explain why the private sector appears to want democracy more than the state sector. Although this might explain the slight increase in democracy support among the private sector in the top half of figure 6.1, it neither explains the steep decline in democracy support among the state sector nor the absence of a positive trend in democracy support among the private-sector middle class in the bottom panel of the figure. In fact, the perceptions of private-sector actors regarding the benefits of closer ties with the West vary by region. In the East, for example, many in the private sector did not expect to benefit from joining the EU. But even if we focus exclusively on Western Ukraine, the results in figure 6.1 are unchanged. 34. I code individuals as having working-class backgrounds if neither parent held a middleclass job, using questions on parental occupation. 35. Note that the results in the bottom half of figure 6.1 are for all those completing higher education and entering white-collar occupations. They do not distinguish between individuals who come from middle-class backgrounds (i.e., at least one parent held a middle-class job) and those who do not.

180

chapter 6

design in this section is robust for the cohort entering the workforce between 2004 and 2007, it is unclear whether these results can be extrapolated to later career decisions. Unfortunately, there are limitations to what the ULMS data can tell us about the effect of moving into or out of the public sector after the initial choice of employment. Whereas it is possible to measure regime preferences for new labor market entrants uncontaminated by prior employment, we cannot do so for those who have already entered the workforce. Supplementary analysis on this question is presented in section D.5 of the online appendix. The bottom line is that any short-run effect on regime preferences of switching from the state to the private sector (and vice versa) is too weak to detect in the present sample, given that the time period is limited and employment is highly stable. At the same time, however, pooled cross-sectional results do suggest that, in the long-run, state employment is associated with weaker support for democracy.36 Why? I have argued that state dependency increases the opportunity costs of supporting democratization. Change in the political control of the state threatens the status, benefits, and rents that public employment provides. Such worries will be especially pronounced where corruption and informality are widespread in the public sector and where reforms to state and markets are closely intertwined. Figure 6.3 shows that a public-sector job is also associated with less support for free markets among new labor market entrants in wave 3 who were initially more similar to their private-sector counterparts. Although there is no evidence that support for a market economy was already declining among future state workers, the pretreatment trend is less clear than in the preceding figures so we must be cautious about a causal interpretation. Still, these findings lend some support to the notion that economic and political preferences are intertwined and may be simultaneously affected by state dependency. Even new public employees—whose alternative employment prospects are probably better than others who have been in the public sector longer—are less likely to want to compete in a free market, where their jobs might be jeopardized or their rents threatened. In sum, these findings suggest that reticence about the transition to a market economy may contribute to limiting public-sector support for democracy. The fact that public-sector workers perceive few alternative sources of economic opportunity further underscores why they might reject regime 36. See table D.3 in the online appendix.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 181

All new labor market entrants

0.8

Private sector in wave 3 0.6

State sector in wave 3

0.4 Wave 1 (2003)

Wave 2 (2004)

Wave 3 (2007)

Mean difference in preference for markets (State − Private)

Preference for free markets

1.0 0.4 0.2

Pre-employment Postemployment

0.0 −0.2 −0.4

Wave 1 Wave 2 Wave 3 (2003) (2004) (2007)

figure 6.3. The preference for a market economy. The dark vertical bars are 90% confidence intervals. The light vertical bars indicate 95% confidence intervals. Treatment occurs between waves 2 and 3. Data source: ULMS.

change, which could endanger their interests. Asked in 2007 to evaluate their chances of finding equivalent alternative employment, those who had recently joined the state sector rated their exit options as worse than did those who were otherwise similar but recently employed in the private sector (52.38 vs. 60.76 on a 100-point scale from “no chance of finding an equally good job” to “absolute certainty,” p = 0.039).37 These differences are also significant when comparing state employees to private-sector employees in the full sample (43.37 vs. 54.66, p < 0.001) and the state- and private-sector middle classes (45.24 vs. 55.28, p < 0.001).38 Lending further support to the argument, poor exit options are, in turn, associated with greater distrust of democratization. This evidence suggests that the political implications of rising economic opportunity are likely to hinge on whether those opportunities are monopolized by the state.

Democratic Breakthrough and the Orange Coalition While the analysis thus far has focused on regime preferences in the abstract, I now turn to two behavioral measures with concrete bearing on regime 37. Unfortunately, this variable was not measured in prior waves. 38. Public-sector employees in state enterprises and budgetary organizations alike believe their prospects of finding alternative employment are worse than either the self-employed or those working in newly established private enterprises. See online appendix figure D.2.

182

chapter 6

outcomes. My analysis divides participants in marches, demonstrations, and other mass activities and events of the Orange Revolution into two groups. The first group comprises participants who supported the Orange side. The second group includes those who took part in rallies for the incumbent regime. Organized by the campaign of the pro-incumbent candidate, Viktor Yanukovych, these “Blue and White” demonstrations were smaller than those on the Orange side, but also drew mass participation. The analysis thus emphasizes the contrast between revolutionaries, who protested under the banner of democracy and against electoral fraud, and counterrevolutionaries, who demonstrated in defense of the incumbent regime. Using panel survey data for this analysis has several inherent advantages. First, a major threat to inference in retrospective surveys of protest participation is that explanatory variables are measured after the fact, potentially introducing posttreatment bias. With panel data, it is possible to eliminate this threat by using lagged values of these variables from prior survey waves. Here I use pretreatment values of the key explanatory variables measured just months before the Orange Revolution, during the summer of 2004. Second, I am able to exploit data from the panel’s first two waves to construct a measure of upward mobility into the middle class on the eve of the Orange Revolution. This measure is consistent with the concept of middle class that I have used throughout, and captures movement into the strata of highly educated whitecollar workers and professionals from the strata of less educated routine and manual laborers. I code as upward mobility into the middle class the change from a blue-collar (routine or manual) job to a white-collar/professional position.39 Because this measure is based on observations of an individual’s occupational status at two points in time, it is less likely to be endogenous to political preferences than typical self-evaluated measures of mobility. In the analysis that follows, I relate this measure to protest propensity, showing how state dependency moderates its effect. Specifically, I estimate the following regression equation using OLS for ease of interpretation. 39. One clear limitation of this measure is that it captures only short-term changes occurring between the first two panel waves. If the true effect of upward mobility is positive over a longer term, the fact that this measure does not capture individuals who joined the middle class in prior years would attenuate its estimated effect, making it more difficult to find a significant result.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 183

yit = β1 mobilityi,t + β2 statei,t−1 + β3 (mobilityi,t × statei,t−1 ) + x i,t−1 γ + it

(6.1)

where yit is a measure of protest (alternately on the Orange or Blue side), mobilityi,t is an indicator for upward mobility defined as a change in class status, statei,t−1 is an indicator for state employment, and x i,t−1 is a vector of control variables. The controls include class, logged income, age, gender, language spoken at home, and region in keeping with past studies of Ukrainian protest behavior (Beissinger 2013; Onuch 2014).40 The coefficient on class represents the effect of being middle class (vs. working class) that is independent of a change in class status during the year leading up to the revolution. I control for income in order to ensure that the results are not confounded by a sectoral wage gap. Again, all of the explanatory variables in the main models are measured in the waves preceding the revolution to mitigate concerns about endogeneity. One additional worry is that participation in the Orange Revolution could be related to individual risk tolerance and that risk-averse individuals may be more likely to choose a public-sector career. To address this concern, I show that the results are robust to controlling for individual risk tolerance, as discussed further below. Appendix table a3 also shows that the results are substantively similar in a baseline specification with no controls and in logistic regressions. The main findings are displayed in figure 6.4. The left side of the plot shows OLS coefficients from a regression predicting participation on the Orange side, while the right side shows the results for participation on the Blue side. Looking at the top of the left plot, we see that the large positive and statistically significant coefficient on “joining the middle class” (β1 = 26.2, p = 0.000) is similar in magnitude to the negative coefficient on the interaction term (β3 = −21.9, p = 0.005). The main effect for state employment, meanwhile, is indistinguishable from zero (β2 = 0.6, p = 0.625). Substantively, then, an individual who had recently gained a middle-class job in the state sector was 21.9 percentage points less likely to be an Orange revolutionary than an individual who had recently gained a middle-class job in the private sector.41 This 40. The results for the Orange side are also robust to controlling for religion, though doing so for the Blue side introduces singularities in the logistic model. For comparison, I have thus excluded religion from the controls. 41. To be sure that this result was not due to some nonlinear age (cohort) effect, I reran the model specifying age as a factor in decades. The results were unchanged. I also compared the age distribution of the middle class (and the state- and private-sector middle classes separately)

184

chapter 6

Orange side

Blue and White side

Joined the middle class State State x joined the middle class Middle class Log of income Age Male Russian language Ukrainian-Russian language Western/eastern Ukraine Kyiv Constant −0.4

−0.2

0.0

0.2

0.4

−0.05

0.00

0.05

figure 6.4. Participation in the Orange Revolution. This figure reports coefficients from OLS models predicting participation in marches, demonstrations, or other mass activities/events during the Orange Revolution on the Orange side (left panel) and Blue and White side (right panel). The horizontal bars are 95% confidence intervals. The reference category for language is Ukrainian speaker. Data source: ULMS.

is a very meaningful effect, roughly similar to the magnitude of residing in Kyiv, where most of the protests took place. The findings in figure 6.4 further confirm that the political consequences of upward mobility into middle-class occupations are conditioned by the role of the state. These results are robust to the inclusion of a variety of controls, which perform as expected. As in previous studies, I find that being a Russian speaker decreased the probability of supporting the revolution, while living in both Kyiv and Western Ukraine increased it. In addition to standard demographic controls, I also investigated whether state employment could be picking up risk avoidance/acceptance. Again, omitting individual risk tolerance from the model would be a problem if risk-avoidant types are more likely to be employed in the public sector and less likely to participate in protest. The measure of risk tolerance, which runs from zero (completely unwilling to take risks) to ten (completely willing to take risks), is positive and highly statistically significant, but does not change the other results. These results are reported in column 3 of appendix table a3. In sum, even after controlling for these other factors, the evidence clearly indicates that upward social mobility on the eve of the revolution substantially increased an individual’s probability with that of the non-middle class to be sure that age (cohort) could not be responsible for observed attitudinal or behavioral differences. In this sample of the working-age population, these distributions are in fact very similar.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 185

of protesting for the Orange side—except among state employees. In other words, the positive effect of upward mobility on protest participation was mostly vitiated by state dependency. Turning next to the right side of figure 6.4, we see a very different pattern for participation in counterrevolutionary demonstrations. Here, upward mobility has a negative, though statistically insignificant, association with protest participation.42 In fact, only state employment and living in eastern Ukraine significantly predict protest in defense of the incumbent regime. This suggests that the regime was best able to mobilize supporters in its historical strongholds and among those whose economic opportunities and life chances depended most on the state. To summarize, using a behavioral measure with direct bearing on regime outcomes, this analysis reinforces the attitudinal evidence in preceding sections. In keeping with the argument in this book, the results show that the effect of upward mobility on Orange protest participation was conditioned by an individual’s relationship to the state. These findings imply that state dependency not only diminishes support for democracy in the abstract, it also affects the choice to take to the streets.

Democratic Retrenchment and the Growing Anti-Orange Coalition In this final empirical section, I examine changes in Ukrainians’ voting behavior. Specifically, I investigate whether state dependency, which I have already shown affects attitudes toward democracy, also translated into changing party support following the Orange Revolution. Within months of taking power, the Orange coalition was consumed by internal infighting and reforms had stalled. Governance challenges led to mounting popular disappointment as change gave way to familiar patterns of corruption and authoritarian rule (Diuk 2014). Against this backdrop, the Party of Regions’ comeback victory in the 2006 parliamentary election—the first held in the wake of the Orange Revolution—quickly became a symbol of Ukraine’s democratic backsliding. Party support provides a noisy but behavioral measure of revealed regime preferences. Moreover, the shifting landscape of party support had real consequences for Ukraine’s democratic institutions during this period. This section 42. The preferred model specification for predicting protest on the Blue side omits the interaction term between upward mobility and state employment. However, the results are very similar if the interaction is included.

186

chapter 6

table 6.3. 2004 Party Support Among Subsequent Defectors to the Party of Regions % Communist forces Victor Yushchenko bloc “Our Ukraine” Socialist Party of Ukraine Social Democratic Party of Ukraine (united) The Green Party of Ukraine Juliya Tymoshenko election bloc Nataliya Vitrenko bloc Peasant Party of ukraine Other Difficult to say/refused to answer

15.3 4.1 4.0 2.9 1.4 0.8 0.6 0.4 1.4 69.2

Note: The table shows party support in 2004, among those who did not support the Party of Regions before the Orange Revolution but subsequently defected to the party by 2007.

takes up the question of who switched to support the Party of Regions by 2007, contributing to Ukraine’s democratic retrenchment. In so doing, it adds to our understanding of the social forces that are most likely to turn against democratizing revolutions. Each survey wave included the standard vote choice question: “If the parliamentary election were held this coming Sunday, for which political party would you vote?” Whereas on the eve of the Orange Revolution in 2004, only about 16 percent of Ukrainians reported supporting the Party of Regions, by 2007 that figure had grown to 47 percent.43 I code as switchers to the Party of Regions those who either supported a different political party or found it difficult to say for which party they would vote in 2003 and 2004, but reported supporting the Party of Regions in 2007. Table 6.3 shows the breakdown of 43. These figures refer to the share of respondents who indicated the Party of Regions out of those who answered the vote choice question in the respective survey year. Nonresponse rates on the vote choice question (combining “difficult to say” and “refused to answer”) were 49% in 2004 (a year in which there was no parliamentary election) and 33% in 2007 (a year in which there was). Reported support for the Party of Regions in 2007 is about 10 percentage points higher than the actual share of votes received by the Party of Regions in early parliamentary elections held in September, while the ULMS was still in the field. This is likely due either to lower turnout among Party of Regions’ supporters or actual fluctuations in party support over the summer months when most of the ULMS interviews were conducted.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 187

parties from which these new Party of Regions voters defected. The first thing to note is that more than half did not support any party in 2004. The next largest share of new Party of Regions supporters came from the Communist Party of Ukraine (KPU), which along with the Social Democratic Party of Ukraine (SDPU(united)) and the Party of Regions had formed the core of the anti-Orange forces. However, the Party of Regions also benefitted from sizable defections from Viktor Yushchenko’s bloc “Our Ukraine” and the Socialists (SPU). Together with the Tymoshenko bloc, these parties were the core of the Orange coalition. I next estimate the probability of becoming a Party of Regions supporter by 2007, among those who had not previously voted for the party. While most analysis of this period has focused on the parties’ appeal to linguistic and regional identities (see, e.g., Kuzio 2011), the analysis in this section reveals that these were not the only significant factors. Shifts in party support also followed distinctive sectoral and class patterns. All independent variables in the model are again measured in 2004, mitigating concerns about endogeneity. Endogeneity would be a problem, for example, if becoming a Party of Regions supporter aided in obtaining state employment or a professional, white-collar job, or in starting one’s own business. Because predictors from the 2004 wave precede the switch, becoming a Party of Regions supporter cannot be their cause. The control variables include age, gender, logged income, religion, the language spoken in a respondent’s home, and indicators for each region. I present results using OLS to ease interpretation. However, the results are unchanged in logistic regressions, which are reported in appendix table a4. Figure 6.5 shows the resulting OLS coefficients. The point estimates represent the increase (decrease) in the probability of voting for the Party of Regions in 2007 that is associated with a one-unit increase in the independent variable. The bars are the associated 95 percent confidence intervals.44 The first row of the figure shows that public employees were significantly more likely than their private-sector counterparts to switch to the Party of Regions after the Orange Revolution, furthering Ukraine’s retreat from democracy. Meanwhile, the second row of the figure implies that the middle class was significantly less likely to defect to the Party of Regions than the working class. 44. I present the main results without controlling for past party support since party support is likely a consequence, in part, of the other covariates. However, the results are unchanged if 2004 party support is included in the model. See online appendix table D.9.

188

chapter 6

State Middle class Log of income Age Male Russian language Ukrainian-Russian language Central Ukraine Eastern Ukraine Western Ukraine Southern Ukraine Catholic Orthodox −0.3

−0.2

−0.1

0.0

0.1

0.2

0.3

figure 6.5. Regression results predicting switching to the party of regions. This figure reports coefficients from OLS models predicting voting for the Party of Regions in 2007. The sample is respondents who did not support the POR before the Orange Revolution. The horizontal bars are 95% confidence intervals. The reference category for language is Ukrainian speaker; for region, it is Crimea; for religion, it is not religious. Data source: ULMS.

The coefficients suggest that public-sector employment increased the probability of becoming a Party of Regions supporter by about 4.3 percentage points, while being middle class decreased it by about 4.5 percentage points. These are meaningful changes over a baseline probability of 21 percent. They also suggest that the effect of being middle class was almost entirely vitiated by state dependency. This is consistent with the preceding results. It is worth noting that class and sector remain significant correlates of voting behavior alongside a robust set of controls.45 The inclusion of logged income, though insignificant, confirms that differences between the state and private sector cannot be attributed to a sectoral wage gap. As expected, those who spoke Russian or a mixture of Russian and Ukrainian at home were more likely to switch allegiance to the Party of Regions. The same was true of 45. A baseline model with no controls is also reported in appendix table a4.

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 189

those living in the party’s traditional strongholds in Eastern Ukraine, while the probability of switching among those living in Central and Western Ukraine was very small. As expected, older Ukrainians were more likely to join the ranks of POR supporters. Neither gender nor religion were significantly related to changing party support. Although the primary basis of party support is clearly region—for example, Ukrainians living in the east were about two and a half times more likely to switch to the POR than Ukrainians living in the center—the effects of class and sector compare favorably with other identity-based determinates of party support. For example, speaking Russian (as compared with speaking Ukrainian) increased the probability of becoming a Party of Regions supporter by about 7.2 percentage points (versus 4.3 percentage points for state employment). In short, even after accounting for other key regional and linguistic determinants of Ukrainian voting behavior, both social class and sector of employment aid in explaining patterns of support for democratic retrenchment after the Orange Revolution.

Conclusion These findings again highlight how extensive state involvement in the economy and a large public sector undermine middle-class support for democratization. The results suggest that in the long run, growth of the middle class may lead to greater support for democracy, but that this process depends also on greater autonomy from the state. Aided by panel data, consisting of reinterviews with the same respondents over a period of significant democratic contestation, this chapter helps to clarify the causal status of relationships that are difficult to assess with cross-sectional data. Its findings sharpen our confidence in the independent causal effect of state dependence on democratic preferences. A particularly strong test of the argument focused on new labor market entrants. Using a difference-in-differences strategy, I showed that taking a job in the state sector caused support for democracy to decline among Ukrainians who entered the labor market between 2003 and 2007, a pivotal time in Ukrainian politics. This decline was especially steep for those who secured white-collar positions. It was steeper still for those who secured white-collar positions but whose parents were not professionals—that is,

190

chapter 6

for those who did not come from a middle-class background but gained a middle-class job in the public sector. Meanwhile the political preferences of those who entered the private sector remained unchanged. Because the first two waves of the survey provide pretreatment measures of political preferences for those who joined the workforce only later, we were able to rule out reverse causality due to political self-selection as well as selection due to time-invariant characteristics correlated with regime preferences. Indeed, the evidence showed clearly that those whose preferences diverged after entering the labor market shared virtually identical prior preferences over democratization. Subsequent tests using the full nationally representative sample focused on costly political behavior. The analysis showed that state dependency attenuated the positive effect of upward mobility on democratic protest participation during the Orange Revolution. It was also associated with mobilizing in defense of the incumbent regime. Later, when the revolution faltered, I found that defection to the Party of Regions was more likely among public employees whose jobs and rents were threatened by the prospect of democratic transition. Ukraine’s private-sector middle class was less likely than its state middle class to shift its votes to the Party of Regions and thereby contribute to democratic retrenchment. Together with the results for new labor market entrants, these findings imply that democratic coalitions will be weaker when direct state economic engagement concentrates economic opportunity and rents in the public sector. How does Ukraine compare to other cases in the post-Soviet region and beyond? While formally civil service protections exist in Ukraine, in practice such protections are weak. Public-sector hiring and promotion are frequently not merit-based. Opportunities to earn informal rents are widespread and officially sanctioned. All of these factors contribute to a high degree of dependence among public-sector professionals. At the same time, corruption and dependence on state contracts are also pervasive in Ukraine’s private sector. If employees in private but closely state-aligned enterprises hold similar preferences to state employees, then the true gap in demand for democracy between workers who are and are not state dependent is likely much larger. The sectoral estimates presented above should be considered conservative. In these respects, Ukraine is again not unique. Weak protections for civil servants, a public sector bureaucracy that is poorly insulated from

r e v o l u t i o n a n d d e m o c r at i c r e t r e n c h m e n t 191

political pressures, and official graft are virtually universal across the postSoviet region. They are also widespread in other autocracies. That said, the Ukrainian case may not be comparable with others in which the public sector has different institutional characteristics. In the presence of stronger civil service protections, where state-sector professionals are competitively recruited, or where there are fewer opportunities to earn informal rents on the basis of one’s official position, all else equal, the state middle class may be more politically independent. In other respects, it is worth noting again that Ukraine provides a tough test of the state dependency argument. Ukraine has struggled to finance its public sector amidst uneven economic performance over much of the postSoviet period. Hence, satisfaction with what autocracy has delivered over the long run as well as support for the current regime is lower in this setting. Ukraine also lacks the resource rents that underpin public spending in other post-Soviet states. The upshot is that although state jobs may be more secure than private-sector jobs, formal salaries and the prestige of public employment are lower in Ukraine. Still, state dependency maintains its consistent negative effect on democracy support, especially among those who joined the public sector at a time when salaries were rising and being paid on time. Ukraine is also a tough test of the argument insofar as other factors like regional and linguistic identities operate strongly on Ukrainians’ political preferences. Yet even in this setting, and alongside a sturdy set of controls for identity-based factors, state dependency still diminished the positive effect of upward mobility on democracy support, in keeping with this book’s argument. Concretely in the Ukrainian case, disaggregating the middle class helps us to better understand why Ukraine’s democratic coalition was not stronger or more successful at consolidating initial democratic gains. In the 2006 parliamentary election, the PORA party, newly created from the Orange Revolution’s most influential civil society organization, declared the entrepreneurial middle class its target constituency. The party failed to surpass the 3 percent threshold to gain parliamentary seats. Finally, the evidence in this chapter adds to our understanding of the political consequences of economic restructuring and privatization in countries that lack consolidated democracy. First, it implies that statist models of development that concentrate rents in the public sector attenuate the positive role that social mobility can play in broadening democratic coalitions. Second,

192

chapter 6

the findings suggest that privatization which reduces state involvement in the economy is likely to improve long-run prospects for democratization. Still, whether privatization policies have a positive impact on democracy’s chances will depend in part, the analysis suggests, on whether they impoverish or provide new opportunities for the state middle class. The next chapter investigates, through the empirical example of Kazakhstan, how regimes engineer the creation of an autocratic middle class alongside significant privatization. In it, I endeavor to show that an autocratic middle class is not just a legacy of the Soviet era, but an ongoing regime strategy.

7 Aligning the Middle Class with Autocracy: Rhetoric and Practice Our key task is to create a strong middle class that will serve as the basis for a robust economy and socio-political stability. — nur ota n pa rt y d o ct r in e , a d o p t e d at t h e x v pa rt y co ng r e s s, 2013

It was simpler when there were peasants, the proletariat, and the intelligentsia. — b o l at ta l ip ov i ch be r e s ba e v, s e cr e ta ry o f the pa rt y nur ota n1

for nearly three decades after independence, Kazakhstan, an oil-rich, post-communist state in Central Asia, had just one leader, Nursultan Nazarbayev. Nur Otan, the regime’s political party, has helped to stabilize his rule and that of his successor, Kassym-Jomart Tokayev. For the past fifteen years, Nur Otan has served to ensure elite cohesion and mobilize society around the regime’s ambitious development goals.2 Planning documents with titles like “Strategy 2050” lay out these goals using familiar rhetoric. Drawing on the examples of Russia, China, Singapore, South Korea, and Taiwan, as 1. In an address to the conference “The New Doctrine of the Nur Otan Party in the Presidential Strategy for the Renewal and Modernization of the Country,” which I attended in Nur-Sultan (formerly Astana) on December 19, 2013. My own translation. 2. Like other dominant parties, it has primarily used economic co-optation to forestall opposition among the Kazakh elite. Where such efforts have faltered, the regime has resorted to repression (McGlinchey 2011). 193

194 c h a p t e r 7

well, of course, as its own Soviet past, the regime has set ambitious objectives. Announced during Nazarbayev’s annual address to the nation in December 2012, Strategy 2050 calls for Kazakhstan to join the world’s thirty most developed countries by mid-century. In support of the new state strategy and in anticipation of its fifteenth anniversary, the Nur Otan party adopted a new party doctrine one year later. How did the party describe its mission in the coming years? It writes: “Our key task is to create a strong middle class that will serve as the basis for a robust economy and socio-political stability.” Kazakhstan is not the only authoritarian state to put forward the explicit strategic aim of increasing the size of the middle class. Among the strategic development goals of Russia’s Strategy 2020 was expanding the size of its middle class to more than half of the population. When the Russian Ministry of Economic Development presented its Strategy in Moscow in 2008, increasing the size of the middle class ranked just behind raising per capita GDP, the economic growth rate, and life expectancy and ahead of higher export volumes and labor productivity. Kazakhstan’s efforts are indicative of, as well as directly influenced by, the efforts of other states, including those just mentioned. This chapter makes two primary arguments. The first is that Kazakhstan’s regime has self-consciously adopted a view of the symbolic and strategic importance of the middle class as a guarantor of stability and source of societal mobilization. The second is that the state’s high degree of economic engagement and effective use of selective incentives have facilitated the growth of a state-dependent middle class that is unsupportive of democracy and political competition. In the sections that follow, I trace Kazakhstan’s pattern of middle-class expansion to a mode of late development and tactics of state capitalism that privilege the economic role of the state. I show that this project is ongoing, and not only a legacy of Kazakhstan’s Soviet past. Although Kazakhstan has privatized a large share of its economy since independence, its development has disproportionately favored growth of the middle class within a few state-dominated industries. Rather than a middle class with its own independent economic base, Kazakhstan’s middle class remains largely a client of the state. The chapter consists of five main sections. Using evidence from Kazakhstan, the chapter begins by examining the discursive context in which fostering middle-class growth has become a priority of post-Soviet states. As the epigraphs to the chapter suggest, though symbolically important today,

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 195

the middle class was conspicuously absent from the official class constellation of the communist era. After the October revolution, there were peasant and proletariat classes, plus the intelligentsia, but formally no middle class. The middle class, as such, had to be created in the post-Soviet period. In pursuit of economic development, Kazakhstan’s government has sought to strengthen the middle class, while at the same time keeping it economically dependent on the state and forestalling democratization. In the second section of the chapter, I describe the size and composition of the middle class that has begun to emerge in post-independence Kazakhstan. This is not an easy task given the relative paucity of data. To do so, I draw on a broad range of sources: the World Bank and Statistical Agency of the Republic of Kazakhstan (Stats RK), a nationally representative survey conducted by the European Bank for Reconstruction and Development (EBRD), personal interviews I conducted in Almaty and Nur-Sultan (known as Astana until 2019) while a visiting researcher at the Academy of Public Administration under the President of the Republic of Kazakhstan, as well as secondary sources that have not generally been available to western academics. The discussion helps to illustrate how the rhetoric and reality of middle-class formation in Kazakhstan have diverged. The third section presents data on income sources and employment benefits, which advance the argument that Kazakhstan’s middle class is neither diverse nor independent. These data are presented alongside public opinion data on citizens’ employment preferences, which show that by providing opportunities and inducements that are not generally available in the private sector, for many in Kazakhstan, the state has remained an attractive place to work—official wages notwithstanding. In the fourth section, I show that lack of autonomy from the state is associated with less support for democracy and political competition, particularly among the middle class. While the private-sector middle class is more likely than the working class to support democracy in Kazakhstan, the public-sector middle class is not. In fact, the results suggest that by the definition employed in chapter 2, there is virtually no support for democracy among Kazakhstan’s state middle class. By contrast, among the nascent private-sector middle class, one in five already supports it. Finally, in the fifth section, I consider state efforts to use the middle class to mobilize society. In this context, I draw on both my own observations of the public-sector work life of middle-class professionals and election

196 c h a p t e r 7

observation reports from the Organization for Security and Cooperation in Europe (OSCE). Rather than mobilize society toward greater freedom and innovation, both kinds of evidence suggest that the public-sector middle class is disciplined to support the regime. Kazakhstan’s post-independence political culture is informed by both its Soviet past and the experience of its neighbors in South and East Asia, which have often served as a model for the country’s policies and social prescriptions. This makes Kazakhstan an interesting case for my theory insofar as it begins to suggest how the argument might travel beyond the post-communist region and help us to make sense of the modernization experiences of other developing states.

Social Engineering—Creating a Middle Class Despite its natural resource-based economy and greater authoritarianism, the Kazakh case is indicative of a variety of post-Soviet experiences that have made fostering the development of the middle class a policy goal and political priority. As in the other republics, the early post-independence period was accompanied by a sharp decline in living standards. At the end of its first decade of independence, Kazakhstan’s ranking on the UN Human Development Index was still fourteen places below its ranking in 1991 (Olcott 2002, 130). Among the most dramatically affected were educators, health professionals, members of the scientific and cultural establishment, and civil servants—those who had formed the core of the Soviet middle class. The pauperization of this group only heightened the urgency of establishing a loyal new middle class. Already by the mid-1990s, the government of Kazakhstan had embraced the importance of fostering a domestic middle class and begun looking for its historical precedents. The country’s academic establishment, still closely allied with the state, produced fascinating studies around this time on the possible bases for a post-independence Kazakh middle class. One volume, tellingly titled “The Middle Class—A ‘Test’ of Modernity,” opens its discussion of one of the first post-independence studies of the middle class, based on sociological surveys of urban Almaty in 1994 and 1998, with a wide-ranging discussion of the historical antecedents of an imagined future Kazakh middle class (Sagadiev & Bekturganova 1998).3 3. Archived in the library of the presidential foundation in Almaty.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 197

The study highlights three key groups as comprising the historical middle class of Kazakh society. First is the administrative and bureaucratic class (chinovnichij klass) that was recruited from within traditional Kazakh hierarchies of status and power to serve the Russian empire (189). Second are the educated, urban Kazakhs and pre-revolutionary intelligentsia, who formed the core of the Kazakh nationalist opposition to the tsar (and who, the authors hasten to add, were the product of an indigenous tradition of learning and enlightenment that long predated the spread of Russian schooling in the eighteenth century) (206, 235). Third are the Kazakh “businessmen” and “entrepreneurs” (kazakhskoe predprinimatel’stvo)—described as the risktakers and power brokers of traditional warring, nomadic society—who used force to expropriate “herds, property, and people.”4 The key factors, then, for the making of a post-independence Kazakh middle class were the wholesale destruction of the second two groups in the October Revolution and the survival of the former administrative class. The strategic importance of the middle class, the study concluded, should become a guiding principle of reform (469). Rather than perceive the rise of a middle class as inherently threatening to the regime, Kazakhstan’s political elite viewed it as a political project wholly consistent with the goal of maintaining power. Of the two popular notions of the middle class—one as an agent of change, the other as a bastion of conservatism—the dominant image of the middle class in political and academic discourse in Kazakhstan is of a stabilizing and politically conservative social force. Nazarbayev himself described the middle class in a 2004 interview as that “which has something to lose, which does not want revolution, does not want changes in policy and laws and is a stabilizer and bulwark of the state” (cited in Daly 2008, 57). “One of the criteria of belonging to the middle class,” a study by a Kazakh academic concluded, “is that [such a person] never opposes the existing order” (Asylkhanova 2010, 219). These statements convey some sense of why, and in what spirit, the Kazakh regime embraced the project of building a middle class. Much as hopes for building communism had been pinned on a proletariat that was largely lacking, so were hopes for building a stable post-communist order pinned on the middle class long before anything of the kind existed. The class in question changed, but many of the tropes remained the same. In 1998, when 4. The implication apparently being that Kazakhstan’s might-makes-right capitalism of the 1990s could be better understood through this historical precedent.

198

chapter 7

Kazakhstan’s poverty rate was still about 40 percent, Nazarbayev declared that, “First and foremost, the state must represent the interest of the middle class” (cited in Daly 2008, 71). During its second decade of independence, Kazakhstan became more prosperous by virtually every metric. Buoyed by high oil prices, per capita GDP rose from about $1,000 in the late 1990s to more than $10,000 a decade later, rapidly making Kazakhstan an upper-middle-income country.5 Poverty sharply contracted. Between 2001 and 2018, the share living below the poverty line decreased from 46.7 percent to 7.4 percent, while the corresponding share of those in extreme poverty, with incomes insufficient to procure basic staples, fell from 16.1 percent to just 0.2 percent.6 During the first part of the decade, real wages rose more than 10 percent a year and from 2006 to 2013 increased a total of 64 percent in real terms. As a result, the share of those living on at least $10 a day, the World Bank’s income-based definition of reaching the middle class, more than doubled (World Bank 2015). Twenty years after the collapse of communism, the Foundation of the First President, closely aligned with the presidential administration, convened a conference on how “social engineering” (sotsial’naia inzheneriia) could be used to “reinvent” (rekonstruktsiia) Kazakhstan’s future. Social engineering, the political leadership of the conference explained, consisted of influencing social perceptions and motivations.7 Still on the agenda was the task of building a middle class—a priority the Kazakh experts who participated deemed necessary to ensure Kazakhstan’s stable political development and burnish the country’s carefully cultivated international image (Ivatova 2010; Nurmagambetov & Azhmetov 2010; Urankhaev, Salimov & Zhubaniiazova 5. Source: The World Bank’s World Development Indicators, GDP per capita (current US$). 6. According to the Statistics Agency of the Republic of Kazakhstan: www.stat.gov.kz. 7. The conference proceedings were published under the editorship of Marat M. Tazhin, then a senior advisor to the president and the secretary of Kazakhstan’s National Security Council, as well as president of the Association of Sociologists of the RK. Tazhin, also a former Foreign Minister (from 2007 to 2009), had served in the presidential administration since 1992. His opening remarks to the conference emphasized Nazarbayev’s role as engineer-inchief: “The methodology of social engineering is universal for all social systems. However, the experience of countries like Great Britain, the United States, Turkey, Singapore, Malaysia, and France demonstrate that it is a delicate (tonkij) mechanism and a most subtle (ves’ma iskusnyj) instrument and that only a leader who is capable of understanding deeply the innate nature (vnutrennaia priroda) of his society and facilitating its socialization (sposobstvuiushchij ego sotsializatsii) ... can master it. In our country, the First President clearly represents such a figure.”

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 199

2010).8 One issue raised concerned the effectiveness of existing party structures in mobilizing the middle class. Ivatova (2010) writes: [Kazkhstan’s nascent] middle class does not have political leaders or organizations that it can trust to represent its interest, even though many parties, especially those which call themselves parties of the “right” advertise themselves as representing its views. In actuality, in Kazakhstan, there are no party structures that advance the interests of this social group (162). While the right-wing or market-oriented parties mentioned were a fragmentary group composed largely of pocket opposition forces that were often supportive of the president’s initiatives, the regime nonetheless appears to have been unwilling to cede ground. When the Nur Otan party moved to rebrand itself in 2013, its new party doctrine prominently described the party’s mission in the first of the doctrine’s five sections as the creation of a strong middle class. The doctrine foresees that the middle class will serve as both an economic engine and source of political stability. At least in an aspirational sense, it also expects that business will form the bulk of the middle class and that it will be a source of innovation. When I surveyed the heads of the regional governors’ staff (rukovoditeli apparatov akimov oblastej, g. Nur-Sultan, g. Alamaty) in late 2013, it was clear that these ideas had diffused to the political leadership in the regions. Using a short questionnaire administered during a training program at the Academy of Public Administration in Nur-Sultan, I asked these senior civil servants about the process of middle-class formation underway in Kazakhstan and the government’s efforts to facilitate its expansion in their regions. In response to a question asking what role the middle class should play in Kazakhstan, a majority talked about some kind of fundamental, stabilizing role (e.g, “the primary base on which government will rely,” “it should be the country’s primary economic force,” “a big part of stability and stable development,” “the [social] group that ensures a certain degree of stability, conservatism”).9 8. On the importance of international image to Kazakhstan’s ruling regime, see also Schatz (2006). 9. In Russian, the language in which all responses were recorded, the phrases used were: “osnovnuiu sostavliaiushch’iu na kotoruiu budet opirat’sia gosudarstva,” “dolzhen byt’ osnovnoj ekonomichekoj siloj strany,” “bol’shuiu doliu dlia stabil’nost’ i ustojchivogo razvitiia,” “rol’ sloia obespechivaiushchego nekuiu stabil’nost’, konservativnost’.”

200

chapter 7

The Empirical Middle Class in Kazakhstan The preceding discussion suggests that Kazakhstan’s regime has consciously adopted a symbolic view of the middle class as a guarantor of stable social and political development. The following sections turn to the empirical reality. In particular, I consider the size, composition, and economic autonomy of the middle class as it has developed over the post-independence period in Kazakhstan.

The Stakes of Size Studies by Kazakh researchers have provided widely varying estimates of the size of the middle class. Using diverse criteria, estimates range between 15 and 70 percent of the population (Azirkhanov 2002).10 At the upper bound of these estimates are studies that focus exclusively on subjective measures of class status—an analytical approach employed frequently by those closely aligned with the state. A 2011 study by the Kazakh Institute of Socio-Economic Information and Forecasting (KISEIP),11 for example, finds that 68 percent identify as middle class, 15 percent as lower middle class, and 12 percent as upper middle class.12 Another study conducted in the same year by the Kazakh Institute of Sociological Research under the President of the Republic of Kazakhstan found similarly that 72 percent of respondents reported being middle income (cited in Kudasheva 2012, 64). These figures are largely unchanged from 2005, when 69 percent self-identified as middle class, according to data from the State Statistics Agency (Kudasheva 2012, 63). 10. In general, these claims rest on poorly defined empirical bases and are difficult to appraise. Ruzanov (2010), for example, writes that the core of the middle class in Kazakhstan is between 15% and 30% of the population, but provides no operational definition of that core or empirical basis for his assertion (36). Daly (2008) meanwhile reports that Kazakh experts give figures ranging between 18% and 60% of the population, but does not provide citations to particular studies. Rather, these figures seem to come from a short analytical piece on the website www.centrasia.ru (http://www.centrasia.ru/newsA.php?st=1193213700). In fact, the only study described at any length in that report found that about 17% of Kazakhstan’s urban population in 2007 could be considered middle class according to criteria based on education, occupation, and income higher than the regional median. About 45% of Kazakhstan’s population is rural. Below I review some of the more empirically grounded figures cited in existing studies and reports. 11. Personal communication with Aiman Zhusupova, Coordinator of Social Research at the Institute of World Economics and Politics, Presidential Foundation of the Republic of Kazakhstan, whom I interviewed on November 15, 2013 in Almaty. 12. The other response options were: upper (3%) and lower (2%).

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 201

While such figures provide a convenient basis for asserting that Kazakhstan has already become the middle-class society it aspired to be in the 1990s, they are sharply at odds with objective measures. Ileuova (2012), for example, finds that among those who identify themselves as middle class, 40 to 45 percent are objectively poor, with incomes that on a per capita basis are not far from the poverty line (2–3). This perspective is shared by Kudasheva (2012), who notes that many of the self-identified middle class in the survey by the state statistics service also report having difficulty affording basic items like clothes and shoes (63). Analyzing the 2011 data from the presidential administration’s Institute of Sociological Research, Kudasheva further finds that the share of those who reported being middle class varied little among income groups. By objective criteria, the share of Kazakhstan’s population that can be considered middle class remains small. The World Bank defines the middle class in strictly income-based terms, which may serve as a benchmark for other income-based measures. Using an absolute definition of $10 a day as the lower threshold, the Bank reports that Kazakhstan’s middle class grew from less than 10% of the population in 2006 to more than 20% of the population in 2013 (World Bank 2015, 15). It is important to note, however, that the World Bank’s figure of $10/day—a threshold it uses throughout the developing world—is approximately equal to what Kazakhstan’s State Statistics Agency reported as the monthly income of the lower middle class in 2006 (Daly 2008, 7). We can more precisely estimate the share of Kazakhstan’s population that may be described as middle class using data from the European Bank for Reconstruction and Development’s Life in Transition Survey. These data confirm that by any objective criteria, the middle class remains a minority. Indeed, the LiTS suggests that a more appropriate estimate is close to the lower bound of the range given above. Using the same educational and occupational definition of the middle class that I have used in all of the preceding chapters, the Kazakh middle class comprised about 20% of the population in 2006. Again, this definition includes those with at least some higher education and nonmanual, white-collar occupations (the first three categories of the ISCO-8 classification scheme), as well as university students. In major metropolitan areas, the figure rises to 31%. The size of Kazakhstan’s middle class is constrained both by the distribution of higher education in the population (20% overall) and the structure of the country’s labor market, with just 34% employed in white-collar professions. This figure testifies to the high share of unskilled labor in Kazakhstan’s economy, particularly in agriculture and extractive industries. At the same time, it is worth recalling that in historical terms Kazakhstan’s middle class is not too small to be consequential.

202

chapter 7

table 7.1. Industry of Employment of the Kazakh Middle Class % Agriculture, hunting and forestry Construction Education Electricity, gas, and water supply Finance Fishing, fish farming Health Hotels and restaurants Manufacturing Mining and quarrying Other community, social, and personal service activities Public administration, military, social security Real estate, renting, and business activities Transport, communications Wholesale and retail trade and repair services

1.4 4.7 34.5 6.1 5.4 0.7 9.5 0.7 2.7 2.0 5.4 13.5 0.7 8.1 4.7

Note: This table shows the distribution of the middle class by industry of employment. The data come from the EBRD LiTS, 2006.

van de Walle (2014) notes that during Europe’s age of democratization the middle classes were in fact quite narrow. Similarly, the sociologically defined middle class studied by Ansell & Samuels (2014) in nineteenth century Britain was also no more than a minority.

Composition The data further show that middle-class growth in Kazakhstan has been concentrated in just a few state-dependent sectors. Table 7.1 shows the distribution of the Kazakhstani middle class by industry, using LiTS data. The table strikingly shows that fifteen years after independence, the vast majority of white-collar employment requiring higher education continues to be in the state sector. Although three-quarters of all employment is now formally in the private sector,13 roughly six in ten middle-class jobs remain in education (34.5%), public administration (13.5%), and health care (9.5%), sectors dominated by the state. Further, and in terms of change over time, World Bank data suggest that employment in public administration and in sectors such as health and 13. Source: EBRD structural change indicators, private sector share in employment.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 203

education with a high share of public employees contributed almost a third of the jobs created over the ten-year period from 2003 to 2013 (12). Moreover, according to the World Bank the substantial real wage increases cited earlier were driven primarily by higher salaries in the energy and so-called “budget” sectors (World Bank 2015; Kudasheva 2012). Again, these are sectors with a high degree of state involvement—a point I return to in the next section. The upshot is that poverty has declined most sharply and living standards increased most rapidly in regions with extractive industries and in the largest cities Nur-Sultan and Almaty with very high concentrations of public wage earners. Another piece of evidence concerns the dynamics of small-business and self-employment income. As wages were rising in the public sector, from 2006 to 2010, the share of income from self-employment and small-business ownership actually fell from 11.8 to 9.6 percent (Kudasheva 2012). Indeed, according to the World Bank, the number of self-employed has been inversely related to the health of the country’s economy as a whole. While the economy was growing from 2002 to 2007, the number of self-employed declined steadily by about 2 percent a year. Only after the financial crisis of 2008–2009 began did the number of self-employed briefly rise again by about 1 percent a year (12). This suggests that self-employment has remained a survival strategy in the most underdeveloped parts of the economy rather than an engine of greater prosperity and economic growth. These realities are echoed in the responses of the senior civil servants I surveyed about the situation in their respective regions. When asked which groups constitute Kazakhstan’s middle class, they were just as likely to mention civil servants and public-sector employees like teachers, doctors, and university professors as businessmen and entrepreneurs. Their responses also suggest that Kazakhstan’s middle class remains a largely urban phenomenon. Neither rural self-employment nor small farms are yet the basis for middle-class growth in the regions. Asked what the government would need to do for the middle class to grow in their regions, several respondents noted insufficient support for family farms and rural small business. This is a key point since self-employment and entrepreneurship represent a sizable share of total employment: among the economically active population, fully 29 percent, with more than half of these jobs in the agricultural sector (World Bank 2015, 14). This is not to say that the Government of Kazakhstan has made no efforts toward improving the climate for small business. Across the country, the

204

chapter 7

government has opened one-stop small-business help centers. Throughout the former Soviet states, partnerships with international development banks and aid agencies have produced networks of centers such as these. Nestled among the small retail shops on an upper floor of a large food market in Nur-Sultan was one such center. And near the ATMs at the local Kazkom bank hung a modest poster urging small-business owners who were experiencing problems with state regulatory structures (read: harassment and shakedowns) to call a government help line. On a more programmatic scale, the government has established (as a subsidiary of the large state holding company Beiterek) the fund “DAMU” to support small and medium-sized business.14 Since 2010, the fund has played a key role in the state program Business Road Map (Dorozhnaia karta biznesa (DKB)), which in turn forms part of the government’s unfortunately named “program of forced industrial and innovative development” (programmy po forsirovannomu industrial’no-innovatsionnomu razvitiiu RK). DAMU offers training, subsidized interest rates, loan guarantees, and capital investment with a long ten-year horizon. In honor of Global Entrepreneurship Week, the DAMU fund has co-sponsored a large annual conference for entrepreneurs, which I attended in November 2013.15 At that time, DAMU advertised supporting more than 2,000 projects. In sum, even as the government has broadened its efforts to become a key lender to small business, the share of household income from self-employment and entrepreneurship has barely budged and remains about 10 percent.16 Small business more often remains as a strategy of last resort than a lift into the middle class.

Diverse and Independent? One needs only to look around here in [Nur-Sultan] to see the beginnings of a diverse and independent middle class. — u.s. s e cr e ta ry o f stat e co n d o l e e z z a r i ce , nur - sulta n (f o r m e r ly, a sta na) , o cto be r 20 0517 14. See http://baiterek.gov.kz. 15. The Global Entrepreneurship week events were held at the Ritz-Carlton Almaty, 18-24 November 2013. 16. Data provided by personal communication with Sara Alpysbaeva, Director of the Center of Macroeconomic and Applied Economic-Mathematical Studies at the Institute of Economic Studies, whom I interviewed on October 16, 2013 in Nur-Sultan.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 205 table 7.2. Source of Income % Wages from working in a public organization or public enterprise Wages from working in a nongovernmental organization Subsistence farming Wages from working for an individual employer or on a small farm Self-employment income Interest from savings or investments Profit from my own business

45.7 44.1 35.8 13.3 11.4 1.2 1.1

Note: This table shows Kazakhstanis’ sources of income in 2010 as reported by the Statistics Agency of the RK. The column sums to more than 100% as respondents could indicate multiple sources of income. Source: Kudasheva 2012, p. 84.

The evidence presented in the preceding section contrasts sharply with the hopes that many in the West continue to harbor for the “transition economies.” Yet instead of fostering the growth of a diverse and independent middle class, the style of development pursued in Kazakhstan has led to the growth of a middle class that is heavily reliant on the state. Nowhere is this more evident than in Nur-Sultan, the country’s capital since 1997 (until 2019, Astana)—a city purpose-built to project bureaucratic power and showcase the state’s achievements. Income data from Kazakhstan’s state Statistics Agency give a further sense of the extent to which households rely on the government for economic opportunities. Table 7.2 shows the percent of households receiving income from several sources. Respondents were permitted to choose multiple categories, so the column total exceeds 100 percent. Two points are worth noting. First is the large share of households whose income depends on wages from public organizations or public enterprises (45.7%). By looking at income sources within a given household, this figure arguably gives a clearer sense of the breadth of state dependency than do either the share of public employment (30%) or share of the public sector in GDP (35%) reported by the EBRD.18 Furthermore, these figures concern only wages. Transfers and subsidies by the state, which are not included, would only accentuate the picture. Second is the large share of households engaged in the production of 17. Quoted in Daly (2008, 15). 18. Source: EBRD structural change indicators, private sector share in GDP (%) for 2010 and private sector share in employment (%) for 2009.

206

chapter 7

food for their own subsistence (35.8%)—a reminder that as Kazakhstan has become a middle-income country, prosperity has not always been broadly distributed. Employment preferences in Kazakhstan further reflect the state’s prominent role in providing economic opportunity and improving life chances. Relative to the OECD, the share of total employment in government is markedly higher only in the Nordic countries. Even more notably, Kazakhstan’s central government employs greater than a quarter of all young people, ages 18–35. This figure is substantially above the OECD average. According to Ileuova (2012), who heads the market research firm Strategy, young people in Kazakhstan strive to work for the state, since it is the only place they see mobility. Not long ago, she notes, many wanted to become businessmen (3). In recent years, the sovereign wealth fund Samruk-Kazyna, which unites more than forty-five national companies operating in various sectors of the economy, has become among the most prestigious places to work. Samruk’s investments cover a wide range of activities and industries, but focus on strategic sectors like oil and gas (KazMunaiGaz), electricity (KEGOC, SamrukEnergo), mining (TKS), chemicals (OKhK), and infrastructure (Kazakh Railways, KazakhTelekom, Air Astana). It is also among the largest sources of capital investment in the Kazakh economy (along with Beiterek, the state holding company responsible for DAMU).19 The footprint of SamrukKazyna is huge. As of 2013, it employed more than 300,000 people in over 500 subsidiary companies.20 Alone, KazMunaiGaz, Kazakhstan’s third largest oil producer, has at least a minority stake in virtually all major oil and gas projects in the country. Together with its subsidiaries, it employs over 34,000. Many of Samruk’s constituent national companies are reputed to pay high salaries and provide generous benefits, some of which I describe below. And there are still other national companies that operate outside of Samruk. In addition to the employees that these companies and their subsidiaries hire directly, the state exerts additional leverage over the hiring decisions 19. Personal interview with Daulet Mynzhasarov, Director of the PSTsA Rating Agency and political commentator, November 14, 2013. 20. This information was provided in a presentation on personnel management in the national companies, which I attended at Samruk-Kazyna headquarters in Nur-Sultan on October 28, 2013.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 207

of private firms. This occurs through several mechanisms. One is the State Procurement Law. The law, which covers all subsoil contract holders, requires a tender for goods and services costing more than 31 thousand dollars (Olcott 2007, 45). The provision extends to the hiring of external experts, allowing the state, as Olcott writes, “to ensure that local cadre and service providers get ‘sufficient’ consideration.” Another tactic is to task the implementation of strategic plans, formalized as legislation, to one of the national companies. These entities, directed by the state, in turn, gain “substantial ability to pressure their consortia partners to hire and purchase locally (even beyond the minimum requirements of their contracts)” (27). While these tactics are less direct than outright state ownership, they are visible to citizens and help to reinforce regime stability. Contrasting these patterns of state dependency with traditional notions of middle-class autonomy, Kazakh political commentator Dosym Satpaev writes: [O]ne of the most important criteria of the middle class is that it consists of people who are economically independent of the state. But the truth is this: In Kazakhstan, at the official level it is announced that the number of people like this is growing, but the peculiarities of the existing economic and political system prevent it. After all, the exceptionally high rate of growth of an inefficient bureaucracy, the absence of law, and the de facto monopolization of the financial-industrial groups (FPG), as well as corruption destroys small and medium-sized business, the basis for any middle class. Here the number employed in the state administration grows exponentially while the share of small business declines. The result is the mistaken impression that state officials can become the core of the Kazakh middle class. With this in mind, they are raising salaries and building housing for those who work for the state (Satpaev 2008). Satpaev’s remarks reinforce the point that rising wages and job creation in the state sector rather than private-sector entrepreneurship are responsible for a large share of middle-class growth in Kazakhstan today. They further suggest, in line with the argument I have pursued throughout the book, that the political consequences of a middle class that expands under these conditions will differ from the consequences of middle class expansion where the state plays a secondary role. I next turn to the consequences of Kazakhstan’s

208

chapter 7

state-led development, which have privileged a middle-class core of state officials and concentrated rents in the public sector.

The Benefits of Being (State) Middle Class States like Kazakhstan invest in an autocratic middle class by raising publicsector salaries, providing generous benefits, and offering opportunities to earn informal rents. Since there is no centralized selection process for new entrants to the civil service, obtaining a government job is typically done through connections. The absence of a strong merit-based framework for retention and promotion means that official corruption is widespread. In addition, many public-sector jobs in Kazakhstan, as well as jobs at state enterprises and their subsidiaries, come with benefits that are not generally available to private-sector employees. These include but are not limited to housing, which Satpaev mentions, medical care, transportation subsidies, and access to credit on preferential terms. Even as Kazakhstan contemplates civil service reforms to improve government efficiency, it aims to mandate the provision of state housing for all civil servants. In Kazakhstan, I lived in government housing for civil servants along with many employees of the Academy of Public Administration, which is a state institution under the presidential administration. The building’s facilities were newly constructed and very decent. Nur-Sultan is an expensive city and I lived with many young professionals from the regions. On NurSultan’s private real estate market, they would have found it difficult to secure housing. Although the Academy was associated with the presidential administration, the pay was not overly generous. The point was that working within the state created opportunities. In the meantime, it also supplied highly subsidized housing and free transportation to and from the workplace. For those at higher levels, employment granted access to preferred medical clinics associated with the state organs. In a country where access to quality health care is low (Olcott 2002, 202–203), these were significant perks. The benefits offered by state enterprises are no less generous. For example, at the state gas giant KazMunaiGas, the state’s developmentalist agenda is reflected in a variety of job training programs focused on providing upward mobility. These include a Center for the Development of Personnel, a program to increase the professional qualifications of “young specialists,” internships with foreign firms, advanced training opportunities at the Kazakh-British University, and other educational grants (Olcott 2007, 45, 46).

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 209

KazMunaiGas is also involved in the construction of medical centers, rehabilitation and rest homes, recreational facilities, and worker housing.21 As in the Soviet era, the benefits of employment at KazMunaiGas extend beyond retirement. The company makes lump sum payments to new retirees and pays annual bonuses on special birthdays and national holidays. A newspaper subscription, paid vacation every three years, and death benefits ensure that employees are cared for until they reach the grave (44-46). Notably, these benefits have survived even as shares in the company came to be traded on the London Stock Exchange. More programmatically, Bissenova (2013) describes how various state housing policies have targeted benefits toward the expansion of the publicsector middle class. Her work provides a fascinating portrait of how these benefits have enhanced the social mobility of her informants, many of whom are employed in the less well-remunerated social sphere. Two types of state housing programs have targeted the public sector. Under the public housing scheme, priority for housing subsidies was based on a system that awarded points both for current and past employment in the state sector. This extended to public-sector workers of the social sphere (e.g., teachers, health care workers), employees of state enterprises, and civil servants. The programs also cultivated loyalty among those just entering the workforce by giving additional points to young families. In turn, to remain participants in the program, applicants were “discouraged” from changing their place of employment—a fact that underscores the excludability of these benefits (Bissenova 2013, 15). At a speech in 2011, Nazarbayev announced that 350,000 families could be expected to gain housing through the program by 2014 (quoted in Bissenova 2013, 16). A second workplace-based program provided additional benefits by combining elements of the old Soviet housing allocation system (based on place of employment) with the new market system of private ownership. This meant that state workers received an exceptionally valuable perk: at least initially, they were free to sell the highly subsidized real estate they acquired through their workplaces at a profit. The example of a state medical clinic in Nur-Sultan gives a general sense of the size of these subsidies. While doctors, nurses, and laboratory workers paid $350 per square meter, Bissenova reports that apartments in the same building sold on the private market for 21. While some of these are for the benefit of local communities, in practice they disproportionately benefit the company’s employees.

210

chapter 7

$1000 per square meter (17). In addition, when the financial crisis hit in 2008 and construction companies went bankrupt, public-sector beneficiaries of these programs found themselves in a more secure position than others who had purchased apartments directly on the market. As Bissenova concludes, “The state invests in the middle class and the middle class reciprocates in its investment in the state” (24).

Democracy and Kazakhstan’s Middle Class The political upshot, I have argued, is that pathways to the middle class through the state create constituencies that benefit from the status quo and are more likely to oppose democratization. Because democratization both threatens the stability of public-sector jobs as well as the unique access of state employees to benefits and rents, I have argued that the public-sector middle class will be less supportive of democracy. The choice of reliable, nationally representative surveys from Kazakhstan is seriously limited. I turn back to the EBRD data from chapter 3 to illustrate the central argument in the context of Kazakhstan’s state-dependent economy. Using only the Kazakh country sample, I again regress the binary measure of democracy support described in chapter 3 on the key explanatory variables middle-class and public-sector employment, plus the controls (age, age squared, and gender). The baseline model reported in the first column of table 7.3 confirms the expected pattern. Using simulation to calculate more readily interpretable predicted probabilities from these logistic regression coefficients, I find that state employment has a negative effect on democracy support that is especially large among the middle class. While the predicted probability of democracy support22 among the private-sector middle class is 21.05 percent, with a 95 percent confidence interval of [14.68, 29.31], among the state middle class it is just 3.19 percent [0.87, 11.69]. This is a substantively important and sizable difference, which implies that sector of employment strongly shapes the Kazakh middle classes’ preferences over democracy. One possibility, however, is that differences in material well-being rather than state employment per se account for observed sectoral variation in democracy support. Could the state middle class be less democratic because 22. As elsewhere, these predicted probabilities are calculated with all other covariates in the sample held at their observed values (Hanmer & Kalkan 2013).

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 211

it enjoys greater wealth from informal rents and formal benefits and is thus more fearful of redistribution? If this is true, and the source of economic opportunity does not matter independently, then we would expect introducing a control for material well-being to diminish the apparent effect of state employment. Column 2 adds a proxy for well-being, an additive index of the number of consumer items that a household possesses. In fact, the coefficients in column 2 change very little after controlling for consumption. The results confirm two things. First, material well-being has a marginally significant positive effect on democracy support, in line with modernization theory. Second, material well-being alone is insufficient to explain democracy support; dependence on the state matters independently and, indeed, matters more than material well-being. Column 3 adds additional controls. First, I investigate whether ethnicity confounds the effect of being middle class or state employed. If ethnicity is correlated with either of these variables and also with democracy support, then the preceding results may be biased. In the Soviet era, much of Kazakhstan’s educated professional class was ethnically Russian. During the breakup of the Soviet Union and its aftermath, many ethnic Russians returned to Russia. In addition, Kazakhstan is home to ethnic Uzbeks, Uygers, Dungans, Kyrgyz, Ukrainians, Tatars, Belarusians, and Germans.23 Because Kazakh is the sole state language (Russian holds the status of an “official language”), ethnic Kazakhs have received preferential treatment in government employment and education (Freedom House 2004). Given that Kazakhstan’s ethnic minorities may be less likely to be employed in the state sector today, the possibility that ethnicity confounds the apparent effect of state employment on the middle class merits further examination. Minority ethnicity may be measured in two distinct ways. One approach is to simply focus on respondents who identify with a non-titular (that is, non-Kazakh) ethnicity. Another approach is to focus on respondents who self-identify as ethnic minorities. I use both and discuss below their implications. Because the experiences of Russians and other slavic groups

23. Of those who consider themselves members of an ethnic minority in Kazakhstan, 48% are ethnic Russians, 8% are Dungan, 7% are Uygur, 7% are Uzbek, 5% are Ukrainian, 4% are Tatar, 3% are Belarusian, 3% are German, and 10% are some other ethnicity. Interestingly, many ethnic Russians surveyed in Kazakhstan did not self-identify as an ethnic minority.

table 7.3. Logistic Regression Results Predicting Democracy Support Dependent Variable: Democrat Base Model (1) State employee

Middle class State × middle class

−0.452 (0.329)

Add Income (2) −0.479 (0.330)

Additional Controls (3) −0.476 (0.332)

0.622∗∗ (0.259)

0.533∗∗ (0.265)

0.565∗∗ (0.271)

−1.653∗∗ (0.824)

−1.640∗∗ (0.824)

−1.682∗∗ (0.826)

0.129∗ (0.074)

0.134∗ (0.077)

Consumption index

Russian/Slavic ethnicity

−0.018 (0.318)

Non-Kazakh ethnicity

−0.005 (0.303)

Dependent Variable: Strong Opposition Regional Controls (4) −0.646∗ (0.361) 0.704∗∗ (0.309) −1.580∗ (0.871)

−1.044∗∗ (0.488)

Non-Kazakh minority

−0.531 (0.471)

Urban

Constant

Controls Region fixed effects Observations

−0.474 (0.387)

−0.688 (0.430)

0.208 (0.221)

0.404 (0.275)

−2.805∗∗∗ −2.866∗∗∗ −2.872∗∗∗ −2.181∗∗ (0.758) (0.760) (0.779) (0.920) 

 979

 979

  979

0.033 (0.021) 0.086∗∗ (0.037) −0.108∗∗ (0.045)

0.213∗∗ (0.089)

Russian/Slavic minority

Metropolitan

Alternative DV (5)

0.871∗∗∗ (0.071) 

979

880

Note: ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01. Controls for age, age squared, and gender were included in all models. The reference category for metropolitan and urban is rural. In columns 1–4, the models are logit. Column 5 is OLS. Source: EBRD LiTS, 2006.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 213

differ in important ways from the experiences of other non-Kazakh groups, I distinguish between them in the analysis. Second, life in Kazakhstan’s major urban centers differs dramatically from life in its auyls (Kazakh for villages). While rates of state employment in the population are similar across rural, urban, and metropolitan areas, high rural unemployment means that state jobs make up a larger share of total employment outside of cities. If Kazakhstan’s large rural population holds systematically different values than its urbanites and rurality is also related to the other variables of interest, it may confound the results. Third, Kazakhstan’s regions differ markedly in their level of economic and social development. Most of the country’s GDP is concentrated in the cities of Nur-Sultan and Almaty, the main oil-extracting region of Atyrau, and the industrial regions of Karagandy and South Kazakhstan. Kazakhstan’s regions also have widely divergent economic bases, from agricultural producers like Kostanay, Akmola, and North Kazakhstan to extracting regions like Mangystau, Aktobe, West Kazakhstan, and Kyzylordy. While state employment makes up a fraction of total employment in the grain-producing region of Kostanay,24 in the rich extractive region of Mangystau, the state’s role in the local labor market is substantial.25 Given these differences, does democracy support vary systematically by region, net of the other factors already discussed? Could other types of regional variation explain away the apparent effect of state dependency? I next explore the role of ethnicity, living in an urban area, and region. As we see in column 3, neither living in an urban center nor ethnicity substantially affects the other coefficients. In fact, none of these variables even reaches conventional levels of statistical significance. Once other factors are considered, there is little evidence that living in an urban area impacts democracy support. There is also little evidence that ethnic self-identification has an independent effect on regime preference. In light of this evidence, it seems unlikely that rural to urban migration, as Kazakhs came to the cities and Russians returned, has made Kazakhstan’s cities any more (or less) democratic.

24. The share of state employment is 13% in Kostanay according to the EBRD survey, and 17% according to the Statistical Agency of the Republic of Kazakhstan. 25. The corresponding figures are 40% in the EBRD survey and 39% according to the Statistical Agency of the Republic of Kazakhstan. In the city of Nur-Sultan, the figure is greater than 50%.

214

chapter 7

Column 4 approaches the definition of ethnic minority status differently— this time focusing on self-identification as a minority. It also adds controls for each of Kazakhstan’s regions. Again, if those who belong to a non-Kazakh ethnic group and self-identify as minorities are less likely to work for the state (or reach the middle class) and are also more supportive of democracy, the previous findings could be biased. Instead, I find that being a minority is associated with less not more favorable attitudes toward democracy. The results also suggest that minority status does not confound the effect of state dependency. While subjective self-identification as a minority carries the same negative sign as the non-titular ethnicity variables examined in column 3, the estimated effect is much larger (βSlav = −1.04, βNonSlav = −0.53), and for Slavs this estimate is highly statistically significant. These results seem to imply, as in Ukraine,26 that post-Soviet (ethnic) minorities are more likely to see democracy as a system of majority rule rather than minority rights. Subjective self-identification as a minority heightens this effect. Crucially, however, whether we control for ethnicity or minority self-identification, the other coefficients remain unchanged and the effect of state employment on regime preferences, especially for the middle class, continues to be stronger than the effect of these other identity-based factors. It is also worth noting that, net of regional heterogeneity, the main results in column 4 remain the same. Further, the model suggests that after accounting for the characteristics of individuals residing in a region (e.g., class, sector of employment, consumption, age, gender, and minority status), level of development does not explain how supportive a region’s citizens are of democracy.27 For example, while Eastern Kazakhstan and Northern Kazakhstan share identical ratings on an index of regional competitiveness (Temirova & Abdimomynova 2016) and very similar regional income (GRP) per capita,28 all else equal, Northern Kazakhstan is more democratic while Eastern Kazakhstan is less. Additionally, there is no evidence that regional reliance on extractive industry contributes to lower levels of democracy support. This finding mirrors the country-level evidence presented in the book’s conclusion. 26. See the evidence on ethnicity in chapter 6. 27. This is true whether level of development is measured by an index of competitiveness that includes GRP per capita, investment, poverty rate, number of small businesses, and industrial production; GRP alone; or various social indicators. 28. See the Statistical Agency of the Republic of Kazakhstan, http://stat.gov.kz/api/ getFile/?docId=ESTAT080539.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 215

Kazakhstan’s resource-driven economy is centered in oil and gas producing regions such as Atyrau and Mangistau and regions rich in mineral resources like Eastern Kazakhstan, Karaganda, and Pavlodar. After accounting for sectoral differences in employment and other individual characteristics, these findings suggest that living in an extractive region is unrelated to democracy support. For example, residents of Eastern Kazakhstan, a mining region, and Zhambyl, an agricultural region, share similar attitudes toward democracy, all else equal. To summarize, then, controlling directly for ethnic identity and geographic location, both class and state employment maintain a sizable independent effect on support for democracy. After accounting for all other covariates (age, gender, income, majority/minority ethnicity, living in an urban center, and region), I estimate that being middle class increases support for democracy on average by 5.2 percentage points, while state employment decreases it by an average of 7.2 points. Moreover, as in the baseline model already described, state employment has an even more pronounced effect for the middle class: the model in column 4 implies that state employment decreases middle-class support for democracy by 16.7 percentage points. To better understand the substantive importance of state employment, we can compare the magnitude of its effect with that of other variables. In fact, its effect is equal to or larger than that of either income or minority ethnic status. Increasing consumption from the first to the 75th percentile raises the probability of democracy support by 3.8 percentage points. Identifying as a Slavic minority lowers it by 7.3 points, while the effect of identifying as a non-Slavic minority is indistinguishable from zero. The impact of state employment on middle-class support for democracy thus eclipses these other factors. In short, these results confirm two things. First, that social class adds to our understanding of potential democratic coalitions in Kazakhstan and, second, that state employment’s negative effect on democracy support cannot be attributed to public-sector workers’ demographic profile. The point is not that state economic engagement is the only variable that shapes attitudes toward democracy. Clearly, multiple factors contribute. This evidence, however, suggests that the state’s role in shaping regime preferences is substantial, and may in fact be stronger than other factors linked to identity and place, which have received greater attention. Finally, column 4 presents the results of an OLS model predicting the importance of a “strong political opposition,” measured using a five-point

216

chapter 7

Likert scale coded to vary between zero and one, where a higher score represents stronger support for political competition. The pattern remains the same as with the other dependent variable. The state middle class is again significantly less likely to approve of a strong political opposition than the private-sector middle class. Again, in fact, the state middle class is no more democratic than the working class. Given that Kazakhstan’s official privatization figures belie the state’s actual role in the economy, I next investigate whether these results extend to the strategic sectors. By holding shares in an extensive network of subsidiaries and investments of its national companies, the government of Kazakhstan has effectively extended its leverage over parts of the economy that have nonetheless attracted extensive private capital. Breaking the data down further by industry confirms the same basic pattern. If we look at employees of strategic sectors (i.e., the extractive industries, utilities, and finance), where state engagement is high but where many companies are technically private, we find weaker support for political competition among the middle class. In particular, middle-class employees in industries dominated by subsidiaries of the national companies with varying levels of private participation are 20 percentage points less likely to agree to the importance of a strong political opposition than middle-class employees in industries where state participation is slight (the average scores for the two groups on the political opposition item are significantly different at the 0.1 level, t = −1.81, df = 28.61, p = 0.081).29 A potential weakness of this analysis is that the number of cases is too small to introduce additional controls. The results, such as they are, imply that the true effect of state-dependent employment in Kazakhstan is probably larger than estimates that rely on official distinctions between public and private. In other words, the already large sectoral differences calculated above are a conservative estimate.

The Middle Class and Societal Mobilization Rhetorically and in practice, Kazakhstan’s regime has cast the middle class as an agent of social mobilization. At the same time, it has disciplined its 29. In the comparison group where state participation is low, I include the following industries: hotels and restaurants; real estate, renting, and business activities; wholesale and retail trade and repair services; domestic labor; agriculture, hunting, and forestry; and fishing and fish farming. However, the latter three categories contribute very few middle-class observations.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 217

representatives to advance the modernization goals of the state.30 When I interviewed him in Nur-Sultan in December 2013, the director of Nur Otan’s public policy institute described a two-pronged strategy involving the middle class.31 The first task he said was to build a robust middle class that valued freedom, rule of law, and solidarity, which he saw as core middle-class values; the second was to instill the values that would help Kazakhstan achieve technological breakthroughs. He told me that the party aimed, through the middle class, to mobilize society behind the regime’s modernization efforts. Though he cited developmental states like South Korea and Taiwan as examples, the language he employed—of societal mobilization to achieve the goals of the state—also echoed the old communist rhetoric. In practice, it is hard to see how building a robust middle class that values freedom is a priority, given the quotidian experience of the middle class in jobs where the state has greatest leverage. Public-sector jobs frequently still consist of a high degree of control on the part of the state as well as rigid hierarchies and a culture of deference to authority. In some cases, workplace-based control goes by the new name of performance management and productivity enhancement. Though many of these concepts and practices are borrowed from the West or successful recent modernizers in East Asia, they receive their own local interpretation. These practices inculcate discipline, heightening the effectiveness of selective incentives. For example, all 300,000 Samruk employees fall under a comprehensive performance-based review and rewards system, which directly affects both pay and career perspectives.32 At the Academy of Public Administration under the Presidential Administration, where I was affiliated during my fieldwork in Nur-Sultan, at one point employees were asked to account for their time each day in 50-minute intervals. Cameras filmed all movement through the halls; magnetic cards were used to record movement into and out of the building; computer logins and logouts were tracked;

30. Alima Bissenova has also seen in the implementation of Kazakhstan’s public housing program an effort to discipline the middle class to moderate its expectations, lengthen its time horizons, and depend for its well-being on the state (Bissenova 2013). 31. Personal interview with Nurbek Sayasat, Director of the Institute of Public Policy of the Nur Otan Party, December 19, 2013. 32. This information was provided in a presentation on personnel management in the national companies, which I attended at Samruk-Kazyna headquarters in Nur-Sultan on October 28, 2013. I did not observe the corporate side of performance management in Kazakhstan, but it would be a worthy subject for future ethnographic research.

218

chapter 7

and web filtering software33 was used to limit access to undesirable internet content, including opposition news websites, Twitter, and YouTube.34 Such supervision is disciplining. State selective incentives are in turn used to both mobilize state workers for the regime and demobilize them from participating in opposition political activity. Civil servants supporting the opposition have been fired from jobs (D’Anieri 2007, 228). There are also widespread accounts of state organizations mandating support for Nazarbayev and other forms of workplace-based mobilization. As one doctor was told unironically by a supervisor tasked with signature-gathering for a regime candidate: “I’m warning you: this is voluntary!” (Human Rights Watch 1999). Indeed, pressing the public-sector middle class into the state’s service at election time—a practice chapter 3 discussed during Russian parliamentary and presidential elections—also remains common in Kazakhstan.35 In effect, by targeting selective incentives to public-sector professionals, the regime gains not one vote, but the cooperation of those who have the human and social capital to also net additional votes. This occurs both through their ability to manipulate vote counting, which frequently takes place in public institutions, and by enlisting them in various forms of campaigning and voter persuasion. The OSCE observer mission reported active public-sector involvement in Kazakhstan’s 2015 early presidential election: Many Nur Otan regional offices are located in state institution buildings. Government officials, including university rectors appointed by the 33. Among the features of this software, the company’s website advertised, were to: “Get detailed usage reporting with Kerio Control Statistics. This component lets managers and admins view the Internet activity details of individual users—from a list of all sites visited and when, to the specific search terms users enter on search engines and regular websites with search capabilities. Use these granular usage insights to refine traffic-shaping rules, monitor employee performance, and more. Best of all, these highly granular reports can automatically run on a schedule and be emailed to you, ready for your review—no need to actively pull reports each week.” Source: http://www.kerio.com/products/kerio-control. 34. The blocking of YouTube resulted in a funny, if awkward, moment at an international conference on civil service reform hosted by the Academy, when a South Korean speaker, invited to talk about online and e-governance initiatives tried to show a video as part of his presentation. It took several awkward minutes before anyone present was willing to tell him that access to YouTube was prohibited from the Academy’s network. 35. See also Forrat (2018) on the role of Russian schoolteachers in elections and Larreguy, Montiel Olea & Querubin (2017) on the institutionalized alliance between Mexico’s teachers and the PRI.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 219

President, took an active role in the incumbent’s campaign meetings. This, together with the incumbent’s institutional advantage and the fact that he drew on a broad network of public sector employees, blurred the line between State and party (OSCE 2015, 8). This blurring of the line between state and party, state and society is at issue particularly for public sector professionals, since they can be organized at lower cost to the regime. Over the years, the OSCE has documented how state workers often serve on local election commissions that are themselves often located in public institutions like schools. They are recruited to volunteer in political campaigns. They are dragooned to take part in pro-regime rallies and mobilize others including subordinates and students to do the same. And they are pressured through their workplaces to vote for the correct party or candidate (OSCE 2015; OSCE 2012; OSCE 2006; Human Rights Watch 1999). The regime in turn uses the public sector’s compliance in these activities to demonstrate and augment its own power and perceived popularity among citizens (Wedeen 1999).

Conclusion This chapter has advanced two primary arguments: first, that Kazakhstan’s regime has self-consciously adopted a view of the symbolic and strategic importance of the middle class as a guarantor of stability and source of societal mobilization, and second, that the state’s high degree of economic engagement has resulted in a middle class that is unsupportive of democracy and political competition and overwhelmingly state dependent. The growth of the middle class in Kazakhstan has not been broad-based. Rather, state strategies to target particularistic benefits and raise wages in the public sector and strategic industries closely aligned with the state have fueled Kazakhstan’s middle-class expansion. In short, Kazakhstan’s public-sector middle class is the frontline of regime efforts, borrowing the language of Schatz & Maltseva (2012), “to persuade [society] of the rectitude of rule, generating a minimum compliance that ensures stability and regime survival” (47, emphasis in the original). In particular, I have shown that soon after the collapse of communism, Kazakhstan’s leaders already envisioned a new challenge of social engineering: the formation of a broad middle class that would ensure economic and political stability. This task was associated from the beginning with the project of building an independent and prosperous state. More than twenty years later,

220

chapter 7

this view was adopted as party doctrine, and I found it well entrenched among the high-ranking regional officials whom I surveyed. While Kazakhstan made great strides in alleviating poverty during the 2000s as the economy boomed, living standards rose most rapidly in the cities and regions with high concentrations of public workers or extractive industries. At the same time, the share of household income from self-employment and entrepreneurship has remained level at just 10 percent. The political consequences of statedependency among the middle class in Kazakhstan are stark: I find that the predicted probability of support for liberal democracy among the state middle class is just 3 percent, while among the private-sector middle class, it is substantially higher (21%). Finally, I show that the preferences of middle-class employees in industries dominated by an extensive network of subsidiaries of the national companies, many of them with varying degrees of private ownership, closely resemble those of state employees. By way of conclusion, I consider two points that impact how the argument and evidence presented in this chapter are likely to travel. First is the issue of official privatization and hidden state control of the economy. Second is the impact of natural resource endowments. Officially, Kazakhstan’s private sector has accounted for approximately 65–70 percent of domestic economic output since the mid-2000s.36 By this yardstick, the extent of privatization in Kazakhstan’s economy is lower than in more democratic post-Soviet countries like Georgia, Armenia, and Kyrgyzstan, and compares favorably with more autocratic Tajikistan, Turkmenistan, Uzbekistan, and Belarus. It is also worth underscoring that the official extent of privatization in Kazakhstan belies the state’s actual underlying role in the economy through subsidiaries of its strategic national companies.37 And yet, as Junisbai (2012) rightly notes, Kazakhstan’s lack of progress on most measures of democracy since the early 1990s appears puzzling for theories that expect privatization to be associated with greater political pluralism. The evidence in this chapter implies at least one possible demand-side explanation for this apparent anomaly. The findings in this chapter suggest that private employment in closely aligned 36. Source: EBRD structural change indicators, private sector share in GDP. Note also that this figure is lower than the share (80%) cited by Junisbai (2012); in fact, the extent of privatization in Kazakhstan is lower than in Georgia, Azerbaijan, Armenia, or Kyrgyzstan, according to the most recent data from the EBRD. 37. Radnitz (2010) relatedly argues that revenues from the commanding heights of the economy have continued to flow to Nazarbayev and members of his family, helping to forestall the influence of an independent capitalist class, despite extensive formal privatization.

a l i g n i n g t h e m i d d l e c l a s s w i t h au t o c r a c y 221

strategic sectors effectively stifles support for democracy and political competition in ways that resemble the effect of employment in the public sector. Further, if people who are employed in technically private but closely statealigned enterprises hold similar preferences to state employees, then the true gap in demand for democracy between workers who are and are not state dependent is likely much larger. The estimates based on official distinctions between public and private presented above should thus be considered conservative. A second issue concerns whether resource rents alone are responsible for the loyalty of Kazkhstan’s state middle class. The importance of Kazakhstan’s extractive industries has risen substantially since the 1990s when they made up roughly 15 to 20 percent of GDP.38 Since 2000, natural resources have contributed twice that share toward the country’s economic output. It is tempting to conclude that the dynamics of state-dependent middle-class growth documented in this chapter are peculiar to Kazakhstan’s resource economy. There is little evidence, however, for this view: either sub-nationally in this chapter or cross-nationally in this book’s concluding chapter. Across the postcommunist autocracies, in the resource and non-resource economies alike, the state middle class more closely resembles the working class in its minimal enthusiasm for democracy. Where the post-communist resource states likely have an advantage is in also capturing the support of those whose jobs are formally outside the public sphere. For the private-sector middle class employed in strategic industries, state capitalism effectively generates the discipline and dependence that limits demand for democracy.

38. Source: The World Bank World Development Indicators “Total natural resources rents (% of GDP).”

8 Conclusion The 1989 middle-class revolutions were not uniformly successful. ... The nonsuccessful cases with their newly restored authoritarian political systems and oligarchic state capitalism are probably more important in thinking about the political repercussions of the middle-class expansion and its transformational potential in other parts of the world. — g r z e g o r z e k ie rt ( 2010, 111 )

over recent decades, middle-class expansion in the developing world has reignited old debates. The great majority of new entrants to the middle class reside in countries that are nondemocratic. This raises several related questions. Will an expanding middle class favor democratization? Does the addition of the middle class, as Aristotle wrote in the fourth century BCE, indeed “turn the scales,” making the prospects for democracy stronger? Under what conditions will growth of the middle class produce bottom-up pressure for democratization? This book has proposed a framework for thinking about the state’s role in middle-class formation under autocracy and its impact on individual incentives to support democracy. The preceding chapters developed an explanation for why growth of the middle class may be only tenuously related to democratization. Rather than approach the middle class as a unified actor, this book told a tale of two middle classes with divergent political interests. Drawing on an earlier macrosociological literature on democratization in Europe, I connected the political loyalties of the middle classes to their position vis-à-vis the state. However, in contrast to these earlier studies, my argument focused on the micro-level mechanisms that affect bottom-up demand for democracy and on qualities of today’s late-developing authoritarian states. This account suggested that 222

c o n c l u s i o n 223

education, occupation, and income do not have an unconditional effect on democratic attitudes. Instead, the middle classes’ interests, attitudes, and actions are also shaped by expectations about how changes in control of the state might jeopardize their future prospects and by more immediate forms of regime retaliation. Two factors were shown to affect how rising middle classes view their interests in these cases: first, the state’s power to bestow or withhold benefits, in particular through public-sector employment, and second, individuals’ alternatives in resisting state incentives. These options are limited when an extensive public sector crowds out private-sector alternatives. By tying future benefits to regime continuity, state capitalism tends to create its own middleclass constituencies, encouraging distrust of democratization and dividing potential democratic coalitions. This argument draws attention to the simultaneous part states play in shaping social stratification and patterns of political incentives. In so doing, it has offered a distinct perspective on debates about economic modernization and redistributive preferences, which have often sidestepped these critical roles of the state. Moreover, while the existing democratization literature has often emphasized class interests, scholars have been preoccupied principally with workingclass movements and the strategic choices of elite actors (e.g., Therborn 1977; O’Donnell & Schmitter 1986; Rueschemeyer, Stephens & Stephens 1992; Collier 1999). Far less often have the middle classes been the subject of empirical research. To the extent that the middle classes have appeared in empirical research, the subject has been business owners and entrepreneurs, rather than professional strata. Studying groups identified a priori as potential change agents has left out of the literature large segments of the middle class, who are neither working class, nor economic or political elite. This leaves a significant blindspot, since, as Banerjee & Duflo (2008) find, the developing world’s middle classes are not distinguished by their entrepreneurship, but instead by stable, salaried employment. This book looked beyond business owners and entrepreneurs to the salaried middle classes and has shown how white-collar professionals can be a source of regime resilience.

Findings I went on to test the implications of the loyal middle class argument in four distinct contexts. Using multiple methods, in each of these analyses, I confirmed that how one gains and maintains middle-class status matters

224

chapter 8

when the regime is nondemocratic. In chapter 3, a multicountry study of the political preferences of the post-communist middle classes, I found clear evidence that, under autocracy, dependence on the state attenuates support for democracy, especially among the middle class. More specifically, I showed that, across the post-Soviet nondemocracies, being middle class is associated with more support for democracy only among those with greater economic autonomy from the state. Fewer private-sector alternatives were correlated with more status quo bias. Careful tests suggested that these results are unlikely to be explained by self-sorting into different career paths based on preexisting political preferences. An important implication of these findings is that middle-class growth in the state sector cannot be expected to enhance popular support for democratization. At the same time, the results suggest that, on average, expanding the private-sector middle class is a good strategy to promote democracy in autocratic settings. The large multicountry survey evidence in chapter 3 confirmed the generalizability of the argument across the post-Soviet space. It also showed the argument’s explanatory power across a variety of autocratic regime types—from more consolidated authoritarianism in Belarus and Uzbekistan to countries like Ukraine teetering on the verge of democracy. Including both democratic and nondemocratic countries in the analysis revealed why prior research on democracies (where there is no democratic deficit among the state middle class) has missed the contingent relationship between middle-class growth and the emergence of democratic values that this book examines. Subsequent chapters “thickened” these generalizations in several ways (Coppedge 2012, 6) and demonstrated that the framework travels to different authoritarian regime types. In chapter 4 on protest in Russia, I showed that the argument explains not only attitudes but also consequential collective action in an authoritarian system with limited political competition. Using data on actual protest participation and a novel research design, I found that the state middle class was far less likely than the private-sector middle class to participate in anti-regime protests, due in part to ideology but also selective incentives. In fact, not only did Russia’s state middle class take part at much lower rates, the public-sector protesters who did participate were notably less likely to identify with the pro-democracy coalition. Moreover, similarly situated protesters from the public- and private-sector middle classes articulated different goals. While the private-sector middle class tended to frame its demands in terms of political rights and freedoms,

c o n c l u s i o n 225

the state middle class placed greater emphasis on improving welfare under the current system. I traced these patterns of participation to the interaction of individual resources and state selective incentives, showing how the power of incentives varied with available alternatives to state employment. In particular, I found support for two mechanisms to explain patterns of participation: the first based on negative incentives, like the threat of dismissal, and the second based on positive inducements grounded in differing economic and social protections. In sum, these findings suggested that cleavages between the public- and private-sector middle classes weakened potential coalitions in support of democratization and helped to reinforce the stability of the Russian regime. The following chapter helped to build the case that public employment itself shapes democratic attitudes by showing that those who choose public and private employment in Russia hold initially very similar political orientations. Using original survey data, I found that less democratic types are not more likely to gravitate to the public sector. In fact, I found that equal proportions of those seeking public and private employment believed that authoritarianism was sometimes preferable. Rather than make career choices on a political basis, informal networks of recruitment, more so than any other factor, appear to drive career decisions. Views on the importance of political freedom, order, national security, and strong economic performance have little relation to the choice of a public-sector position. What differences do exist in regime preferences across sectors plausibly arise only later. Chapter 6 showed how Ukraine’s slow privatization and the regime’s vast power over public-sector patronage shaped its political trajectory during a period of democratic revolution and retrenchment. Using a panel survey, spanning the Orange Revolution, I investigated how changing economic opportunities and career incentives affect democratic preferences, as demonstrated on the streets and at the ballot box. This chapter offered a particularly demanding test of my theory both because Ukraine has not been able to afford the perks of public employment that its resource-rich post-Soviet neighbors provide and because this was, at least nominally, a democratizing period. Nonetheless, using data from a different source and period, I replicated the main attitudinal and behavioral results of the preceding chapters. By exploiting repeated interviews with the same respondents, I was also able to directly address critical questions about causal order and endogeneity— most importantly, whether state employment shapes political preferences or political preferences are already reflected in individuals’ career trajectories.

226

chapter 8

The chapter’s analyses showed three things. First, using a difference-indifferences design and following the same respondents over time, I found that the regime preferences of students with previously identical orientations to democracy diverged once they entered the labor market. New labor market entrants, including those taking up middle-class professions, became less supportive of democracy after joining the state sector. What is more, the preferences of new public sector workers shifted quickly, consistent with changing material incentives that affected how individuals viewed their future prospects under democracy, rather than long-run socialization. Second, the chapter demonstrated that rising middle-class groups were more likely to take part in pro-democracy protests only if they were employed outside the state sector. Individuals with careers tied to the state, instead, rallied in greater numbers behind the nondemocratic incumbent. Third, I showed that the patterns of shifting party support, which facilitated Ukraine’s democratic retrenchment, were shaped not only by regional and linguistic identities but also by class and sector. As the revolution faltered, state employees whose jobs and rents were threatened by the prospect of democratic transition were more likely to defect to the Party of Regions than private-sector employees with similar characteristics. Together, the evidence in this chapter suggested that while the initial choice of public- versus private-sector employment was unrelated to democracy support, state-sector pathways to the middle class subsequently affected regime preferences as well as protest and voting behavior. Finally, chapter 7 evaluated the argument through a mix of survey data, interviews, archival evidence, and thick description, focusing on the vocabulary and practice of state-led middle class expansion in a hegemonic party system. Through a close examination of Kazakhstan, I illustrated some of the strategies states use to cultivate alliances with expanding middle classes: proactive efforts to ideologically align a dominant political party with the middle classes, expanding job opportunities and creating rapid upward mobility in the public sector, and targeting real wage growth and excludable benefits, such as subsidized loans and workplace-based housing schemes, to state-sector professionals. The evidence also highlighted how Kazakhstan’s regime has viewed building an autocratic middle class as a project of social engineering—and a source of societal mobilization. In addition, this chapter probed a variety of pathways into the middle class that, while dependent upon the state’s economic intervention, formally take place outside of the state sector. I found that, in industries dominated by

c o n c l u s i o n 227

the state, the regime preferences of the private-sector middle class closely resembled those of state employees. This chapter thus extended the book’s basic framework to show how regimes successfully wield new forms of state capitalism—deploying strategic equity investments, national companies, and extensive networks of their subsidiaries, many of them with varying degrees of private ownership—to diversify dependence. The results suggest that relying on official distinctions between public and private may actually underestimate the true effect of state-dependent employment. The state’s retreat from these more indirect forms of engagement may thus also be a precondition to building a stronger base of support for democracy in the private sector. Although the crux of this book is in explaining variation in the middle classes’ regime preferences, I have compared the political preferences and behavior of the middle classes throughout with those of other societal actors. Doing so allowed me to reach more reliable conclusions about the conditions under which the middle classes’ regime preferences are (and are not) distinctive. Beyond simply finding weak support for democracy among the state middle class, the preceding chapters have shown that the state middle class is no more democratic than the working class. Hence, an expansion of the state middle class is unlikely to herald more democracy. To put it another way, these findings suggest a new set of predictions about democratization: namely, that it will be more likely to take place in countries with a growing middle class and a lower (or shrinking) state presence in the economy. Though not designed to parse the mechanisms behind working-class regime preferences, this study admits several explanations for the patterns found in preceding chapters. One interpretation, particular to the postcommunist region, is that the working class was for more than sixty years the target of Soviet ideological mobilization. As such, its democratic deficit may be a legacy of the communist period. Yet another interpretation is that when there is a high degree of state economic engagement, the working classes (whether state or private) judge their best prospects for upward mobility as regime dependent. Put positively, the working class may see the regime as an engine of advancement, where the public sector supplies the dominant pathway to the middle class. Put negatively, the working class may observe the value of loyalty and the price of opposition in shaping life chances, and choose the path of least resistance. Another explanation, along the lines of elitecompetition theory, is that demands for democracy arise only among those

228

chapter 8

who already have something to lose. A fourth, in line with modernization theory, is that people are more likely to embrace democratic values only as they gain higher education and occupations that allow for greater autonomy. Among individuals with low levels of education and unskilled occupations, these theories predict weak support for democracy, in keeping with the evidence I present.

Contribution and Implications The argument and evidence in this project contribute to several important areas of scholarly inquiry. First, this project deepens our understanding of how state economic engagement influences public opinion and political behavior in autocratic settings. Not long ago, mainstream political science considered the study of political attitudes and behavior under authoritarian rule theoretically uninteresting, not to mention impractical. Most theories of authoritarianism have focused on elite actors and a limited selectorate; ordinary citizens often appear as undifferentiated masses (peasants, the proletariat, the poor, or middle class), to the extent that they appear at all. Today’s emphasis on regime hybridity has been accompanied by a growing recognition that authoritarianism is sustained not only by coercion but also by mass appeal. This literature has renewed interest in public opinion’s important, understudied influence on authoritarian regime dynamics—the “imperative for gaining and sustaining public support” in Hale’s (2015) words. This book contributes to a shift in the authoritarianism/hybrid regimes literature from authoritarian institutions to the actors in authoritarian systems, including social groups and their preferences. It takes seriously the logic of ordinary citizens’ preferences under autocracy and sheds light on why some accept, while others reject, democracy. Scholars of the advanced industrial economies have long understood that public-sector policies create their own constituencies (e.g., Pierson 1994). These constituencies are durable, since the benefits of public-sector policies are concentrated (among public employees and the direct recipients of government transfers), while their costs are dispersed. In the context of developed Europe, scholars recognize that this makes beneficiaries a powerful constituency for the status quo. This project offers new insight into a similar dynamic under authoritarianism. The evidence presented in the preceding pages shows that, under autocracy, state-sponsored development through

c o n c l u s i o n 229

the public sector also helps to create its own pro-status quo middle-class constituencies. This base of support gives authoritarian regimes resilience and contributes to their stability.

Mobilization Studying individual preferences is arguably important in its own right for understanding “pent-up demand” for democracy. But this book also finds that sector of employment and social class, as well as expressed support for democracy, predicts politically consequential protest behavior (see also Beissinger 2013; Smyth, Sobolev & Soboleva 2013a). The state-employed post-communist middle classes are not merely more fearful of expressing support for democracy. They are also less likely to take to the streets during episodes of anti-regime contention. Indeed, a closer look at middle-class formation in autocratic settings aids our understanding of potential protest coalitions. Given the increasingly mobilized nature of democratic transitions (Karatnycky & Ackerman 2005), this is an important contribution to the literatures on both protest mobilization and democratization. While many have assigned a “middle-class” character to mass urban uprisings against authoritarianism, this project has problematized the perspective that a growing middle class necessarily increases the likelihood of successful civic revolution. A close examination of protest coalitions in Russia suggested that cleavages within the middle class serve to weaken potential pro-democracy groups and undermine the middle class’s capacity for oppositional collective action. The analysis in chapters 4 and 6 clearly demonstrates that states have advantages in demobilizing middle classes that depend on public employment, leading to lower rates of anti-regime protest participation. Moreover, with their greater organizational skills and resources, the state middle classes are often called upon (for example, in Kazakhstan in 1999, in Ukraine in 2004– 2005, and in Russia in 2011–2012) to organize and participate in rallies for incumbents, as well as tabulate votes and falsify elections. Disaggregating the middle classes reveals that the state middle class may be an important swing group in the social coalition supporting nondemocratic incumbents, especially at times of mobilized contention. As theories of critical mass and informational cascades remind us, decisions to participate in risky collective action are interdependent and hinge on beliefs about how many others are likely to join (Kuran 1995; Lohmann 1994). This implies that

230

chapter 8

the demobilization of public-sector workers also discourages others, making it even less likely that protests will achieve critical mass.

Development and Democratization Another major area of scholarly interest to which this study contributes is the literature on development and democratization. Since Lipset’s (1959) pioneering work, empirical tests of the modernization hypothesis have only gained in methodological sophistication. However, the causal mechanisms that explain the link between development and democracy have remained the subject of vigorous debate. A long tradition going back to Aristotle, and emphasized in Lipset’s early work, points to the pivotal intervening role of the middle class. Surprisingly, this middle-class mechanism has been rigorously tested only by Easterly (2001), who provides cross-national evidence that a larger middle class is associated with more civil liberties and political stability. However, the design of Easterly’s research is macro-macro, leaving the individual-level mechanisms, those that help to specify why and under what conditions a large middle class may matter, unclear and unspecified. This study’s findings demonstrate that growth of the middle classes is likely to provide a social base of support for democratization only when development creates certain conditions. In particular, I have shown that where the state plays a limited role in the economy and the private sector is large, growth of the middle classes can create bottom-up pressure for democratization. In contrast to Moore (1966), this book’s evidence shows that not just the capital-owning bourgeoisie, but broader strata of white-collar managers and professionals in the private sector can furnish support for democratization. Indeed, those who were outside of state structures in Russia, Ukraine, and Kazakhstan were more likely to be carriers of certain democratic norms. The intensity and consistency of their preference for democratic institutions were greater. And, as we saw in Russia, they expressed greater concern about official corruption and had less confidence that the courts would protect their interests. The same incentives that helped to demobilize the state middle class contributed to grievances among private-sector professionals excluded from special privileges. Still, in a state where the middle class remains divided, as in Russia, such anti-regime grievances may remain marginal. Protests are unlikely to achieve the critical mass necessary to dislodge authoritarianism. All of this implies that an expanding middle class is much more likely to risk democratization when development reduces state dependency.

c o n c l u s i o n 231

These findings also aid in clarifying the appropriate micro-foundational assumptions for macro-level theories of democratization. A variety of evidence in the preceding pages reveals patterns that are puzzling from the perspective of redistributive theories, and, in fact, contradict these theories’ individual-level assumptions. In particular, these theories do not expect the middle class to be more democratic than the working class, since the working class has more to gain from redistributive taxation under democracy.1 Across repeated tests, however, I found that the private-sector middle class’s support for democracy was stronger than that of the working class. Both the fact that the middle class remains a minority in the countries studied and that my operational measure may hypothetically have captured some elites further cut against redistributive explanations. Similarly, these theories do not expect becoming more affluent in a country where the middle class remains small to be associated with increasing support for democracy. Yet in Ukraine, where this project examined change over time in political preferences, and where the middle class remains a minority with income above the median, I found that individuals whose wellbeing rose became more, not less, supportive of democracy if their economic opportunities were secured in the private sector. I also found again that the private-sector middle class was significantly more supportive of democracy than the working class—a particularly striking result during a period of political revolt when redistributive theories predict that the middle class would side with authoritarian elites. Finally, redistributive theories cannot easily explain why the state- and private-sector middle classes’ regime preferences would diverge in the way I have shown, given their similar location in the income distribution. These theories focus on aggregate relationships, and their micro-level machinery is silent on the source of such differences (Meltzer & Richard 1981). Chapter 3 showed, moreover, that the post-communist state and private-sector middle classes held very similar views on redistribution, yet differed over the desirability of democratization. By contrast, each of these findings is anticipated by a framework which assumes that high human and social capital enhance the middle classes’ labor market alternatives and 1. This prediction hinges on the assumption that the income of the middle class is higher than that of the working class. While I do not use income as my primary measure of class, expenditure and consumption data suggest that this assumption holds in the forgoing empirical analyses. Unsurprisingly, occupation and education proxy in the expected way for incomebased class.

232

chapter 8

prospects under democracy, but emphasizes that these prospects relative to current prospects depend importantly on an individual’s relationship to the state. These findings also raise questions for other prominent theories of democratization. While values-based modernization theory suggests that rising affluence, increased education, and occupational specialization will be associated with greater support for democracy, this literature largely ignores the intervening role of the state. As Ross (2001) put it succinctly, “the modernization effect does not work through the state” (337). The upshot is that these theories are overly rigid and homogenizing of the middle class. On the one hand, this study found support for modernization theory’s prediction that the middle class will be more democratic than the working class—among individuals employed in the private sector. On the other, the attitudes of the state middle class documented here are anomalous from the modernization perspective. Value change occurs, this evidence would suggest, conditional on institutional settings and material incentives. Although it possesses the prerequisites for democratic value formation, the state middle class is not more democratic. It is also more likely to disengage,2 or to use the resources that would allow it to participate effectively in democratic politics to participate in nondemocratic politics instead. This means more middle-class participation in activities like protest and voter mobilization, but on the side of the regime. Like redistributive theories, then, these theories are ill-suited to explain why qualitatively different forms of modernization have divergent effects on regime preferences, as seen in the preceding chapters. Confronted with deviant (or “off-the-line”) cases—those where modernization has failed to produce pressure for democratization—this literature points to Cold War politics and cultural explanations. However, these theories provide little guidance as to why their implied mechanism—through rising affluence, education, and occupational specialization—sometimes fails at the individual level. The evidence in these pages suggests that the variation that valuesbased theories assign to culture may, in part, be explained more concretely: as arising from modernization trajectories that limit economic autonomy from the state. In a similar vein, this study suggests a micro-foundational explanation for the finding that economic development benefits democracy only when the 2. On the deliberate disengagement of more educated citizens from politics in autocratic settings, see also Croke et al. (2016).

c o n c l u s i o n 233

level of state economic engagement is low (Tang & Woods 2014). Indeed, the cross-national association between development and democratization is weak in post-WWII samples, when state intervention in the economy peaked. While some scholars have looked to international factors to explain development’s varying effects, this study suggests that domestic factors contributing to weak societal demand for democracy may also have been significant. After all, membership in the communist bloc is co-linear with a set of economic institutions that would have, according to the evidence in this book, stifled demands for democracy. In short, this book’s findings demonstrate how state economic engagement limits the size of constituencies that prefer and pursue democracy. This book thus provides a very concrete, micro-level explanation for one of comparative politics’ enduring puzzles: why we often see development without democratization. My account emphasizes why the social bases of support for democratization are often absent, despite high levels of economic growth. Creating a middle class on its own is not enough. To enhance the prospects for democracy, development must create a middle class that is independent of the state. The distinctive contribution of this book has been to show concretely, at the individual level and for a broad set of post-communist cases, how state intervention in the economy can cultivate an autocratic middle class. These findings advance our understanding of the role of the middle classes in democratic transition and their participation in contentious politics. They contribute to newer literatures on the rise of state capitalism in contemporary autocracies. And they provide new insights into the micro-level consequences of state-directed development, a subject that has been studied primarily at the institutional and macro-sociological level. The micro-level perspective in this book complements those comparative historical analyses by showing that, where state economic engagement is extensive, potentially pivotal middle-class actors perceive their interests as aligned with authoritarian stability.

Beyond the Post-Communist Region How unique are these insights to the ex-communist countries? What is the scope and applicability of the theory? At the outset of this book, I introduced two broad scope conditions: growth of the middle class under autocratic state institutions and extensive state economic engagement in the context

234

chapter 8

of late industrialization. By these two criteria, this book’s framework would extend to Algeria, Venezuela, and Malaysia, or Brazil in 1980. In each of these cases government investment has equaled approximately half or more of total investment.3 It would not, however, extend to cases like Peru, where the government has traditionally played a limited role in the economy, or Chile, where even under Pinochet government investment in 1980 made up only about 30% of the total. In fact, since the fall of communism, public-sector size in the average post-Soviet country has converged with that of other developing economies. Today, the former Soviet states are not alone in their high rates of public employment. This is especially true if we compare them with other nondemocracies. Two decades of privatization, notwithstanding, the public sector still contributes a large share of GDP in many of the world’s most prominent autocracies. Data on the number, composition, and share of output supplied by SOEs and government investment as a share of total investment from the Fraser Institute confirm that the state’s role in the post-Soviet countries is not exceptional. In comparison with Venezuela, Angola, the Democratic Republic of Congo, Malaysia, Nigeria, Cambodia, China, and Egypt—to name just a few—the state, in fact, plays a more modest economic role in the post-Soviet region.4 The post-Soviet states are also not unique in playing a central role as employers and engines of upward mobility. Such conditions are, in fact, widespread in the world’s other most autocratic regions. In sub-Saharan Africa, for example, Birdsall (2015) notes that nascent middle classes are highly dependent on the government for their livelihoods and social mobility. This dependence arises “either directly in the case of civil servants, or indirectly in the case of employees of state-owned banks and other enterprises heavily dependent on government subsidies” (221). Citing data from the ILO, Birdsall (2007) notes that the average share of formal employment in the public sector in sub-Saharan Africa is actually higher than in the post-communist countries.

3. According to the Fraser Institute’s measure of government enterprises and investment, in Algeria this figure was 61.3%, in Venezuela it was 60.6%, and in Malaysia it was 45.2% (all in 2010). In Brazil in 1980, the corresponding figure was 49.2%. 4. According to the 2012 Economic Freedom of the World data on government enterprises and investment archived at http://efwdata.com/.

c o n c l u s i o n 235

Behar & Mok (2013) make a similar point regarding the high proportion of state-sector employment in the Middle East. Indeed, they find that the proportion of public employment is similar in the Middle East, North Africa, the Caucasus, Central Asia, and Central and Eastern Europe (11). High rates of public employment are common across developing regions. Cross-national survey evidence further suggests that public-sector careers are often attractive, compared to private-sector positions. State jobs frequently offer greater security, the opportunity to earn informal rents, and better wages than in the private sector. State-sector wage premiums tend on average to be higher in less-developed countries and working hours are often shorter (Sharunina 2013). In other words, public-sector jobs pay better than their private-sector equivalents. Where they do not, the state often turns a blind eye to bribe-taking and corruption, allowing its employees to supplement their official earnings. This has been the strategy of many post-communist autocrats and, certainly, public-sector corruption has been rampant in other regions. The World Bank reports that across the Middle East and North Africa, a majority of young people aspire to public-sector positions (cited in Behar & Mok 2013, 13). In Indonesia, civil service jobs are so desirable that they are bought and sold (Kristiansen & Ramli 2006). In Kazakhstan, Nurmagambetov & Azhmetov (2010) find that survey respondents at all levels of educational attainment saw state jobs as a source of greater opportunity than either working for a private company or starting one’s own business. Many educated Russians prefer public-sector positions. About 40 percent of public-sector employees, but only 26.7 percent in the private sector, have a college-level education. In Russia, the popularity of majoring in state administration has only increased over the post-Soviet period; it is nearly as popular today as majoring in economics or business. Even in China, where the formal size of the state has shrunk significantly over the past two decades, public-sector jobs remain prestigious. This is true especially among those entering the workforce with the highest levels of education. Lee & Zhang (2013) report that government employment has consistently been the most preferred career option among Chinese university graduates (1504). A government job in China “brings employment security, handsome salaries, and superior benefits (especially in terms of housing), in addition to family prestige and personal status” (1504). According to the Beijing Colleges Panel Survey of nearly 5,000 students enrolled at fifteen elite universities in China, government entities ranked among the most desirable

236 c h a p t e r 8

organizations to work for upon graduation. Only private-sector entities with foreign investment and state-owned enterprises were more popular among college graduates. All of this suggests that the state sector plays an important role in the labor market for educated, high-skilled workers beyond the post-communist region. According to Behar & Mok (2013), this role is extensive enough to crowd out private-sector employment options in many developing nondemocracies. In sum, then, the basic conditions for this book’s main argument—a large state sector that dominates the labor market for highskilled professions, limiting alternatives and providing an attractive avenue of social mobility—exist well beyond the post-communist region.

Resource States Over recent years, high commodities prices have contributed to an expanding middle class alongside strong economic growth in developing countries. There is little doubt that the economic and political conditions in resourceproducing states are conducive to the mechanism this book advances. Nontax revenues help to sustain high public spending and a large public sector. Such largess also translates into better benefits and higher wages for publicsector workers. Greater rewards mean that the state wields strong incentives. Larger budgets mean that resource states encounter fewer constraints to keep public enterprises lean and competitive. Focusing on the variety of ways that an expansion of the middle classes can be achieved helps to clarify why high-income oil-producing countries remain autocracies. The Gulf countries, with their massive public employment rolls, are a case in point. The ILO (2013) finds that: ... countries in the region have continued using public employment and working conditions in the public sector to boost incomes ..., in particular for young people. ... The large and well-paying public sector has put those nationals that were seeking employment in the private sector at a disadvantage. Often, the conditions are not comparable, either in terms of working conditions or in terms of flexibility of employment (84). Scholars of the resource curse, such as Ross (1999), have posited that the structure of production in resource states inhibits the formation of an autonomous workforce (see also Ross 2001, 336). This book dovetails nicely with that perspective, and provides clear evidence of how this mechanism

c o n c l u s i o n 237

Probability of being a ‘democrat’

0.30

Non-middle class

Middle class

0.25

Non-middle class

Middle class Private

Private

0.20 0.15

State Private

State

Private

State

State

0.10 0.05 0.00

(a) Resource states

(b) Non-resource states

figure 8.1. Pattern of democracy support similar across post-soviet resource and non-resource states. This figure compares the predicted probabilities of being a ‘democrat’ for middle-class and non-middle-class respondents by sector of employment in the post-communist resource (left panel) and non-resource states (right panel). The point estimates and 95% confidence intervals are given. The hash marks denote 90% confidence intervals. Data source: EBRD LiTS, 2006 (nondemocracies only).

operates through the public sector. Across a variety of tests, I found that public employment attenuates bottom-up pressures for democratization, especially among the middle class. Consistent with the evidence in this study, governments in the Gulf region have used the public sector to supply the upward mobility that binds growing middle classes to the regime. The more successful and well financed their efforts, the more likely they have been to crowd out private-sector alternatives (Behar & Mok 2013, 6). Considering the example of the Gulf states raises the question: Is the theory I have put forward a theory about resource states? The findings in chapter 6 from Ukraine, which gets no more than 15% of GDP from the exploitation of natural resources, suggest otherwise. Could Ukraine be an exception? The evidence indicates it is not. While resource states enjoy certain advantages in financing the public sector, the mechanism this book describes also applies in non-resource states. Figure 8.1 confirms that the pattern of preferences I found in the cross-national analysis does not depend on the presence of oil. Using the same crossnational EBRD survey data that I analyzed in chapter 3, I plot the main results

238 c h a p t e r 8

separately for the ex-Soviet resource states (left panel) and non-resource states (right panel). Among the former, I include Azerbaijan, Kazakhstan, and Russia.5 Each draws more than a third of GDP from resource-producing sectors. The pattern is strikingly similar in both resource and non-resource states, though the results are less precisely estimated for the three energy exporters. Across both types of economies, the private-sector middle class is somewhat more supportive of democracy than the working class, while the state middle class is not. Further, the results hint that resource states may enjoy an advantage in also buying off the private sector, though these differences are not statistically significant. This is consistent with many accounts of the advantages of possessing large non-tax revenues (see, e.g., Chaudhry 1989; Karl 1997). It also accords with this book’s findings from Kazakhstan, where support for democracy among the private-sector middle class was minimal. In sum, my framework applies to both resource and non-resource states. However, resource states can be expected, on average, to provide more effective inducements to a larger number of private-sector workers—maximizing satisfaction (i.e., the sense that autocracy delivers) and minimizing exit options beyond the state’s reach.

Nature of the State and State Institutions In thinking about the argument’s scope, several additional observations are in order about the nature of the state and how the argument is likely to operate under a broad set of state institutions. The first point is that some authoritarian regimes are more circumstantial than others. When a regime is long-standing and ideological, workplacebased socialization is likely to play a more significant role in shaping the preferences of the state middle classes than it did in these cases. Since the collapse of communism, none of the post-Soviet regimes has been ideological in nature. Indeed, it could be argued that the ideological significance of communism had already waned and been replaced by a more opportunistic logic of loyalty, based on careerism rather than ideology, by the late Soviet period (Pop-Eleches & Tucker 2014). In terms of longevity, the post-Soviet cases range from a high degree of regime continuity over the past two decades

5. Additionally including Uzbekistan does not change the results.

c o n c l u s i o n 239

(e.g., Kazakhstan, Uzbekistan, Tajikistan) to fairly frequent regime rotation (e.g., Ukraine). In more circumstantial and temporary regimes, working for the state poses less of a problem in terms of signaling loyalty to the opposition should it win democratizing elections. This dynamic likely helps to explain greater statesector support for democratic political forces in the Latin American countries that experienced frequent regime turnover. For example, in countries such as Chile, Argentina, and Brazil, which experienced multiple coups d’état and military dictatorships, middle-class professionals within the state were more likely to ally with other social forces for democratization. The second point is that some institutional configurations of the state sector give autocrats more leverage than others. As previously discussed, three institutional features lend themselves to a high degree of dependence among public-sector professionals: 1) the absence of merit-based recruitment, 2) lack of employment protections, and 3) tolerance of official corruption (i.e., the ability to earn informal rents by exploiting one’s professional position). All three of these factors are virtually universal in the post-Soviet cases examined in this book. Indeed, even where political opinion is prohibited by law as grounds for dismissal (as it is, for example, in Russia and Azerbaijan), actual practice regularly deviates. Institutional characteristics such as these are also widespread in other autocratic settings. However, in cases where public-sector institutions produce public rather than private goods and are insulated from political pressure, the middle class may be more politically independent. Low-quality bureaucracies are clearly more conducive than high-quality bureaucracies to the mechanisms of dependence that this book highlights.

Market Backlash and Nationalizations In recent years, there has been a pronounced backlash against the market. This backlash is exemplified by nationalizations in a diverse set of authoritarian regimes. After years of retrenching state enterprises, China is once again acquiring private companies in sectors from metals to manufacturing— forcing the private sector to recede as the state advances. Turkey under Erdogan has nationalized hundreds of companies in the name of rooting out corruption. In Bolivia under Morales, companies in the oil, natural gas, telecoms, mining, aerospace, banking, and manufacturing sectors were brought

240

chapter 8

under public ownership, with the government taking credit for the creation of nearly 700,000 new state jobs. These trends evince a return to ideas implicit in the developmental state tradition and an earlier era of industrial policy in which the state led. Indeed, trends toward greater state involvement in key industries continue apace not only in prominent autocracies. Rising state control of the economy has accompanied democratic backsliding in countries such as Hungary and Poland. Since Viktor Orbán came to power, Hungary has renationalized or otherwise participated in the transfer of ownership of scores of companies— in particular, across the energy, natural gas, and other utilities sectors. Schools and hospitals have been brought once again under the supervision of the central government. In Poland, the country’s ruling Law and Justice party has renationalized banks and power companies under its program of “repolonisation” and has set its sights on other industries. Predictably, increasing state ownership has provided the governments of Hungary and Poland key levers of economic influence, allowing them to consolidate power to the detriment of democracy. In both countries’ state sectors, managers and professionals have been purged and replaced by loyalists on a scale unseen since the communist era. The argument in this book has clear implications for this wave of market backlash and nationalizations. An enlarged state sector and the expansion of government monopolies gives autocrats (and would-be autocrats) a very powerful tool to induce loyalty among those who owe their economic opportunities and livelihood to the state. By creating the conditions for an autocratic middle class, the return of economic statism is likely to result for the time-being in more resilient authoritarianism.

Economic Crises and the Durability of Big States One further implication of this book’s findings is that autocrats who are forced to carry out cuts to public-sector benefits are likely to confront waning support from key middle-class constituencies. Where government jobs fail to provide employment security, decent salaries, and superior benefits, especially in comparison with private-sector alternatives, the result is a gradual hollowing out of regime influence. As their purchasing power deteriorated, African civil servants played a prominent role in their countries’ democratic movements of the early 1990s (Bratton & van de Walle 1997). This was also the

c o n c l u s i o n 241

case in Tunisia and Egypt, where government employees joined protests that toppled two of the region’s long-standing authoritarian regimes in 2010–2011. The alliance of public- and private-sector middle classes resulted in demonstrations that were truly mass (Beissinger, Jamal & Mazur 2015). Particularly among the Egyptian middle class, the loss of subsidies as the state retreated and increasing reliance of state workers on second salaries in the private sector led to the breakdown of selective incentives. The challenge, of course, is that high rates of public employment incur substantial fiscal costs, which may not be sustainable in the long run—all the more so if rent-seeking and lower worker productivity decrease the public sector’s efficiency. While state firms’ preferential access to finance can forestall the effects of declining productivity and corruption in the short-to-medium term, longer-term prospects are less certain. Haggard & Kaufman (1995) argue that economic crises diminish the capacity of authoritarian leaders to influence how they exit. The evidence in this book suggests one concrete channel through which prolonged crises may mobilize pivotal groups within the state middle class. In cases such as these, and when economic crises occur, the state may rapidly become a target. As economic opportunities decline, it is relatively more straightforward for state workers to blame the government for their deteriorating economic position. For private-sector workers, the causal logic is more attenuated. Thus, in cases characterized by a high degree of state economic engagement, revolutionary cascades may occur quickly. Indeed, Tang, Huhe & Zhou (2017) find that economic crises are most likely to produce democratic transition when state involvement in the economy is widespread. At the same time, however, low support for democracy among the state middle class cautions against the conclusion that such crises, even if they result in transitions, will necessarily produce stable democracies. Without a broad popular base of support for democracy, crises may merely lead to regime cycling. Of course, crises can also contribute to consolidation of the state’s economic position. Larger state-backed firms often weather crises better than private players. Crises frequently provide auspicious moments for government intervention. Popular backlash against the market supplies political cover for creeping nationalization. Recent consolidation, growing in part out of the 2008–2009 global financial crisis, has put two-thirds of Russia’s banking sector in state hands (Szakonyi 2019). State banks have at the same time diversified their holdings into different sectors of the economy, well beyond those

242

chapter 8

previously seen as strategic. In 2018, Putin’s so-called May decree made no mention of privatization, in contrast with his decree in 2012, which promised to roll back government involvement.6 By the logic of this book, these moves are likely to further limit popular demands for democratization—even if they introduce distortions that in the long run are unsustainable.

Conclusion Where this study leaves off, questions remain about the newest forms of state economic engagement and their impact on regime preferences. State capitalism’s contributions to middle-class formation call for ever more subtle delineations of dependence on the state. What is the most appropriate way to capture newer, more indirect forms of state control? Tsai (2005), for example, finds that the state employment backgrounds of Chinese private entrepreneurs affect their political orientations. In chapter 3, I showed similarly that having previously worked in the post-communist public sector had a lingering effect on regime preferences. At the same time, a growing generation of state capitalists lack any formal ties to the state. This is true of the post-communist region and of China, and in emerging markets more generally. The challenge of studying state capitalism is that it intentionally obscures the state’s involvement. When state influence is exercised through subsidized loans, minority equity investment, interlocking shell companies, and subsidiaries, states seek to hide the origin of financial flows. As they enter national statistics, these financial flows and the firms they support are frequently reified as private. The analyses of Kazakhstan in chapter 7 suggested that indirect state influence in strategic sectors has an effect on regime preferences that resembles more traditional forms of direct state employment. Where states dictate terms to formally private companies that supply economic opportunity, the logic of dependence that this study underscored may operate similarly. When states exercise overt control over the selection of a firm’s senior management and countermand the profit motive by blocking layoffs, that firm’s employees may well see themselves as better off under autocracy. What is clear from this study is that these newer, more indirect ways that autocrats co-opt the middle classes are one nexus for understanding how popular demands for democratization do and do not arise. At the end of the day, the 6. See “What Is the State’s Share in Russia’s Economy,” Russia Matters, June 26, 2019, https://www.russiamatters.org/analysis/what-states-share-russias-economy.

c o n c l u s i o n 243

legal difference between public and private may ultimately matter more to investors than ordinary citizens where the state’s economic influence remains pervasive. This research opens several avenues for future investigation into how and when autocrats target patronage to the middle class in state sectors. While the evidence in this book speaks primarily to the consequences of dependence as a structural feature of statist economies, future studies could extend this line of inquiry by examining the conditions that produce different patterns of redistributive public employment and exploring variation in the size of wage premia for the state-employed middle class.7 Preceding discussions suggest several testable propositions. First, where the state is large enough to crowd out private-sector alternatives, wage premia for the state-employed middle class are likely to be lower. Second, autocrats may be more inclined to target patronage to state sectors that can effectively coordinate pro-regime mobilization in proximity to elections and when faced with mobilized opposition. Such targeted distribution can be used to mobilize teachers and civil servants with ties to local communities, who can in turn efficiently mobilize others. With its lower ideological barriers to mobilization and greater organizational capacity, the state middle class is an attractive target. In conclusion, authoritarian states are themselves often the engine of an expanding middle class. In this light, many millions of citizens across the developing world achieve and sustain middle-class status and stability thanks to opportunities granted by the state. This study has shown that dependence on the state for economic opportunities and life chances weakens middle-class support for democracy and demobilizes regime opponents. This evidence suggests that where state- and quasi-state-owned enterprises or an expansive public sector are the principal avenues of social mobility into the middle class, a democratizing middle class is unlikely to form. These pages thus shed new light on how authoritarian states hold together the social coalitions that keep them in power, even as they grow.

7. While a large body of research studies public-sector wage differentials in democracies, far less is known about policy heterogeneity across the skill dimension (i.e., across educational and occupational groups) even in democratic settings (Gomes 2018; Christofides & Michael 2013; Mueller 1998; Poterba & Rueben 1993; Katz & Krueger 1991). Still, recent work in political economy (e.g., Murillo & Calvo 2019; Alesina, Baqir & Easterly 2000) provides an excellent starting point for such an investigation and extensions to public-sector wage policy and redistributive public employment in autocracies.

appendix i

Regression Results

Non-middle class

Probability of being a ‘democrat’

0.50

Private

0.45 Middle class

0.40 0.35 0.30

Middle class

State

Non-middle class

State

Private

0.25 0.20 0.15 0.10 0.05 0.00

figure a1. Main results after relaxing the definition of democracy support. The left panel of this figure shows the predicted probability of being a ‘democrat’ given that a respondent is middle class (squares) or non-middle class (circles), after relaxing the definition of democracy support. The second panel shows the predicted probability of being a ‘democrat’ for middle-class and non-middle-class respondents by sector of employment. The point estimates and 95% confidence intervals are given. The hash marks denote 90% confidence intervals. Data source: EBRD LiTS, 2006 (nondemocracies only).

245

246

appendix i

table a1. Main Results: Middle-Class Status, State Employment, and Democracy Support Dependent Variable: Democrat Nondemocracies Democracies Nondemocracies Nondemocracies Base Model Base Model by State Type Mechanism Test (1) (2) (3) (4) Middle class State employment

0.522∗∗∗ (0.132) −0.229∗∗∗ (0.080)

0.375∗∗∗ (0.072) −0.047 (0.046)

Civil servants/directly employed by the state State educators State health professionals Other state employees

0.520∗∗∗ (0.132) −0.344 (0.231) −0.325∗ (0.196) −0.361 (0.225) −0.189∗∗ (0.087)

Career share in state sector Years in state sector Current state employment Democracy Age/10 Age/10 squared Gender

0.371∗∗∗ (0.122) −0.047∗∗∗ (0.013) 0.219∗∗∗ (0.068)

0.043 (0.067) −0.012∗ (0.007) 0.160∗∗∗ (0.038)

−0.404∗∗ (0.176)

0.029 (0.098)

0.367∗∗∗ (0.122) −0.047∗∗∗ (0.013) 0.218∗∗∗ (0.069)

Armenia MC × state MC × civil servant MC × state educator MC × state health MC × other state MC × career share in state sector

−0.248 (0.387) −0.206 (0.273) −0.437 (0.473) −0.554∗∗ (0.229)

0.459∗∗∗ (0.130)

−0.245∗ (0.134) −0.004 (0.012) −0.138 (0.115) 0.396∗∗∗ (0.128) −0.048∗∗∗ (0.014) 0.199∗∗∗ (0.070) −3.352∗∗∗ (0.450)

−0.249 (0.187)

Full Sample (5) 0.469∗∗ (0.125)

−0.181∗ (0.081)

0.587∗∗∗ (0.055) 0.101∗ (0.057) −0.018∗∗∗ (0.006) 0.170∗∗∗ (0.033)

−0.368∗∗ (0.178)

r e g r e s s i o n r e s u lt s 247 table a1. (continued) Dependent Variable: Democrat Nondemocracies Democracies Nondemocracies Nondemocracies Base Model Base Model by State Type Mechanism Test (1) (2) (3) (4) MC × democracy State career share × democracy MC × state career share × democracy Constant Sample Observations Log likelihood Akaike inf. crit.

Full Sample (5)

−2.679∗∗∗ (0.257)

−1.410∗∗∗ (0.151)

−2.669∗∗∗ (0.257)

−2.585∗∗∗ (0.271)

−0.069 (0.141) 0.153∗ (0.093) 0.350∗ (0.203) −2.138∗∗∗ (0.132)

All nondem 8,715 −3,090.596 6,195.191

All dem 17,546 −8,657.106 17,328.404

All nondem 8,715 −3,089.425 6,204.849

All nondem 8,417 −2,881.371 5,782.721

Full sample 26,256 −11,750.700 23,523.410

Note: ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01. Logistic coefficients with standard errors in parentheses. Data source: EBRD LiTS, 2006.

table a2. Relative Risk Estimates (Case-Control Model with Contaminated Controls, Moscow Sample)

Middle class vs. Non-middle class State vs. non-state State middle class vs. non-state middle class

Est.

95% CI

2.31 0.62 0.74

[1.91, 2.83] [0.51, 0.74] [0.56, 0.98]

Source: The Levada Center and the Foundation for Public Opinion (FOM). Notes: This table reports the fractional increase in the risk of protest participation for several groups relative to the given baseline. The results are based on a case control with contaminated controls design, using a random sample of protesters and a random sample of the population of Moscow and Moscow region. The same results are presented in figure 4.2.

Russian speaker

East

West

Kyiv

Log of income

Middle class

State × upward mobility

Upward mobility

State

0.012 (0.011) 0.324∗∗∗ (0.061) −0.208∗∗ (0.076)

(1)

−0.073∗∗∗ (0.013)

0.006 (0.011) 0.262∗∗∗ (0.065) −0.219∗∗ (0.079) 0.049∗∗∗ (0.014) 0.027∗∗∗ (0.008) 0.131∗∗∗ (0.023) 0.097∗∗∗ (0.016)

Orange Side OLS (2)

−0.074∗∗∗ (0.013)

0.009 (0.011) 0.264∗∗∗ (0.065) −0.222∗∗ (0.078) 0.047∗∗ (0.014) 0.027∗∗∗ (0.008) 0.134∗∗∗ (0.023) 0.092∗∗∗ (0.016)

(3)

table a3. Regression Results Predicting Participation in the Orange Revolution

0.034∗∗∗ (0.007) 0.009 (0.007)

−0.0004 (0.007) 0.004 (0.004) 0.001 (0.012)

0.019∗∗∗ (0.006) −0.012 (0.020)

Blue Side OLS (4)

−1.858∗∗∗ (0.324)

0.080 (0.238) 1.702∗ (0.745) −1.306 (0.908) 0.737∗∗ (0.243) 0.564∗∗ (0.176) 1.839∗∗∗ (0.325) 1.270∗∗∗ (0.247)

Orange Side Logistic (5) (6) 0.214 (0.203) 2.403∗∗∗ (0.543) −1.229 (0.742)

Dependent Variable:

2.060∗∗∗ (0.609) 1.245 (0.821)

0.183 (0.501) 0.316 (0.363) 0.601 (1.150)

2.189∗∗ (0.751) −14.344 (914.690)

Blue Side Logistic (7)

2,052 0.017

0.051∗∗∗ (0.009)

Note: ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001.

Observations R2 Log likelihood

Constant

Risk tolerance

Age

Male

Ukrainian-Russian speaker

1,922 0.102

−0.107∗ (0.054)

−0.021 (0.018) −0.003 (0.011) −0.003 (0.005)

1,890 0.110

−0.020 (0.019) −0.010 (0.011) −0.001 (0.005) 0.008∗∗∗ (0.002) −0.137∗ (0.054) 1,922 0.032 −472.055

−0.012 (0.028)

0.0003 (0.009) 0.002 (0.006) −0.006∗ (0.002)

1,922 −115.653

−358.956

−6.550∗∗∗ (1.239)

2,052

−2.913∗∗∗ (0.167)

−0.231 (0.382) −0.051 (0.212) −0.065 (0.095)

1,922

−8.544∗∗ (2.635)

0.685 (1.046) 0.115 (0.398) −0.424∗ (0.178)

250 a p p e n d i x i table a4. Regression Results Predicting Switching to the Party of Regions Dependent Variable: Switched to POR in 2007 OLS (1) State Middle class

(2)

0.048∗∗ 0.043∗∗ (0.015) (0.015) −0.068∗∗∗ −0.045∗ (0.019) (0.020)

Upward mobility Log of income Center East South West Russian speaker Ukrainian-Russian speaker Catholic Orthodox Male Age Constant Observations R2 Log likelihood

−0.017 (0.010) −0.191∗∗∗ (0.043) 0.126∗∗ (0.042) −0.022 (0.042) −0.242∗∗∗ (0.047) 0.072∗∗∗ (0.021) 0.076∗∗ (0.026) −0.003 (0.034) 0.022 (0.018) −0.012 (0.015) 0.019∗∗ (0.006) 0.275∗∗∗ 0.189∗∗∗ (0.011) (0.081) 2,993 0.007

2,621 0.168

Logistic (3)

(4)

(5)

0.038 (0.020) −0.114∗∗∗ (0.022) −0.075 (0.055) −0.002 (0.014) −0.264∗∗∗ (0.055) 0.106∗ (0.054) −0.036 (0.055) −0.317∗∗∗ (0.061) 0.095∗∗∗ (0.026) 0.073∗ (0.034) 0.011 (0.045) 0.033 (0.023) −0.003 (0.019) 0.013 (0.009) 0.294∗∗ (0.111)

0.297∗∗ (0.093) −0.453∗∗∗ (0.130)

0.321∗∗ (0.110) −0.357∗ (0.152)

−1.461∗∗∗ (0.071)

−0.126 (0.072) −1.224∗∗∗ (0.267) 0.529∗ (0.244) −0.086 (0.249) −2.440∗∗∗ (0.404) 0.515∗∗∗ (0.148) 0.547∗∗ (0.179) −1.047 (0.638) 0.144 (0.119) −0.094 (0.109) 0.139∗∗ (0.044) −1.093 (0.563)

0.252 (0.135) −0.841∗∗∗ (0.163) −0.482 (0.382) −0.003 (0.097) −1.452∗∗∗ (0.311) 0.457 (0.285) −0.084 (0.293) −2.721∗∗∗ (0.484) 0.609∗∗∗ (0.175) 0.511∗ (0.220) −0.690 (0.662) 0.210 (0.143) −0.030 (0.131) 0.098 (0.058) −1.286 (0.737)

1,706 0.223

2,993

2,621

1,706

−1,498.968 −1,097.666

(6)

−759.273

Note: The sample is respondents who did not support the POR before the Orange Revolution. ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001

bibliography

Abelson, Robert P., & John W. Tukey. 1970. Efficient Conversion of Non-Metric Information into Metric Information. In The Quantitative Analysis of Social Problems, ed. Edward R. Tufte. Reading, MA: Addison-Wesley pp. 407–417. Abramov, Alexander, Alexander Radygin & Maria Chernova. 2017. “State-Owned Enterprises in the Russian Market: Ownership Structure and Their Role in the Economy.” Russian Journal of Economics 3:1–23. Abramson, Paul, & Ronald Inglehart. 1995. Value Change in Global Perspective. Ann Arbor, MI: The University of Michigan Press. Acemoglu, Daron, & James A. Robinson. 2006. Economic Origins of Dictatorship and Democracy. New York: Cambridge University Press. Acemoglu, Daron, Simon Johnson, James A. Robinson & Pierre Yared. 2005. “From Education to Democracy?” The American Economic Review 95(2):44–49. Achen, Christopher H. 1975. “Mass Political Attitudes and the Survey Response.” American Political Science Review 69(4):1218–1231. Achen, Christopher H. 1983. Toward Theories of Data: The State of Political Methodology. In Political Science: The State of the Discipline, ed. Ada W. Finifter. Washington, DC: The American Political Science Association pp. 69–93. Achen, Christopher H. 1992. “Social Psychology, Demographic Variables, and Linear Regression: Breaking the Iron Triangle in Voting Research.” Political Behavior 14(3):195–211. Alesina, Alberto, Reza Baqir & William Easterly. 2000. “Redistributive Public Employment.” Journal of Urban Economics 48(2):219–241. Almond, Gabriel, & Sidney Verba. 1963. The Civic Culture: Political Attitudes and Democracy in Five Nations. Boston: Little, Brown. Amin, Magdi, Ragui Assaad, Nazar al Baharna, Kemal Dervis, Raj M. Desai, Navtej S. Dhillon, Ahmed Galal, Hafez Ghanem, Carol Graham & Daniel Kaufmann. 2012. After the Spring: Economic Transitions in the Arab World. New York: Oxford University Press. Anderson, John. 2007. “Putin and the Russian Orthodox Church: Asymmetric Symphonia?” Journal of International Affairs 61(1):185–201. Angrist, Joshua D., & Jörn-Steffen Pischke. 2009. Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton, NJ: Princeton University Press. Angrist, Michele Penner. 2005. Party Systems and Regime Formation. In Authoritarianism in the Middle East: Regimes and Resistance, ed. Marsha Pripstein Posusney & Michele Penner Angrist. Boulder, CO: Lynne Rienner Publishers pp. 119–138.

251

252 b i b l i o g r a p h y Ansell, Ben, & David Samuels. 2008. “Inequality and Democratization.” Paper presented at the 2008 Meeting of the Midwest Political Science Association. Ansell, Ben W., & David J. Samuels. 2014. Inequality and Democratization: An Elite-Competition Approach. New York: Cambridge University Press. Aristotle. 1999. Politics. Kitchener: Batoche Books. Translated by Benjamin Jowett. Arriola, Leonardo R. 2013. Multiethnic Coalitions in Africa: Business Financing of Opposition Election Campaigns. New York: Cambridge University Press. Asylkhanova, Gulnar Zhanabekovna. 2010. Stratifikatsionnaia sistema sovremennogo obshchestva Kazakhstana. PhD thesis, Pavlodar State University. Pavlodar, Kazakhstan. Avraamova, E.M. 2008. “Srednii’ klass epokhi Putina.” Obshchestvennye nauki i sovremennost’ 1:28–36. Azirkhanov, B. 2002. “Osnovnye problemy i podkhody v izuchenii sovremennogo sostoianiia protsessov formirovaniia srednego klassa Kazakhstana.” Evrazijskoe soobshchestvo 1(37): 24–28. Banerjee, Abhijit V., & Esther Duflo. 2008. “What is Middle Class about the Middle Classes around the World?” Journal of Economic Perspectives 22(2):3–28. Banerjee, Ritwik, Tushi Baul & Tanya Rosenblat. 2015. “On Self Selection of the Corrupt into the Public Sector.” Economics Letters 127:43–46. Barfort, Sebastian, Nikolaj Harmon, Frederik Hjorth & Asmus Leth Olsen. 2015. “Dishonesty and Selection into Public Service in Denmark: Who Runs the World’s Least Corrupt Public Sector?” Working Paper. University of Copenhagen, Department of Economics. http:// www.econ.ku.dk/english/research/publications/wp/dp_2015/1512.pdf. Barro, Robert J. 1999. “Determinants of Democracy.” Journal of Political Economy 107(S6): S158–183. Bayat, Asef. 2009. Life as Politics: How Ordinary People Change the Middle East. Stanford, CA: Stanford University Press. Behar, Alberto, & Junghwan Mok. 2013. “Does Public-Sector Employment Fully Crowd Out Private-Sector Employment?” IMF Working Paper WP/13/146. June. http://www.imf.org /external/pubs/ft/wp/2013/wp13146.pdf. Behar, Alberto, & Junghwan Mok. 2015. “Does Public Employment Reduce Unemployment?” Topics in Middle Eastern and African Economies 17(2):111–121. Beissinger, Mark. 2013. “The Semblance of Democratic Revolution: Coalitions in Ukraine’s Orange Revolution.” American Political Science Review 107(3):574–592. Beissinger, Mark, Amaney Jamal & Kevin Mazur. 2015. “Explaining Divergent Revolutionary Coalitions: Regime Strategies and the Structuring of Participation in the Tunisian and Egyptian Revolutions.” Comparative Politics 48(1):1–24. Beissinger, Mark, & Stephen Kotkin, eds. 2014. Historical Legacies of Communism in Russia and Eastern Europe. New York: Cambridge University Press. Bellin, Eva R. 2000. “Contingent Democrats: Industrialists, Labor, and Democratization in Late-Developing Countries.” World Politics 52(2):175–205. Bellin, Eva R. 2002. Stalled Democracy: Capital, Labor, and the Paradox of State-Sponsored Development. Ithaca: Cornell University Press. Bian, Yanjie. 1994. “Guanxi and the Allocation of Urban Jobs in China.” The China Quarterly 140:971–999.

b i b l i o g r a p h y 253 Birdsall, Nancy. 2007. “Do No Harm: Aid, Weak Institutions, and the Missing Middle in Africa.” Center for Global Development. Working Paper 113. March. http://www.cgdev.org/sites /default/files/13115_file_Aid_Institutions_Africa.pdf. Birdsall, Nancy. 2015. “Does the Rise of the Middle Class Lock in Good Government in the Developing World?” European Journal of Development Research 27(2):217–229. Birdsall, Nancy, Carol Graham & Stefano Pettinato. 2000. “Stuck in the Tunnel: Is Globalization Muddling the Middle Class.” Center on Social and Economic Dynamics Working Paper 14. Brookings Institution, Washington DC. Bissenova, Alima. 2013. “The Expectations of Paternalism: the Boom, the Bust, and the Mediating State in an Authoritarian Context.” Paper Presented at the annual meeting of the American Association of Anthropology. Chicago. Boix, Carles. 2003. Democracy and Redistribution. New York: Cambridge University Press. Boix, Carles. 2011. “Democracy, Development, and the International System.” American Political Science Review 105(4):809–828. Borisov, Gleb V., & Christopher A. Pissarides. 2016. “The Intergenerational Transmission of Human Capital and Earnings in Contemporary Russia.” IZA Discussion Paper No. 10300. http://ftp.iza.org/dp10300.pdf. Bortolotti, Bernardo, & Mara Faccio. 2009. “Government Control of Privatized Firms.” The Review of Financial Studies 22(8):2907–2939. Box-Steffensmeier, Janet M., Henry E. Brady & David Collier, eds. 2008. The Oxford Handbook of Political Methodology. Oxford: Oxford University Press. Bratton, Michael, & Nicolas van de Walle. 1997. Democratic Experiments in Africa: Regime Transitions in a Comparative Perspective. New York: Cambridge University Press. Breen, Richard. 2005. Foundations of Class Analysis in the Weberian Tradition. In Alternative Foundations of Class Analysis, ed. Erik Olin Wright. Cambridge: Cambridge University Press pp. 31–50. Bremmer, Ian. 2010. The End of the Free Market: Who Wins the War between States and Corporations? New York: Penguin. Brooks, Clem. 1994. “Class Consciousness and Politics in Comparative Perspective.” Social Science Research 23:167–195. Brown, David, & David Martin Jones. 1995. Democratization and the Myth of the Liberalizing Middle Classes. In Towards Illiberal Democracy in Pacific Asia, ed. Daniel A. Bell et al. Basingstoke, UK: Macmillan pp. 78–106. Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. New York: Cambridge University Press. Bunce, Valerie J., & Sharon L. Wolchik. 2011. Defeating Authoritarian Leaders in Postcommunist Countries. New York: Cambridge University Press. Calvo, Ernesto, & Maria Victoria Murillo. 2004. “Who Delivers? Partisan Clients in the Argentine Electoral Market.” American Journal of Political Science 48(4):742–757. Camp, Roderic Ai. 2001. Democracy through Latin American Lenses: An Appraisal. In Citizen Views of Democracy in Latin America, ed. Roderic Ai Camp. Pittsburgh, PA: University of Pittsburgh Press pp. 3–23. Campbell, David E. 2013. “Social Networks and Political Participation.” Annual Review of Political Science 16:33–48.

254 b i b l i o g r a p h y Capoccia, Giovanni, & Daniel Ziblatt. 2010. “The Historical Turn in Democratization Studies: A New Research Agenda for Europe and Beyond.” Comparative Political Studies 43(8/9):931–968. Chaisty, Paul, & Stephen Whitefield. 2013. “Forward to Democracy or Back to Authoritarianism? The Attitudinal Bases of Mass Support for the Russian Election Protests of 2011–2012.” Post-Soviet Affairs 29(5):387–403. Chaudhry, Kiren Aziz. 1989. “The Price of Wealth: Business and State in Labor Remittance and Oil Economies.” International Organization 43(1):101–145. Cheeseman, Nic. 2015. “ ‘No Bourgeoisie, No Democracy’? The Political Attitudes of the Kenyan Middle Class.” Journal of International Development 27(5):573–715. Chen, An. 2002. “Capitalist Development, Entrepreneurial Class and Democratization in China.” Political Science Quarterly 117(3):401–422. Chen, Jie. 2013. A Middle Class without Democracy: Economic Growth and the Prospects for Democratization in China. New York: Oxford University Press. Chen, Jie, & Chunlong Lu. 2011. “Democratization and the Middle Class in China: The Middle Class’s Attitudes toward Democracy.” Political Research Quarterly 64(3):705–719. Christofides, Louis N., & Maria Michael. 2013. “Exploring the Public-Private Sector Wage Gap in European Countries.” IZA Journal of European Labor Studies 2(15):1–53. Chubb, Judith. 1982. Patronage, Power and Poverty in Southern Italy: A Tale of Two Cities. Cambridge: Cambridge University Press. Collier, Ruth B. 1999. Paths toward Democracy: The Working Class and Elites in Western Europe and South America. New York: Cambridge University Press. Collier, Ruth B., & David Collier. 2002. Shaping the Political Arena: Critical Junctures, the Labor Movement, and Regime Dynamics in Latin America. Notre Dame, IN: University of Notre Dame Press. Converse, Philip E. 1964. The Nature of Belief Systems in Mass Publics. In Ideology and Discontent, ed. David E. Apter. New York: The Free Press pp. 206–261. Coppedge, Michael. 2012. Democratization and Research Methods. Cambridge: Cambridge University Press. Copsey, Nathaniel. 2006. “Europe and the Ukrainian Parliamentary Elections of 2006.” European Parties Elections and Referendums Network (EPERN) Election Briefing No. 24. http://www.sussex.ac.uk/sei/documents/epern-election-briefing-no-24.pdf. Croke, Kevin, Guy Grossman, Horacio A. Larreguy & John Marshall. 2016. “Deliberate Disengagement: How Education Can Decrease Political Participation in Electoral Authoritarian Regimes.” American Political Science Review 110(3):579–600. Dahl, Robert A. 1971. Polyarchy: Participation and Opposition. New Haven: Yale University Press. Dahrendorf, Ralf. 1967. Society and Democracy in Germany. New York: Norton. Dal Bó, Ernesto, Frederico Finan & Martin A. Rossí. 2013. “Strengthening State Capabilities: The Role of Financial Incentives in the Call to Public Service.” The Quarterly Journal of Economics 128(3):1169–1218. Dalton, Russell J., Doh C. Shin & Willy Jou. 2008. How People Understand Democracy. In How People View Democracy. Baltimore: Johns Hopkins University Press pp. 1–30. Daly, John C. K. 2008. “Kazakhstan’s Emerging Middle Class.” Central Asia-Caucasus Institute Silk Road Studies Program, Washington DC, March 2008.

b i b l i o g r a p h y 255 D’Anieri, Paul. 2007. Understanding Ukrainian Politics: Power, Politics, and Institutional Design. London: Routledge. Darden, Keith. 2008. “The Integrity of Corrupt States: Graft as an Informal State Institution.” Politics and Society 36(1):35–60. Diamond, Larry. 1987. “Class Formation in the Swollen African State.” Journal of Modern African Studies 25(4):567–596. Diamond, Larry. 1992. “Economic Development and Democracy Reconsidered.” American Behavioral Scientist 35(4/5):450–499. Diamond, Larry, & Marc F. Plattner, eds. 2008. How People View Democracy. Baltimore: Johns Hopkins University Press. Diuk, Nadia. 2014. “Finding Ukraine.” Journal of Democracy 25(3):83–89. Dixit, Avinash, & John Londregan. 1996. “The Determinants of Success of Special Interest in Redistributive Politics.” Journal of Politics 58:1132–1155. Dong, Hsiang-Kai Dennis. 2014. “Individual Risk Preference and Sector Choice: Are RiskAverse Individuals More Likely to Choose Careers in the Public Sector?” Administration & Society 49(8):1121–1142. Duch, Raymond M. 1993. “Tolerating Economic Reform: Popular Support for Transition to a Free Market in the Former Soviet Union.” American Political Science Review 87(3): 590–608. Easterly, William. 2001. “The Middle Class Consensus and Economic Development.” Journal of Economic Growth 6(4):317–335. Ekiert, Grzegorz. 2010. The End of Communism in Central and Eastern Europe: The Last Middle-Class Revolution? In Political Power and Social Theory, Volume 21, ed. Julian Go. Bingley, UK: Emerald Group Publishing Limited pp. 99–123. Elster, Jon. 2015. Explaining Social Behavior: More Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press. Erikson, Robert, & John H. Goldthorpe. 1992. The Constant Flux. A Study of Class Mobility in Industrial Societies. Oxford: Oxford University Press. European Bank for Reconstruction and Development, World Bank. Life in Transition Survey (LiTS) 2006, Ref. ECA_2006_LITS_v01_M. Dataset downloaded from https://www. ebrd.com/what-we-do/economic-research-and-data/data/lits.html on January 28, 2013. Feldman, Stanley. 1989. “Measuring Issue Preferences: The Problem of Response Instability.” Political Analysis 1(1):25–60. Finifter, Ada W., & Ellen Mickiewicz. 1992. “Redefining the Political System of the USSR: Mass Support for Political Change.” American Political Science Review 86(4):857–874. Finkel, Steven E. 1995. Causal Analysis with Panel Data. London: Sage. Fish, M. Steven, & Omar Choudhry. 2007. “Democratization and Economic Liberalization in the Postcommunist World.” Comparative Political Studies 40(3):254–282. FOM. 2009. “The Labor Market: The Position of Workers in the Labor Market.” The Foundation for Public Opinion. Dominanty No. 38. http://bd.fom.ru/pdf/d38trud.pdf. Forrat, Natalia. 2018. “Shock-Resistant Authoritarianism: Schoolteachers and Infrastructural Capacity in Putin’s Russia.” Comparative Politics 50(3):417–449. Fox, Richard L., & Jennifer L. Lawless. 2005. “To Run or Not to Run for Office: Explaining Nascent Political Ambition. To Run or Not to Run for Office: Explaining Nascent Political Ambition.” American Journal of Political Science 49(3):642–659.

256

bibliography

Franzese, Robert J. 2007. Multi-Causality, Context-Conditionality, and Endogeneity for Empirical Evaluation of Positive Theory in Comparative Politics. In Oxford Handbook of Comparative Politics, ed. Carles Boix & Susan Stokes. Oxford: Oxford University Press pp. 27–72. Freedom House. 2003. Freedom in the World 2004—Kazakhstan. https://www.refworld.org /docid/473c549c23.html. Freedom House. 2013. Nations in Transit 2013: Democratization in East Central Europe and Eurasia. Lanham: Rowman and Littlefield. Frye, Timothy. 2003. “Markets, Democracy, and New Private Business in Russia.” Post-Soviet Affairs 19(1):25–45. Frye, Timothy. 2006. “Ownership, Voting, and Job Creation in Russia.” European Journal of Political Economy 22(2):452–471. Frye, Timothy, Ora John Reuter & David Szakonyi. 2014. “Political Machines at Work: Voter Mobilization and Electoral Subversion in the Workplace.” World Politics 66(2):195–228. Frye, Timothy, Ora John Reuter & David Szakonyi. 2019. “Hitting Them with Carrots: Voter Intimidation and Vote Buying in Russia.” British Journal of Political Science 49(3):857–881. Fukuyama, Francis. 2014. Political Order and Political Decay: From the Industrial Revolution to the Globalization of Democracy. New York: Farrar, Straus, and Giroux. Gabris, Gerald T., & Gloria Simo. 1995. “Public Sector Motivation as an Independent Variable Affecting Career Decisions.” Public Personnel Management 24(1):33–51. Gandhi, Jennifer, & Adam Przeworski. 2007. “Authoritarian Institutions and the Survival of Autocrats.” Comparative Political Studies 40(11):1279–1301. Gans-Morse, Jordan, Alexander S. Kalgin, Andrei V. Klimenko & Andrei A. Yakovlev. 2017. “Motivations for Public Service in Corrupt States: Evidence from Post-Soviet Russia.” National Research University Higher School of Economics Working Paper No. WP BRP 13/PSP/2017. http://dx.doi.org/10.2139/ssrn.3072181. Gerber, Alan S., & Gregory A. Huber. 2009. “Partisanship and Economic Behavior: Do Partisan Differences in Economic Forecasts Predict Real Economic Behavior?” American Political Science Review 103(3):407–426. Gerber, Theodore P., & David R. Schaefer. 2004. “Horizontal Stratification of Higher Education in Russia: Trends, Gender Differences, and Labor Market Outcomes.” Sociology of Education 77(1):32–59. Gerschenkron, Alexander. 1962. Economic Backwardness in Historical Perspective: A Book of Essays. Cambridge, MA: Belknap Press of Harvard University Press. Gerschewski, Johannes. 2013. “The Three Pillars of Stability: Legitimation, Repression and Co-optation in Autocratic Regimes.” Democratization 20(1):13–38. Gervasoni, Carlos. 2010. “A Rentier Theory of Subnational Regimes: Fiscal Federalism, Democracy, and Authoritarianism in the Argentine Provinces.” World Politics 62(2): 302–340. Gibson, James L. 1995. The Resilience of Mass Support for Democratic Institutions and Processes in Nascent Russian and Ukrainian Democracies. In Political Culture and Civil Society in Russia and the New States of Eurasia, ed. Vladimir Tismaneanu. Armonk, NY: M. E. Sharpe pp. 53–111.

b i b l i o g r a p h y 257 Gibson, James L. 1996. “Political and Economic Markets: Changes in the Connections Between Attitudes toward Political Democracy and a Market Economy within the Mass Culture of Russia and Ukraine.” Journal of Politics 58(4):954–984. Gibson, James L., Raymond M. Duch & Kent L. Tedin. 1992. “Democratic Values and the Transformation of the Soviet Union.” Journal of Politics 54(2):329–371. Gimpelson, Vladimir, & Daniel Treisman. 2002. “Fiscal Games and Public Employment: A Theory with Evidence from Russia.” World Politics 54(2):145–183. Goldthorpe, John H., & Gordon Marshall. 1992. “The Promising Future of Class Analysis: A Response to Recent Crititiques.” Sociology 26(3):381–400. Gomes, Pedro. 2018. “Heterogeneity and the Public Sector Wage Policy.” International Economic Review 59(3):1469–1489. Gontmakher, Yevgeny, & Cameron Ross. 2015. “The Middle Class and Democratisation in Russia.” Europe-Asia Studies 67(2):269–284. Gorodnichenko, Yuriy, & Klara Sabirianova Peter. 2007. “Public Sector Pay and Corruption: Measuring Bribery from Micro Data.” Journal of Public Economics 91(5):963–991. Green, Donald, Bradley Palmquist & Eric Schickler. 2002. Partisan Hearts and Minds. New Haven, CT: Yale University Press. Greene, Kenneth F. 2007. Why Dominant Parties Lose: Mexico’s Democratization in Comparative Perspective. New York: Cambridge University Press. Greene, Kenneth F. 2010. “The Political Economy of Authoritarian Single-Party Dominance.” Comparative Political Studies 43(7):807–834. Grzymala-Busse, Anna. 2007. Rebuilding Leviathan: Party Competition and State Exploitation in Post-Communist Democracies. New York: Cambridge University Press. Gurr, Ted Robert. 1970. Why Men Rebel. Princeton: Princeton University Press. Haggard, Stephan, & Robert R. Kaufman. 1995. The Political Economy of Democratic Transitions. Princeton, NJ: Princeton University Press. Haggard, Stephan, & Robert R. Kaufman. 2016. Dictators and Democrats: Masses, Elites, and Regime Change. Princeton, NJ: Princeton University Press. Hakhverdian, Armen, & Quinton Mayne. 2012. “Institutional Trust, Education, and Corruption: A Micro-Macro Interactive Approach.” Journal of Politics 74(3):739–750. Hale, Henry E. 2007. Correlates of Clientelism: Political Economy, Politicized Ethnicity, and Post-communist Transition. In Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. Herbert Kitschelt & Steven I. Wilkinson. Cambridge: Cambridge University Press pp. 227–250. Hale, Henry E. 2015. Patronal Politics: Eurasian Regime Dynamics in Comparative Perspective. New York: Cambridge University Press. Handley, Antoinette. 2015. “Varieties of Capitalists? The Middle-Class, Private Sector and Economic Outcomes in Africa.” Journal of International Development 27(5):609–627. Hanmer, Michael J., & Kerem O. Kalkan. 2013. “Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variable Models.” American Journal of Political Science 57(1):263–277. Hanna, Rema, & Sheng-Yi Wang. 2017. “Dishonesty and Selection into Public Service: Evidence from India.” American Economic Journal: Economic Policy 9(3):262–290.

258

bibliography

Hedstrom, Peter, & Richard Swedberg. 1998. Social Mechanisms: An Introductory Essay. In Social Mechanisms: An Analytical Approach to Social Theory, ed. Peter Hedstrom & Richard Swedberg. New York: Cambridge University Press. Hellman, Joel. 1998. “Winner Take All: The Politics of Partial Reform in Postcommunist Transitions.” World Politics 50(2):203–234. Hertel-Fernandez, Alexander. 2017. “American Employers as Political Machines.” The Journal of Politics 79(1):105–117. Hertog, Steffen. 2013. “State and Private Sector in the GCC after the Arab Uprisings.” Journal of Arabian Studies 3(2):174–195. Hicken, Allen. 2011. “Clientelism.” Annual Review of Political Science 14:289–310. Hough, Jerry F. 1994. “The Russian Election of 1993: Public Attitudes toward Economic Reform and Democratization.” Post-Soviet Affairs 10:1–37. Houston, David J. 2014. “Public Service Motivation in the Post-Communist State.” Public Administration 92(4):843–860. Huber, Evelyne, & John D. Stephens. 1999. “The Bourgeoisie and Democracy: Historical and Contemporary Perspectives.” Social Research 66(3):759–788. Human Rights Watch. 1999. “Kazakhstan: Freedom of the Media and Political Freedoms in the Prelude to the 1999 Elections.” http://www.hrw.org/reports/1999/kazakhstan /Kaz1099b-05.htm. Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Century. Norman: University of Oklahoma Press. Ileuova, G. 2012. Osobennosti sotsial’noj mobil’nosti v usloviiakh sovremennogo Kazakhstana. In Sovremennye realii politologicheskoj nauki v kontekste sotsial’noj modernizatsii Kazakhstana. Astana, Kazakhstan. ILO. 2013. “Global Employment Trends 2013.” The International Labour Oranization. International Lobour Office. Geneva. http://www.ilo.org/wcmsp5/groups/public/---dgreports /---dcomm/---publ/documents/publication/wcms_202326.pdf. INDEM. 2005. “Corruption Process in Russia: Level, Structure, Trends.” http://www.indem. ru/en/Publicat/2005diag_engV.htm. Inglehart, Ronald. 1977. The Silent Revolution: Changing Values and Political Styles Among Western Publics. Princeton, NJ: Princeton University Press. Inglehart, Ronald, 1990. Culture Shift in Advanced Industrial Society. Princeton, NJ: Princeton University Press. Inglehart, Ronald, & Wayne Baker. 2000. “Modernization, Cultural Change, and the Persistence of Traditional Values.” American Sociological Review 65(1):19–51. Inkeles, Alex. 1969. “Participant Citizenship in Six Developing Countries.” American Political Science Review 63(4):1120–1141. International Trade Administration, U.S. Department of Commerce. 2017. “Ukraine Country Commercial Guide.” https://www.export.gov/article?id=Ukraine-State-Owned-Enter prises. Ivatova, L. M. 2010. Politicheskaia institutsionalizatsiia srednego klassa v kontekste povysheniia imidzha Respubliki Kazakhstan. In Fenomen sotsial’noj inzhenerii: opyt i rekonstruktsiia budushchego: Materialy mezhdunarogoj nauchno-prakticheskoj konferentsii, 21–22 maia 2010 g., ed. M. M. Tazhin. Almaty, Kazakhstan: Fond Pervogo Prezidenta Respubliki Kazakhstan pp. 160–165.

b i b l i o g r a p h y 259 Jackson, John E., Jacek Klich & Krystyna Poznanska. 2005. The Political Economy of Poland’s Transition: New Firms and Reform Governments. New York: Cambridge University Press. Johnson, Dale L. 1985. Class and Social Development: Toward a Comparative and Historical Social Science. In Middle Classes in Dependent Countries, ed. Dale L. Johnson. Beverly Hills, CA: Sage pp. 13–41. Jones, David M. 1998. “Democratization, Civil Society, and Illiberal Middle Class Culture in Pacific Asia.” Comparative Politics 30(2):147–169. Junisbai, Barbara. 2012. “Improbable but Potentially Pivotal Oppositions: Privatization, Capitalists, and Political Contestation in the Post-Soviet Autocracies.” Perspectives on Politics 10(4):891–916. Kalinin, Kirill. 2013. “Study of Social Desirability Bias at the Russian Presidential Elections.” Unpublished manuscript. Karakoc, Ekrem. 2018. Inequality after the Transition: Political Parties, Party Systems, and Social Policy in Southern and Postcommunist Europe. Oxford: Oxford University Press. Karatnycky, Adrian, & Peter Ackerman. 2005. “How Freedom Is Won: From Civic Resistance to Durable Democracy.” Washington, DC: Freedom House. Karl, Katherine A., & Cynthia L. Sutton. 1998. “Job Values in Today’s Workforce: A Comparison of Public and Private Sector Employees.” Public Personnel Management 27(4):515–527. Karl, Terry Lynn. 1997. The Paradox of Plenty: Oil Booms and Petro States. Berkeley: University of California Press. Katz, Lawrence F., & Alan B. Krueger. 1991. Changes in the Structure of Wages in the Public and Private Sectors. In Research in Labor Economics, ed. Ronald Ehrenberg. Greenwich, CT: Jai Press pp. 131–172. Keogh, Ruth H., & D.R. Cox. 2014. Case-control studies. Cambridge: Cambridge University Press. Kerbo, Harold R. 1982. “Movements of ‘Crisis’ and Movements of ‘Affluence’: A Critique of Deprivation and Resource Mobilization Theories.” Journal of Conflict Resolution 26(4): 645–663. King, Gary, & Langche Zeng. 2001. “Logistic Regression in Rare Events Data.” Political Analysis 9(2):137–163. Kitschelt, H., Z. Mansfeldova, R. Markowski & G. Toka. 1999. Post-Communist Party Systems: Competition, Representation, and Inter-Party Collaboration. Cambridge: Cambridge University Press. Kitschelt, Herbert, & Philipp Rehm. 2014. “Occupations as a Site of Political Preference Formation.” Comparative Political Studies 47(12):1670–1706. Kitschelt, Herbert, & Steven I. Wilkinson, eds. 2007. Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition. Cambridge: Cambridge University Press. Koesel, Karrie, & Valerie Bunce. 2012. “Putin, Popular Protests, and Political Trajectories in Russia: A Comparative Perspective.” Post-Soviet Affairs 28(4):403–423. Kohli, Atul. 2004. State-Directed Development. New York: Cambridge University Press. Koo, Hagen. 1991. “Middle Classes, Democratization, and Class Formation: The Case of South Korea.” Theory and Society 20(4):485–509. Kornai, János. 1992. The Socialist System: The Political Economy of Communism. Oxford: Oxford University Press.

260

bibliography

Krastev, Ivan, & Stephen Holmes. 2012. “An Autopsy of Managed Democracy.” Journal of Democracy 23(3):33–45. Kristiansen, Stein, & Muhid Ramli. 2006. “Buying an Income: The Market for Civil Service Positions in Indonesia.” Contemporary Southeast Asia 28(2):207–233. Krueger, Alan B. 1988. “The Determinants of Queues for Federal Jobs.” Industrial and Labor Relations Review 41(4):567–581. Kudasheva, Tat’iana Viktorovna. 2012. Modelirovanie i otsenka vliianiia ekonomicheskikh faktorov na stratifikatsiiu naseleniia Respubliki Kazakhstan. PhD thesis, Kazakh National University Al-Farabi. Almaty, Kazakhstan. Kudrin, Alexey, & Evsey Gurvich. 2015. “A New Growth Model for the Russian Economy.” Russian Journal of Economics 1(1):30–54. Kuran, Timur. 1995. Private Truths, Public Lies: The Social Consequences of Preference Falsification. Cambridge, MA: Harvard University Press. Kurlantzick, Joshua. 2015. State Capitalism: How the Return of Statism is Transforming the World. New York: Oxford University Press. Kuzio, Taras. 2011. Ukraine: Muddling Along. In Cental and East European Politics: From Communism to Democracy, ed. Sharon L. Wolchik & Jane Leftwich Curry. Plymouth, UK: Rowman & Littlefield, pp. 335–368. Lancaster, Tony, & Guido Imbens. 1996. “Case-Control Studies with Contaminated Controls.” Journal of Econometrics 71:145–160. Landes, David. 1998. The Wealth and Poverty of Nations. New York: Norton. Lankina, Tomila, & Alisa Voznaya. 2015. “New Data on Protest Trends in Russia’s Regions.” Europe-Asia Studies 67(2):327–342. Larreguy, Horacio, Cesar E. Montiel Olea & Pablo Querubin. 2017. “Political Brokers: Partisans or Agents? Evidence from the Mexican Teachers’ Union.” American Journal of Political Science 61(4):877–891. Lee, Ching Kwan, & Yonghong Zhang. 2013. “The Power of Instability: Unraveling the Microfoundations of Bargained Authoritarianism in China.” American Journal of Sociology 118(6):1475–1508. Leiulfsrud, Hakon, Ivano Bison & Heidi Jensberg. 2005. “Social Class in Europe. European Social Survey 2002/3.” NTNU Social Research Ltd., Trondheim. http://www. europeansocialsurvey.org/docs/methodology/ESS1_social_class.pdf. Lentz, Bernard F., & David N. Leband. 1989. “Why So Many Children of Doctors Become Doctors: Nepotism vs. Human Capital Transfers.” Journal of Human Resources 24(3): 396–413. Leventoglu, Bahar. 2014. “Social Mobility, Middle Class, and Political Transitions.” Journal of Conflict Resolution 58(5):825–864. Levinson, Aleksei. 2012. “Eto ne srednij klass—eto vse.” http://www.vedomosti.ru/opinion /news/1509376/eto_ne_srednij_klass_eto_vse. Lewis, Gregory B., & Sue A. Frank. 2002. “Who Wants to Work for the Government?” Public Administration Review 62(4):395–404. Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic Development and Political Legitimacy.” American Political Science Review 53(1):69–105. Lipset, Seymour Martin. 1960. Political Man: The Social Bases of Politics. Garden City, NY: Doubleday & Company, Inc.

b i b l i o g r a p h y 261 Liu, Hanzhang, & Yuhua Wang. 2017. “Elite School Networks and Working for the Government: Natural Experimental Evidence from China.” Unpublished manuscript. Harvard University. http://scholar.harvard.edu/files/yuhuawang/files/vocation.pdf. Lohmann, Suzanne. 1994. “The Dynamics of Informational Cascades: The Monday Demonstrations in Leipzig, East Germany, 1989–91.” World Politics 47(1):42–101. Mach, Bogdan W., & John E. Jackson. 2006. “Employment Change, Attitude Evolution and Voting during Poland’s Transition: Longitudinal Evidence.” European Journal of Political Economy 22(2):472–502. Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico. New York: Cambridge University Press. Magaloni, Beatriz. 2010. “The Game of Electoral Fraud and the Ousting of Authoritarian Rule.” American Journal of Political Science 54(3):751–765. Makarkin, Aleksei. 2011. “Protest srednego klassa.” http://www.politcom.ru/print.php? id=13024. Maleva, Tat’iana, & Lilia Ovcharova. 2008. Rossijskie srednie klassy: nakanune i na pike ekonomicheskogo rosta. In Rossijskie srednie klassy: nakanune i na pike ekonomicheskogo rosta, ed. A. E. Shastitko et al. Moscow: Ekon-Inform. Maleva, Tat’iana, & Lilia Ovcharova. 2009. “Sotsial’naia modernizatsia i srednij klass.” Naselenie i Obshchestvo 381–382. Manza, Jeff, & Clem Brooks. 1996. “Does Class Analysis Still Have Anything to Contribute to the Study of Politics.” Theory and Society 25(5):717–724. Mares, Isabela. 2015. From Open Secrets to Secret Voting: Democratic Electoral Reforms and Voter Autonomy. New York: Cambridge University Press. Mares, Isabela, Aurelian Muntean & Tsveta Petrova. 2018. “Economic Intimidation in Contemporary Elections: Evidence from Romania and Bulgaria.” Government and Opposition 53(3):486–517. Mares, Isabela, & Lauren Young. 2016. “Buying, Expropriating, and Stealing Votes.” Annual Review of Political Science 19:267–288. Marwell, Gerald, & Pamela Oliver. 1993. The Critical Mass in Collective Action: A Micro-Social Theory. Cambridge: Cambridge University Press. Marx, Paul. 2011. “The Unequal Incidence of Non-standard Employment Across Occupational Groups.” IZA Discussion Papers, No. 5521, Institute for the Study of Labor (IZA), Bonn. http://nbn-resolving.de/urn:nbn:de:101:1-201104113948. McConnell, Christopher, Yotam Margalit, Neil Malhotra & Matthew Levendusky. 2018. “The Economic Consequences of Partisanship in a Polarized Era.” American Journal of Political Science 62(1):5–18. McGlinchey, Eric. 2011. Chaos, Violence, Dynasty: Politics and Islam in Central Asia. Pittsburgh, PA: University of Pittsburgh Press. McGregor, Richard. 2010. The Party: The Secret World of China’s Communist Rulers. New York: Harper. McIntosh, Mary E., Martha Abele MacIver, Daniel G. Abele & Dina Smeltz. 1994. “Publics Meet Market Democracy in Central and East Europe, 1991–1993.” Slavic Review 53:483–512. McIntosh, Mary E., Martha Abele MacIver, Richard B. Dobson & Steven A. Grant. 1993. “The Meaning of Democracy in a Redefined Europe.” Paper presented at the 1993 Annual Meeting of the American Association for Public Opinion Research, St. Charles, Illinois.

262

bibliography

McMann, Kelly M. 2006. Economic Autonomy and Democracy: Russia and Kyrgyzstan. New York: Cambridge University Press. Medina, Luis Fernando, & Susan C. Stokes. 2007. Monopoly and Monitoring: An approach to Political Clientelism. In Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. Herbert Kitschelt & Steven I. Wilkinson. New York: Cambridge University Press pp. 68–83. Meltzer, Allan H., & Scott F. Richard. 1981. “A Rational Theory of the Size of Government.” Journal of Political Economy 89(5):914–927. Milanovic, Branko, & Shlomo Yitzhaki. 2002. “Decomposing World Income Distribution: Does the World Have a Middle Class?” Review of Income and Wealth 48(2):155–178. Miller, Arthur H., Vicki L. Hesli & William M. Reisinger. 1994. “Reassessing Mass Support for Political and Economic Change in the Former USSR.” American Political Science Review 88(2):399–411. Miller, Warren E. 1999. “Temporal Order and Causal Inference.” Political Analysis 8(2):119–139. Ministry of Economic Development of the Russian Federation. 2011. “Sostoianie bytovoj korruptsii v Rossijskoj Federatsii.” http://www.indem.ru/corrupt/doklad_cor_INDEM_ FOM_2010.pdf. Mills, C. Wright. 1956. The Power Elite. New York: Oxford University Press. Moore, Barrington. 1966. Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World. Boston: Beacon Press. Mueller, Richard E. 1998. “Public-Private Sector Wage Differentials in Canada: Evidence from Quantile Regressions.” Economics Letters 60(2):229–235. Munck, Gerardo L., & Jay Verkuilen. 2002. “Conceptualizing and Measuring Democracy: Evaluating Alternative Indices.” Comparative Political Studies 35:5–34. Murillo, Maria Victoria, & Ernesto Calvo. 2019. Non-Policy Politics: Richer Voters, Poorer Voters and the Diversification of Electoral Strategies. New York: Cambridge University Press. Musacchio, Aldo, & Sergio G. Lazzarini. 2012. “Leviathan in Business: Varieties of State Capitalism and Their Implications for Economic Performance.” Harvard Business School Working Paper 12-108. http://www.hbs.edu/faculty/Pages/download.aspx?name=12108.pdf. Musacchio, Aldo, & Sergio G. Lazzarini. 2014. Reinventing State Capitalism: Leviathan in Business, Brazil and Beyond. Cambridge, MA: Harvard University Press. Naughton, Barry, & Kellee S. Tsai. 2015. State Capitalism, Institutional Adaptation, and the Chinese Miracle. New York: Cambridge University Press. Nelson, Joan M. 1993. “The Politics of Economic Transformation: Is Third World Experience Relevant in Eastern Europe?” World Politics 45(3):433–463. Neundorf, Anja. 2010. “Democracy in Transition: A Micro Perspective on System Change in Post-Socialist Societies.” Journal of Politics 72(4):1096–1108. Nunnally, J. C. 1978. Psychometric Theory. 2nd ed. New York: McGraw-Hill. Nurmagambetov, A. A. & S. K. Azhmetov, 2010. Rol’ formiruemogo srednego klassa v transformatsii obshchestvennykh otnoshenij i formirovanii grazhdanskogo obshchestva v Kazakhstane. In Fenomen sotsial’noj inzhenerii: opyt i rekonstruktsiia budushchego. Materialy mezhdunarogoj nauchno-prakticheskoj konferentsii, 21-22 maia 2010, ed. M. M. Tazhin. Almaty, Kazakhstan: Fond Pervogo Prezidenta Respubliki Kazakhstan pp. 234–239.

b i b l i o g r a p h y 263 O’Donnell, Guillermo A. 1973. Modernization and Bureaucratic-Authoritarianism: Studies in South American Politics. Berkeley: Institute of International Studies, University of California. O’Donnell, Guillermo, & Philippe C. Schmitter. 1986. Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies. Baltimore: Johns Hopkins University Press. O’Dwyer, Conor. 2004. “Runaway State Building: How Political Parties Shape States in Postcommunist Eastern Europe.” World Politics 56:520–553. Olcott, Martha Brill. 2002. Kazakhstan: Unfilfilled Promise. Washington, DC: Carnegie Endowment for International Peace. Olcott, Martha Brill. 2007. “KazMunaiGaz: Kazakhstan’s National Oil and Gas Company.” Policy Report published by The James A. Baker III Institute for Public Policy of Rice University. http://bakerinstitute.org/media/files/page/9820ee52/noc_kaz_olcott.pdf. Oliveros, Virginia. 2016. “Making it Personal: Clientelism, Favors, and the Personalization of Public Administration in Argentina.” Comparative Politics 48(3):373–391. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. Onuch, Olga A. 2014. “Who Were the Protesters?” Journal of Democracy 25(3):44–51. OSCE. 2006. “OSCE/ODIHR Election Observation Mission Final Report. Republic of Kazakhstan—Presidential Election 4 December, 2005.” http://www.osce.org/odihr/ elections/kazakhstan/18153?download=true. OSCE. 2012. “OSCE/ODIHR Election Observation Mission Final Report. Republic of Kazakhstan—Early Parliamentary Elections 15 January, 2012.” http://www.osce.org/ odihr/elections/89401?download=true. OSCE. 2015. “Statement of Preliminary Findings and Conclusions. OSCE Election Observation Mission, Republic of Kazakhstan—Early Presidential Election 26 April 2015.” http:// www.osce.org/odihr/elections/kazakhstan/153566?download=true. Ovcharova, Lilia. 2012. Russia’s Middle Class: At the Center or on the Periphery of Russian Politics. In Russia: Insights from a Changing Country, ed. Sabine Fisher. Paris: EUISS. Pavolvskii, Gleb. 2011. Genial’naia Vlast’! Slovar’ Abstraktsii Kremlia. Moscow: Evropa. Peisakhin, Leonid. 2018. “Contested Nationhood: Imperial Legacies and Conflicting Political Identities in Ukraine.” Unpublished Manuscript. New York University–Abu Dhabi. Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. New York: Cambridge University Press. Pop-Eleches, Grigore. 2007. “Historical Legacies and Post-Communist Regime Change.” The Journal of Politics 69(4):908–926. Pop-Eleches, Grigore, & Graeme Robertson. 2014. “After the Revolution.” Problems of PostCommunism 64(4):3–22. Pop-Eleches, Grigore, & Joshua A. Tucker. 2017. Communism’s Shadow: Historical Legacies and Contemporary Political Attitudes. Princeton, NJ: Princeton University Press. Pop-Eleches, Grigore, & Joshua Tucker. 2014. “Communist Socialization and Post-Communist Economic and Political Attitudes.” Electoral Studies 33:77–89. Pop-Eleches, Grigore, & Michael Donnelly. 2013. “Income Measures in the World Values Survey: Problems and Solutions.” Princeton University. Unpublished manuscript.

264

bibliography

Poterba, James M., & Kim S. Rueben. 1993. “The Distribution of Public Sector Wage Premia: New Evidence Using Quantile Regression Methods.” NBER Working Paper 4734. https:// www.nber.org/papers/w4734. Pressman, Steven. 2007. “The Decline of the Middle Class: An International Perspective.” Journal of Economic Issues 41(1):181–199. Prior, Markus. 2010. “You’ve Either Got It or You Don’t? The Stability of Political Interest over the Life Cycle.” Journal of Politics 72(3):747–766. Prothro, James W., & Charles M. Grigg. 1960. “Fundamental Principles of Democracy: Bases of Agreement and Disagreement.” Journal of Politics 22(2):276–294. Przeworski, Adam. 1991. Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. New York: Cambridge University Press. Przeworski, Adam, Michael E. Alvarez, Jose Antonio Cheibub & Fernando Limongi. 2000. Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990. New York: Cambridge University Press. Radnitz, Scott. 2010. “The Color of Money: Privatization, Economic Dispersion, and the PostSoviet ‘Revolutions.”’ Comparative Politics 42(2):127–146. Ravallion, Martin. 2010. “The Developing World’s Bulging (but Vulnerable) Middle Class.” World Development 38(4):445–454. Rechel, Bernd, Erica Richardson & Martin McKee, eds. 2014. Trends in Health Systems in the Former Soviet Countries. Copenhagen: World Health Organization. Remington, Thomas. 2011. “The Russian Middle Class as Policy Objective.” Post-Soviet Affairs 27(2):97–120. Remington, Thomas, Irina Soboleva, Andrey Sobolev & Mark Urnov. 2012. “The Middle Class in Four Russian Regions: Government Policies and Behavioral Strategies.” The Higher School of Economics, Moscow. https://www.hse.ru/pubs/share/direct/ document/53695111. Resnick, Danielle. 2015. “The Political Economy of Africa’s Emergent Middle Class: Retrospect and Prospects.” Journal of International Development 27(5):573–587. Richter, Kaspar. 2006. “Wage Arrears and Economic Voting in Russia.” American Political Science Review 100(1):133–145. Robertson, Graeme B. 2011. The Politics of Protest in Hybrid Regimes: Managing Dissent in Postcommunist Russia. New York: Cambridge University Press. Robinson, James A., & Thierry Verdier. 2013. “The Political Economy of Clientelism.” Scandinavian Journal of Economics 115(2):260–291. Róna-Tas, Ákos. 1997. The Great Surprise of the Small Transformation: The Demise of Communism and the Rise of the Private Sector in Hungary. Ann Arbor: University of Michigan Press. Rosenfeld, Bryn. 2017. “Reevaluating the Middle Class Protest Paradigm: A Case-Control Study of Democratic Protest Coalitions in Russia.” American Political Science Review 111(4):637–652. Rosenfeld, Bryn, Kosuke Imai & Jacob N. Shapiro. 2016. “An Empirical Validation Study of Popular Survey Methodologies for Sensitive Questions.” American Journal of Political Science pp. 783–802. Ross, Michael L. 1999. “The Political Economy of the Resource Curse.” World Politics 51(2):297–322. Ross, Michael L. 2001. “Does Oil Hinder Democracy?” World Politics 53(3):325–361.

b i b l i o g r a p h y 265 Rosstat. 2016. “Russia in Figures 2016: Statistical Handbook.” Official publication of the Russian Federal State Statistics Service. http://www.gks.ru/free_doc/doc_2016/rusfig /rus16e.pdf. Rosstat. 2017. “Russia in Figures 2017: Statistical Handbook.” Official publication of the Russian Federal State Statistics Service http://www.gks.ru/free_doc/doc_2017/rusfig /rus17e.pdf. Rozenas, Arturas, & Yuri M. Zhukov. 2019. “Mass Repression and Political Loyalty: Evidence from Stalin’s ‘Terror by Hunger.”’ American Political Science Review 113(2):569–583. Rueschemeyer, Dietrich, Evelyne Huber Stephens & John D. Stephens. 1992. Capitalist Development and Democracy. Cambridge: Polity Press. Russian Academy of Sciences, Institute of Sociology. 2014. “Srednij klass v sovremennoj Rossii: 10 let spustia.” http://www.isras.ru/files/File/Doklad/Analit_doc_Sredny_klass /full.pdf. Rustow, Dankwart A. 1970. “Transitions to Democracy: Toward a Dynamic Model.” Comparative Politics 2(3):337–363. Ruzanov, Aleksander. 2010. “Srednij klass Kazakhstana: ot potrebitel’skogo otnosheniia—k sozidaniiu i professionaliszmu.” National Business 75(5):32–37. Sagadiev, K. A., & B. I. Bekturganova. 1998. Srednij klass—‘test’ na sovremennost’: istoriia, teoriia, statistika. Almaty, Kazakhstan: Assotsiatsiia sotsiologov i politologov (ASiP). Satpaev, Dosym. 2008. “Mify i real’nost’ o srednem klasse v Kazakhstane. Voprosy i otvety.” http://risk.kz/11034. Schatz, Edward. 2006. “Access by Accident: Legitimacy Claims and Democracy Promotion in Authoritarian Central Asia.” International Political Science Review 27(3):263–284. Schatz, Edward, & Elena Maltseva. 2012. “Kazakhstan’s Authoritarian ‘Persuasion.”’ Post-Soviet Affairs 28(1):45–65. Schumpeter, Joseph. 1942. Capitalism, Socialism, and Democracy. New York: Harper. Scott, James C. 1969. “Corruption, Machine Politics, and Political Change.” American Political Science Review 63(4):1142–1158. Shaffer, Frederic Charles. 1998. Democracy in Translation: Understanding Politics in an Unfamiliar Culture. Ithaca: Cornell University Press. Sharunina, A. V. 2013. “Iavliaetsia li rossijskij ‘budzhetnik’ ‘neudachnikom’? Analiz mezhsektornykh razlichij v oplate truda.” Ekonomicheskij Zhurnal 17(1):75–107. Shevtsova, Lilia. 2012. “Implosion, Atrophy, or Revolution?” Journal of Democracy 23(3):19–32. Shin, Eui Hang. 1999. “Social change, Political Elections, and the Middle Class in Korea.” East Asia 17(3):28–60. Shleifer, Andrei, & Daniel Treisman. 2014. “Normal Countries: The East 25 Years After Communism.” Foreign Affairs. Full version available at http://scholar.harvard.edu/shleifer/ publications/normal-countries-east-25-years-after-communism. Sklar, Richard L. 1965. “Contradictions in the Nigerian Political System.” Journal of Modern African Studies 3(2):201–213. Slater, Dan. 2009. “Revolutions, Crackdowns, and Quiescence: Communal Elites and Democratic Mobilization in Southeast Asia.” American Journal of Sociology 115(1):203–254. Smyth, Regina, A. S. Sobolev & I. V. Soboleva. 2013a. “Patterns of Discontent: Identifying the Participant Core in Russian Post-Election Protest.” http://www.ucrs.uu.se/digitalAssets/ 254/254668_3discontentconfsmythpatternsofdiscontent.pdf.

266

bibliography

Smyth, Regina, A. S. Sobolev & I. V. Soboleva. 2013b. “A Well-Organized Play: Symbolic Politics and the Effect of the Pro-Putin Rallies.” Problems of Post-Communism 60(2):24–39. Sørenson, Aage B. 2000. “Toward a Sounder Basis for Class Analysis.” American Journal of Sociology 105(6):1523–1558. Stokes, Susan C. 2005. “Perverse Accountability.” American Political Science Review 99:315–325. Sunden, Jed. 2005. “A Revolution? We’re Still Waiting.” Kyiv Post. April 28. http://www. kyivpost.com/opinion/op-ed/a-revolution-were-still-waiting-22644.html. Sundhaussen, Ulf. 1991. “Democracy and the Middle Classes: Reflections on Political Development.” Australian Journal of Politics and History 37:100–117. Svolik, Milan W. 2012. The Politics of Authoritarian Rule. Cambridge: Cambridge University Press. Szakonyi, David. 2019. “Banking on Politics: How Russia’s Banking Sector Clean-Up Affects Regime Stability.” PONARS Eurasia: Point & Counterpoint. http://www.ponarseurasia. org/point-counter/article/banking-politics-how-russias-banking-sector-clean-affects-reg ime-stability. Tang, Min, & Dwayne Woods. 2014. “Conditional Effect of Economic Development on Democracy—the Relevance of the State.” Democratization 21(3):411–433. Tang, Min, Narisong Huhe & Qiang Zhou. 2017. “Contingent Democratization: When Do Economic Crises Matter?” British Journal of Political Science 47(1):71–90. Temirova, A. B., & A. Sh. Abdimomynova. 2016. “Sotsial’no-Ekonomicheskaia Differentsiatsiia regionov Respubliki Kazakhstan.” Vestnik Rossikskogo universiteta druzhby narodov. Economic Series. pp. 58–69. Teorell, Jan. 2010. Determinates of Democratization. New York: Cambridge University Press. Therborn, Goran. 1977. “The Rule of Capital and the Rise of Democracy.” New Left Review 103. Thurow, Lester. 1987. “A Surge in Inequality.” Scientific American 256:30–37. Tikhonova, N. E. 2008. Srednij klass v sotsiologicheskom izmerenii. In Srednij klass: problemy formirovaniia i perspektivy rosta. Moscow: Institute Sovremenogo Razvitiia pp. 23–26. Tikhonova, N. E., & S. V. Mareeva, eds. 2009. Srednij klass: teoriia i real’nost’. Moscow: Al’fa-M. Tsai, Kellee S. 2005. “Capitalists without a Class: Political Diversity among Private Entrepreneurs in China.” Comparative Political Studies 38(9):1130–1158. Tucker, Joshua. 2006. Regional Economic Voting: Russia, Poland, Hungary, Slovakia and the Czech Republic, 1990–1999. Cambridge: Cambridge University Press. Tucker, Joshua A. 2007. “Enough! Electoral Fraud, Collective Action Problems, and PostCommunist Colored Revolutions.” Perspectives on Politics 5(3):535–551. Tucker, Joshua A., Alexander C. Pacek & Adam J. Berinsky. 2002. “Transitional Winners and Losers: Attitudes Toward EU Membership in Post-Communist Countries.” American Journal of Political Science 46(3):557–571. Ulfelder, Jay. 2005. “Contentious Collective Action and the Breakdown of Authoritarian Regimes.” International Political Science Review 26(3):311–334. Urankhaev, N. T., E. R. Salimov & A. Zhubaniiazova. 2010. Sotsial’naia proizvodstvennaiia infrastruktura kak faktor obespecheniia ekonomicheskoj bezopasnosti Respubliki Kazakhstan. In Fenomen sotsial’noj inzhenerii: opyt i rekonstruktsiia budushchego: Materialy mezhdunarogoj nauchno-prakticheskoj konferentsii, 21–22 maia 2010 g., ed. M. M. Tazhin. Almaty, Kazakhstan: Fond Pervogo Prezidenta Respubliki Kazakhstan pp. 347–352.

b i b l i o g r a p h y 267 van de Walle, Nicolas. 2014. “Barrington Moore in the Tropics: Democracy and the African Middle Class.” Paper prepared for the Political Economy of Regime Transitions Workshop at Lehigh University. van Stekelenburg, Jacquelien, Stefaan Walgrave, Bert Klandermans & Joris Verhulst. 2012. “Contextualizing Contestation: Framework, Design, and Data.” Mobilization 17(3): 249–262. Vassilev, Rossen. 1999. “Modernization Theory Revisited: The Case of Bulgaria.” East European Politics and Societies 13(3):566–599. Volkov, Denis. 2012. “The Protesters and the Public.” Journal of Democracy 23(3):55–62. Walgrave, Stefaan, & Joris Verhulst. 2011. “Selection and Response Bias in Protest Surveys.” Mobilization 16(2):203–222. Wang, Yuhua. 2018. “Are College Graduates Agents of Change? Education and Political Participation in China.” Unpublished manuscript. Harvard University. https://ssrn.com/ abstract=2905288. Way, Lucan. 2015. Pluralism by Default: Weak Autocrats and the Rise of Competitive Politics. Baltimore: Johns Hopkins University Press. Wedeen, Lisa. 1999. Ambiguities of Domination: Politics, Rhetoric, and Symbols in Contemporary Syria. Chicago: University of Chicago Press. Weeden, Kim A., & David B. Grusky. 2005. “The Case for a New Class Map.” American Journal of Sociology 111(1):141–212. Weil, Frederick D. 1985. “The Variable Effects of Education on Liberal Attitudes: A Comparative-Historical Analysis of Anti-Semitism Using Public Opinion Survey Data.” American Sociological Review 50(4):458–474. Weitz-Shapiro, Rebecca. 2012. “What Wins Votes: Why Some Politicians Opt Out of Clientelism.” American Journal of Political Science 56(3):568–583. Welzel, Christian. 2013. Freedom Rising: Human Empowerment and the Quest for Emancipation. Cambridge: Cambridge University Press. Welzel, Christian, & Ronald Inglehart. 2008. “The Role of Ordinary People in Democratization.” Journal of Democracy 19(1):126–140. Wengle, Susanne. 2015. Post-Soviet Power: State-led Development and Russia’s Marketization. New York: Cambridge University Press. Weyland, Kurt. 1996. Democracy without Equity: Failures of Reform in Brazil. Pittsburgh: University of Pittsburgh Press. Weyland, Kurt. 2004. “Neoliberalism and Democracy in Latin America: A Mixed Record.” Latin American Politics and Society 46(1):135–157. White, Stephen. 2011. “Elections Russian Style.” Europe-Asia Studies 63(4):531–556. Whitefield, Stephen, & Geoffrey Evans. 2001. Attitudes towards the West, Democracy, and the Market. In Democratic Consolidation in Eastern Europe: International and Transnational Factors, ed. Jan Zielonka & Alex Pravda. Vol. 2. Oxford University Press pp. 231–253. Wolfinger, Raymond E. 1972. “Why Political Machines Have Not Withered Away and Other Revisionist Thoughts.” Journal of Politics 34(2):365–398. Wooldridge, Jeffrey M. 2003. Introductory Econometrics: A Modern Approach. 2nd ed. Australia: Thomson South-Western.

268

bibliography

World Bank. 2007. “Ukraine: Poverty Update.” Human Development Sector; Ukraine, Belarus and Moldova Country Unit; Europe and Central Asia Region. Report No. 39887–UA. http://web.worldbank.org/archive/website01022A/WEB/IMAGES/UKRAINEP.PDF. World Bank. 2015. “Kazakhstan: Low Oil Prices; Opportunity to Reform.” World Bank, Economic Update No. 1. Washington, DC. Yang, David D. 2007. “Classing Ethnicity: Class, Ethnicity, and the Mass Politics of Taiwan’s Democratic Transition.” World Politics 59(4):503–538. Yastrebov, G. A. 2011. “Kharakter stratifikatsii rossijskogo obshchestva v sravnitel’nom kontekste.” Vestnik Obshchestvennogo Mneniia 4(110):15–27. Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge University Press.

index

Page numbers in italics refer to figures and tables. Acemoglu, Daron, 1, 100 Africa, 240; electoral competition in, 45; middle class in, 24, 17n24, 234; public-sector employment in, 101 agriculture, 158, 201, 203, 205–6, 213 Algeria, 234 Angola, 234 Ansell, Ben W., 4, 18–19, 202 area studies, 24 Argentina, 8, 239 Aristotle, 1, 222, 230 Armenia, 88n31, 220 asset mobility, 16n13 association, freedom of, 56 attitudes, behavior vs., 20, 30–31, 33, 60, 98, 102–3 authoritarianism, autocracies: bureaucracy and, 23; career paths and sympathy for, 134, 136, 138, 140, 144–45, 146, 153; challenges to, 105; competitive, 8, 28; defections policed by, 43; democracy support under, 2, 78–86; democratic rhetoric of, 57; economic interventionism coupled with, 3; electoral and political mobilization by, 23, 26, 36, 44–45, 218; elite interests aligned with, 21, 231; employment under, 5–6, 10–11, 38, 39–40; favoritism under, 27, 63, 243; in Gulf region, 236; higher education under, 15; institutional buttressing of, 22; in Latin America, 239; middle class co-opted by, 17–18, 21, 103, 105, 106, 107,

131, 242; middle class cultivated by, 3, 4–5, 25, 27, 29, 32, 35–36, 38, 39, 44–45, 49, 99, 133, 160, 193–222, 226, 228–29, 233, 240, 243; middle-class divisions under, 3, 19; mobility under, 6, 18, 47, 48, 105, 237; in modern Europe, 12–13; political economy of, 23, 39–47; post-Soviet resurgence of, 28, 140, 143, 185; privatization vs. nationalization by, 192, 234, 239, 240; range, 8, 161, 162, 224; rent seeking encouraged by, 16, 42, 105–6, 190–91, 235, 239; satisfaction with, 46, 51–52, 67, 191, 238, 242; social conditioning to, 53–54; state capitalism in, 4; state dependency linked to, 5, 6, 14, 27, 32, 33, 37–39, 44, 47–55, 63, 66–68, 70, 85–86, 90, 92, 97–99, 101, 105–6, 156–57, 189, 224, 239; survey research on, 71–72, 78–81, 83–84, 141–54, 225; theories of, 228. See also protests Azerbaijan, 3, 11, 101, 132, 162, 238, 239 Ashmetov, S. K., 235 Banerjee, Abhijit, 69, 223 Behar, Alberto, 235, 236 behavior, attitudes vs., 20, 30–31, 33, 60, 98, 102–3 Beissinger, Mark, 157n3 Belarus, Belarusians, 3, 11, 101, 212; autocracy in, 220, 224; health care in, 41; public-sector employment in, 132, 162 Bellin, Eva R., 8, 45–46 269

270 Beresbaev, Bolat Talipovich, 193 Birdsall, Nancy, 234 Bismarck, Otto von, 21 Bissenova, Alima, 209, 211, 217n30 Bolivia, 239–40 Brazil, 8, 104n3, 234, 239 brokers, 18, 38, 43, 45 budget sector, 64, 76, 109n8, 113; as career choice, 149, 151, 153, 154, 175; corruption in, 42; state administration vs., 63; wages in, 3, 52, 129, 203 Bunce, Valerie, 110 Cambodia, 234 Campbell, David E., 15 Capoccia, Giovanni, 12n9 career paths, in public sector, 29, 33, 34, 37; authoritarian attitudes and, 134, 136, 138, 140, 144–45, 146, 153; financial incentives for, 136–37, 138–39; models predicting, 147, 150; networks and, 135, 137–38, 139, 146, 149, 151, 152–53, 225; political values and, 135, 138, 140, 148; self-selection into, 54–55, 76, 85, 89–90, 97, 103, 126n32, 134–54; in state administration, 4, 32, 136, 142, 145, 146, 149–53, 235 case-control method, 31, 102, 113, 117 Caucasus, 235 censorship, 218 Central Asia, 28, 235 Chaisty, Paul, 110 Chen, Jie, 24 Chile, 8, 234, 239 China, 19, 234; nationalized industries in, 239; public-sector employment in, 106n7, 137, 141, 235–36, 242; statedependent middle class in, 24, 25, 69; social networks in, 137–38 Choudhry, Omar, 26–27 Chun Doo-hwan, 104 civil service, 3, 63, 64, 243; in Africa, 234, 240; attitudes shaped by, 76, 154; as career choice, 86, 152; in China, 106n7, 137; in democracies, 8n4; in Indonesia, 235; in Kazakhstan, 30, 32, 35, 196, 199,

index 203, 208, 209, 218; in Russia, 101, 109, 113, 154; Ukrainian deficiency in, 160, 161, 190–91. See also patronage clientelism, 43, 44, 48, 105n5, 138. See also patronage Collier, David, 7, 61–62 Collier, Ruth B., 7, 61–62 communications sector, 23 communism: collapse of, 1, 2, 26, 66; indoctrination and conditions under, 47, 53, 68; labor market control under, 39–40; middle class under, 2, 63, 82; post-Soviet influence of, 128–29, 187, 217 Congo, Democratic Republic of, 234 Converse, Philip E., 168 Coppedge, Michael, 254 corruption: autocracies stabilized by, 38–39, 42, 110, 180, 241; in budget sector, 42; career choice and, 138–39, 152n7; democracy impeded by, 48; in health and education sectors, 42; informal economies linked to, 59; in Kazakhstan, 207, 208; in Latin America, 104n3; nepotistic, 98, 139n3, 153; in new democracies, 136; official toleration of, 3, 42, 132, 159, 235, 239; public opinion of, 123, 144, 149, 230; in Russia, 42–43, 76, 109–10, 122, 123, 154, 230; in Ukraine, 159–60, 161, 166, 185, 190–91, 230. See also rents Crimea, 38, 166 Czech Republic, 10n7 Dahrendorf, Ralf, 13, 21 Dalton, Russell J., 57 Daly, John C. M., 200n10 defense industry, 41, 132 democratization, democracies: civic revolution linked to, 2; coalitions for, 108, 126–29, 131, 223, 225, 240–41; after communism’s fall, 66–67, 96; demographic differences and attitudes toward, 76–83, 94; development with and without, 5, 25–26, 99, 104, 133, 155, 230–33; disillusionment with, 155, 170, 185; economic performance and, 93; economic

i n d e x 271 statism and, 21–27; education and, 13–15, 38, 74, 81–82, 98; employment in, 10–11, 92–93; ethnic and regional differences in support for, 213–15; inequality and, 16, 17; in Latin America, 8, 36, 239; meritocracy linked to, 48; middle-class differences and, 3, 6, 19, 42–43, 50, 54, 67, 83, 99, 101, 105, 115, 133, 224–25, 229, 230; mobility linked to, 17, 18, 27, 190, 191; negative views of, 6–7, 8–9, 42, 46–50, 52, 68, 75, 84; neoliberalism linked to, 8; in nineteenth-century Europe, 12–13, 18–19, 22, 202; obstacles to, 6–7, 9, 24; overpromising during, 48; privatization linked to, 26–27, 35, 48, 191–92, 241–42; redistribution and, 68, 78, 95; rent seeking reduced by, 9, 16–17, 47–48, 157, 180, 190, 210; in resource vs. non-resource states, 221, 237, 243; retrenchment from, 35, 48, 133, 140, 155–56, 161, 168, 185–89, 225, 226, 240; rhetoric of, 57, 162, 167; state dependency and, 31, 32, 34, 36, 39, 43–65, 67, 70, 75, 85–93, 97, 99, 107, 134–54, 157, 158, 161–62, 164–65, 171–80, 185, 191, 194–95, 211, 216, 219, 220–21, 223–24, 226, 237; support for, defined, 33, 55–57, 67, 71–73, 83, 244; theories of, 4, 20, 28–31, 104–5, 227–28, 232; waves of, 1; working-class views of, 47, 97, 166, 221, 227, 231, 238. See also modernization theory; protests Diamond, Larry, 24 difference-in-differences design, 31, 35, 156, 164–65, 173, 189, 226 differential attrition, 32, 55, 90, 97 Duflo, Esther, 69, 223

202–3; regime preferences linked to, 39; rent seeking in, 63; as social measure, 7; state involvement in, 41, 49; wages in, 94 Egypt, 234, 241 Ekiert, Grzegorz, 222 elections: in Africa, 45; in autocracies, 23, 26, 36, 44–45; career paths and, 151; commissions administering, 45, 219; in Kazakhstan, 218–19; patronage linked to, 18, 44, 45, 243; public opinion of, 72; in Russia, 30, 33, 131, 218 electoral and political mobilization, 23, 26, 36, 44–45, 218 elite-competition theory, 7, 18, 47, 62, 77, 227–28 employment contracts, 41–42, 50, 109, 122 energy sector, 3, 41; autocracy and, 236–37; in Kazakhstan, 36, 201, 203, 206, 213–15, 220, 221, 238; state control of, 49, 141, 158, 159; volatility of, 38 Engels, Friedrich, 64 England, 12, 21 Erdogan, Recep, 239 Estonia, 9–10 European Bank for Reconstruction and Development (EBRD), 70, 85, 97, 136, 195, 201, 211 exit options: of long-term public employees, 89, 107, 179; public employees’ life cycle and, 122, 124–25; of middle class, 47, 52; retaliatory firing and, 106; state dependency linked to, 6, 39, 53, 132, 180–81, 223; in specialized fields, 153; in state-dominated fields, 141. See also mobility expression, freedom of, 56

Easterly, William, 69, 78, 230 Eastern Europe, 2, 28, 50, 66, 235 East Germany, 171 economic diversification, 27 education: democratization linked to level of, 13–15; employment in, 63, 76, 86, 87, 148, 152; of middle class, 62, 73–74, 82, 98; in post-Soviet states, 63, 82, 132, 135, 159,

financial sector, 41, 132; in Africa, 45; in Russia, 141, 241–42; in Ukraine, 158, 159 Fish, M. Steven, 26–27 flexibility, of work arrangements, 49, 141, 143 Foundation for Public Opinion (FOM), 112 France, 12–13, 21

272 fringe benefits, 41, 50, 109, 122, 131, 207, 208–9, 211 Frye, Tim, 23, 110 Fukuyama, Francis, 12 Gans-Morse, Jordan, 76, 137, 148, 154 GeoRating surveys, 112 Georgia, 8, 9, 90, 132–33, 162, 220 Gerber, Theodore P., 136 Germany, Germans, 12–13, 21, 212 Gerschenkron, Alexander, 22–23 Gervasoni, Carlos, 46n9 Gorodnichenko, Yuriy, 136, 159, 166 Greene, Kenneth F., 23, 26 Gulf region, 236–37 Haggard, Stephan, 241 Hale, Henry E., 228 Handley, Antoinette, 24 health care, 49, 202–3; employment in, 63, 76, 86, 87, 148, 152; as fringe benefit, 41, 50, 109, 208; in hospitals, 41, 42; in Russia, 132, 135, 141; in Ukraine, 159; wages in, 94 Higher School of Economics (HSE), 141, 145, 146 Holmes, Stephen, 101 hospitals, 41, 42 housing, for state employees, 41, 50, 109, 131, 207, 208–9, 211, 235 Houston, David J., 154 Huhe, Narisong, 241 Hungary, 10n7, 11n8, 26, 240 Huntington, Samuel P., 100 Ileuova, G., 201, 206 Imbens, Guido, 111, 117 Indonesia, 24, 235 inequality, 2, 16, 17, 71, 178. See also redistribution inflation, 2 informal economy, 59, 77, 122, 136 informational cascades, 131, 229 Inglehart, Ronald, 14

index International Labor Organization (ILO), 62, 64 Ivatova, L. M., 19 Jones, David M., 24–25 Jou, Willy, 57 judicial independence, 56, 72, 151 Junisbai, Barbara, 220 Kaufman, Robert R., 241 Kazakhstan, 8, 230; Asian models for, 193, 196, 217; censorship in, 218; civil service in, 30, 32, 35, 196, 199, 203, 208, 209, 218; corruption in, 207, 208; counterrevolutionary forces in, 229; democracy support in, 195, 211, 212, 214–15, 238; development goals in, 193–94, 204, 208–9, 216–17; elections in, 218–19; energy sector in, 36, 201, 203, 206, 213–15, 220, 221, 238; ethnic and regional differences in, in k212–14, 215; government procurement in, 207; hidden state influence in, 242; middle-class size and composition in, 200–204; middle class as symbol and goal in, 35–36, 194–95, 196–99, 208, 219–20, 226; poverty in, 198, 201, 203; privatization in, 194, 216, 220; public-sector employment in, 202, 206, 213, 215, 235; regime stability in, 193, 239; small business in, 203–4; sovereign wealth fund of, 206; state dependency in, 205–7, 211, 212, 220; urban vs. rural life in, 213 Kitschelt, Herbert, 19 Koesel, Karrie, 110 Kornai, János, 23, 40 Krastev, Ivan, 101 Kuchma, Leonid, 156n2, 159–60 Kudasheva, Tat’iana Viktorovna, 201 Kyrgyzstan, Kyrgyz, 212 labor contracts, 41–42, 50, 109, 122 labor unions, 43, 46, 110 Lancaster, Tony, 111, 117 land reform, 19

i n d e x 273 Lankina, Tomila, 110 Latin America, 3, 57; democratization and democratic failure in, 8, 36, 239; educational level in, 63; privatization in, 48; protests in, 104n3; public-sector employment in, 101; regime turnover in, 239 Latvia, 9–10 Lee, Ching Kwan, 235 Leventoglu, Bahar, 18 Life in Transition Survey (LiTS), 70, 85, 96, 97 Lipset, Seymour Martin, 13, 230 Lithuania, 9–10 Liu, Hanzhang, 137, 139 lustration, 48, 50 Magaloni, Beatriz, 27 Malaysia, 24–25, 234 Marcos, Ferdinand, 104 Mares, Isabela, 43n6 Marx, Karl, 64 McMann, Kelly M., 46 Meltzer, Allan H., 16, 17 Mexico, 23, 26, 43n7, 218 middle class: in Africa, 17n14, 24, 234; authoritarian cultivation of, 3, 4–5, 25, 27, 29, 32, 35–36, 38, 39, 44–45, 49, 99, 133, 160, 193–222, 226, 228–29, 233, 240, 243; in China, 24, 25, 69; under communism, 2, 63, 82; contrasting views of, 1; co-optation of, 13, 17–18, 21, 103, 105, 106, 107, 131, 242; defined, 7, 58–63, 73–74, 81, 97–98, 104, 105, 165–66, 198, 201; divisions within, 3, 6, 19, 42–43, 50, 54, 67, 83, 99, 101, 105, 115, 133, 224–25, 229, 230; education of, 62, 73–74, 82, 98; ethnicity and, 212–14; exit options of, 47, 52; in Kazakhstan, 32, 35–36, 194–205, 207–12, 216–17, 219–21, 226, 238; in mass uprisings, 2; mobility and, 18, 27, 47, 154, 182, 184; in modernization theory, 13–15, 22, 78, 130, 232; in protests, 31, 33–34, 52, 88, 100, 102, 104, 108–25, 127, 128, 130–31,

224, 226, 229, 232, 241; in redistributive theories, 15–17; rent seeking by, 42; in resource vs. non-resource states, 236–38; in Russia, 25, 28, 29, 30, 44, 51, 54, 69, 88, 100–101, 103, 104, 108–25, 127, 128, 130–31, 133, 142; state dependency and, 2–5, 8–9, 11, 17, 20, 24, 25, 29, 33, 36, 41, 44, 47–49, 69, 70, 81, 101, 103, 120–23, 133, 156, 158, 188–89, 211, 214–16, 220, 224, 227, 234; wages and, 23; in Western Europe, 12–13, 202 Middle East, 3, 36, 63, 101, 235 minority rights, 56, 72, 151, 167 mobility, 121, 165; in autocracies, 5–6, 18, 47, 48, 105, 237; democratization linked to, 17, 18, 27, 190, 191; informal networks as impediment to, 137; in post-communist states, 71, 77, 85, 94, 182, 184–85, 190, 206, 208–9. See also exit options modernization theory, 20; democracy support in, 78, 228; empirical support for, 230; middle class in, 13–15, 22, 78, 130, 232; shortcomings of, 14, 22; testing of, 163; values-based variant of, 4, 5, 7, 13, 14, 20, 22, 61, 130, 232; working-class values viewed by, 47 Mok, Junghwan, 235, 236 Moldova, 9–10, 162 monopolies, 23, 132, 240 Moore, Barrington, 4, 7, 13, 21, 62, 230 Morales, Evo, 239–40 Moscow State Institute for International Relations (MGIMO), 142, 145, 146, 149, 151 Moscow State University (MGU), 141–42, 146 Munck, Gerardo, 56 Muntean, Aurelian, 43n6 Nazarbaev, Nursultan, 193–94, 197–98, 209, 218 neoliberalism, 8, 48 nepotism, 98, 139n3, 153 Nigeria, 234

274 North Africa, 235 Nurmagambetov, A. A., 235 October Revolution (1917), 35, 195, 197 O’Donnell, Guillermo, 13–14n11 O’Dwyer, Conor, 10n7 Olcott, Martha Brill, 207 omitted variable bias, 31, 165 Orbán, Viktor, 240 Organization for Economic Cooperation and Development (OECD), 63 Organization for Security and Cooperation in Europe (OSCE), 196, 218–19 panel data, 25, 29, 30, 71, 99, 156; advantages of, 31; cross-sectional data vs., 163, 164 partisanship, 38, 139, 140 patronage, 138; electoral, 18, 44, 45, 243; autocracies stabilized by, 38–39, 43–44, 48, 130–31, 243; middle-class divisions linked to, 105; for middle class vs. working class, 44–45, 243; networks linked to, 139; party strength linked to, 10n7; regime preference shaped by, 26; supervision linked to, 43; in Ukraine, 159–60, 225. See also civil service; clientelism; rents Pavolvskii, Gleb, 109 Peisakhin, Leonid, 158 pensions, 107, 125 perestroika, 96 permanent income, 59, 60n20 Peru, 234 Peter, Klara Sabarianova, 136, 159, 166 Petrova, Tsveta, 43n6 Pinochet Ugarte, Augusto, 234 Poland, 10n7, 26, 240 political economy, of authoritarianism, 23, 39–47 Pop-Eleches, Grigore, 59n18 post-communist states: ideology eclipsed by opportunism in, 41, 238; inequality in, 2, 16; inflation in, 2; middle-class divisions in, 19; middle-class support for

index democracy in, 70; poverty in, 2; public opinion data in, 28; small-business help centers in, 204; unemployment in, 2, 78, 155, 213 poverty, 2, 69, 156, 198, 201, 203, 220 preferences, revealed vs. latent, 56 press freedom, 72, 151 privatization, 23, 225; democratization linked to, 26–27, 35, 48, 191–92, 241–42; in Kazakhstan, 194, 216, 220; in Latin America, 48; limits to understanding, 26–27; public-sector employment resistant to, 234; in Ukraine, 158, 160 protests: critical mass of, 131, 229–30; ideology and, 121; in Latin America, 104n3; middle class in, 31, 33–34, 52, 88, 100, 102, 104, 108–25, 127, 128, 130–31, 224, 226, 229, 232, 241; in Russia, 31, 33–34, 44, 52, 67, 98, 100–102, 110–33, 224–25, 229; state dependency and, 44, 45, 157, 230; state workers in, 119; studies of, 110–15 Putin, Vladimir, 7, 101, 113, 129; authoritarianism of, 143; modernization pursued by, 108–9; nationalization under, 37–38, 242 Radnitz, Scott, 220 Ravallion, Martin, 69, 74, 82 redistribution, 5, 15–17, 20, 22, 68, 77, 78, 95, 178, 231 Rehm, Philipp, 19 rents: democratization as threat to, 9, 16–17, 47–48, 157, 180, 190, 210; in education, 63; from natural resources, 162, 191, 221; in public sector, 6, 11, 17, 37, 42–43, 48, 50, 105, 106, 109, 123, 132, 136, 138–39, 162, 190–91, 208, 235, 239; underreporting of, 77n11. See also patronage resource curse, 132n41 retaliation, 50–51, 68, 75, 88–89, 92, 177; protections from, 132n39; against protest participation, 106, 121–22, 131 Reuter, Ora John, 23, 110 revealed preferences, 56 reverse causation, 31, 85, 88, 178, 190

i n d e x 275 Rice, Condoleezza, 204 Richard, Scott F., 16, 17 risk tolerance, 84n27, 85, 162n9, 176n27, 177, 183, 184 Robertson, Graeme B., 126 Robinson, James A., 1, 100 Róna-Tas, Ákos, 11n8, 26 Ross, Michael L., 232, 236 rule of law, 8n4, 42–43, 48, 50, 173, 217 Russia, 7, 11, 238; authoritarian resurgence in, 140, 143; chauvinism in, 145; communists in, 128–29; corruption in, 42–43, 76, 109–10, 122, 123, 154, 230; economic growth in, 108–9; elections in, 30, 33, 100, 131, 218; grievances in, 123; job security in, 122–23, 239; middle-class divisions in, 3, 19, 42, 115, 133, 224–25, 230; mixed ownership in, 38; nationalized banks in, 241–42; perestroika in, 96; positive vs. negative inducements in, 46, 52; private health care in, 41; protests in, 31, 33–34, 44, 52, 67, 98, 100–102, 110–33, 224–25, 229; public-sector employment in, 34, 37, 76, 101, 109, 113, 134–54, 162, 225, 235; radicalism in, 129; research methods in, 30–31; state-dependent middle class in, 25, 29, 44, 69, 101, 103; state involvement in, 37–38, 49–41, 241–42; top-down modernization in, 37; Ukraine’s ties to, 158; wages in, 136 Russian Orthodox Church, 145 Rustow, Dankwart A., 163 Ruzanov, Aleksander, 200 Samruk-Kazyna, 206, 217 Samuels, David J., 4, 18–19, 202 Satpaev, Dosym, 207 Schaefer, David R., 136 selection bias, 20, 68, 75, 85 Shin, Doh C., 57 Sklar, Richard L., 24 Slovakia, 10n7 Smyth, Regina, 110, 131 Sobolev, A. S., 110, 131

Soboleva, I. V., 110, 131 social desirability bias, 33, 51n11, 68, 71, 77n11, 80n17, 88–89, 177–78n31 socialization: ideological, 33, 47, 54, 67, 68, 76, 78, 96, 132n40, 227; through state employment, 53–54, 75, 88, 91, 238 Social Origins of Dictatorship and Democracy (Moore), 21 Society and Democracy in Germany (Dahrendorf), 21 Soviet Union, 2, 66; Communist Party membership in, 140; education in, 15; ideological mobilization by, 227; nostalgia for, 145; social networks in, 137 speech freedom, 56, 72, 151 Sharunina, A. V., 42, 136 Shevtsova, Lilia, 110 Shleifer, Andrei, 63 state administration, 7, 64, 86, 160; budget sector vs., 63; as career choice, 4, 32, 136, 142, 145, 146, 149–53, 235; in China, 24; growth of, 10n7, 41, 207 state employment, defined, 63–65 state banks, 23 strategic industries, 41, 220–21 Suchinda Khraprayun, 104 Suharto, 24 Szakonyi, David, 23, 110 Tajikistan, 220, 239 Tang, Min, 25–26, 241 Tatars, 212 Tazhin, Marat M., 198n7 Tikhonova, N. E., 50, 109 Tokayev, Kassym-Jomart, 193 transportation sector, 23, 41, 49, 132, 141, 159 Treisman, Daniel, 63 Tsai, Kellee, 19, 242 Tucker, Joshua, 94n39 Tunisia, 241 Turkey, 239 Turkmenistan, 10, 220 Tymoschenko, Yulia, 160, 188

276

index

Ukraine, Ukrainians, 8, 9, 29, 90, 212, 224, 230; corruption in, 159–60, 161, 166, 185, 190–91, 230; counterrevolutionary forces in, 155–56, 157, 160, 161, 164, 168, 169, 185–90, 226, 229; democratic retrenchment in, 155–56, 161, 168, 185–89, 225, 226; educational level in, 165–66; ethnic and regional differences in, 157–58, 162, 184, 187–89, 191, 214, 226; mobility in, 165, 190; Orange Revolution in, 7, 30, 35, 67, 155–56, 160–70, 176, 181–91, 225–26, 248–49, 250; public opinion data in, 163, 165; public-sector employment in, 132–33, 136, 158, 160, 161, 162, 165, 166, 172–75, 188, 190–91, 225, 237; regime preferences in, 166–81; regime turnover in, 239; state dependency in, 34–35, 52, 172–81, 190–91, 231; voting behavior in, 185–91 unemployment, 2, 78, 155, 213 Union of Soviet Socialist Republics (USSR), 9 unions, 43, 46, 110 universal suffrage, 56 Uzbekistan, 8, 11, 220, 224, 239 Uzbeks, 212

Veblen, Thorstein, 21 Venezuela, 234 Verkuilen, Jay, 56 Voznaya, Alisa, 110

values-based modernization theory, 4, 5, 7, 13, 14, 20, 22, 61, 130, 232 van de Walle, Nicolas, 12, 17n14, 202

Zhang, Yonghong, 235 Zhou, Qiang, 241 Ziblatt, Daniel, 12n9

wages, 12, 178; in budget sector, 3, 52, 129, 203; informal, 77; in Kazakhstan, 198, 203, 205, 220; middle-class growth linked to, 23; politically based, 40, 44; in postSoviet states, 94, 109, 162; in public vs. private sector, 11, 42, 52, 77, 95n44, 136, 235, 243; for teachers, 94. See also fringe benefits Wang, Yuhua, 137, 139 Weitz-Shapiro, Rebecca, 45 Welzel, Christian, 14 Whitefield, Stephen, 110 Woods, Dwayne, 25–26 World Bank, 235 xenophobia, 145 Yanukovych, Viktor, 156, 160, 161, 163, 169, 176, 182 Yushchenko, Viktor, 160–61, 187

a note on the type

This book has been composed in Arno, an Old-style serif typeface in the classic Venetian tradition, designed by Robert Slimbach at Adobe.