The Regional Roots of Russia's Political Regime 2016028182, 9780472130184, 9780472122462

Citation preview

Page ii → Page iii →

The Regional Roots of Russia’s Political Regime William M. Reisinger and Bryon J. Moraski University of Michigan Press Ann Arbor

Page iv → Copyright В© 2017 by William M. Reisinger and Bryon J. Moraski All rights reserved This book may not be reproduced, in whole or in part, including illustrations, in any form (beyond that copying permitted by Sections 107 and 108 of the U.S. Copyright Law and except by reviewers for the public press), without written permission from the publisher. Published in the United States of America by the University of Michigan Press Manufactured in the United States of America A CIP catalog record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Reisinger, William M. (William Mark), 1957– author. | Moraski, Bryon, author. Title: The regional roots of Russia’s political regime / William M. Reisinger and Bryon J. Moraski. Description: Ann Arbor : University of Michigan Press, 2017. | Includes bibliographical references and index. Identifiers: LCCN 2016028182| ISBN 9780472130184 (hardcover : alk. paper) | ISBN 9780472122462 (e-book) Subjects: LCSH: Russia (Federation)—Politics and government—1991– | Regionalism—Political aspects—Russia (Federation) Classification: LCC JN6693.5.R43 R46 2017 | DDC 320.947—dc23 LC record available at https://lccn.loc.gov/2016028182

Page v → To our families.

Page viii →Page ix →

Preface The project that led to this volume began in 2000, the same year Russians elected Vladimir Putin as their new president. Bryon defended his dissertation on legislative electoral system design in Russia’s regions that year, while Bill had begun collecting data on gubernatorial elections in Russia for a nascent project on how democratization unfolds at different levels: individual, regional, and national. Given our mutual research interests, Bill suggested that we pool our resources and collaborate in an effort to better understand electoral politics in Russia’s regions. Over the next 15 years, President Putin consolidated his position, prepared for a 2008 transferal of the presidency to a hand-picked successor, and returned to office in 2012. To facilitate these changes, the Kremlin eliminated gubernatorial elections and established an executive hierarchy of power that helped undermine the competitiveness of national elections. As Putin transformed Russian politics, the importance of the regional dimension grew clearer, and our project took on new dimensions. Also during those years, much changed for us personally. Bryon landed a tenure-track job at the University of Florida, was promoted with tenure, and then soon assumed positions in the department’s administration. Along the way, he and his wife had two wonderful daughters. Bill became department chair and then spent five years as Associate Provost and Dean of International Programs at the University of Iowa. During this period, however, our interest in Russian political developments beyond Moscow’s Ring Road remained constant. While long in the making, this book is a culmination of our joint effort to make sense of the complicated relationships between Page x →the Russian Federation’s political trajectory and those of its constituent regions. The authorship order reflects Bill’s role in initiating the project; otherwise, authorship for the book, as with our previously published manuscripts, is equal. This project has required extensive work to transfer data on Russia’s regions from their sources into our datasets. We have received valuable assistance in this task from numerous students at the University of Iowa and the University of Florida. We thank Merideth Bentley, Brooke Berget, Alexandra Huffman, Alanna Karpa, Branislav Kovalcik, Lauren McCarthy, Kathleen McNulty, Katherine Otto, John Rejowski, Justin Sugg, Elizabeth Veen, Joy Woods, and Hyemin Yoo for working with care on several waves of data entry. We also thank Saskia van Wees for her help formatting the manuscript and Kathlyn Cunneen for proofreading the text. We are grateful to John O’Loughlin and Thomas Remington for sharing shapefiles of Russia’s regions. Joe Aufmuth assisted in modifying the shapefiles for our use in chapter 8. We appreciate the financial support provided by our universities at several points over the course of the project. The University of Iowa’s Honors Research Scholar Program and its Center for Russian, East European and Eurasian Studies supported the initial data collection and analyses. A University of Florida Faculty Enhancement Opportunity Grant funded Bryon’s participation in the 2010 ICPSR Workshop on Spatial Regression Analysis. The University of Iowa’s International Programs and the Department of Political Science made possible travel to Russia and to present preliminary findings. Our analyses and conclusions have benefitted from suggestions we have received from Andrei Akhremenko, Michael Bernhard, Frederick Boehmke, Irina Busygina, Ernesto Calvo, Vladimir Gel’man, Magda Giurcanu, Robert Grey, David M. Hedge, Yoshiko M. Herrera, Tomila Lankina, Michael Martinez, Valentin Mikhailov, Sara Mitchell, Petr Panov, Cameron Ross, Inga Anna-Liisa Saikkonen, Gulnaz Sharafutdinova, Daniel Smith, Rostislav Turovskii, and Stephen White. We thank them as well as various anonymous reviewers. We also received valuable feedback from students and staff at the Politics Faculty of the National Research University–Higher School of Economics in Moscow. We are grateful to its dean, Andrei Melville, for inviting Bill to present our work. Chapter 4 is derived, in part, from an article published in Democratization on July 24, 2007, available online at http://dx.doi.org/10.1080/13510340701398303. An earlier version of the analyses in chapter 5 appeared in William M. Reisinger and Bryon J. Moraski, “DeferencePage xi → or Governance? A Survival Analysis of

Russia’s Governors Under Presidential Control,” in William M. Reisinger, ed., Russia’s Regions and Comparative Subnational Politics, Routledge Research in Comparative Politics, 40–62 (New York: Routledge, 2013). We are grateful for permission to reprint portions of that earlier work here. Finally, our most important debt of gratitude is to our families. Bill thanks Kathy and Mark for their encouragement before and during the time he has worked on this project. Bryon is grateful to Jayne, Blaire, and Joslyn for their love and understanding as well as the reminders that life is about much more than the words one writes on a page.

Page xii →Page 1 →

One Introduction In March 2014, Russia expanded its territory for the first time in its post-Soviet history by annexing Ukraine’s Crimean Peninsula. This not only rejuvenated President Vladimir Putin’s popularity at home but also created new electoral strongholds: regions largely comprised of, and more importantly governed by, ethnic Russians who are nostalgic for the Soviet past and grateful for a Russian present. The addition of these new subjects of the Russian Federation came after a decade and a half of President Putin’s recentralization policies. During this period, President Putin’s policies converted an unruly country governed as much by regional barons as federal policy into one with a much more hierarchical system of executive power. The regional barons now seek to support the national electoral authoritarian regime by providing the Kremlin with unassailable electoral victories. In this book, we examine post-Soviet Russian politics by analyzing election results at the subnational level. By focusing on Russia’s constituent regions, we augment the current understandings of two closely related but analytically distinguishable questions: Why did democracy fail to take root? And, how did Putin’s authoritarian regime develop? Russia and its future ought to command the interest of everyone concerned with world affairs. Its sprawling land mass connects east, central and west Asia to Europe. It is a nuclear power, a key regional power across Eurasia, and a permanent member of the United Nations Security Council. It is a major exporter of oil, natural gas, and other strategic fuels and minerals. Not surprisingly, the political course Russia follows influences the world in a substantial way. To understand where Russia’s policies may take Page 2 →it, we need the best possible understanding of where it has been, particularly the road Russia has followed since it emerged from the Soviet Union in 1991. That road was dramatic, to say the least. It required simultaneously rebuilding Russia’s state, its economy, and its political and judicial systems, as well as adjusting to a new role in international relations (Kuzio 2001). Russians suffered through horrific privations in the early years as the switch to market economics produced hyperinflation and unemployment as well as heightened crime and mortality. Although hopes were initially high that the new post-Soviet Russia would establish a stable democratic polity, those hopes were dashed. Russia’s first post-Soviet president, Boris Yeltsin, ended his second term (at the end of 1999) in ill-health and politically powerless. Yeltsin’s chosen successor, Vladimir Putin, solidified a more effective undemocratic political regime that continues to rule. While Putin’s regime has achieved widespread acceptance within Russia and has lasted more than 15 years, its prospects for long-term stability are currently under debate. Many insightful explanations are available for how and why Russia underwent these changes. Extant explanations range widely. Some highlight factors peculiar to Russia, while others place Russia in comparative perspective. Some focus on Russia’s ordinary citizens, while others privilege the role of political and economic elites or of transnational factors. Of course, multiple influences played a part, a point made by Way (2010, 247) in describing the consolidation of authoritarianism in Russia as “over-determined.” Such a description, however, implicitly shrugs at the challenge of understanding Russia’s post-Soviet political change. We argue that this change has yet to receive an adequately comprehensive explanation. In particular, what remains understudied is the extent to which the course of Russia’s political change since 1991 was shaped by subnational politics and federal relations. In this book, we address these dynamics, specifically with an eye to enriching our understanding of Russia’s postcommunist political course and, by extension, of democratization and authoritarianism more generally. Using careful multivariate analyses of regional electoral data, we are able to reveal the extent to which regional variations influenced the course of political change in Russia. Elections are important events that mobilize elite and public energy even in nondemocratic settings. In large federal states such as Russia, elections occur at multiple levels, and the national-level elections play out differently in different regions. We analyze both

gubernatorial election results and the regional results for federal elections. Page 3 →Our analyses include elections from 1991 to 2012 for all of Russia’s regions, using statistical techniques to describe and explain variation among them as well as trends over time. In addition, we analyze patterns of Kremlin gubernatorial appointments during the 2005–12 period when gubernatorial elections were eliminated. Our findings illuminate Russian politics by emphasizing the critical role that the subnational level plays in constructing and maintaining a national regime. We bring data from post-Soviet Russia to bear in addressing two concerns prominent in current comparative politics research. One is regime type, democracy versus authoritarianism. The other is subnational politics and how it varies in ways that influence national-level politics. Our approach, though motivated by scholarly questions, also allows us to weigh in on policy-relevant questions: Why and how did Russia regress from its democratic ambitions of the immediate post-Soviet period? How did Vladimir Putin build his particular nondemocratic regime without altering the constitutionally allocated powers of the presidency? And, how solidly built is his regime? To adequately answer any of these questions requires a deep understanding of Russian politics as it operates outside Moscow’s Ring Road. Before turning to our analyses, we will use this chapter to expound on Russia’s place in efforts to understand political regimes and regime change. We will then present our research questions and describe the approach we use in answering them. The chapter closes by indicating how our arguments are laid out in the subsequent chapters.

The Democratic Hope for Russia Democracy—struggles for it and resistance to it—is a dominant issue in world affairs. Democratic governance has commandeered the international high ground as the only legitimate form of political rule: even nondemocracies insist on being called democracies. In addition to their many advantages over authoritarianism in promoting freedom, self-determination, and other nonmaterial benefits (Dahl 1998, ch. 5), democracies better meet their citizens’ physical needs. Countries governed democratically experience fewer human rights violations and repression (Rummel 1998; Moeller and Skaaning 2013), their citizens live longer and healthier lives (Shin 1989; Przeworski et al. 2000, ch. 5; Franco, ГЃlvarez-Dardet, and Ruiz 2004; Ruger 2005; Besley and Kudamatsu 2006), and women and minorities have more opportunities (Jaquette 2001; Paxton, Hughes, and Painter 2010). Page 4 →Also noteworthy is the evidence from scholars of international relations that democracies do not wage war against other democracies (Bueno de Mesquita 2006; Doyle 2012) and that, when democracies do go to war, they are likely to be victorious (Reiter and Stam 2002). For these reasons, those who care about Russia want democracy for its citizens, just as those who care about democracy must prize success in Russia. Gorbachev’s reforms and the subsequent end of the Soviet Union brought to the fore the question: can democracy take root in Russia? (Eckstein et al. 1998). This question prompted a flurry of research from established experts on Russia, from scholars whose previous work had examined other countries or regions and from many talented new PhDs (for a review of the first decade of such work, see Laitin 2000, esp. 120–24). This research employed, and extended, all of the previous approaches to democratization. However, as this research was proceeding, it became increasingly clear that democracy was not consolidating but receding—even to those who, during the Yeltsin years, granted Russia the status of a democracy. When the new post-Soviet Russian regime had congealed and been tested in the 1996 presidential election, it was a hybrid, mixing freedoms and elements of electoral competition, on the one hand, with clear elements that subverted the democratic process, on the other. Hopes for a democratic future in Russia could be maintained, though, because the situation was fluid, indeed chaotic. Unfortunately, no democratic progress was made during Yeltsin’s second term as president, from 1996–1999. As Reisinger (2003, 273) noted, “The [1999–2000] federal elections were not more cleanly run, not better served by mass media, not better organized by strong parties, and not more a forum for national debate than those that came earlier.” The arrival of Putin accelerated the diminishment of democracy. Colton (2005, 103–4) wrote of democracy’s “attenuation”: “RussiaВ .В .В . has significantly regressed and is by standard measures further removed from being governed

democratically than it was at the beginning of the 1990s.” Gitelman (2005, 255) argued that “Putin’s intolerance of political competitors and growing appetite for control have moved Russia further away from consolidated democracy, at least for now.” In the 2000s, it therefore became common to label Russia’s political regime as a hybrid on the authoritarian side of the spectrum: “electoral authoritarianism” (Ross 2005) or competitive authoritarianism (Levitsky and Way 2002). Stoner-Weiss (2010, 255) wrote of “Russia’s turn away from democracy,” dating scholarly consensus about this turn to 2005. By the late 2000s, Levitsky and Way (2010, 371) had changed Russia’s designation from competitive to “full” authoritarianism. As we show in figure 2.1, quantitative Page 5 →measures of the extent of democracy characterizing the world’s countries show Russia declining in a series of steps throughout the post-Soviet period. While Russia’s scores in 2014 exceed the world’s most repressive regimes, they are well below the global average. Russian politics clearly operates by authoritarian not democratic rules. Telling the story of post-Soviet Russian politics therefore involves understanding both why democracy failed to take root and how Putin’s authoritarian regime did take root. In this book, we highlight subnational dynamics as a way to enhance perspectives that remain only on the national level. Given the lack of democracy, how were Putin and his team able to establish a workable and stable regime from the political system and circumstances that Yeltsin bequeathed them? We provide a new perspective on the politics between the Kremlin and regional leaderships and its critical role in creating and strengthening the new regime.

Why We Study Elections In The Third Wave: Democratization in the Late Twentieth Century, Huntington (1991, 174) concluded that “The lesson of the third wave is elections are not only the life of democracy; they are also the death of dictatorship.” Published in the midst of communism’s collapse across Eastern Europe, this claim captured a prevailing optimism that the spread of competitive elections would give birth to a new age of democracy. While not all observers of international politics at the time would go so far as to agree with Fukuyama (1992) that the fall of communism heralded the triumph of Western liberalism, the notion that countries in transition from authoritarianism were also in transition to democracy underpinned the assumptions of many scholars and policy makers (for a critique, see Carothers 2002). Of course, authoritarian rule continued to exist and operate during this “age of democratization” (Brownlee 2007), but elections were commonly understood as instruments of democracy (Powell 2000). However, as time progressed and regimes that had been deemed democratizing failed to consolidate democracy, scholars began questioning the validity of assumptions, concepts, and theories from the comparative democratization literature. Drawing on experiences in Africa, Barkan (2000) considered whether a conceptual distinction exists between “protracted transitions” and unconsolidated democracies. For him, a unique configuration of structural conditions and constraints that includes short, “imposed” prior experience with democratic rule and agrarian economies Page 6 →differentiated African countries transitioning from authoritarian rule from most other cases, both historical and contemporary (ibid., 228–31). Half a decade earlier, Bunce (1995) challenged the ability of what Carothers (2002) would later call the “transitions paradigm” to apply to postcommunist cases. For Bunce, the end of state socialism differs from earlier cases of democratization in Latin America and southern Europe in meaningful ways that area studies scholars were best positioned to address (e.g., the nature of the previous authoritarian regime). Even earlier, at least one Soviet expert bucked the democratization bandwagon: Roeder (1994) depicted the establishment of post-Soviet regimes primarily as a choice among authoritarian options: autocracy (e.g., Azerbaijan, Russia), oligarchy (e.g., Georgia), and exclusive republic (e.g., Estonia). While a more liberal regime was a possibility with Armenia and Moldova (1990–1993) categorized as balanced republics (ibid., 66), Roeder’s perspective clearly differed from Huntington’s by recognizing that the onset of competitive elections alone does not signal the end of authoritarian rule. Still, elections are fundamental to how democracy operates. From Schumpeter (1942 [2003], 269) on, the holding of competitive popular elections to fill the principal positions of political power has been offered as a “minimalist” standard of democracy. Even those who value higher standards for a democracy see competitive elections as a critical starting point. For example, Dahl (1998, 85) lists elected officials and free, fair,

and frequent elections as the first two of the six institutions required for democracy. Powell (2000, 4) explains the importance of elections as follows: Elections are not the only instruments of democracy. They must be helped by other organizations and by rules that encourage communication and cooperation. But elections seem to be the critical democratic instruments. They claim to establish connections that compel or greatly encourage the policymakers to pay attention to citizens. There is widespread consensus that the presence of competitive elections, more than any other feature, identifies a contemporary nation-state as a democracy. More recent discussions, and by non-Americans, accept this conception as well. Merkel (2004) lists elections as the first of five areas that must meet democratic standards if a regime is to be considered democratic. Moeller and Skanning (2010, 271) begin their typology of democracies with competitive elections, then free and fair elections, followed by civil liberties and Page 7 →the rule of law. The indicators we will use in our analyses, therefore, track what we see as the core of democracy—competitive elections. Some analysts, moreover, view the holding of elections as itself a mode of transition in which de jure competitive elections provide the institutions, rights, and processes to advance democratization beyond minimal standards (Lindberg 2009, 9). Under other circumstances, though, elections may forestall democratic advancement or even open the door for a return to authoritarianism. According to Karl (1986, 9), “elections can be used domestically to arrest liberalization when a regime employs them as a means of limiting basic individual liberties and collective rights.В .В .В . Carefully staged вЂdemonstration’ elections can, moreover, ratify existing power arrangements.” It is thus important to understand how a country’s elections actually work. Do they strengthen or undermine competition, individual rights, and the operation of formal institutions? Elections also matter in nondemocracies. By the start of the twenty-first century, regimes combining competitive elections with autocratic practices had emerged as the most common type of authoritarianism (Diamond 2002). As democracies grew in number during the final decades of the twentieth century, with democratic ideals attracting increased public support on every continent, authoritarian regimes needed to adopt a “democratic disguise” (Brooker 2013, 9). Elections have been the central component of this strategy. To be effective—that is, if these elections are to provide the regime with a mantle of democratic legitimacy at home and abroad, or even just “plausible deniability”—they must permit multiple candidates or parties to compete, including some that can plausibly be seen as in opposition to the incumbents. Regimes that operate such elections have been analyzed using several partly overlapping terms, including electoral authoritarianism (Schedler 2006) and competitive authoritarianism (Levitsky and Way 2010).1 Electoral or competitive autocracies are therefore distinct from “closed authoritarian” regimes (Howard and Roessler 2006; Levitsky and Way 2010), which either have no elections for top positions or the elections are entirely uncompetitive. At the same time, electoral authoritarian regimes are not democracies. The incumbent leadership goes to great effort so that an opposition victory is highly unlikely. Elections in these regimes, despite presenting voters with multiple parties or candidates to choose from, are sufficiently unfair and fraudulently conducted that the regime’s preferred candidates are almost certain to prevail (on the many techniques available, see, e.g., Calingaert 2006; Case 2006; Alvarez, Hall, and Hyde 2008; Birch 2011; Schedler 2013, ch. 2; van Ham 2015). Page 8 →Nonetheless, elections are highly consequential in electoral autocracies for two reasons. One is that the regime may miscalculate or poorly carry out its tasks, making elections the spark for a significant challenge to the regime (Howard and Roessler 2006; Tucker 2007; Bunce and Wolchik 2011; Hyde and Marinov 2012; Kaya and Bernhard 2013). The other is that managing the competitive but unfair elections changes how the regime operates. In Schedler’s (2013, 1) words, even though such regimes “betray the democratic spirit of plural elections, they change their own inner dynamic by admitting them.” Recent scholarship (summarized in Gandhi and Lust-Okar 2009, 405) investigates how authoritarian elections may serve the regime: by coopting elites by distributing spoils in a comparatively fair or efficient way; by attracting to the regime those able to connect with the public; by dividing the opposition by giving some a place partially on the inside of the tent; by learning more about where support and opposition to the regime are concentrated; by demonstrating the regime’s strength

through a dominant electoral victory; by providing a safety valve by allowing regime opponents to express discontent; and, should worse come to worse, by making it more likely the autocrat will survive a collapse of the regime. “The coexistence of meaningful democratic institutions and authoritarian incumbents creates distinctive opportunities and constraints for actors, which—in important areas of political life—generate distinct patterns of political behavior,” argue Levitsky and Way (2010, 27). Elections, then, operate across both democratic and authoritarian contexts, providing an admirable vehicle for studying the political trajectories of Russia and its constituent regions. Elections constitute a common thread running through the fabric of post-Soviet Russian politics. As we show in chapter 2, even as the outcomes of Russia’s elections grew more and more certain, their role in the regime’s political evolution remained strong. In subsequent chapters, we analyze election results to reveal variation in opportunities, constraints, and behavior across time periods and regions as well as to link these differences back to national developments.

Subnational Politics and National Regime Dynamics Two decades after the 1991 dissolution of the Soviet Union, we have few agreed-upon conclusions about why Russia partially democratized under President Yeltsin only to revert to more authoritarian practices under presidents Putin and Medvedev. Although passionate proponents of different explanations exist, no one view predominates. One reason for the Page 9 →lack of consensus is Russia’s complexity, spanning 11 time zones and consisting of over 100 nationalities. Indeed, just as Russia’s national political trajectory has changed over time, so too have the political trajectories of Russia’s regions. Our goal is to link regional political trajectories and the national one. We submit that examining territorial dimensions of power can shed new light on how the dynamics of political control change over time (see also Gibson 2012), and that these developments are fundamental to understanding uncertain transitions from authoritarian rule (O’Donnell and Schmitter 1986). Comparative political science was slow to incorporate subnational politics. Addressing countries as the objects of comparison reflected the importance of states in world affairs as well as theoretical interest in understanding how institutions and cultures differ among countries. Of course, scholars have long noted that a country’s politics flow from factors that vary subnationally (as well as transnationally). Dahl (1971, 12–13), for example, began his study of comparative contestation and participation by expressing his regret that he could not incorporate subnational variety. Even long after Dahl was writing, good data on subnational politics remained less readily accessible, and only recently have statistical techniques been developed that allow a scholar to analyze multiple, nested levels of data. The latter problems are being continually redressed, with the result that, over the last roughly 15 years, attention to subnational dynamics has blossomed. Two big advantages flow from the enhanced ability to study subnational politics. One is that scholars can pursue comparative questions by analyzing multiple subnational units within a single country (US states or German lГ¤nder, for example), and doing so may be more practical in terms of cost or theoretical controls than analyzing multiple countries. The other is that politics in all countries, whether formally federations or not, is shaped importantly by subnational variation and trends. Assigning one value of a variable to a country may hide more than it reveals, particularly if that value is simply the average of subnational or individual-level data (Snyder 2001). Decomposing the nationwide value is not valuable solely for what one learns about subnational units. It also can reveal mechanisms driving changes in national characteristics. Within-country variation can take many forms, of course, not necessarily the boundaries of subnational units. Even so, we argue that those boundaries are critical for understanding a wide range of phenomena and are particularly valuable to study with regard to democracy and authoritarianism. Whether states are federal or unitary, they organize state administration geographically, giving structure to formal and informal politics. Page 10 →This in turn infuses subnational patterning into whatever are the key sources of strength for the national regime. In democracies of any size, voting and therefore control of national institutions proceeds subnationally. For aspiring democracies, localities will differ in the extent to which they have made the transition from authoritarianism. For example, as we will show, Russia in the 1990s was composed of meaningful subpolities that varied in their regime type, from electoral democracies (e.g., Nizhnii Novgorod) to uncompetitive authoritarian

regimes (e.g., Tatarstan). As Gibson (2012) notes, regime juxtaposition is politically consequential in democratic states as tensions emerge when local and national interests fail to coincide. Yet even for an authoritarian regime without competitive elections, the geographically based subdivisions of the ruling party and state agencies make subnational dynamics valuable to study. For analyzing authoritarian regimes with unfair but competitive elections (Schedler 2013), among them Russia under Putin, election results at the subnational level are vital, and we must consider the degree to which subnational authoritarian enclaves may facilitate a national authoritarian revival.

The Gains from Statistically Analyzing All of Russia’s Regions Our investigation of the regional roots of Russia’s national regime employs data on all of Russia’s regions. Although the number of Russia’s regions has varied slightly over time (see chapter 2), its composition of over 80 units, and the multiple elections each has held, allow us to make rigorous comparisons using statistical methods. We utilize statistical techniques to test hypotheses from the literatures on both democratic and authoritarian elections. Since these literatures expect different types of behavior, we can leverage them to determine whether general trends across Russia’s regions conform to authoritarian or democratic contexts. We also can identify regional characteristics that lead, or at least permit, certain regions to defy the federal norm—that is, to establish regional authoritarianism despite a national democratic trajectory or to preserve political competitiveness in an otherwise increasingly hegemonic context. In other words, what factors make some regions more vulnerable and responsive to national pressures and demands than others? In using Russia’s relatively large number of cases to test the applicability of comparative hypotheses, we do not neglect that subnational administrative units represent important contexts in their own right. Their locations; their particular combinations of economic, social, and cultural characteristics;Page 11 → their histories—these create filters for how the national derives from the individual or local. This explains the call from political geographers to take context seriously and view a region as more than the sum of its components (O’Loughlin 2000, 2003; Goodin and Tilly 2006). In Russia, these different configurations have influenced, and been influenced by, the evolution of federalism. Thus, alongside hypotheses drawn from general theories in comparative politics, we examine context-specific hypotheses drawn from extant work on Russian politics. For example, we ask whether electoral competitiveness varied systematically across Russia’s regions in accordance with the different constitutional statuses in use. Also, did the opportunity for some regions, namely Russia’s republics, to hold gubernatorial elections earlier than other regions have a systematic effect on the consolidation of authoritarian enclaves? And, if so, to what extent did these enclaves resist democratization under Yeltsin or facilitate a renewal of authoritarianism under Putin? Answers to these questions help us determine whether longitudinal developments in Russia’s regional elections factored into President Putin’s 2004 decision to eliminate gubernatorial elections as well as the degree to which the appointment era proved amenable to pro-Kremlin results in national contests during Putin’s second term in office, as well as Medvedev’s presidency. Our analysis, then, spans the staggered onset, institutionalization, and elimination of Russia’s gubernatorial elections and continues through the establishment of an appointment system, one that lasts until Putin’s return to the presidency in 2012. We intentionally limit our analysis to Russia. As Collier et al. (2004, 204) stress, contextual specifics should guide one’s concepts and measures whether comparing across nations, across disaggregated subunits, or over time.2 Focusing only on Russia allows us to identify relationships that could easily be missed in a study that aggregates Russia’s subnational units into a larger cross-national investigation of subnational politics. It also allows us to hold constant a host of “thick” concepts (e.g., political culture, national institutional configurations) that large-N, cross-national studies must measure, sometimes with relatively “thin” indicators (Coppedge 1999, 469). To be fair, our quantitative analyses also occasionally must rely on narrow indicators of thick concepts. Still, conducting statistical analyses of all subnational units within a single country over time offers a practical solution: one that combines advantages from both large-N and small-N approaches (ibid., 475). It is worth noting that our interest in context applies as much to temporal specificity as it does to geographic specificity. While elections play critical roles in both democratizing and electoral authoritarian regimes, Page 12 →the legitimacy of the Soviet regime was based on ideology, performance, and coercion. So while our work

remains cognizant of Soviet history and legacies (cultural, economic, and social), it makes little sense to extend our statistical analysis back in time to the Soviet era. Yet, just as we disaggregate the Russian Federation into meaningful geographic subunits, we also disaggregate the post-Soviet era into meaningful time increments. For example, we intentionally analyze gubernatorial elections held during the Yeltsin era (chapter 3) separately from those held during the Putin era (chapter 4) based on the belief that a change of presidents in Russia constitutes a significant change in the country’s political context. Similarly, our survival analysis of gubernatorial appointments (chapter 5) explicitly distinguishes Putin-era from Medvedev-era appointments. At other points in the book, we develop time-specific hypotheses (see Pierson 2004, 12–13), including “first-mover” advantages (chapter 3) and the possibility of learning or adaption (chapter 8). Thus, although one could delve deeper into the thicket of regional politics than we do,3 we believe that our approach strikes a reasonable balance between generalizability and specificity.

Drawing from Comparative Theories of Democratization and Authoritarianism As we pursue the two related questions noted above, why Russia’s democracy failed to take root and how Putin’s authoritarian regime arose, we will investigate propositions from the literatures on democratization and authoritarianism, which have mostly developed separately. We will examine both cross-sectional comparisons of Russia’s regions and over-time comparisons of patterns and relationships. Our task thus requires theoretical eclecticism. In our analytical chapters, we explain the reasons for the specific explanatory variables we analyze. Here, though, we can present overviews of the two relevant literatures. Democratization The broadest distinction in comparative democratization theory is between structural preconditions for, or prerequisites of, democratization on the one hand and political agency, including the choice of institutions, on the other (e.g., Reisinger 1997; Grzymala-Busse and Luong 2006; Haerpfer et al. 2009). The analytical pros and cons of these two camps are well known, thanks to a long-running debate begun by Rustow (1970). Beyond this Page 13 →distinction, the proposed influences range widely. Jepsen (2011) finds 22 distinct factors that fall into six categories: modernization; economic preconditions; social preconditions; timing, sequence, and politics; agency and advocacy; and external effects. Malone (2011) reviews seven groups of theories, those stressing economic development; elites; political culture and civil society; education; state institutions; national unity or social divisions; and international context. Most scholars now acknowledge that the various approaches are not mutually exclusive (Grzymala-Busse and Luong 2006) and frequently combine explanatory factors from multiple approaches. Almost all the factors that have been put forward as influences on democratization have been examined by one or more analysts of post-Soviet Russia. More work is needed, however, to understand what Russia’s trajectory tells us about comparative theories. Excellent studies of post-Soviet Russia disagree on the influence of the same factor, be it political culture, the oil curse, institutional choices, or others. It also happens that factors agreed to be relevant comparatively can be hard to assess in the Russian case. For example, comparative research strongly supports that a country’s level of socioeconomic development, or modernization, is related to democracy, although a debate continues about whether rising development promotes democratization or only stabilizes existing democracies (Przeworski et al. 2000; Boix and Stokes 2003; Epstein et al. 2006; Houle 2009; Kennedy 2010; Teorell 2010). As we show in chapter 2, Russia as a whole is either just below or clearly above the level of national income usually associated with stable democracies. Whether its nationwide level of socioeconomic development is relevant is thus murky at best. Because Russia’s subnational regions differ quite sharply in their level of development, our analyses may illuminate this factor in a way that the national level cannot. Our analytical chapters employ measures of regional economic and social development to control for this possibility. At the same time, however, we stay attuned to the possibility that regions’ economic and social structures may impede democratic outcomes. For example, since regions that were wealthier at the onset of Russia’s transition likely achieved this status thanks to the advantages accrued under the previous regime, one might reasonably expect authoritarian continuity. Such an outcome should be even more likely in regions rich in natural

resources as the distribution of rents may be substituted for political representation (e.g., Mahdavy 1970; Herb 2005). We also consider the importance of short-term strategic decisions and performance. Drawing on the work of transitologists (e.g., O’Donnell and Schmitter 1986), for example, we examine whether decisions made Page 14 →in the midst of Russia’s transition (e.g., the timing of regional elections) facilitated or impeded the emergence of competitive politics in certain regions. Similarly, since elections are generally understood as providing voters with an opportunity to hold governing elites accountable for public policies, especially those that influence their economic well-being (e.g., Fiorina 1981), we look for evidence of retrospective voting: Did voters in regions where average incomes, for example, were on the decline punish regional incumbents, the national party of power, or presidential candidates of the party of power? At the same time, we look for instances where the relationship between regional performance and election outcomes is indicative of authoritarian mobilization: Did worse conditions actually favor incumbents, thus suggesting voter vulnerability rather than electoral accountability? Accounts of Russia’s democratic failure frequently point to multiple causes. For example, Fish (2005) cites three key factors: Russia’s abundance of and economic reliance on oil and other exportable resources, its failure to carry economic liberalization far enough, and the weak legislature in the 1993 constitution. Analyzing Russia at the nationwide level, unless done as part of a cross-national study, makes it very difficult to sort out each factor’s relative impact. Examining subnational patterns, then, not only makes this possible while holding constant characteristics shared by all the regions as constituents of the same state, but it also helps move the study of democratization beyond what Gibson (2012, 10) labels a “nation-fixation,” where the answers found speak more about a country’s “best-known” places rather than the country as a whole. Authoritarianism Our other key question is why the current regime under Vladimir Putin took on the form it did. From 2000 on, Russia’s political regime has allowed significant—albeit declining—individual freedom and retained the basic institutions of a democratic polity—multiparty or multicandidate elections for many, although not all, significant political offices, and an elected legislature—while eliminating all political opposition of substance. The major tools with which the regime maintains its authoritarian control, many beginning prior to Putin, include a large executive-branch staff that duplicates functions assigned by the constitution to the cabinet (Chaisty 2005; Fish 2005, ch. 7; Sakwa 2011); state ownership of major business entities, especially in the energy and natural resources sectors (Fish 2005, ch. 6; Weinthal and Luong 2006; Shevtsova 2007, 118–31; Sutela 2012); Page 15 →control of television and most other major media outlets (Oates 2006; Mickiewicz 2008; Burrett 2011); strong control by the central executive over regional and local governments (Hyde 2001; Chebankova 2010; Ross 2010); and the development of a dominant political party, United Russia, capable of providing clear-cut victories for the Kremlin’s preferred candidates in federal, regional, and local elections (Smyth 2006; Gill 2012; Makarenko 2012; Roberts 2012a). Russia’s regime rests not only on its formal institutions but on a wide range of informal practices, with rules understood by those engaging in them (Wilson 2005; Ledeneva 2006; Smyth, Lowry, and Wilkening 2007; Sakwa 2011; Gel’man 2012). The party United Russia, for example, while dominant in the sense of maintaining the regime’s authority across a large country, is not a ruling party that could oust Putin; rather, power emanates from Putin and his executive-branch staff (Roberts 2012b). Russia’s formally independent court system remains susceptible, in high-profile cases, to behind-the-scenes pressure from political elites (Hendley 2009; Solomon 2010). State-sponsored organizations replace or undermine spontaneous civil society organizations and NGOs (Horvath 2011; Chebankova 2012). Personal connections among members of the political elite both create incentives and structure conflict, leading many to depict Russia’s regime as neopatrimonial (Huskey 2005; Lynch 2005; Van Zon 2008; Whitmore 2010; Pain 2011; Robinson 2012) or patronal (Hale 2010). Informal patron-client ties also support the regime far below the top-level elite (Robinson 2007; Solomon 2008; Turovskii 2009; Hale 2010). Corrupt informal practices are extensive (Dininio and Orttung 2005; Wedel 2005; Cheloukhine and King 2007; Holmes 2008; Frye 2010).

The importance of complicated informal ties makes Putin’s role vital. Many aspects of Russian politics fit with what theorists call a personalist regime, in some ways resembling the extreme form known as “sultanism” (Linz 1975, 259–60; Linz and Stepan 1996; Geddes 1999). Putin’s role in setting policy and balancing among informal groups, the emphasis placed on preserving his options on any policy matter, and the lack of any rules for leadership succession make personalism a hallmark of Russia’s current regime (Gill 2012; Isaacs and Whitmore 2014; Baturo and Elkink 2015). The literature on authoritarianism and on systems that combine authoritarian and democratic features is not as large as the work on democratization but is nonetheless strong and has been growing rapidly (e.g., Magaloni 2006; Wintrobe 2007; Gandhi 2008; Levitsky and Way 2010; GГ¶bel 2011; Hale 2011; Webb 2011; Kailitz 2013; Schedler 2013). Recent work has sought to better understand how ostensibly democratic institutions such as Page 16 →legislatures, elections, and parties perform valuable functions for authoritarian regimes. The comparative literature has added much to our understanding of policy dynamics within authoritarian regimes, including about elections, as we discuss below, but has less to say about our question of how and why a certain type of regime emerges over time in a country. Surprisingly, in our view, the comparative literature on how subnational politics fits into nondemocratic regimes remains scant, over a decade after Snyder’s (2001) argument for such attention. Studies of subnational authoritarianism in Russia are numerous, by contrast (see Reisinger 2013, 12–13 and 18–20), though most have aimed to illuminate the regions’ influence on sovereignty and economics (e.g., Treisman 1999; Kahn 2001; Kahn 2002; Herrera 2005; Stoner-Weiss 2006; Sharafutdinova 2010a). We join a smaller number of analysts who specifically seek to trace the role of subnational dynamics in Russia’s overall political trajectory (e.g., Golosov 2011b, 2011a; Gel’man 2013). We focus on the rise and preservation of regional political regimes and how they interact with one another as well as with the national regime. Specifically, we argue that a country’s political trajectory reflects, if not depends on, the degree to which its constituent regions implement or resist the political agenda of the ruling national elite. Our analysis highlights the persistence of authoritarian enclaves during Russia’s experimentation with electoral democracy in the 1990s as well as the survival of democratic enclaves despite Russia’s authoritarian resurgence in the Putin era. By examining Putin’s regime as resting on relations between the Kremlin and the regions, we can see its development more clearly.

Structure of the Book and Main Findings In chapter 2, we provide a narrative overview of post-Soviet Russian political developments. It provides the reader with the background and context for the empirical analyses in subsequent chapters. In particular, we explain the key features of each federal election as well as how regional governors switched from appointed to elected to appointed again to elected once more. We also describe how the formally democratic institutions failed to take hold and the path by which Putin’s nondemocratic regime was established. The chapter concludes by placing Russia’s trajectory in comparative perspective, drawing explicit contrasts to other countries that transitioned from communism in the late 1980s and early 1990s and taking into consideration approaches that focus on interval-level rankings and Page 17 →regime categorizations. It is notable that a powerful correlate of democracy worldwide—the level of economic development—provides little help in understanding Russia’s trajectory since independence. Chapter 3 is the first of two chapters in which we analyze results from gubernatorial elections. This chapter examines elections in the 1990s while Yeltsin was president. We analyze the variation among the regions and over time in two features of these elections that bear on democracy: how competitive were the elections in the sense of the incumbent facing the potential of losing and did an incumbent in fact lose? The patterns we find show that a region’s economic performance failed to systematically influence competitiveness, while higher levels of socioeconomic development appear to have dampened competition rather than promoted it. On the other hand, the availability of natural resources, as expected, is associated with less competition. In addition to these structural effects, agency also matters. As scholars of institutional design would expect, regions with runoff electoral systems had significantly higher numbers of candidates. While controlling for ethnicity, we demonstrate that an interaction effect between constitutional status and election timing shaped competition levels in the 1990s. Not only did electoral competition vary among Russia’s republics, but that variation reflected differences in the

opportunity and willingness of regional leaders to hold founding elections early on in Russia’s transition: holding founding elections early was something that republic leaders could do, but not all of them did. Chapter 4 then examines the gubernatorial elections held from 2000 to 2005, during the period when Putin was working to rein in the autonomy of regional governors in a variety of ways. Again focusing on whether an incumbent was turned out, we examine Kremlin efforts to unseat governors not seen as loyal or to support those who were. The dynamics of Russia’s gubernatorial elections were no longer systematically influenced by natural resource wealth, levels of socioeconomic development, violent crime, or electoral rules. Rather, gubernatorial elections during the early years of Putin’s reign functioned more like a game played by elites. Kremlin opposition to incumbents spurred more effective candidates—a measure of the election’s competitiveness calculated from each candidate’s vote share—and the number of effective candidates alone determined the fates of the incumbents. Although the Kremlin was able to influence the competitiveness of the field in some cases, it overall struggled to use elections, specifically, as a tool to shape the gubernatorial corps. That difficulty paved the way for the elimination of gubernatorial elections from 2005 to 2012. Chapter 5 completes our investigation of changes in the composition of Page 18 →Russia’s gubernatorial corps by examining the Kremlin’s choice of retaining a governor in office or removing the governor, a question for which the appropriate statistical method is survival analysis, also known as event-history analysis. We employ this technique to examine all gubernatorial tenures during the appointment period from 2005 to 2012. We show that the governors’ political effectiveness, specifically their ability to provide large vote totals for Kremlin candidates or parties in federal elections, was the predominant factor in Kremlin decision-making. The region’s size, ethnic composition, and economic performance have little effect. The appointment system thus reached its evident goal: to incentivize governors to show political loyalty toward the Kremlin, particularly by delivering the votes in federal elections. Nonetheless, the results in the 2011–12 cycle were disappointing to the Kremlin, and it gave up its appointment power to search for new methods of regional control. In chapters 6 through 8, we turn from gubernatorial elections to federal elections. The federal elections are for the president and the lower house of the national legislature. Different election results across the regions reflect important political processes at the regional level. In chapter 6, we examine patterns of turnout: the proportion of each region’s eligible voters who vote. Turnout levels have been a sensitive matter because election results based on high turnout are considered more legitimate, as well as because turnout below a minimum threshold invalidated the election. On the other hand, turnout levels above a certain level reflect coercion or falsification on the part of regional authorities. Our analyses indicate which regions built political machines sufficient to generate high turnout and when that occurred. In the first four federal elections, the high-turnout regions are those with more elderly and more sympathy with the Communist Party. This suggests that individual voters faced less elite pressure in most regional settings, allowing those who did not wish to vote to abstain. In subsequent elections, into the early 2000s, age and ideology begin slipping as correlates of regional turnout. From 2003 on, they have virtually no effect. Instead, the factor most strongly associated with high turnout becomes the percent of a region’s population that is non-Russian. The ethnic regions located in Russia’s North Caucasus lead the way, but those located elsewhere also see remarkable turnout levels. By contrast, regions located in Russia’s northwest area, predominantly nonethnic regions, anchor the lower end of the turnout distribution in most elections. In chapter 7, we analyze cross-regional patterns in the extent of pro-Kremlin voting in each federal election. We take into account both the share of votes the Kremlin’s party or candidate received and the level of Page 19 →turnout by analyzing pro-Kremlin votes as a proportion of all eligible voters in the region. With this measure, we identify regions where the authorities are willing and able to generate artificially high levels of support for the Kremlin’s party or presidential candidate. We find that some regions are consistently predisposed toward more authoritarian rule. Not only did these successfully resist the encroachment of democratic practices, they possessed electoral machines that could be subsequently mobilized for the cause of authoritarianism when the national elite redirects the country’s political path. In a manner frequently neglected, the voting returns from this minority of regions with effective machines had a disproportionate impact on the nationwide outcome. Even when the region’s population is comparatively small, providing the Kremlin with the votes of 90% of every

adult resident has an impact. By making common cause with the leaders of these regions, the Kremlin ensured victories during Putin’s early years. Yet the country also possesses regions that provide fertile ground for democracy, eluding the ability of their governors’ machines to regiment their voting. These regions serve as examples of democracy’s potential to thrive under the right circumstances, even in countries that have become authoritarian. In chapter 8, we address the issue of influence among the regions. We use spatial analytic techniques to explore whether regional electoral deference to the Kremlin spread in the 1990s and 2000s through a diffusion process. Did some sets of regional authorities draw lessons from others in identifiable ways to help explain the trends over time? We find that deferential election results were much more likely in regions with larger non-Russian nationalities. Election results in regions with more access to resource rents, like Russia’s gas-rich regions, were not demonstratively more deferential. Governors who experienced less competitive gubernatorial elections in the contest preceding a presidential election were significantly more likely to deliver results that were deferential to the Kremlin. This effect continued into the appointment era, suggesting that the Kremlin appears to have kept around governors with demonstrably greater electoral prowess, ostensibly for its own advantage. Finally, even after we control for economic and social characteristics often associated with determining elections, proximity to more deferential regions made deference to the Kremlin more likely. When it was appropriate to label Russia’s political regime a hybrid, it functioned to a large degree as a spatially distributed hybrid of politically competitive and uncompetitive regions. Our analysis not only shows that electoral deference to the Kremlin was more likely in regions with more non-Russians, but that its spread over time corresponded with important spatial dimensions. Page 20 →Ultimately, Russia’s experience suggests that the rise of many competitive authoritarian regimes rests in the hands of what scholars of American elections call “swing states.” An initial goal of the Putin administration was to control election outcomes in these swing states. First, the Kremlin sought to consolidate control over the regions through the existing mechanism of gubernatorial elections. The limitations of this approach led to replacing popularly elected governors with presidential appointees. The gubernatorial appointment process gave the Kremlin direct leverage over regional regimes, and the Kremlin used that leverage to reorient the country’s political trajectory. This reorientation is most evident during national elections. The Kremlin could and did reward or punish appointed regional executives based on their regions’ level of support during national contests. Yet we also emphasize that the new level of organizational power that gubernatorial appointments granted the Kremlin was neither uniform nor unobstructed. Factors that vary by region constrain the center’s control, and the consequences of these constraints came to the fore during the 2011 Duma elections as a dozen regions proved less deferential than they had been previously. So, while the Kremlin’s control over Russia’s governors includes those presiding over more democratically and more authoritarian inclined regions, the regions in the middle comprise the country’s political battleground. In democracies, winning in swing states merely determines the electoral victor. In competitive authoritarian regimes, like Russia, controlling the swing states dictates the regime’s trajectory. In our conclusion, we remind the reader of our major findings and how we came to them in the preceding chapters. In regions where the elections were more democratically conducted, we demonstrate that Russian voters behave in ways that correspond to expectations. Looking only at nationwide electoral results, however, obscures this because of regional variation in whether voters have been able to vote their wills. In the first half of the 1990s, when free, fair, and competitive elections were more widely spread, the Yeltsin administration proved unable to use them to solidify electoral-based power. Instead, the elections prompted Yeltsin to promote noncompetitive elections in regions willing to support him. Putin pushed this model nationwide, in a process we illuminate in chapter 8.

Page 21 →

Two Post-Soviet Russia’s Political Trajectory The formative stage for contemporary Russian politics extended from the last years of the USSR through roughly 1996. By that time, the boundaries of the new state had been set, the commitment to market economics was in place, a new constitution had been ratified, and an initial version of federal relations was in place, including gubernatorial elections. What failed to materialize over the succeeding years, however, was the widely hoped-for economic and social recovery or the democratization of politics. This failure set the stage for Vladimir Putin to construct a quite different, partially liberal but nondemocratic political regime. We begin this chapter with a look at the diversity characterizing the regions comprising the Russian Federation. Then, because so many of Russia’s major post-Soviet clashes and changes are connected to federal elections, we provide an overview of Russia’s federal elections over the entire period. We follow that by describing the changing framework for electing Russia’s regional leaders; we analyze the results of these elections in chapters 3 and 4. We then examine Russia’s political transition and change under Yeltsin and Putin from a comparative perspective, showing that Russian politics is more authoritarian than its location and socioeconomic characteristics would predict. Unlike most postcommunist countries, where the level of democracy has either stayed at approximately the same level from the early 1990s on or has seen its democracy level increase, Russia has experienced repeated declines in its democratic features. Page 22 →

The Regions of Russia The Russian Federation is the world’s largest country by territory, its over 16 million square kilometers of land mass exceeding runner-up China’s by three-quarters. If you stand on the coast of Maine, you are closer to Moscow than if you were standing in Vladivostok (Richardson 2009, 30). In the course of its expansion, the Russian Empire became home to numerous nationalities and ethnic groups besides Russians. The Empire’s administrative division was not federal but territorial, divided into dozens of “governates” having little or no formal autonomy. When the Bolshevik Party took control of most of Czarist Russia’s territory, its leader, Vladimir Lenin, decided that a federal structure was the answer to “the nationalities question,” that is, to the potential for separatism by non-Russian ethnic groups (Gleason 1990, 31–33). Thus, the country became a union of four republics in 1922. The largest of these was the Russian Republic, officially known as the Russian Soviet Federated Socialist Republic (RSFSR). Other parts of the former Russian Empire were added to the union in the succeeding years (Fainsod 1953, 307–11). Within the larger republics, including the RSFSR, certain territorial units received special status recognizing a nationality group: autonomous socialist republics, autonomous oblasts, and autonomous okrugs. When Russia became independent at the end of 1991, Soviet-era territorial delineations remained largely in place. The republics lost the designation “autonomous socialist,” and the republic of Chechnya-Ingushetia split into two. The autonomous okrugs and oblasts were placed on the same formal level as the republics. The “autonomous” regions are so called because each is located within a different region. The result was 89 regions or “members of the federation.” In the mid-2000s, six small autonomous okrugs were merged into their surrounding regions. Maps 2.1 and 2.2 show the 83 Russian regions as of 2012, the final year of the elections we analyze. Two regions were added in 2014 with the annexation of Crimea. Today, each of the resulting 85 regions has one of six statuses: oblast (46), krai (9), city of federal significance (3), republic (22), autonomous okrug (4), or autonomous oblast (1). A region’s status is part of its official name: for example, Tomsk Oblast or Zabaikalskii Krai. The cities of federal significance are Moscow, St. Petersburg, and, since 2014, Sevastopol.

Each of these cities is surrounded by a separately governed region (Moscow Oblast, Leningrad Oblast, and the Republic of Crimea, respectively). The republics and the autonomous regions are named after a non-Russian ethnic group—for example, the Tatar Republic—andPage 23 → tend to have relatively small proportions of ethnic Russians. (The newly annexed Republic of Crimea, however, departs from both of these principles—its residents are predominantly ethnic Russians, and Crimea is not the name of an ethnic group.) At the time of Russia’s independence, the RSFSR constitution of 1978 still operated. This constitution specified the legislature as the supreme authority within the RSFSR.1 It was as chair of the legislature that Yeltsin initially became the leader of Russia in 1990. The 1978 constitution also provided for a government, or Council of Ministers, headed by a prime minister chosen by the legislature, as well as a Supreme Court and subsidiary judicial bodies. From 1989 to 1992, a series of amendments were adopted, including one that created the post of president. At the Soviet level, the post was elected indirectly, and was filled by Mikhail Gorbachev. In June 1991, at what was then the subnational level, Russia’s president was elected directly by popular vote, which led to Yeltsin’s victory. Even with amendments, however, the Soviet-era constitution that Russia inherited did not adequately resolve the tension between the “supreme” status of the legislature and the newly created presidency. As we describe below, a new constitution resolving this tension in favor of the president came into force in December 1993 (Belyakov and Raymond 1994; Metcalf 1996; Sharlet 2003). It provided for a bicameral legislature, a president, and a prime minister chosen by the president but approved by the legislature. It created, in other words, a formally semi-presidential system. Except in 1993, legislative elections have only been for seats in the lower house of the legislature, the State Duma. Thereafter, seats in the upper house were ex officio for regional leaders or chosen by those leaders. The 1993 constitution gives strong powers to the president. Although the constitution did provide mechanisms for checks on presidential power, they were much weaker than the president’s tools for controlling the other branches. Using the coding rules developed by Shugart and Carey (1992, 150) to classify national executives’ formal powers, Metcalf (1996, 142) calculated that Russia’s presidency scores 8 for powers to create or influence law (legislative powers) and 10 for nonlegislative powers such as naming the cabinet or dismissing the legislature. The French semi-presidential system, by comparison, scores one for legislative powers and four for nonlegislative powers. Worldwide, while many countries give their presidents more nonlegislative powers than Russia’s, Russia’s score of eight for legislative powers is truly high: 88% of the systems in Shugart and Carey’s sample have lower scores than Russia. Many scholars argue that this institutional design Page 24 →is not so much semi-presidential as super- or even hyper-presidential (Frye 1997; Ishiyama and Kennedy 2001; Colton and Skach 2005; Fish 2005, ch. 7). Nonetheless, the interlocking presidential-parliamentary control of the government makes it more practical for the president to secure a supportive majority in the legislature. Elections to the federal legislature therefore matter for the Russian executive. Map 2.1. Western regions (Subjects of the Russian Federation) in 2012 (Source: Modified by authors from a map by Ezhiki available on Wikimedia Commons [https://commons.wikimedia.org/wiki/File:Russia_-_blank_map_(2008-01).svg].) Map 2.2. Eastern regions (Subjects of the Russian Federation) in 2012 Page 25 →

An Overview of Russia’s Federal Elections Russia’s formally competitive federation-wide elections span 1991 to 2012.2 The 1991 presidential election occurred while the USSR continued to exist, but we include it with post-Soviet Russian elections because it was not conducted in the manner of a Soviet-era election, and its outcome—Boris Yeltsin as president—remained when Russia became an independent country at the end of 1991. Federal elections for the presidency and the legislature have been closely tied to Russia’s broader political trajectory over the entire 1991–2012 period. In table 2.1, we summarize the federal elections, the winning

candidates and parties and major challengers. Presidential elections subsequently occurred in 1996 (two rounds), 2000, 2004, 2008, and 2012. Yeltsin won reelection in 1996. Putin became acting president at the start of 2000, Page 27 →when Yeltsin resigned, and was then elected to four-year terms in 2000 and 2004. In 2008, Putin’s close associate, Dmitrii Medvedev, was elected president. Since Russia’s constitution limits presidents only to two consecutive terms, Medvedev’s tenure permitted Putin to run for a third presidential term, which he won in 2012. In 2008, Russia’s constitution was amended to lengthen the presidential term from four years to six years. The changes did not apply to those currently in office (i.e., President Medvedev). The 2012 presidential election, therefore, was the final one on the previous four-year cycle. The next one is scheduled to occur in 2018. Page 26 → Table 2.1. Presidential Candidates and Key Parties in Federal Elections, 1991–2012 Victorious Main “Party of Presidential Other Key Parties Rival(s) Power” Candidate Boris Nikolai June 1991 Yeltsin Ryzhkov — — [57%] [17%] December Russia’s Liberal Democratic Party of Russia (LDPR) [23%] — — 1993 Choice [16%] Communist Party of the Russian Federation (CPRF) [12%] December Our Home is CPRF [22%] — — 1995 Russia [10%] LDPR [11%] Gennadii Zyuganov June (CPRF) Yeltsin 1996, [32%] — — [35%] 1st round Alexander Lebed [15%] July 1996, Yeltsin Zyuganov — — runoff [54%] [40%] Unity [23%] December Fatherland/All CPRF [24%] — — 1999 Russia (OVR) Union of Right Forces [9%] [13%] March Vladimir Zyuganov — — 2000 Putin [53%] [29%] December — 2003 March 2004

United Russia [38%]

Nikolai Kharitonov Putin [71%] — (CPRF) [14%]

December — 2007 March 2008

—

Dmitrii Medvedev [71%]

—

United Russia [64%]

Zyuganov — [18%]

CPRF [13%] LDPR [12%] Motherland [9%] — CPRF [12%] LDPR [8%] A Just Russia [8%] —

December — 2011

—

March 2012

Zyuganov — [17%]

Putin [64%]

United Russia [49%]

CPRF [19%] A Just Russia [13%] LDPR [12%] —

Note: Percentages indicate the percent of the total nationwide votes received. In the case of the legislative races, the figures are for the party-list voting only. Federal legislative elections occurred in 1993, 1995, 1999, 2003, 2007, and 2011, in each case in December. These elections determine membership in the lower house of Russia’s parliament, the State Duma.3 Through 2003, half the Duma’s members came from 225 single-member districts (SMD) with plurality voting, while the other 225 were selected through a proportional representation (PR) system with party lists and a single, nationwide constituency. For the PR voting, voters selected their preferred political party from those that made it onto the ballot. Each party receiving more than a minimum threshold of votes received seats in the Duma proportionate to its share of the votes. A federal law passed in 2002 raised the threshold for PR seats from 5% of the valid votes cast to 7%. In 2005, the law was again changed to eliminate the SMD voting. In the 2007 and 2011 elections, PR voting determined all 450 seats in the Duma. In the same reform that extended presidential terms, the length between legislative elections became five years. The most recent Duma election was held in September 2016.4 From 1991 to 2006, federal election law specified minimum turnout levels in order for an election or referendum to be valid (Sakwa 1995). For Duma elections, 25% of eligible voters needed to participate, while for presidential elections and referenda it was 50%. Another noteworthy aspect of Russian electoral law from 1991 to 2006 was that ballots included the option of selecting “Against All” rather than a candidate or party (on the legislative basis for this option, see McAllister and White 2008, 69–71). This option provided an outlet for protest voting by those who did not want simply to disengage from politics (for analyses, see Oversloot, van Holsteyn, and van den Berg 2002; Liubarev 2003; Hutcheson 2004; McAllister and White 2008). The 1993 law establishing the procedures for elections to the Duma specified that the initial elections in December of that year would be for a two-year term, with elections in December 1995 for what then became the standard four-year terms. With the next presidential electionPage 28 → scheduled for June 1996, this created a pattern in which the Duma elections occurred in the shadow of an approaching presidential race: six months ahead in 1995 /1996 and, thanks to President Yeltsin’s 1999 resignation, three months ahead in 1999/2000, 2003/2004, 2007 /2008, and 2011/2012. These legislative and presidential pairs form five electoral cycles through 2012—that is, until the 2008 changes in term lengths eliminated this pattern. Table 2.1 includes a column specifying one or more political parties as what Russians began calling the “party of power” in a given election. The essence of the term is when one of the parties on the ballot represents, or is perceived to represent, the political forces in control of the federal executive branch. In other words, the term connotes “the party of the powers-that-be.” The extent of executive-branch involvement with the party of power has varied. So, too, did whether the party leadership was based in the Kremlin (symbolizing the offices of the presidency) or in the White House (symbolizing the prime minister and the agencies comprising the government). Despite this variation, the appellation party of power has held meaning for both voters and elites during electoral periods. Only Unity and its successor, United Russia, proved successful in giving the president a solid base of legislative support. United Russia, of course, did more than that; it also became a hegemonic party, dominating not just the federal legislature but also those at the regional and local levels.

The Rise, Demise, and Partial Second Coming of Gubernatorial Elections When the Soviet Union collapsed in 1991, the Yeltsin leadership scrambled to establish governing institutions, including executive branches, for each of Russia’s regions. Under Soviet rule, power in the regions was formally parliamentary, with the head of the legislature listed as the top political authority of the region. Yet

Soviet political authority, especially executive authority, had not, in fact, been located in the legislature. Rather, a region’s political leader was the first secretary of its branch of the Communist Party (Hough 1969; Moses 1974). During the Gorbachev-era reforms of the late 1980s and the system collapse in 1990 and 1991, regional political authority became confused, with regional legislatures meant to have their executive committees strengthened and the Communist Party’s regional committee meant to play a lesser role (Willerton 1990; Moses 1992). Then, Yeltsin’s August 1991 decree banning the operation of the Communist Page 29 →Party on the territory of Russia—famously signed in front of a humiliated Mikhail Gorbachev—left the regions’ top politicians scrambling. Creating post-Soviet structures of regional executive authority and determining who held the top position were, therefore, more complicated than simply mailing the regional Party leaders name tags with the title “governor” and asking them to change the official stationery. Differences between republics in Russia and other regions became increasingly evident after Yeltsin’s populist call in Tatarstan in 1990 for regions “to take as much independence as they could swallow” (quoted in Kahn 2001, 377). Those republics that issued declarations of sovereignty prior to the Soviet Union’s collapse are commonly described as possessing greater leverage than other regions in their subsequent dealings with the center (Shlapentokh, Levita, and Loiberg 1997; Treisman 1997; Kahn 2002). Indeed, republics, as a group, were permitted to ratify their own constitutions, which in turn produced significant variation in their executive-legislative frameworks. While the differences included the number of legislative chambers (Moraski 2006c, 140–41), the most conspicuous, perhaps, is that some moved quickly to elect independent chief executives (often called “presidents”) directly, while others, at least initially, opted for parliamentary arrangements.5 In the second half of the 1990s, many republics preserved higher levels of sovereignty, while several other regions sought to establish their own, through a series of bilateral treaties with President Yeltsin. Yeltsin received the power to appoint executives for regions other than republics in late 1991, as a temporary measure to deal with the transition (Slider 1994; Clark 1998).6 However, he needed to make the appointments as quickly as possible (since economic chaos was erupting). He struggled to identify plausible prospects who would not spark a backlash from the region. In the end, a majority of those Yeltsin appointed had been first or second secretary of the regional Party (Tolz and Busygina 1997). Republics had the right to select their chief executives autonomously, usually through a vote of the regional legislature. In the early post-Soviet years, a few regions held elections for the governor’s seat (see chapter 3). Yeltsin did authorize an initial round of gubernatorial elections in April 1993, expecting that his appointees would win election. However, when five of the seven elections resulted in the defeat of his appointees (Solnick 1998, 50), Yeltsin postponed further elections until 1996. In that year, federal law made direct election of governors mandatory for all regions. From 1991 through mid-1995, 28 of Russia’s 89 regions selected their executive through popular elections (Tolz and Busygina 1997, 411). From late 1995 Page 30 →through 1999, 102 valid gubernatorial elections took place in 87 of the 89 regions. Of these elections, four held in fall 1996 were declared invalid and were reheld the following spring. Once elections became the norm from 1996 on, incumbent governors increasingly figured out how to use them to solidify their standing within the region. The extra legitimacy of having been popularly elected rather than chosen by Moscow gave many governors greater ability to oppose Kremlin policies and plans. The institution of gubernatorial elections, in the context of the chaotic 1990s, thus contributed to Putin’s commitment to strengthening the center’s control over the regions. In September 2004, Putin reacted to the terrorist tragedy in the town of Beslan in the Republic of North Ossetia (Phillips 2008; Harding 2012) by halting gubernatorial elections, effective in 2005. The Russian president gained the power to dismiss any governor and nominate a candidate for the vacancy. Although a candidate nominated by the president had to be approved by the regional legislature, United Russia dominated regional legislatures, at that time and since, making approval pro forma. In common usage, therefore, many describe the Russian president as “appointing” the governor. We examine the Kremlin’s use of this power in chapter 5. Following large-scale protests against electoral fraud in the December 2011 legislative elections, the Kremlin

announced that gubernatorial elections would be brought back. The legislature that resulted placed a limit on who could compete in the elections. Commonly referred to by Russian analysts as the “municipal filter,” candidates can be placed on the ballot only if they receive support from 10% of municipal legislators in the region. The municipal filter prevents those not in the elite from competing and raises a significant barrier even to challengers from within the elite. Five regions held gubernatorial elections in October 2012: Amur, Belgorod, Bryansk, Novgorod, and Ryazan. In each case, the sitting governor was easily reelected. In 2013, discussions began about eliminating the controlled gubernatorial elections for the republics, on the grounds that elections could exacerbate ethnic tensions and unrest.

The Multiple Transitions to Post-Soviet, Capitalist, Formally Democratic Russia The Russian Soviet Federated Socialist Republic (RSFSR) was one of the primary constituent units of the Soviet Union since the latter’s formationPage 31 → in 1922. The boundaries of the RSFSR at the end of World War II largely correspond to the borders of today’s Russian state. A noteworthy change came in 1954, when the Crimean Peninsula was transferred from the RSFSR to the jurisdiction of the Ukrainian Soviet Socialist Republic as a “gift.” While 51% of the Soviet population lived in the RSFSR, it also constituted 76% of the Soviet Union’s territory and recognized the most numerous and politically prominent nationality (USSR Central Statistical Administration 1983, 9 and USSR State Committee for Statistics 1989, 19). Yet, these features made Russia less distinct from the USSR as a whole and, ironically, less important as a level of governance than the 14 other union republics comprising the USSR. During the liberalization of the late 1980s, debates about Russian nationalism among intellectuals and cultural figures reemerged. By 1989, many in the Russian public had come to believe the Soviet regime was poorly serving their nation (Dunlop 1990; Szporluk 2000, ch. 7). Boris Yeltsin, in his rise to the leadership of the RSFSR, would use that belief as he sought to turn the Russian Republic into a sovereign state. In March 1989, Yeltsin achieved a dominant victory in the race to represent Moscow in the Soviet legislature. A year later, he won a seat in the RSFSR parliament as well. (Holding seats in both legislatures was allowed). Despite the best efforts of Soviet leader Mikhail Gorbachev, the Russian parliament then selected Yeltsin as its speaker, which at the time was the republic’s highest political office and, notionally, the head of state, even though internationally recognized statehood was two years off. In March 1990, at the same time as the elections to the RSFSR parliament, elections were held at the regional level within the RSFSR. The results of these elections reflected and to some extent congealed the regions’ quite diverse political climates. For example, case studies by Moses (1992) found two regions, Kemerovo and Volgograd, in which antiestablishment (reformist) elites dominated in 1991, other regions in which conservative party-state office holders could no longer dominate the region’s affairs (Kaliningrad, Kaluga, Novosibirsk, Perm, Tomsk, and Yaroslavl), but also groups of regions in which conservative officials continued to dominate (Belgorod, Bryansk, Kirov, Kursk, Orel, Rostov, Saratov, Smolensk, Tambov, Vologda). Though outside of Moses’s sample, leaderships in the republics and autonomous regions tended to fall into the latter category. Their Communist Party secretaries were mostly able to survive the turmoil of the early 1990s and secure governorships in post-Soviet Russia. According to Slider (1996, 243–44), local executives and enterprise directors dominatedPage 32 → many regional assembly elections. And, since regional chief executives often appointed these officials to their executive posts, their presence in the assemblies dramatically impeded the ability of these nascent institutions to check gubernatorial power. The regional variation in political competitiveness would be a critical component of future struggles that both presidents Yeltsin and Putin faced. The appeal of Russian sovereignty was not the sole source of Yeltsin’s popularity. Another was the support he received from pro-democracy Russians. Initially, members of Russia’s democratic movement distrusted Yeltsin. By 1990, however, as Gorbachev had moved in a conservative direction, the democracy movement came to see supporting Yeltsin as the best way to defeat the Communist Party’s hold over Soviet politics. In March 1991, in response to Gorbachev creating a presidency for the Soviet Union, Yeltsin oversaw a public referendum that supported the same post for Russia. Constitutional amendments in April then led to the first

election in June. To win, a candidate had to receive over 50% of the votes, either initially or through a runoff between the top two vote getters in the first round. The president would be the head of state and serve a term of five years. The amendments also created a vice presidency, filled by the person on the ballot with the winning presidential candidate. (The 1993 constitution would reduce the length of presidential terms to four years and eliminate the position of vice president.) Six candidates were on the ballot. Yeltsin’s most serious challenger was a former Soviet prime minister. Yeltsin received a majority, 59%, on the first ballot. An obscure candidate who outperformed expectations by receiving 8% of the votes was Vladimir Zhirinovskii, a hotheaded Russian nationalist (Fedarko and Aikman 1994; Kartsev 1995; Solov’ev and Klepikova 1995). Zhirinovskii and his political party, the Liberal Democratic Party of Russia (LDPR), have remained visible in Russian politics, though without gaining any direct influence on policy. The major issues for the RSFSR under Yeltsin’s leadership in 1990 and 1991 were economic reform and Russian sovereignty. While Yeltsin could do little to implement radical economic reform until Russia became independent, he committed himself to supporting it. Russian sovereignty drove politics during this period. In June 1990, the Russian parliament declared the republic to be sovereign, as some union republics had done earlier. Although somewhat short of declaring that Russia was independent of the Soviet Union, it came at a time when a few union republics were seeking outright independence and put the continued existence of a Soviet Union into question: “the idea of a Union without Russia was too absurd to merit Page 33 →a moment’s consideration” (Brown 2009, 554). For Hale (2005, 61), the actions by Russia, in particular, as the Union’s core ethnic region were primary determinants of the USSR’s dissolution. The RSFSR’s sovereignty declaration emboldened leaders of the autonomous regions within it to decide that they deserved sovereignty as well, especially to the extent that sovereignty implied regional control over profits from exploiting minerals and other natural resources within one’s territory. Yeltsin himself, as noted above, encouraged regional leaders to assert sovereignty. Famously, he said in August 1990 that the republics should “take as much autonomy as they could swallow.” Each of the 16 autonomous republics within the RSFSR thereafter passed its own declaration of sovereignty. For some republics, this act was pro forma and did not foreshadow conflict with the central Russian authorities after the end of the USSR. For a few, however, efforts to realize some level of genuine autonomy or sovereignty posed a major challenge to post-Soviet Russia and Yeltsin’s regime (on why some republics pursued this and others did not, see Giuliano 2011, esp. ch. 7). Throughout 1991, struggles over sovereignty and independence dominated Soviet politics: from the use of violence against pro-independence protestors in Latvia and Lithuania in January, to the failed August putsch, to Gorbachev’s resignation on December 25, which signaled the end of the USSR. Strengthening democracy and reforming the economy took back seats. Yeltsin’s bold resistance to the August 1991 putsch had burnished his credentials as a “democrat” and convinced many observers that the struggle was between supporters and opponents of democracy. Yet the coup plotters actually were seeking to hold together the Soviet Union, and their failure hastened the break-up. Throughout 1991, almost every major event strengthened the view that large multinational unions (empires) were illegitimate. Post-Soviet Russia, when it began to operate at the beginning of 1992, remained a multinational union, even if it was substantially less culturally diverse than the Soviet Union had been. Many of its ethnically non-Russian citizens therefore saw the post-Soviet Russian Federation as little better than the Soviet Union. Simultaneously, many ethnic Russians identified with and felt patriotic about the now defunct USSR rather than the new country of Russia. The challenges of determining Russia’s state institutions and developing legitimacy for them would be formidable. Like other postcommunist countries, Russia at the start of 1992 had to negotiate multiple, fundamental transitions: from the USSR to Russia, from a centrally planned to a market economy, and from an authoritarian system to a consolidated liberal, multiparty democracy. With regard to democracy, although Russia had popularly elected rulers at this time, political parties Page 34 →did not exist and the formal democratic institutions remained untested. Nor were these institutions supported by an independent judiciary and the widespread commitment to the rule of law, an independent media, a lively civil society, or other important sources of democratic strength. The complexity of undertaking all these diverse transitions simultaneously was huge. Each, ideally, would be undertaken after the others had been resolved, meaning that no one right sequence was, or is, obvious.

Against the advice of some advisors, Yeltsin chose to postpone political reform. Instead, he initiated rapid economic reform. On January 2, 1992, legislation that had been passed by the Russian parliament a few weeks earlier went into effect. Prices, including for currency, were freed to be set by the market, which barely existed in an orderly sense. Given the few producers of most goods, this meant prices would skyrocket. Savings were wiped out, and those on a fixed income, notably pensioners, soon saw their savings vanish. These and the other “shock therapy” policies also disrupted patterns of internal trade, so that many goods became hard to obtain at any price. Later that year, the government began selling off (“privatizing”) state assets through an auction process. Although each Russian citizen received a voucher they could use to bid in an auction, most citizens quickly sold their vouchers for much-needed cash. Insider elites and those already wealthy became the new owners. The economic tribulations produced corresponding declines in the quality of life. Life expectancy, birth rates, infant mortality, and health care in general suffered while support for education and culture declined. Fueled by the economic chaos, official corruption and organized crime became rampant. The economic transition and the suffering continued for almost a decade. They echo in Russian politics even today. What is even more notable about their impact, however, was how immediately it began. Within days of the price hikes on January 2, 1992, shortages became apparent, and politicians began opposing the new policies. Even with all the other changes going on, the main battle line for Russian politics quickly came to be between supporters and opponents of market reforms. In addition, it took little time for two unexpected leaders of the opposition to emerge: the speaker of the parliament and Yeltsin’s former deputy, Ruslan Khasbulatov, and Yeltsin’s vice president, Alexander Rutskoi. On the one hand, both were acting in a calculated way to bolster their popularity and political power (Treisman 2011, 48–52). On the other hand, they were defending what they saw as the constitutional balance of power between parliament and the presidency. Elite clashes and public outcry caused by the economic reforms dominated politics in 1992 and 1993. Page 35 →The process of establishing Russia’s federal system also began early in 1992. Treaties defining the relationship between the central government and the regions were signed in March. One covered the ethnic republics, another the oblasts and krais, and a third the autonomous okrugs and oblasts (Lankina 2004). Despite considerable effort, Yeltsin could not convince Tatarstan and Chechnya to sign the treaty for republics. Even with these two important republics outside the treaty system, the treaties reduced substantially the danger of Russia splitting up in the way the Soviet Union recently had. That is not to say, of course, that they immediately put an end to struggles over regional sovereignty and the federal center’s influence. In the fall of 1992, Russia’s Constitutional Court issued a ruling that would influence the development of party politics (Thorson 2012, 82–91). In the immediate aftermath of the August 1991 putsch, Yeltsin had outlawed the Communist Party from operating on the territory of Russia. What had formerly been the property of the Communist Party became the property of the Russian state. A group of communist supporters filed a case with Russia’s Constitutional Court asking that it overturn Yeltsin’s decree as unconstitutional and undemocratic. The Yeltsin administration argued that the Communist Party had hardly been a democratic political party; it was, in effect, a shadow state; freeing Russia from its claws was entirely constitutional and democratic. When the Court’s ruling came down on December 1, it specified that Yeltsin had the power to ban the Communist Party and to turn its property into state property. What he could not do, though, was to prevent communist parties from organizing themselves and competing in elections. In the following months, numerous political parties were established that used the word communist in their names or otherwise professed support for communists. The Communist Party of the Russian Federation (CPRF) was the strongest among them, and it became the primary antiestablishment political party in the Russian party structure. It was, and continues to be, led by Gennadii Zyuganov, who had made his career through 1991 in the apparatus of the Soviet Communist Party. The level of confrontation between Yeltsin and the opponents of reform remained high in 1993. A series of dramatic events in March almost resulted in Yeltsin’s impeachment. A referendum the next month produced stronger support for Yeltsin than for the legislature but did not eliminate the gridlock. The struggle reached its climax in September. Convinced that he had no chance of negotiating with the Khasbulatov-led parliament, Yeltsin issued a decree disbanding it and setting new elections for December. Page 36 →Yeltsin’s announcement was quickly decried as unconstitutional, including by the chair of the Constitutional Court.

Legislative leaders vowed to remain in the legislative building—the same White House where Yeltsin had resisted the putsch in 1991—until Yeltsin backed down. On October 3, a crowd of anti-Yeltsin demonstrators that had marched to the White House became unruly and, together with armed defenders from within the White House, attacked and seized the nearby mayoral building and then attacked the main television center across town. Some 187 died in the resulting fighting. Yeltsin convinced the military to shell, then storm, the White House and arrest those inside. The image of Russian troops shelling the White House, on orders from Yeltsin himself, and the tragedy of the deaths, reduced Yeltsin’s popularity greatly even as it solidified his power (for more on these events, see Kutsylo 1993; Buzgalin and Kolganov 1994; McDonnell 1994; Zheleznova, Panova, and Surkov 1994; Shevtsova 1996; Remnick 1997, 37–83). Yeltsin then issued a draft constitution that created a bicameral legislature and a presidency with much greater constitutional power than before. He called for elections to be held on December 12 both for a plebiscite on approving the constitution and to elect members of the parliament based on the to-be-approved constitution. The largest thrust of the new constitution was the strong role it gave to the federal president, as noted above. The 1993 legislative elections took place two months after the bloody showdown between Yeltsin and his opponents in the legislature. Neither political parties nor the electorate had long to prepare for the political contest to select the members of the newly configured parliament, and the public was disillusioned, depressing turnout nationwide. Also, voters in the Republic of Tatarstan boycotted this election in large numbers to protest its lack of autonomy within the Russian Federation. At the same time, however, these elections included a referendum on Yeltsin’s constitution, which required 50% turnout to be valid. With little preparation and no prior experience with multiparty elections, Russian voters in 1993 had four choices to make when they entered the voting booth: (1)В their preferred party (or bloc of parties running together) from those on the ballot for the nationwide PR ballot; (2)В their preference for their district’s representative in the SMD race; (3)В their preference for a “senator” to represent their region in the Federation Council; and (4)В whether they were for or against the constitution that provided the framework for the legislature to which they were electing members. Thirteen parties were on the ballot. Although officially unaffiliated with President Yeltsin, the party known as Russia’s Choice saw itself as Page 37 →representing loyalty to Yeltsin. Even though Yeltsin never affiliated himself with Russia’s Choice, it did enjoy tacit support from the president in the form of more access to state resources than its rivals (Colton and McFaul 2000, 202; Golosov 2004, 30; White 2007, 25). The party’s second-place finish in the party-list vote, with 16%, together with additional seats gained from single-member districts gave it a two-seat plurality in the legislature. Overshadowing Russia’s Choice, however, was the much more popular, and nationalist, LDPR, which received 23% of the party-list votes. The CPRF, which also opposed Yeltsin and his policies, came in third. Yeltsin’s 1992 treaties with the regions had been multilateral. Although the oblasts and krais were jealous of the privileges allowed to the republics, the 1993 Russian Constitution did not formalize republican privileges, spurring calls in the Republic of Tatarstan to boycott the December election. In response to continued demands for greater sovereignty from the subjects of the federation, Yeltsin began to deal with republics and other regions in a more bilateral fashion (DeBardeleben 1997; Frommeyer 1999; Solnick 2002, 189–92). From 1994 to 1998, Yeltsin struck deals, in the form of bilateral “treaties,” with most of the then 89 regions. These agreements specified formulas for sharing tax revenue and profits from mineral extractions. Fiscal transfers from the federal budget to the regions were less transparent components of the bilateral bargaining. Rather than operating collectively, governors sought the best deals they could from the federal center. This increased the center’s ability to restructure previous deals that had been in the regions’ favor. The Republic of Chechnya, which had never signed the 1992 treaty for republics, continued to resist integrating into the Russia Federation. In late 1994, Yeltsin became frustrated and launched a military invasion (Lapidus 1999). The resulting war lasted two years and resulted in tens of thousands of deaths. The Russian military failed to win a decisive victory. The fighting also changed Chechnya into a hotbed of radical Islamic fighters from

Chechnya as well as outside the country, ensuring that the 1996 ceasefire would not last long (Wilhelmsen 2005). With the country’s economic situation still quite severe, the war further eroded Yeltsin’s popularity in the run-up to the 1995/1996 election cycle. With more time to organize than had been the case in 1993, a staggering 43 parties appeared on the ballot for the December 1995 Duma elections. The four that passed the 5% threshold included three that had run in 1993: the CPRF, the LDPR, and a party named Yabloko, which represented pro-democracy, pro-social policies. The fourth, Our Home is Russia, had been created earlier that year with the approval of President Page 38 →Yeltsin and was headed by Prime Minister Viktor Chernomyrdin. Compared to Russia’s Choice in 1993, Our Home represented centrist and less economically radical policies. Like Russia’s Choice two years earlier, though, Our Home enjoyed an advantage in access to state resources and the media due to Kremlin backing. Observers widely considered it a party of power (Badovskii and Shutov 1997, 36; White, Wyman, and Oates 1997, 771; Gel’man and Elizarov 1999, 34; Colton and McFaul 2000, 202; Easter 2001, 56; Remington 2008, 172–73). In 1995, in contrast to two years earlier, the protest votes went overwhelmingly to the CPRF, which received the highest proportion of the votes cast nationwide, over 22%, and three times as many seats in the Duma as the next party. Zhirinovsky’s LDPR fell to 11% in the party-list voting. Despite its financial advantages, Our Home is Russia placed third with 10% of the party-list votes. The prospects for Yeltsin to work effectively with the Duma dimmed. The big question, then, as 1996 began was whether Yeltsin would be able to beat off a challenge in the June election from CPRF leader Zyuganov. The CPRF’s strong showing in the Duma election combined with Yeltsin’s single-digit public approval ratings at the start of the year made Zyuganov’s chances seem high. Yeltsin was able, however, to rally support throughout the first half of 1996. He had tremendous advantages in what Russians call “administrative resources,” that is, the financial, personnel, and other resources the executive branch has at its disposal to influence elections in ways that go beyond incumbency advantage (Nikolaev 2000; Vorontsova and Zvonovskii 2003). Among other factors accounting for Yeltsin’s resurgence was his success at framing the race as a referendum on a return to communism. Yeltsin received 35% of the votes, with Zyuganov second at 32%. Yeltsin then made an alliance with the third-place finisher, General Alexander Lebed, who had received 14.5% in the first round. Yeltsin received 54% of the votes in the second round. By 1996, then, the basic contours of Russia’s post-Soviet political system were established, including the provisions for electing regional governors. The October 1993 crisis had led to a new constitution, which remains in force today with only minor changes. Yeltsin’s victory in the 1996 election ensured that the basic commitment to market economics would continue as well as that election outcomes would matter. Although the Kremlin would alter key features of federal relations in the years ahead, Russians had broadly accepted the idea of Russia as a state with the RSFSR’s borders, and only in Chechnya was independence still a goal. Change hardly stopped in 1996, of course, but the subsequent stability looks impressive. Page 39 →What remained unanswered, though, was whether the big transitions would succeed. Could the political system, formally dedicated to being a democracy, support a genuinely democratic political regime while grappling with substantial challenges of governing? Would the fledgling market economy stabilize, grow, and end the emiseration Russians had experienced since the late 1980s? Would the federal system bolster economic and political progress as well as the country’s social stability?

The New Institutions Fail Their Tests The remaining years of the 1990s, most of Yeltsin’s second term, convinced many Russian and foreign observers that the answers to all three of these questions were no. Though Yeltsin was a dynamo in the first half of 1996, while campaigning for reelection, his health kept him from playing a strong leadership role in his second term. He suffered a series of heart attacks, three in 1995 plus one between the first and second rounds of the 1996 election (Colton 2008, 314, 372, 376–83). Heart troubles kept him largely out of politics throughout the second half of 1996. Open-heart surgery in November 1996 led to an improvement by 1997, but he had to conserve his

strength and minimize his time spent working throughout the rest of his term. By 1996, despite much progress in structural economic reform, there were few signs of the economy improving. Inflation had been brought down to manageable levels, although one way the government did so—by not paying state employees their salaries—undid some of the public’s gain. Most other indicators had yet to reverse their downward trends: unemployment was still climbing, real wages were at 50% of what they had been in 1991, and gross domestic product was at 60% of the 1989 value (World Bank 2014). An ill and frequently absent president and a hostile, communist-dominated legislature made it very difficult to grapple with such serious economic problems. It is no surprise that the second half of the 1990s saw Russia’s mass public and members of the elite alike grow increasingly unhappy with the political system and express distrust of its main institutions. The very word “democracy” grew unpopular because it was the name of the political system that was serving the public so poorly. In August 1998, Russia experienced a serious fiscal and monetary crisis sparked by an earlier crisis in Asian currencies. Hyperinflation returned for several months. The economic tribulations produced corresponding declines in the quality of life. Life expectancy, birth rates, infant mortality, Page 40 →and health care in general suffered while support for education and culture declined (Field and Twigg 2000). Although the effects of the crisis would help Russia’s economy begin a period of strong growth, the short-term pain exacerbated the political struggles between Yeltsin and the legislature. Prior to the crisis, Yeltsin had dismissed the experienced centrist Prime Minister Chernomyrdin and forced the State Duma to approve a young, reform-oriented official, Sergei Kirienko, by threatening to use his power to disband the Duma. When the crisis hit, Yeltsin quickly fired the prime minister he had fought so hard for and nominated Chernomyrdin to return to the post. The legislature balked, and eventually Yeltsin had to nominate foreign policy expert Yevgenii Primakov instead. That fall, the State Duma initiated impeachment proceedings against Yeltsin, which dragged on until it failed to pass in May 1999. On the eve of the impeachment vote, Yeltsin fired Primakov and named the interior minister, Sergei Stepashin, to replace him. In August, the musical chairs of prime ministers finally ended when Yeltsin again fired the prime minister and named Putin to the post. The disarray in national politics did nothing to improve the federal system or encourage good governance at the regional and local levels. Central elites were growing increasingly concerned that the pendulum had swung toward the regional governors, whose popular election gave them formal independence of the federal government. Many were consolidating political control in their regions, including control over branch offices of federal bureaucracies, especially those involved in law enforcement. Regional policies ignored or openly contravened federal laws. Yeltsin’s need for the governors’ support in his struggles with the legislature enhanced their bargaining power—their ability to “swallow” a great deal of autonomy. The sense that many regions had become fiefdoms flouting Moscow’s rules was widespread. As Russian president Dmitrii Medvedev (2010) said casually to representatives of social organizations from the Caucasus regions, “In the 1990s, we had no real authority anywhere, not in the Caucasus and not in the country as a whole, to be frank about it.” The sense grew that this posed a genuine threat to the integrity of the Russian state (Taylor 2011, 126–27; Alexseev 1999; Lapidus 1999; Stoner-Weiss 1999). Nor could defenders of regional authority claim that the onset of gubernatorial elections had bolstered Russia’s democracy or provided citizens with good governance. The need to win elections increased the incentives for electoral fraud and the building of political machines. As we discuss in chapter 3, genuine electoral competition had begun to emerge in some of Russia’s regions during this period. To many, however, regional Page 41 →politics posed a bigger threat to Russian democracy than the gridlock at the national level. By the time Yeltsin named Putin as prime minister in August 1999, the center’s weak control over the regions would take on a new light and, in Putin’s eyes, indeed threaten Russia’s continued existence as a state (Taylor 2011, 255). In the years following the Chechen War of 1994–96, radical groups had strengthened their presence in neighboring regions of the north Caucasus area as well as in Chechnya. Just days before Putin assumed office, armed insurgents crossed from Chechnya into neighboring Dagestan and held a portion of its territory. In early September, a series of terrorist bombings in Russian cities killed several hundred people. Putin

ordered air strikes against Chechnya and a full-scale invasion soon after. This time, Russian forces had more success in subjugating the Chechen resistance. By early 2000, Russian troops had control, with the Chechen capital flattened and heavy casualties among the Chechen population as well as the fighters. For several years, everyone had understood that the 1999–2000 electoral cycle was to be particularly critical for Russia’s future. Yeltsin was constitutionally obligated to step down in 2000, and his health was too poor to push the issue. Several regional governors sought to maximize their influence over the federal center in the years ahead—which included presidential ambitions on the part of some—by creating political parties to compete in the 1999 Duma elections. Moscow mayor Yurii Luzhkov created the Fatherland Movement in December 1998, and despite many regions’ distrust of the capital city, gained the support of some 11 governors by the fall of 1999. In April 1999, the leaders of Tatarstan and St. Petersburg and other regions established a rival pro-regions party, All Russia. In August 1999, these two parties merged to create an electoral bloc called Fatherland/AllRussia (following the Russian spelling, abbreviated as OVR) to run as a single listing on the PR ballot. The next month, Putin’s team created a political party, originally called Unity/Medved (“the Bear”) but mostly referred to as Unity. Unity’s role was to give Putin organized support within the State Duma, in particular by sapping the strength of OVR in the December balloting. By election day, Putin’s popularity had become very high, thanks to his hardline response to the terrorist incidents in September. Unity outperformed expectations in that December’s legislative elections. When those elected convened in early 2000, Putin maneuvered his party into the most influential position (Remington 2003). No one any longer doubted that Unity was the party of power with the full backing of the president and his team. Page 42 →On the last day of December 1999, Yeltsin surprised almost everyone by stepping down six months before the end of his term. This move elevated Putin to the presidency and caused the 2000 presidential election to occur in March, three months earlier than in 1996. Putin was personally popular, had achieved the strongest position in the legislature, and was running as the incumbent. No regional governor saw any purpose in challenging him for the presidency. Instead, they became eager to show their support for Putin in hopes of securing or maintaining federal largess. Although 11 candidates were on the ballot, CPRF leader Zyuganov represented Putin’s only serious challenger, and he secured only 29% of the votes. Putin received 53% and won in the first round. As Putin began his presidency, he was helped substantially by the turn-around in the Russian economy, which had begun the previous year (Rutland 2005; Hedlund 2008; McFaul and Stoner-Weiss 2008; Treisman 2011, 232–37). Economic output, real wages, household income, and state revenue all began a steady climb that continued until the 2008 world financial crisis. The causes for the turn-around included the devaluation of the ruble caused by the 1998 crisis, reforms adopted in 1998 and 1999 in response to that crisis, rising world oil and natural gas prices, and reforms Putin introduced, including a simplified tax code. Although the economic upturn began before Putin took charge, its impact became clear to Russians only after he was in charge. The strong economics have provided Putin with high levels of public popularity. At the same time, the Russian economy came to rely over this same period to an even higher extent on exports of natural resources. Oil and natural gas accounted for a growing portion of Russia’s entire budget revenue, exceeding one half by 2012 (Gel’man 2010b, 9; US Energy Information Agency 2013). With world prices for these commodities rising from the late 1990s on, apart from a temporary dip in 2008–10, Russian officials had less incentive to diversify the economy, strengthen manufacturing, and, in other ways, modernize.

A New Regime Is Constructed Putin came into the presidency sharing with the majority of other elites and the populace a conviction that Russia was on the wrong course in fundamental ways. Putin’s response was not to alter or replace the constitution or formal institutions of power such as elections for top offices. Nor did he contemplate changing Russia’s commitment to the free market or the basic concept of Russia as a federal state. Instead, he made Page 43 →important changes in all three areas less through institutional changes than by constructing new power relationships. That is, he changed the political regime governing Russia.

Whether one views Russia’s political system in the 1990s as mostly or hardly democratic, the political regime Putin constructed in the 2000s is authoritarian. Formal institutions required for democracy such as elections and parliaments continue to operate and play important roles in the political system. Yet they do not produce democratic conditions such as genuine political competition and effectual public participation. Putin, immediately upon obtaining the presidency, set out to tame the social forces and institutions that might have promoted democratic politics. Among other things, he increased the government’s control of the most strategic, and profitable, sectors of the economy; eliminated any political opposition from rich business people, or oligarchs; gained control over the major media outlets and stifled opposition throughout the media; cracked down on civil society organizations that might be vehicles of public opposition; weakened judicial independence through threats or bribes to judges and lawyers (Pomeranz 2010; Solomon 2010; White 2011b, 344–47); and undermined the development of a responsible party system capable of generating changeovers in control of the country’s executive power. The change in the political party system was about more than the series of laws that reduced the number of political parties and gave the government the ability to ban certain parties. Crucial as well was the creation of a dominant party. Unity, which had so successfully garnered votes in 1999 and created a pro-Putin bloc within the Duma, needed to expand its role and extend its reach throughout the country. Putin began working to bring former rivals into the fold. A big step came in 2001, when the leadership of OVR, the other party of power in 1999, agreed to merge into Unity. The resulting political party was renamed United Russia. In the years that followed, a major focus of Putin’s leadership was to bring into United Russia’s membership as many as possible of Russia’s governors and other key elites, along with their political clients (Reuter and Remington 2009; Reuter 2010). Once this happened, the leadership had a vehicle for ensuring easy wins in federal, regional, and local elections, which in turn allowed them to control legislation. Some rival parties, including the CPRF and LDPR, continued to compete in elections but without the prospect of victory. All these facets of Putin’s authoritarian regime were connected with relations between the Kremlin and Russia’s regions and major cities. Putin had come to power sharing the views of many that centrifugal forces had gained the upper hand, threatening Russia’s unity. In his efforts to create Page 44 →a stronger Russian state, Putin gave a key place to strengthening central authority over regional policies. The shorthand for this was “strengthening the vertical [dimension] of power” or “power vertical.” A key early step came when, in a May 13, 2000, decree, Putin established seven federal administrative districts, each composed of from six to 18 contiguously located regions (Hyde 2001, 725–27). These seven districts were the Northwest, Central, Southern, Volga, Urals, Siberian, and Far East. In 2010, Medvedev created an eighth, the North Caucasus district, which contains seven regions that had previously been part of the Southern district. Each district has a presidential representative assigned to it. The districts did not become a new level of governance between the regions and the Kremlin but rather streamlined the Kremlin’s oversight of the regions—in particular to ensure that the president’s domestic and foreign policies are implemented in the regions and that appropriate federal personnel are appointed—and provided a mechanism to pressure regional leaderships to repeal laws that conflicted with federal law (Orttung 2004, 22; Petrov 2005, 57). Another step taken in 2000 was the change in the membership of the upper house of the federal legislature. Rather than the regional governors and legislative speakers serving as each region’s two members, Putin had the law changed so that each of those officials chose someone else to be the member. Reforms also made it hard for the upper house to block legislation. The “senators” now had incentives to be friendly to the Kremlin in hopes of receiving largess. Another part of this process was the annulment of the bilateral treaties signed under Yeltsin. Fiscal federalism, the rules for sharing or transferring funds from one level to another, was altered to eliminate what had previously been notable asymmetries among the regions and to capture for the center certain lucrative types of revenue. In the context of these changes and through informal pressures and inducements, regional leaders along with their followers did indeed join United Russia, further solidifying the power vertical. Putin also began, fairly early in his administration, to float ideas about reducing the total number of regions (Chebankova 2010, 171–78). Eighty-nine regions is more than any other federal state in the world, significantly more than the 50 US states. Moreover, their ranges of size, population, and budget may lead the world as well.

Although some proposals were floated for a quite radical reduction in number, Putin moved slowly and only was able or willing to push through the consolidation from 2005 to 2008 of six small autonomous okrugs into the larger regions around them, leaving the federation with 83 subjects. Page 45 →The December 2003 election to the State Duma saw little competition. United Russia received 38% of the party-list vote—a record to that point in Duma elections. Candidates running under its label won numerous SMD mandates while even more switched to the party of power after winning election. Together, this gave United Russia a two-thirds majority in the State Duma, allowing it to pass constitutional amendments without votes from other parties. The CPRF received only 13% of the party-list votes, barely more than half of its 1999 vote total. In the presidential race in March 2004, Putin had no viable rivals. Even CPRF leader Zyuganov opted to forego the race; Nikolai Kharitonov represented the communists. Despite the lack of a strong challenger, the Kremlin portrayed the elections as a referendum on Putin’s leadership and devoted substantial resources to securing Putin’s landslide. Still, Putin’s forceful handling of Chechnya failed to resolve the conflict, and Russia experienced an uptick of terrorist attacks in 2004. These included the downing of two commercial flights in August and the siege of a school in Beslan on September 1 where over 1,100 people were taken hostage and more than 300 died. Following the 2004 Beslan tragedy, Putin made two major institutional changes to strengthen the Kremlin’s hand. One, as noted above, was the end of gubernatorial elections. The other was to end the election of half of the membership in the State Duma from single-member districts in favor of all 450 members being chosen through proportional representation. These changes further strengthened the vertical of power but weakened avenues for democracy. Without gubernatorial elections, Russians lacked the formal right to control a critical political office. Both Freedom House and POLITY, prominent Western sources for democracy ratings, downgraded Russia’s score the following year. Alongside the support of most regional governors, the Kremlin worked hard to achieve majorities for United Russia in all the regional legislatures. They began by passing federal laws to alter regional institutional settings. In 2001, revisions to the law on political parties made it impossible for a party to operate in only one or a small number of regions (Ross 2002, 111–12; Moraski 2006b; Golosov 2011a, 628). This strengthened the role of a small number of national parties. To overcome many politicians’ preference to run without a party label at all, a 2002 law required that all regional legislatures select at least half of their members using party-list proportional representation (Golosov 2003; Moses 2003). The use of a minimum threshold in these party-list elections meant that any votes for minor parties boosted the number of seats held by large parties such as United Russia. Page 46 →These institutional measures were not fully successful, in large part because most governors, even those strongly supporting the Kremlin in federal elections, wanted a pliable regional legislature rather than one with a solid majority for the Kremlin’s party (Golosov 2011a). The switch to Kremlin appointments of governors in 2005 increased the incentives for governors to promote United Russia in regional legislative elections. At the same time, the Kremlin also removed two mechanisms with which governors could keep United Russia from dominating the regional legislatures. To reduce the total number of small parties, it raised the required number of members for a party fivefold at the end of 2004 (Golosov 2011a, 632–33). The next year, it banned the use of electoral blocs (Golosov 2014b). Over the course of the next several years, United Russia’s dominance of regional legislatures solidified and remains in place to this day. Kremlin control of regional affairs grew correspondingly stronger. In addition, the Kremlin enhanced its control over a key function for any authoritarian regime, coopting elites, since its party was now the primary gatekeeper of who would serve in regional legislatures (Golosov 2014a). Although United Russia’s nationwide dominance was unquestioned prior to the 2007 Duma race, federal and regional officials pushed hard to produce high turnout and high vote totals for this party of power. It proved easy to accomplish. Not only did Putin enjoy continued high popularity among the public, he had unified control of the federal organs and established much stronger tools for influencing the regions than he had inherited. In addition, the public recognized and supported his party, United Russia, far more widely than other parties. This was the first

election affected by a new rule that eliminated the SMD portion of the Duma and allocated all 450 Duma seats on the basis of PR. Because over 8% of the votes went to seven parties that failed to pass the threshold, United Russia’s two-thirds vote percentage translated into 70% of the seats. Other parties were afterthoughts. Putin’s constitutionally maximum second consecutive term as president was scheduled to end in 2008. Speculation had been pervasive for several years prior about whether Putin would step down when his term ended—even though he consistently maintained that he would—and, if he did, whom he would anoint as his preferred successor. In 2005, he had designated two frontrunners to succeed him by making each a first vice-prime minister: Dmitrii Medvedev and Sergei Ivanov. In late 2007, Putin and United Russia announced their support for Medvedev to become the next president, while Medvedev made clear that he would nominate Putin as the prime minister. Putin thus would remain active in running federal politics Page 47 →but from a new position. In early March 2008, with Putin’s and United Russia’s backing—and therefore without even needing to campaign—Medvedev received almost as high a vote total as Putin had in 2003. Zyuganov of the CPRF was a very distant second. Medvedev was inaugurated just shortly before the global economic crisis hit. The financial shockwaves hit Russia hard, especially as it was accompanied by a decline in world energy prices. Crude oil prices fell from a peak of $140 a barrel in early 2008 to under $40 later that year. Natural gas prices also fell in 2008. While natural gas prices stayed lower than their pre-2008 levels, oil prices rose slowly throughout 2009 and 2010, staying just under or above $100 a barrel until the middle of 2014. Medvedev made it his watchword to modernize the economy. Yet, perhaps because high oil revenues allowed for a quick recovery from the crisis, fundamental economic reforms were few during Medvedev’s presidency. In September 2011, with a new election cycle approaching, Medvedev and Putin announced that they would switch positions. Putin would run for president and then appoint Medvedev as prime minister. With the presidential term having been changed from four to six years, this would permit Putin to serve as president potentially until 2024. The disdain this showed for the public’s intelligence hit a chord, especially with urban professionals and the young. Former regime supporters were, in any case, growing dissatisfied with the country’s progress in modernizing the economy. This discontent soon became public in the run-up to the December 2011 Duma election. Internet satire and criticism helped spark public protests in the larger cities. When the election occurred, United Russia’s 49% vote total rested on multiple clear-cut instances of fraud. Public outrage ramped up immediately, leading to largescale demonstrations that continued into the spring of 2012. These protests sent ripples through Russia’s political regime. For example, President Medvedev announced several moves aimed at pacifying the protesters, including the return to a new form of gubernatorial elections, as discussed above. The regime, however, was not seriously shaken by the opposition. Putin ran a mostly boring campaign for the March 2012 presidential race,7 refused to debate his opponents, and won easily in the first round. Following his victory, Putin put an end to concessions to the opposition and began cracking down in a variety of ways designed to raise the costs of opposition. These included arrests and harsh sentences for demonstrators, arrests of opposition leaders such as the anticorruption blogger Alexei Navalny on trumped-up charges, and a law requiring NGOs that receive foreign funding to declare themselves “foreign agents” (see, for example, Kramer and Corke 2013, 20). Page 48 →Events in early 2014 redoubled these trends. In February, after months of protest in Kyiv’s Maidan Square, a rapid flare-up in violence led Ukrainian president Viktor Yanukovych to flee the country. His supporters vacated the parliament, which then chose an opposition government. Russian leaders portrayed this as the result of Western, especially American, plotting aimed at integrating Ukraine into NATO. Russia responded by stealthily seizing control of the Crimean peninsula and, in March, annexing it to Russia. Not long after, rebels in eastern Ukraine took up arms against the central government, some seeking integration into Russia, others independence, with Russia providing financial and military support. Western governments responded to Russia annexing Crimea by imposing financial sanctions on top leaders and government-owned corporations. Patriotic fervor and antiWestern feelings soared, as did support for Putin’s job performance. A series of new policies further cracked down against criticism or opposition to the leadership, now treating opposition as sedition.

Russia’s current political regime, which Putin molded out of the institutions and practices begun in the Yeltsin era, has shown enough stability to weather the 2008–10 financial downturn, an outbreak of political protest in 2011–12, and another financial downturn from late 2014 on caused by foreign sanctions and declining oil prices. Elite unity and public support for the leadership were both high in 2016. Nonetheless, the stability of the political regime is being debated as perhaps never before. The regime depends on Putin to give direction with input in recent years from a quite narrow circle of advisors, while the role of formal institutions shrinks (Baturo and Elkink 2015; Melville and Mironyuk 2016). Reforms that would promote economic and social modernization are unlikely without a change in leadership at the top, which Putin’s indispensability makes unlikely for many years. Dissatisfaction is high in numerous quarters, including among regional governors who have seen federal funds reduced and their share of the cost for providing social services to their residents grow.

Russia’s Trajectory in Comparative Perspective The political developments in Russia since 1991 have driven it downward in the estimation of those who judge how democratic countries are. How does Russia’s regime compare to those in other countries over a similar period, particularly other countries that emerged from communist rule in the late 1980s or early 1990s? One way to gauge this is to compare Russia’s Page 49 →annual scores on measures of democracy produced for most countries by expert judgments. Fig. 2.1. Russia’s Level of Democracy in Comparative Perspective. (Source: Freedom House various years; Marshall, Gurr, and Jaggers 2015.) The scores presented in figure 2.1 were created using the procedure developed by Hadenius and Teorell (described in Teorell 2010, 33). This involves combining POLITY’s measure of combined democratic and autocratic features (the polity2 score, see Marshall, Gurr, and Jaggers 2015) with the average of Freedom House’s (various years) measures of civil liberties and political rights. The resulting index ranges from 0 (least democratic) to 10 (most democratic).8 The figure shows the trends from 1972 to 2013 for the world as a whole, and through 2014 for Russia and other postcommunist countries, grouped geographically. The horizontal dashed black line indicates the level above which countries, by this measure, are considered democracies. A broad double line indicates the trend for all countries in the world. This line illustrates the acceleration in 1989 of the so-called third wave of democratization (Huntington 1991) and its leveling off from the mid-2000s. On average, the countries of North Central Europe (which include the Baltic countries) and Mongolia score higher than the line for democracy,Page 50 → higher than the world average, and well above the other postcommunist countries throughout the postcommunist period. The countries in South Central Europe (Romania, Bulgaria, and the former Yugoslavia) take a decade before their average score rises into democratic territory, but by 2012 these countries are only slightly below their neighbors to the north. The Central Asian countries are distinctly more authoritarian than countries in the other regions or the world average. The post-Soviet countries in the Caucasus (Armenia, Azerbaijan, and Georgia) and those in non-Russian Eastern Europe (Belarus, Moldova, and Ukraine) are in the middle, averaging since the mid-1990s either somewhat above or below five, or midpoint on the scale. Russia is given credit, in the immediate aftermath of the collapse of the USSR, for a degree of democracy higher than the world average. However, while the world average continues to climb rather steadily after 1991, Russia’s score follows a downward trend. Its score crosses the world average in 1997 and in 2012 is almost two points below that average out of the ten-point scale. The fraud surrounding the 1996 presidential election begins the decline in Russia’s scores. When Putin replaces Yeltsin, the score bumps upward a bit, but as Putin and United Russia consolidate control in 2003–2004, another decline ensues. The reforms that Putin initiated following the Beslan tragedy reduce the score yet again. Scholars who assign world political regimes to different types, as opposed to scoring them on an ordinal or interval-level scale, consider Russia’s regime as authoritarian throughout most of the post-Soviet period. For example, Levitsky and Way (2010) placed Russia into their category of competitive authoritarianism—a hybrid regime indicating a comparatively liberal regime in which the power holders abuse state power to give themselves

unfair electoral advantages—in the early 1990s but by 2008, when they did a second round of scoring, they placed Russia in the fully authoritarian camp. Other regime-classifying projects do not define categories based on the degree of democratic elements. Instead, they begin by distinguishing democracies from nondemocracies and then subclassify the latter. Cheibub, Gandhi, and Vreeland (2010) place Russia with the group of civilian dictatorships throughout the postcommunist period. Geddes, Wright, and Frantz (2013) categorize Russia as having a democratic regime in 1992 and 1993 but as a personal (not party, military, or monarchical) autocracy from 1994 through the end of their data collection in 2010. Wahman, Teorell, and Hadenius (2013) break down civilian regimes further than the others in classifying Russia’s regime as multiparty authoritarian from 1991 on. Fig. 2.2. Russian Per Capita Income and Democracy Trends, 1990–2011. (Source: Income figures come from the Penn World Tables: Heston, Robert, and Aten 2012 and Feenstra, Inklaar, and Timmer 2013. Democracy scores are from figure 2.1.) Page 51 →It will be helpful at this point to show that a powerful correlate of democracy worldwide—level of economic development—helps little in understanding the trajectory of Russia’s polity since independence. As discussed in chapter 1, scholars debate whether prosperity makes democratic progress more likely (Epstein et al. 2006; Boix 2011) or primarily prevents a new or fragile democracy from failing (Przeworski et al. 2000). Was Russia’s level of economic development at the time of independence simply too low for democracy to “stick” or did the economic collapse caused by marketization lower it to such a level? Figure 2.2 shows Russia’s annual per capita income levels from 1990 to 2010 in constant 2005 dollars. We use the same source for per capita income as Przeworski et al. (2000): the Penn World Table (Heston, Robert, and Aten 2012; Feenstra, Inklaar, and Timmer 2013). Horizontal dotted lines at $8,100 and $11,300 indicate the levels at which, the comparative evidence shows, democratic failure is, respectively, unlikely and highly unlikely. Figure 2.2 also includes for comparison the measure of Russia’s level of democracy that was shown in figure 2.1. Clearly, Russia’s level of economic development cannot explain much about its political trajectory. Russia’s per capita income during the two Page 52 →years prior to the break-up of the Soviet Union exceeded the level at which new democracies have excellent odds of stabilizing. The economic transition to capitalism produced a seven-year nose-dive in per capita income, yet the level in the worst year, 1998, was only just below the $8,100 level. This economic crash unquestionably complicated the functioning of new institutions and destabilized Russian society. Yet if this economic decline is to be seen as a cause of Russia’s failure to democratize, it is not for the reasons argued by modernization theory. Russia did not suddenly become less urban or educated. Also, figure 2.2 shows that even in the year of lowest per capita income, 1998, Russia’s income level was relatively high by worldwide standards. From 1999 on, per capita income rose steadily (except 2008–2009) as Russia’s democracy declined. We will, however, return to economic development in subsequent chapters because the development trends for Russia as a whole mask sharp interregional variation, and, as we argue in chapter 1, it is interregional variation that matters. So it will remain important to examine level of development as a relevant part of each region’s political context.

Conclusion Over post-Soviet Russia’s first two decades, the chaos of the early years gave way to a flawed and unstable political order, which was replaced by a more clearly defined and resilient authoritarian political order under Putin. Two dynamics played major roles in shaping the regime: the dozen federal election campaigns from 1993 to 2012 and relations between the federal center and regional leaderships. In addition, these two dynamics were tightly connected with each other. To get Russia’s post-Soviet political story right, therefore, demands close attention to electoral trends across Russia’s regions. Beyond this, Russia’s post-Soviet political trajectory offers comparative lessons for those who study both democratization and authoritarian regime building. Regularly scheduled multicandidate elections and federal institutions can be associated with the deepening of democracy (Lindberg 2009). That they did not in

Russia’s case deserves note, and how they factored into the rise of the authoritarian regime deserves explanation. Subnational politics and federal relations are central to this explanation.

Page 53 →

Three Explaining Competition in Russia’s Early Gubernatorial Elections Ruling an authoritarian regime is a lucrative occupation. Political power and a monopoly on the means of coercion allow authoritarian rulers not only to extract rents and levy taxes but also to amass substantial personal fortunes, and those who overthrow an authoritarian regime can be tempted to repeat the sins of their predecessors. Without a dispersion of force and resources, or intervention from external actors, one leader or group may again prevail over the rest. While such conditions may emerge at the national level, or in the capital, some political actors may be able to monopolize power in smaller, seemingly peripheral domains, leading to the rise of “miniautocracies” (Olson 1993, 573). Thus to understand the environment into which Russia’s fledgling democratic institutions were introduced and to assess why they foundered, we need to understand how elections worked across the country during the early post-Soviet years. Gubernatorial elections provide an excellent window into regions’ levels of political competition. The executive leaders of the regions had and continue to have significant influence over citizens’ lives. We want to know, then, where regional politics was competitive enough that voters could make meaningful choices and have their votes determine the outcomes. In our analyses, we employ two straightforward indicators of electoral competitiveness: how close were the election results, a measure known as the effective number of candidates, and whether the incumbent was Page 54 →retained or voted out. We estimate models of cross-regional variation in these measures of competitiveness. The models include an array of conventional arguments from the transitology and comparative democratization literature. In addition to allowing us to assess how well explanations developed for comparing countries’ performance at the regional level, we use this initial empirical chapter to paint a picture of Russia’s electoral landscape that we will build upon as we move from understanding the regional correlates of regional electoral competition to understanding the Kremlin’s decision to eliminate gubernatorial elections (chapter 4) and regional correlates of electoral competition in federal contests (chapters 6 and 7). In brief, our analysis finds mixed results for hypotheses drawn from cross-national studies of political transitions and comparative democratization when seeking to explain competition levels at the regional level in Russia. For the most part, economic performance fails to systematically influence the competitiveness of Russia’s gubernatorial elections during the Yeltsin-era. Meanwhile, both higher levels of socioeconomic development and greater availability of natural resources dampen competition in these regional contests, measured in terms of the effective number of candidates. The choice of electoral rules also influences the number of candidates, though it appears to have no independent effect on turnover. On the other hand, regional levels of party development and violent crime levels influence the likelihood of turnover in elections held after 1994, but not the number of gubernatorial candidates. Moreover, the former only proves significant when the latter is excluded from the model. This chapter also emphasizes the importance of regionally specific measures and hypotheses. For example, although we use levels of violent crime as an intuitive indicator of cross-regional variation in societal trust, which is something identified as important in the cross-national literature, the measure itself differs from those used in cross-national studies where public opinion data are much more readily available. Unlike survey questions that directly ask respondents to express levels of trust in others but fail to capture whether people actually behave in ways that are less trusting, our measure identifies behavior (acts of violent crime) that should breed both the attitudes and actions of distrust. More emblematic of regionally specific explanations, however, are the effects of the constitutional status of Russia’s regions (i.e., being a republic or not) and the timing of the first gubernatorial elections. Here we draw on the question of electoral timing within the transitology literature. Our analysis reveals that the failure to take into account this consideration could lead one to miss an importantPage 55 → dimension of electoral politics across Russia’s regions. Specifically, we demonstrate that republics did not consistently experience less competitive elections than other Russian regions, as one might expect based on initial differences in constitutional status.1 The variation also does not depend on the concentration level of non-Russian

minorities. Rather, this characteristic emerges in republican elections where leaders managed to hold founding elections during Russia’s transition from communism, which we define as the 1991–93 period.2 In republics where the first gubernatorial elections were held after the onset of Russia’s new regime, demarcated by adoption of the 1993 Russian Constitution, the likelihood of turnover was actually higher, even higher than in nonrepublics during this period.

Electoral Competitiveness While scholars debate whether electoral competitiveness is a sufficient criterion to denote a democratic regime (cf. Schumpeter 1942 [2003]; Dahl 1998; Przeworski 1999; Diamond 2008; Pateman 2012), they agree that it is a central minimal requirement. A competitive polity both enables and enjoys a comparatively broad degree of involvement in politically influential acts, including but not limited to voting, campaigning, and lobbying. It also enables and enjoys organized rivalries for key positions of power (Dahl 1971; Vanhanen 1997). The core of democratic political competition is the presence of competitive elections. Of course, finding competitive elections in a region does not indicate by itself a democratic environment. Competitive elections may occur in a regime that fails to qualify as a democracy on other dimensions. For example, incumbents may lose despite blatant attempts to skew the electoral playing field in their favor. Elite splits (Langston 2006), opposition coalescence (Van der Walle 2006), and manipulative skill (Case 2006) introduce the possibility of an opposition victory in what is not a genuinely democratic setting. Focusing on subnational elections complicates matters even more since a country’s constituent regions may not only vary significantly among themselves when it comes to regime type (Snyder 2001; Gibson 2005) but national actors also may take actions that exogenously influence the trajectories of the regional regimes. For example, in the next chapter we argue that the Kremlin’s inability to effectively influence gubernatorial elections within the norms of electoral democracy explains the decision to eliminate gubernatorial elections altogether. Before moving to that Page 56 →analysis, however, this chapter examines regional gubernatorial elections during the Yeltsin presidency to determine whether certain regional characteristics or decisions were systematically correlated with more liberal or autocratic election outcomes. As noted earlier, we measure competitiveness in Russia’s gubernatorial elections with two indicators. First, we assess the degree to which those elections have presented voters with a choice among two or more viable candidates, where viable means enjoying sufficient support that winning is possible. As is standard in the literature, evidence on viability comes post hoc, from the electoral returns themselves. When one candidate receives a very high proportion of the vote despite the ballot having multiple candidates, an individual voter’s choice is barely more consequential than when the ballot contains a single candidate. The most broadly used indicator in the literature on elections and parties is the “effective number of candidates (or parties)” index (Laakso and Taagepera 1979). As values of the measure decrease toward 1, they indicate increasingly inadequate voter choice. Higher scores indicate more competitive races. We employ a modification of this measure proposed by Golosov (2010), as explained in appendix 1. Our second indicator of electoral competitiveness captures whether control of the government changes hands, a common litmus test for a democratic system (e.g., Huntington 1991, 266–68). Turnover from one governing administration to another provides clear evidence that elections were competitive. It indicates more than this, however. A region having competitive elections but still awaiting a turnover in the executive branch is likely to miss out on some or all of the benefits that theorists expect from elections. Faith in elections as the path to power can hardly take root if no one has seen a power-holder peacefully step down from office after losing an election. In addition, turnover enhances the commitment of politicians in democratizing systems to work within the system, since those who fall from power might make a comeback (di Palma 1990). Solnick (1998, 72–73) makes this point about Russia based on examining gubernatorial elections in 1996–97. Hahn’s (1997) study of gubernatorial elections found turnovers to be a sign that mass electoral behavior was shaping regional politics. Accordingly, we want to distinguish between regions where turnover never or rarely occurred and those where it did, thereby cautioning the governing team to either consider popular interests or take steps to undermine or minimize competition in the next round of elections. Our measure of turnover, also described in appendix 1, takes

into account every Yeltsin-era gubernatorial election—that is, those from 1991 through the end of 1999.3 Page 57 →

Correlates of Democracy and Considerations for Subnational Competition If democratically elected governments are more fragile than normal in the years right after an authoritarian regime ends (Przeworski and Limongi 1997, 174), we certainly must examine which regions suffered relatively more or less during the transition after Soviet power. Citizens for whom new, putatively democratic, politics was accompanied by greater misery would naturally be less likely to support democracy in principle or to oppose ambitious elites following nondemocratic courses (Treisman 1999, 92). We expect, then, less political competition in regions where post-Soviet economic and social emiseration was comparatively high. Although democracies are meant to allow citizens to replace their leaders during times of economic or social troubles, this requires that those in control of the executive branch are unable to minimize or control challenges from one or more rival parties or candidates. During a transition period, however, with fledgling institutions, including the party system, we should instead expect that, ceteris paribus, emiseration eases the task of elites who would thwart electoral institutions. That is, executive turnover becomes less likely to occur. We will test whether this shows up as a tendency across Russia’s regions. To measure how regional economic situations vary, we include in our analyses the real monthly income (average per capita income as a percentage of the previous year) for each region. To measure noneconomic social misery, we also include the rate of violent crime in each region. Although violent crime differs from social trust, Rothstein (2011, 197), for one, notes the negative impact of the former on the latter in Jamaica. Similarly, Varshney (2001) also suggests that violent settings can dampen civic development as individuals withdraw from intergroup interactions: in violent Indian cities where criminals escaped the law, people migrated away from the kinds of ethnically heterogeneous neighborhoods that could help undermine ethnic conflict into communally homogeneous ones (Varshney 2001, 378). We also want to examine the impact of different types of regional economies. A number of notable works contend that natural resource wealth, in particular, sustains, if not creates, authoritarian regimes (Mahdavy 1970; Ross 2001; Wantchekon 2002; Smith 2007). Although recent studies question the generalizability of this assertion (Herb 2005; Dunning 2008; Jones Luong and Weinthal 2010; Haber and Menaldo 2011), much has been made of the ability of an increasingly authoritarian Russia to capitalize on its natural resource wealth, especially its oil and gas reserves (Goldman Page 58 →2010). As oil prices rose in the 2000s, President Putin freed his country from its reliance on foreign financial assistance (Hill 2004, 15), real disposable income and consumer spending in Russia jumped dramatically, and unemployment dropped appreciably (McFaul and Stoner-Weiss 2008, 78; Rutland 2008a). Treisman (2011) argues that energy exports at rising prices were a key explanation for Putin’s rise in popularity and power. We expect, then, that regions rich in natural resources may be different, just as resourcerich countries can be. To control for this possibility, we calculate available resource rents by composing a scale using the standardized values of regional levels of oil and natural gas production, two natural resources that received the most credit for Russia’s economic boom during the Putin era (Hill 2004; McFaul and StonerWeiss 2008; Rutland 2008b). Although we transform these data using the natural log, which reduces the influence of the high-output regions, the results were similar using the unmodified variable (unreported but available from the authors). Appendix 1 further explains and provides sources for these variables. Pressure in favor of democracy should be more likely to arise when a society has a relatively large proportion of people who expect that democracy will enhance their life prospects and those of their children. Yet democracy provides more tangible benefits to some members of a society than others. Urban and highly educated citizens are most likely to derive benefits from democracy’s intangible values (Reisinger et al. 1994; Miller, White, and Heywood 1998; Gibson 2001). These citizens also possess resources—proximity to like-minded others, intellectual and communication skills to effectively organize and promote their interests—that help them place demands on their government, increasing the extent to which a democratic system is beneficial in tangible ways. If present in large numbers, such citizens increase the likelihood that nascent democratic institutions will stabilize. Although individual-level relationships need not be found at higher levels of aggregation,4 these reasons have

been used to explain the empirical pattern that democracy occurs most commonly in countries with larger proportions of such people (Lipset 1959; Inkeles and Smith 1974; Burkhart and Lewis-Beck 1994; Teorell 2010). Hence, the variety of democratic success across Russia’s regions might be due to the variety in the type of people residing there. This line of argumentation, known broadly as modernization theory (Lerner 1958; Lipset 1960; Deutsch 1961; Inglehart and Welzel 2010), links urbanization and education levels to broader socioeconomic changes, including industrialization and postindustrialization. From its beginning, this literature has frequently employed crossnational or other cross-sectionalPage 59 → comparisons, with key variables such as level of industrialization measured at a single time. Yet the measurements from a single time are understood to reflect cumulative changes over decades or longer. Modernization theory proposes that those cumulative changes will underlie a crosssectional correlation between modernity and democracy. We can look for that pattern across the Russian regions, using our data to capture the variation among the regions resulting from social changes up until the time of Russia’s transition from Soviet rule. Although several recent studies provide evidence that modernizing trends in the nineteenth and early twentieth centuries influence patterns even today (e.g., Darden and GrzymalaBusse 2006; Remington 2011; Lankina 2012), we are only concerned with the pattern of socioeconomic circumstances facing political actors during the post-Soviet period. The social and economic changes highlighted by modernization theory tend to be strongly correlated with each other, creating problems for statistical analyses that employ them as separate explanatory variables. To address this, we will use an index variable made from seven features of each region that are highly intercorrelated and provide a useful single measure of each region’s overall socioeconomic development or modernization. The seven are: the percent of the region’s population with at least some higher education; the population density; the percent of the region’s residents living in cities or urban areas; residential telephones per 1,000 urban residents; the density of hard-surfaced roads, in kilometers per 1,000 square kilometers of territory; the number of theater-goers per 1,000 residents; and the number of museum-goers per 1,000 residents. The procedure for combining them is explained in appendix 1. Moscow and St. Petersburg receive markedly higher scores on our development index than the other regions both because they are purely urban regions and due to their historical development as the two capitals of Russia.5 For this reason, we report coefficients not from ordinary least squares regression but from robust regression, which down-weights the impact of extreme outlier cases when calculating coefficients (Andersen 2008). While the establishment of the new regime may not have been their doing, Russia’s regional elites still had plenty of room to make decisions that would influence their regions’ trajectories (McAuley 1997). These include tackling questions about institutional choice (Moraski 2003, 2006a; Bilev 2013) and how regional variations in election rules and political organizations might influence regional competition levels. From Duverger (1954) on, scholars have identified differential effects on electoral behavior under first-past-the-post electoral systems and absolute majority (or double-ballot)Page 60 → contests. Specifically, where runoffs are permitted, more candidates are likely to seek office on the belief that they have a shot of making it into the second round. Part of the motivation behind this decision is the expectation that voters are less likely to vote strategically since the rules enable them to pick the lesser of two evils in the runoff (see also Rae 1967; Taagepera and Shugart 1989; Cox 1997). In Russia’s regions, specifically, work by Bilev (2012, 2013) suggests that the different systems significantly influenced levels of gubernatorial competition levels. In established democracies, political parties are crucial for structuring and shaping electoral competition. However, Russia’s postcommunist party system struggled to produce parties with the organizational clout to play those roles (Moser 1999; Yargomskaia 1999; Rose 2000; Malinova 2001; Slider 2001; Smyth 2006; Hanson 2003; Fish 2003; Hale 2006). In many of Russia’s regions, political parties developed slowly because elites turned instead to informal patronage networks to mobilize votes. Yet the degree to which political parties have been vehicles for candidates has varied across the regions, especially during the Yeltsin era (Golosov 1999; Slider 2001; Golosov 2004). One should expect greater party development to correlate with greater levels of democracy by providing more formal links between the public and elected officials as well as more opportunities to hold politicians accountable. We measure regional party development as the percentage of deputies in regional

parliaments affiliated with a political party. This variable provides an indication of the degree to which regional elites were operating outside traditional patronage relations and parties were fulfilling their expected role as vehicles linking elites and voters (Weiner and LaPalombara 1966, 400; Fish 2003, 187). At first glance, focusing on regions within a single country might lead one to conclude a certain degree of uniformity on questions associated with state borders and the definition of the nation. Yet as already discussed, Russia’s status as an ethnofederal state (Hale 2006) and the fact that this institutional arrangement contributed to the splintering of other former communist states (Bunce 1999) caution against such a conclusion. Yeltsin’s willingness to grant republics special privileges (Kahn 2002) and his decisions to go to war in Chechnya, twice, poignantly illustrate that Russia has yet to resolve the nationalities question. Although Yeltsin’s decision to encourage republics within Russia to take as much sovereignty as they could swallow may have helped him in his battle with Gorbachev, it also led some republics to begin the process of writing their own constitutions. As a result, not only were republics exempted from President Yeltsin’s appointment of regional chief executives, but a Page 61 →few, like Udmurtia, postponed creating popularly elected executives for many years, while one, Dagestan, passed on this option altogether. One marked difference among Russia’s regions, then, was not so much whether the regions had popularly elected governors (all but one of them did) but how much experience the regions had with gubernatorial elections. In other words, although Russia’s regions held largely the same number of regional legislative elections and participated in the same number of national elections, their number of gubernatorial elections varied. President Yeltsin’s decision to phase in direct gubernatorial elections in Russia’s nonrepublic regions also added to the variation. This variation is noteworthy since work by Lindberg (2006, 2009) finds that repeated elections themselves can yield democratizing effects above and beyond the institutionalization of electoral processes. At the same time, some comparative studies of democratization (e.g., Huntington 1996; Bova 1997) have argued that the ethnic composition of a polity influences its propensity to democracy because of the differing historical traditions that ethnic groups inherit. And, of course, much center-periphery conflict in post-Soviet Russia has been related to ethnic claims for self-determination (Lehmann 1997; Lapidus 1999; Kolsto 2000). Although less ethnically complex than the USSR, Russia contains many sizable non-Russian ethnic communities. Ethnic Russians comprise about 75% of the Russian citizenry, but their presence in the regions ranges from 10% in Dagestan to 97% in Lipetsk, Kursk, Tambov, and Orlov. Fourteen regions have a minority of ethnic Russians. Russia’s republics, in particular, have enjoyed politically relevant resources of power that have been unavailable to regions with other appellations. First, Russian president Boris Yeltsin did not appoint chief executives in the republics in 1992 as he did elsewhere, which granted republican governments greater independence during Russia’s transition. Second, the bilateral treaties that emerged between the republics and the center provided republics with an opportunity to negotiate special prerogatives. Third, most republics adopted their constitutions without significant federal oversight. To control for the different levels of sovereignty that characterized the early post-Soviet period, we employ a dichotomous variable where 1 identifies republics and 0 otherwise. Doing so allows us to differentiate republics from other regions on the basis of the enhanced status that the former enjoyed, including the ability to determine when (or even whether) to hold direct elections for the regional executive. Below we take this one step further and distinguish between republics where incumbent leaders got out in front of liberalization, by holding elections early on (i.e., prior Page 62 →to the adoption of the 1993 constitution), versus those where direct elections for their regional executives were postponed until after the country’s national political trajectory was more firmly established. Such timing matters because, according to the literature, it can play a critical role in the preservation of authoritarian practices. From our perspective, the fast-track to elections in republics is reminiscent of the “liberation from above” that occurred across the Soviet Union.6 To the extent that Yeltsin’s concessions to Russia’s republics resemble a negotiated transition (i.e., a pact between the national government and vital regional interests),7 the historical record for such transitions is at best a mixed bag. The conventional view is that negotiated transitions advance democracy’s prospects as vital economic, political, and societal (or, in this case, regional) interests buy into the new system (O’Donnell and Schmitter 1986).

However, Bunce (2003, 173) argues that “transitions in the postcommunist region that combined pacting with demobilized publicsВ .В .В . were precisely the transitions that were most likely to continue authoritarian rule.” One mechanism, so far rarely considered, is the possibility that concessions granted by the central government to regional interests may have contributed to the rise of authoritarian enclaves.8 A more common consideration in work comparing Russia’s regions is the variation in the size of their nonRussian populations. In previous research, we have found that regional support for the Kremlin’s preferred party or candidate correlates strongly with the percentage of non-Russians in the regions (Reisinger and Moraski 2010). This result may reflect the ability of Russia’s republics to build political machines and authoritarian regimes within their borders thanks to the liberty they enjoyed from the federal government during the Yeltsin years. Hale (2003) suggests that ethnic minority mobilization has been higher in republics with a larger concentration of non-Russians. More specifically, the geographic concentration of ethnicity provided governors a capacity not only to monitor the ethnic vote but also to reward members of the titular ethnic group with “preferential treatment in education, state employment, territorially concentrated investment, and status” (Hale 2007, 231). Saikkonen’s (2016, 449) analysis of the regions’ political trajectories tentatively supports this contention. She finds that the effect of minority ethnicity on the emergence of hegemonic regimes in Russia was stronger in the republics than in other regions, though the effect is significant at only the .10 level. It is notable as well that, at least in the 1990s, the political machines in Russia’s ethnic regions tended to be nonpartisan; they substituted for political parties Page 63 →(Hale 2006, 2007). Our models will therefore include the ethnic composition of Russia’s regions.

Analyses We begin by distinguishing between elections taking place after the adoption of the 1993 Constitution and those held during the early transition years. Elections held prior to December 1993 operated in a qualitatively different context than later elections. In several republics (Adygeya, Chechnya, Kabardino-Balkaria, Marii El, Mordovia, Tatarstan, Yakutia) and in the two cities of Moscow and St. Petersburg, valid elections preceded Mikhail Gorbachev’s resignation on December 25, 1991. In a limited number of other regions (the oblasts of Amur, Bryansk, Chelyabinsk, Krasnoyarsk, Lipetsk, Penza, Orlov, Smolensk), Yeltsin permitted his appointees to test the electoral waters on April 11, 1993, prior to the national referendum held April 25. Russia’s October 1993 Constitutional Crisis and the subsequent December 1993 elections and constitutional referendum provide a natural endpoint to the transition period. First, the crisis gave President Yeltsin the upper hand in pushing through Russia’s 1993 Constitution. Yeltsin also used it as an opportunity to remove recently elected governors of their duties, specifically Lodkin in Bryansk and Sumin in Chelyabinsk. And second, the 1993 Duma elections inaugurated genuine multiparty competition in the Russian Federation. Although some elections were held in 1994, including North Ossetia’s in January, gubernatorial elections become more widely institutionalized in 1995 and 1996. Table 3.1 provides descriptive statistics for our measures of electoral competitiveness: Golosov’s index of effective candidates and turnover. Gubernatorial candidates were classified as incumbents if they headed the regional executive going into the election in question (i.e., were appointed governors or were chairs of regional Supreme Soviets).9 Given the different temporal contexts, table 3.1 also presents the results of t-tests assessing whether elections held prior to the December 1993 Constitution were significantly less competitive than elections held during the rest of the Yeltsin period.10 According to the t-tests, they are not. Given that, one approach would be to pool all of the Yeltsin-era elections and analyze them collectively. Unfortunately, data from elections during the 1991–93 period are less complete than for those held in 1994 and after. Most notable is the absence of comparable party development Page 64 →data. For cases between 1994 and 1999, we measure regional party development as the percentage of deputies in regional parliaments affiliated with a political party. These data are not available for the 1991–93 periods. Also missing are data on the type of electoral system—relative or absolute majority—governing ten gubernatorial elections during the early years of transitions.11 At the same time, while the results of table 3.1 suggest that competitiveness levels for gubernatorial elections across the two periods (i.e., with and without party system data) fail to differ significantly

from one another, one cannot say the same for cases with and without electoral system data: the average score on the Golosov index for the 10 elections with missing data is 1.28 while the average score for the remaining 120 elections is 1.89. This difference in means is significant at the .001 level for a two-tailed test.12 With these limitations in mind, we focus our multivariate analysis of competitiveness levels on elections held after January 1994.13 Of course, elections from the 1991–93 period still matter, and the results of these elections also inform our analysis. Before beginning the multivariate analysis, we pause to consider one important question about these elections: Did a difference in constitutional status—specifically, being a republic—influence the competitiveness Page 65 →of a region’s gubernatorial elections during the initial years of transition? This question flows from the debate about whether the republics’ different status and powers helped their elites build authoritarian political machines. Tables 3.2 and 3.3 present the average values of our competitiveness measures—effective number of candidates and turnover, respectively—for elections held in republics versus the average for elections in regions holding a different constitutional status.14 Each table presents one column with information from the 1991–93 period and another for the 1994–99 period. Table 3.1. Descriptive Statistics for Yeltsin-Era Gubernatorial Competitiveness and Difference of Means between Elections Held in 1991–93 versus 1994–99 Effective Number of Candidates Turnover Average 1.84 0.49 Standard Deviation 0.71 0.50 N 130 130 Maximum 1 4.69a Minimum 0 1.00b Percent of Elections with Turnovers, 1991–99 — 49.2% N 64 Average for Elections Held during 1991–93 1.75 0.59 N 22 22 Average for Elections Held during 1994–99 1.86 0.47 N 108 108 t-test of Difference between Periods в€’0.64 1.42 Score for Russian Presidential Election, 1991 1.67 Score for First Round of Russian Presidential Election, 1996 2.75 Score for Russian Presidential Election, 2000 1.84 Note: The effective number of candidates is estimated using Golosov’s (2010) method. a

The Republic of Mordovia (1991).

b

The republics of Ingushetia (1993), Kabardino-Balkaria (1997), Kalmykia (1995), Karelia (1994), and Tatarstan (1991, 1996). According to table 3.2, no statistically significant difference emerges between the effective number of candidates in republican gubernatorial Page 66 →elections and those in nonrepublic regions during the 1991–93 period. While the average effective number of candidates was lower in the 1991–93 period than the 1994–99 period (1.75 compared to 1.86), the average is actually higher for republics than for other regions during the 1991–93 period. This relationship then reverses for elections held between 1994 and 1999. In other words, republics became less competitive than other regions only during the 1994–99 period, with the difference of means between the types of regions attaining significance at the .10 level for a two-tailed test. On the other hand, the

percent of elections producing turnover was higher in the early elections than in the later ones (see table 3.3). Despite no statistically significant difference in the effective number of candidates during the 1991–93 period, turnover was significantly less likely (at the .01 level for two-tailed tests) to occur in republics than in other regions during this time period. That is, republican incumbents during the early years of transition were more likely to defeat their challengers (despite facing more effective competitors than their counterparts in other regions). Yet this difference was temporary: the effect fails to hold for the 1994–99 period. The following multivariate analyses help explain these results. Table 3.2. Difference of Means in Effective Number of Candidates for Gubernatorial Elections, Republics versus Other Regions across Periods 1991–93 1994–99 Regional Average 1.75 1.86 Standard Deviation 0.90 0.67 Maximum 4.69 4.08 Minimum 1 1 N 22 108 Average for Republics N Average for Nonrepublics N t-test +

1.83 12 1.67 10 0.41

1.64 23 1.92 85 в€’1.77+

Indicates significance at the .10 level for two-tailed tests.

Table 3.3. Difference of Means for Turnover in Gubernatorial Elections, Republics versus Other Regions across Periods 1991–93 1994–99 Regional Average 0.59 0.47 Standard Deviation 0.50 0.50 N 22 108 Percent with Turnover 59.1 47.2 N 13 Percent without Turnover 40.9 N 9

51 52.8 57

Average for Republics N Average for Nonrepublics N t-test

0.39 23 0.49 85 в€’0.875

**

0.33 12 0.90 10 в€’3.26**

Indicates significance at the .01 level for two-tailed tests.

A first step in the multivariate analysis is differentiating the regions on the basis of their experience with gubernatorial elections. While the vast majority of Russia’s regions began holding gubernatorial elections

between 1994 and 1997, two republics—Dagestan and Udmurtia—defied the norm altogether. Dagestan relied on a parliamentary executive until the imposition of gubernatorial appointees during Putin’s second term in office. Udmurtia, meanwhile, did not institute an independently elected governor until 2000, after President Yeltsin left office.15 On the other end of the spectrum is the city of Moscow, which held three valid gubernatorial elections prior to Yeltsin stepping down.16 In between these extremes are four regions (all republics) for which the last Yeltsin-era elections were not only inaugural gubernatorial elections but founding elections—that is, the election was to fill a newly created position and not to replace a Yeltsin appointee. To capture these differences in electoral experience, we employ a dichotomous variable where a score of 1 indicates that an election was a region’s first. We also continue to differentiate between regions that are republics from other regions by utilizing a dummy variable where a score of 1 indicates a republic. With these variables in hand, we then distinguish republican founding elections from first elections in nonrepublic regions—that is, in regions where a Yeltsin-appointed governor was the incumbent—by including an interaction term that multiplies “first election” and “republic.” Page 67 →Together, these variables control for a lack of independence among the cases included in the analysis (i.e., multiple elections from the same region) since all of the regions except the city of Moscow held only one or two elections during the Yeltsin era. Our analysis of Yeltsin-era gubernatorial elections includes several additional independent variables designed to capture the expectations detailed in the previous section. To control for variations in ethnic composition, we include the percentage of non-Russian residents in each region based on 1989 census data. Also, like Saikkonen (2016), we examine the possibility that the effect of non-Russian ethnicity mattered more in republics than in other regions by including an interaction term that multiplies percent non-Russian by the dichotomous variable for republic. In addition to the interaction terms and their components, the models include the regional development index and the measure of regional oil and gas production. As indicators of the regions’ evolving economic and social conditions, we use average monthly income per capita (measured as a percent of the previous year’s income), regional unemployment levels, and the number of violent crimes per 1,000 people. Aside from regional development and socioeconomic conditions prior to the elections, our investigation considers variation in the electoral rules and regional party development. In some regions, gubernatorial candidates had to receive an absolute majority of the votes cast to be declared the winner. If they did not, then the two top votegetters would compete in a second round (or runoff). In other regions, however, a candidate could win with a relative majority (or plurality) of the votes (see Central Electoral Commission of the Russian Federation 1997, 2001). We use a dichotomous variable where a score of 1 indicates rules that require a second round if no candidate wins an absolute majority to control for differences in the electoral rules. We measure regional party development, meanwhile, as the percentage of deputies in regional parliaments affiliated with a political party between 1994 and 1999. Table 3.4 presents three equations estimated using robust regression.17 In each instance, the dependent variable is Golosov’s index for the effective number of candidates in elections held after January 1994 but prior to January 2000. When possible, we measure the variables using data from the previous year for elections held prior to July 1 and data from the election year for those cases where the election was held July 1 or after.18 1996 data from Russia’s less populous autonomous okrugs are unavailable for income, unemployment, and violent crime. Since many of these regions also held elections in 1996 and 1997, excluding them reduces our N to 94.19 The first Page 69 →equation, then, tests the model with the full model and the smaller N due to missing data. To assess the effect of the missing cases on the analysis, the remaining equations exclude the variables with missing data, allowing for an analysis of 104 elections. Page 68 → Table 3.4. Regression Analysis of Effective Candidates in Russian Gubernatorial Elections, 1994–99 Equation 1 Equation 2 Equation 3

.203

в€’.193

в€’.548

First Election

(.732) .073 в€’.086 (.588)

(.690) в€’.230 в€’.080 (.581)

(.060) в€’.334 в€’.077 (.606)

Republic Г— First Election (Founding Election)

в€’.040 .409 (.278)

в€’.076 .467 (.180)

в€’.070 .566 (.103)

.264 .021

.307 .006

.321 .001

(.300) .063 в€’.010 (.329) в€’.136 в€’.170 (.027) в€’.090 в€’.477 (.028) в€’.189

(.782) .023

—

—

—

—

—

—

Republic

Non-Russians as % of Population

(.083) .313 −.029 Non-Russian × Republic (.065) −.699 −.214 Socioeconomic Development Index (.015) −.071 −.916 Natural Log of Oil and Natural Gas Production (.172) −.121 −.002 Violent Crime (.317) −.034 −.006 Real Income (.315) −.069 .005 Unemployment (.764) .178 −.001 Percent of Regional Deputies with Party Affiliation, 1994–99 (.728)

Absolute Majority (runoff)

Constant R2 Deviance Number of Cases

.093 .518 (.000) .296 1.954 (.001) .183 25.40 94

— −.161 (.042) −.092 −.420 (.056) −.182

в€’4.46e-4 в€’2.25e-4 (.873) (.938) .093 .419 (.001) .264 1.491 (.000) .147 27.10 104

.092 .390 (.003) .259 1.588 (.000) .135 29.13 104

Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Beta weights with an absolute value of .15 or higher are bolded.

Generally speaking, the models explain modest amounts of variation: the R2s range from .14 to .18. Since we are examining the population rather than a sample, we draw attention to the significance levels to indicate the strength of a coefficient’s value. In addition, we report standardized coefficients, drawn from ordinary least squares regression, which allows one to compare the relative substantive impact of each explanatory variable. The most consistent findings across the equations relate to our index of regional development and the use of runoff elections. Regarding the former, the table presents strong evidence that more developed regions did not produce more competitive races. In fact, the relationship between development and effective candidates is negative and significant at the .05 level for two-tailed tests. Although the standardized coefficients are not high, the results suggest that, if anything, competition levels are higher in regions at lower levels of development when holding other factors constant. This outcome contradicts expectations that competitive elections are more likely to take place in Russia’s more urban, better educated, and more connected regions. Like Bilev (2013), we also find that the electoral rules governing gubernatorial elections matter. Whether regions had a runoff is consistently significant in the expected direction with beta weights well above .15. In accordance with the literature on majoritarian elections, requiring a runoff in elections where no one candidate receives an absolute majority of the vote bred more competitive elections—in terms of the effective number of candidates—than elections where the victor is simply the person receiving the greatest share of the vote (i.e., plurality contests).20 According to Equation 1, the concentration of ethnic minorities in Russia’s republics also appears to shape competition levels: the interaction of percent non-Russian and republic is negatively correlated with the effective number of candidates. Although the effect is only significant at the .10 level for two-tailed tests, its effect is the most powerful of the variables in the equation with a standardized coefficient of в€’.70. Interestingly, while controlling for this interaction term and the other variables, regions with larger shares of non-Russians experienced greater competition, again significant at the .10 level and a relatively large beta weight (over .30). Meanwhile, the dichotomous variable distinguishing republics from other regions fails to meet standard levels of significance or demonstrate much of an impact according to its standardized coefficient. These findings initiallyPage 70 → support the contention that lower competition levels in republics are correlated with higher concentrations of minorities in these regions (see also Hale 2003; Saikkonen 2016). However, regression diagnostics reveal that using the interaction term introduces substantial multicollinearity to the model: the variable inflation factors (or VIFs) for the interaction term and its components are all above .10 with the VIF for the interaction term above .27. In addition, Equation 2 reveals that the findings from including the interaction term are not robust. Social and economic conditions more proximate to the gubernatorial elections—like income, crime rates, and unemployment—and party development fail to reach standard levels of significance in Equation 1 and only the beta for unemployment exceeds .15. These results suggest that variation in the number of candidates competing in Russia’s gubernatorial elections between January 1994 and January 2000 had little in common with conventional explanations from the democratization literature. The insignificant impact of resource wealth in Equation 1, meanwhile, appears to challenge assertions that the availability of that wealth will undermine competition levels.21 Nevertheless, this effect, like that of minority concentration in republics, varies across the equations. The regression results change appreciably when we remove the variables that both reduce the number of cases under investigation and also appear insignificant (see Equation 2). For one, the effect of resource wealth becomes significant at the .05 level. In addition, neither the interaction term for minority concentration in republics nor the percent of non-Russians proves significant in the respecified equation with the larger sample. Moreover, their standardized coefficients diminish dramatically. What might explain these changes? As alluded to previously, issues with data availability were most common for Russia’s autonomous okrugs during the earlier part of Russia’s post-Soviet transition. It is plausible, then, that these changes reflect the inclusion of additional okrug elections (up from one to ten) after removing the variables that reduced the number of cases under investigation. Indeed, while many republics may have developed political machines and authoritarian regimes, Golosov (2011b, 401) points out that autonomous okrugs also possessed characteristics that may undermine competitive politics. These traits include small populations, dense ethnic networks, and, in some

cases, resource rents. On the latter, Khanty-Mansiisk and Yamal-Nenets reside at the high end, though both were excluded from Equation 1, which likely explains the stronger effect of resource wealth in Equations 2 and 3. With these considerations in mind, we reestimated Equation 2. First, we Page 71 →changed the interaction term for minority concentration so that it defined ethnic regions more broadly (i.e., republics and okrugs) and added this new component (a dichotomous variable for ethnic regions) to the model. This modification failed to alter the findings, however, aside from reducing the significance level of resource rent availability (unreported but available from the authors). Given this null finding and the multicollinearity in the model, we removed the interaction term for minority concentration while leaving the percent of non-Russian residents in the model (see Equation 3). The results of this analysis are particularly noteworthy because, under this simpler specification, both the variable for republic proves significant at the .10 level while it and the variable capturing republican founding elections yield standardized coefficients above .30, higher than that of the dichotomous variable for runoff elections. The results of Equation 3, then, suggest that the timing of elections, during a transition or after it, may influence competition levels. First, incumbents in republics where elections were held while Russia was still transitioning from Soviet rule (i.e., prior to 1993) faced fewer challenges in the next round of elections (after 1993) than did leaders of other regions, even when controlling for percent non-Russian. In addition, Equation 3 highlights a possible difference between republican founding elections and other gubernatorial elections: on average, competition was likely higher in initial elections where no previously elected incumbent, or presidential appointee, held the post of governor. This finding suggests that Kremlin support during the Yeltsin era may have dampened the amount of competition Yeltsin appointees faced. Since we explore the possibility of a similar dynamic during Putin’s first term in the following chapter, this point merits particular attention. Taken together, then, the regression analysis underscores the value of hypotheses developed for the specific regional contexts under investigation, as opposed to general theories taken from cross-national studies, for understanding subnational politics in Russia. In table 3.5, we move from an analysis of candidate number to turnover. Since the effective number of candidates competing in an election should influence the probability of electoral turnover, the preferred estimation technique would be a structural (or simultaneous) equation model (SEM). However, SEM is a large sample technique, requiring Ns above 200 (Kline 2005, 111), which is significantly more than the number of cases at our disposal. Instead, we present two sets of logistic regression equations: each set presents models with and without Golosov’s index of effective candidates. Table 3.5 presents two logistic regression analyses for gubernatorial turnover between 1994 and 1999. These equations use the full model presented in Equation 1 of table 3.4, except for the interaction term for minorityPage 72 → concentration in republics.22 The model without the effective number of candidates produces a reasonable pseudo-R2 and a statistically significant likelihood ratio П‡2. In the second equation, which adds the Golosov Index, both the pseudo-R2 and the likelihood ratio П‡2 improve. Among the independent variables included in the models, levels of violent crime and the interactive term denoting republican founding elections are consistently significant. Note that odds ratios range from 0 to infinity with values below 1 indicative of negative relationships and those above 1 capturing positive relationships. Dividing 1 by the odds ratio makes the negative relationships comparable to positive ones. Thus, the odds of turnover were 24 to 35 times higher in a region that was a republic holding a founding election when compared to other regions. Meanwhile, Page 73 →for each additional violent crime per 100,000 people in a region, the odds of turnover decreases by about 2% (e.g., 1/.985 = 1.02). The finding that an increase in violent crime dampens the likelihood of turnover supports the assertion that regions suffering from higher levels of violent crime are unlikely to possess a social climate supportive of the kinds and levels of civic engagement to facilitate political accountability via elections. The outcome that gubernatorial turnover was more likely in republican founding elections extends a key finding from table 3.4. To help interpretation, figure 3.1 provides a graphical representation of the interaction term’s results. As it reveals, not only is the effective number of candidates higher in republics that postponed elections relative to other regions, but so too is the likelihood of turnover, and this occurs even when controlling for the effective number of candidates (Equation 2 in table 3.5). Moreover, status as a republic alone had no direct effect on the likelihood of gubernatorial turnover in elections held during the 1994–99 period nor did the

percent of non-Russians in a region. Table 3.5. Logistic Analyses of Turnover in Russian Gubernatorial Elections, 1994–99 (Full Model) Equation 1 Equation 2 Odds Ratio (p 95% Confidence Interval Odds Ratio (p value) 95% Confidence Interval value) Republic .251 (.370) .012 5.173 .614 (.762) .026 14.484 First Election .670 (.518) .199 2.257 .721 (.625) .194 2.677 Republic × First Election (Republican 35.366 (.028) Founding Election) Non-Russians as % of .996 (.897) Population Development .642 (.159) Index Natural Log of Oil-Gas .009 (.143) Production Violent Crime .985 (.021) Real Income 1.034 (.183) Unemployment .950 (.383) Percent of Regional Deputies with 1.020 (.119) Party Affiliation, 1994–99 Absolute Majority 3.232 (.042) (Runoff) Effective Number of — Candidates Constant .081 (.276) .200 Pseudo R2 Likelihood 25.99 (.007) Ratio χ2 (significance) Number of 94 cases

1.456

859.046 24.361 (.073)

.745

796.475

.933

1.062

.991 (.783)

.925

1.060

.347

1.190

.672 (.335)

.299

1.508

1.86e-5

4.816

.019 (.228)

3.22e-5

11.779

.973 .984 .846

.998 1.087 1.066

.982 (.013) 1.050 (.088) .614 (.263)

.968 .993 .814

.996 1.111 1.058

.995

1.045

1.016 (.216)

.991

1.043

1.041

10.030

1.587 (.455)

.472

5.341

5.185 (.002)

1.843

14.582

.003 (.043) .298

1.1e-5

.831

.001

7.440

38.71 (.000) 94

Note: Numbers in parentheses are p-values. Odds ratios in bold are significant at conventional levels for a twotailed test. Fig. 3.1. Adjusted Predictions of First Election, by Republic, on the Likelihood of Turnover (Full Model).

The margins plot uses Equation 2 from table 3.5 with all other variables held at their means and presents 90% confidence intervals. Among the remaining variables in the models, the electoral rules and the number of candidates competing in the elections both increase the odds Page 74 →of turnover with the runoffs improving the odds by three times and one additional effective challenger increasing the odds by five times. However, the former only seems to matter to the extent that it determines the latter. When the index of effective candidates is used alongside the dichotomous variable for elections governed by absolute majority rules, the independent effect of the rules on turnover dissipates. While the models in table 3.5 reveal that the dichotomous variable, republic, does not influence the likelihood of turnover, it is possible as well that republican status has an indirect effect on turnover since being a republic is a significant predictor of the effective number of candidates (Equation 3 in table 3.4). Interestingly, turnover also may be more likely in regions where real income is higher. The relationship should be taken with a grain of salt, however, since it emerges only in the second equation and then only at the .10 level. Notably, the availability of resource rents seems to have no direct effect on turnover, operating primarily through the dampening of the effective number of candidates. As table 3.4 illustrates, excluding most of the okrugs from the analysis has drawbacks. Given the significance of several independent variables that also reduce the number of cases under investigation in the logit models, any analysis of models that excludes these variables so as to increase the number of cases should be interpreted with caution. We therefore discuss the results presented in table 3.6—with a reduced model and full set of cases—primarily as it relates to the results in table 3.5. The equations in table 3.6 differ from those in table 3.5 in a few notable ways. First, the amount of variance explained drops relative to the full models, as evidenced by the respective declines in the pseudo-R2s: from .20 to .14 in the equations without the number of candidates and from .30 to .24 for the equation with it. In addition, the interaction term for republican founding elections has a much weaker effect in these models. While it is still significant and increases the odds of turnover by twelve times in the model without the effective number of candidates, the effect decreases to five times—and is no longer statistically significant—when the reduced model controls for the number of candidates. As figure 3.2 illustrates, while the likelihood of turnover in the 1994–99 period is higher for republics where the election during this period was the region’s first, the confidence interval for such republics is much wider in this figure than it is in figure 3.1. Note as well that the number of candidates continues to shape the odds of turnover, though the impact drops slightly. A more surprising change is the significant effect of our measure of party development in both equations. These results suggest that social Page 75 →and economic conditions may operate through their influence on party development to determine the likelihood of turnover since the effect only emerges when these conditions are not in the model. To determine whether this is the case or whether the change reflects the addition of the missing okrugs, we estimated the reduced model on the same 94 cases analyzed in table 3.5 (unreported but available from the authors). Party development continues to matter, suggesting that while neither party development nor levels of development and regional conditions influence turnover independently, they may do so jointly. Another noteworthy difference is the significance of the oil and gas index in the first equation. As discussed already, this change makes sense given that two of the most resource-rich regions are okrugs and were Page 76 →excluded from the analysis in table 3.5. Yet the fact that the effect disappears when included alongside the effective number of candidates suggests that the availability of resource rents only indirectly shapes turnover. This outcome lends support for the findings in equations 2 and 3 in table 3.4 that resource wealth in the regions reduces the number of candidates. Table 3.6. Logistic Analyses of Turnover in Russian Gubernatorial Elections, 1994–99 (Reduced Model) Equation 1 Equation 2 Odds Ratio (p 95% Confidence Interval Odds Ratio (p value) 95% Confidence Interval value)

Republic

.184 (.138)

.020

1.721

.387 (.431)

.037

4.104

First Election Republic Г— First Election (Republican Founding Election) Non-Russians as % of Population Development Index

.741 (.584)

.254

2.165

.842 (.774)

.261

2.720

12.356 (.066)

.847

180.357 5.284 (.265)

.283

98.714

1.010 (.590)

.975

1.046

1.008 (.660)

.972

1.047

.711 (.228)

.408

1.238

.757 (.454)

.364

1.571

4.64e-5

1.801

.011 (.117)

3.81e-5

3.113

Natural Log of Oil-Gas .009 (.082) Production Violent Crime — Real Income — Unemployment — Percent of Regional Deputies with 1.028 (.020) Party Affiliation, 1994–99 Absolute Majority 2.093 (.130) (Runoff) Effective Number of — Candidates Constant .205 (.036) 0.135 Pseudo R2 Likelihood 19.37 (.013) Ratio χ2 (significance) Number of cases

104

— — —

1.004

1.053

1.027 (.025)

1.003

1.051

.805

5.443

1.287 (.636)

.452

3.663

4.756 (.001)

1.923

11.764

.014 (.000) .236

.001

.139

.047

.902

33.98 (.000) В

В

104

Note: Numbers in parentheses are p-values. Odds ratios in bold are significant at conventional levels for a twotailed test. Fig. 3.2. Adjusted Predictions of First Election, by Republic, on the Likelihood of Turnover (Reduced Model). The margins plot uses Equation 2 from table 3.6 with all other variables held at their means and presents 90% confidence intervals.

Discussion In this chapter, we distinguished among Russia’s regions on the basis of their electoral competitiveness in the Yeltsin era. The analysis incorporates variables common to cross-national studies of democratization and

transitions as well as variables specific to the post-Soviet context governing regional politics. Throughout the 1990s, Russia’s regions varied appreciably in terms of their levels of socioeconomic development and performance, natural resources within their borders, choice of electoral rules, and progress in party building. At the same time, the politics of center-periphery Page 77 →relations influenced the levels of sovereignty that different regions enjoyed, the timing of their first executive elections as well as whether a presidentially appointed incumbent held office at the time of the election. On the whole, we find limited support for core explanations in the comparative democratization literature: longterm structural development matters but only in ways that undermine competition. The effective number of candidates is lower not only where resource rents are available but also in regions with higher levels of socioeconomic development. By comparison, explanations more closely associated with the literature on transitions—that is, declining social order and collective trust (measured as violent crime levels), institutional choice, and election timing—fare better. Indeed, perhaps the most interesting findings come from contextspecific hypotheses, captured by differences in constitutional status, the timing of founding elections, and the difference between founding elections and other elections (i.e., where presidential appointees held office). Table 3.7 presents descriptive statistics that summarize the main context-specific findings. In elections held during Russia’s initial post-Soviet transition, from late 1991 through the constitutional crisis and first Duma elections in 1993, incumbent turnover was significantly less likely in republics (i.e., regions told to take as much sovereignty as they could swallow) than in other regions (see the top panel of table 3.7). Notice as well that this difference exists even though the effective number of candidates was higher in republican elections than elections held in other regions during this period. For the victors of republican elections held between 1991 and 1993, the decision to hold elections early on paid off in subsequent elections. In the 1994–99 period, republican elections with elected incumbents on the ballot saw fewer effective candidates and a lower probability Page 78 →of turnover than other gubernatorial elections held during this period. In fact, the figures are even less than in elections held between 1991 and 1993.

Republics Other regions

Table 3.7. Descriptive Statistics Illustrating Context-Specific Findings Average Score on Golosov Index Probability of Turnover 1991–93 1.83 33% 1.67 90%

Republican elections with elected incumbents First elections in other regions (with appointed incumbents)

1994–99 1.43

20%В

1.90

48%В

Elections in other regions with elected incumbents 1.96 First republican elections (i.e., founding elections) 2.05

52%В 75%В

More surprising, however, is that both the effective number of candidates and the likelihood of turnover, even when controlling for the number of candidates (see table 3.5), were highest in republican founding elections during the 1994–99 period than in other elections. As table 3.7 highlights, republics that held founding elections between 1994 and 1999 witnessed more competitive contests than other republics with elected incumbents or regions where an appointed governor held office. This finding emphasizes that ethnicity alone cannot explain differences in regional competition levels since Russia’s ethnic republics comprised both the least competitive group in table 3.7 and the most competitive, with electoral timing serving as the key characteristic distinguishing among them. Our results add a new dimension to Bunce’s (2003) argument that negotiated transitions in the postcommunist region facilitated the continuation of authoritarian rule. In Russia, at least, Yeltsin’s concessions to the republics may have given him greater leverage vis-à -vis Gorbachev and it likely helped

preserve the federation while other socialist federations collapsed, but these gains came at the expense of electoral competition. As tables 3.3 and 3.7 demonstrate, turnover was significantly less likely in republican elections between 1991 and 1993 than in elections held in other regions during this period. At the same time, these earlymovers appear to have consolidated their positions going into the 1994–99 period. As table 3.4. emphasizes, nonfounding republican elections witnessed significantly less competition—measured by the effective number of candidates—in their gubernatorial contests between 1994 and 1999 than other regions, including other republics where elites postponed electing a regional chief executive. These results, then, suggest that in a number of cases republican leaders responded to the national liberalization process by getting ahead of it: they held elections early, and those who proved victorious in these elections faced significantly less competition. Moreover, since fewer effective candidates translated into a lower likelihood of defeat in elections once the new federal regime had been established (table 3.5), these incumbents appeared more insulated than their counterparts who waited to hold direct elections until after elections had become the norm. These results offer an explanation for the range of electoral competitiveness within Russia’s republics that go beyond their ethnic composition. While ethnic composition justified and facilitated the holding of early elections, agency mattered as well.23 Roeder’s (2007) analysis of Russia’s republic as segment-states augments Page 79 →this point. According to his work, differences in electoral timing across the republics reflected the ability of leaders to establish “political identity hegemony” in their regions, which is critical if a segmentstate wishes to gain concessions from the national government. Elections were first held in republics where leaders controlled both party and state apparatuses (e.g., Valerii Kokov in Kabardino-Balkaria and Mintimer Shaimiev in Tatarstan). They were followed by republics where regional bosses established political machines after an initial power struggle, though still early on in Russia’s transition (e.g., Rakhimov in Bashkortostan and Iliumzhinov in Kalmykia). Elections were last held in republics where leaders were unable, or took longer, to establish such hegemony (ibid., 99–103).24 Finally, it is noteworthy that turnover in the 1994–99 period was significantly more likely (with a 75% probability) in first elections with no presidentially appointed incumbent than in first elections where a presidentially appointed governor held office (48%). This not only indicates the wisdom of Yeltsin’s decision to postpone additional gubernatorial elections in the regions until after the country’s political trajectory was better established, but it also provides some indication that, during the Yeltsin-era at least, the Kremlin’s backing could help insulate incumbent governors.

Conclusion As we highlight in chapter 1, examining Russian politics through a regional lens grants us an opportunity to compare the performance of expectations drawn from the general literature on transitions and democratization alongside context-specific hypotheses. It is worth emphasizing that the latter hypotheses correspond to characteristics that make the study of democratization across Russia’s regions substantively different from cross-country studies of democratization. As Gibson (2012, 12) notes, questions of sovereignty and vertical interaction between levels of governmental and nongovernmental actors matter. In the Russian case, how these questions were resolved influenced electoral timing and constitutional status. Accordingly, this chapter has analyzed different theories associated with variations in electoral competition using data from Russia’s gubernatorial elections during the Yeltsin era. In contrast to the comparative democratization literature, our analysis finds that economic performance failed to systematically influence the competitiveness of gubernatorial elections during the 1990s while higher levels of socioeconomic development appear to have dampened competition rather than promoted it. On the Page 80 →other hand, the availability of natural resources works as expected, serving more as a curse than a blessing. In addition to these structural effects, our analysis also finds that decisions made during Russia’s transition mattered. As scholars of institutional design would expect, the adoption of runoff electoral systems significantly increased the number of candidates, though it had no independent effect on the likelihood of turnover. Since previous work by scholars of Russian politics have contended that differences in regional regime types

reflect either the presence of large non-Russian populations or variations in constitutional status, we strove to untangle the effects of these closely intertwined characteristics. While controlling for ethnicity, we demonstrate that an interaction effect between constitutional status and election timing shaped competition levels in the 1990s. We not only show that electoral competition varied among Russia’s republics, but we also contend that variation in electoral competition across the regions as a whole reflected differences in the opportunity and willingness of regional leaders to hold founding elections early on in Russia’s transition: holding founding elections early is something that republican politicians could do, but it is not something that all of them did. Thus, while differences in sovereignty among the regions matter, so too does agency. Finally, our investigation indicates that while regional levels of party development and violent crime levels did not determine how many candidates effectively competed in gubernatorial elections, they did affect the likelihood of turnover. If higher levels of violent crime, and thus lower levels of societal trust, diminished the prospects for turnover, higher levels of party development increased them. However, since the latter effect varied depending on the model specification and the former could not be tested across the full range of cases, both findings are suggestive not definitive. As the twentieth century was coming to a close, electoral politics in the Russian Federation appeared to be more competitive than at any point in Russia’s history. The 1999 Duma election, in particular, was hotly contested as the Kremlin raced to assemble a new party of power in support of Russia’s new prime minister, Vladimir Putin, and counter a surging rival, Fatherland-All Russia, which was created by rival elites. Notably, these included several of Russia’s most powerful governors. Many gubernatorial elections, meanwhile, had turned into no-holds-barred contests as regional actors fundamentally understood where true political power in the regions rested (Moses 2002, 907). Tactics included using the media to distribute or encourage false and compromising stories about opponents (kompromat) and publishing phony editions of newspapers with the main stories Page 81 →discrediting one’s rivals (falshivki). The persistence of autocratic leaders in many Russian regions led some observers to classify these regimes as regional fiefdoms (Smirnov 2001, 526). Yet, to the extent that President Yeltsin relied on a decentralized version of federalism to hold the country together, the rise of such subnational authoritarian enclaves should not have been surprising. In this way, the Russian case parallels Argentina and the United States after Reconstruction, where, as Gibson (2012, 162) observes, the high levels of regime autonomy awarded to provincial authorities institutionalized and legally sanctioned subnational authoritarianism. Still, as this chapter highlights, by the late 1990s, genuine electoral competition had begun to emerge in some of Russia’s regions as many candidates effectively competed for the top post in gubernatorial elections while several incumbents lost reelection bids. In the next chapter, we examine attempts by Yeltsin’s successor, Vladimir Putin, to influence the outcomes of these elections.

Page 82 →

Four The Erosion of Competitive Gubernatorial Elections in Russia During the Yeltsin era, gubernatorial elections varied appreciably in terms of their levels of competitiveness. While governors in some regions appeared impervious to removal, other regions were characterized by relatively fierce electoral contests. In this chapter, we examine the correlates of competition in the gubernatorial elections held during President Putin’s first two terms in office. These occurred from March 2000 through February 2005, when Russia held its last gubernatorial election before moving to a system of presidential appointments. Putin’s September 2004 decision to halt gubernatorial elections was a decisive step in the country’s democratic regression. It is worth understanding, then, what convinced Putin to end direct gubernatorial elections and why his decision, which made regional executives accountable to the president rather than to voters, met so little public opposition. In examining these elections, we are particularly interested in when a sitting governor succeeded in being reelected. We begin by considering the same correlates examined in chapter 3. In this chapter, as in that one, we wish to understand how well hypotheses from established literatures like modernization theory, transitology, neoinstitutionalism, and resource wealth perform at the subnational level in Russia. At the time of its independence, Russia not only lacked historical experience with competitive elections genuinely determining access to political power, but its transition had left in power a large share of elites with no commitment to democracy. Page 83 →With these considerations in mind, we also incorporate indicators that capture attempts by both federal and regional leaders to manipulate electoral outcomes through legal but unfair means. Our findings suggest that, on the whole, strategic changes in the timing of elections failed to influence election results. Instead, the results reveal that Kremlin opposition to a sitting governor fueled competition levels, which in turn shaped the likelihood of turnover. However, Putin’s Kremlin was not able to discourage competitors from running against its favored incumbents. The inability of presidential preferences to systematically dampen competition levels in gubernatorial elections reveals the limits of Kremlin influence. Still, this immunity did not necessarily bolster the prospects for democracy. Regional characteristics commonly understood as shaping competition levels and election outcomes in democratic settings prove unrelated to either competition levels or election outcomes across Russia’s regions. For example, incumbents ruling over Russia’s poorer regions were no more likely to be turned out of office than those governing wealthier regions. Putin’s decision to end gubernatorial elections thus makes sense, as does the lack of a broad public outcry: if the president cannot use the electoral process to protect governors whom he supports while still other governors could prevail regardless of regional conditions, then neither the Kremlin nor significant portions of the population were likely to find much value in elections.

Gubernatorial Elections in Putin’s Russia No gubernatorial elections were held between Yeltsin’s 1999 New Year’s Eve resignation and Putin’s election in March 2000. Therefore, the period under investigation runs from March 2000 to February 2005, when Nenets autonomous okrug held the runoff from its January 23 election, the last gubernatorial election prior to the appointment era. Again, elections are our unit of analysis (N = 113). Because of our interest in the ability of incumbents to secure reelection, we exclude cases where the incumbents died in office, declined to run for reelection, or were barred from running for reelection. Such examples constitute 21 cases.1 We discuss them in appendix 2. While some of these cases may be interesting from the perspective that they might represent blatant violations of electoral fairness, we exclude them because they impede our ability to assess how well explanations from the comparative democratization literature perform when trying to understand competition levels and election results in Russia’s gubernatorial contests. As we note in chapter 1, we recognize that conditionsPage 84 → in Yeltsin’s Russia, especially during his second term, and in the first term of Putin’s reign were not democratic as much as what recent scholarship calls competitive (or electoral) authoritarian. Yet we employ independent variables drawn from the comparative democratization literature in order to address whether lessons

learned from the scholarship of comparative democratization apply to the study of elections in more authoritarian contexts, including subnational ones. As in chapter 3, we measure electoral competitiveness using Golosov’s index of effective candidates and the dichotomous variable, turnover, scored as one when an incumbent governor lost the election and zero otherwise. Descriptive statistics reveal significant variation for these two indicators. The numbers of candidates competing in the elections range from 1.02 effective candidates in Kemerovo’s 2001 contest to 3.79 in the Republic of Altai’s 2001 election, with a mean of 1.76 and a standard deviation of 0.64. Meanwhile, incumbent governors have been more likely to win reelection than to be turned out of office—the average rate of turnover equals 0.24 with a standard deviation of 0.43. We begin our multivariate analysis by presenting a basic model resembling those used in chapter 3. We include measures of regional socioeconomic development, resource wealth, economic performance, violent crime, party development, constitutional status, ethnic heterogeneity, and the rules governing the elections. We also control across elections held within the same region by including a variable that captures the total number of (valid) previous elections held in the region. After this, we expand the analysis to consider whether federal and regional actors were able to manipulate the electoral outcomes through tactics that resemble rule changes or electoral practices available to politicians in consolidated democracies. Specifically, we focus on the federal executive’s (i.e., the Kremlin’s) support for or opposition to gubernatorial turnover and regional changes to the gubernatorial election date. Since federal endorsements and election data changes exist in consolidated democracies, they conform to our focus on whether practices at lower levels of government that are generally accepted as adequately democratic, if somewhat unfair, may undermine the prospects for democratization. The act of federal executives promoting or opposing candidates in regional elections is pervasive in democratic federations worldwide. Finding the phenomenon in Russia might, by itself, reflect competitive politics becoming normal. At the same time, heavy-handed attempts by federal actors to influence regional elections might weaken Russians’ faith in the electoral process. The question, however, is not whether federal actors seek Page 85 →to influence regional elections but whether such attempts succeed on a systematic basis. If they do, then popular support for democratic institutions may decline as the voting public perceives its interests as secondary to political games among the elite, especially the federal government. While consolidated democracies enjoy the social capital to prevent such behavior from undermining the democratic system itself, societies where the roots of democratic traditions are shallow do not. At the same time, it is doubtful that popular dissatisfaction with elections was sufficient to doom the democratic process in Russia. In Russia, not only can disillusioned citizens withdraw from politics and abstain from the electoral process, prior to 2007 they could vote “against-all” candidates competing for office. This provided Russian voters an explicit mechanism for expressing their political disenchantment, as opposed to simply their lack of interest (Oversloot, van Holsteyn, and van den Berg 2002; Liubarev 2003; Hutcheson 2004; McAllister and White 2008). In addition to this institutionalized outlet for expressing dissent, Russian citizens during the period under investigation were conspicuously passive in the political arena beyond voting, even when provided significant cause for political protest (e.g., Javeline 2003; White and McAllister 2004). In the Russian context, then, the irony may be that failed attempts by the federal executive to determine regional election outcomes were more corrosive than successful attempts. While successful federal intervention may have alienated voters, such alienation was unlikely to lead to the cessation of elections. Conversely, the inability to manage regional outcomes may have caused the federal executive to question the value of regional elections. To measure the impact of the federal executive, we employ news reports and summaries from Radio Free Europe /Radio Liberty that assess whether the Kremlin took a position on the winning candidates of the gubernatorial elections and whether it seemed to support or oppose turnover (see, for instance, Radio Free Europe/Radio Liberty 2001, 2003). Since the Kremlin took no position on many of the victors of gubernatorial elections, two dummy variables allow us to determine whether the Kremlin had more success when it promoted turnover than when it opposed turnover, or vice versa. Of course, President Putin’s announcement that he planned to replace directly elected governors with presidential appointees could have altered the political landscape such that the

Kremlin’s influence carried more weight following the announcement than it did before. To control for this contextual change, we include a third dummy variable, Post-September 2004 Announcement, to distinguish gubernatorial elections held after the proposed elimination from those that occurred prior. Page 86 →A zero on all three dichotomous variables indicates gubernatorial elections held prior to Putin’s September 2004 announcement and cases where the Kremlin did not openly support or oppose turnover. If any of the three variables prove statistically significant, all else equal, then the results can be interpreted as indicating a systematic difference between these elections and those where the Kremlin took a position or those that were held after the announcement cessation of gubernatorial elections. To capture the Putin administration’s changing levels of experience with influencing gubernatorial elections, we include a variable that measures the number of months between the election in question and Putin’s March 2000 election as president. One might expect that the Kremlin would prove more successful at influencing gubernatorial elections as Putin’s administration gained more experience and became more consolidated over time. Like federal intervention in regional elections, changes in the timing of elections are not necessarily improper and occur in established democracies. Indeed, in parliamentary systems, calling snap elections is a normal stratagem available to a sitting prime minister (Palmer and Whitten 2000; Kayser 2005). Still, moving an election date reduces the fairness of the elections, albeit in a sense that term does not usually have. Scheduling early elections, for example, can unfairly benefit the incumbent by shortening the time available for potential competitors to organize and raise funds. Although Huntington (1991, 183) notes that little empirical evidence exists to support this logic, the Russian context is replete with examples. The most notable is Yeltsin’s New Year’s Eve 1999 resignation, which not only substantially improved the electoral prospects of his handpicked successor but demonstrated little respect for the intent of regularly held, fair elections. While not the first use of the tactic, President Yeltsin’s resignation no doubt recommended this ploy’s utility to many governors.2 In Russia, unlike in the United States, for instance, the timing of gubernatorial elections was unstandardized. Several scholars posit that Russian governors used their positions to alter election dates to increase their reelection chances (e.g., Moses 2002; Sakwa 2002). Such actions, which are clearly designed to undermine the intent of free and fair elections, are important because both voters and aspirants for power will draw conclusions from such tactics about the value of democratic politics. If key elected officials continually seek to diminish public accountability, other political actors—including those with the power to end elections—may conclude that electoral outcomes are more important than observing democratic norms. Thus, the longer a country experiences elections with such corollaries, the less secure that country’s democracy will be. Page 87 →To estimate the effect of election date changes, we first determined the length of terms that incumbent governors were serving.3 We then established whether gubernatorial elections in each region were held in the month anticipated based upon the month and year of the previous election.4 Next, we created two variables. The first depicts whether an election was on time (producing a score of zero) or held early (measured by the number of months). In general, we expect earlier elections to correlate with fewer candidates and less turnover. The other distinguishes elections held on time from those held late (again ranging from zero to the number of months late). While early elections are believed to favor incumbents by giving potential challengers less time to organize and raise campaign money, it is possible that incumbents also might choose to delay elections as a means of increasing their electoral prospects. Specifically, incumbents could postpone elections with the hope that later elections will yield a more favorable electoral environment by, for example, allowing more time for an economic recovery or by allowing enough time to use the judicial system to challenge the candidacies of particularly strong competitors.5 Another timing issue that could influence gubernatorial elections is whether they are held concurrently with federal elections. In Russia, eight gubernatorial elections were held as many as nine months early with the explicit goal of having them coincide with the 2004 Russian presidential election. Marsh (2002, 129) suggests that Russia’s liberals, in particular, try to schedule regional elections concurrent with national elections as a way to increase their electoral support since younger, more liberal voters participate at higher rates during Russia’s national presidential and parliamentary elections. Simultaneous regional and national elections are far from unusual in consolidated democracies, but the political implications can vary. In the United States, a growing trend has been for gubernatorial elections to coincide with midterm elections rather than with presidential

elections. The goal is for American voters to focus on the appeals of the gubernatorial candidates and state economic conditions rather than the presidential campaign and national economic conditions. While incumbent governors of the president’s party traditionally had been punished for national economic conditions, the electoral fate of incumbent governors in the United States have been closely tied to state economic conditions (Salmore and Salmore 1996). Yet, the shift of US gubernatorial elections to coincide with midterm elections has not only curtailed presidential coattails, it has lowered voter turnout in these gubernatorial elections (Jewell and Olson 1988, 209). The question then is whether the proximity of gubernatorial elections to national elections in a transitional state, like Page 88 →Russia, functions in the same fashion as it does in an established democracy, like the United States. Since holding gubernatorial elections simultaneously with national elections can encourage voter turnout, simultaneous elections might also be more competitive and, possibly, have a higher probability of turnover. Russia does not hold concurrent national elections for its legislative and executive branches, however. It also remains unclear as to whether competition should be fiercer during national parliamentary elections or national presidential elections. Shvetsova (2003), for example, notes that parliamentary elections have served as presidential primaries in Russia. Yet, presidential elections are contests between individuals for the top political office, and presidential coattail effects are widely recognized. Given this ambiguity, we measure the impact of simultaneity with both types of national elections.6

Correlates of Competition and Turnover in Putin-Era Elections The first equation in table 4.1 provides the results of regressing only the independent variables discussed in chapter 3 and capturing correlates of democratization on the effective number of candidates competing in the gubernatorial elections.7 As in that chapter, we deploy robust regression due to the presence of outliers and leverage cases that emerged during diagnostics of ordinary least squares models.8 We are not, of course, sampling regions from a larger universe but looking at all the regions in the universe under consideration. Hence we do not need to rely on statistical significance in the way one does when working with samples. Nonetheless, as before, the statistical significance level of the coefficients is reported as another way of assessing impact. We also draw attention to the standardized coefficients, which clarify how substantially changes in an explanatory variable produce changes in the dependent variable.9 With an R2 of 0.10, the basic model explains little variance. In fact, only one variable meets conventional levels of statistical significance and then only at the .10 level for a twotailed test: higher levels of party development—measured as the percent of regional parliamentary deputies behaving in office as though they are affiliated with a political party—are correlated with more competitive elections.10 The availability of resource rents only moderately dampens competition.11 Other factors from socioeconomic conditions to constitutional status, electoral experience, and ethnic heterogeneity12 fail to matter substantively. Page 89 → Table 4.1. Robust Regression Analysis of Effective Candidates in Russian Gubernatorial Elections, 2000–2005 Equation 1 Equation 2 Equation 3 в€’.055 в€’.045 в€’.072 Socioeconomic Development Index (.594) (.668) (.493) в€’.087 в€’.079 в€’.122 .001 .001 .001 Violent Crime (.243) (.273) (.305) .130 .128 .115 в€’.309 в€’.131 в€’.172 Natural Log of Oil and Natural Gas Production (.174) (.548) (.423)

в€’.158

в€’.059

в€’.090

.092 (.668) .122 в€’.001

.106 (.643) .057 в€’.001

.177 (.445) .112 в€’.001

Non-Russians as % of Population

(.807) в€’.113 .002

(.862) в€’.091 .004

(.831) в€’.101 .005

Real Income

(.744) .049

(.540) .057

(.422) .086

.004 (.209) .213 в€’.121 (.442) .021 в€’.179 (.340) в€’.134 .730 (.000) .421 в€’.073 (.657) в€’.021 .176 (.591) .043 .007 (.371) .121 .011

.003 (.231) .203 в€’.125 (.430) .019 в€’.134 (.484) в€’.084 .754 (.000) .433 в€’.044 (.785) .018 .296 (.324) .129 .003 (.624) в€’.036 .010

(.603) .081 в€’.036 (.721) в€’.031 в€’.350 (.298) в€’.292

(.636) .087 в€’.057 (.568) в€’.061

Republic

.006 Percent of Regional Deputies with Party Affiliation, 1997–2003 (.051) .234 −.050 Absolute Majority (runoff) (.753) −.007 −.108 Total Previous Elections (.399) −.104 Kremlin Supported Turnover

—

Kremlin Opposed Turnover

—

Election Occurred after September 2004 Announcement

—

Months since Putin’s 2000 Election

—

Early Election (in months)

—

Late Election (in months)

—

Proximity to Duma Elections

—

Proximity to Presidential Election

—

—

Constant

1.209

1.220

— −.129 (.692) −.146 .984

R2 Deviance Number of cases

(.079)

(.127)

(.221)

.102 25.68 91

.299 18.25 91

.295 19.02 91

Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Beta weights with an absolute value of .15 or higher are bolded. Page 90 →Subsequent equations in table 4.1 replicate the analysis, adding our measures of Kremlin intervention and electoral timing. Since we cannot include the two proximity variables into the same models due to collinearity, we present two sets of results.13 With R2s around .30, the second and third equations in table 4.1 explain more variance in the effective number of gubernatorial candidates while the regression coefficients reveal some interesting findings. In both models, Kremlin support for gubernatorial turnover—that is, opposition to an incumbent or support for a specific challenger—is strongly associated with a more competitive race, again measured as more effective candidates. By definition, this means the sitting governor was more vulnerable to defeat, as the Kremlin desired. However, the Kremlin’s endorsement of incumbent governors was not sufficient to reduce the number of challengers; instead, the effective number of candidates proved largely immune to the Kremlin’s opposition to turnover. Thus, while the federal executive’s position on specific candidates in gubernatorial elections during this period could increase the competitiveness of the electoral environment, it could not suppress it. Moreover, as the equations in table 4.1 indicate, the level of competition appears unrelated to the Kremlin’s experience with regional elections, as the number of months President Putin was in office at the time of each gubernatorial election fails to matter systematically.14 In other words, Putin may have eliminated gubernatorial elections not just because he wished to avoid humiliating defeats “at the hand of well-entrenched problem governors” (Konitzer 2005, 225) but so he could protect supportive governors, something he could not do in the context of popular elections.15 When controlling for the Kremlin’s influence (Equations 2 and 3), party development continues to be associated with more competitive elections. Also, a number of the null results in table 4.1 are noteworthy for the study of electoral politics in Russia’s regions. It is surprising, for example, that incumbent governors did not systematically benefit from changes to the election date. Neither the variable measuring premature elections nor the variable indicating belated elections significantly affected competition in the regions. In other words, decisions to manipulate election timing yielded no tangible benefit to Russia’s regional executives.16 This is not to say, however, that such actions were inconsequential. Even failed attempts by incumbents to bend the electoral rules to their advantage send signals to the federal government, as well as voters, that regional elites are willing to sacrifice fairness for victory. These signals, then, preserved a more authoritarian political climate, one where political ends take precedence over democratic means. Page 91 →Perhaps most surprising, the regression analyses reveal that regional conditions other than party development did not significantly influence electoral competitiveness. Incumbent executives in Russia’s poorer, less developed, and more crime-ridden regions do not appear to have encountered more effective challengers. Moreover, despite the growing importance of oil and natural gas to the Russian economy as a whole during this period, the regional availability of resource rents fails to significantly reduce competition levels in gubernatorial contests: the very modest effect uncovered in the reduced model disappears in the full model. Even regional characteristics such as a region’s status as a republic, ethnic composition, and use of double ballots failed to significantly shape the effective number of candidates during the early Putin era. In the end, then, competitiveness largely appears to have reflected the federal government’s posture. Of course, the Putin administration was likely less concerned about how many effective candidates competed in gubernatorial elections than with the outcome of the election—whether the incumbent won or lost. Unfortunately for the Kremlin, the two cannot be easily separated from one another. Table 4.2 presents the results from the logistic regression analyses for the dichotomous variable indicating whether the incumbent was voted out of office. Since the effective number of candidates competing in an election could influence the prospects for turnover—by increasing the prospects for a runoff election and thereby engaging more voters located at different

points along the ideological spectrum (and demonstrated in chapter 3)—this variable is included alongside the other independent variables.17 As in the preceding analysis, the models in table 4.2 explain a solid amount of the variance in the dependent variable, with pseudo-R2s around .30. However, the only variable to significantly influence the likelihood of gubernatorial turnover is the effective number of challengers competing in the election.18 This finding suggests that the Kremlin should have had a strong interest in the number of candidates challenging incumbents in gubernatorial contexts since competition levels greatly determined election outcomes. This relationship, in other words, likely pleased the Kremlin in regions where President Putin preferred turnover and worried it in regions where he opposed it, since Putin’s preferences mattered in the former but not the latter. Indeed, aside from increasing the effective number of candidates in races where it preferred more challengers, the Kremlin appeared unable to alter the electoral fate of incumbent governors: when controlling for the effective number of candidates, the Kremlin was unable to systematically undermine an incumbent governor’s bid for reelection Page 93 →by openly opposing the incumbent or publicly endorsing a political challenger. These findings suggest that a more effective route for unseating incumbents may have entailed keeping them from seeking reelection in the first place (see below and appendix 2). Likewise, if the president openly supported an incumbent (or opposed a certain challenger), such intervention was insufficient to rescue that governor since it determined neither the effective number of candidates nor the likelihood of turnover. Again, these relationships hold even when one controls for how many months President Putin was in office. Page 92 → Table 4.2. Logistic Regression Analysis of Turnover in Russian Gubernatorial Elections, 2000–2005 Equation 1 Equation 2 Odds Ratio Odds Ratio 95% Confidence Interval 95% Confidence Interval (p value) (p value) Effective Candidates Index 10.165 (.001) 2.557 40.413 10.533 (.001) 2.507 44.254 Socioeconomic .699 (.691) .120 4.074 .761 (.777) .116 5.003 Development Index Violent Crime .999 (.855) .991 1.008 .999 (.856) .991 1.008 Natural Log of Oil-Gas .243 (.732) 7.30e-5 806.812 .368 (.815) 8.66e-5 1565.19 Production Republic 1.079 (.947) .114 10.202 .856 (.892) .091 8.080 Non-Russian as % of .992 (.662) .955 1.030 .992 (.686) .955 1.031 Population Real Income 1.035 (.274) .973 1.102 1.037 (.260) .974 1.104 Percent of Regional Deputies with Party 1.003 (.853) .972 1.035 1.002 (.913) .971 1.034 Affiliation, 1997–2003 Absolute Majority (Runoff) 2.607 (.298) .428 15.872 3.059 (.235) .484 19.230 Total Previous Elections .494 (.476) .071 3.437 .339 (.293) .045 2.549 Kremlin Supported .672 (.656) .117 3.853 .741 (.738) .128 4.441 Turnover Kremlin Opposed Turnover 1.192 (.827) .247 5.751 1.033 (.968) .208 4.297 Election Occurred after September 2004 .119 (.227) .004 3.749 .035 (.113) 5.43e-4 2.208 Announcement Months since Putin’s 1.032 (.446) .952 1.118 1.068 (.209) .964 1.184 2000 Election

Early Election (in months) .955 (.709) Proximity to Duma Elections Proximity to Presiden-tial Election Constant Pseudo R2 Likelihood Ratio П‡2 (significance) Number of cases

.749

1.216

— .213 (.392)

.006

1.58e-4 (.065) 1.44e-8 .282

.921 (.516)

.718

1.181

.047 (.165)

6.11e-4

3.543

7.360

—

1.735

2.38e-4 (.078) 2.21e-08 0.295

28.39 (.028)

29.47 (0.019)

91

91

2.555

Note: Numbers in parentheses are p-values. Odds ratios in bold are significant at conventional levels for a twotailed test. Just as presidential preferences failed to demonstrate an independent and systematic effect on gubernatorial election outcomes, the likelihood of turnover also appears to have been impervious to regional characteristics and conditions: the likelihood of incumbent turnover proves unrelated to regional development, crime rates, resource wealth, regional party development, the constitutional status of the region, ethnic heterogeneity, or whether the electoral rules offered the possibility of a runoff. Finally, electoral timing—that is, whether the gubernatorial elections were held early or in closer proximity to national contests—failed to significantly determine an incumbent’s chances of reelection.

Contrasting the Yeltsin and Putin Periods Like many gubernatorial elections, Russia’s 2000 presidential election conformed to the letter of the law while flirting with the boundaries of electoral fairness. With Yeltsin’s resignation, Prime Minister Putin became acting president and enjoyed the perks of incumbency, like unparalleled visibility, that made campaigning unnecessary. At the same time, the resignation moved the 2000 presidential election up the calendar by three months, shortening the campaign period available to his opponents. Putin’s victory was decisive, and one would be surprised if the new president did not expect the show of support and powers of the presidency to grant the Kremlin sufficient influence to flex similar electoral muscles in the regions. While gubernatorial elections during Putin’s first term in office were still relatively new to Russia, chapter 3 demonstrated that elected incumbents already had begun to successfully navigate them. Yet one of the major differences between the results in chapter 3 and those presented in this chapter is the declining importance of constitutional status. In the 1994–99 period, gubernatorial elections in republics experienced significantly fewer effective candidates than those in other regions. Since the effective numberPage 94 → of candidates then, as during the Putin era, had a powerful impact on the likelihood of turnover, republican status indirectly reduced the likelihood that republican incumbents would be turned out of office when compared to appointed governors (in nonrepublican regions) or incumbents in republics who had yet to compete in direct elections. Indeed, republican founding elections during the 1994–99 period were more likely to generate turnover than elections elsewhere. It is notable, then, that in the 2000–2005 period neither the effect of constitutional status nor the number of previous elections matter. Taken together, we can conclude that while the power of incumbency worked to the advantage of those republican incumbents who were willing and able to institute elections early in the Yeltsin era, incumbency advantage was no longer the dominion of republican governors by the start of Putin’s first term in office. This temporal difference, then, emphasizes the value of examining Yeltsin-era elections and Putin-era elections separately, as studies that pool elections across the two periods risk overlooking this temporal impact of constitutional status (e.g., Moraski and Reisinger 2003, 2007; Saikkonen 2016). Comparing gubernatorial elections in the Yeltsin and Putin eras also reveals important differences between regional conditions and regional institutions. Specifically, greater availability of resource rents dampened competition levels—measured as the effective number of candidates—during the Yeltsin years. Somewhat

surprisingly, competition levels were also lower in regions with higher levels of socioeconomic development. This result directly challenges the expectation associated with modernization theory that more developed regions would be the home of more competitive elections. We find instead that, among Russia’s regions, more developed polities actually experienced less competition. Runoff elections, on the other hand, certainly yielded greater competition but demonstrated no direct effect on turnover. Instead, the likelihood of turnover was more correlated with levels of violent crime. The finding that higher levels of violent crime lowered the likelihood of turnover suggests that political accountability requires a regional climate supportive of societal trust and civic engagement. In all these instances, then, we find value in applying insights drawn from democratic settings to regional politics in Russia during the 1990s. By the 2000s, however, the importance of hypotheses drawn from democratic settings and the literature on resource wealth fades. The dynamics of Russia’s gubernatorial elections were no longer systematically influenced by rent availability, levels of socioeconomic development, violent Page 95 →crime, or electoral rules. Rather, gubernatorial elections during the early years of Putin’s reign functioned more like a game played by elites with Kremlin opposition to incumbents spurring more effective candidates, and the effective number of candidates alone determining the fates of the incumbents. By the time President Putin assumed office, then, directly elected governors had become the norm across Russia, but electoral success had little in common with factors widely seen as shaping voter behavior in Western democracies.19 Meanwhile, President Putin learned that the Kremlin’s influence in gubernatorial elections, if confined to more democratically acceptable electoral practices, was limited as well: in contrast to the 2000 presidential election, the Kremlin was unable to systematically suppress competition levels in gubernatorial elections. This lesson is not entirely different from President Yeltsin’s initial experience with gubernatorial elections. As chapter 3 notes, Yeltsin appointees struggled to win gubernatorial elections in contests held prior to the adoption of the 1993 Russian Constitution. At that time, Yeltsin responded by postponing additional elections, a decision that paid off: between 1994 and 1999, the likelihood of gubernatorial turnover was actually lower in first elections where a presidential appointee held office than in regions holding their first executive elections but without a presidentially appointed incumbent in office (i.e., republican founding elections during this period). Putin’s experience with gubernatorial elections resembled Yeltsin’s to the extent that the Kremlin’s backing was not enough to save a preferred incumbent from ouster. Rather, the best Putin’s Kremlin could do was to indicate its displeasure with certain incumbents, which facilitated greater electoral competition but did not necessarily result in the election of the Kremlin’s chosen candidate. At this point in Russia’s post-Soviet history, though, gubernatorial elections had become institutionalized. So, unlike Yeltsin in the early 1990s, Putin did not possess the option of temporarily postponing gubernatorial elections and filling those posts with appointees until conditions in select regions tilted in the Kremlin’s favor. Putin’s reaction then, while mirroring Yeltsin’s a decade earlier, proved more drastic: to guarantee the allegiance of Russia’s gubernatorial corps required the wholesale elimination of gubernatorial elections. While this move qualifies as a center-led transition designed to undermine regional actors’ control over politics in their borders, it diverges from those identified by Gibson (2012) in that Putin intervened to end competitive elections rather than to enable them.

Page 96 →Conclusion In this chapter, we conduct multivariate analyses of Putin-era gubernatorial elections prior to their elimination. As in the previous chapter, we focus on the effective number of candidates running in each election and whether the incumbent is turned out. In addition to variables examined in the previous chapter focusing on Yeltsin-era elections, we investigate whether the federal executive’s preferences and changes to gubernatorial election dates had independent impacts on the election results. We find that although the Kremlin could use its influence to increase the competitiveness of Russia’s gubernatorial elections by opposing a sitting executive, it was unable to systematically dampen competition when it supported an incumbent. Moreover, intervention on behalf of the federal executive failed to affect the actual outcome (i.e., turnover) of Russia’s gubernatorial election campaigns. Thus, not only was the Kremlin unable to protect those incumbents whom it supported, its preference

for ousting a sitting governor had, at best, an indirect effect on the election’s ultimate outcome. These findings reveal a stark difference between Putin-era elections and the first gubernatorial elections across Russia’s regions during the 1994–99 period when the leaders of Russia’s republics proved more vulnerable than presidentially appointed incumbents in nonrepublican regions. The preceding analyses do not, however, evaluate the Kremlin’s ability to coerce a governor to resign or to decide against seeking reelection since these cases likely indicate practices of electoral authoritarianism. Such decisions can easily reflect strong-arm tactics from the federal government, and observers of Russian politics generally agree that those governors who had fallen most out of favor with the president were more likely to be pressured not to seek reelection, were “legally” barred from running for re-election, or were lured into resigning by a post in the federal government. In two regions, Sakha and Chukotka, for example, criminal investigations were initiated against recalcitrant governors. In others, like St. Petersburg and Primorye krai, incumbents received posts in the federal government in exchange for their resignations. Thus, we argue that Russia’s president had two options for influencing gubernatorial elections: (1) openly supporting or opposing the campaigns of certain gubernatorial candidates and hoping that such federal intervention would produce the intended election results, or (2) leaving much less up to chance and improvising ways to remove unruly regional leaders. Given not just the uncertainty but the relative ineffectiveness of the first option, it is not surprising that Russia’s president would become frustrated with elections and decide to move from Page 97 →the ad hoc methods of removal associated with the second option to an institutionalized mechanism of control.20 In sum, then, the limited ability of the Kremlin to shape Russia’s gubernatorial election outcomes helps explain Putin’s call for bringing these elections to a halt. Indeed, we argue that this step stemmed from an inherent conflict between Putin’s rhetoric during his first term in office and the political reality on the ground in Russia’s regions: despite declarations and reforms by the Putin administration intended to restore the executive hierarchy in Russia, Russia’s gubernatorial elections were less responsive to the interests of the Kremlin than the presidential administration liked. At the same time, the analyses suggest that regional socioeconomic conditions had no perceptible effect on an incumbent governor’s electoral prospects, which may explain the limited public outcry when Putin decided to halt gubernatorial elections. Many regional voters probably felt as though it would be better for someone to be able to hold governors accountable and, if it could not be them, then the president may be the next best option.

Page 98 →

Five Failed Experiment? Presidential Control over the Tenures of Governors in the Russian Federation, 2005–12 Changes in how Russia’s regional leaders are chosen provide something of a natural experiment. Following more than a decade in which regional governors were elected, Russia’s federal president had de facto power to appoint and fire regional governors from 2005 to 2012.1 Governors served as long as they maintained the confidence of the Russian president. That is, “losing the president’s confidence” was an adequate reason for a governor to be dismissed prior to the end of his or her formal term. Although the president did not need to provide any reason for declining to nominate a sitting governor for a new term, the functional difference between providing no explanation and stating that one has “lost the president’s confidence” is small. Under these rules, any governor could be fired at any time depending on the will of the federal president. Whether a governor left office at the end of a formal term or in the middle did not alter the political basis for that departure: a decision by the Kremlin.2 Moreover, each additional week or month a governor remained in office also reflected a Kremlin decision, that is, the decision not to make a change. For just under eight years, the Kremlin held direct control over who occupied the governorships in Russia’s regions. What do we learn about politics between Russia’s federal center and its Page 99 →regions from examining governors’ tenures during the period of Kremlin control? The switch from elected to appointed governors certainly gave the Russian president substantially more bargaining power vis-Г -vis the governors. Yet it would be a mistake to interpret the situation as entirely under the Kremlin’s control. The regions’ performance in the economic, social, and political spheres remains crucial to the federal leadership’s goals for the country. Also, the governors are important players in national politics, especially due to their role in promoting United Russia’s success in national presidential and legislative elections (on this, see Reuter 2010). We use survival analysis of data from all of Russia’s regions from 2005 to 2012 to estimate the factors influencing when a governor is replaced, including the region’s size and ethnic composition, the governor’s age, and aspects of the governor’s administrative and political performance. We find that the Kremlin targeted for replacement the governors who could not deliver pro-Kremlin votes in federal elections. Yet this tool of political control over the regional leaderships did not prevent the weak showing of the Kremlin’s party in the 2011 legislative election or the mediocre showing of Putin in the 2012 presidential election. The appointment era now looks more like a failed experiment than an inexorable tightening of authoritarian control over the country.

Personnel Control and Replacement under Nondemocratic Rule Scholarship over the past decade has directed increasing effort to understanding the functioning of nondemocracies, whether fully autocratic or hybrids. In particular, more work examines how formal participatory institutions such as elections, parties, and parliaments help the regime’s leadership maintain nondemocratic control (Geddes 2005; Magaloni 2006; Brownlee 2007; Gandhi 2008; Gandhi and Lust-Okar 2009; Blaydes 2010). Among the important roles formal institutions can play is “elite management,” restraining and rewarding political officials so that capable people are put into responsible positions and those capable people have paths for advancing. In her study of the PRI (Institutional Revolutionary Party) in Mexico, for example, Magaloni (2006, 8) concluded that “Autocratic regimes reward with office those politicians who prove most capable in mobilizing citizens to the party’s rallies, getting voters to the polls, and preventing social turmoil in their districts. The autocracy thus forces politicians to work for the benefit of the party and to have a vested interest in the survival of the regime.” Page 100 →The contrasting means of elite management is bureaucratic or corporate. That is, one or more officials at a higher place in the regime hierarchy decide on rewards—such as promotions to higher or more lucrative positions—and punishments without public participation. In single-party regimes, such as the Soviet Union or

China, the party manages “cadres policy.” In the Soviet case, the rules about which level of the party controlled which job was known as the nomenklatura system (Harasymiw 1969; Hill and Frank 1986, 87–92), and memory of it remains among many of Russia’s current leaders. What has occurred over the past decade in Putin’s Russia, then, is a swing from an electoral-based process for managing governors to a bureaucratic one, followed by a swing back to a (more managed) electoral one. While much recent work incorporates subnational leaders and lower officials into its analyses, it focuses on those officials’ role in the hierarchy of an overall national regime. The complexity of Russia’s federal system challenges this way of approaching national-subnational dynamics. The extent to which elite management is centralized or “federalized” within a given regime will matter as well. The Soviet nomenklatura system was centralized.3 The Kremlin could, at least formally, craft a nationwide policy via appointments or firings, including allowing the general secretary to build a nationwide team or network (Moses 1985; Willerton and Reisinger 1991; Willerton 1992). From 2005 to 2012, however, officials within a governor’s region were not under bureaucratic control from the Kremlin. Albeit with significant variation across the regions, Russia’s governors were tasked with elite management in their own region. Many sought bureaucratic control over the elites below them (ending mayoral elections, for instance), but for the most part they were not given it (Gel’man and Lankina 2008). In other words, regime-wide elite management was federalized rather than centralized. During this period, then, integrating the governors was critical to the effectiveness of the Russian political regime. By integration, we mean that (1)В the governors should bend their efforts toward priorities of the national regime, including deferring to the national regime’s needs when those needs conflict with the governor’s personal priorities or ambitions, and (2)В the governors must be effective at regional governance. With regard to the former, Russia’s national regime in comparative perspective lacks either of two sources of elite cohesion that research suggests are highly important. One is a shared ideology that can overcome short-term incentives for cadres in an organization to defect (Hanson 2010; Levitsky and Way 2010; Slater 2010). The other Page 101 →is the elite bond that can arise when most of its members participated in a common period of violent struggle (Slater 2010; Levitsky and Way 2012). In addition, shared ethnicity sometimes can provide a sufficient bond among key elites to undergird a durable regime (Levitsky and Way 2010, 26), but ethnic cohesion in Russia occurs among elites at the level of certain regions (the republics and autonomous okrugs), not for the ruling elite as a whole. Without significant nonmaterial links binding together Russia’s elite, Russia’s regime relies to a comparatively greater extent on the also common practice of pursuing elite cohesion by dispensing patronage or spoils (e.g., Gandhi 2008; Blaydes 2010). A patronage-based system faces the problem, of course, that the pursuit of material gain on the part of the subordinates can conflict with effectiveness at policy-making and implementation. In addition, patronage-based systems encourage a subordinate, such as a governor, to build his or her own political machine at the subnational level, potentially strengthening his or her ability to resist the top level of the regime. During roughly a decade in which gubernatorial elections were first held, the latter issue gained greater salience because of the added legitimacy a governor gains from being popularly elected. Eliminating gubernatorial elections was directed precisely at removing governors’ ability to resist the Kremlin’s aims. What drives the pattern of gubernatorial tenures during the appointment period? Gubernatorial replacements, not counting those caused by an incumbent’s death, occurred 10 times in 2005, 4 in 2006, 8 in 2007, 8 in 2008, 10 in 2009, 19 in 2010, 6 in 2011, and 19 in 2012 through June.4 Sharafutdinova (2010b, 682) argues that the primary goals of the appointment process are “vote delivery” and “societal manageability,” and it is these considerations that drive gubernatorial replacements. Similarly, Turovskii (2010) submits that the primary determinants of replacement are the governors’ abilities to provide electoral results favorable to the Kremlin and to consolidate the regional elite. Indeed, in the absence of any mechanisms for popular input, Turovskii (2010, 69–70) contends that neither the personal popularity of a governor nor the socioeconomic situation in a region matters much. Although electoral results and societal stability may be prevailing considerations for the Kremlin, the relative

weight of these two considerations seems to have varied over time and probably varies from region to region. Turovskii (2010) concludes that the appointment era had, at the time he was writing, evolved in three stages: (1) inertia, when most incumbents were reappointed; (2) experimentation, when the president experimented Page 102 →with instilling some new blood into the regions in the form of outsiders; and, most recently, (3) replacement, as the center became more confident in its ability to control the situation in the country. Dismissals initially were largely confined to sparsely populated regions notorious for poor governance. Once the Kremlin moved into the experimentation stage, the potential risks of change were offset by focusing on regions with passive populations and high public confidence in the center (ibid., 72). Turovskii’s assertion that caution characterized the first two stages of the appointment process complements Sharafutdinova’s (2010, 683) view that the appointment process was not solely driven by a desire to maximize power: if it had been, then Russia’s most powerful regional barons would have been replaced first. Previous studies, then, suggest that a governor and his or her allies at the regional level can influence both the president’s desire to install someone else and the cost to the president for making a change. As Gel’man and Ryzhkov (2011, 453) note, “[D]espite the numerous cases of forced resignations of regional governorsВ .В .В . the hierarchy of the вЂpower vertical’ is far from an army-like chain of command, and it operates according to a different logic.” Both explicit and tacit bargaining occurs between the Kremlin and the governor. Understanding this bargaining requires attention not only to the formal institutional rules, which constrain the governors substantially and the president hardly at all, but also to informal power relations and other contextual constraints. Formal and informal influences do not simply coexist; they can be in tension. Chebankova (2010, 2), for instance, sees a “growing rift between the erected institutional structures [of Russian federalism] and the functioning processes taking place within them.” Meanwhile, the low supply of sufficiently qualified and trustworthy replacements not only represents the kind of contextual constraint that initially tied the president’s hands (Turovskii 2010, 66) but that also may keep the Kremlin from being able to hold governors accountable for the quality of governance in their regions (Sharafutdinova 2010b, 673). Although the research question is when the Russian president replaces a governor, we are assuming that presidential choices reflect a process of explicit or tacit (to them) bilateral bargaining and competition between the Kremlin and each regional governing team. If this assumption is correct, the pattern of gubernatorial change /retention ought to reflect cross-regional differences in the factors that affect decision-making. Although formal institutions and the rules that generate them will always shape the distribution of power resources among actors, we do not examine the influence of any one formal institution or set of institutions. Indeed, what gives rise to Page 103 →our research question is the set of rules governing presidential control over gubernatorial retention and appointment. These rules provide the Russian president with tremendous power vis-Г -vis each governor. However, they are the same for each of the bilateral relationships between the Kremlin and a regional leader. For our purposes, then, they are not variables nor, therefore, do they constitute an explanatory factor in our approach. As we will discuss below, the factors we examine flow from thinking about the relationship of power resources and the resulting Kremlin strategy.

Why Are the Governors Replaced When They Are? While previous work provides qualitative and impressionistic accounts of the considerations driving the gubernatorial appointment process in Russia, scholars have yet to apply methods that allow them to analyze the relative effects of rival explanations. The exception is Reuter and Robertson (2012), who undertake to assess the impact of the election motive relative to factors that one might associate with quality of governance. Yet their logistic regression analysis of annual data does not reveal why a governor is replaced when he or she is replaced, a question at the crux of the matter. The timing and manner of the replacements varied widely. Mintimer Shaimiev governed the Republic of Tatarstan for 22 years in total, including 62 months during the appointment period, until he stepped down voluntarily in January 2010. A protГ©gГ© became the new president, and Shaimiev retained a title as state advisor plus his old office in the presidential building (Pavlov 2010). Governor Nikolai Kolesov of Amur Oblast, appointed from outside the region in May 2007, was fired just 16 months later (Ostrovskaia 2008; Petrov 2008).

Although the law gave the Russian president during 2005–12 the authority to end a governorship at any time, whether to do so was a sensitive political question. Medvedev’s decision to dismiss Moscow’s mayor, Yuri Luzhkov, for example, came after months of speculation about whether it would happen. The president’s decision naturally would have been influenced by numerous factors pertaining to the region, the situation in the country as a whole, the president’s own political standing, and, of course, the governor him- or herself. For insights into the politics of presidential control over gubernatorial power, one must examine the entire pattern in search of those factors that clearly promote or retard Kremlin action. Because the president could oust a governor at any time, regardless of the incumbent’s stated term length, opting to keep or remove someone Page 104 →from office was a continuous process. Survival, or event-history, analysis is the proper technique for examining a pattern of data having this over-time character (Blossfeld, Golsch and Rohwer 2007; Golub 2008; Mills 2011). Thus, with information on the governors, the regions, and the national situation, we use event-history analysis to see what factors influence the survival of a governorship. We examine the pattern of gubernatorial replacements and nonreplacements, month-by-month, from 2005 through June of 2012. We then employ multivariate models incorporating factors that (a) bear on the Kremlin’s desire to replace the incumbent governor, (b) indicate the Kremlin’s political readiness to make a replacement, and (c) provide important controls. We expect that regional characteristics, the incumbent governor’s personal characteristics, his or her political performance from the Kremlin’s perspective, and his or her administrative strengths or weaknesses ought, all else being equal, to shape Kremlin incentives for or against making a replacement. First, our expectations about timing. The frequency of replacements should increase from the third quarter of 2008 on, when the global economic crisis began to harm Russia’s economy. As economic pain grew, so did social discontent. Public dissatisfaction with poorly performing governors was rising, and having a regional governing team that could maintain social stability was of growing importance. Also, the Kremlin itself needed to show that it was responding to the economic downturn. Replacing governors, therefore, had particular value.5 In addition, the rise of public protests following the fraudulent federal legislative election in December 2011 can be expected to have produced more changes, as part of the overall strategy to quell the discontent, particularly with Putin’s campaign for the March 2012 presidential election getting underway. With the return to direct elections announced, President Putin’s reshuffling of the gubernatorial corps from March on took on greater urgency, thus yielding more replacements than earlier. Second, regional characteristics: one would expect the Kremlin to exhibit more care in making a leadership change in the more nationally important regions, since a poorly managed change could cause greater harm than doing nothing. We will measure national importance with a region’s total population.6 The more populous regions, in other words, should see fewer gubernatorial changes. One also should expect the predominantly ethnically non-Russian regions (republics or autonomous regions) to have fewer leadership changes. They are of extra sensitivity to the Russian leadership, albeit in Page 105 →a different sense than are the economically vital regions. Replacing such a governor might lead to public outcry by the ethnic population as an attack on the control of the region by the Kremlin, especially when an ethnic Russian is brought in to be the new governor. To assess this factor, we use the percent of ethnically non-Russian residents in a region rather than categorical variables distinguishing regional statuses. Third, characteristics of the incumbent governor and his or her past experiences will be relevant. Those who came into office prior to 2005, especially those whose control of the regional machine is evidenced by strong electoral victories, owe their loyalty less exclusively to Putin and his team in the Kremlin than do those appointed later. Also, longer-serving governors should have more ways to resist a Kremlin move against them. Countering these possibilities, however, is the very fact that the Kremlin will see its own appointees as inherently more trustworthy. Moreover, quite a few of the governors who took office prior to 2005 actually began serving, like Shaimiev, well before then. The very length of their tenure means their age is high, and they may appear to the Kremlin as less dynamic or flexible.7 The pre-2005 group could, therefore, be more of a target of the Kremlin. However, to the extent that greater age carries political experience in that region, we expect replacements to be fewer.

Fourth is the incumbent governor’s administrative record. As the Kremlin examines the merits of retaining or replacing a particular governor, it would like all the regions to exhibit social stability, economic growth, and the absence of serious infighting among the elites (Gel’man and Ryzhenkov 2011, 454). The Kremlin’s interest in replacing a governor, therefore, should rise in response to economic decline, rising crime, or worsening health statistics. Tracking the regions’ economic, social, and administrative success is the goal behind the Kremlin’s effort since 2007 to gather systematic data on regional performance (Reuter and Robertson 2012). Finally, we expect the incumbent governor’s political performance to matter. As the observers cited above have noted, the Kremlin expects more from a region’s leadership than just successful governance of the region. Governors are expected to provide effective support for the vertical of power, that is, for the nationwide strength of the Kremlin’s party, United Russia, and for the regime more generally. The success of United Russia’s candidates in both regional and federal legislative elections is a criterion by which governors are judged. We therefore expect that, all else being equal, governors of regions in which the Kremlin’s party or candidate fares well electorally will be safer from replacement. Page 106 →

Patterns of Gubernatorial Survival, 2005–12 Our dependent variable is, for each governor, the number of months until he or she is removed from office. The beginning month is January 2005 for those who were governing at the time that the new system took effect or, for those appointed subsequently, their month of appointment. We have collected data through the end of June 2012.8 Out of the 177 governorships analyzed, the 83 governors still in office at the end of June 2012 become “rightcensored” cases. This designation allows the survival analysis procedure to note that the governorships continue yet to draw information from the number of months they were in office through June 2012. We also code as right-censored three governors who died in office,9 one governor who resigned for health reasons and died within weeks,10 and six governors who headed regions that merged with larger regions, since these governors left office for reasons exogenous to the appointment process.11 The total number of other, not right-censored, cases is 84.12 In support of our point about the institutional change in 2005 not providing the Kremlin with carte blanche authority is the overall infrequency of gubernatorial turnover. During the seven and a half years our data cover, the average number of changes was 1.06 per region. Excluding a flurry of 16 replacements after Putin’s reelection in 2012 and before the appointment-era rules ended, that average is .88, less than one per region. For a given region in a given year, the chance of a turnover was .15, about one in six. With the Kremlin holding all the formal levers and having many reasons to prod regional leaderships, one might have expected more frequent replacements. Still, 84 governors were removed over the seven and a half years, and some periods showed higher rates of change than others. Figure 5.1 shows a histogram of the number of regions in each three-month period experiencing a changeover in governor. Lighter bars indicate quarters in which federal elections occurred. With regard to our expectation that replacements would be relatively more frequent from the third quarter of 2008 on, we find instead that the pattern is only somewhat higher, and the variability in both periods is more striking than the difference between them. Those quarters in which more than three replacements occurred suggest that change has been a particular Kremlin priority during certain periods. There were three such quarters preceding the December 2007 legislative elections: the third quarter of 2005, and the second and third quarters of 2007. The first three quarters of 2010 saw 18 changes, a period when Medvedev was seeking to rejuvenate the gubernatorial corps. Fig. 5.1. Frequency of Gubernatorial Changeovers by Quarter, 2005 through Mid-2012 Page 107 →What, though, makes it likely that a given governor will continue in office or be removed during a particular month? Figure 5.2 shows the slope of the survivor function, along with bands indicating the 95% confidence interval. Governors remaining in office for the entire period from January 2005 through June 2012 have tenures of 90 months. The probability of survival is one, by definition, at the very start of a governorship.

From there, that probability can only decline. Thus, survivor functions slope downward. Those in office for less than two years have a 90% or more likelihood of continuing in office. Those in office for over five years have less than a 57% likelihood. For those with the maximum tenure of 90 months, the likelihood is 40%. Among all governors during this period, their tenure lasts on average 42.8 months (median = 42) or roughly three and a half years.13 One can see from figure 5.2 that the odds of a governor remaining in office decline (the slope of the survivor function declines) noticeably from about 31 months through 50 months, then again between 60 and 68 months and from 83 months on. For governors in office in January Page 108 →2005 who remain so, Medvedev takes over as Russian president in their 41st month and the economic crisis begins in their 44th month. To some extent, then, the downward slope from 31 to 50 months reflects an upsurge of replacements during the run-up to and the aftermath of the 2007–2008 election cycle. The 50th month for governors in office since the start of the appointment era is February 2009, when the governors of four regions—Orel, Pskov, Voronezh Oblasts, and the Nenets Autonomous Okrug—were replaced on the same day. For that same group of governors, their 58th month in office is October 2009 and their 67th is July 2010. Replacements picked up speed during this period. Fifteen governors are replaced during this period, 14 of whom began their tenure prior to the appointment period, including such powerhouses as Shaimiev of Tatarstan and Eduard Rossel of Sverdlovsk. As noted earlier, Turovskii (2010) depicts this as a time of greater Kremlin confidence in its ability to replace entrenched governors. The many replacements in the final months of the appointment period affect the final downward turn in slope at the right of figure 5.2. Fig. 5.2. Survivor Function for Gubernatorial Tenure, January 2005–June 2012 The same type of graphs can depict differences in governors’ survival probability based on different values of key variables. For instance, do governors in republics and autonomous regions have higher rates of survival because of the greater prospect of social resistance or political-machine resistance in those regions? To distinguish the two groups of regions, we Page 109 →use 30% of the regional populace being non-Russian as the cutoff point. (All regions above 30% have a constitutional status of republic or autonomous region.) Figure 5.3 shows the survivor functions for each group. The line for the predominantly Russian regions is higher at every number of months in office, at times substantially so. In these charts, a lower line indicates a lower probability of survival in office. The bivariate pattern in figure 5.3 does not suggest extra resources lay in the hands of the ethnic regions’ governors. They were actually replaced at a higher rate than other governors. Fig. 5.3. Survivor Function for Gubernatorial Tenure, by Ethnic Composition of Region The survivor functions shown in figure 5.3 illustrate how ethnic composition has a bivariate impact on the probability of survival. We turn now, though, to multivariate analysis so that we can examine the impact of size and ethnic composition along with characteristics of the governor and his or her performance. To estimate our models, we use the Cox proportional hazards technique, a partial likelihood method (Blossfeld, Golsch, and Rohwer 2007, 223–46; Cleves, Gould, and Guttierrez 2010, 129–228; Mills 2011, 86–113). We begin by examining the results of models that relate only to the Kremlin’s interest in replacing a governor. We then introduce explanatory variables that indicate a governor’s ability to resist being replaced, allowing us to contrast the relative influence of the two types of factors. Page 110 →Although observers of Russian politics have reason to be interested in whether presidents Medvedev and Putin replaced governors at different rates, the observation of such an effect cannot be conclusive since the change in presidential leadership coincided with the onset of the global economic downturn. Thus, to the extent that a different pattern in gubernatorial fates before and after May 2008 exists, it may result either from different political personalities or different political incentives. Still, it is clear that the onset of economic crisis and a new presidency initiated a new period in the appointment era and that the prospects for gubernatorial survival (i.e., staying in office) should differ across the two periods. An intuitive approach for capturing this variation would be to add a dichotomous variable to the multivariate analysis to distinguish governors who served after April 2008 from those who served only under Putin. This approach, however, distorts the effects of the new presidency and

economic crisis on gubernatorial survival because we are interested in the total number of months that all of Russia’s appointed governors have survived rather than how long a subset of governors—those in office during the Medvedev presidency and the economic crisis—have survived. By definition, the dummy variable distinguishes Putin-appointed governors who failed to survive into the Medvedev era not only from Medvedev’s appointees but also from Putin-appointed governors whose tenures continue beyond Putin’s presidency. As a result, the variable gives Putin full credit for the governors he fired but no credit for those who survived his presidency. Even worse, our focus on total time in office means that the Putin-era tenures of these governors are carried over into the Medvedev era, thus inflating the survival rates of Medvedev-era governors. We can, fortunately, distinguish between the fates of governors under President Putin and under President Medvedev without adding a biased explanatory variable to our multivariate model. A right-censored Cox proportional-hazard model allows one to add a cluster function that captures a clustering of subjects that could likely result in nonindependent observations. (Gharibvand and Liu [2009, 1] offer the example of mice from the same litter.) Our multivariate analysis, then, uses a cluster that separates the governors into two categories: governors who were removed under Medvedev and those removed or right-censored under Putin (before or after Medvedev). Doing so allows us to estimate the effects of the covariates that interest us while controlling for dependence among these observations. Table 5.1 presents our multivariate analysis, which includes three reduced models and one full model. The three reduced models allow us Page 111 →to assess step-by-step the effects of regional characteristics, gubernatorial performance, and the governors’ ability to resist replacement on the governors’ survival rates. For each model, we present, in separate columns, the hazard ratio (exponentiated coefficient) and robust standard error for each independent variable. Estimated hazard ratios greater than one indicate an increased hazard of having the event, namely a gubernatorial replacement (Mills 2011, 94). Estimated hazards less than one indicate a decreased hazard.Page 112 → Hazard ratios of one indicate no association between the covariate and the hazard. The final four rows of table 5.1 present two common goodness-of-fit estimates for Cox proportional hazard models, the likelihood ratio test and the score (logrank) test as well as their significance levels. Table 5.1. Cox Estimations Clustering Governors Based on the President(s) They Served Reduced Model Reduced Model 2 Reduced Model 3 Full Model 1 Hazard Robust Hazard Robust Hazard Ratio Robust S.E. Hazard Ratio Robust S.E. Ratio S.E. Ratio S.E. Regional Characteristics Population 1.012 .160 1.191* .105 (standardized) Percent Non1.004*** .000 1.001 .009 Russian Performance Crime Rate 1.034*** .009 1.033*** .010 Change Life Expectancy 0.966 .068 0.928 .067 Change GRP per capita 1.018*** .002 1.020*** .004 Change Recent Presidential 0.858* .062 0.516* .135 P.O.E.

Ensconced Prior Tenure (Months) Governor’s Age Interaction Term NonRussian*Pres. P.O.E. N 177 Likelihood ratio 2.11 test p .349 Score (logrank) 2.22 test p .329

1.004

.006

1.004

.004

0.991

.014

1.007

.036

1.006***

.000

160

177

160

49.92

5.39

60.11

.000

.068

.000

76.50

5.60

89.14

.000

.061

.000

Note: POE = the percent of a region’s eligible voters having voted for the Kremlin’s candidate, either Putin or Medvedev, in the most recent presidential election, standardized by the election in which they occurred. * and *** indicates significance at the .05, and .001 levels, respectively, for two-tailed tests. In the first reduced model, governors of larger regions are removed at a slightly higher rate. By subtracting one from the hazard ratio and multiplying that by 100 (Mills 2011, 95), we learn that, holding the region’s ethnic composition constant, a one-unit increase in the standardized population size results in a 12 percent increase in the hazard to the governor: (1.012 в€’ 1) Г—100 = 1.2. The percentage of non-Russians in a region has a statistically significant but modest positive impact. According to the exponentiated coefficient, a 1% increase in the percentage of non-Russians results in a .4% increase in the hazard, holding population constant. Comparing this result to figure 5.3, we see that controlling for population reduces the extent to which the governors of ethnically non-Russian regions were at risk of being removed. These two variables by themselves, however, do not provide adequate goodness-of-fit. The second reduced model in table 5.1 presents the results when only gubernatorial performance variables are included in the model. We present these equations separately, rather than in an additive fashion, due to high collinearity between the percentage of non-Russians in a region and our measure of the governor’s political performance. To capture the governors’ administrative performance, we calculate the percent change in several measures of regional standard of living for the two years preceding when the governor is dismissed, when possible. We use three indicators of quality of life change: overall crime rate, life expectancy, and gross regional product per capita (GRP).14 The data on crime and life expectancy were unavailable at the time of writing beyond 2011 and for gross regional product, beyond 2010. We therefore use the most recent two-year period. Because the data are annual, we use the previous full year’s number as the numerator if the governor’s departure occurs prior to July 1, and the year of the departure, if it occurs in the second half of the year. We expect annual increases in the crime rate to increase the hazard of replacement while increases in life expectancy and gross regional product per capita to lower the hazard. We base our measure of a governor’s political performance on federal election results, specifically those that cut closest to home for the Kremlin, presidential elections. We calculate the votes received by the Kremlin’s candidate (Putin in 2004 and 2012 and Medvedev in 2008) as a percentage of the total eligible voters in that region (POE; for more on this measure, Page 113 →see chapter 7). Using POE incorporates both high vote totals and high levels of turnout. Because the elections in 2004, 2008, and 2012 were at different stages of Putin’s

efforts to construct a nationwide regime and differed in the average levels of POE, what constituted a particularly high or low level of POE varied by election. We therefore standardized the POE scores within elections to reflect that a score of, for instance, 60% had a different meaning in 2004 than in 2008 or 2012. The second reduced model in table 5.1 is significantly stronger than the first. It also supports the expectation that a governor’s performance influences his or her survival. The hazard ratio for change in crime is significant and positive; it indicates that a unit increase in the region’s crime rate exposes a governor to a 3.4% higher likelihood of removal. Life expectancy increases reduce the hazard to the governor, as expected. The effect of positive economic change, however, goes against the expectation, causing a small increase in hazard in this model. The strongest impact in this model comes from POE, the measure of the governor’s political performance. A one-unit increase in this measure lowers the likelihood of removal by over 14%. The third reduced model estimates the ability of more ensconced governors to resist replacement. As previously discussed, governors who served prior to 2005 did so by winning a popular election, which suggests that they possessed their own basis of legitimacy (either among the population, among regional elites, or both). While age and prior tenure are certainly correlated (Pearson r = 0.512), we include age of the governor in the model because those serving in office prior to the appointment era were not necessarily so old as to justify removal on the basis of age-related factors. Including age in the model, then, allows us to differentiate more carefully among the governors and to attain a more accurate estimate of the impact of prior tenure. It turns out that both variables have only a modest impact. The impact of age goes against the expectation that older governors would be at higher risk. In this model, each additional year of a governor’s age reduces his or her risk by just less than 1%. This model does not, however, include the variables shown to matter in the earlier models, so these patterns may reflect that misspecification. The final equation in table 5.1 presents the full model. Simultaneously including indicators for the regional characteristics, gubernatorial performance, and the degree to which governors are ensconced in their regions produces a model with the best goodness-of-fit measures so far. Both the likelihood ratio test and the score (logrank) test attain significance below the .001 level. Page 114 →Besides including all of the variables examined previously, the full model adds an interaction term: the percent of the region’s residents who are ethnically non-Russian multiplied by POE. Including the interaction term and its components in the model allows us to determine whether a synergistic relationship might exist between the percentage of non-Russians from a region and the level of support from the region for Kremlinbacked presidential candidates. Remember, our expectations were that the Kremlin would move more cautiously in regions with more non-Russians and would more likely reward governors overseeing high levels of proKremlin electoral support. The results of the full model appear to justify the inclusion of the interaction term. Its hazard ratio is statistically significant, and governors from regions producing highly deferential election results and having higher percentages of non-Russians are more likely to be removed from office. Interpreting the fixed effects for the interaction term is less straightforward than for the other variables in the analysis. With all other covariates held constant, the product of a 1% increase in POE and a 1% increase in the percentage of nonRussians yields just under a 1% increase in the likelihood of gubernatorial removal. While the increase initially may seem trivial, since increases in the components of the interaction effect have a multiplicative effect on the term itself, they also have a multiplicative effect on the hazard ratio. In this case, the function for converting the interaction effect into the hazard ratio involves multiplying the product of the two components’ increases by a factor of 10 [x Г— (.6) Г— (.6) = 3.6 where x = 10]. That is, a 10% increase in both components would yield a 360% increase in the hazard. In the full model, we find that the governors of larger regions are more at risk of being removed: 19% higher for each standard deviation of population size. The region’s ethnic composition has only a minor impact, when controlling for the other variables in the model. The measures of the region’s socioeconomic performance remain little changed from Reduced Model 2, including the unexpected result that an improved economic situation is associated with a modestly higher chance of being removed. The measure of political performance becomes even stronger than in the previous model, now indicating that a unit change in support for the Kremlin is

associated with a 48% lower risk of removal. Neither of our measures of the governor’s political roots in the region is statistically or substantively significant in the full model. What the pattern of replacements from 2005–12 shows, therefore, is that the Kremlin’s decision-making was based primarily on a governor’s Page 115 →performance in office, more so than on the governor’s personal situation or the region’s characteristics. Moreover, the type of gubernatorial performance that has shaped the Kremlin’s choices most strongly is, by a wide margin, the governor’s ability to provide the Kremlin with strong voting results in federal elections. This is not to deny that some governors were removed during the 2005–12 period because the level of regional governance had sharply deteriorated. The patterns we find, though, strongly suggest that the Kremlin was focused on how well the regional leaderships were playing their assigned roles in maintaining the national-level political regime’s control of the political space. We also find little to suggest that, overall, the Kremlin treated more carefully governors who had stronger roots in their region. Given the weight the Kremlin assigned to governors’ political performance, the positive hazard rate for a region’s population size likely reflects how important it was for the Kremlin to replace an ineffective governor of a populous region with a governor who could do a better job at harvesting the large number of potential votes there. Creating an effective vertical of power has been at the heart of Putin’s efforts to construct and maintain a political regime that can control a large, widespread, and socially complex country. He did not want to or could not create a regime based solely on a single party that penetrates all the regions. United Russia provided an electoral vehicle for controlling legislatures, but it was federalized in the sense that, in many regions, it rested on machines the governors had previously established. In addition to consolidating the position of a ruling party, Putin sought to provide carrots and sticks to incentivize regional leaderships in the desired direction. Prior to 2005, the Kremlin’s efforts to influence the outcomes of gubernatorial elections had proven modest (see chapter 4). The decision to end the elections was likely related to this difficulty in controlling regional elites and policies when governors were elected. The switch to bureaucratic control over the governors certainly amplified the Kremlin’s voice when it told governors what it wanted. What it wanted, our results show, was favorable election results. Despite this, the 2011–12 federal election cycle demonstrated that, in many regions, the leaderships were politically weak or inept, or both. Medvedev’s subsequent announcement that gubernatorial elections would be reinstated was seen as a concession to the social forces that had become highly disgruntled. In addition, however, seven years of stronger Kremlin control over regional leaderships had not definitively strengthened the vertical of power. A new tack was in order. How interesting, then, that the new mechanism for controlled gubernatorial elections was hardly in place Page 116 →before a move began to return some regions to presidential appointment, with the corresponding law signed in April 2013. At this point, it remains unclear what trusses and struts can solidify Russia’s vertical of power.

Conclusion We asked why Russia’s governors were replaced when they were and tackled this question using data on all governors of Russian regions who were in office during the period from January 2005 through June 2012. Using survival or event-history analysis, we examined the tenures of those who left the governorship during this period as well as those for whom we did not know when their tenure would end, because they remained in office at the end of the period we study. Although the Russian president changes from Putin to Medvedev and back during this period, the two politicians are closely allied and represent a single presidential administration. Thus, the patterns we find over the seven and a half years shed light on the Kremlin’s strategies and its relations with regional leaders. When governors were replaced during the 2005–12 appointment period was influenced by a combination of factors: the region’s size and ethnic composition and aspects of the governor’s administrative and political performance. The Kremlin relatively rarely took advantage of its ability to fire a governor, reflecting the political complications associated with doing so rather than the Russian president’s formal power.

Replacements were more frequent in some periods than others, notably in the run-up to a federal election cycle or when the appointment rules were about to end in favor of a new form of elections. The hazard to a governor of being replaced in a given month was most strongly affected by how well the governor had performed in the electoral sphere. Regional importance and socioeconomic performance were also considerations, but less influential when controlling for political performance. Despite a law that gave Russia’s central executive authorities bureaucratic control over the governors, the Kremlin faced a complicated challenge of elite management throughout the country. The Russian president had the upper hand but not political carte blanche. Rather, our analyses identify patterns indicative of power plays and bargaining, suggesting that seven and a half years of presidential appointments failed to solve the challenge of effective elite management in a nondemocratic regime. Given the complexities of Russia along with the absence of a unifying ideology or party structure, the question of whether it can be solved lingers.

Page 117 →

Six Mobilized Voter Turnout and the Spread of Regional Authoritarianism We have now traced the cross-regional differences in gubernatorial voting patterns. In this chapter, we turn to a different indicator of regional political differences: participation in major elections—that is, voter turnout. Any polity holding elections benefits when a high proportion of the eligible population votes. Among other reasons, high turnout indicates public support for the political system of which the elections are part and thereby legitimizes those in power. As noted in chapter 2, post-Soviet Russia’s leaders have been particularly concerned about turnout, in part because until 2006 elections with inadequate turnout were invalid and required new elections. Procedurally, voting is easy in Russia. Every Russian must register his or her place of residence with the police. Having done that, they need do nothing extra to be eligible to vote.1 All citizens aged 18 and above are eligible to vote, whether residing in the country or not, unless they are imprisoned or have been declared legally incompetent. While they must show identification at the polling place, this is an insubstantial barrier since all Russian citizens have a government-issued internal passport for identification. Nevertheless, not all eligible voters take part in any given election, in Russia or elsewhere. In this chapter, we examine the cross-regional patterns in turnout as well as how those patterns have changed over time. In Russia’s early elections,Page 118 → turnout levels shed light on Russians’ political behavior. The patterns are similar to comparative findings from other countries, including that voting is more common among the elderly and that, as the easiest form of political participation, it is more common among populations with fewer political resources. However, our analyses make clear that over time, especially in the 2000s, turnout levels cease to reflect individual political behavior. They instead reveal the ability of authoritarian regional elites to mobilize (or fabricate) high turnout levels.

What Does Turnout Reveal? Given our concern with differences in the regions as political sites, we want to understand why regional turnout levels vary, sometimes widely, during a given election as well as over time. We do not assume that turnout is, in and of itself, an indicator of democracy. Dahl’s (1971) influential depiction of popular participation as one of the two dimensions on which real-world democracy rests led some cross-national researchers (e.g., Vanhanen 1984, 1990, 1997; Lijphart 2012) to employ turnout levels as an indicator of democracy’s strength or quality. The drawbacks of doing this are apparent. For one thing, contexts vary widely, including between countries with mandatory voting and those without. In addition, as Jackman (1987, 406) has pointed out, it is unclear whether lower turnout reflects political apathy and alienation or public satisfaction with politics. Nor has the wide-ranging literature on comparative turnout (for reviews, see Blais 2006; Geys 2006; Smets and van Ham 2013) agreed on the key contextual factors that might allow one to use higher- or lower-than-expected turnout as an indicator. In authoritarian contexts, moreover, the political meaning of turnout and abstention can be quite different than in democracies (Karklins 1986; Roeder 1989; Power and Roberts 1995; Shi 1999). Achieving high turnout helps a competitive authoritarian regime by demonstrating the strength of the leadership (Magaloni and Kricheli 2010). If not an indicator of democracy, turnout nonetheless helps illuminate important aspects of a country’s politics. In democracies, it is a low-cost way for citizens to participate in politics and therefore tells scholars much about popular attitudes and values as well as about institutional and structural features of a polity (Blais 2007; Dalton 2013, 42–44). In nondemocracies, where elections are by definition less free and fair, turnout levels provide murkier insights into popular outlooks because voters may be pressured or enticed to vote when they otherwise would not have. We Page 119 →ask whether Russia’s regional turnout patterns correspond to the presence in the region of democracy or authoritarian control. That is, do regions vary in ways that conform to existing knowledge on voter mobilization in democratic settings given their social composition, the nature of the

election, or the strength of party organizations or interest groups? If not, what factors might suggest authoritarian control, and how do these relationships change from one election to another?

Regional Turnout Patterns over Time How have regional turnout levels varied among the regions and across different elections? Our data come from official Russian sources (as detailed in appendix 1). Although, conceptually, turnout is easy to understand—as a ratio of those who voted to those who might have voted—it can be measured in several ways (see Geys 2006, 638–39). For our data, the numerator is the sum of the valid and invalid votes cast. The denominator is the total number of those on official voter lists, or eligible voters. Table 6.1 provides the national level, regional mean, and regional range for each of Russia’s federal elections—that is for the presidency or Duma, from 1991–2012. Nationwide, turnout was highest in the 1991 presidential race, prior to the end of the Soviet Union. It was lowest in the very next election, when public disillusionment from the October 1993 crisis was high. In most other elections, the nationwide percentage was in the sixties. In all the electoral cycles, turnout for a Duma election is lower than for the presidential election it precedes. The averages of the regional turnout levels track the nationwide turnout level pretty closely. They will not match it, of course, because each region has a different adult population size and therefore a different share in the country’s potential and actual voters. In every election but one, the average regional turnout is somewhat higher than the nationwide average. Arithmetically, this can happen when less populous regions have relatively higher turnout. Of more note is the breadth of the variation across the regions. In 1993, the gap between the high and low regions exceeded 50%, or half the 100-point scale; in half of the 12 other elections, the gap exceeded 40% . In both 1991 and 1993, Tatarstan’s leadership encouraged residents to boycott the federal elections (Walker 1996; McAuley 1997, 58 and 78), causing its turnout levels to be far below any other region, especially in 1993. If one excludes Tatarstan, the regional spreads in 1991 and 1993 are less extreme. Even so, in all the post-Soviet federal elections,Page 120 → the range from high region to low region is broad enough to merit investigation. Table 6.1 also shows that the distributions of regional turnout grow broader over time. If one excludes Tatarstan in 1991 and 1993, the range is less than 30% throughout the 1990s. The range exceeds 35% in the 2000 presidential election and broadens even further thereafter, reaching 52% in 2011 before declining slightly in 2012. The standard deviations of the regional distributions jump to almost 9 in 2003 and 10 or more Page 121 →subsequently. For comparison, the gap in turnout across US states in the 2004 presidential election was 29% versus the 47% gap in Russia that same year, while the American standard deviation was 7.0 versus Russia’s 11.1 (McDonald 2011). Table 6.1. Regional Patterns in Voter Turnout Rates in Russian Federal Elections Nationwide Mean Regional Election Voter Turnout (std. High Region Low Region Difference Turnout dev.) 1991 Presi76.2% 85.7% 49.1% 74.7% 36.6%(Tatarstan) dential (6.5) (Karachaevo-Cherkassia) (22.7% excl. Tatarstan) 56.1% 70.2% 13.8% 56.4% 1993 Duma 54.3% (7.8) (Karachaevo-Cherkassia) (Tatarstan) (30.5% excl. Tatarstan) 65.0% 75.5% 53.1% 1995 Duma 64.4% 22.4% (4.7) (Altai Repub., Belgorod) (Sverdlovsk) 1996 69.8% 79.0% 59.6% Presidential, 69.7% 19.4% (3.9) (Bashkortostan) (Murmansk) round 1

1996 Presidential 68.8% round 2 1999 Duma 60.5% 2000 Presi68.6% dential 2003 Duma 55.7% 2004 64.3% Presidential 2007 Duma 63.7% 2008 69.7% Presidential 2011 Duma 60.1% 2012 Presi65.3% dential

68.6% (4.8)

83.4%(Ingushetia)

56.6% (Murmansk)

62.7% (5.5) 69.7 (5.3) 56.9% (8.7)

78.4% (Kabardino-Balkaria) 92.8% (Ingushetia) 87.1% (Chechnya)

49.9% 28.5% (Leningrad Oblast) 57.3% 35.5% (Evensk) 43.9% 43.2% (St. Petersburg)

65.7% (11.1)

97.6% (Kabardino-Balkaria)

50.7% (Krasnoyarsk)

65.9 (11.8) 70.4% (10.3) 61.6% (12.9) 66.7% (10.0)

99.5% (Chechnya) 92.8% (Mordova) 99.5% (Chechnya) 99.6% (Chechnya)

51.5% (St. Petersburg) 52.9% (Ivanovo) 47.1% (Irkutsk) 53.1% (Vladimir)

26.8%

46.9% 48.0% 39.9% 52.4% 46.5%

Source: See appendix 1. Note: Percentages for Duma elections are those for the party-list voting. Fig. 6.1. Distributions of Regional Turnout Levels. For each election, the turnout levels for regions at the 25th and 75th percentiles define the box, with the line through the middle indicating the level of the median region. The lines extending out from each box show the distance from the 25th or 75th percentile to the “adjacent values,” which are defined as the 75th percentile plus 1.5 times the distance between the 75th and 25th percentiles. Even more extreme outliers are shown as dots or diamonds. Presidential elections have grey boxes, while clear boxes indicate Duma elections. (Source: See appendix 1.) Each election’s full distribution of turnout values therefore deserves attention. FigureВ 6.1 illustrates these distributions using “box-and-whisker” plots. For each federal election from 1991 through 2012, it shows the median, the range in which the bulk of the regions fall, and extreme outliers. (Presidential elections have grey boxes, while clear boxes indicate Duma elections.) Tatarstan’s extremely low values in 1991 and 1993 are indicated in the figure. No region stands out in any subsequent election to the same extent for its low values. Throughout the 2000s, though, what is Page 122 →remarkable is the growth in high outliers, regions that appear as dots above the high adjacent value in the given plot. From 2004 on, each election features several regions that exceed 90% turnout—20–30 percentage points above the median region’s turnout.2 Of interest as well are the cases at the low end. In elections such as 2004–2008, with so many regions exceeding 90% turnout, quite a few barely exceed 50%. Ten regions fall between 50 and 55% in 2004 and 2007, two in 2008 and one in 2012. In 2011, despite the campaign to vote for anyone other than United Russia and even as quite a few regions reported extremely high turnout, turnout fell in most regions compared to 2007. That year, two-fifths or 34 of the regions fell below 55%, and eight were below 50%. Contrasting the distributions for the first round of the 1996 presidential race with the second round, three weeks later, shows an intriguing increase in the high outliers as well as the low end of the spectrum. The regional median declines between the rounds, as do the lowest values. Both Vladimir and the Nenetsk Autonomous Okrug see their turnout fall by at least 5%. Three regions see their turnout in the second round increase by more than 5%: Mordova, Kabardino-Balkaria, and, with a 13.5% increase, Ingushetia. All three of these regions have the status of republics. The top-most dot in the plot

for the 1996 second round indicates Ingushetia, which went from 46% to 80% to lead all regions. The other high outlier is Bashkortostan, which had the highest turnout in the first round and only increased by 1.5% for the second round. As we discuss in more detail below, the distribution of turnout levels depicted by the box-and-whisker chart is approximately normal (again, when excluding Tatarstan in 1991 and 1993) in each of the races prior to 2000. From 2000 on, the turnout distributions all contain a notable skewness caused by the minority of particularly high values. In addition to how substantially the levels vary, the levels themselves are politically important. We can illustrate this by comparing the two rounds of the 1996 presidential race. The regions that increased their turnout between rounds of the election also voted significantly more strongly for Yeltsin. Among the 10 regions that increased their turnout by more than 1.5%, the average vote share for Yeltsin in the second round was almost 15 points higher than the remaining regions.3 The same relationship holds in comparing regions that increased their turnout dramatically between the 1990s and the 2000s. The 15 regions that had turnout at least eight percentage points higher in the 2004 presidential election than in the first round of 1996 gave Putin 85% of their votes, 15 points higher than other regions (t-test = в€’7.45 [.000]). By contrast, those same regions differedPage 123 →insignificantly in their voting for Yeltsin in 1996’s first round. When we turn to examining voting results more fully in chapter 7, we will see that the high turnout regions are consistently more strongly supportive of the Kremlin. Table 6.2 provides information on turnout in elections for governor. It distinguishes four periods. In 1991–93, the principle of electing regional executive leaders was not yet established, and only 24 regions held them. In the second group are elections held subsequently during Yeltsin’s presidency, though we include the elections held in December 1999 with the first years of Putin’s presidency. As explained in chapter 2, the elections held at that time already exhibited primarily post-Yeltsin dynamics, including the rapid organization of a pro-Putin party with the backing of many regional governors. The third period goes through the end of 2002. We separated elections held from 2004 through January 2005 as a fourth period to capture the increasing control the federal leadership was obtaining over regional affairs—the strengthening of the so-called vertical of power. The turnout patterns in gubernatorial races show several similarities to those in the federal elections. Overall turnout is highest in the earliest years, prior to the conflict between Yeltsin and the legislature. The range of turnout levels between the highest and lowest regions in each time period is significant, even larger than for the federal elections. The high region in each time period is a republic, and the ethnically non-Russian regions dominated the high end in federal elections as well. St. Petersburg’s low turnout level for its gubernatorial election in 2003 parallels its basement Page 124 →showing in the 2003 and 2007 Duma races. More generally, regions in which turnout is high for the gubernatorial elections also tended to be among the highest in turnout for federal elections and vice-versa. The correlation between turnout in gubernatorial elections held in the late 1990s and turnout in the 1999 Duma election is .56 (.000), while between the gubernatorial elections held in the early 2000s and the 2003 Duma election the correlation is .63 (.000). Table 6.2. Regional Patterns in Voter Turnout Rates in Russian Gubernatorial Elections Election Mean Regional Turnout (std. dev.) High Region Low Region Difference Ingushetia Cheliabinsk 1991–93 65.4% February 1993 April 1993 61.2% n = 24 (12.6) 92.7% 31.5% Kabardino-Balkaria Tiumen 1994–Sept. 1999 56.1% January 1997 December 1996 67.4% n = 87 (11.9) 97.7% 30.3%

Dec. 1999–2002 56.2% n = 77 (10.8) 2003–2005 n = 45

55.7% (12.2)

Kabardino-Balkaria January 2002 85.9% Chechnya October 2003 87.7%

Vladimir December 2000 51.9% 34.0% St. Petersburg September 2003 59.1% 28.6%

Source: See appendix 1. We see, then, clear-cut differences among the regions in their propensity to have high or low turnout levels across both regional-level and federal-level elections. These differences are comparatively very large, which suggests they reflect important political differences. A region’s status as a republic or autonomous region seems implicated in the explanation somehow, but we must also consider the factors that previous research has shown to influence comparative turnout levels.

Arguments for Why Regions Differ in the Proportion of Their Residents Who Vote Our interest is not in the individual-level question of why a citizen makes (and follows through on) the decision to vote. Turnout, as distinct from a decision whether or not to vote, is an aggregate-level phenomenon: “It is a feature of an electorate not a voter” (Franklin 2004, 16). Yet we cannot neglect the individual characteristics of the regions’ residents. Each potential voter engages in a “voting calculus” in deciding whether or not to vote. Personal characteristics that have been found to operate on individual choice may then help us understand turnout aggregated to the regional level. In the absence of fraud, regional turnout is the sum of the decisions made by all the potential individual voters. If the elderly tend to vote in greater numbers than the young, a region with more elderly as a share of its population will feature higher turnout than a region with comparatively few elderly. This makes it fair to incorporate the aggregate values of individual-level factors. When doing so, however, it is important not to imagine individuals as entirely self-mobilizing, making their decision to vote or not in isolation. An individual’s decision to vote, even when no undue pressure is at work, is very much a social decision (Franklin 2004, ch. 2). Moreover, various kinds of pressure always come into play, some democratically legitimate, others illegitimate. We must therefore be attentive not only to regional social composition but also to the forces that mobilize or deter voters. Page 125 →When comparing countries’ turnout levels, their institutions, especially the rules for holding elections and translating votes into seats, matter greatly (Perez-Linan 2001; Geys 2006; Blais 2007). However, because the institutional features of the elections we study are constant across Russia’s regions, they will not help us explain cross-regional patterns. In seeking to explain those patterns, therefore, we will look at the composition of regional populaces as well as at how the social, economic, and political contexts vary. The Characteristics of the Regions’ Residents Which characteristics of a region’s residents might influence regional turnout? The comparative literature has generated many hypotheses; Smets and van Ham (2013) review over 170 factors proposed in the literature. Perhaps the most prominent group of such factors relate to an individual’s socioeconomic status. Those with higher standing tend to have increased access to education, which in turn can increase interest in and understanding of politics (or what has been called “cognitive mobilization” [Dalton 1984; Inglehart 1990, 337]). In addition, higher socioeconomic standing provides an individual with increased resources of various sorts, including time, money, and civic skills (Brady, Verba, and Schlozman 1995).4 Survey-based studies of voting participation in Russia’s early elections show mixed support for this perspective. Wyman et al. (1995, 597) found that their respondents with only elementary and those with higher education recalled voting in the 1993 Duma election in the same proportion. McAllister and White (1998, 24), however, find with a multivariate analysis that higher education gives a statistically significant, though modest, boost to turnout in the same election. In his analysis of Russians’ voting behavior in the 1995 and 1996 elections, Colton (2000, 40–47) finds that education and the extent of one’s social connection both make voting more likely.

In comparing regions, differences in regional socioeconomic development can stand in for the individual-level resources hypotheses. Cross-national research has found that countries at higher levels of socioeconomic development tend to have higher turnout levels (Russett 1964; Powell 1982, 37–38; Blais 2006, 116–17). As we noted in chapter 2, Russia’s regions vary sharply in their levels of socioeconomic development. More developed regions will have more residents who belong to the middle or upper levels of class or socioeconomic position. The comparatively higher education, interest, or resources may dispose them to participate more readily, producing higher turnout levels. Regional differences in education Page 126 →and resources are captured in part by the extent to which a region’s population is urban or rural. Highly urban regions are likely to have more residents with higher education as well as with access to information and social connectedness, all of which can bolster an individual’s political resources.5 Yet the comparative literature also suggests a rival hypothesis about how regional development might relate to regional turnout. In this view, different forms of political participation stem from different sources. Inglehart (1990, 338–41) and others (e.g., Barnes et al. 1979) have pointed to the difference between low-complexity and high-complexity forms of participation. Elites can mobilize large numbers of people into low-complexity acts. More difficult forms of participation, such as demonstrating or joining social movements, do not involve as many people. Voting falls into the first category and, presumably, requires no more resources than literacy. By contrast, higher levels of education and more resources make people less likely to follow elite guidance and more likely to engage in the second category of participation. This line of argument suggests that regions with a higher share of the well-educated and better-off citizens may well exhibit lower levels of voter turnout because a smaller share of their populaces are amenable to elite mobilization and a larger share have access to other forms of political engagement. In a different setting, Blaydes (2010, 115–23) found that district literacy levels in Mubarak’s Egypt correlated negatively with turnout: turnout was higher in districts where more eligible voters were illiterate. This finding illustrates the importance of taking regime type and the level of aggregation seriously as even the most basic individual-level resource may correlate with turnout in an unexpected way when aggregated and examined within authoritarian elections. For some scholars, socioeconomic status is less critical in individuals’ decisions to vote than social learning or culture (Easton and Dennis 1969; Sniderman 1975; Inglehart 1990; Putnam 1993). Voting, like other acts of political participation, is a learned behavior influenced by formative experiences in youth and early adulthood. Those experiences rather than an individual’s situation at the time of an election may be key. Formative experiences may be of many sorts, of course. The propensity to vote may be stronger in one culture than another, shaped by lessons passed down over long time periods through families (Eckstein 1988; Inglehart 1988; Putnam 1993). In the Russian context, this suggests that we incorporate a measure of the extent to which different regions are primarily of Russian cultural heritage or nationality. In interpreting this measure, however, we bear in mind that the regions with substantial non-Russian populations all have a distinct formal-institutional status (republic or autonomous region) Page 127 →along with other differences that might better explain the relationship between regional nationality composition and turnout. A different take on the role of social learning is Inglehart’s (1977, 1990) well-known thesis that political behavior can be shaped by the extent to which a society’s situation causes people in their teens and young adult years to favor material security or less materialistic, more aspirational values. Generational cohorts have differing value structures depending on whether the period when that generation came of age is relatively benign. The arguments about resources and social learning both suggest that the age composition of a region’s populace may influence its turnout levels. We will use a measure of the proportion of each region’s population at the time of an election that is elderly, or a “pensioner” in Russian terms.6 The elderly have the resource of time and have gained political experience that the young may lack. In terms of social learning, they were raised under the Soviet regime, when voting was mandatory and a prominent social obligation. Using survey data, White and McAllister (2004, 93–94), for example, find that those who came of age after the end of the Soviet Union are significantly less likely to vote. Pensioners also came of age prior to the economic crisis Russia experienced from the late 1980s through the late 1990s. Therefore, higher proportions of the elderly ought, ceteris paribus, to be associated with higher turnout.

The Characteristics of the Election Cross-national as well as subnational research documents that the characteristics of one election versus another matter strongly for turnout (Cox and Munger 1989; Franklin 2004; Blais 2006; Geys 2006, but see Smets and van Ham 2013 for a null finding). Specifically, one should expect higher turnout levels when elections are more competitive, when turnout rates are more likely to influence election results, and when the election results themselves are more likely to have tangible effects on policy outcomes. Not only do such circumstances raise an individual’s interest in participating, they call forth much greater effort from societal or elite forces seeking to mobilize voters. For gubernatorial elections, differences in the competitiveness of each race should influence cross-regional turnout differences. For the federal elections, this factor will not influence cross-regional differences for a given election but may well influence differences from one election to another. As noted above, turnout for presidential elections is consistently higher than for the legislative election in the same electoral cycle. From this perspective, such a pattern makes sense because Page 128 →the Russian presidency has substantially greater influence on policy than the State Duma. Similarly, the presence or absence of concurrent elections has been demonstrated to influence turnout levels. In the United States, for example, changing the dates of gubernatorial elections to coincide with midterm elections rather than presidential elections has not only shortened presidential coattails but also lowered voter turnout levels (Pacek 1994; Lewis-Beck 1998). One study of Russia (Marsh 2002, 129) suggests that liberal politicians try to schedule regional elections to coincide with national elections as a way to increase their electoral support since younger, more liberal members of the electorate participate at a lower rate when there is no federal election. Or, it may be that a region holding a gubernatorial election at the same time as a federal election will generate additional turnout because of interest in the former race. A Region’s Economic and Social Conditions at the Time of the Election For citizens in a setting with a poor quality of life, especially when life quality has notably declined prior to an election, will that awareness produce higher turnout—so as to register a protest, or perhaps even “throw the bums out”—or will it cause voters to become disillusioned and less engaged in politics, hence less likely to vote? Following Rosenstone’s (1982) findings from the United States that bad economic times depress electoral turnout, it may be that regions with weaker economic conditions will experience lower voter turnout (see also Radcliff 1992). Other studies, however, have found no strong relationship (e.g., Southwell 1988; Blais and Dobrzynska 1998), perhaps because the two possible outcomes of misery offset each other. For comparable reasons, scholars have investigated whether crime retards political engagement, including turnout, and found some evidence that it does (Brehm and Rahn 1997; Akhmetkarimov 2008). We therefore incorporate measures of regional economic fortunes and crime levels in the year of each election. Regional Political Contexts In reviewing possible influences on cross-regional turnout levels, we now move away from those that extrapolate from an image of individual citizens independently deciding whether or not to go to the polls. As noted above, voters never make this decision entirely independently, but the role Page 129 →of elites in some settings can be particularly strong. Elites can mobilize citizens to vote in a range of ways, from contacting and advertising through material inducements to coercion. Thus, alongside information about what kind of people a region’s potential voters are, information on the regional political context is vital. In other words, the key may not be the factors that distinguish the residents of one region from each other as individuals but what unites them and distinguishes them (in the aggregate) from residents of other polities. Party and Non-Party Sources of Regional Political Competitiveness

In established democracies, the party system is a central aspect of a society’s political organization, since political parties organize and shape electoral competition. Settings with more than one electorally competitive party are necessary to create close elections that will attract voters. If opposition parties are essentially irrelevant in a region, their supporters will have less incentive to vote, ceteris paribus, than in a more competitive region. In

addition, parties can work to promote or suppress voter turnout as part of their electoral strategies (see, for example, Huckfeldt and Sprague 1992). Various studies have shown that contacting by political parties increases the likelihood of an individual voting (e.g., Rosenstone and Hansen 1993; Wielhouwer and Lockerbie 1994; Perez-Linan 2001; Karp, Banducci, and Bowler 2008). Well into the 2000s, post-Soviet Russia’s political parties had shallow roots in society and weak organization, with the exception of the Communist Party. Hough (1998, 688) portrays the initial incarnation of post-Soviet parties as “highly personalistic and ephemeral.” Meanwhile, Rose (2000) depicts a tendency for Russian parties to “float” above society, often supplying candidates and policy rather than responding to voter demands. Other research (e.g., Fish 1995; Miller et al. 2000; Moser 2001), gives Russia’s parties more credit for influencing the course of the transition through the early 2000s. Still, only the Communist Party of the Russian Federation effectively got citizens to the polls until the spread of Putin’s United Russia party throughout the country, which eliminated other parties’ influence. In Russia’s regions, the degree to which political parties have been vehicles for candidates has varied (Golosov 1999, 2004). Clem and Craumer (1998), in particular, found that if a party with strong organizational structures at the grassroots (like the Communist Party) was popular in a region, that region had higher voter turnout. But these conclusions Page 130 →stem from bivariate analysis, which raises questions about whether the impact of party development holds when, say, socioeconomic conditions are controlled for. However, the impact of party development may well be attenuated. Hale (2006) has shown how “party substitutes” have functioned in Russia, especially at the regional level, and other countries. For politicians seeking to win an election, parties are not the sole organization providing them with funding and personnel with which to advertise and to mobilize turnout. In Russia’s regions, those resources can also be provided to candidates by what Hale (2005, 152) calls “politicized financial-industrial groups” or by political machines led by a region’s governor. Still, since political parties have developed slowly in the regions, variation in party development represents an important consideration when comparing turnout rates. In our multivariate analyses we include, for appropriate elections, a measure of the importance of political parties in the regional legislature as well of support for the Communist Party. Patronage, Ethnic Ties

One of the most powerful means of mobilizing citizens around the world to vote are political machines structured by patron-client relations, or patronage. Patronage in this sense refers to networks of “individualized, reciprocal political relations” (Willerton 1992, 6; see also Hale 2015, 9–10). Patronage has a long history in Russia. During the Soviet era, it dominated the nomenklatura system of the Communist Party (Willerton 1992; Hale 2015, ch. 3; on regional patterns, see Willerton and Reisinger 1991). Factors such as common service with an important official in a particular geographic region, one’s succession of job assignments, and ascriptive traits, like ethnic identity and gender, became important determinants of recruitment and mobility (Barghoorn and Remington 1986). The Soviet Union’s collapse did not dismantle these existing patron-client ties. Instead, patronage networks have continued their importance (Willerton 1998; Glinski and Reddaway 1999; Hale 2015). Patronage norms have even survived the removal of nomenklatura members from positions of political power. Although Russia’s patronage networks have their roots in the Soviet era, Hale (2003) argues that the ability of regional executives, especially those in Russia’s ethnic regions, to convert them into political machines is a more recent phenomenon. Some of Russia’s regions have been characterized by rivalries between patronage networks headed by different politicians. In other regions, however, the governor heads the dominant network and its Page 131 →associated electoral machine.7 When this is the case, the governor’s team has a range of resources for exerting pressure and enticing support. As we noted in chapter 2, Russians refer to these resources as “administrative resources.” Indeed, a strong regional machine and the use of administrative resources may give regional leaderships the ability not only to get voters to the poll but to inflate turnout figures fraudulently. Through ballot stuffing or manipulation of election figures (i.e., either the nominator or the denominator of the ratio that is turnout), authorities can produce turnout levels beyond the number of actual voters who participate.

[M]any students of regional politics in RussiaВ .В .В . tend to view unusually high levels of voter turnout as a clear manifestation of what is euphemistically referred to as an “administrative resource.” The term applies to different kinds of pressures exerted upon voters by the regional authorities in order to provide desired electoral returns, the very ability to exert such pressures being associated with the measure of authoritarianism possessed by the given regional regime. From this perspective, exceptionally high voter turnouts witnessed in some of the regions stemmed from forced voter mobilisation. (Golosov 2006, 53) Myagkov, Ordeshook, and Shakin (2009) use district-level data to make a strong case that turnout was inflated in this way in many parts of the country. As we analyze regional turnout levels, then, we want to determine whether certain regional characteristics are capturing differences in the ability of the regional executive to mobilize voters. One such characteristic is the ethnic composition. In the early post-Soviet years, the regions with larger non-Russian populations demanded and acquired greater leeway in local matters in exchange for their support in the consolidation of Russia’s federal system (Solnick 1995; Treisman 1999; Kahn 2002). In many of these regions, the elite leadership team maintained itself during the transition to post-Soviet Russia and was therefore in a stronger position to bargain with the center.8 It would therefore be more likely to be in a position to organize voting within its region. In fact, despite the importance of party organizations to patronage in other contexts (e.g., Kitschelt 1995; Chandra 2004), the ethnic networks buttressing regional political machines in Russia operated during the 1990s as party substitutes (Hale 2007, 231). Indeed, Golosov (2014c) has traced how governors could switch their machines from one political party to another. Thanks to the confluence of ethnicityPage 132 → and geography, the governors of Russia’s ethnic regions are in a strong position to monitor ethnic voting in their regions effectively, then reward or punish accordingly (Hale 2007, 231). That republics dominated the high-turnout regions in tables 6.1 and 6.2 supports the idea that their elites were influencing turnout. Thus to the extent that these ethnic networks function as regionalized power pyramids (Hale 2015, 110–11), they also represent potentially valuable prizes to a central government interested in driving voters to the polls. When elite mobilization is strong, the expectations discussed above that urban, educated, and otherwise highresource individuals will vote at a higher rate can be reversed. Although less educated individuals may inherently be less likely to participate in politics, they are more likely to be responsive to powerful patrons. Rural residents have been shown in some settings to be similarly more influenced by elites. Monroe (1977, 76) contends that the ability of local politicians to distribute government jobs in the rural counties of Illinois explain why these counties enjoyed higher levels of voter turnout. Thus, local politicians who play prominent roles in the provision of services and jobs should be more able to determine rates of voter turnout. Darr and Hesli (2010) find that traditional rural social networks are the most powerful mobilizer of voters in Kyrgyzstan. Kuenzi and Lambright (2011), likewise, find that political parties get more “bang for their buck” by focusing their mobilization efforts on rural settings in Africa, thanks in part to higher levels of social connectivity and lower levels of anonymity when compared to urban areas. According to Fish (1995), the link between dependency in the workplace and city size proved particularly prominent in the success of conservative politicians during the 1989 and 1990 elections to the Soviet and Russian Congress of People’s Deputies, respectively. Similarly, Myagkov, Ordeshook, and Shakin (2009, 90–92) find the implausibly high turnout levels in the 2000s to be significantly more common in rural districts. These findings, together with the arguments noted above by Inglehart and others about voting as a form of participation more amenable than others to elite mobilization, produce an expectation that our measure of a region’s socioeconomic development will correlate negatively with turnout. Finally, we will examine a regional economic characteristic that is not as much about the state of the economy at the time of the election as about the availability to the regional elite of financial resources. This is whether or not the region has significant fuel extraction, specifically the availability of oil and gas. The expectation that this can help elites control a populationPage 133 → has garnered a wide, though still controversial literature (as reviewed in previous chapters).

Analyses

Federal Elections We begin by noting that the relative closeness of the races has no or very little influence on turnout levels among the federal elections. For one thing, only a few of the 13 elections were at all closely contested. Among presidential elections, the closest was the first round in 1996, when Yeltsin received 35% to Zyuganov’s 32% . Yeltsin won the runoff by 14 percentage points, and the next closest presidential election, in 2000, had a gap between first and second places of 24 percentage points. The gaps were substantially larger in 1991 and after 2000. As table 6.1 shows, however, that 1996 first-round race produced turnout that was only marginally higher than in 2000, 2004, and 2012, equal to 2008 and substantially lower than in 1991. The federal legislative elections in the 1990s were close in a multiparty sense. Yet the closest race, in 1999, had a lower overall turnout and lower regional average than in 1995 as well as in 2007, when United Russia’s dominant victory was a foregone conclusion. Differing levels of excitement about the elections are thus not driving differing numbers of voters to the polls. The larger point that characteristics of the elections matter is, however, supported by the contrast between turnout in 1991 and 1993. In the former election, many Russians were eager to support the status of Russia by seeing it get a president. The 1993 election took place soon after the October crisis, with its bloodshed, as well as during a period when the economy was in terrible shape. Enthusiasm for politics was low. Similarly, lower turnout in legislative elections than in presidential ones shows that election characteristics matter at one level. Next we present a multivariate analysis for each federal election. We assess regional socioeconomic development with an index variable (explained more fully in appendix 1) that incorporates seven intercorrelated measures: higher education, urban population, population density, roads, telephone distribution, and attendance at museums and theaters.9 As an indicator of cultural distinctions between Russians and non-Russians, we include the percentage of non-Russian residents in each region. This Page 134 →is a more effective measure than distinguishing the ethnic and nonethnic regions because regions in the former category range widely in how nonRussian their populations are. To capture regional age distribution, we employ pensioners as a proportion of the population. We measure regional suffering by including the rate of violent crime in the region and the region’s change in income compared to the previous year. To estimate whether resource rents are available in a region, we again create a scale using the standardized values of regional levels of oil and natural gas production, two natural resources that received the most credit for Russia’s economic boom during the Putin era (Hill 2004; McFaul and Stoner-Weiss 2008; Rutland 2008b).10 We have transformed these data using the natural log, which reduces the influence of the high-output regions. If energy-rich regions differ systematically from others, we expect that it will be in increased mobilized turnout. Our model also includes two political features of each region. One is the support for the Communist Party, which had the strongest track record for mobilizing supporters. We use the proportion of votes received in the given election by the CPRF or its presidential candidate. The other is the percentage of deputies in the regional assembly who were affiliated with a political party, a measure that captures the strength in each region of political parties.11 The results of regressing regional turnover on these variables are presented in table 6.3. Each cell presents the coefficient from robust regression, the statistical significance of that coefficient, and the standardized coefficient or beta weight. Robust regression reduces the impact on one’s results of cases that are extreme outliers. Although we are not drawing inferences from a sample to a population, the significance levels indicate when a coefficient’s value is strong. Standardizing the coefficients allows us to examine how the relative substantive impact of each explanatory variable changes across the elections. They indicate by how many standard-deviation units turnout is changed when the explanatory variable changes by one standard deviation. We draw attention to standardized coefficients with an absolute value above .15. Looking across the first rows in table 6.3, with the results for our index of regional socioeconomic development, we see clear evidence against resources theory: regions with higher socioeconomic development levels do not produce higher turnout. The relationship is negative every year, strongly so for quite a few elections. That is, turnout is higher in regions at lower levels of development when holding the other factors in the model constant.

Indeed, the bivariate correlations between development and turnout Page 135 →are also negative for each election, increasingly strongly so under Putin. Contrary to expectations of resources theory, Russia’s more urban, better educated, and more connected regions are not engaging more actively in voting as a form of political participation. Clem and Craumer (1998, 62) identified this pattern from the early 1990s elections, and it has continued to the present. The non-Russian proportion of the regional population shows a striking chronological pattern. It has an insignificant influence in the 1991–95 elections and the first round of 1996, then becomes strongly influential each election thereafter. Although this measure incorporates variation even within the categories of republics /autonomous regions and other regions, it captures that distinction fairly well. The status of the regions does not change from election to election. Yet this electoral outcome clearly underwent a systematic change in the second half of the 1990s. While the Republic of Tatarstan set the standard for low turnout in 1991 and 1993, it and other ethnic regions became markedly higher in their turnout levels from 1996 on. If the increase in the role of ethnicity over time is surprising, so is the decline and then disappearance of the impact of a region’s age distribution. The 1991–2003 elections show the pattern hypothesized above: regions with a larger elderly proportion of their population will have higher turnout, all else equal. However, that relationship declines from 1996 to 2003 and disappears entirely from 2004 to 2012. Of the two variables that measure living conditions within the region at the time of an election, violent crime and income change, only the former has a noteworthy impact. More crime-prone regions have lower turnout in most of the elections except 2012.12 Oil and natural gas production is associated with lower turnout in 1991–95 but has only a limited impact in subsequent elections. The negative relationship in 1991 and 1993 is partially the result of the election boycotts in oil-producing Tatarstan. Like the proportion of pensioners, the variable measuring support for the CPRF initially has the expected impact. In the 1991–95 elections, more support for the communists is associated with higher turnout. But the positive influence of this variable on turnout levels ceases after 1995. Moreover, unlike the proportion of pensioners, this variable actually reverses its sign, becoming substantially negative from 1999 on. Our measure of interparty competitiveness at the regional level, the proportion of party-nominated deputies in the regional legislatures, shows little effect in most of the elections. With the exception of CPRF mobilizing in the first half of the 1990s, then, we see little effect of party mobilization on regional turnout. Page 136 → Table 6.3. Turnout in Federal Elections Regressed on Regional Socioeconomic and Political Characteristics (A) 1991–99 1991 1993 1995 1996, 1 1996, 2 1999 в€’2.453 в€’1.983 в€’1.309 в€’1.452 в€’.693 в€’2.260 Index of socioeconomic (.000) (.001) (.040) (.024) (.364) (.005) development [в€’.326] [в€’.181] [в€’.225] [в€’.278] в€’.101 [в€’.362] .042 .044 в€’.054 .009 .064 .161 Non-Russians as % of (.060) (.183) (.064) (.764) (.040) (.000) population в€’.045 в€’.074 в€’.056 .091 .317 .482 .586 .429 .078 .314 .383 .472 Pensioners as a % of the (.000) (.003) (.594) (.043) (.025) (.004) population .278 .188 .151 .330 .325 .293 в€’.026 в€’.054 в€’.039 в€’.032 в€’.039 в€’.038 Violent crime (.002) (.000) (.001) (.019) (.011) (.002) [в€’.236] [в€’.190] [в€’.290] [в€’.319] [в€’.355] [в€’.346]

Income as % of the previous year

в€’.474 (.001) [в€’.303]

в€’.047 (.399) [в€’.155] в€’.309 (.030) [в€’.189]

.070 (.162) .167 в€’.036 (.857) в€’.020

.054 (.326) .126 .021 (.925) в€’.003

в€’.040 (.615) в€’.117 в€’.186 (.284) в€’.033

.186 (.001) .265

.179 (.027) .380

.068 (.236) .189

.002 (.959) .017

в€’.012 (.819) в€’.018

в€’.142 (.072) [в€’.235]

—

в€’.015 (.510) [в€’.180]

в€’.019 (.393) в€’.088

в€’.000 (.990) в€’.024

в€’.007 (.760) в€’.075

в€’.019 (.461) в€’.048

60.997 (.000) .30 72

45.753 (.000) .36 79

69.287 (.000) .27 76

58.535 (.000) .27 76

57.655 (.000) .28 76

57.572 (.000) .21 85

—

—

−.121 Oil and Natural Gas Production, (.262) logged [−.240] Voting for CPRF or CPRF candidate Party-affiliated regional deputies Constant Adjusted r-squared Number of cases (B) 2000–12

2000 в€’1.201 Index of Socioeconomic (.052) Development [в€’.215] .106 Non-Russians as % of (.000) Population .627 .360 Pensioners as a % of the (.005) Population .229 в€’.026 Violent Crime (.000) [в€’.488] в€’.007 Income as % of the (.874) Previous Year в€’.026 .015 Oil and Natural Gas (.888) Production, logged в€’.070 в€’.081 Voting for CPRF or (.126) CPRF Candidate [в€’.207] .003 Party-affiliated Regional (.860) Deputies в€’.012 63.716 Constant (.000) Adjusted r-squared .45 Number of cases 88 Source: See appendix 1.

2003 в€’2.978 (.007) [в€’.241] .153 (.000) .575 .365 (.113) .199 в€’.013 (.123) [в€’.237] .121 (.254) .108 .119 (.397) .036 в€’.469 (.012) [в€’.183] .057 (.057) .092 34.813 (.012) .31 87

2004 в€’1.971 (.134) в€’.147 .269 (.000) .614 в€’.227 (.432) .039 в€’.028 (.009) [в€’.276] .012 (.921) в€’.029 в€’.156 (.547) в€’.071 в€’.189 (.303) в€’.145 в€’.046 (.232) в€’.092 70.548 (.000) .52 87

2007 в€’2.882 (.037) [в€’.184] .163 (.002) .323 в€’.137 (.668) в€’.047 в€’.033 (.012) [в€’.230] .044 (.809) .019 в€’.323 (.126) в€’.119 в€’1.420 (.000) [в€’.492] в€’.004 (.915) в€’.009 77.906 (.002) .60 82

2008 в€’1.575 (.202) в€’.112 .125 (.009) .283 .052 (.857) в€’.023 в€’.022 (.061) [в€’.170] .453 (.246) .067 в€’.305 (.144) в€’.080 в€’1.121 (.000) [в€’.540] .033 (.349) .029 35.753 (.429) .53 82

2011 2012 .150 в€’.230 (.918) (.823) в€’.033 в€’.024 .283 .233 (.000) (.000) .556 .526 .425 .307 (.224) (.218) .068 .051 в€’.028 в€’.006 (.037) (.557) [в€’.197] в€’.049 .019 .149 (.942) (.450) .020 .085 в€’.273 в€’.239 (.204) (.128) в€’.114 в€’.068 в€’.875 в€’.505 (.000) (.001) [в€’.401] [в€’.385] в€’.026 в€’.053 (.513) (.070) .051 [в€’.164] 63.324 45.625 (.053) (.042) .64 .61 82 82

Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Positive beta weights of .15 or over appear in bold. Negative beta weights of в€’.15 or lower appear in bold and in square brackets. Page 137 → Taken together, the performance of the independent variables in the model indicate that Russia’s federal elections changed their character over the country’s first two decades. This change is observable in the late 1990s, but becomes dramatically different in the 2000s. Factors indicative of political competition and democratic processes lose their influence, and some of them are reversed. Gubernatorial Elections What about voter turnout in the elections we analyzed in the previous chapters—the elections for regional governors from 1994 to 2005? They occur across the transition period from when federal elections are less to more controlled. Are the same regional characteristics associated with high or low turnout in gubernatorial elections as in federal elections? Table 6.4 shows the results of multivariate analyses using the same model as in table 6.3. Since each region has several gubernatorial elections from 1994 to 2005 but at no set intervals, we separate them into two multiyear periods corresponding to the Yeltsin and Putin presidencies. The first covers elections held from 1994 through 1999, 74 in total. The second includes elections from 2000 through January 2005, the last gubernatorial election prior to the appointment period. For regions which held two elections during the same period, we use the average of the turnout from the two elections. Page 138 →We see in table 6.4 patterns similar to those found for the federal elections. Regional socioeconomic development has no influence in the 1990s but emerges as a modestly negative influence in the early 2000s. Regions with higher proportions of ethnically non-Russian residents have strongly higher turnout levels in both time periods. The proportion of pensioners in the populace has a positive and strong influence on turnout in the earlier period but becomes noninfluential by the later one. In all these respects, the pattern for gubernatorial elections matches closely those from the federal elections at the equivalent time. The other socioeconomic and politicalPage 139 → variables in the model play no significant part in explaining turnout for gubernatorial elections. Table 6.4. Turnout in Gubernatorial Elections Regressed on Regional Socioeconomic and Political Characteristics 1994–99 2000–2005 1.181 в€’2.291 Index of socioeconomic development (.498) (.189) .079 [в€’.160] Non-Russians as % of population

Pensioners as a % of the population

Violent crime

Income as % of the previous year

.390 (.000) .606 1.255 (.003) .444 в€’.006 (.840) в€’.039 в€’.207 (.189) в€’.146

.264 (.000) .521 в€’.014 (.971) в€’.046 в€’.015 (.296) в€’.113 .020 (.908) .037

в€’.187 Oil and Natural Gas Production, logged (.639) в€’.061 .006 Voting for CPRF or CPRF candidate (.969) в€’.000

.096 (.673) .032 в€’.305 (.319) в€’.066

в€’.012 (.848) в€’.016

.066 (.172) .137

Adjusted r-squared

32.813 (.065) .30

50.242 (.026) .30

Number of cases

74

85

Party-affiliated regional deputies Constant

Source: See appendix 1. Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Positive beta weights of .15 or over appear in bold. Negative beta weights of в€’.15 or lower appear in bold and in square brackets. Are Ethnic Regions Different? The preceding analyses make clear that Russia’s ethnic regions are distinct. Table 6.5 shows the t-tests of whether the average turnout levels for republics and autonomous regions (AOs) are significantly different from the average for other regions. A negative and significant coefficient from a t-test indicates that the average turnout among the ethnic regions is much higher than among the remaining regions. In the 2000 presidential election, for instance, the average turnout level among ethnic regions was 72% compared to 68% for all other regions. In the 2011 parliamentary elections, the gap is 73% to 57%. Clearly, the ethnic regions have been the driving forces behind the increase in extremely high regional turnouts shown in figure 6.1. Another useful way to contrast turnout patterns in ethnic and nonethnic regions is to examine whether turnout levels approximate a normal curve, overall as well as for each of the two subtypes of regions.13 We use a statistical test that assesses the likelihood that a distribution is skewed (D’Agostino, Belanger, and D’Agostino 1990; Royston 1991; StataCorp 2011, 2002–6). Table 6.6 shows the probabilities, first, for all the regions, Page 140 →then just for the ethnic regions, and then just for the nonethnic regions. Non-normal distributions have low values, indicating confidence in the ability to reject the hypothesis of normality. We have put those below .05 in bold. Table 6.5. Differences in Federal Turnout between Ethnic Regions and Other Regions t-test (significance) 1991 .696 (.488) 1993 в€’.090 (.929) 1995 .926 (.357) 1996, 1st round .701 (.485) 1996, 2nd round в€’1.168 (.246) 1999 в€’3.735 (.000)

2000

в€’3.266 (.000)

2003 2004 2007 2008

в€’5.250 (.000) в€’6.628 (.000) в€’6.216 (.000) в€’5.367 (.000)

2011 2012

в€’6.512 (.000) в€’6.458 (.000)

Note: Results of two-tailed t-test of difference in mean turnout between republics or AOs and all other regions. Negative t-scores indicate that the ethnic regions have a higher average turnout score than the nonethnic regions. Statistically significant scores are bolded. The skewing of results toward high turnout levels that is visually apparent in figure 6.1 begins to reach significance in 1999 and remains strong thereafter. A minority of regions had turnout far above the others. Yet the distribution of values among the republics and AOs has no significant skewness except in 2000 and 2003. The nonethnic regions approximate normality from 1991 through 2000. In each of the subsequent elections, though, the distribution is significantly skewed. Although this may seem like an anomaly, that minority of regions with the very high turnout levels represents most of the ethnic regions. The highest-turnout ethnic regions are less distant from the others and thus do not create a tail in the distribution. Among the nonethnic regions, by contrast, a small minority had turnout levels distinctly higher than the rest, creating a skewed distribution. In the context of the overall distribution being skewed from 1999 on, the absence of skewness among ethnic regions clarifies that they are clustered at the top of the overall distribution. However, the clear and intriguing differences between republics and nonrepublics do not rest only on the formal status of republic or nonrepublic.Page 141 → The percent of a region’s residents who are non-Russian is associated with higher turnout even within the set of ethnic regions (cf. Hale 2003; Saikkonen 2016). For every election in which turnout and percent non-Russian has a significant bivariate correlation (from the 1996 runoff election on), the correlation between the same two variables just within the group of ethnic regions is also significant and as strong or stronger. For example, in the 2004 presidential race, the correlation across all regions is .72 (.000). For ethnic regions only, it is .65 (.000). For all other regions, it is .07 (.612). Republics and AOs differ in the proportion of their population that is of the titular nationality versus Russian versus another ethnicity. Table 6.6. Normality of Turnout Distributions: Overall, for Ethnic Regions and for Nonethnic Regions (A) 1991–99 Type of Region 1991 1993 1995 1996, 1st rd. 1996, 2nd rd. 1999 .000 .000 Overall [excl. Tatarstan] .917 .102 .323 .042 [.148] [.441] .000 .000 Republics & AOs [excl. Tatarstan] .425 .714 .222 .380 [.249] [.183] Other Regions .294 .975 .400 .207 .020 .304 (B) 2000–12 Type of Region 2000 2003 2004 2007 2008 2011 2012 Overall [excl. Tatarstan] .000 .000 .000 .000 .009 .000 .000 Republics & AOs [excl. Tatarstan] .019 .042 .251 .712 .649 .578 .434 Other Regions .525 .004 .002 .004 .029 .001 .001 Spatial Patterning of Regional Turnout Levels

Subsequent chapters will present more detail about the forces driving the over-time shift in the characteristics of Russia’s federal elections. We can foreshadow some of those findings here by examining spatial patterning. We use membership in the eight federal administrative districts (discussed in chapter 2) as an indication of geographic proximity. We use all eight for analyzing all the elections even though the eighth, the North Caucasus District, was only broken out of the Southern District in 2010. Table 6.7 presents descriptive statistics for how turnout varied in each election by federal district. If the highturnout and low-turnout regions were distributed randomly across the country, the F-statistics would be small and statistically insignificant. Yet the F-statistics are significant for each election except 1999, indicating that regions in a given federal district tend to be similarly high or low. Turnout levels cluster geographically, in other words. In each election except 1995 and the first round in 1996, the district with the highest average regional turnout is the North Caucasus, which is entirely made up of republics. From 1991 to 2000, the low-turnout regions are to be found in either the Far East or the Urals. From 2003 on, however, it is the Northwest district that trails the other federal districts. It resists the trend for turnout to rise from 2003 on. Between 2003 and 2004, the gap between the North Caucasus and the Northwest districts jumps from 13.7 to 29.2 percentage points and remains large. Both the high turnout in the North Caucasus and the low turnout in the Northwest are politically important. Given that both regional leaderships and the Kremlin want turnout to reach satisfactory levels, receiving over 85% from the North Caucasus region is a big step toward these goals. On the one hand, the potential voting pool in the North Caucasus district is much smaller than in the Northwest district (5.7 million vs. 11 million adult residents). Page 142 →On the other, the higher turnout in the former meant that in years such as 2004 and 2011, the gap in actual votes was small. Since most votes cast in the North Caucasus go to the Kremlin’s party or candidate, the political impact of the federal district’s high turnout is substantial. The distinctiveness of regions by their location may, however, go beyond just the high and low outliers. Table 6.8 shows, for each election, whether each federal district is significantly different in its turnout from the regions elsewhere in the country, as well as in what direction its distinctiveness lies. Insignificant values are omitted. As noted above, the North Caucasus republics have significantly higher turnout in almost every election. The Volga regions are higher than typical from the second round in 1996 through 2003 but then become more typical thereafter (as other regions’ turnout levels rise to meet theirs). Prior to 2003, the Urals, Siberian, and Far Eastern districts each have elections in which their turnout levels are distinctly lower than typical. Logistical barriers to participation may have been at play, with many of the regions in these districts being vast and cold. From 2003, though, it is the Northwest district, as noted, that becomes a low outlier by virtue of resisting the trend for heightened turnout. With regard to the regions Page 143 →in the Central district, which has 27% of the country’s eligible voters in 2012 and includes the city of Moscow and Moscow Oblast—the two most populous regions in the country—we see another case of the directionality switching from the early elections to the later ones. Through 1996, the regions of the Central district are among the most participatory. During these years, the regions in the Central district tend to have three characteristics that, based on table 6.3, are associated with higher turnout: a sharply Page 144 →higher proportion of pensioners in the population, less violent crime, and more support for the CPRF. (T-tests for the difference in means between the Central regions and others are significant for all three, and all three have significant coefficients in table 6.3 for some or all of the first four federal elections.) The Central regions lose their distinctiveness from 1999 to 2003. From 2004 to 2008, however, they become distinctively low. They, like the regions in the Northwest district, are resisting the rise in turnout levels that occurs elsewhere. Table 6.7. Distribution of Average Turnout Values by Federal Administrative District (A) 1991–2000 District 1991 1993 1995 1996, 1st rd. 1996, 2nd rd. 1999 2000 (No. of regions) 81.4 61.6 68.4 72.3 75.2 66.7 79.2 Highest district average N. Cauc. N. Cauc. Central Central N. Cauc. N. Cauc. N. Cauc.

72.8 Far East Mean of district averages 76.01 2.54 F-statistic from ANOVA (.020) Lowest district average

49.0 Urals 55.7 2.95 (.009)

60.4 Urals 64.4 2.67 (.016)

66.9 Urals 69.6 3.36 (.003)

65.9 Far East 68.7 5.82 (.000)

59.6 Urals 62.6 1.49 (.183)

(B) 2003–12 District (No. of regions)

2003

Highest district average

2004

2007

2008

2011

2012

66.7 87.2 84.9 85.3 86.4 83.2 N. Cauc. N. Cauc. N. Cauc. N. Cauc. N. Cauc. N. Cauc.

53.0 NW Mean of district averages 57.1 2.86 F-statistic from ANOVA (.010) Lowest district average

58.0 NW 67.2 8.35 (.000)

58.2 NW 67.1 5.80 (.000)

63.9 NW 71.5 4.07 (.001)

55.2 NW 62.9 7.22 (.000)

61.1 NW 67.6 5.18 (.000)

Table 6.8. Distinctiveness of Federal Administrative Districts (A) 1991–99 District 1991 1993 1995 1996, 1 1996, 2 1999 (No. of regions) Northwest (11) −2.82−3.10−3.64−3.23−1.76 Central (18) (.003) (.001) (.000) (.001) (.041) −1.66 −1.70 Volga (14) (.051) (.047) Southern (6) −2.23 −1.82 −4.11−1.87 North Caucasus (7) (.014) (.036) (.000) (.033) [2.37] [2.58] [1.89] Urals (6) (.010) (.006) (.031) [2.61] [2.42] Siberian (12) (.005) (.009) [1.81] [1.82] [1.96] Far Eastern (9) (.037) (.036) (.027) (B) 2000–2012 District (No. of regions) Northwest (11) Central (18) Volga (14) Southern (6)

2000

2003

2004

2007

[1.62] [2.55] [2.39] (.055) (.006) (.010) [2.08] [2.07] (.020) (.021) в€’1.80 в€’2.28 (.038) (.013)

2008

2011

2012

[2.30] [1.80] [2.02] (.012) (.038) (.023) [1.68] (.048)

66.6 Far East 69.9 7.32 (.000)

North Caucasus (7)

в€’5.70в€’3.26в€’6.50в€’5.07в€’4.41в€’6.58в€’5.26 (.000) (.001) (.000) (.000) (.000) (.000) (.000)

Urals (6) Siberian (12) Far Eastern (9)

[1.93] (.029)

Note: Results of one-tailed t-test of difference in mean turnout between regions in the given federal administrative district and all other regions. (Negative t-scores indicate that the given federal administrative district has a higher average turnout score than elsewhere.) Blank cells indicate that the t-test did not reach statistical significance of .05. If the significance level is .01 or lower, the t-score is bolded. Square brackets indicate districts with scores significantly lower than elsewhere. The overall pattern in table 6.8 suggests that the federal elections fall into three periods: 1991 through 1996’s first round, 1996’s second round though 2000, and 2003 on. As discussed above, regional turnout levels in Russia’s early post-Soviet elections correspond to explanations that stress more democratic forms of mobilization.14 Then, from the middle of the 1990s into the early part of the 2000s, the pattern ceases to fit with such explanations. To a growing extent, what drives high turnout levels are indicators of elite pressure and fraud of the sort documented by Myagkov, Ordeshook, and Shakin (2009, 71–137).

Conclusion Voter turnout across Russia’s regions has varied substantially, both cross-sectionally for each election and over time. From the early 1990s through the 2012 presidential election, turnout shows a clear trend. It ceases to be an aggregate reflection of individual voters’ behavior. Instead, it becomes an aggregate reflection of regional elites’ willingness and ability to organize voting at high levels or fraudulently report the level of turnout. Although turnout nationwide fluctuates within a reasonably narrow band from 1993 to 2012, the distance between the high-turnout and low-turnout regions is large. The gap between high-turnout and low-turnout regions in gubernatorial elections is even more substantial and correlates with the pattern in federal elections. For most of the elections (excluding 1991 and 1993), the noteworthy outlying regions are at the high end of the scale, including those that report turnout above 90%. From the second round of the 1996 presidential election on, the number of regions able to mobilize high voter turnout grows, and the turnout levels they are able to produce increases. In the first four federal elections, the high-turnout regions are those with more elderly and more sympathy for the Communist Party. This suggests that individual voters faced less elite pressure in most regional Page 145 →settings, allowing those who did not wish to vote to abstain. In subsequent elections, into the early 2000s, age and ideology begin slipping as correlates of regional turnout. From 2003 on, they have virtually no effect. Instead, the factor most strongly associated with high turnout becomes the percent of a region’s population that is non-Russian. The ethnic regions located in Russia’s North Caucasus lead the way, but those located elsewhere also see remarkable turnout levels. By contrast, regions located in Russia’s northwest area, predominantly nonethnic regions, anchor the lower end of the turnout distribution in most elections. Even though many of the ethnic regions producing the highest turnout are relatively small in population, they contribute in an important way to the federal leadership’s nationwide goals for turnout by virtue of turning out 80–90% or more of their eligible voters. The switch over time in what determines regional turnout, particularly the way that the elderly and communist supporters cease to matter, strongly suggests that nationwide patterns reflect regional elites’ skills. Leaders of ethnic regions demonstrate the ability and willingness to provide high turnout earlier than other regional leaderships. When the latter also become better able to promote higher turnout in the 2000s, the former continue to exceed them, pushing their turnout levels to particularly high levels. Of course, regional leaders promote high turnout because it benefits them. In the abstract, for example, we might

expect a given region’s leadership to discourage turnout in order to deny a challenger extra votes. That we instead find a steady growth in the prevalence of notably high turnout levels means governors expect the extra votes produced by higher turnout to go toward candidates or parties they support. Clearly, then, we need to examine not just how many people have voted in Russia’s elections but for whom the votes have been cast. In the next chapter, we build on our analyses of regional turnout by delving into the patterns of regional voting in federal elections.

Page 146 →

Seven The Regional Role in Federal Election Outcomes For a decade and a half, Russia’s federal elections have held no suspense about their outcome. The Kremlin’s candidate or party has won without serious challenge. Less often noted, however, is the more complex and interesting story of which regions provide those votes. In some regions, support has been tepid even when the Kremlin’s control was strong. In other regions, with the number growing over time, the support for the Kremlin, including the proportion of voters who came to the polls, approached the levels of Soviet-era elections (see Swearer 1961; Zaslavsky and Brim 1978; Karklins 1986): more than 99% in both turnout and proParty voting. In other words, the official results coming from some regions is, like Ivory Snow detergent, “99 and 44/100% pure” (Zimmerman 1987, 336). We have introduced the federal elections and their contexts in chapter 2. Here, we reiterate that each of Russia’s federal elections has mattered politically, particularly for the country’s leadership. Even when the possibility of the Kremlin losing the election was far-fetched, achieving targeted results was a political priority. Elections play important roles in stabilizing and operating a regime that is not a democracy. Achieving a high level of public participation in the elections, turnout, is part of this. Chapter 6’s analyses of turnout in federal elections brought forth some of the key patterns in regional voting in Russia’s federal elections. Of course, who receives the votes is critical as well. We turn to that in this chapter. We use a measure of the results of federal elections that reflects both Page 147 →the reported turnout level and the level of voting for the Kremlin’s candidate or party. We examine the distributions across all of Russia’s regions in each of the 13 federal elections from 1991 to 2012. High scores on this measure are strongly associated with nondemocratic political conditions in a region and a greater ability of the regional executive-branch leadership to control the voting outcomes. We find nondemocratic regions are a feature of Russian federal elections throughout the 1990s, but not to the extent and consistency as under Putin, especially from 2004 to 2008. Although the regions that deliver extraordinary vote totals to the Kremlin tend to be small in population size and economic output, their votes alter the outcomes disproportionately. A region’s level of socioeconomic development helps explain its voting returns in the 1990s but largely ceases to do so from 1999 on. More developed regions in the 1990s are more pro-reformist and their support for the Kremlin comes from that. In the 2000s, these regions are generally less supportive of the Kremlin. While socioeconomic development falls away as an influence on regional voting, the proportion of ethnically non-Russian residents predicts support for the Kremlin over the entire period but particularly sharply in the 2000s. This reflects but goes beyond the distinction between republics or AOs and regions with other statuses. Regions that give the Kremlin the extraordinarily strong electoral support are geographically clustered in every election but especially so from 2003 to 2012: the regions in the North Caucasus provide the Kremlin with electoral support at levels significantly exceeding other regions, even when the other regions’ average levels of deference are objectively high. The regions in the Central and Northwestern districts, which include the cities of Moscow and St. Petersburg, resist the trend toward high Kremlin support.

Interpreting Regional Election Results When competitive elections are relatively fairly conducted and reported, data on regional differences illuminate the political preferences of geographically clustered groups of Russian citizens, whether they are voting for presidential candidates or for “democratic” parties versus “communist” or “nationalist” parties. Numerous excellent studies in the 1990s and early 2000s used federal election results broken down by region (or, even further, by electoral district) to examine Russian electoral behavior. Data from 13 federal elections across four presidencies are available. This ought to have allowed scholars to generate even deeper insights by examining whether Russians’ voting behavior changes from 1991 to 2012.

Page 148 →Unfortunately, violations of democratic election procedures and a noncompetitive electoral environment characterize Russia’s recent federal elections to such an extent that the official voting results tell us little about Russian voters’ preferences or decision-making. A wide array of careful data analysis and reporting has demonstrated the problems beyond doubt. They fall into three broad categories. One involves elite behavior that, while legal or quasi-legal, chokes off political competition in favor of a preferred candidate or party. This includes vast differences in funding and media access available to incumbent or incumbent-favored competitors (Zadorin 2000; Oates 2003; White and Oates 2003; Sharafutdinova 2007; Golosov 2009). It also includes various “dirty tricks” such as the once trendy practice of encouraging an unknown to run who has the same or almost the same name as a viable challenger (Maksimov 1999; Minchenko 2001; Miroshnichenko 2003, 258–65; Stoliarov 2003, 216–22; Wilson 2005, 62; Smirnov 2008). In a second category are the means by which executive officials control who votes and how. The common term for such levers is “administrative resources” (see, e.g., Nikolaev 2000; Minchenko 2001; Stoliarov 2003, 216–22; Vorontsova and Zvonovskii 2003; Mikhailov 2004, 198–99; Zvonovskii 2004; Baker and Glasser 2005, 322; Wilson 2005, ch. 4; Buzin and Liubarev 2008; White 2011a). A few examples of what practices the term covers would include pressuring employees of a firm dependent on government contracts (Hale 2007, 228–29), having a court strike a rival candidate or party off the ballot, using state officials to get out the vote, and thuggery against a rival’s supporters (Kirkow 1998, 116–17). A third category consists of the cases when the results are falsified, which can be done through any of numerous means (Lowenhardt 1997; Paramonov and Kirichenko 2007; Myagkov, Ordeshook, and Shakin 2009; Lukinova, Myagkov, and Ordeshook 2011). These practices vary by region. Indeed, the regional level seems to be the key locus for how federal campaigns will play out and how the elections will be managed. Beryozkin, Myagkov, and Ordeshook (2003, 170) put it this way: As the sole focal point of political power within their domain, it is only reasonable to suppose that a governor is uniquely positioned to coordinate political activity there—the endorsements of newspaper editors, the actions of industrial managers, the efforts of directors of collective farms, and so on. As we will see below, regional vote totals and turnout levels in recent years are supportive of the Kremlin to a highly improbable extent if Page 149 →they actually reflected voter behavior. Prior to Putin’s presidency, some regions were notable for violating democratic rules of electoral conduct. By the time of the federal elections in 2003–2004, however, such violations are much more widespread. Regions still differ in their support for the Kremlin but within much narrower bounds and around a much higher average level. Notwithstanding Putin’s high popularity among Russians (Whitefield 2005; Rose, Mishler, and Munro 2011, 125; Levada Analytical Center 2012), the change in voting results is too sharp and rapid to rest primarily on a change in voter preferences. While thus not a source of information on the Russian mass public, data from the federal elections allows us to track the strength and activities of the regions’ governing political organizations. Our strategy, therefore, is to deploy the regional results of federal elections as indicators of one kind of strength on the part of a region’s leadership: the ability to deliver turnout levels and vote totals for a particular party or candidate that exceed what could occur if elites were not pressuring voters or falsifying results. Such strength is organizational strength—state employees and others loyal to the region’s leadership have been formed into some type of political “machine” to bring the reported results about. In a loose sense, the term political machine can include any political organization, from those violating no democratic norms to those that entirely control elections in fully closed authoritarian regimes, including the former Soviet Union. However, political science scholarship on political machines (e.g., Key 1936; Banfield and Wilson 1963; Scott 1972; Cornelius 1977) has used the term to characterize situations in the middle, such as Russia’s hybrid regime: a machine in this sense is an organization that uses what Russians would come to call administrative resources to dominate the electoral field when elites cannot be fully certain of electoral outcomes. (For comparisons of Russia’s regions to political machines elsewhere, see Hale 2003, 2010; Gel’man 2013). In some regions, the organization that had operated during Soviet times remained extant through the introduction of competitive elections and the end of the Soviet Union. Machine politics was quickly established. In other regions, the transition disrupted older

patterns and no machine could be built for many years. A feature commonly found in studies of political machines is that they use a political party as a framework to unite the machine’s leadership, mid-level officials, and rank-and-file supporters. Putin’s United Russia party extended its representation throughout the regions in the mid-2000s, and most governors formally joined the party (Reuter 2010). Most studies, however, argue that the party itself is less central to controlling elections than parties for politicalPage 150 → machines elsewhere (Slider 2010; Makarenko 2012; Roberts 2012a; Robinson 2012). United Russia is only one tool with which executive officials, both federal and regional, control elections. In the presence of highly improbable turnout or vote totals, we assume that they reflect not only the regional leadership’s capacity to produce those results but their intent to curry favor with federal executive authorities. When such high totals are absent, we cannot distinguish whether the leadership lacks the capacity or the intention to amplify them. There is little question, however, that most regional leaders have wanted the Kremlin to see them in a favorable light, ceteris paribus. And when the elections in question are for federal offices, the regional leaders lose little if anything by helping the Kremlin meet its goals. As shown above, the practice of regions seeking federal approval by producing favorable electoral results begins not under Putin but at least with the 1995 elections. Based on his fieldwork in Dagestan at the time of the 2003 Duma elections, Ware (2005, 590) notes a logic that can surely be found in many regions: First, it appears that there was massive ballot stuffing in most or all districts, and second, that there was widespread tampering with electoral protocols. Together these manipulations appear to have achieved three tactical and three strategic political objectives. From a tactical standpoint, it appears that the irregularity of the party list election (a) artificially inflated Dagestan’s total voter turnout; (b) massively skewed the result in favour of United Russia, the party of power; and (c) also somewhat skewed the result in favour of the KPRF. The achievement of these three tactical objectives seems to have secured the following three strategic goals: (1) the election successfully seated all Dagestanis who could possibly have won a place in the Duma; (2) the election evidently provided Kremlin officials with a demonstration of the loyalties and abilities of Dagestani officials; (3) it may have left officials in Moscow indebted to officials in Dagestan, thereby facilitating subsequent federal support, particularly in the form of budgetary subsidies. As discussed above, the 1999 Duma race complicates the general pattern because two parties of power were available, and many of the regions that most clearly delivered high vote totals did so for OVR, which did not assume executive power as Unity did. Nonetheless, those regional leaderships that delivered exceedingly high votes for OVR, particularly Ingushetia,Page 151 → had their eyes in one way or another on the prize of the federal executive. From 2005 to 2012, of course, the Kremlin’s favor was a formal requisite, since the Russian president de facto appointed and could remove governors from office. Gel’man (2010a, 12) notes the importance of election results in how the Kremlin judged its appointees: “The capacity to control sub-national electoral politics by any means possible, rather than effective management, was the best predictor of the survival of appointed governors and city mayors.” In the words of the Russian geographer Kozlov (2008, 9), “the characteristic sign of the вЂmanageability’ of regional political regimes is the monopolization of votes. The political stakes of regional leaders are directly expressed by election results. This sharply differentiates regions and allows one to distinguish groups with various levels of political loyalty.”

Measuring Regional Support for the Kremlin For regional leaderships seeking to show their loyalty to the executive leaders of the country in Moscow, they needed their teams to produce both an adequate level of turnout and an adequate share of votes cast for the Kremlin’s candidate or party. Even if a region showed a high vote share for, say, Putin, its performance could well be judged as poor if that high vote share rested on low overall turnout. In other words, a regional leadership producing either relatively high vote share or relatively high turnout will be judged as more loyal than leaderships providing neither and that higher levels of both together are even better. We therefore seek a single number that will measure regional performance in both dimensions.

That number is how many raw votes the Kremlin’s presidential candidate or its preferred party (see table 2.1) received as a percentage of all of a region’s eligible voters. (We henceforth abbreviate “percentage of eligible” as POE.) Whereas the vote share going to a candidate is calculated from the votes received as a percentage of all the votes cast, POE uses all the possible voters in the region as the denominator, thus taking into account the level of turnout. In other words, POE = Vi / EV, where Vi is the raw number of votes cast in that region for candidate or party i and EV is the number of eligible voters in the region. POE must be equal to or less than the vote share for any single candidate or party, and it will usually be substantially lower. In addition, it is easy to show that Vi / EV = Shi Г— T, where Shi is the vote share of candidate or party i and T is the turnout in the region. When data are available, we calculate POE using the raw numbers of votes Page 152 →cast over the raw number of eligible voters (Vi / EV). For other elections, though, we can multiply each region’s vote share for the Kremlin by its turnout to produce the POE figure. While our measure allows for a region to have a high level of vote share but low turnout level or vice versa, in fact the two strongly correlate. Table 7.1 shows the correlations between regional turnout and the vote share received by the Kremlin’s candidate or the party of power. In the early and middle 1990s, the Communist Party held advantages in mobilizing their supporters (White, Rose, and McAllister 1997; Colton 2000). Consequently, voting for Yeltsin or the parties he backed correlates negatively with turnout levels. It is therefore interesting to see that this correlation weakens between the rounds of the 1996 presidential election and is positive thereafter. Indeed, from 2003 on, the correlation is strikingly positive. As a contrast, the correlation at the state level between turnout and US President Obama’s vote share in 2012 was only modestly positive: .24 (.087) (Leip 2012; McDonald 2012). Precisely because our measure is a relatively simple one, we need to validate that it captures what we want it to. Petrov and Titkov (2008) of the Carnegie Moscow Center compiled a quantitative index of electoral democracy using a variety of electoral data from each region as well as an index of regional democracy based on experts’ ratings of how each region is performing in various areas. One must be cautious in judging regions’ comparative level of political noncompetitiveness from the inverse of Page 153 →scales that quantify democratic progress. A polity can be judged nondemocratic for several reasons, including the presence of anarchy or instability. Nonetheless, as a practical matter, measures of regional machine strength or electoral noncompetitiveness should correlate negatively with measures of democracy. Table 7.2 presents bivariate correlations between three of Petrov and Titkov’s measures and POE. The rating of the quality of elections is a component of the index of regional democracy. We present it separately because of its inherent relevance. Table 7.1. Correlations between Pro-Kremlin Vote Share and Regional Turnout, 1991–2012 Pro-Kremlin Party or Candidate Correlation (significance, two-tailed) Yeltsin, 1991 в€’.09 (.427) Russia’s Choice, 1993 в€’.49 (.000) Our Home is Russia, 1995 в€’.20 (.058) Yeltsin, 1996, 1st round в€’.26 (.015) Yeltsin, 1996, 2nd round в€’.10 (.357) Fatherland-All Russia or Unity, 1999 (whichever is higher) +.31 (.004) Putin, 2000 +.47 (.000) United Russia, 2003 +.69 (.000) Putin, 2004 +.70 (.000) United Russia, 2007 +.90 (.000) Medvedev, 2008 +.80 (.000) United Russia, 2011 +.92 (.000) Putin, 2012 +.86 (.000)

Source: See appendix 1. Table 7.2 shows that high vote totals for Yeltsin in 1991 and for Russia’s Choice in 1993 occurred in the regions judged to have more politically open or democratic conditions. In such regions, elite manipulation of the vote would, by definition, be lower than elsewhere, so the vote can be seen as reflecting the preferences of those regions’ voters. In 1991 and 1993, then, voting for Yeltsin and the party backing him was concentrated in a subset of regions that also tended to be more democratic. The reasons for such a correlation have been studied elsewhere (Slider, Gimpel’son, and Chugrov 1994; Sobianin, Gel’man, and Kaiunov 1994; Clem and Craumer 1995). The positive correlations in 1993 arise from a distribution that lacks any region with the high levels of POE that indicate elite control. In 1991, some regions have high POE, but their correlation with the Petrov and Titkov measures is statistically insignificant. In addition, the correlation Page 154 →for the lower-POE regions is also statistically insignificant. Therefore, the positive correlations shown in the first row of table 7.2 reflect a few regions that were strongly pro-Yeltsin. Among them are Sverdlovsk, Yeltsin’s home region, and the city of Moscow. Table 7.2. Correlations of POE with Petrov and Titkov’s Measures of Regional Democracy, 1991–2012 Experts’ Ratings Quality of POE for: Quantitative Index of Electoral Democracy Democratic Index of Regional Democracy Elections Yeltsin, +.17 +.33 +.13 1991 Russia’s +.32 +.45 +.32 Choice, 1993 Our Home is в€’.30 в€’.30 в€’.53 Russia, 1995 Yeltsin, в€’.07 в€’.00 в€’.29 1996, 1st rd. Yeltsin, в€’.04 +.04 в€’.27 1996, 2nd rd. Unity or Fatherlandв€’.35 в€’.46 в€’.42 All Russia, 1999 Putin, 2000 в€’.27 в€’.27 в€’.52 United Russia, 2003 Putin, 2004 United Russia, 2007 Medvedev, 2008 United, Russia, 2011 Putin, 2012

в€’.54

в€’.57

в€’.87

в€’.56

в€’.58

в€’.84

в€’.55

в€’.60

в€’.78

в€’.55

в€’.55

в€’.74

в€’.54

в€’.61

в€’.75

в€’.46

в€’.51

в€’.69

Note: Coefficients in bold are significant at .01 or greater.

In contrast to the positive correlations from the 1991 and 1993 elections, each election thereafter produces a negative correlation between POE and the Petrov and Titkov measures. Except in the first round of the 1996 presidential election, for two of the measures, all of the negative correlations are strong and statistically significant. A high level of POE, therefore, is strongly associated with nondemocratic political conditions in a region and a greater ability of the regional executive-branch leadership to control the voting outcomes.

Patterns of Regional Support Percent of Eligible Voters Figure 7.1 shows the distributions of POE for all federal elections from 1991 to 2012. For each election, we present a box-and-whisker plot of the type introduced in chapter 6. Again, for 1999, we employ the POE score for whichever of the two rival parties of power received the most votes in the given region. Clear box charts with diamond outliers indicate legislative elections, while shaded boxes with circles portray the presidential results. The far left plot in figure 7.1 represents Yeltsin’s run for the presidency in 1991. His national popularity was high, as he ran against Gorbachev and for an independent Russia, largely ignoring the opponents on the ballot. The median line is just about 40% of eligible voters, a strikingly high level for him to average across the country. It is also noteworthy, however, that he has high support in some regions and low support in others. More regions are extreme low outliers (shown by dots), in fact, than are extreme high outliers. Several of the low outlier regions are ethnic republics, Tatarstan and Tyva, for instance, in which regional elites promoted a boycott of the election as part of a campaign for greater autonomy or even sovereignty. It is therefore difficult to judge the extent to which the low support for Yeltsin reflects elite control of the voting in the region. The three regions giving Yeltsin over 60% POE should be treated as representing highly improbable and potentially suspicious outcomes. One of them is Yeltsin’s home region of Sverdlovsk. The other two are Chechnya-Ingushetia, at the time not yet split into separate regions, and Chelyabinsk. Fig. 7.1. Distributions of Regional POE Levels in Federal Elections, 1991–2012. POE = percent of the region’s eligible voters who voted for the Kremlin’s candidate or party. For each election, the POE levels for regions at the 25th and 75th percentiles define the box, with the line through the middle indicating the level of the median region. The lines extending out from each box show the distance from the 25th or 75th percentile to the “adjacent values,” which are defined as the 75th percentile plus 1.5 times the distance between the 75th and 25th percentiles. Even more extreme outliers are shown as dots or diamonds. Presidential elections have grey boxes, while clear boxes indicate Duma elections. (Source: See appendix 1.) Page 155 → The low POE distributions in 1993 and 1995 reflect both the plummeting of Yeltsin’s popularity following the onset of economic reforms and the disillusionment with politics generated by the 1993 political standoff and shelling of the White House. Even though legislative races should produce lower distributions than presidential races, the POE distributions for the 1993 and 1995 elections are much lower even than for subsequent legislative elections. Similarly, by comparison with 1991, the first round of the 1996 presidential election shows a much lower median regional score and a lower POE level for the high outliers (no region exceeded 50%). The median regional POE scores in the legislative elections, the clear boxes, are consistently lower than for the presidential elections that followPage 156 → them. The median for 2007’s legislative race, however, does exceed the median scores in the 1996 and 2000 presidential elections. Support for United Russia peaked in 2007, its median regional POE score falling substantially in 2011 (even according to the official statistics) from 38 to 25. The pattern from 1996 on is for the median POE scores to rise until 2008, then dip somewhat in 2011–12. For the presidential races, the median regional POE score jumps up between the first and second round in 1996 (partly reflecting the reduction to two candidates, partly irregularities noted in the second round), then continues upward under Putin and then Medvedev in 2008, before falling in Putin’s 2012 race. Even more noteworthy, perhaps, is the number and upper reach of the high outliers from 2000 on. Putin’s

victory in the first round in 2000 is helped by Ingushetia delivering to him 78% of all eligible voters. Quite a few regions will exceed that number, however, in 2008 and 2012. In the Duma races of 2003, 2007, and 2011, numerous regions provide United Russia with votes from half or more of their eligible voters and, in 2007 and 2011, many provide votes from more than 70%. The leaderships of these regions have demonstrated their effectiveness at supporting the Kremlin in the electoral sphere. Many of the high POE regions in the more recent years were already producing relatively high POE scores in the mid-1990s and vice versa. Figure 7.2 shows lines that trace how the POE in a given election correlates with the POE in each of the elections prior to it. The thick unbroken line, for example, shows how POE in 2012 correlates with each election noted on the x-axis: 1991 (barely positive), 1993 (strongly negative), 1995 (strongly positive), and so on, until it reaches .93 for the correlation between the December 2011 and March 2012 POE figures. The lines for almost every election follow remarkably similar paths. The 1991 election’s POE levels do not correspond in any strong way to those of later elections. The 1993 POE levels are inversely related to those of almost every election from 1995 on. The two lines that do not dip below zero in 1993 are those for the two rounds of the 1996 election. This tells a story of some regional leaderships being able to create the organizational wherewithal to produce deferential results and maintaining that capability while others lag behind consistently (although what it means to lead and to lag changes from one election to another). As we explore in chapter 8, the “early adopters” of high POE in the mid-1990s influence another group of regions to back the Kremlin with high POE in the 1999–2000 elections. Even as the POE levels decline between the 2007–2008 and 2011–12 cycles (figure 7.1 above), the intra-regional correlations between these elections remain over .90. As Page 157 →POE levels moved up from 2000–2008 then down in 2011 and 2012, the regions maintained their rank order to a high degree. Fig. 7.2. POE Correlated with POE from Prior Federal Elections. POE = percent of the region’s eligible voters who voted for the Kremlin’s candidate or party. Deferential Regions For some of our analyses, we are interested in separating the regions where the leadership has strong authoritarian control over its electoral space from those that lack this control. Below a certain level, POE scores are more likely to be influenced by the public’s preferences and less by elite mobilization. The difference between, say, Irkutsk’s 32% of eligible voters for Putin in 2004 and Ulianovsk’s 40% is less indicative than the gap between the Koryak AO’s 60% and Tatarstan’s 68%. To move a region’s POE from an already notable three-fifths of all adults to more than two-thirds implies striking organizational strength. Relatively low scores do not, of course, rule out the regional leadership having employed administrative resources to influence the outcome. It suggests, however, that the leadership either does not seek to or is unable to manufacture higher levels of turnout and pro-Kremlin voting. As scores increase, POE more effectively measures Page 158 →the regional leadership’s ability to provide the Kremlin with favorable electoral outcomes. We refer to this as “deference to the Kremlin.” In chapter 8, we will explore why regions move into the group that provides high POE for the Kremlin. We therefore need to assign a cutoff point to the POE distributions. Because legislative elections have systematically lower values than presidential elections, we use a different cutoff for each: 50% POE for presidential elections and 35% for legislative elections. To reach either cutoff requires rather high totals for both turnout and pro-Kremlin votes. To reach the presidential cutoff of 50%, for example, a region with 70% turnout would need 71% proKremlin voting. For one party’s vote in a legislative race (multiparty proportional representation) to reach 35% POE, it might receive 50% of the votes cast with 70% turnout. With these cutoff points, the risk is small that we will label a region as deferential when enthusiasm for the Kremlin’s candidate within that region actually accounts for the high score. Any combination of voting and turnout that produces over 50% of the eligible voters supporting one candidate is highly unlikely in democratic settings. Lyndon Johnson’s landslide victory in 1964, for example, rested on 38% of the eligible voters.

Ronald Reagan was reelected as US president 20 years later in another notably lopsided outcome with votes from 31% of the eligible population. Similarly, even dominant legislative victories in democratic multiparty settings seldom exceed 35% of eligible voters. Margaret Thatcher’s Conservative Party won the 1983 British general election in what was widely treated as a landslide, having received votes from about 31% of the eligible voters. In 2011, the Hungarian Fidesz Party won an outright majority of seats against nine other parties, the largest victory in post-1989 Hungary, and their votes came to 31% of the eligible voters. Our chosen cutoff points, therefore, are conservative, in the sense of being deliberatively high. POE figures above these cutoffs must reflect not only the regional leadership’s capacity to produce those results but their intent to curry favor with federal executive authorities. Table 7.3 shows, for each election, which regions were deferential, where they are located, their contribution to the total votes compiled by the Kremlin’s candidate or party, and their share of Russia’s population and GDP. In 1991, the largest number of deferential regions comes from the North Caucasus and Volga areas. Note that the Volga regions that are deferential are primarily ethnically Russian regions, which will not be the case from 2000 on. Also in 1991, the percent of the Russian population living in the deferential regions is larger than the proportion of the regions that are not deferential. In other words, many relatively populous regions fall into Page 159 →the deferential category. This is not repeated from 1999 on. The 1993 and 1995 legislative elections produce no deferential regions, nor does the first round of the 1996 presidential election. In the second round of the 1996 race, however, five regions present high enough POE results to be deferential. Although the reduction in candidates to two makes higher support for one of the candidates more plausible, the 50% cutoff still constitutes a very high bar. Four of the five are ethnic regions; the other is Moscow. They give Yeltsin almost a tenth of all the votes he receives, even though they have less than 8% of the country’s population. In 1999, only two republics provide deferential support, one in the North Caucasus (Ingushetia’s 59% POE for Fatherland/All-Russia) and one in Siberia (Tyva’s 47% POE for Unity). With Putin’s election in 2000, Ingushetia becomes even more notably deferential, from 59% to 78%—but in favor of Putin. This switch is remarkable, particularly given that the republic’s support for Putin’s Unity Party a few months earlier had been only seven-tenths of one percent! Tyva falls below the threshold of deference in 2000 although it continues to back Putin strongly, and three newly deferential republics appear from the North Caucasus and the Volga areas. By 2003, the number of deferential regions in the Duma election is four times as many as in 1999, and they provide United Russia with 14.5% of its votes even though they have only 7% of the country’s population. The eight deferential regions in 2003 include all four of those that had produced deferential results in 2000, plus one more from each of the North Caucasus and Volga districts as well as two Siberian regions: Tyva is back on the program, along with the Aga-Buryat Autonomous Okrug. The following year’s presidential election sees deferential results for Putin increase to 26 regions. Deferential regions in 2004 are found in each federal district except the Northwestern. The Volga District has six deferential regions, although only Mordova was also deferential in 1991. The Volga District’s representatives on the deferential list in 2004, in other words, are all ethnic regions. The 2007 and 2008 elections show even more prevalent deference to the Kremlin. Two-thirds of United Russia’s votes come from deferential regions, exceeding their 60% of the population. Every federal district produces deferential regions. The 2008 pattern is similar to that of 2004, but with two additional regions and with the share of Medvedev’s vote being much larger, almost two-fifths, than the share for Putin four years earlier. In 2011, the number of deferential regions declines substantially from 2007, just as the number in 2012 is significantly lower than in 2008. In 2011, though, the extent to which the vote contribution exceeds the population in those regions is quite a bit larger than in 2007 (15% vs. 8%). Page 160 → Table 7.3. Deferential Regions, by Election

1991

Deferential Regions, by Federal Administrative Districts 12 of 89 Regions (13%): Central: Tula

Vote Contribution Percent of Population Percent of GDP —

17.8%

—

North Caucasus: Dagestan, Ingushetia, Karachaevo-Cherkassia, Chechnya Volga: Mordova, Perm, Nizhnii Novgorod, Penza, Samara

1993 1995

Urals: Sverdlovsk, Chelyabinsk — —

1996, 1st — Round 1996, 2nd 5 of 89 Regions (6%): Round Central: Moscow

— —

0 0

0 0

—

0

0

9.9%

7.9%

12.4%

1.3%

1.2%

0.1%

5.2%

3.6%

North Caucasus: Ingushetia, Kalmykia Urals: Yamal-Nenetsk AO

1999

Far Eastern: Chukotkskii AO 2 of 88 Regions (2%): North Caucasus: Ingushetia

2000

Siberian: Tyva 4 of 89 Regions (4%): 7.0% North Caucasus: Dagestan, Ingushetia, Karbardino-Balkaria

2003

Volga: Tatarstan 8 of 89 Regions (9%): 14.5% 7.0% North Caucasus: Dagestan, Ingushetia, Karbardino-Balkaria, Chechnya

3.6%

Volga: Mordova, Tatarstan

2004

Siberian: Tyva, Aga-Buryat AO 26 of 89 regions (29%): Central: Orlov

28.8%

21.7%

26.3%

Southern: Adygeya, Kalmykia, Rostov North Caucasus: Dagestan, Ingushetia, Karbardino-Balkaria, Karachaevo-Cherkassia, North Ossetia, Chechnya Volga: Bashkortostan, Mordova, Saratov, Tatarstan, Udmurtia, Komi-Perm AO Urals: Tiumen, Yamal-Nenetsk AO Siberian: Tyva, Taimyr AO, Evensk AO, Kemerovo, Aga-Buryat AO, Sakha/Yakutia

Far Eastern: Koryaksk AO, Chukotsk AO 2007

55 of 85 Regions (65%): 67.6% 59.5% Northwest: Belgorod, Bryansk, Novgorod, Pskov, Komi, Vologda

53.3%

Central: Moscow Oblast, Orlov, Tambov, Voronezh, Kursk, Tula, Lipetsk Southern: Krasnodar, Astrakhan, Rostov, Adygeya, Kalmykia North Caucasus Dagestan, Ingushetia, Karbardino-Balkaria, Karachaevo-Cherkassia, North Ossetia, Chechnya Volga: Bashkortostan, Mariel, Mordova, Tatarstan, Udmurtia, Chuvash, Kirov, Penza, Saratov, Ulianovsk Urals: Sverdlovsk, Tiumen, Kurgan, Khanti-Mansiisk AO, Yamal-Nenets AO, Chelyabinsk, Altai Rep. Siberian: Buryatia, Tyva, Zabaikalsk/Chita, Krasnoyarsk, Ust-Orda AO, Kemerovo, Omsk, Aga-Buryat, Sakha/Yakutia

2008

Far Eastern: Kamchatka, Khabarovsk, Amur, Evrei AO, Chukotsk AO Page 161 →Deferential Regions, by Vote Contribution Percent of Population Percent of GDP Federal Administrative Districts 28 of 83 Regions (34%): 38.8% 31.3% 33.4% Central: Tambov, Lipetsk Southern: Rostov, Belgorod, Komi, Krasnodar North Caucasus: Dagestan, Chechnya, Ingushetia, Karbardino-Balkaria, Karachaevo-Cherkassia, North Ossetia Volga: Bashkortostan, Tatarstan, Mordova, Penza, Mari-El, Kirov, Saratov Urals: Khanti-Mansiisk AO, Yamal-Nenetsk AO Siberian: Omsk, Altai Rep., Tyva, Tiumen, Kemerovo,

2011

Far Eastern: Sakha/Yakutia, Chukotsk AO 24 of 83 Regions (29%): 41.4%

26.3%

27.2%

Northwestern: Komi Central: Tula, Belgorod, Tambov Southern: Krasnodar, Adygeya, Kalmykia North Caucasus: Dagestan, North Ossetia, Chechnya, Ingushetia, Karachaevo-Cherkassia, KarbardinoBalkaria Volga: Tatarstan, Bashkortostan, Mari-El, Saratov, Mordova, Penza Urals: Tiumen, Yamal-Nenets AO Siberian: Kemerovo, Tyva Far Eastern: Chukotsk AO

2012

16 of 83 Regions (19%):

22.6%

16.9%

14.0%

Central: Tambov North Caucasus: Chechnya, Karbardino-Balkaria, Karachaevo-Cherkassia, Ingushetia, North Ossetia, Dagestan Volga: Mordova, Bashkortostan, Tatarstan Urals: Tiumen, Yamal-Nenets AO Siberian: Kemerovo, Tyva Far Eastern: Sakha/Yakutia, Chukotsk AO Note: “Vote contribution” stands for the percentage of all the votes received nationwide by the Kremlin’s party or candidate that came from the deferential regions. The far-right column in table 7.3 shows that, in each election from 1999 on, the deferential regions’ importance to the Kremlin’s control of the electoral space exceeds their importance in the national economy. Their share in the country’s gross domestic product is lower, often substantially lower, than their vote contribution. This would also be true in 1996’s second round except for Moscow’s appearance as a deferential region in that election. The city of Moscow does not produce deferential results in any of the other federal elections. Although the deferential regions are small in population size and economic output, the extraordinary number of pro-Kremlin votes they deliver Page 162 →alters the nationwide outcomes disproportionately. In the 2011 Duma election, for instance, if the 24 deferential regions had produced the same POE levels as the median region, United Russia would have received some six million fewer votes and would have received 48% of the seats in the Duma rather than a majority. (United Russia’s official vote share of 49% earned them 53% of the seats in the Duma because three parties on the ballot failed to pass the threshold.) Although many commentators have referred to the Kremlin’s need to offset poor vote totals in Moscow and St. Petersburg with good returns from other regions, the deferential regions provide enough votes to compensate for many regions. The 24 deferential regions provided United Russia with the same number of votes as the 51 lowest-POE regions, even though the latter had 2.6 times as many eligible voters. In 2012, an even fewer number of deferential regions—16 out of 83 or 19% of the regions, representing 14% of the country’s eligible voters—provided Putin with almost 23% of his total nationwide votes. Putin’s victory in 2012 would have been much less impressive had those regions voted at the level of the median regional POE. The deferential regions gave him as many votes as he received from Moscow City, Moscow Oblast, plus 18 other low POE regions. In his study of voting and turnout data, Zvonovskii (2004, 41) wrote: “In essence, what is occurring today in Russia’s electoral space can be seen as the formation within the RF of two distinct countries. One of them (about 15–20% of the populace), live under firm administrative influence, which guarantees the mobilization of electoral support in favor of the candidates and parties that have secured the support of local leaders. The population of the other retains relative independence in its electoral choice.” By our measure, which as noted is a conservative measure, the percentage of the population living under firm administrative influence at the time Zvonovskii was writing (the 2003 election) was only 7%. Since then, however, a much higher proportion of Russians reside in deferential regions: 60% in the 2007 Duma election, down to 16% in Putin’s 2012 victory.

What Explains the Regional Differences? To explain these regional differences, we begin by replicating the analyses in table 6.3, using POE as the dependent variable. That is, we regress POE on our index of socioeconomic modernization, ethnically nonRussian residents as a percentage of the region’s population, the rate of violent crime, income as a percentage of the previous year, the proportion of deputies Page 163 →in regional parliaments who have a party affiliation,

and our measure of oil and gas production. We omit the vote share received by the CPRF or its candidate that we used in table 6.3 because POE includes as one of its components the vote share of the main rival to the CPRF, and the two must vary together arithmetically. The more votes for Putin, the fewer available for Zyuganov. Table 7.4 shows the results of regressing each election’s POE scores on these variables. For each election, we present the r-squared from the ordinary least squares (OLS) regression as a measure of goodness of fit. They range from .1 in 1991 up to a high of .63 in 2004. The number of cases varies across the election, reflecting the availability of data and, from 2007 on, the elimination of some autonomous okrugs and the resulting decrease in the total subjects of the Federation from 89 to 83. Within each cell, we report for each variable the coefficient and significance level from robust regression, a technique that reduces the impact on the standard errors of extreme outlier cases (Andersen 2008), as well as the standardized coefficient from OLS. We draw attention to variables that reach an absolute value of at least .15 in their standardized coefficient. The measure of regional development has statistically significant and strongly positive coefficients from 1993 through the second round of 1996. In those years, regions with higher proportions of urban educated professionals supported Yeltsin and the pro-Yeltsin parties to a higher degree than communist or nationalist candidates and parties. This matches our expectations based on Yeltsin’s identification with democracy, capitalism, and a more pro-Western foreign policy. As table 7.3 shows, deferential regions, those where the voting is highly controlled, were absent or few during these elections, so the performance of the socioeconomic development measure matches the expectation that arises from the nature of the voters. It is thus striking that this measure’s coefficients are negative from 1999 onward. In fact, in 2007, the coefficient is relatively strongly negative. No longer does the extent of a region’s social modernization matter in its voting returns. Even when it has a small impact, it shows that more developed regions are returning vote totals that are less supportive of the Kremlin. The performance of the regions with high percentages of ethnically non-Russian residents is the most consistently strong influence on the results in table 7.4. This variable’s coefficients are significant in every election except the two rounds of the 1996 election. In 1991 and 1993, ethnic diversity is associated with lower levels of POE. Partly, this reflects the efforts by some highly non-Russian regions to boycott these elections. In 1995 and from 1999 on, however, the proportion of ethnically non-Russian residents strongly predicts support for the party of power or the Kremlin candidate. The strength of its impact during the Putin period reduces most other variables in the model to insignificance. Page 164 → Table 7.4. POE Regressed on Regional Socioeconomic and Political Characteristics (A) 1991–99 1996, 1996, 1991 1993 1995 1999 1st 2nd Index of socioeconomic development

Non-Russians as % of population

Pensioners as a % of the population

Violent crime

.188 1.676 1.076 4.363 (.863) (.000) (.003) (.000) в€’.006 .524 .265 .539 в€’.021 в€’.023 .021 в€’.042 (.660) (.130) (.170) (.232) [в€’.193] [в€’.247] .339 в€’.021 .234 .036 .024 в€’.091 (.310) (.606) (.781) (.642) .093 .132 .001 в€’.064 в€’.065 .003 в€’.009 .037 (.000) (.589) (.162) (.035) [в€’.334] .311 [в€’.194] .288

5.280 (.000) .526 .005 (.928) .003 в€’.148 (.608) в€’.108 .023 (.364) .107

в€’1.809 (.016) в€’.092 .071 (.009) .378 в€’.054 (.727) [в€’.151] в€’.008 (.477) в€’.140

.031 Real income as a % of the previous year — — (.334) .085 −.067 −.001 .068 Oil and natural gas production, logged (.652) (.984) (.141) −.002 −.000 .170

.032 (.631) .005 .172 (.101) .220

в€’.068 (.487) в€’.084 .179 (.246) .158

.072 (.388) в€’.073 в€’.195 (.040) в€’.126

—

в€’.022 в€’.029 в€’.075 в€’.132 в€’.010 (.047) (.017) (.006) (.001) (.705) .238 в€’.148 [в€’.267] [в€’.340] в€’.006

Adjusted r-squared

42.502 (.000) .16

9.305 (.000) .34

4.204 (.244) .25

20.931 (.034) .31

46.840 (.002) .30

9.546 (.283) .14

Number of cases

72

79

76

76

76

86

Party-affiliated regional deputies Constant

(B) 2000–12 2000

2003 2004 2007 2008 2011 .260 в€’1.002в€’1.098 в€’2.073 в€’.185 в€’.097 Index of socioeconomic (.818) (.194) (.326) (.041) (.908) (.956) development в€’.010 в€’.117 в€’.064 [в€’.199] в€’.044 в€’.071 .180 .169 .405 .163 .404 .562 Non-Russians as % of population (.000) (.000) (.000) (.000) (.000) (.000) .608 .704 .755 .604 .670 .746 .303 в€’.137 в€’.154 в€’.086 в€’.377 в€’.034 Pensioners as a % of the (.199) (.394) (.534) (.701) (.312) (.936) population .008 .012 в€’.086 [в€’.208] в€’.147 в€’.047 в€’.013 .007 в€’.020 .026 в€’.022 в€’.014 Violent crime (.302) (.240) (.033) (.007) (.137) (.381) [в€’.263] в€’.116 [в€’.171] в€’.136 в€’.120 в€’.136 .118 в€’.064 .199 .356 в€’.478 .461 Real income as a % of the (.191) (.404) (.067) (.009) (.328) (.146) previous year .120 .014 .081 .032 .117 .039 .007 в€’.081 в€’.227 в€’.162 в€’.410 в€’.206 Oil and natural gas production, (.965) (.352) (.078) (.135) (.028) (.253) logged в€’.047 в€’.098 в€’.101 [в€’.185] [в€’.154] в€’.115 Page 165 →2000 2003 2004 2007 2008 2011 .016 Party-affiliated regional deputies (.667) в€’.006 12.621 Constant (.337) Adjusted r-squared .34 Number of cases 88

в€’.006 (.774) .035 25.952 (.011) .43 87

в€’.055 (.108) в€’.002 21.746 (.135) .63 87

в€’.072 (.011) в€’.036 в€’4.137 (.814) .56 82

в€’.036 (.436) в€’.043 в€’2.020 (.971) .51 82

2012 в€’2.106 (.038) в€’.100 .229 (.000) .647 в€’.133 (.574) в€’.122 в€’.023 (.023) .013 в€’.007 (.970) .063 в€’.099 (.350) в€’.108 2012

в€’.070 в€’.030 (.143) (.289) в€’.049 [в€’.152] в€’25.68142.848 (.504) (.047) .58 .57 82 82

Source: See appendix 1. Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Positive beta weights of .15 or over appear in bold. Negative beta weights of в€’.15 or lower appear in bold and in square brackets.

We showed in table 7.2 above that the POE measures have a strong negative correlation with Petrov and Titkov’s expert ratings of regional democracy. As a check on the robustness of our multivariate model, we regressed Petrov and Titkov’s two ratings, pertaining to different times, on the same variables as those shown in table 7.4. We reversed the Petrov and Titkov scores so that higher scores indicate less democracy. In table 7.5, we present the results alongside those for POE in 1999–2003. The patterns are remarkably similar. The proportion of the variance explained is similar, regions’ proportions of ethnically non-Russian residents have the most powerful impact, regional socioeconomic development is an insignificant factor, and the coefficients for the other variables tend to be similar whether democracy rating or actual election results serve as the dependent variable. The kind of dominance of the electoral space indicated by a high POE score is by definition an impairment of democracy. The POE patterns we find, then, do provide a mapping of the decline of regional democracy from the 1990s to the 2000s. Of equal interest is that they indicate a strong regional leadership using its power to provide support for the federal authorities. To clarify how the coefficients for table 7.4’s key variables change over time, we present the standardized coefficients for those variables in a line chart in figure 7.3. The solid black horizontal lines at .2 and в€’.2 indicate the bands we use to suggest a strong substantive impact of that variable on POE levels in that election. The difference between the 1993–96 period Page 166 →and the period from 2000 on is clear, as is the visually striking rise from 1999 on in the extent to which POE levels correlate with a region’s percentage of nonRussian residents (the bivariate correlation reaches .79 in 2011). After the mid-1990s, the influence of a region’s level of socioeconomic development plummets and remains insignificant. Do the remarkably powerful effects of percent non-Russian reflect just a few outliers? The republics in the North Caucasus regions provide the three most non-Russian populaces in the 1989, 2002, and 2010 censuses (Ingushetia, Chechnya, and Dagestan, all above 95% non-Russian). All six North Caucasus republics are in the top 13 in 1989, the top eight in Page 167 →2002, and the top nine in 2010. Perhaps the impact of percent non-Russian derives actually from the distinctiveness of these six republics? This turns out not to be the case. When excluding these six cases, the standardized coefficients for percent non-Russian remain strongly negative in 1991, and strongly positive in 1995 and from 1999 on. Table 7.5. Petrov and Titkov Ratings Regressed on Regional Socioeconomic and Political Characteristics Inverse of POE Democracy POE 2000 Inverse of Democracy Rating, 1999–2002 POE 2003 1999 Rating Index of в€’1.809 в€’.134 .260 в€’.110 в€’1.002 socioeconomic (.016) (.358) (.818) (.483) (.194) development в€’.092 в€’.097 в€’.010 в€’.074 в€’.117 Non-Russians .071 .032 .180 .039 .169 as % of (.009) (.000) (.000) (.000) (.000) population .378 .643 .608 .740 .704 Pensioners as в€’.054 в€’.030 .303 .019 в€’.137 a % of the (.727) (.331) (.199) (.554) (.394) population [в€’.151]в€’.065 .008 .096 .012 в€’.008 в€’.003 в€’.013 в€’.003 .007 Violent crime (.477) (.219) (.302) (.121) (.240) в€’.140 в€’.125 [в€’.263] [в€’.158] в€’.116 Real income .072 .012 .118 .002 в€’.064 as a % of the (.388) (.461) (.191) (.886) (.404) previous year в€’.073 .076 .120 .018 .014

Oil and natural в€’.195 gas (.040) production, в€’.126 logged Party-affiliated в€’.010 regional (.705) deputies в€’.006 9.546 Constant (.283) Adjusted rsquared Number of cases

в€’.008 (.663) в€’.071

.007 (.965) в€’.047

в€’.025 (.237) в€’.114

в€’.081 (.352) в€’.098

в€’.002 (.744) .007 2.843 (.107)

.016 (.667) в€’.006 12.621 (.337)

.001 (.897) .001 1.802 (.323)

в€’.006 (.774) .035 25.952 (.011)

.14

.41

.34

.41

.43

86

86

88

88

87

Source: See appendix 1. Note: Each cell lists the coefficient from robust regression with its significance probability and, below that, the standardized coefficient, or beta weight. Positive beta weights of .15 or over appear in bold. Negative beta weights of в€’.15 or lower appear in bold and in square brackets. Fig. 7.3. Trends in Standardized Coefficients of Key Variables, 1991–2012. (Source: See table 7.4.) It seems clear, therefore, that the boundary between Zvonovskii’s (2004) “two Russias” relates to regional ethnic homogeneity or heterogeneity. We need to probe this distinction in more detail. We first confirm, in table 7.6, that the difference between primarily ethnically Russian oblasts, krais, and cities, on the one hand, and the republics and AOs, on the other, is substantial. They are significantly different in every one of the elections, in fact. The ethnic regions have lower levels of POE in 1991 and 1993, indicated by the positive coefficients. From 1995 forward, however, the average POE level among the republics and AOs is significantly and strongly higher than among the nonethnic regions. The t-test statistics are notably larger from 2003 on. They do not decline in 2011 and 2012 even though the absolute support levels for United Russia and Putin declined from previous elections. Page 168 →We therefore need to ask whether some of the patterns shown in table 7.4 and figure 7.3 are driven primarily by the distinction between the two types of regions included in the full sample. It turns out that this is only partly the case. We ran the same regression model as above on data only from the oblasts, krais, and cities. This reduced the number of cases to 56 or 57, depending on the election. The resulting goodness of fit for the model was higher than those shown in table 7.4 in the 1993–96 elections but quite a bit smaller for other elections. The latter reflects the importance of the percent non-Russian variable in the models in the 2000s. Figure 7.4 plots the standardized coefficients from these new regressions. In comparing the trends to those in the earlier figure, a few things stand out. First, the impact of the index of socioeconomic development is even stronger during the 1990s among this smaller sample of regions than among the entire group. During those years, when elite control of the elections is less, particularly within the nonethnic regions, the socioeconomic structure of a region’s voters matters a great deal. To the extent that Yeltsin and his policies were retaining any support during this period of economic collapse and social instability, it was among populaces that were more likely to support the general reform orientation. As in figure 7.3, however, the impact of socioeconomic development plummets in 1999 and beyond. That trend is not driven by developments in the ethnic regions alone. In addition—quite remarkably—the effect of percent non-Russian Page 169 →remains strong during the Putin era even in a sample that excludes all republics and AOs. Among the latter, the average percent non-Russian in 2010 was 54% (median = 50%). Among the nonethnic regions examined in figure 7.4, by contrast, the average was 9.4% (median=7.5%). The most ethnically diversified of the nonethnic regions was 68% Russian. The

standard deviation for this variable among the ethnic regions is 25.6 while among the nonethnic regions it is only 6.4. Even given the lower levels and narrower variation among the nonethnic regions, the percent non-Russian variable retains some influence in understanding the cross-regional levels of POE. The standardized coefficients for percent non-Russian in figure 7.4 are smaller than for the full set of regions in figure 7.3 but still positive and strong. From 1999 on, then, the presence in a region of sizable non-Russian ethnic groups has influenced its regional politics and its relationship to the Kremlin. Table 7.6. Differences in POE between Ethnic Regions and Other Regions t-test (significance) 1991 2.55 (.012) 1993 1.96 (.053) 1995 в€’4.10 (.000) 1996, 1st round в€’3.48 (.001) 1996, 2nd round в€’2.84 (.006) 1999 в€’3.14 (.002) 2000 в€’3.68 (.000) 2003 в€’6.63 (.000) 2004 в€’7.98 (.000) 2007 в€’6.79 (.000) 2008 в€’6.36 (.000) 2011 в€’6.75 (.000) 2012 в€’6.63 (.000) Note: Results of two-tailed t-test of difference in mean turnout between republics or AOs and all other regions. Negative t-scores indicate that the ethnic regions have a higher average POE than do the nonethnic regions. Fig. 7.4. Trends in Standardized Coefficients of Key Variables, 1991–2012, for Nonethnic Regions Only As the box-and-whisker plots in figure 7.1 showed, the POE distribution for most elections included extreme outliers. The presence of those outliers suggests that the distributions are not statistically normal. Testing for skewness in the distributions is valuable as a way to verify how unusual Page 170 →the patterns are. Also, by comparing the extent to which the distributions are skewed for the ethnic and nonethnic subsamples, we learn more about this key distinction. In table 7.7, we provide a test for the skewness in the distribution of POE values for each election, first for all regions, then for the ethnic regions only, and then for the nonethnic regions only (D’Agostino, Belanger, and D’Agostino 1990; Royston 1991; StataCorp 2011, 2002–6). Low values indicate non-normal distributions, that is, confidence in the ability to reject the hypothesis of normality. We have put those below .05 in bold. The overall distributions are significantly skewed in each election except that of 1991, which confirms what is visually presented in figure 7.1. Less intuitive, however, is that the skewness is absent among the republics and AOs. Even though in most of the elections, certainly from 2004 on, they lead the country in providing high POE voting returns, the entire group is high enough that the regions that are extremely high outliers compared to the entire set of regions are not so among just the ethnic regions. The top POE providers from 1999 to 2003 do stand out from the other ethnic regions, a pattern we will discuss in a later chapter. Among the nonethnic regions, relatively fewer provide high POE levels, meaning that the regions doing so cause the distribution for the nonethnic regions to depart from normality. Spatial Patterning of POE Levels

The manner in which high and low POE-producing regions have been distributed across Russia also deserves attention. Average POE levels by federalPage 171 → administrative district help illuminate this. It should be noted that the spatial pattern we find is only partially related to the strong relationship between POE and ethnic regions. Seven of Russia’s eight federal districts contain two or more ethnic regions; the Central District being the exception. Only in the North Caucasus District do ethnic regions predominate. In the others, they form a minority of the regions. Table 7.7. Normality of POE Distributions: Overall, for Ethnic Regions and for Nonethnic Regions (A) 1991–99 1996, 1996, Type of Region 1991 1993 1995 1999 1st rd. 2nd rd. Overall .400 .006 .000 .001 Oblasts, Krais, Cities .104 .000 .009 .000 Republics and AOs .840 .072 .003 .623

.011 .108 .328

.000 .080 .000

(B) 2000–2012 Type of Region 2000 2003 2004 2007 2008 2011 2012 Overall .000 .000 .000 .000 .000 .000 .000 Oblasts, Krais, Cities .411 .000 .023 .000 .002 .000 .000 Republics and AOs .006 .002 .105 .247 .679 .283 .119 Table 7.8 lists, for each election, the districts with the highest and lowest average POE scores, the average across the eight districts of the country, and a test of whether the differences among the district averages are statistically significant. In 1991, for example, the highest average POE is found among the regions in the North Caucasus (48.5%), and the lowest in the Siberian District (32%), with the eight districts’ averages centering around 40%. In 2012, the district with the highest average is still the North Caucasus, but its level has increased substantially, to 71% as the average of seven regions. Among the patterns in table 7.8 that stand out is the consistency of high POE levels being produced by the North Caucasus regions. Only in the 1993 legislative election and the first round of the 1996 presidential Page 172 →election does any other district exceed it. By contrast, the district with the lowest average varies among five districts during the seven elections from 1991 to 2000. From 2003 to 2012, however, the low POE scores are clustered in the Central and Northwestern districts. The gap between districts with high-scoring and low-scoring regions is 17 percentage points in 1991 and less than 9 percentage points in 1996’s first round. By 2004, however, that gap has jumped to 39 percentage points and remains similarly high through 2012. Deference to the Kremlin is strongly geographically clustered throughout the period but especially so from 2003 to 2012. Table 7.8. Distribution of Average POE Values by Federal Administrative District (A) 1991–99 District 1996, 1996, 1991 1993 1995 1999 (No. of regions) 1st rd. 2nd rd. 48.5 10.5 13.4 29.9 43.4 24.6 Highest district average N. Cauc. NW N. Cauc. Urals N. Cauc. N. Cauc. 31.8 3.3 4.7 21.2 31.8 14.5 Lowest district average Siberian N. Cauc. Far East Central Siberian Urals Mean of district averages 40.8 7.6 6.9 24.8 37.2 18.6 F-statistic from ANOVA 5.24(.000) 5.77(.000) 3.30(.004) 1.28(.268) 3.41(.003) 1.06(.400)

(B) 2000–12 District (No. of regions) Highest district average Lowest district average

2000

2003

2004

2007

2008

2011

2012

53.0 N. Cauc. 32.2 Siberian

41.2 N. Cauc. 18.6 Central

78.6 N. Cauc. 40.3 Central

76.3 N. Cauc. 33.4 NW

72.5 N. Cauc. 42.7 NW

72.6 N. Cauc. 20.5 NW

71.3 N. Cauc. 36.0 NW

24.0 4.97 (.000)

50.3 9.53 (.000)

46.2 8.55 (.000)

50.9 6.17 (.001)

34.5 9.52 (.000)

45.6 7.48 (.000)

Mean of district averages 38.7 7.06 F-statistic from ANOVA (.000)

We can depict how distinctive are the regions in each federal administrative district by testing in which elections a given region’s average POE score is statistically significantly different from the average for all the other regions. Table 7.9 shows the coefficients of t-tests, omitting results when the district’s average is not significantly different from scores elsewhere. It uses brackets to distinguish when the district’s average POE score is Page 173 →significantly lower than the remainder of the regions. (Note that the sign of the t-test is positive when regions in that district have lower scores and negative when they have higher scores.) Three of the eight districts stand out in none of the elections: the Southern, Volga, and Far Eastern districts. The Northwestern District provides the Kremlin with higher than average support in 1993, the 1996 second round, and 2000. It provides significantly lower support than the average from 2007 to 2012. When the regions in the Central District are significantly distinct, it is because their average POE scores are lower than elsewhere. This extends across the change from Yeltsin to Putin. The North Caucasus regions are distinct in almost every election, although they are significantly lower than other regions in 1993 while being significantly higher in every other race. The regions in the Urals District, including Yeltsin’s home region of Sverdlovsk, provide higher than average support in each of Yeltsin’s three races. The regions of the Siberian District provide lower than average support in three presidential races, 1991, 1996 second round, and 2000. Table 7.9. Distinctiveness of the Average POE of Federal Administrative Districts (A) 1991–99 District 1991 1993 1995 1996, 1 1996, 2 1999 (No. of regions) Northwest (11) в€’3.27 в€’2.01 Central (18) [1.87] [1.98] Southern (6) North Caucasus (7) в€’2.38[4.40] в€’4.61 в€’2.34в€’2.19 Volga (14) Urals (6) в€’2.70 в€’1.99в€’1.84 Siberian (12) [3.85] [2.17] Far Eastern (9) (B) 2000–2012 District 2000 2003 2004 2007 2008 2011 2012 (No. of regions) Northwest (11) в€’2.23 [2.39] [1.79] [2.21] [1.96] Central (18) [2.04] [2.72] [2.40] [2.15] [2.07] Southern (6)

North Caucasus (7) в€’5.27в€’5.36в€’7.49в€’6.53в€’5.66в€’7.50в€’6.36 Volga (14) Urals (6) Siberian (12) Far Eastern (9)

[2.82]

Note: Results of one-tailed t-test of difference in mean turnout between regions in the given federal administrative district and all other regions. Blank cells indicate that the t-test did not reach statistical significance of .05. A negative t-score indicates the given district’s average score is significantly higher than elsewhere. Square brackets indicate districts with scores significantly lower than elsewhere. In the 1990s under Yeltsin, the spatial patterning of deference to the Kremlin was more complicated and labile. Across the three electoral cycles of 2003–2004, 2007–2008, and 2011–12, however, the pattern is consistent: the regions in the North Caucasus are providing the Kremlin with electoral support at levels significantly exceeding other regions, even when the other regions’ average levels of deference are objectively high. The regions in the Central and Northwestern districts, which include the cities of Moscow and St. Petersburg, are resisting the trend toward high POE. They are less extreme on the low end than the North Caucasus republics on the high end, hence the asymmetry in the overall distributions we discussed above. Chapter 8 examines the changing dynamics of these spatial patterns in further detail.

Discussion The potential for Russia’s nominally democratic institutions to produce a functioning and consolidated democratic regime faded in the 1990s. The 1991 presidential race and the 1993 and 1995 parliamentary elections took place with comparatively little manipulation of results from regional leaderships.1 Yeltsin’s strong finish in the 1991 presidential election, strong enough that 12 regions meet our criterion for being deferential, depended very little on support from those deferential regions. The 1993 and 1995 elections, in contrast, produced antiYeltsin control of the legislature so Page 174 →that, even after the use of force in 1993 and the new, strongly presidential constitution, Yeltsin found himself in a situation of divided government. With Yeltsin’s popularity sliding fast, the possibility was very real that Yeltsin would lose the 1996 presidential race to Zyuganov of the Communist Party. For Yeltsin and his supporters, this would mark not just a rotation from one party to another but a defeat for the transformations Yeltsin had initiated and a potential return to Soviet rule. Yeltsin therefore turned to selected regional leaderships and struck deals to gain their support in creating high Yeltsin vote totals in their regions. Particularly striking were the regions Dagestan and Karachaevo-Cherkessia, where a Zyuganov majority in round one became a Yeltsin victory in round two. Regional socioeconomic characteristics continued to matter for Yeltsin’s vote share in both rounds of 1996. It is during this pair of elections, however, when the Kremlin begins constructing federal authority through deals that allow regional leaders to maintain undemocratic power within the regions in exchange for ensuring that the Kremlin receives extraordinary votes in federal elections. The fears in the first half of the 1990s that a communist or extreme nationalist (LDPR) form of authoritarianism would emerge via elections proved unfounded, yet the seeds were sown for capitalist, right-center authoritarianism. Both during his efforts to create a political party to support him in the fall of 1999 and after he became president, Putin turned to many of the same regions that had assisted Yeltsin in 1996 and struck deals to get their backing. From 1999 through 2003, a period when rising oil revenue bolstered state coffers, Putin was able to bring increasing numbers of regions into deferential status. He thus came to dominate Russia’s electoral space, which was the crucial step in solidifying his regime. The result, however, bears little resemblance to how Putin described his plans for a “vertical of power.” Formal rules and institutions play a smaller role than touted. The key new institution, that of presidential representatives to the federal administrative districts, facilitated Kremlin pressure and bargaining but established no new rules or procedures. Moreover, a small number of regions have disproportionate impact in securing Kremlin election victories. In critical elections, including 2012, the Kremlin’s victories are owed to a small

number of regions representing a small proportion of the Russian population and the Russian economy. The Russian political regime is therefore more fragile than it would appear even in periods when Putin’s popularity is high. Were the Kremlin to lose those regions’ support, it would not lead to a simple rotation in power between rival parties, as in a democracy. It would instead threaten the dominant electoral victories that Page 175 →deter elite challenges to the regime. Maintaining the support of regional leaders who are effective at delivering votes is a complex task, as chapter 5 revealed. The vertical dimension of the current regime is vulnerable to rapid decay if multiple complex federal-regional relationships cannot be maintained.

Conclusions In this chapter, we have analyzed regional results from each of 13 federal elections from 1991 to 2012. Our measure of POE, voting for the Kremlin’s preferred candidate or party as a proportion of the eligible voters in each region, provides a useful way to examine how the Kremlin was able to ensure its dominance in Russia’s electoral space over time and in which parts of the country that dominance emerges earlier or later. Regions where the leadership has sufficient political control to deliver high, even extraordinary, vote totals for the Kremlin exist in the 1990s but are much rarer compared to the post-2000 elections. Regional socioeconomic development level corresponds to more support for the Kremlin in the early post-Soviet elections but corresponds thereafter with less deference to the Kremlin. The regions with higher proportions of ethnically non-Russian residents reliably support the Kremlin earlier than others and in a more dominant fashion during the Putin period. An important part of the story of post-Soviet Russia’s political evolution is the transformation of regional politics. As the cross-sectional analyses in this chapter illustrate, the regions in which democratic, or at least politically competitive, politics existed in the early and mid-1990s mostly lose that quality by the mid-2000s. In the next chapter, we analyze the spatial pattern created as more and more—but not all—regions become deferential to the Kremlin.

Page 176 →

Eight The Diffusion of Deference In this chapter we ask, did the regions’ spatial relations, as opposed to just their composite characteristics, shape the spread of regional authoritarianism in Russia, as measured by electoral deference to the Kremlin? And, if so, to what extent does diffusion explain the increase of electoral deference over time? Under Yeltsin, spatial clustering in the distribution of deference is largely limited to the 1991 presidential election—that is, prior to Russia’s transition from authoritarian rule. Under Putin, meanwhile, spatial autocorrelation first emerges in the 2003 Duma elections and continues to characterize Russian federal elections through 2012 (i.e., the last federal election under investigation). To explain this development, we identify the regions that were deferential in 2000. We then examine whether regions in the same “neighborhood” as these deferential regions were more likely to act deferentially in subsequent elections than regions not in their neighborhood. When this pattern of diffusion is examined carefully, it demonstrates how, when democracy regresses, authoritarian enclaves can set a new standard, one to be learned from and followed. In Russia, deferential behavior from elites governing two of the country’s more independent and authoritarian regions sent clear signals about the direction of center-periphery politics in the country, and, as we show, Russia’s more ensconced governors were quick studies. Thus, while previous work (Fox 1994; GarretГіn 1995; Mozaffar and Vengroff 2002; Gibson 2005; Magaloni 2006, ch. 4; Magaloni and Kricheli 2010, 408) has emphasized the importance of authoritarian enclaves in otherwise democratic states or democratic enclaves in otherwisePage 177 → authoritarian states, it may be the opportunistic behavior of regional actors in the middle that determine the country’s political trajectory.

Diffusion of Regional Political Dynamics Since Putin needed to bargain with the regional leaders if he hoped to create a strong national foundation for the current regime, we investigate whether spatial factors influenced its evolution. At the beginning of Putin’s first term in office, many important sources of state power were either de facto in the hands of the regional leaders or being contested between the regions and the center. Putin had a number of resources, including sufficient control of the central political institutions, to make important institutional changes that weakened the governors. Yet regional leaders also had resources. If the Kremlin hoped to neutralize potential rivals, distribute material rewards, and sort out qualified elites, it needed to demonstrate its electoral prowess throughout the country, which in turn required coopting existing regional machines and encouraging the construction of new ones. The bargaining that played out across Russia’s regions was both generalized and specific. Part of Putin’s strategy involved offering all regional leaders certain “deals,” such as increased leeway within their regions in exchange for support for Putin and United Russia. Another part of the strategy involved specific bargaining about levels of federal financial support for each region. Although Putin’s changes to federal laws, political norms, and expectations applied to all regions, those regions did not respond to each at the same speed or to the same extent. Governors had to learn that, to stay in the Kremlin’s good graces, they needed effective electoral machines—that is, organizations “dedicated to mobilizing votes or demobilizing opposition for the purpose of winning elections” (Keiser 2001, 666). As previous chapters have shown, some governors had strong regional machines prior to Putin’s presidency. The others needed time to establish them, if they could.1 Governors also had to learn that they could not influence central politics by supporting the Kremlin’s opponents and that it was going to be more lucrative to exchange pro-Kremlin vote returns for rewards from the center. While the end of gubernatorial elections in 2005 dramatically enhanced the Kremlin’s position, it did not produce immediate conformity. As Ross (2011, 645) notes, governors learned to control the media, courts, and regional electoral commissions at varying paces. A regional leadership’s success at delivering votes for Kremlin candidates resembles, then, Page 178 →policy diffusion: behavior that is initially adopted in one political unit, then another and another. Since regional

administrations are not unitary actors, the success of this policy depends on a variety of factors, including the degree to which the governor’s authority has been consolidated in regions. We consider this below.2 Yet our work is not simply about which regions were more or less successful at implementing a policy of electoral deference. Several analysts (Goode 2007; Reuter and Remington 2009; Gel’man and Ryzhenkov 2011; Sharafutdinova 2013) have already provided compelling explanations, largely complementary to ours, of how Putin co-opted regional leaders and recentralized federal power. None, however, sheds light on where and when regions began to defer to the Kremlin’s strategy, as we seek to do. Substantial research, especially but not exclusively on US states (e.g., Shipan and Volden 2008; Sugiyama 2008; Pacheco 2012), shows that policy innovations tend to “diffuse” from one subnational unit to another. Generally speaking, diffusion indicates that one unit doing something increases the odds of another unit doing so and that multiple units doing it increases those odds further (Elkins and Simmons 2005, 36–38). In general, political scientists have struggled to clearly identify the mechanisms of diffusion, partly because the number of mechanisms is many and the conceptual boundaries between them are few. With this in mind, Elkins and Simmons (2005, 39) distill the multitude of possible diffusion mechanisms into two separate, though potentially complementary categories: adaptation and learning. Diffusion by adaptation occurs when the policy decision of one actor alters the value of the policy for other actors. Commonly, the benefits of adopting (or the costs of not adopting) a policy increase as a direct result of the proportion of other actors who have adopted that policy (Elkins and Simmons 2005, 36–38). In contrast, diffusion based on learning occurs when the behavior of one actor provides information to other actors (like the costs and benefits of such behavior) without significantly altering the conditions in which the other actors operate. Fundamentally, then, learning involves adopting policies that have been successful elsewhere. For example, governors observing other regions’ support for the hegemonic party and competitive authoritarian regime could emulate techniques for establishing an effective regional organization capable of delivering votes or disqualifying challengers (Golosov 2011a, 627). In Elkins and Simmons’s (2005) terms, then, the spread of regional deference may be understood as a product of learning if it is initiated by a Page 179 →small number of influential early adopters, leaving the externalities of other regional actors largely unchanged. Adaptation is the better explanation if the number of deferential regions is high (or has grown significantly), since failing to keep pace with other regions in the production of deferential results could undermine a region’s relationship with the federal executive. But why would one expect the transmission of such lessons to assume a spatial dimension? For Russia at least, we believe the pattern reflects the Putin administration’s recentralization policies. The federal administrative districts created in 2000 (see chapter 2) are important because they delineate official regional “neighborhoods,” institutionalize federal oversight on the basis of these neighborhoods, and facilitate communication within them. Thus, to the extent that Russia’s governors learned about the potential benefits of supporting the Kremlin, one should expect the transmission of this information to have been significantly more likely within these districts’ boundaries. Also, as we argue below, while this institutional reform probably facilitated the diffusion of deferential electoral results, it is unlikely that the reform alone was the cause. For one reason, the presidential representatives’ influence in the regions was fairly limited. Putin’s presidential representatives possessed few formal powers and little to no public support. Instead, the envoys relied heavily on informal power, particularly their direct line to the president, but this power itself was constrained since an envoy could not run to the president every time a regional official ignored him (Orttung 2004, 23). For another, while the creation of federal districts occurred simultaneously for all regions, the spread of electoral deference differed across the regions over time. Indeed, the regions that we identify as its catalysts, KabardinoBalkaria and Tatarstan, behaved deferentially prior to Putin’s decree. From this perspective, then, the spread of regional deference is better understood, at least initially, as the result of learning.

Spatial Analyses of Regional Deference to the Kremlin

While Przeworski and Teune (1970) are widely known for urging researchers to “replace proper names with variables,” recent work has made the case that geographic characteristics and spatial relations among polities matter (e.g., O’Loughlin et al. 1998; Beissinger 2002; Pevehouse 2002; Gleditsch and Ward 2006). Although the study of election results across Russia’s regions has privileged compositional explanations,3 a few studies indicate that spatial relations matter there as well (Clem 2006). White, McAllister,Page 180 → and Yun (2002), for example, investigate whether regional context influenced the 1999–2000 election cycle. The authors identify Far Eastern regions that over- and underperform in their levels of support for the pro-Kremlin party (Unity) and candidate (Putin) relative to what compositional variables alone would predict. Lankina and Getachew (2006) provide evidence that diffusion may have influenced variations in the level of democracy across Russia’s regions. By relying on the regions’ proximity to the most eastern capital of a European Union (EU) state, Helsinki, and the flow of EU aid to Russia’s regions, they find that regions closer to Helsinki received more aid and were more democratic than more distant regions. Meanwhile, Gel’man and Lankina (2008) find that the extent of authoritarian diffusion at the local level is the product of countervailing external forces—with United Russia serving as a domestic agent of authoritarianism and the European Union as an international agent of democracy—as well as regional neighborhood effects. For the latter, regional actors were not only well aware of authoritarian developments in neighboring regions (i.e., the elimination of mayoral elections) but also that a bandwagon effect was less likely in regions more proximate to the EU and receiving EU resources aimed at municipal development. Like these works, our study goes beyond the compositional approach to understanding regional variation in Russian elections by seeking evidence that regional executives use national elections to demonstrate their deference to the Kremlin and that deferential behavior has spread from one geographic unit to another over time. To determine whether deferential election results diffused across Russia’s regions, we employ both standard data analysis and techniques of exploratory spatial data analysis (Anselin 1994), which combine more traditional statistics with maps to test hypotheses regarding spatial patterns. Testing for Spatial Autocorrelation The term for when a set of spatial features, regions, in our case, and their associated data values tend to be clustered together in space or dispersed is spatial autocorrelation (Esri 2014). Spatial autocorrelation can be measured with a statistic known as Moran’s I. We begin by examining the global Moran’s I values for each federal election. Significant values indicate the presence of spatial patterning that requires investigation (Anselin 1996, 115). Thus, we begin by examining whether regional results during Russian federal elections actually reflect significant spatial patterns that merit explanation. Page 181 →When exploring spatial effects, geographic neighborhoods can be defined on the basis of contiguity or distance. Due to the wide variation in the territorial size of Russia’s regions—from approximately 1,100 square kilometers (the city of Moscow) to over three million square kilometers (Sakha)—defining neighbors on the basis of distance is problematic. A distance value large enough to capture Sakha’s immediate neighbors will categorize many regions as Moscow’s neighbors even though they are separated from Moscow by several other regions. On the other hand, smaller distance values would treat Sakha as a spatial island—that is, as though it lacks other regions that share its boundaries—which is clearly not the case. Therefore, instead of defining the regions’ neighborhood on the basis of distance, we define spatial neighbors using the notion of contiguity. The spatial weights matrix used in the analysis to test for spatial randomness is a “first-order queen,” meaning that the neighbors of any given region “A” are defined in relation to those other regions that share a common boundary with “A” in any direction. Under this scenario, Kaliningrad, a Russian exclave with no other Russian regions on its borders, is a spatial island and is excluded from the Moran’s I analysis. The only other region that raises an issue for the contiguity option is Sakhalin, which is in fact an island. However, rather than exclude Sakhalin, we coded it as possessing one neighbor, Khabarovsk Krai, which lies directly across the Strait of Tatary (a little over four and a half miles away at the closest point). Likewise, we added Sakhalin to the list of Khabarovsk’s neighbors. Table 8.1 presents the global Moran’s I statistics for all federal elections from 1991 to 2012. The null hypothesis of spatial randomness in our deference measures can be rejected in 1991 and every federal election

from 2003 on. In 1991, the Moran’s I is significant at the .01 level for a two-tailed test (based on 999 permutations). In 1996 (using the second round), the Moran’s I equals .122 and is significant at only the .10 level. The 1999–2000 election cycle merits particular attention due to the lack of spatial autocorrelation. As a reminder, the 1999 Duma election was a particularly competitive contest with Fatherland/All Russia and Unity winning support from different regional heads. Moreover, the resignation of President Yeltsin altered the dynamics of the 2000 presidential race by making Vladimir Putin acting president and moving up the date for the presidential election from June to March. To increase the likelihood of detecting spatial autocorrelation during the Duma elections, we cast a wider net by measuring deferential behavior for either party rather than for Unity alone. The lack of significance for Moran’s I, then, is quite robust. Meanwhile, as we note below, the 2000 presidential contest witnessedPage 182 → an increase in the number of deferential regions despite occurring just three months after the Duma elections. Perhaps more importantly, the increase in deference emerges in the absence of divided loyalties. As we point out in chapter 7’s analysis of the percent of eligible voters (POE), only two regions yielded levels indicative of deferential support in 1999, and the recipients of that support differed: Ingushetia’s 59% POE went for Fatherland/All-Russia while Tyva’s 47% POE was for Unity. Yet in 2000, Ingushetia becomes even more notably deferential, from 59% to 78%, and notably the increase favors Putin even though the republic’s support for Unity had been less than one percent in 1999. Despite this change and the emergence of three newly deferential regions in 2000, regional support for Putin in this race fails to demonstrate significant spatial autocorrelation. Only in the 2003–2004 cycle does systematic spatial patterning return, reaching significance at the .05 level in 2003 and the .001 level for President Putin’s first reelection in 2004. As table 8.1 indicates, the spatial autocorrelation continues through 2012.4 These results, then, justify further examination of the spatial patterns. Proximity to Deferential Regions Spatial analysis can examine absolute or relative location. An example of the former is Lankina and Getachew’s (2006, 2008) work, with its focus on regions’ distances from the European Union.5 Yet, as noted above, there are problems with a focus on distance given the vastly different sizes of Page 183 →Russia’s regions. Instead, we focus on relative location, that is, on neighboring regions as potentially influential actors. As the literature suggests, similar phenomena may occur in proximate polities because those sharing spatial neighborhoods either possess similar cultural, political, or socioeconomic characteristics or because the ideas are likely to spread first to spatial neighbors. With this in mind, then, one might expect the probability of any region becoming deferential or more deferential to grow as the amount of deference in its neighboring regions increases. To examine this, we define a region’s neighborhood as its federal administrative district, for the reasons explained above. For each presidential election, we calculate the mean deference level by federal district and assign that score to each region in that district. Table 8.1. Global Moran’s I Statistics for Deference Measures from Federal Elections (Empirical Pseudo-Significance Based on 999 Permutations) Federal Election Moran’s I Permuted Mean Permuted Std. Dev. Pseudo p-value Yeltsin 1991 0.194** в€’0.013 0.071 0.009 Yeltsin 1996, 2nd Round 0.122+ в€’0.012 0.065 0.056 Unity-OVR 1999 в€’0.015 в€’0.012 0.052 0.983 Putin 2000 в€’0.032 в€’0.013 0.060 0.739 United Russia 2003 0.140* в€’0.011 0.066 0.040 Putin 2004 0.325*** в€’0.009 0.078 0.001 United Russia 2007 0.203** в€’0.007 0.076 0.005

Medvedev 2008

0.157*

в€’0.013

0.074

0.025

United Russia 2011 Putin 2012

0.295** 0.189*

в€’0.010 в€’0.011

0.076 0.074

0.001 0.012

Note: Excluded elections represent those without deferential regions. +

indicates significance (i.e., presence of univariate spatial autocorrelation) at the .10 level. * indicates significance at the .05 level. ** indicates significance at the .01 level. *** indicates significance at the .001 level. All tests are two-tailed. Our investigation of the spatial dimension of deference examines the bivariate relationship between deference in one election and the previous election’s district average and includes a multivariate analysis incorporating a variety of social and political factors. Given the importance of the federal executive to contemporary Russian politics, we focus our attention primarily on the presidential elections. Still, due to the more dramatic changes between the 1999–2000 and 2003–2004 election cycles, we also discuss those legislative elections to more carefully unpack the timing of the trends uncovered. Additional information on specific deferential regions during other legislative elections can be found in chapter 7, especially table 7.3. For the multivariate analysis, we again use our regional development index to serve as a single measure of each region’s level of socioeconomic development or modernization,6 which among other things captures expectations that regions with more highly educated and urban voters behave differently from others (Colton 2000, 76–88; Clem 2006; Rose, Mishler, and Munro 2006, ch. 6), as do less populated regions since they have higher proportions of residents who are dependent on the state (White, McAllister, and Yun 2002, 147). Given the patterns shown in previous chapters, we also include the percentage of non-Russian residents in each region. Although the persistence of authoritarian politics in Russia’s ethnic regions may reflect cultural effects, several ethnically non-Russian regions also have greater access to resource rents, frequently associated with authoritarian rules. To estimate available resource rents, we create a scale using the standardized values of regional levels of oil and natural gas production, two natural resources that received the most credit for Russia’s economic boom during the Putin era (Hill 2004; McFaul and Stoner-Weiss 2008; Rutland 2008b).7 To control for variation in the regions’ economic situations, we include the percentage of unemployed residents in Page 184 →each region as a percentage of the economically active population, except in 1991 when we use average per capita income because unemployment data are unavailable.8 Since we believe a key factor in the rise and spread of deference is the consolidation of gubernatorial power, we estimate the degree to which governors had consolidated their authority in their regions at the time of each presidential election from 1996 on. To do this, we employ two measures: number of months in office (or tenure) at the time of the presidential election and Golosov’s index of the effective number of gubernatorial candidates competing in the regional executive election preceding the national presidential election in question. Since some Russian governors had held office even prior to the collapse of the Soviet Union, the former measure distinguishes these cases from those who only recently assumed the position. The latter measure, meanwhile, captures the level of competition that elected governors face, with higher values indicative of more competitive elections. Yet, as chapters 3–5 highlight, the electoral mandates of Russian governors varies dramatically over time, with some governors elected prior to the adoption of Russia’s 1993 federal constitution and other regions postponing direct elections for their executives. Udmurtia, for example, did not hold a gubernatorial election until after the 2000 presidential election, while Dagestan had only indirectly elected governors for the period under investigation, 1991–2012. While nothing precluded indirectly elected governors from developing powerful electoral machines for, say, the purposes of regional legislative elections, such machines were not tested in the same way as those utilized for regional gubernatorial elections. To control for the difference between elected and indirectly elected governors, we created a dichotomous variable such that regions receive a value of 1 if the governor in office at the time of the presidential election in question was elected prior to that presidential election, and a 0 otherwise. To receive values for the Golosov index of

effective gubernatorial candidates for all of the regions, we multiply the index by this dichotomous variable so that regions that held elections receive values equal to the Golosov index and those that did not receive scores of 0.9 During the appointment era, then, governors who continue in office after having been elected in the region’s last gubernatorial continue to receive scores equal to the Golosov index during that election on this variable. Meanwhile, those without any electoral mandate during the appointment era (2005–12) receive values of 0, similar to governors holding office in the absence of elections in the pre-1996 period. Aside from controlling for the compositional characteristics conventionallyPage 185 → found in the literature on electoral behavior in Russia’s regions or the larger literature on competitive authoritarianism, we also must consider the degree to which the policy of providing deferential election results to the party or candidate of power has coincided with the consolidation of Putin’s regime across Russia, including United Russia’s entrenchment in regions. To test whether the diffusion of deference reflects the spread of the party of power’s influence in the regions, we include a dummy variable that indicates whether the incumbent governor had formally joined United Russia prior to the March 2004 election.10 Table 8.2 begins with bivariate regression analyses in which the only independent variable is the previous election’s district averages on deference.Page 186 → It also presents models with the additional variables just discussed. We report standardized coefficients (beta weights) along with the significance probabilities. The standardized coefficients indicate the extent to which change in the explanatory variable corresponds to change in the dependent variable. A standardized coefficient of .20, for instance, indicates that a change in the explanatory variable of one standard deviation leads to the dependent variable changing by a fifth of a standard deviation. Table 8.2. Regression of Deference in Presidential Elections on Average Deference among Regions in the same Federal District in the Preceding Election 1996 2000 2004a 2004b 2008 2012 .25 .50 .66 .61 .62 Federal district average in prev. election — (.019) (.000) (.000) (.000) (.000) No. of observations 89 89 89 83 83 .05 .24 .43 .36 .37 Adj. R2 .018 .306 .284 .277 .269 .187 (.880) (.004) (.002) (.003) (.021) (.072) .260 .258 .573 .585 .524 .556 Percent non-Russian (.111) (.029) (.000) (.000) (.000) (.000) .376 .040 .011 в€’.002 в€’.083 в€’.023 Development index (.000) (.660) (.868) (.977) (.341) (.785) в€’.069 в€’.014 в€’.052 в€’.039 в€’.063 в€’.100 Gas-oil production index, logged (.533) (.875) (.430) (.572) (.440) (.193) .265 .137 .055 .032 в€’.002 .125 Percent unemployed (.060) (.250) (.566) (.750) (.982) (.209) [в€’.654в€’.031 [в€’.238[в€’.234[в€’.348[в€’.388 Effective number of gubernatorial candidates (.001)] (.764) (.002)] (.003)] (.008)] (.014)] .588 [в€’.155.010 .015 .404 .477 Governor elected (.003) (.091)] (.888) (.840) (.016) (.030) Tenure [в€’.275.093 в€’.047 в€’.045 в€’.094 [в€’.170 (months) (.013)] (.379) (.527) (.562) (.420) (.255)] в€’.031 Governor joined United Russia by March 2004 — — — — (.651) No. of observations 78 88 88 84 82 83 Federal district average in prev. election

Adj. R2

.31

.39

.66

.65

.53

.59

Source: Percent non-Russian data come from the 1989 census for 1996 and 2000, from the 2002 census for 2004 and 2008, and from the 2010 census for 2012. Data on governors’ affiliation with United Russia are from Reuter 2010. Note: Coefficients are standardized. Numbers in parentheses indicate the variable’s significance probability. Positive beta weights of .15 or over appear in bold. Negative beta weights of в€’.15 or lower appear in bold and in square brackets. For the 2008 and 2012 equations, the logged index of gas and oil production uses 2007 data because regional breakdowns cease being published thereafter. The full models in table 8.2—estimated using ordinary least squares—perform well, explaining almost a third of the variance in 1996 and two-thirds in 2004. The average deference value for a given district at time t1 strongly predicts its member regions’ deference values at t2.11 This result supports our expectation that the probability of any region becoming deferential or more deferential would increase as the amount of deference in its neighboring regions increased. The neighborhood lag receives strong support from 2000 through 2008, dropping to significance at the .10 level in 2012 as support for the Kremlin decreases substantially across the federation. Diffusion may occur because polities are spatially proximate or because polities share common characteristics. In previous chapters, we identified at least one political characteristic that robustly correlates with electoral deference in federal elections: having a high percentage of residents who are ethnically non-Russian.12 Table 8.2, likewise, demonstrates this effect from 2000 on. Among our other compositional variables, the influence of gas and oil production on deference, surprisingly, fails to influence deference levels despite Russia’s growing prominence as an energy supplier and emergence as a competitive authoritarian regime. Two variables in table 8.2, meanwhile, demonstrate period-specific effects: the development index and percent unemployed. Higher values on each are significantly correlated with more deferential election results during the second round of the 1996 presidential election. After this election, however, neither matter. Since higher values on the development index indicate more educated, more urban, and more densely populated regions, the positive relationship between this variable and deference in 1996 may initially be surprising. However, given that the second round of the 1996 presidential campaign pitted the incumbent reformer, President Boris Yeltsin, against the unrepentant Communist Party candidate, Gennadii Zyuganov, the outcome makes sense as more educated, more urban, and more populated regions likely represented Kremlin strongholds. The finding that the regions’ deferential results correlate strongly with higher proportions of unemployed residents, on the other hand, makes sense intuitively as regions with more unemployed residents provide a larger pool of potentiallyPage 187 → malleable voters. The surprise here, perhaps, is the variable’s lack of significance in Putin-era presidential elections. We expected that governors more deeply ensconced in their position would have a greater ability to produce exceptionally high regional vote totals for the Kremlin, and our measures yield interesting results. In 1996, deference was significantly more likely in regions with elected governors; however, this effect could be countered by competition levels since deference was less likely in regions where elected governors faced greater competition. Regardless of whether the governor was elected or not, regions with governors in office longer going into the 1996 election were significantly less deferential. These divergent effects reflect the complicated status of the governors’ mandates, such as differences among Yeltsin-appointed governors, indirectly elected executives in the republics, and directly elected governors in the republics as well as those in nonrepublic regions that were permitted to hold elections prior to June 1996. By the March 2000 presidential election, the differences in electoral timing are less prominent with all of Russia’s regions but two holding direct elections. And, just as that election fails to exhibit spatial autocorrelation, two of the variables capturing gubernatorial embeddedness—the number of effective candidates and tenure—also fail to significantly influence deference levels. In fact, the only variable capturing gubernatorial embeddedness to matter is the dummy indicating whether the governor was elected. In contrast to 1996, deference is higher in regions where governors were unelected, indicated by the negative standardized coefficient of в€’.155, significant at the .10 level for a two-tailed test. As

map 8.2 below indicates, this outcome likely reflects highly deferential results coming out of Dagestan in 2000 since it and Udmurtia were the only two regions that fell into this category at the time. The level of competition in gubernatorial elections matters again in 2004, however. From that election on, regions where governors experienced lower levels of competition in their most recent contest were significantly more likely to provide the Kremlin with support indicative of deferential results. The coefficient is consistently negative and significant. In 2008 and 2012, the level of competition in a region’s last gubernatorial election not only continues to matter, but so does whether the governor in office at the time of the presidential election in question was still the victor of that election. That is, during the appointment era, regions with previously elected governors still in office (i.e., receiving the Kremlin’s permission to continue to govern their regions) also exhibited significantly higher levels of deference than those regions where their governors had been Page 188 →replaced. An arguably more interesting, if tenuous finding is the modest effect of tenure during the 2012 presidential election. When controlling for whether the governors in office in 2012 had been elected (i.e., prior to the appointment era) as well as the level of competition that these incumbents had faced when elected, longer tenure appears associated with less deference. While this result does not meet standard levels of significance, the standardized coefficient of в€’.17 may suggest that longer serving governors who had not been elected were less able to produce pro-Kremlin electoral results in 2012 than their previously elected counterparts. Future research may wish to explore this further. Did it change things if a region’s governor had been an early backer of Putin, as indicated by joining United Russia prior to the March 2004 presidential election? Since data are missing for four regions and because we wish to present models as similar as possible across all the elections, we test for the effect of this variable in a separate equation, labeled 2004b in the table. As table 8.2 reveals, gubernatorial affiliation with United Russia fails to significantly influence regional levels of electoral deference in 2004.13 In sum, the multivariate regression analysis presents four key findings. First, deferential election results were much more likely to emerge in regions with larger non-Russian nationalities. Although expected, this result was not guaranteed. Putin made it a top priority to rein in regional politicians, and the governors of Russia’s ethnic regions could have responded with defiance. That they fell in line so readily illustrates their confidence in navigating the waters of Putin’s electoral authoritarian regime. After all, since many of the country’s longest serving governors were in its ethnic regions, many already had effective electoral machines in place. Mobilizing these machines on the Kremlin’s behalf, then, was a low-cost survival strategy relative to openly defying the Kremlin. Second, while resource wealth may have facilitated the consolidation of Russia’s electoral authoritarian regime, it has not determined the emergence of deferential election results at the regional level in Russia. Election results in regions with more access to resource rents, like Russia’s gas-rich regions, were not demonstratively more deferential. Third, the effect of gubernatorial election characteristics on deference levels seems to vary in conjunction with the presence of spatial autocorrelation. Such characteristics matter in 1996, 2004, 2008, and 2012, but not in 2000 (see table 8.1). More tellingly, when the characteristics did matter, regions that experienced less competitive gubernatorial elections in the contest preceding a presidential election were significantly more likely to deliver results that were deferential to the Kremlin. More notable, Page 189 →perhaps, this effect continued into the appointment era, suggesting that the Kremlin appears to have kept around governors with demonstrably greater electoral prowess, ostensibly for its own advantage. Finally, even after we control for economic and social characteristics often associated with determining elections, proximity to more deferential regions made deference to the Kremlin more likely. This finding emphasizes the degree to which the behavior of regional politicians influences one another. The political fortunes of regional politicians regularly depend on the policies of national politicians. Yet regional decisions about how much to support or oppose such policies are not made independently. How some regional politicians respond may depend on their perceptions of how other, possibly more influential and more prominent, regional politicians behave. Our

work suggests that this interdependence holds for even one of the most important of political questions, the direction of a country’s political regime. Mechanisms of Transmission Bearing in mind the distinction between adaptation and learning, one can combine knowledge of the particular regions with tools like standard deviations maps to assess whether adaptation or learning makes the most sense. For adaptation to prevail, the probability of any region becoming deferential or more deferential should grow as the proportion of already deferential regions increase. On the other hand, for learning to predominate, the probability of any region becoming deferential or more deferential should increase as deferential behavior by other regions provides information that such behavior can lower costs or increase benefits in a region’s relations with the Kremlin. To provide a visual representation of deference’s ebbs and flows, we map the distribution of deference toward the Kremlin in presidential elections over time. Maps 8.1–8.6 are standard deviation maps created in GeoDaв„ў.14 Next to each map, we present the mean value of deference for the presidential election in question and list the regions falling within one of the three standard deviational units above the mean. In Russia’s early presidential elections (1991, 1996, 2000), the average percentage of eligible votes going to the Kremlin’s candidate stayed significantly below our cutoff, ranging from 36% to 40%. In fact, a region that scored more than two standard deviations above the mean in 1996, with 47% of eligible votes going to Yeltsin, would have fallen below the regional means from 2004 on. Maps 8.1–8.8 confirm that deference to the Kremlin is not merely a Page 190 →phenomenon of the Putin years. Both the mean value and the number of regions lying two and three standard deviations above the mean was higher in Yeltsin’s first election in 1991 than in Putin’s in 2000. Note as well that although the mean value of our deference measure in 1996 falls below the 2000 mean, the number of extreme outliers—those regions lying more than two deviations above the mean—is higher in 1996 than in 2000. Still, maps 8.1–8.3 reveal diminishing levels of deference across the federation. The number of highly deferential regions drops consistently from twelve to six to two. This trend conforms to conventional assessments of the post-Soviet period, in which elections are seen as more or less competitive and suggests that, at least during the Yeltsin era, acts of deference were not contagious. The insignificance of the global Moran’s I also supports this conclusion. One noteworthy exception to the downward trend between 1991 and 2000, however, is the mean value of our deference measure. While it Page 191 →drops in 1996 relative to 1991, the deference mean in 2000 is higher than it was in 1996 despite fewer regions proving deferential in 2000. The 2000 election appears to represent a critical turning point. Yet fewer regions were deferential in 2000, which means the diffusion observed in 2004 did not result from adaptation, which would expect the policy of deference to spread because more regions (not fewer) had become deferential between 1996 and 2000, thus increasing either the benefits or the costs of not behaving in a similar manner (Elkins and Simmons 2005, 36–38). Map 8.1. Standard Deviation for Deference in the 1991 Presidential Election (Yeltsin) Map 8.2. Standard Deviation for Deference in the 1996 Presidential Election, Second Round (Yeltsin) The adaptation hypothesis, though, remains a viable explanation for the diffusion that occurs between 2004 and 2008. Between 2000 and 2004, the number of deferential regions increases dramatically, from four out of 89 regions to 28 out of 89. Just as important, the cessation of gubernatorial elections grants the Kremlin the ability to hold governors directly accountable for their regions’ performance in national elections. It stands to reason, Page 192 → then, that one would witness an increase in the proportion of regions behaving deferentially during the appointment era. Yet, the median vote for the Kremlin’s presidential candidate across the regions increases by just over one percentage point, and the proportion of deferential regions jumps only slightly, going from 28 out of 89 to 28 out of 83. With these trends in mind, we return to the learning hypothesis and explore the possibility that Russia’s highly deferential regions in 2000 functioned as leaders in the diffusion of deferential election results. Map 8.3. Standard Deviation for Deference (broadly defined) in the 1999 Legislative Elections (Unity or Fatherland/All Russia)

As map 8.4 indicates, the deference leaders in 2000 were Dagestan, Ingushetia, Kabardino-Balkaria, and Tatarstan. Of these, Ingushetia had been highly deferential in 1991 and 1996. In each presidential election, its support for the Kremlin was more than two standard deviations from the mean. Dagestan, meanwhile, was clearly deferential in 1991 but not in 1996. Since both regions had behaved deferentially during the period in which deference to the Kremlin was declining, it is unreasonable to identify either as catalysts of diffusion. Both may have been deferential leaders in the 1990s, but they were leaders without followers. On the other hand, Page 193 →Kabardino-Balkaria and Tatarstan are new additions to the highly deferential camp. As noted in previous chapters, Tatarstan, in particular, stood out among Russia’s regions over the course of the 1990s not only for its lack of competitive politics but also as a source of opposition to the Kremlin. Given the tumultuous relationship between Tatarstan and the Kremlin and the fact that Shaimiev helped to found Fatherland/All Russia, which rivaled the Kremlin in 1999 (Slider 2001), the region’s move from potential rival to Kremlin supporter in a matter of a few months is striking (see maps 8.3 and 8.4). Indeed, other regions throughout the federation likely perceived these suddenly deferential results as a signal that the game of Russian politics would be played differently in the 2000s than it had been in the 1990s. Map 8.4. Standard Deviation for Deference in the 2000 Presidential Election (Putin) These seemingly powerful yet electorally deferential regional leaders also appeared to have had the new administration’s ear. When Putin established a high-level commission to reform local government in 2001, both Kokov from Kabardino-Balkaria and Shaimiev enjoyed seats on the commission’s two key working groups, on “Development of Inter-Budgetary Relations” and “Economic Development” (Lankina 2003). As knowledge Page 194 →of this and similar influence became widespread, other governors likely learned from it. Map 8.5 captures the initial expansion of deferential behavior during the 2003 legislative elections while map 8.6 reveals an explosion of deferential results for the 2004 presidential contest. Given the emergence of a highly popular president actively pursuing reforms that directly weakened the regions’ position vis-Г -vis the center, Russia’s weaker governors appear to have taken a cue from Kabardino-Balkaria and Tatarstan: showing deference to the Kremlin during federal elections represented one potentially effective method for earning an audience with the Kremlin and its increasingly powerful occupant. Map 8.5. Standard Deviation for Deference in the 2003 Legislative Elections (United Russia) Although Russia’s 2012 presidential election represents the first time since 2000 that the number of electorally deferential regions decreased, the evidence indicates more continuity than change. Most notable, perhaps, is that the regional mean for the percentage of eligible voters counted toward the Kremlin’s candidate (49.1%) is on par with those from 2004 (48.3%) and 2008 (49.4%). When contrasted with Putin’s 2000 victory, with a Page 195 →regional mean of 37.8% and deferential support in only four regions, the 2012 results highlight the continued effectiveness of the Kremlin’s electoral machine. Map 8.6. Standard Deviation for Deference in the 2004 Presidential Election (Putin)

Conclusion Using electoral deference to the Kremlin as an indicator of regional authoritarianism, we have used Russia’s federal election results—particularly those from presidential elections—to better understand the dynamics of Page 196 →the country’s authoritarian turn. Russia provides a particularly interesting case of authoritarian revival due to its federal nature and experimentation with electoral democracy. When it was appropriate to label Russia’s political regime a hybrid, it functioned to a large degree as a spatially distributed hybrid of politically competitive and uncompetitive regions. Our analysis not only shows that electoral deference to the Kremlin was more likely in regions with more non-Russians but that its spread over time corresponded with important spatial dimensions. Deference spread at significantly higher rates among regions sharing the same federal district. By Page 197 →the 2004 presidential election, the number of regions providing deferential results more than two standard deviations above the mean jumps from four to seven. More to the point, these regions come from the same federal districts, Caucasus and Volga, as the two regions we identify as deference leaders,

Kabardino-Balkaria and Tatarstan. The diffusion of deference, then, rested initially on regional leaders learning from one another. While deferential behavior in subsequent elections could reflect a combination of adaptation and learning, by 2008 it had become the norm across the regions rather than the exception. Indeed, by the 2012 election, the process of diffusion appeared to have run its course. Map 8.7. Standard Deviation for Deference in the 2008 Presidential Election (Medvedev) Map 8.8. Standard Deviation for Deference in the 2012 Presidential Election (Putin)

Page 198 →

Nine Conclusion In the preceding chapters, we have analyzed the trajectory of Russia’s political regime using subnationallevel patterns. We treat the more than 80 regions of Russia as cases for comparison at multiple times from the end of the Soviet Union to 2012. Doing so allows us to ask questions about the regions’ own trajectories as well as how well the regions bore the weight of economic, political, and social changes at key moments in Russia’s post-Soviet history: its initial transition from communism, the institution of gubernatorial elections, the transition from President Yeltsin to President Putin, and the recent reversal of political rights and civil liberties. Our approach to understanding Russian politics contrasts with three other types of studies: studies at the nationwide level only; those that analyze political dynamics between the federal leadership and “the regions” treated as a single actor; and still others that examine Russia’s regions as separate cases but do so qualitatively or using a method that allows for only one or a few regions to be analyzed. Work of all three types has advanced understanding of Russia’s trajectory, and we cite them throughout the book. We find particular advantages, though, to our quantitative analyses of data from all or most of the regions for a span of two decades. They allow us to take advantage of the diversity among the regions, their different situations at similar moments in time, and the different ways they responded to central stimuli over time in order to depict evolving federal relations. In a sense, our analyses are both subnational and nationwide, providing a valuable way to reveal links between the two levels of analysis. Despite Vladimir Putin’s important role at the top, our work emphasizesPage 199 → the degree to which the regime rests equally importantly on 80-plus regional struts. Each strut is a political relationship involving interdependence between the Kremlin and a regional leadership that is in large measure responsible for ensuring the proper electoral outcomes. As we illustrate, these ties of mutual dependence were forged election by election, beginning while Boris Yeltsin was president but accelerating in the 2000s. Advantageous conditions, including rising prices for oil and other exports, helped the Kremlin ensure that it has the upper hand—that the interdependence is asymmetrical in its favor. Yet formal institutions and laws play little role in defining how the Kremlin and the regions are linked. If recent, less advantageous, conditions continue in the years ahead, the Kremlin could lose its upper hand and risks spreading resistance rather than loyalty.

Findings It will be helpful to recap the findings from chapters 2 through 8. In chapter 2, we reviewed Russia’s postSoviet political history. One reason for a historical overview was to introduce major people and developments necessary to understand the subsequent analyses. These included the three presidents, major political parties and their candidates for president, the institutional rules governing elections, parties and federal-regional relations, and noninstitutional factors such as economic collapse, the course of oil prices, and President Yeltsin’s declining health. We stressed the complexity of Russia’s multiple transitions away from the Soviet Union, which means that institutional choices, economic conditions, elite bargaining, and public behavior all helped determine the course of events and informed our quantitative analyses. The chapter also foreshadowed the results of our later analyses, at the national rather than subnational level. We introduced each federal election and the phases of gubernatorial elections to highlight how election outcomes shaped the ruling regime. We explained the context of the decision to cease gubernatorial elections in favor of Kremlin appointments, which we analyze in chapter 5. We related the trend in Russia’s overall democracy level to trends elsewhere among the postcommunist countries as well as worldwide. Russian democracy, judged to be superior to most of its neighbors in the mid-1990s, becomes worst among all postcommunist countries except those in Central Asia. We end the chapter with data showing that Russia’s level of economic development, that is, its national income, was not low in terms of global correlations between income and democracy. Thus, Russia’s failure to democratize and Page 200 →its

decline into authoritarianism are not easily explainable by its standing vis-Г -vis other countries. Chapter 3 focused on the electoral competitiveness of the regions’ gubernatorial elections during the 1990s. We analyzed both the effective number of candidates in these elections and whether the incumbent governors held on to office. Regions with higher levels of socioeconomic development had more residents likely to have supported calls for democratization prior to 1991. Their presence did not, however, create conditions for more electoral competition within their regions; in fact, our analyses showed that more developed regions experienced less competition, as did those with natural resource wealth. Regions with runoff electoral systems had significantly more competitive races without a clear increase in the likelihood of turnover. Regions where political parties had a stronger foothold made turnover more likely. We found that an interaction effect between constitutional status and election timing shaped competition levels in the 1990s. Republic leaders who chose to hold founding elections early were better able to bring the region’s elites into their machine and give themselves an easier path to reelection while warding off criticism from the center. In chapter 4, we turned to the results of gubernatorial elections under Putin, from 2000 to 2005, when appointment of governors became the rule. To the regional characteristics employed in chapter 3, we added potential explanatory factors connected to attempts by both federal and regional leaders to manipulate electoral outcomes through legal but unfair means. When the Kremlin wanted to oust a sitting governor, it could use its influence to increase the competitiveness of the election but seldom to cause the governor to lose the race. It also had little success in deterring competitors when it supported an incumbent. Thus, not only was the Kremlin unable to protect those incumbents whom it supported, its preference for ousting a sitting governor had, at best, an indirect effect on the election’s ultimate outcome. When the opportunity came along for making a major change, due to the 2004 terrorist attack in Beslan, Putin opted to replace gubernatorial elections with Kremlin appointments, hoping to create a more performance-based administrative system. While Putin’s elimination of gubernatorial elections underscores previous assertions that democracy collapses when political elites decide to disassemble it (e.g., Bermeo 2003, 234), our work finds that the federal executive’s inability to effectively influence regional election results may make such decisions more likely. This finding complements Mainwaring and PГ©rez-LiГ±ГЎn’s (2013, 5–6) point that, while democracy can endure economic, social, and political Page 201 →challenges when powerful political actors are normatively predisposed to it, it becomes endangered when these actors lack a democratic predisposition and begin to see democracy as an obstacle to their personal agendas. We use event-history, or survival, analysis in chapter 5 to assess the appointment system, which lasted until 2012. Although a region’s performance in the economic and social spheres was meant to be a primary criterion for the Kremlin to retain a governor or not, a successful governor also needed to meet the political criteria of preventing unrest and, most critically, delivering large vote totals for Kremlin candidates and parties in federal elections. Our analyses showed that a region’s size, ethnic composition, and economic performance had little effect on a governor’s tenure and that the most important type of performance was political, especially delivering votes in federal elections. The nationwide authoritarian regime remained a coalition of subnational political machines exchanging loyalty to the Kremlin for autonomy and resources. In chapters 6–8, we turned to the federal election results, beginning with voter turnout. Turnout was politically sensitive and varied substantially among the regions in most elections. In the first four federal elections, the highturnout regions are those with more elderly and more sympathy with the Communist Party. In other words, the motivation to vote in most regions reflected individual preferences or party mobilization, rather than elite pressure (what came to be called “administrative resources”) or the manipulation of reported results. From 2003 on, regional age and ideological characteristics have virtually no effect. Instead, the factor most strongly associated with high turnout becomes the percent of a region’s population that is non-Russian. This indicates that leaders of ethnic regions first demonstrated the ability and willingness to provide high turnout. When other regional leaderships became better at promoting higher turnout in the 2000s, the ethnic regions pushed their turnout levels to even higher levels. In chapter 7, we considered the extent to which the votes cast went for the Kremlin’s preferred candidate or

party. Votes for the Kremlin as a proportion of all a region’s eligible voters proved a useful indicator of regional machine control and deference to the Kremlin’s wishes. As with turnout alone, the regional patterns in these voting results change markedly from the Yeltsin to the Putin periods. Regions with higher levels of socioeconomic development had more voters preferring Yeltsin to his communist rival and more preferring to vote against Putin and his authoritarian regime in the 2000s. Our development measure helps explain voting returns in the 1990s but largely ceases to do so from 1999 on. Social conditions and party development likewise have the expected impacts in the 1990s but come Page 202 →to make no difference from 2003 on. As was found for turnout by itself, the proportion of ethnically non-Russian residents predicts support for the Kremlin over the entire period but particularly sharply in the 2000s. Regions in the North Caucasus provide the Kremlin with electoral support at levels significantly exceeding other regions, even when the other regions’ average levels of deference are objectively high. The regions in the Central and Northwestern districts, which include the cities of Moscow and St. Petersburg, resist the trend toward high Kremlin support. We investigated these geographical patterns more directly in chapter 8. We used spatial analysis techniques to examine how, over time, the exceedingly high delivery of votes for the Kremlin spreads from a few regions to a majority. Deferential election results were much more likely in regions with larger non-Russian nationalities. Controlling for that, such election results were not more prevalent in regions with higher resource rents, like Russia’s gas-rich regions. Deference spreads from election to election at significantly higher rates among regions sharing the same federal district. By the 2004 presidential election, the number of regions providing deferential results more than two standard deviations above the mean jumps from four to seven. More to the point, these regions come from the same federal districts, Caucasus and Volga, as the two regions we identify as deference leaders, Kabardino-Balkaria and Tatarstan. The diffusion of deference, then, rested initially on regional leaders learning from one another. While deferential behavior in subsequent elections could reflect a combination of adaptation and learning, by 2008 it had become the norm across the regions rather than the exception. Although far from exhausting all that can be learned through comparing Russia’s subnational units, the foregoing analyses demonstrate how much such comparison adds to the story of Russia’s post-Soviet political evolution.

Subnational Analyses and Comparative Politics Russia’s territorial size and the complexity of the federal system bequeathed to it make subnational dynamics of particular importance for understanding the country’s political change. In addition, the situation during early post-Soviet years, when fundamental institutions and practices of the state, the economy, and the polity were all unclear or in flux, meant that a large measure of what was to be done had to be chosen by those outside the Kremlin. With our subnational focus, we have been able to disaggregate national electoral results not only to illustrate how politically meaningful Page 203 →subunits may deviate significantly from national trends but also to consider the implications of regional changes over time for the national regime’s trajectory. For example, in the early 1990s, when regions were contesting the new system under construction, deviant regions could be found at the high and low ends of our measure of support for the Kremlin’s candidate or party during national elections. However, once the rules of the game were established, we observe outliers only on the upper end of the spectrum. Our approach, then, tackles a critical, yet underappreciated dynamic of contemporary Russian politics: how President Putin—a politician who had never held elected office prior to assuming the Russian presidency—grappled with the legacy of a weak central government, a legacy that was itself bequeathed to Putin by his predecessor, Boris Yeltsin. It also explicitly avoids the bias of viewing a country’s politics solely through the lens of politics in the capital. While one cannot deny the importance of certain powerful, national politicians, those in office change, as they have in Russia. Putin himself moved from being a handpicked successor beholden to Yeltsin’s inner circle to a political force in his own right. Not surprisingly, during his first term in office, Putin largely played by the rules that he inherited from Yeltsin when it came to controlling regional political elites, allowing gubernatorial elections to persist though intervening in conspicuous ways. However, once reelected and confronted with the prospects of identifying his own successor,

Putin dramatically changed the game and eliminated gubernatorial elections in favor of presidential appointees. Placing the fate of the governors in the hands of the president facilitated a presidential succession process that transferred the formal powers of the presidency to Dmitrii Medvedev, as governors from electorally deferential regions were rewarded with longer tenures. With President Medvedev in office and not planning to seek reelection himself, the pace of gubernatorial replacement increased, dislodging some of the country’s most powerful regional barons (see chapter 5). As this account emphasizes, how center-periphery relations are established, maintained, and change are fundamental to the functionality of the national regime, including electoral authoritarian regimes. Unfortunately, the return of authoritarian rule in Russia has sparked more interest in the man at the top than examinations of subnational complexities. Our work combines both national and subnational considerations by demonstrating that the Kremlin’s relations with the regions have hinged on a desire to establish and maintain a perception of electoral invincibility, which, as Magaloni’s (2006, 15) work on Mexico highlights, discourages intra-elite splits. It also explicated the different tactics that Russia’s presidents have Page 204 →deployed in different scenarios over the last 20 years and should therefore help lay the foundation for understanding future developments. The gain from subnational analyses, though, is not limited to providing a finer-grained portrayal of a single country’s experiences. Comparative scholarship can benefit as well. Given the shared histories and cultural similarities that exist across subnational units of the same country, the risks associated with underspecified models for statistical analyses should be reduced relative to cross-national statistical work. Put differently, subnational tests of core theories in the comparative politics literature should be less likely to suffer from the omitted variable bias possible in large-N statistical analysis than cross-national studies, all else equal, since hard-to-measure historical and cultural factors are controlled for by case selection. Although a common critique of single-country studies centers on their limited generalizability, the generalizability of theory also requires that a theory perform well across a battery of tests and under in a variety of settings. Moreover, while the discipline’s main outlets are chocked full of, at times contradictory, cross-national tests, subnational ones are few and far between. Our work, then, takes a step toward filling this gap. For theorists of regime change, in particular, the unfolding of post-Soviet Russian politics ought to provide many lessons. Yet, as discussed in chapter 1, Russia as a single country is hard to score on key variables. It is not a clear-cut case in favor of this or that theory about democratization or democratic failure. Introducing subnational variation allows Russian experiences to inform debates about what causal mechanisms influence regime change. We will now discuss this with regard to this book’s two motivating questions.

The Failure of Democracy in Russia Our first motivating question—why democratic institutions and practices did not prevail in Russia—has received many answers from those focusing on the country as a whole. Yet factors posited to influence Russia’s democratization nationwide actually played out in different ways in different regions. Many distinctions among the regions influenced their levels of political openness and electoral competitiveness. We have shown that, during the 1990s, when Boris Yeltsin led Russia, gubernatorial elections developed into mechanisms of accountability in some regions but, in others, gave voters little opportunity to punish regional rulers for poor economic and social conditions. Rather, executives governing regions with worse Page 205 →conditions—and in the greatest need of electoral accountability—enjoyed greater latitude than their peers in better-performing regions, who faced more challengers and a higher likelihood of turnover. This dynamic harmed Russian democracy in two ways. In the suffering regions, an inability to replace the governor devalued elections in the eyes of the public, setting the stage for mass indifference toward their cessation. In better-off regions, the uncertainty of a swinging electoral pendulum bred elite indifference. As a result, Putin could rely on two sets of allies from different corners of the Russian Federation, voters in some regions and politicians in others, as he moved to rein in governors by eliminating their independent bases of power. Equally important, federal elections, particularly Yeltsin’s campaign for reelection in 1996, taught the

Kremlin the need to garner support from the relatively small (at that time) group of regional autocrats who could control their region’s voting results and produce supermajorities for the Kremlin. Leaders in other regions saw advantages to allying with the center while solidifying their regional bases. This in turn helped reduce elites’ motivations to engage in party building or pursue their ambitions through electoral channels. Were these cross-regional differences influenced by structural factors, such as those associated with modernization theory? That is, do Russia’s more modernized regions have a higher propensity to exhibit great political openness and competition? Although modernization may be understood very broadly, we follow Lipset’s (1959) influential treatment and focus our attention on socioeconomic differences among Russia’s regions. Specifically, we employ a composite measure that combines demographic (e.g., population density, level of urbanity) and educational dimensions alongside the indicators of infrastructural development (e.g., road density, telephone access) and opportunities for leisure (e.g., theater and museum attendance). The level of modernization proves a mixed bag for explaining politics across Russia’s regions, serving as a useful factor in our multivariate models, but only in the early years of the post-Soviet era and only with regard to voting patterns, as opposed to the more complicated process of locking in electoral competitiveness. At the same time, chapter 8’s finding that the country’s northwestern regions resisted the diffusion of deference may reflect their comparatively higher level of modernization. Ethnic composition represents an important demographic characteristic of Russia’s regions which has influenced both regional regime types and regional voting in national contests. Previous research has paid significant attention to this regional characteristic. Understanding the causal impact of this variable, however, has been complicated by the fact that it Page 206 →may capture individual incentives shaping the behavior of politicians and voters alike (i.e., agency) as well as a key structural difference among the regions: more populated regions with large non-Russian minorities existed as autonomous republics during the Soviet era and enjoyed considerably more leverage vis-Г -vis Russia’s emerging federal government during the country’s transition away from communism, with many republics institutionalizing these differences in their constitutions throughout the 1990s. This produced an asymmetrical form of federalism (Lapidus 1999; Ross 2000; Stepan 2000). Our work adds new insight by exploring more directly this nexus of agency and structure. Specifically, we find that structural differences created an opportunity for agency to matter as republics determine the pace of governmental reform in their borders, including the timing of elections to the regional chief executive. Equally important, we find that not all actors reacted uniformly to this opportunity, with early movers in the republics in a better position to lock in noncompetitive politics than politicians elsewhere. Similar to the presence of asymmetrical federalism, electoral rules help to structure politics. Yet, as the literature on neoinstitutionalism emphasizes, these rules are also the product of agency. Like other scholars of regime change, we find that institutional choice matters, at least initially. The use of runoff elections in Russia’s gubernatorial contests increased the effective number of candidates competing in the Yeltsin era, which in turn strongly determined the likelihood for turnover. Yet the effects of institutional choice prove temporary as the impact of runoff elections disappears for Putin-era elections. In these cases, agency—specifically, the actions of the federal executive—prevails.

The Spread of Russian Authoritarianism Our second overarching question concerns the way Vladimir Putin constructed a new authoritarian regime. The regional dynamics that eroded democratic practices in the 1990s helped shape Putin’s strategy in the subsequent decade. Although Putin spoke of strengthening legality and an administrative hierarchy, he followed the path of creating a dominant coalition of regional machines with strong governors supporting the Kremlin. Elections played a key role in this process, even though they ceased, by Putin’s second term, being elections as understood in democratic settings. Elections were vital, first, because they shaped Putin’s ability to rein in regional governors. Putin inherited a federal system built on gubernatorialPage 207 → elections and sought to refashion these elections for his own purposes. Although Putin occasionally employed extra-electoral mechanisms to induce change in the gubernatorial corps, he initially tended to seek influence within the existing parameters: the Kremlin openly advocated for the reelection of some governors and opposed the reelection of others. Yet we

demonstrate that the president’s influence over gubernatorial elections was inconsistent, making the move toward gubernatorial appointees in 2005 an attractive option. Federal elections were also critical. The national regime needed dominant victories in federal legislative and presidential elections. By the time Putin consolidated his rule, seeking reelection in the 2004 presidential election, regional outliers provided results in the Kremlin’s favor that nearly doubled the level of support supplied by such outliers to Yeltsin during the first round of his 1996 reelection campaign (see figure 7.1). To achieve such dominance without a traditional nationwide political party, Putin needed governors to deliver the votes. He employed a variety of resources to incentivize regional leaders to deliver votes in federal elections and to sanction those who would not or could not. Our analyses of successive federal elections track how this process unfolded. We show that regions that led the way in delivering votes in Putin’s favor were those that had been both more independent and more authoritarian during the Yeltsin era. Thus, regions that had been authoritarian enclaves during the Yeltsin regime became models of regional behavior for a Putin regime that prized deferential election results. Other regions were quick to follow this lead, functioning during Putin’s ascendancy like “swing states.” These were regions where competitive politics allowed democracy to sprout roots, but also where democracy’s roots remained shallow and vulnerable to national conditions. As our analysis reveals, regions were more likely to swing in the authoritarian direction, by delivering votes to the Kremlin, the closer they were located to regions that we identified as leaders of the practice. By Putin’s second term, federal elections were widely understood as tests of loyalty where governors were expected to deliver votes to the Kremlin or face removal from office. Still, we find that the geographic composition of Russia’s regime continued to be a composite of regime diversity, only now one with democratic enclaves: economically prosperous regions that have resisted the push to become cogs in the Kremlin’s electoral authoritarian wheel. Thus, after a quarter century of political struggle and change, post-Soviet Russia’s political regime is authoritarian with strong personalist elements. A key factor of this new authoritarianism is the Kremlin’s abilityPage 208 → to conduct formally competitive elections while controlling the outcomes. Authoritarian reliance on elections is not new, of course (Brooker 2013), but recent scholarship provides fresh insights into how incumbents build ruling parties and utilize elections as means of distributing patronage and releasing pent-up popular pressure (see, among others, Magaloni 2006; Brownlee 2007; Gandhi and Lust-Okar 2009; Blaydes 2010). Among the factors that make Russia such an interesting case is the democratic opening that followed the Soviet Union’s collapse. Thus, although elections in countries like Egypt, Iran, and Mexico (under the Institutional Revolutionary Party, or PRI) operated in regimes commonly understood as durably authoritarian, Russia experienced renewed authoritarian elections that occurred on a vastly uneven terrain, thanks, at least in part, to the democratic experimentation of the Yeltsin years. As chapter 3 reveals, in some regions, democratic practices began to establish roots during the 1990s, while in others authoritarianism proved much more resilient. For Hale (2015), fluctuations in Russia’s regime are symptomatic of politics in Eurasia as a whole and should not be seen as simply a post-Soviet manifestation. Rather, periods of political liberalization followed by periods of repression are best understood as artifacts of the regular ebbs and flows associated with changes in the underlying patronage pyramids that define the region. From this perspective, President Putin’s reliance on a power vertical is less an indication of democracy’s failure than the revival of standard practice. Leaving open the possible impact of such longer historical and cultural roots, our work delves into the dynamics by which regional elections in Russia have influenced, and been influenced by, national developments. Like Gibson (2012), we explore political consequences of regime juxtaposition—that is, the tensions between local and national interests when local and national regime types fail to coincide. Yet our work diverges from Gibson’s by considering the conditions that allow subnational authoritarianism to emerge (or, perhaps more accurately, continue) despite a transition widely seen as inaugurating a national democratic trajectory. In fact, the case of Russia demands that one explore the relationship between subnational authoritarian enclaves and a national authoritarian revival. So while Gibson analyzes how those governing subnational authoritarian enclaves insulate their territories against encroachment by a national democratic regime, our analysis of Russia highlights how subnational authoritarian enclaves may contribute to the resurrection of a national authoritarian regime.

Russia’s presidents—like those elsewhere—have sought to maximize Page 209 →votes for themselves, their preferred candidates in times of succession, and a preferred party in national legislative contests. Doing so increases the probability of victory during freer and fairer elections and an image of invincibility during more authoritarian ones. To succeed in this task, we argue that Russia’s leaders have sought support from regional elites and, when possible, sanctioned them. Indeed, the Kremlin’s goal of establishing a vertical hierarchy of power during Putin’s tenure can be understood as seeking to rein in and control political subordinates who possess interests and objectives that do not necessarily always coincide with those of the Kremlin.1 From this perspective, one can draw a parallel between the Kremlin’s reliance on governors to a party’s reliance on local brokers. As Stokes et al. (2013, 76) note, brokers are locally networked and locally powerful individuals who “can leverage this influence over voters to obtain resources from party leaders.” During the 1990s, President Yeltsin’s reliance on governors was rooted in his struggle with Soviet president Mikhail Gorbachev and the consolidation of power after the Soviet Union’s collapse. These considerations shaped, among other things, decisions about when and where to introduce direct gubernatorial elections, as reflected in the willingness to allow republics the freedom to make these decisions on their own and the staggered onset of gubernatorial elections in other regions. At the ballot box, a reliance on regional governors was particularly evident in the 1999 Duma elections with the establishment of Unity, built explicitly in association with select governors as a way to counterbalance another governor-led party, Fatherland-All Russia. During President Putin’s first term, the Kremlin’s efforts include attempts to influence gubernatorial election results, but the ineffectiveness of its efforts within the realm of the democratic process likely contributed to its decision to eliminate these elections entirely. More to the point, as chapter 5 reveals, governors who were willing and able to demonstrate their electoral value during national contests during the appointment era outlasted their comrades. While the parallel between governors and brokers has obvious limits (among other things, a defining characteristic of brokers is their ability to interact, personally, with voters), our work complements Stokes et al.’s point that political machines are not unitary actors. Rather, they consist of individuals who must be monitored and sanctioned if leaders (of a party or a country) wish to avoid unpleasant electoral surprises. To the extent that the ability of Russia’s governors to establish their own regional networks is generalizable, future research may wish to pay more attention to this intermediate level of aggregation and the role that such “meta-brokers” play in Page 210 →determining national election results. Likewise, future research may seek to analyze the links that bind regional networks together, since one should expect subregional variation to emerge, just as we have identified instances of subnational variation. In other words, while we have focused explicitly on relations between the Kremlin and governors, it is worth emphasizing that regional networks can be disaggregated and that the size and effectiveness of these networks hinge on the governors’ relationships with local actors who serve as brokers in a more conventional sense.

How Stable Is Russian Politics? Experts largely were caught by surprise by the unrest in 2011–12; most of the circumstances prompting color revolutions elsewhere were absent in Russia (Stoner-Weiss 2010). Indeed, the demonstrations and other oppositional activities proved short-lived. By the end of 2012, the regime had clamped down, prosecuting enough opposition leaders to disrupt efforts to form a movement. The annexation of Crimea in 2014 and subsequent developments drove Putin’s popularity to record high levels, and the opposition grew yet more splintered. Few analysts, however, see in Putin’s renewed control any evidence that his regime is responding adequately to Russia’s fundamental challenges. Observers therefore disagree about Russia’s prospects for political stability in the years ahead. Pessimists see a conservative leadership and political elite who will bring stagnation to the economy and polity even as changes in Russian society and global trends demand fundamental economic and political reforms (e.g., Trenin et al. 2012). Successful reform would require political liberalization, causing many in the elite to lose power and privilege. Nor will the overall leader push through needed changes. In her afterword to her book-length study of Putin, for instance, Gessen (2013, 292) predicts that the president will never reform Russia’s political system gradually and from within because he is too traumatized by the collapse that Gorbachev’s perestroika

impelled and abhors any perception of weakness. This means, the pessimists believe, we can expect instability and perhaps a dramatic change in the not-too-distant future. For example, Russian political analyst Lilia Shevtsova (2012, 251) reacted to Putin’s reelection in 2012 in this way: Indeed, Putin’s return to the Kremlin means that Russia is starting to repeat the logic of the final Soviet years, characterized by the politicalPage 211 → system’s degradation and a growing gap between the authorities and society. There are no guarantees that the country will manage to avoid fragmentation. So far, the difference between the Soviets’ decline and that of Putin’s Russia is that, unlike the Soviet ruling elite, which had grown too old and flabby to survive, today’s authorities are ready to fight for their power until the bitter end. Similarly, writing in the aftermath of Russia’s 2014 annexation of Crimea, and in the midst of Western sanctions on Russia, Dawisha (2014, 348) warns that economic decline and a weakening of Putin’s stature among the middle class would result in greater coercion as the means to maintain control: “Putin will not go gentle into the night” (ibid., 349). A report commissioned by a foundation associated with former prime minister Aleksei Kudrin (Dmitriev et al. 2015) sees prospects for instability whether or not the fighting in eastern Ukraine drags on. Russian analysts affiliated with the Carnegie Moscow Center, which at that time included Shevtsova, pointed out following the 2011–12 election cycle that an abrupt change might well make things worse: “If the authorities begin to lose control and the country starts to implode before an alternative is developed, the emergence of a dictatorial regime is not to be ruled out. The more hardline elements of the ruling elite might try to save themselves by resorting to raw force and isolating the country internationally. Needless to say, recourse to brute force will only make the exit from the crisis more painful” (Trenin et al. 2012, 18). On the other hand, more optimistic observers (e.g., Pain 2011; Hahn 2012) suggest that, while regime change in the short or medium run might be plausible, it need not produce a tragic collapse; preconditions are in place for positive movement toward democracy. A third set of observers (e.g., Anderson 2010; Gill 2012) view Russia’s regime as likely to be stable even beyond the medium run while the country travels a path toward further economic and social modernization and the political system prevents a return to the chaos and suffering experienced in the 1990s. In other words, Russia’s “authoritarian modernization” (on this term, see Trenin 2010) can produce a better future, as has happened in other countries, notably China. The majority in Russia appreciate the better standard of living they have experienced since the late 1990s, the relative social stability, and the symbolism used by the Putin regime. Discontent among intellectuals and the white-collar class can be managed, this line of thinking goes, until such time as a peaceful transition to full democracy can occur. In this view, if Russia can follow the path of South Korea, how much better for Russians Page 212 →and the rest of the world than some misguided premature switch to Western notions of democracy. Of course, a path of authoritarian modernization will only work if the country modernizes. Many observers give Russia low prospects for this (e.g., Trenin 2010; Lo and Shevtsova 2012; Maslov 2012; Sutela 2012). Certainly, Russia’s economy is less diversified, its technology and infrastructure less advanced, and its integration into world markets lower than South Korea’s or even China’s. As is frequently noted (e.g., Gel’man and Marganiya 2010; Sutela 2012; Trenin et al. 2012), the Russian government’s ability to balance its budget while maintaining high military spending and the social programs to which it is pledged still rests on a narrow revenue foundation. The sharp decline in world oil prices at the end of 2014, combined with borrowing limits imposed as part of sanctions for annexing Crimea, narrowed Russia’s choices (Zubarevich 2015). To overcome the economy’s structural problems, Russian leaders will have to engage in “creative destruction and investment-based growth” (Sutela 2012, 229), despite the risks not just to their public popularity but to elite unity and, therefore, to regime stability. Predicting the political future is a treacherous business, for—as Gel’man (2015, 129–30) observes—world developments are vulnerable to unexpected events and these events may produce a different logic than what one can glean from possible scenarios available in the present. Still, the ability of Russia’s

leaders to address the economy’s structural problems may challenge and perhaps upend Russia’s current federal politics. The sinews of the Putin regime run through Russia’s regions; social and economic modernization, if it is to occur, must run through the regions as well. Despite years of budget surplus and much talk about modernizing life outside a few major cities, vast disparities among the regions remain in terms of wealth, educational opportunities, infrastructural capabilities, and access to investment capital. In 2013, six regions, representing less than 15% of the population, received 35% of the country’s fixed-assets investment funds: Moscow city, Moscow oblast, Krasnodar in which Sochi is located, and three oil-producing regions (Russian Federal State Statistics Service 2014). Those six plus another eight regions—one-sixth of all the regions—received over half of all investment funds. As for foreign investment, Moscow city received 57% of the country’s total in 2013, with runner-up St. Petersburg at less than 8% (ibid.). Moreover, the regional leaderships have only limited ways to generate their own investment capital that could lead to modernization. Even during the years when the federal budget was booming, little was done to Page 213 →promote regional modernization. The shrinking of that budget from 2014 on makes central investment even less likely. Yet demands on the regional leaderships to manage pensions, health care, environmental degradation, and other challenges have grown. Naturally, the Kremlin also expects governors to adeptly manage regional and federal elections and diffuse societal discontent. In this context, the regime’s reliance on informal political interdependencies with each of over four score regions may prove to be one of its weaknesses not the success story it had seemed to be.

Page 214 → Page 215 →

Appendix 1 Data Sources and Construction of Indicators Gubernatorial Election Results Data used to calculate gubernatorial turnover, the effective number of candidates, and turnout in gubernatorial elections come from: Central Electoral Commission of the Russian Federation 1997; McFaul and Petrov 1998; Golosov 1999; Central Electoral Commission of the Russian Federation 2001; Golosov 2008; and Electronic Archives of the Russian Federation’s Central Election Commission. For data collected in the early 2000s, we utilized data from http://www.fci.ru, last accessed on February 17, 2003 (see also Moraski and Reisinger 2003). However, the web address has changed since then. More recent data are available at http://www.vybory.izbirkom.ru. For hard copies of older data printed from http://www.fci.ru, please contact the authors. Missing data after consulting the previous sources come from the archives of the Independent Institute of Elections located at http://www.vibory.ru and a second electronic archive of gubernatorial election data (1991–2000) at http://www.cityline.ru/politika/vybory. The second site was last accessed on December 20, 2005. For hard copies printed from these sites, please contact the authors. Gubernatorial Turnover To calculate turnover, we began with a dichotomous variable where regions score a one if the victor was formally the head of the regional executive Page 216 →(e.g., a Yeltsin-appointed governor, chair of the Supreme Soviet prior to founding elections in most republics, and prime minister in Udmurtia prior to the 2000 gubernatorial election). Turnover, then, is calculated as one minus this score. Effective Number of Candidates The most common measure for the effective number of candidates is the index developed by Laakso and Taagepera 1979): where n is the number of parties (or candidates) with at least one vote and is the square of each party’s (or candidate’s) share of all votes. However, Niemi and Hsieh (2002, 77–78) point out that the conventional measure of the effective number of candidates fails to capture reasonably the number of viable candidates. For example, in a hypothetical single-member-district (SMD) election in which one candidate receives 40%, another 37%, two candidates receive 11%, and one candidate 1%, the effective number of candidates score equals 3.11. In reality, though, the number of viable candidates in this scenario is two, not three. Or, where one candidate receives 70% and six candidates receive 5% each, the effective number equals 1.99. In practice, though, only one candidate is truly viable. Such anomalies in the index for effective number of candidates have led scholars to propose alternatives (e.g., Molinar 1991; Dunleavy and Boucek 2003). Niemi and Hsieh (2002, 81) themselves propose counting the winning candidates (i.e., the first-place candidate in an SMD election) and then every candidate receiving at least 70% of the vote garnered by the last (or only) winner. Applying the 70% rule to Niemi and Hsieh’s examples, the first election yields a value of two and the second a value of one. While Niemi and Hsieh defend the threshold as reasonable and as producing values that more accurately reflect reality, they also acknowledge that 70% is an arbitrary cutoff. Golosov (2010, 182–83), by contrast, proposes an alternative index to the effective number of candidates, which allows one to avoid criticisms about using an arbitrary cutoff. Like Niemi and Hsieh, Golosov begins by counting the largest candidate (or party) as receiving a score of one and then uses the largest party’s vote share to calculate values for the remaining parties. However, for Golosov, smaller parties receive a value equal to the fraction of their vote share relative to what the first-place party receives. The aggregate score for Golosov’s index, then, is the sum of these values. As Golosov (2010, 183) notes, this index is dynamic: “smaller parties’ [or candidates’] scores change not only if their seat- or vote-shares change in

absolute terms, but also if they change in relation to the seat- or vote share Page 217 →of the largest party.” Returning to the anomalies identified by Niemi and Hsieh (2002), Golosov’s index produces a value of 2.5 for the first distribution and a value of 1.43 for the second. Both of these scores almost split the difference between those resulting from the effective number of candidates index (3.11 and 1.99) and Niemi and Hsieh’s 70% rule (2 and 1). In other words, Golosov’s index captures the presence of candidates or parties receiving smaller shares of the vote but does not overweight them. Since we sought to capture overall electoral competitiveness in Russia’s gubernatorial contests, including the presence of smaller, if nonviable candidates, we employed Golosov’s measure.

Type of Gubernatorial Electoral System Elections with runoffs were straightforward from available election data and the corresponding sources. For all elections where no runoffs were held, we consulted the following sources to determine the electoral system: Orenburg Oblast 1995; Central Electoral Commission of the Russian Federation 1997; Orlov Oblast 1997; Central Electoral Commission of the Russian Federation 2001; Golosov 2008; Panorama Informational-Expert Group 2013; Liubarev 2015.

National Turnout Levels These are available on the Russia Votes website of the Centre for the Study of Public Policy, University of Aberdeen: http://www.russiavotes.org

National Election Results, by Region The percent of eligible voters who voted for the Kremlin candidate or party was calculated directly from the raw votes received and the number of eligible voters when those numbers were available (from 1995 on). For the 1991 and 1993 elections, we imputed the data with the formula: vote share Г— turnout/100. Data from the following sources were used to calculate the percentage of eligible voters voting for the Kremlin’s candidate and regional turnout in national contests. Page 218 → 1991 Vote share: McFaul and Petrov 1998, vol. 1, 379–81 and 393–96. Turnout: ibid., 379–81. Vote share: DeBardeleben and Galkin 1997, tables 3.1–3.4, supplemented with data from McFaul and 1993 Petrov 1995, passim. Turnout: McFaul and Petrov 1998, vol. 1, 387–90. Vote share and eligible voters: ibid., 408–11. Turnout: ibid., 393–96. Raw votes: Natsional’naia 1995 Sluzhba Novostei 1995; Central Electoral Commission of the Russian Federation 2008. 1996, Vote share and eligible voters: McFaul and Petrov 1998, vol. 1, 419–22. Turnout: ibid., 408–11. Raw 1st votes: Sharavina, Khramchikhina, and Snegirevoi 1996, passim. round 1996, Vote share and eligible voters: McFaul and Petrov 1998, vol. 1, 422–25. Turnout: ibid., 425–28. Raw 2nd votes: Sharavina, Khramchikhina, and Snegirevoi 1996, passim. round Vote share, eligible voters and raw votes: Golosov 2008. Turnout: Central Electoral Commission of the 1999 Russian Federation 1999. Vote share, eligible voters and raw votes: Golosov 2008. Turnout: Central Electoral Commission of the 2000 Russian Federation 2000. Vote share and turnout: Central Electoral Commission of the Russian Federation 2012. Eligible voters and 2003 raw votes: Golosov 2008. Vote share and turnout: Central Electoral Commission of the Russian Federation 2012. Eligible voters: 2004 Veshniakov and Zastrozhnaia 2004, 133–68. Raw votes: Golosov 2008.

Vote share: Statistics for all regions except Kamchatka and Perm are from Golosov 2008. Because final data for the 2007 Duma election in Kamchatka were unavailable we used preliminary results collected at 7 2007 pm, an hour before the polls officially closed (Regnum News Agency 2007). Electoral data for Perm Krai were obtained from the regional server (Permskii Krai 2007). Turnout, eligible voters and raw votes: Golosov 2008. Vote share, turnout, eligible voters and raw votes: Central Electoral Commission of the Russian Federation 2011 2012.Page 219 → Vote share, turnout, eligible voters and raw votes: Central Electoral Commission of the Russian Federation 2012 2012.

Socioeconomic Development Index To construct this index for each year, we standardized the values of seven variables, then took the average of each region’s standardized value across the seven variables using Stata’s egen, rowmean command. If a region has a missing value on one or more of the variables for the given year, its value on the index is the mean of the other variables. The seven variables are: the percent of the region’s population with at least some higher education; the population density; the percent of the region’s residents living in cities or urban areas; residential telephones per 1,000 urban residents; the density of hard-surfaced roads, in kilometers per 1,000 square kilometers of territory; the number of theater-goers per 1,000 residents; and the number of museum-goers per 1,000 residents. Note that the higher education data are from the 1989 census for 1990–97, from the 2002 census for 1998–2006, and from the 2010 census for 2007 on. Otherwise, each variable is from the same year as the overall index. Data sources for the seven components are below. Percentage of the region’s population with at least some higher education: 1989 census values come from: Orttung 2000, passim. 2002 census values are from: Russian Federal State Statistics Service 2005. Data for 2010 come from: Russian Federal State Statistics Service 2012. Population density measured as people per square kilometer: Regional population at the end of the indicated year (in thousands) is divided by territory in thousands of square kilometers. For cases through 2004, we use territory measured in 1986. With regional mergers in 2005, we move to territory as measured in 2006. Data on regional population in the indicated year come from: Russian Federal State Statistics Service various years. Data on regional territorial size as 1986 of come from: Russian Federal State Statistics Service 1996. Territorial size as of 2006 comes from: Russian Federal State Statistics Service 2007. Page 220 →Percentage of a region’s total population that is urban in the indicated year: Russian Federal State Statistics Service various years. Number of telephones connected to the public network per 1,000 urban residents in the indicated year: Russian Federal State Statistics Service various years. Note that Ingushetia and Chechnya share the same figure in 1990. Density of hard-surfaced roads, in kilometers per 1,000 square kilometers of territory: Russian Federal State Statistics Service various years. Note that this source omits figures for Moscow and St. Petersburg until 2012, so we assigned them high values of from 500 to 850 to reflect that they are entirely urban regions. For Chechnya and Ingushetia, a combined figure is given until 2005, and we assign that figure to both regions. No figures are reported for these two regions from 1992 to 1994, so we extrapolate from the 1990, 1991, 1995, and 1996 figures. For Komi-Perm in 1995, the value is interpolated from the 1990 and 1996 data. For 2003 and 2004, we use its 2002 value. Number of theater-goers per 1,000 people in the indicated year: Russian Federal State Statistics Service various years. Number of museum-goers per 1,000 people in the indicated year: Russian Federal State Statistics Service various years.

Other Measures Real monthly income (average per capita income as a percentage of the previous year for each region):

Russian Federal State Statistics Service various years. Percentage of pensioners in a region’s population 1995, 1999, 2003, and 2006: Ibid. Number of crimes per 100 people age 14 and older, by region, in 1995 and 2000–2006: Ibid. Regional levels of oil production in thousands of tons in the indicated year: Ibid. Note that the figures for regions that contain autonomous okrugs (AOs) exclude the AOs’ amounts. Regions that subsume an AO will therefore see a jump up in the year the AO ceases to exist. Regional levels of natural gas production in millions of cubic meters in the indicated year: Ibid. Note that the figures for regions that contain autonomous okrugs (AOs) exclude the AOs’ amounts. Regions Page 221 →that subsume an AO will therefore see a jump up in the year the AO ceases to exist. Level of registered unemployment as a percentage of the economically active population for the years 1993–2000: Ibid. Percentage of the region’s population that is ethnically Russian in 1989, 2002, and 2010: McFaul, Petrov, and Ryabov 1999, table 2; Russian Federal State Statistics Service 2005, 2012. The percentage of deputies in regional assemblies who are affiliated with a political party from 1994–99 and from 1997–2003: Kynev and Liubarev 2011, 482–88.

Page 222 →

Appendix 2 Cases Excluded from Analyses of Putin-Era Gubernatorial Elections Incumbent governors did not compete in 21 of the 113 gubernatorial elections analyzed for the 2000–2005 period. These cases are listed in chronological order with a brief description of why the incumbent did not run. Accounts come from Radio Free Europe’s Analysis, Newsline and Russian Political Weekly reports (e.g., Radio Free Europe/Radio Liberty 2001, 2003). Besides these cases, we also exclude the October 2003 election in Chechnya, which was internationally denounced as undemocratic, reducing the N to 91. Also note that we include in the analysis Udmurtia’s first gubernatorial election (in October 2000). Since the region had employed a parliamentary-type executive, we treat the prime minister, Aleksandr Volkov, as the incumbent. The excluded elections are: Kursk oblast, October 2000—Rutskoi was a candidate for reelection but was removed from the ballot by regional court, presumably with the Kremlin’s encouragement. Kaluga oblast, November 2000—Sudarenkov did not run for reelection. Ivanovo oblast, December 2000—Tikhomirov did not run for reelection. Kamchatka oblast, December 2000—Biryukov decided not to run and declined to provide an explanation. Page 223 →Krasnodar krai, December 2000—Kondratenko declined to seek reelection, citing poor health. Chukotka autonomous okrug, December 2000—Nazarov withdrew his candidacy among a police investigation and in the context of substantial competition from the oligarch, Roman Abramovich. Evenk autonomous okrug, April 2001—Governor Bokovikov decided not to run, citing an inability to cope with the region’s economic shortfall and the lack of private or public sector support in ensuring the delivery of services. Primorye krai, May 2001—Nazdratenko resigned under pressure from Putin and received a new post in the federal government. Republic of Sakha, December 2001—Nikolaev withdrew his candidacy among legal attempts to remove him as a candidate (under the context that he could not serve a third term) and pressure from the Kremlin. Republic of Ingushetia, April 2002—Aushev resigned from office. It was rumored that the Kremlin coerced the resignation. Aushev denied the rumor. Krasnoyarsk krai, September 2002—Election to replace Lebed, who was killed in a helicopter crash. Taimyr autonomous okrug, January 2003—Election to replace Khloponin, who resigned to run for Lebed’s seat in Krasnoyarsk. St. Petersburg, September 2003—Yakovlev resigned under pressure from Putin and was granted a new post in the federal government. Magadan, November 2003—Election to replace Tsvetkov, who was murdered in Moscow. Kirov oblast’, December 2003—Sergeenkov was legally prohibited from running a third term. Sakhalin oblast, December 2003—Election to replace Farkhutdinov, who was killed in a helicopter crash. Republic of Chechnya, October 2004—Akhmad Kadyrov was assassinated in May 2004. Astrakhan oblast, December 2004—The incumbent, Anatolii Guzhvin, died suddenly in August. Putin named Guzhvin’s first deputy, Aleksandr Zhilkin, acting governor. Zhilkin won the election in the first round. Bryansk oblast, December 2004—A local court struck the name of the incumbent, Lodkin, off the ballot a week before the election. Ulyanovsk oblast, December 2004—The Kremlin recommended Page 224 →that the incumbent, Shamanov, withdraw his candidacy for a second term; Shamanov subsequently left Ulyanovsk to become an aide to then-Prime Minister Mikhail Fradkov. Nenets autonomous okrug, January 2005—A St. Petersburg court cancelled the incumbent’s, Vladimir Butov, registration as a candidate. The formal charges stemmed from a case that had seemed settled, in which Butov was accused of beating a traffic cop.

Page 225 →

Notes Chapter 1 1. Both terms are meant to denote regimes that are neither democracies nor closed authoritarian regimes. Howard and Roessler (2006, 367) suggest that electoral authoritarianism be seen as the larger category, with competitive authoritarianism and what they call hegemonic authoritarian regimes as subtypes. 2. Indeed, some scholars remain unsure of how useful comparisons to countries elsewhere in the world, such as those in Latin America or Asia, can be since even basic political concepts (like voting and parties) can have different meanings in diverse cultures. With this point in mind, Munck (2004, 115) notes the value of constructing “system-specific” indicators rather than “common indicators.” Still, few scholars have the linguistic and financial resources to conduct an in-depth longitudinal study of the domestic politics in multiple countries from different continents. 3. For example, other scholars have investigated the dynamics of regional legislative elections in Russia (e.g., Panov and Ross 2013; Golosov 2014b). We focus on elections and appointments to regional executives because these posts represent the pinnacle of regional power and are the primary representatives of the regions in interactions with the federal executive (see, among others, Hale 2015, 110–14). Moreover, as Gandhi (2008, 186–87) emphasizes, in the absence of democracy, legislative power hinges on the amount of legitimacy and level of policy-making concessions that autocrats (who hold executive office) cede to legislators.

Chapter 2 1. The preeminence of the republic’s legislature was only formal. The 1978 constitution also subordinated the republic to the Soviet constitution, Article 6 of which made clear that the Communist Party of the Soviet Union was in control (Hill 1985, 91–92). Page 226 →2. For more information on Russia’s federal elections, see: Sakwa 1995; Remnick 1997, 317–54; White, Rose, and McAllister 1997; Gel’man and Golosov 1999; Colton 2000; Gel’man, Golosov, and Meleshkina 2002; Tucker 2002; Colton and McFaul 2003; Hesli and Reisinger 2003; Sakwa 2005; Ivanchenko and Liubarev 2007; Buzin and Liubarev 2008; White 2011a. Studies of the major political parties include Urban and Solovei 1997; McFaul and Petrov 1998; Colton and McFaul 2000; March 2002; Hale 2004; Makarenko 2012; Roberts 2012a. 3. In 1993 only, the public voted as well for the members of the upper house, the Federation Council, which has two members for each of the regions. Subsequently, the Federation Council members have been selected through different means. For more on the Federation Council, see Remington 2003a; Chaisty 2006, ch. 4; Turovsky 2007. 4. The system was again changed in February 2014 (Rossiiskaia Federatsiia 2014), a year after President Putin called for a return to the two-track rules for electing members of the Duma (Herszenhorn 2013). In 2016, therefore, 225 seats will be selected through PR party voting and 225 through SMD districts. 5. As we highlight in chapter 3, these decisions led to important variation in the timing of the republics’ first executive elections. 6. Yeltsin also began appointing presidential representatives to each of the federal subjects (Busygina 1997). Their powers and role were never clearly defined, and they had less influence than Putin’s representatives from 2000 on, who oversee a group of subjects. We discuss the latter below. 7. His campaign, however, did dominate the television advertising, utilizing both scare tactics (https://www.youtube.com/watch?v=hAqqJ-uQRZQ) and sex (https://www.youtube.com/watch? v=Noo0lzJILaM). 8. Because the polity2 scores range from в€’10 to +10, we transformed them to the 0–10 scale by adding 10 to each score and dividing by 2. To put the Freedom House scores—which range from 1 to 7, with 1 being most democratic—into the proper scale, we first subtract each score from 7 to invert their order, then multiply by 10/6. With both ratings transformed into the 0–10 scale, we average them for the results

shown in figure 2.1.

Chapter 3 1. While Stepan (2000) expects Russia’s asymmetrical federalism to facilitate authoritarian rule in republics, Matsuzato (2004) argues that variation still exists among Russia’s republics. Some republics, like Tatarstan in the mid-Volga region, emerged as regional leaders while spatially proximate regions differed in their abilities to emulate their neighbors. In the mid-Volga region, for example, Bashkortostan’s leaders emulated the behavior of Tatarstan’s president, Mintimer Shaimiev, more successfully than did leaders in republics like Chuvashia, Mordovia, Marii El, and Udmurtia. We share Matsuzato’s concern with spatial relationships among Russia’s regions and investigate them explicitly in chapters 7 and 8. 2. We employ O’Donnell and Schmitter’s (1986, 6) definition of a transition as “the interval between one political regime and another.” As they note, however, much depends on how one defines regime (ibid., 73). We focus on the national regime. For us, the interval begins with the 1991 collapse of the Soviet Union and Page 227 →ends with the adoption of the December 1993 Russian Constitution. The former begins Russia’s path as a newly independent state, with uncertainty prevailing not just about its economic and political future but also about relations between the center and its constituent regions. The latter, meanwhile, indicates the installation of a new political and legal framework, including the status of subjects of the federation, albeit one that was still subject to modifications (see chapter 2). 3. In all but two cases, Irkutsk in 1997 and Nizhnii Novgorod in 1997, the variable captures instances where the sitting governor competed and either won or lost. In Irkutsk’s 1997 gubernatorial election, the incumbent, Yurii Nozhikov, did not run and did not handpick a successor (Minchenko 2001, 84). We therefore code this case as turnover. In Nizhnii Novgorod, Boris Nemtsov’s departure for Moscow in March 1997 raised succession questions. The two main candidates were Ivan Skliarov, a former Nemtsov deputy serving as mayor of the capital, and Nemtsov’s most recent deputy, Yurii Lebedev. According to Sharafutdinova (2010, 58), Nemtsov did not openly support either. Therefore, we also treat this as a case of turnover. 4. Assuming that individual-level relationships will be mirrored at a higher level is an “individualist fallacy”; see Scheuch 1969; Hannan 1991. 5. More specifically, Moscow and St. Petersburg have values on our index of socioeconomic development that are over four standard deviations above the mean. Thanks to Russia’s particular development path, prior to and during the Soviet period, these two cities concentrate far more of the country’s wealth, educated people, cultural and educational institutions, media, and other facets of socioeconomic development than would be the case in most other countries. 6. While Huntington (1991, 125) calls the transition process in the Soviet Union transformation, Kitschelt et al. (1999, 39) disaggregate the process, observing that some newly independent states experienced negotiated transitions (Armenia, Georgia, and the Baltic states) while others underwent preemptive reform (Russia, Ukraine) or witnessed regime continuity (Azerbaijan, Belarus, and the Central Asian states). By rotating the analytical lens to center-periphery relations, we feel justified applying the moniker “negotiated transition.” Regardless, cases of preemptive reform are expected to be even more conducive to the preservation of authoritarianism than negotiated transitions, so our choice is the more conservative of the two. 7. O’Donnell and Schmitter (1986, 37) define a pact “as an explicit, but not always publicly explicated or justified, agreement among a select set of actors which seeks to define (or, better, to redefine) rules governing the exercise of power on the basis of mutual guarantees for the вЂvital interests’ of those entering into it.” The parade of sovereignties, then, conforms to this definition; see chapter 2. 8. Jones Luong (2002, 63) argues that Soviet policies and institutions not only created but also politicized regional interests in Central Asia—specifically, Kazakhstan, Kyrgyzstan and Uzbekistan—to an extent that they superseded other sociopolitical cleavages (e.g., tribe, religion, and nationality). In her view, this regionalism served as an important source of political stability and continuity with the Soviet system (ibid., 100).

9. We also code Aslan Dzharimov of Adygeya, who had served as the region’s party secretary and chaired the regional parliament, as the incumbent for that region’s 1991 election.Page 228 → 10. All measures and analyses only utilize data from valid gubernatorial elections. For Chechnya, our data treats the 1991 and 1997 election as separate founding elections and exclude the 1995 election. 11. See CEC (1997, 2001). The indeterminate cases are elections in the cities of Moscow (1991) and St. Petersburg (1991), the oblasts of Penza (1993) and Orlov (1993), and the republics of Chechnya (1991), Ingushetia (1993), Kalmykia (1993), Tuva (1992), and Yakutia (1991). Information is also missing for North Ossetia. Since its election was held on January 16, 1994, it also falls, if less neatly, into the transition period. 12. No statistically significant difference emerges across these two samples when it comes to their mean values on turnover. 13. This cutoff, then, removes the republic of North Ossetia. See footnote 10. 14. For the gubernatorial election variables, there are only 19 republics since neither the republic of Dagestan nor the republic of Udmurtia had held popular elections to select the top executive. Udmurtia’s first gubernatorial election occurs in October 2000. 15. For Udmurtia’s first presidential election, we code the sitting prime minister, Volkov, as the incumbent. 16. The republics of Chuvashia and Kalmykia and Amur oblast also held three regional executive elections under Yeltsin; however, in each region one of the elections was declared invalid. Meanwhile, we do not include Chechnya’s 1995 election, which took place during the first Chechen War. 17. Diagnostics of ordinary least squares models revealed heteroskedasticity (evidenced by the BreuschPagan/Cook-Weisberg test), outliers (according to the studentized residuals), and problematic cases in terms of leverage or discrepancy along different independent variables in the model (identified using the Cooks’ D [or dfbeta command in STATA]). 18. Since we lack data on real income in 1994 as well as sufficient data to construct a development index for 1994, we rely on 1995 values for the five elections held in 1994. We also do not have data on oil and gas production in 1997 and 1998; thus, we use 1996 data for elections held in these years. 19. The period analyzed is essentially February 1994 through December 1999 since we lack data on the electoral system governing North Ossetia’s January 1994 election. See footnote 6. 20. Including dummy variables in the analyses presented in tables 3.3 and 3.4 to control for any previous electoral turnover failed to substantively change the results. The overall fit of the models improves, but the variable fails to attain significance. The performance of other variables remains largely the same, though the effect of the oil-gas production index falls below the .10 level of significance in these equations (unreported, but available from the authors). 21. Saikkonen (2016, 447), for example, finds resource wealth to be associated with regional democracy as measured with expert ratings. 22. Given the multicollinearity in the previous model and that the interaction term for minority concentration in republics was not significant in preliminary analyses (unreported), we chose to report models with only percent non-Russian to control for ethnic composition, though it also emerges as insignificant. 23. See Mainwaring and PГ©rez-LiГ±ГЎn (2013) for a recent work arguing that the Page 229 →study of regime change should revolve around political actors, including their normative preferences about democracy, rather than structural conditions. 24. Roeder’s analysis, then, emphasizes the importance of Soviet-era legacies. Our decision to separate republics (and okrugs) from other regions in this chapter’s analysis is one, though admittedly modest, step toward recognizing these legacies. It is also worth noting that our work controls for factors that Roeder identifies as impeding or facilitating political identity hegemony in the republics: these regions’ ethnic composition, economic conditions, and their leaders’ access to lootable resources, gas and oil in particular (Roeder 2007, 152–60).

Chapter 4

1. We also exclude Chechnya’s 2003 presidential election, which was widely viewed as fully controlled. 2. The strategic scheduling of early elections actually preceded Yeltsin’s move. Moses (2002, 907) describes it as the “Belgorod alternative” due to its implementation in that region’s May 1999 gubernatorial election. 3. The terms ranged from four years in the majority of the nonrepublic regions to seven years in the republic of Kalmykia. Several republics, like Tatarstan and Bashkortostan, provided five-year terms for their chief executives, while the republic of Mordovia employed a six-year term. 4. The majority of regions (70%), however, held their elections as scheduled and scored zeros. 5. Again, cases where the governor died in office (Krasnoyarsk), left office for another position (Taimyr), or was externally removed from office (Primorye) are excluded from the analysis (see appendix 2). 6. Like Neto and Cox (1997, 158), however, we recognize that the proximity of elections is a matter of degree: where elections are held concurrently, proximity is maximal. To capture different points along the scale between simultaneous elections and gubernatorial elections held at the midpoint of a presidential term in office, we denote the gubernatorial election by Gt, the date of the preceding presidential election by Pt-1, and the date of the succeeding presidential election by Pt+1. We then calculate the proximity value of the gubernatorial election to the national presidential election using the following formula: Proximity = 2 Г— |(Gt – Pt-1)/(Pt+1 – Pt-1) – ВЅ| This formula expresses the time between the previous presidential election and the gubernatorial election as a fraction of the presidential term. Subtracting ВЅ from this fraction and then taking the absolute value shows how far away from the midterm the gubernatorial election was held. The formula yields a score of zero for gubernatorial elections held at the midpoint of a president’s term indicating that they are the least proximal elections. Meanwhile, the most proximate are those held concurrently with the presidential election and are indicated as such by a score of one. We measured the proximity of each gubernatorial election to the Duma elections similarly. 7. To limit potential collinearity among the social and economic variables, the Page 230 →analyses in this chapter use real income in place of unemployment. To the extent that either mattered during the Yeltsin era, it was the former (see chapter 3). 8. Studentized residuals revealed up to nine outliers (those with an absolute value above two) while an investigation of hat values identified nine high leverage cases. In addition, we examined the Cooks’ D (or dfbeta command in STATA), which assesses leverage and discrepancy on each independent variable in the regression model. Across the variables in the equation, the number of potentially problematic cases ranged from 8 to 23. 9. Since robust regression analysis in STATA does not provide standardized coefficients, those reported are derived from the ordinary least squares (OLS) models. In instances with poor probabilities of significance (i.e., p-values approaching .80), the coefficients from the robust regression analysis and the standardized coefficients may have opposite signs. This occurs twice in table 4.1, both times in Equation 3. Still, we believe the benefit of reporting standardized coefficients drawn from the OLS models outweigh the drawbacks as long as the beta weights are considered as an additional piece of information to be examined alongside the significance levels derived from the robust regression results. 10. As in chapter 3, our measure of party development comes from Kynev and Liubarev (2011, 482–88). For the 1994–99 period, we use their measure of party development as the percentage of deputies in regional assemblies who are affiliated with a political party. For the 1997–2003 period, however, Kynev and Liubarev also provide a measure that captures what they call “genuine party affiliation.” This percentage includes deputies who did not formally declare an affiliation but in practice acted as though they were affiliated with a party. It is notable that little correlation exists between the measures of official party affiliation and “genuine” party affiliation for the 1997–2003 period; the correlation coefficient is only .113. Theoretically, though, one can reasonably expect party-affiliated behavior to drive electoral competitiveness as much or more than running on a party label. Unreported analyses confirm this expectation. The 1997–2003 measure of official party ties not only fails to yield a statistically significant

effect on competitiveness in Putin-era gubernatorial elections but none of its beta weights in the corresponding equations come close to the .15 threshold. Table 4.1, then, reports the better performing measure of party development as a stricter test of the Kremlin’s influence in regional elections. 11. See chapter 3 for fuller descriptions of the other independent variables. 12. As in chapter 3, an interaction term estimating the concentration of minorities in republics performs poorly; instead, both the dichotomous variable republic and percent non-Russian perform better across all specifications (unreported but available from the authors). As a result, we present models with these two components but not the interaction term. We also conducted the analysis with only republics and the results largely resemble those reported. 13. Thanks to differences in how we measure proximity and Putin’s tenure, however, we are able to include them in the same model. The results with Putin’s tenure excluded from the equations are largely the same as with it included (unreported but available from the authors). 14. It is possible that this null result reflects the failure of the equations in table 4.1 to interact presidential support (or opposition) with presidential tenure. That is, the Kremlin may have become more influential over time when it supported challengersPage 231 → than when it supported incumbents (or vice versa), which is something that the equations presented in table 4.1 do not capture. With this in mind, we reanalyzed the data while adding interaction terms (tenure-support turnover and tenure-opposed turnover) to the equations. Yet neither interaction term proved statistically significant, the variable for Kremlin support of turnover remained significant and the other variables in the models remained insignificant (unreported but available from the authors). 15. Of course, appointing governors also might allow the Kremlin the opportunity to expand the base of its support in the regions, which is something we examine in the following chapters. 16. The only aspect of election timing that demonstrates even a substantive effect is the proximity of the gubernatorial elections to national legislative elections. As expected, gubernatorial contests closer to Duma elections tended to witness fewer effective candidates, as evidenced by the standardized coefficient of –.29. 17. The measure for number of months late that a gubernatorial election was held is excluded from these models because zeros for the variable predicts failure (i.e., no turnover) perfectly. 18. None of the other independent variables prove statistically significant even with the effective number of candidates left out of the model (unreported but available from the authors). 19. Indeed, higher income levels, if anything, promoted turnover (see chapter 3, table 3.5, Equation 2). 20. Of course, this option becomes more likely in polities where incumbent rulers are, at best, normatively indifferent to democracy. This point conforms to recent arguments by Mainwaring and PГ©rez-LiГ±ГЎn (2013), although they focus more on organizations and movement than on individual leaders. Specifically, where the most powerful political actors in a country are predisposed to democracy, competitive politics have better prospects even when confronted by major economic, social, and political challenges. On the other hand, one can expect competitive elections to become endangered when powerful actors rank other priorities higher than democratization and see elections as obstacles in their pursuit of these other priorities (ibid., 5–6).

Chapter 5 1. Those in office in January 2005 continued in office until the end of their term, when they needed a presidential nomination to be reappointed. They also had the option of requesting an expression of confidence from President Putin prior to their term expiring. Those who sought and received it began a new four-year term immediately. 2. Future research may wish to consider whether the means of dismissals are correlated with their timing. Turovskii (2010, 61), for example, distinguishes between “gentle” and “harsh” procedures for removing a governor. 3. In Treisman’s (2007, 23–25) terms, the Soviet system was administratively but not politically decentralized. Post-Soviet Russia had “appointment decentralization” before 2005 and from 2012 on. 4. Our information on the names and tenures of the regional leaders comes Page 232 →from Rulers.org

(2013), modified and augmented by media reports from a variety of sources. 5. The onset of the economic crisis falls too close in time to the start of the Medvedev presidency to be able to treat his tenure in office as a time-related factor. Medvedev began his presidency in early May of 2008, and only a few months later the economic downturn began. Although we derive our expectation of more frequent replacements over the last three years from the country’s economic problems, Medvedev occupying the presidency might also be relevant. We cannot, though, disentangle the two. 6. Regions’ population sizes are highly correlated with their gross regional product (Kendall’s taub = .62). We opt to employ population because a few resource-rich regions such as Khanti-Mansiisk rank very high in gross regional product despite their small population size. Other than their resources, little suggests that such regions are highly important to the national leadership. 7. The governors’ birthdates are taken, when possible, from the official website of the governor or the regional executive branch, augmented as necessary from other sources. 8. President Medvedev signed the bill on gubernatorial elections in May 2012 and it came into effect on June 1, 2012 (see Gubernatory.ru 2012). 9. The three governors who died in office are Mikhail Yevdokimov of Altai Krai, Viktor Shershunov of Kostroma, and Igor Yesipovsky of Irkutsk. 10. Valerii Kokov of Kabardino-Balkaria. 11. These governors are Gennady Savel’yev (Komi-Permyak), Boris Zolotarev (Evensk), Oleg Budargin (Taimyr), Valery Maleyev (Ust-Orda), Bair Zhamsuyev (Agin-Buryat), and Oleg Kozhemyako (Koryak). 12. Since we are interested explicitly in whether governors “survive” as governors, as opposed to as politicians within the regime, we do not right-censor governors who received promotions to higher office (cf. Reuter and Robertson 2012, 1028). Promotions have been a means of coaxing governors to give up office without disrupting elite stability, for example, Dagestani leader Magomedov being dismissed as governor but appointed as a presidential aide in January 2013, and some have likely indicated that the Kremlin deemed the politicians as valuable to the regime but poor regional managers. In either case, they reflect a Kremlin judgment not to retain them in their posts. 13. This neglects a governor’s time in office prior to the appointment era. The hazard of being fired by the president only begins in 2005, and that is what we seek to explain. We do, however, use length of time in office prior to 2005 as an explanatory variable below. 14. For cases where data are missing for the governor’s last year in office, we use data from the most recent year available during the governor’s tenure.

Chapter 6 1. This carries over from Soviet-era practices. The rules for voter eligibility in post-Soviet Russia appear in the federal law “On Basic Guarantees of Electoral Rights and the Rights of Citizens of the Russian Federation to Participate in a Referendum”Page 233 → (Rossiiskaia Federatsiia 2002 [2014]). While this law has been amended or replaced several times from 1992 on, the provisions for voter eligibility have remained the same. 2. The same trends are evident when comparing turnout at the level of voting districts, or raions. See Myagkov, Ordeshook, and Shakin 2009, 84–87. 3. This feature of the second round of the 1996 election was quickly identified and has been analyzed as a factor contributing to Yeltsin’s victory. See Kozlov and Oreshkin 1996; Oreshkin 2001. 4. Brady, Verba, and Schlozman stress that participative resources are not connected only to socioeconomic position but to other factors such as memberships. 5. Such logic led an early analysis of voter turnout (Milbrath 1965) to contend that urban areas enjoy higher rates of political participation than rural areas. Empirically, however, other works show that the relationship between urbanity and participation at the subnational level is far from clear-cut (Nie, Powell, and Prewitt 1969; Johnson 1971; Verba and Nie 1972; Monroe 1977; Schulz 1979, 12). 6. As a carryover from Soviet times, the definition of a pensioner in Russia includes people of lower ages than would be the case in Western countries. Russian women retire at 55, and men at 60. Raising the age

continues to be discussed frequently but has remained unchanged through 2012. 7. This dynamic is one reason why we argue that a fuller understanding of Russia’s electoral politics requires an analysis of competition and turnover in Russia’s gubernatorial elections. 8. Indeed, our analysis in chapter 3 suggests that one factor preserving the regional leadership in the republics was the timing of founding elections, since leaders who failed to seize the opportunity to hold executive elections early on in Russia’s transition were significantly more likely to experience gubernatorial turnover in their republics than those who held founding elections prior to 1994. 9. Moscow and St. Petersburg have the highest values by far on our index of socioeconomic development, over four standard deviations above the mean. Because of Russia’s particular development path, prior to and during the Soviet period, these two cities concentrate far more of the country’s wealth, educated people, cultural and educational institutions, media, and other facets of socioeconomic development than would be the case in most other countries. In addition, our measure includes urban residents as a proportion of the regional population, and these “cities of federal status” are 100% urban. Diagnostic tests show that these two outliers do have a strong impact on the role that our socioeconomic development index plays in our analyses. For this reason, we report coefficients not from ordinary least squares regression but from robust regression, which down-weights the impact of extreme outlier cases when calculating coefficients (Andersen 2008). In addition, we have estimated our models while excluding Moscow and St. Petersburg as cases. The direction and approximate strength of other variables in these models generally remains unchanged. Because Moscow and St. Petersburg really do lead the other regions in socioeconomic development, it is important that we keep them in the analyses, outliers though they be. 10. We construct the scale using data from 1995, 2000, 2004, and 2007. The Cronbach alpha scores range from .77 to .79. Page 234 →11. As explained in chapter 4, for elections from 2000 on, we use Kynev and Luibarev’s (2011) measure of the percent of regional parliamentary deputies behaving in office as though they are affiliated with a political party. 12. The Republic of Tyva is a high outlier on this variable. For example, it scores 719 crimes against people per 100,000 population in 2004, versus 492 for the next highest region. However, removing Tyva from the analyses changes the results in only a minor way. 13. Looking at a different level of analysis than we do, Myagkov, Ordeshook, and Shakin (2009, 82–87) show that districts (raions) located in republics have a skewed distribution while those elsewhere are normally distributed. 14. We are not claiming that these elections are free of elite pressure or fraud. We instead are noting that these phenomena do not, in the 1990s, undermine the operation of other influences, while in the 2000s they do. On fraud throughout Russia’s post-Soviet period, see Wilson 2005; Myagkov, Ordeshook, and Shakin 2009.

Chapter 7 1. The referendum on the constitution, which was held concurrently with the 1993 parliamentary election, was likely the subject of fraud but that fraud was coordinated by federal authorities not regional.

Chapter 8 1. On the different forms of political machine across Russia’s regions and major cities, see Gel’man 2013. 2. In this vein, we acknowledge the work of others who have disaggregated electoral manipulation to the local (or raion) level. According to Goodnow et al. (2014), for example, local level concentrations of nonethnic Russians significantly influence electoral manipulation in titular ethnic regions and Russiandominated regions alike, making the practice more decentralized than regional-level analyses depict. This emphasis on disaggregating regional election returns mirrors our own efforts to disaggregate national ones. But just as the Kremlin need not maintain uniform control across the federation to win elections, Russian governors need not establish uniform control within their respective regions to yield deferential election

results. The appropriate level of disaggregation, then, depends on the phenomenon under investigation, and our focus is on the spread of regional deference to the Kremlin across the federation over time rather than the intra-regional mechanism by which these results are attained. 3. For example, the emergence of a Russian “red belt” during the 1990s, identified as regions where the Communist Party of the Russian Federation performed best, was explained by those regions having less educated, poorer, and older voters living in agriculturally based economies (Clem and Craumer 1995). 4. Since our deference measure gives the same value to regions at or below the mean, one might suspect that having many cases with the same values could increase the likelihood of finding the spatial autocorrelation in the data. If anything, our measure deflates the spatial autocorrelation in federal voting patterns across Page 235 →the regions. For example, the Moran’s Is for the percentage of eligible votes going to Putin are .249 in 2000, .297 in 2004, and .238 in 2012. In 2008, the Moran’s I for the percentage of eligible votes going to Medvedev is 0.222. In all cases, the pseudo p-values are significant at the .01 level or higher for two-tailed tests based on 999 permutations. 5. To examine distance from the West, we used the longitude and latitude of the regional capitals, alongside and in lieu of the federal districts measure. When both measures were included our neighborhood variable performed much as reported in our tables. The location variable, however, failed to prove statistically significant in any models except those for 2000. These results are available from the authors. 6. We introduced this index in chapter 3 and the procedure for combining the components may be found in appendix 1. Still, as a reminder, the index consists of seven components: the percent of the region’s population with at least some higher education; the population density; the percent of the region’s residents living in cities or urban areas; residential telephones per 1,000 urban residents; the density of hardsurfaced roads, in kilometers per 1,000 km2 of territory; the number of theater-goers per 1,000 residents; and the number of museum-goers per 1,000 residents. 7. We construct the index using data from 1995, 2000, 2004, and 2007. The alphas range from .77 to .79. 8. For most of our models, unemployment outperforms income so we use it as the best means for testing the robustness of our spatial variables. 9. To check the robustness of this specification, we compared the model’s performance with and without the Golosov index for the second round of the 1996 presidential election—that is, during the most troublesome election when it comes to the number of regions having not held gubernatorial elections. The models perform in similar fashions with the main differences conforming to expectations: first, the goodness of fit is better with the full specification as the adjusted R2 increases from .21 to .31 thanks to the additional variable, which itself is powerful and statistically significant (see table 9.2). Second, without controlling for the number of effective candidates, the variable for whether the governor is elected no longer matters while the significance level of tenure drops to the .10 level. 10. We do not employ gubernatorial party affiliation data for elections prior to 2004 because they rarely proved to bind the governors’ hands. Even among Russia’s “red governors,” few had close ties to the Communist Party’s national or provincial organization and many incentives existed that kept governors independent of national party ties (see, among others, Slider 2001). 11. As a check on the robustness of our measure of regional deference to the Kremlin, we have employed the same models on the percentage of eligible votes received, without cutoffs. The direction and magnitude of the coefficients, as well as the overall goodness of fit, are very similar. Results are available from the authors. 12. In 2004 and 2008, the correlation between the average deference by district in the preceding elections and the percent non-Russian is high enough so that the regression results have some multicollinearity. It is not extreme, though, and the larger finding holds that a region’s neighbor’s performance in the preceding election corresponds to a region’s deference level in the subsequent election. 13. It also fails to matter in 2008. Results are available from the authors. Page 236 →14. GeoDaв„ў is a trademark of Luc Anselin. The software incorporates licensed libraries from the Environmental Systems Research Institute’s (ESRI) MapObjects LT2. It is available from the website of the GeoDa Center for Geospatial Analysis and Computation at the Arizona State University: http://geodacenter.asu.edu/software

Chapter 9 1. As Sharafutdinova (2010) emphasizes, a key difference between the Yeltsin and Putin administrations entails the tactics these patrons have used to address the principal-agent problem: ensuring that agents are serving the principal’s interests and not (just) their own. This problem is exacerbated by an asymmetric distribution of information (with the agents possessing more than the principal), the costs of monitoring agents, and the fact that, when regional politicians are themselves elected, these politicians serve multiple principals (the Kremlin, constituents among the regional elite, and voters) and can play the principals off of one another (674). Eliminating gubernatorial elections in favor of presidential appointments eliminated the latter concern.

Page 237 →

References Akhmetkarimov, Bulat. 2008. “Revisiting the Profile of the American Voter in the Context of Declining Turnout,” Alternatives: Turkish Journal of International Relations 7, nos. 2–3 (Summer and Fall), 111–33. Alexseev, Mikhail A. 1999. Center-Periphery Conflict in Post-Soviet Russia: A Federation Imperiled (New York: St. Martin’s Press). Alvarez, R. Michael, Thad E. Hall, and Susan D. Hyde, eds. 2008. Election Fraud: Detecting and Deterring Electoral Manipulation (Washington, DC: Brookings Institution Press). Andersen, Robert. 2008. Modern Methods for Robust Regression, Quantitative Applications in the Social Sciences no. 152 (Thousand Oaks, CA: Sage). Anderson, Richard D., Jr. 2010. “When the Center Can Hold: The Primacy of Politics in Shaping Russian Democracy,” Communist and Post-Communist Studies 43, no. 4 (December), 397–408. Anselin, Luc. 1994. “Exploratory Spatial Data Analysis and Geographic Information Systems,” in Marco Painho, ed., New Tools for Spatial Analysis, 45–54 (Luxembourg: Eurostat). Anselin, Luc. 1996. “The Moran Scatterplot as an ESDA Tool to Assess Local Instability in Spatial Association,” in Manfred Fischer et al., eds., Spatial Analytical Perspectives on GIS: GISDATA 4, 111–25 (London: Taylor & Francis). Badovskii, D. V., and A. Iu. Shutov. 1997. “Regional Elites in Post-Soviet Russia: Aspects of Political Involvement,” Russian Politics and Law 35, no. 1, 30–53. Baker, Peter, and Susan Glasser. 2005. Kremlin Rising: Vladimir Putin’s Russia and the End of Revolution (New York: Scribner). Banfield, Edward C., and James Q. Wilson. 1963. City Politics (Cambridge, MA: Harvard University Press and the MIT Press). Barghoorn, Frederick C., and Thomas F. Remington. 1986. Politics in the USSR, 3rd ed. (Boston: Little, Brown). Barkan, Joel D. 2000. “Protracted Transitions Among Africa’s New Democracies,” Democratization 7, no. 3 (Autumn), 227–43. Page 238 →Barnes, Samuel H., et al. 1979. Political Action: Mass Participation in Five Western Democracies (Beverly Hills, CA: Sage). Baturo, Alexander, and Johan A. Elkink. 2015. “Dynamics of Regime Personalization and Patron–Client Networks in Russia, 1999–2014,” Post-Soviet Affairs, prepublication version, accessed at http://www.tandfonline.com/doi/full/10.1080/1060586X.2015.1032532 Beissinger, Mark R. 2002. Nationalist Mobilization and the Collapse of the Soviet State (New York: Cambridge University Press). Belyakov, Vladimir V., and Walter J. Raymond. 1994. Constitution of the Russian Federation (Lawrenceville, VA: Brunswick Publishing and Novosti). Bermeo, Nancy. 2003. Ordinary People in Extraordinary Times: The Citizenry and the Breakdown of Democracy (Princeton: Princeton University Press).

Beryozkin, Andrei, Mikhail Myagkov, and Peter Ordeshook. 2003. “Location and Political Influence: A Further Elaboration of Their Effects on Voting in Recent Russian Elections,” Eurasian Geography and Economics 44, no. 3 (April), 169–83. Besley, Timothy, and Masayuki Kudamatsu. 2006. “Health and Democracy,” American Economic Review 96, no. 2 (May), 313–18. Bilev, Gavril. 2012. Checking the Boss: Legislative Autonomy and Executive Contestation in the Russian Regions, 1992–2005. PhD thesis from the Department of Political Science (Providence, RI: Brown University, May). Bilev, Gavril. 2013. “89 Russias: The Role of Regional Parliament in Shaping Executive Contestation,” paper presented at the Conference on Subnational Research in Comparative Politics, Center for Latin American and Caribbean Studies and Watson Institute for International Studies, Brown University, Providence, RI, May 9–10. Birch, Sarah. 2011. Electoral Malpractice (New York: Oxford University Press). Blais, Andre, and Agnieszka Dobrzynska. 1998. “Turnout in Electoral Democracies,” European Journal of Political Research 33, no. 2 (March), 239–61. Blais, Andrei. 2006. “What Effects Voter Turnout?” Annual Review of Political Science 9, 111–25. Blais, Andrei. 2007. “Turnout in Elections,” in Russell J. Dalton and Hans-Dieter Klingemann, eds., The Oxford Handbook of Political Behavior, 621–35 (New York: Oxford University Press). Blaydes, Lisa. 2010. Elections and Distributive Politics in Mubarak’s Egypt (New York: Cambridge University Press). Blossfeld, Hans Peter, Katrin Golsch, and Gotz Rohwer. 2007. Event History Analysis with Stata (Florence, KY: Psychology Press). Boix, Carles. 2011. “Democracy, Development, and the International System,” American Political Science Review 105, no. 4 (November), 809–28. Boix, Carles, and Susan C. Stokes. 2003. “Endogenous Democratization,” World Politics 55, no. 4 (July), 517–49. Bova, Russell. 1997. “Democracy and Liberty: The Cultural Connection,” Journal of Democracy 8, no. 1 (January), 112–26. Brady, Henry E., Sydney Verba, and Kay Lehman Schlozman. 1995. “Beyond SES: A Resource Model of Political Participation,” American Political Science Review 89, no. 2 (March), 271–94. Brehm, John, and Wendy Rahn. 1997. “Individual-level Evidence for the Causes Page 239 →and Consequences of Social Capital,” American Journal of Political Science 41, no. 3 (July), 999–1023. Brooker, Paul. 2013. Non-Democratic Regimes, 3rd ed. (New York: Palgrave Macmillan). Brown, Archie. 2009. The Rise and Fall of Communism (New York: HarperCollins). Brownlee, Jason. 2007. Authoritarianism in an Age of Democracy (New York: Cambridge University Press). Bueno de Mesquita, Bruce. 2006. “Central Issues in the Study of International Conflict,” in Donald A. Ritchie and Barry R. Weingast, eds., The Oxford Handbook of Political Economy, 831–51 (New York: Oxford University Press).

Bunce, Valerie. 1995. “Should Transitologists Be Grounded?” Slavic Review 54, no. 1 (Spring), 111–27. Bunce, Valerie. 1999. Subversive Institutions: The Design and Demise of Socialism and the State (New York: Cambridge University Press). Bunce, Valerie. 2003. “Rethinking Recent Democratization: Lessons from the Postcommunist Experience,” World Politics 55, no. 2 (January), 167–92. Bunce, Valerie J., and Sharon L. Wolchik. 2011. Defeating Authoritarian Leaders in Postcommunist Countries (New York: Cambridge University Press). Burkhart, Ross E., and Michael S. Lewis-Beck. 1994. “Comparative Democracy: The Economic Development Thesis,” American Political Science Review 88, no. 4 (December), 903–10. Burrett, Tina. 2011. Television and Presidential Power in Putin’s Russia (New York: Routledge). Busygina, Irina M. 1997. “The President’s Representatives: Problems of Establishing and Prospects for Developing an Institution,” Russian Politics and Law 35, no. 1, 15–29. Buzin, A. Iu., and A. E. Liubarev. 2008. Prestuplenie bez Nakazaniia: Administrativnye Tekhologii Federal’nykh Vyborov 2007–2008 godov, Pravo i Vybory (Moscow: NIKKOLO M). Calingaert, Daniel. 2006. “Election Rigging and How to Fight It,” Journal of Democracy 17, no. 3 (July), 138–51. Carothers, Thomas. 2002. “The End of the Transition Paradigm,” Journal of Democracy 13, no. 1 (January), 5–21. Case, William. 2006. “Manipulative Skills: How Do Rulers Control the Electoral Arena?” in Andreas Schedler, ed., Electoral Authoritarianism: The Dynamics of Unfree Competition, 95–112 (Boulder, CO: Lynne Rienner). Central Electoral Commission of the Russian Federation. 1997. Vybory Glav Ispolnitel’oi Vlasti Sub’’ektov Rossiiskoi Federatsii, 1995–1997: Elektoral’naia Statistika (Moscow: Ves’ Mir). Central Electoral Commission of the Russian Federation. 1999. Vybory i Referendumy, accessed at http://www.izbircom.ru/1911/xod/index.shtml Central Electoral Commission of the Russian Federation. 2000. Vybory i Referendumy, accessed at http://www.izbircom.ru/2217/xod/index.shtml Central Electoral Commission of the Russian Federation. 2001. Vybory v Organy Gosudarstvennoi Vlasti Sub’’ektov Rossiiskoi Federatsii 1997–2000: Elektoral’naia Statistika v 2 Tomakh (Moscow: Ves’ Mir). Central Electoral Commission of the Russian Federation. 2008. Dannye ProtokolovPage 240 → Territorial’nykh Izbiratel’nykh Komissii ob Itogakh Golosovaniia po Federal’nomu Okrugu (1995–2-TIK-usech.xls) (Moscow: Central Electoral Commission of the Russian Federation), accessed at http://cikrf.ru/banners/vib_arhiv/gosduma/1995/index.html, on May 22, 2013. Central Electoral Commission of the Russian Federation. 2012. Vybory i Referendumy. Accessed at http://www.vybory.izbirkom.ru/region/region/izbirkom Chaisty, Paul. 2005. “Majority Control and Executive Dominance: Parliament-President Relations in Putin’s Russia,” in Alex Pravda, ed., Leading Russia: Putin in Perspective, Essays in Honour of Archie Brown, 119–38 (New York: Oxford University Press).

Chaisty, Paul. 2006. Legislative Politics and Economic Power in Russia (New York: Palgrave Macmillan). Chandra, Kanchan. 2004. Why Ethnic Parties Succeed: Patronage and Ethnic Headcounts in India (New York: Cambridge University Press). Chebankova, Elena. 2012. “State-Sponsored Civic Associations in Russia: Systemic Integration or the вЂWar of Position’?” East European Politics 28, no. 4 (December), 390–408. Chebankova, Elena A. 2010. Russia’s Federal Relations: Putin’s Reforms and the Management of the Regions (New York: Routledge). Cheibub, JosГ© Antonio, Jennifer Gandhi, and James Raymond Vreeland. 2010. “Democracy and Dictatorship Revisited,” Public Choice 143, nos. 1–2 (April), 67–101. Cheloukhine, Serguei, and Joseph King. 2007. “Corruption Networks as a Sphere of Investment Activities in Modern Russia,” Communist and Post-Communist Studies 40, no. 1 (March), 107–22. Clark, William A. 1998. “Presidential Prefects in the Russian Provinces: Yeltsin’s Regional Cadres Policy,” in Graeme Gill, ed., Elites and Leadership in Russian Politics: Selected Papers from the Fifth World Congress of Central and East European Studies, Warsaw, 1995, 24–51 (New York: St. Martin’s Press). Clem, Ralph S. 2006. “Russia’s Electoral Geography: A Review,” Eurasian Geography and Economics 47, no. 4, 381–406. Clem, Ralph S., and Peter R. Craumer. 1995. “The Politics of Russia’s Regions: A Geographical Analysis of the Russian Election and Constitutional Plebiscite of December 1993,” Post-Soviet Geography 36, no. 2, 67–86. Clem, Ralph S., and Peter R. Craumer. 1998. “Regional Patterns of Voter Turnout in Russian Elections, 1993–1996,” in Matthew Wyman et al., eds., Elections and Voters in Postcommunist Russia, 40–67 (London: Cass). Cleves, Mario A., William W. Gould, and Roberto G. Guttierrez. 2010. An Introduction to Survival Analysis Using Stata, 3rd ed. (College Station, TX: Stata Press). Collier, David, Henry E. Brady, and Jason Seawright. 2004. “Critiques, Responses and Trade-Offs: Drawing Together the Debate,” in Henry E. Brady and David Collier, eds., Rethinking Social Inquiry: Diverse Tools, Shared Standards, 195–227 (Lanham, MD: Rowman & Littlefield). Colton, Timothy J. 2000. Transitional Citizens: Voters and What Influences Them in the New Russia (Cambridge, MA: Harvard University Press). Colton, Timothy J. 2005. “Putin and the Attenuation of Russian Democracy,” in Page 241 →Alex Pravda, ed., Leading Russia: Putin in Perspective, Essays in Honour of Archie Brown, 103–18 (New York: Oxford University Press). Colton, Timothy J. 2008. Yeltsin: A Life (New York: Basic Books). Colton, Timothy J., and Michael McFaul. 2000. “Reinventing Russia’s Party of Power: вЂUnity’ and the 1999 Duma Elections,” Post-Soviet Affairs 16, no. 3, 201–24. Colton, Timothy J., and Michael McFaul. 2003. Popular Choice and Managed Democracy: The Russian Elections of 1999 and 2000 (Washington, DC: Brookings Institution Press). Colton, Timothy J., and Cindy Skach. 2005. “The Russian Predicament,” Journal of Democracy 16, no. 3 (July), 113–26.

Coppedge, Michael. 1999. “Thickening Thin Concepts and Theories: Combining Large N and Small in Comparative Politics,” Comparative Politics 31, no. 4 (July), 465–76. Cornelius, Wayne A. 1977. “Leaders, Followers and Official Patrons in Urban Mexico,” in Steffen W. Schmidt et al., eds., Friends, Followers and Factions: A Reader in Political Clientelism, 337–53 (Berkeley: University of California Press). Cox, Gary W. 1997. Making Votes Count: Strategic Coordination in the World’s Electoral Systems (New York: Cambridge University Press). Cox, Gary W., and Michael C. Munger. 1989. “Closeness, Expenditures, and Turnout in the 1982 U.S. House Elections,” American Political Science Review 83, no. 1 (March), 217–31. D’Agostino, Ralph B., Albert Belanger, and Ralph B. D’Agostino, Jr. 1990. “A Suggestion for Using Powerful and Informative Tests of Normality,” The American Statistician 44, no. 4 (November), 316–21. Dahl, Robert A. 1971. Polyarchy: Participation and Opposition (New Haven: Yale University Press). Dahl, Robert A. 1998. On Democracy (New Haven: Yale University Press). Dalton, Russell J. 1984. “Cognitive Mobilization and Partisan Dealignment in Advanced Industrial Democracies,” Journal of Politics 46, no. 1 (February), 264–84. Dalton, Russell J. 2013. Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, 6th ed. (Washington, DC: CQ Press). Darden, Keith, and Anna Grzymala-Busse. 2006. “The Great Divide: Literacy, Nationalism, and the Communist Collapse,” World Politics 59, no. 1 (October), 83–115. Darr, Benjamin, and Vicki Hesli. 2010. “Differential Voter Turnout in a Post-Communist Muslim Society: The Case of the Kyrgyz Republic,” Communist and Post-Communist Studies 43, no. 3 (September), 309–24. Dawisha, Karen. 2014. Putin’s Kleptocracy: Who Owns Russia? (New York: Simon and Schuster). DeBardeleben, Joan, and Aleksander A. Galkin. 1997. “Electoral Behavior and Attitudes in Russia: Do Regions Make a Difference or Do Regions Just Differ?” in Peter J. Stavrakis et al., eds., Beyond the Monolith: The Emergence of Regionalism in Post-Soviet Russia, 57–80 (Washington, DC: The Woodrow Wilson Center Press). Page 242 →Deutsch, Karl W. 1961. “Social Mobilization and Political Development,” American Political Science Review 53, no. 3, 493–514. di Palma, Guiseppe. 1990. To Craft Democracies: An Essay on Democratic Transitions (Berkeley: University of California Press). Diamond, Larry. 2002. “Thinking About Hybrid Regimes,” Journal of Democracy 13, no. 2 (April), 21–35. Diamond, Larry. 2008. The Spirit of Democracy: The Struggle to Build Free Societies Throughout the World (New York: Henry Holt). Dininio, Phyllis, and Robert Orttung. 2005. “Explaining Patterns of Corruption in the Russian Regions,” World Politics 57, no. 4 (July), 500–29. Dmitriev, Mikhail E., et al. 2015. “Mezhdu Krymom i Krizisom: Pochemu Izmenilos’ Sotsial’nye Ustanovki Rossiian?” (Moscow: Komitet Grazhdanskikh Initiativ, March 31). Accessed at http://komitetgi.ru

/analytics/2147/ Doyle, Michael W. 2012. Liberal Peace: Selected Essays (New York: Routledge). Dunleavy, Patrick, and FranГ§oise Boucek. 2003. “Constructing the Number of Parties,” Party Politics 9, no. 3 (May), 291–315. Dunlop, John B. 1990. “The Return of Russian Nationalism,” Journal of Democracy 1, no. 3 (Summer), 114–22. Dunning, Thad. 2008. Crude Democracy: Natural Resource Wealth and Political Regimes (New York: Cambridge University Press). Duverger, Maurice. 1954. Political Parties (New York: Wiley). Easter, Gerald M. 2001. “Networks, Bureaucracies, and the Russian State,” in Klaus Segbers, ed., Explaining Post-Soviet Patchworks, Vol. 2: Pathways from the Past to the Global, 39–58 (Burlington, VT: Ashgate). Easton, David, and Jack Dennis. 1969. Children in the Political System (New York: McGraw-Hill). Eckstein, Harry. 1988. “A Culturalist Theory of Political Change,” American Political Science Review 82, no. 3 (September), 789–804. Eckstein, Harry, et al. 1998. Can Democracy Take Root in Post-Soviet Russia? Explorations in State-Society Relations (Lanham, MD: Rowman & Littlefield). Elkins, Zachary, and Beth Simmons. 2005. “On Waves, Clusters, and Diffusion: A Conceptual Framework, ” The ANNALS of the American Academy of Political and Social Science 598, no. 1 (March), 33–51. Epstein, David L., et al. 2006. “Democratic Transitions,” American Journal of Political Science 50, no. 3 (July), 551–69. Esri. 2014. GIS Dictionary (Redlands, CA: Esri), accessed at http://support.esri.com/en/knowledgebase /GISDictionary/term/spatial%20autocorrelation (June 13, 2014). Fainsod, Merle. 1953. How Russia Is Ruled (Cambridge, MA: Harvard University Press). Fedarko, Kevin, and David Aikman. 1994. “Rising Czar?” Time 144, no. 2 (July 11), 38–43. Feenstra, Robert C., Robert Inklaar, and Marcel P. Timmer. 2013. The Next Generation of the Penn World Table (dataset available from Groningen Growth and Development Centre), accessed at www.ggdc.net/pwt Field, Mark G., and Judyth L. Twigg, eds. 2000. Russia’s Torn Safety Nets: Health and Social Welfare during the Transition (New York: St. Martin’s Press). Page 243 →Fiorina, Morris. 1981. Retrospective Voting in American National Elections (New Haven: Yale University Press). Fish, M. Steven. 1995. Democracy from Scratch: Opposition and Regime in the New Russian Revolution (Princeton: Princeton University Press). Fish, M. Steven. 2003. “The Impact of the 1999–2000 Parliamentary and Presidential Elections on Political Party Development,” in Vicki L. Hesli and William M. Reisinger, eds., The 1999–2000 Elections in Russia: Their Impact and Legacy, 186–212 (New York: Cambridge University Press).

Fish, M. Steven. 2005. Democracy Derailed in Russia: The Failure of Open Politics (New York: Cambridge University Press). Fox, Jonathan A. 1994. “Latin America’s Emerging Local Politics,” Journal of Democracy 5, no. 2 (April), 105–16. Franco, Alvaro, Carlos ГЃlvarez-Dardet, and Maria Teresa Ruiz. 2004. “Effect of Democracy on Health: Ecological Study,” BMJ 329, no. 7480 (December), 1421–23. Franklin, Mark N. 2004. Voter Turnout and the Dynamics of Electoral Competition in Established Democracies Since 1945 (New York: Cambridge University Press). Freedom House. various years. Freedom in the World (Washington, DC: Freedom House), accessed at http://www.freedomhouse.org/report-types/freedom-world Frye, Timothy. 1997. “A Politics of Institutional Choice: Post-Communist Presidencies,” Comparative Political Studies 30, no. 5 (October), 523–52. Frye, Timothy. 2010. “Corruption and Rule of Law,” in Anders Aslund et al., eds., Russia After the Global Economic Crisis, 79–94 (Washington, DC: Peterson Institute for Internal Economics). Fukuyama, Francis. 1992. The End of History and the Last Man (New York: Free Press). Gandhi, Jennifer. 2008. Political Institutions Under Dictatorship (New York: Cambridge University Press). Gandhi, Jennifer, and Ellen Lust-Okar. 2009. “Elections Under Authoritarianism,” Annual Review of Political Science 142, nos. 1–2, 403–22. GarretГіn, Manuel Antonio. 1995. “Redemocratization in Chile,” Journal of Democracy 6, no. 1 (January), 146–58. Geddes, Barbara. 1999. “What Do We Know About Democratization After Twenty Years?” Annual Review of Political Science 2, 115–44. Geddes, Barbara. 2005. “The Role of Elections in Authoritarian Regimes,” paper presented at the American Political Science Association, Washington, DC, September, http://www.allacademic.com/meta /p41799_index.html, accessed September 17, 2012. Geddes, Barbara, Joseph Wright, and Erica Frantz. 2013. New Data on Autocratic Breakdown and Regime Transitions (dataset available from Penn State University, http://dictators.la.psu.edu), accessed May 2, 2013. Gel’man, Vladimir. 2010a. “The Dynamics of Sub-National Authoritarianism: Russia in Comparative Perspective,” in Vladimir Gel’man and Cameron Ross, eds., The Politics of Sub-National Authoritarianism in Russia, 1–18 (Aldershot, UK: Ashgate). Gel’man, Vladimir. 2010b. “Introduction: Resource Curse and Post-Soviet Eurasia,”Page 244 → in Vladimir Gel’man and Otar Marganiya, eds., Resource Curse and Post-Soviet Eurasia: Oil, Gas and Modernization, 1–22 (Lanham, MD: Lexington Books). Gel’man, Vladimir. 2012. “Subversive Institutions, Informal Governance, and Contemporary Russian Politics,” Communist and Post-Communist Studies 45, nos. 3–4 (September-December), 295–303. Gel’man, Vladimir. 2013. “Politics, Governance, and the Zigzags of the Power Vertical: Toward a Framework for Analyzing Russia’s Local Regimes,” in William M. Reisinger, ed., Russia’s Regions and Comparative Subnational Politics, Routledge Research in Comparative Politics, 25–39 (New York: Routledge).

Gel’man, Vladimir. 2015. Authoritarian Russia: Analyzing Post-Soviet Regime Changes (Pittsburgh: University of Pittsburgh Press). Gel’man, Vladimir, and Vitalii Elizarov. 1999. “Russia’s Transition and Founding Elections,” in Vladimir Gel’man and Grigorii V. Golosov, eds., Elections in Russia, 1993–1996: Analyses, Documents and Data, 19–46 (Berlin: Edition Sigma). Gel’man, Vladimir, and Grigorii V. Golosov, eds. 1999. Elections in Russia, 1993–1996: Analyses, Documents and Data (Berlin: Edition Sigma). Gel’man, Vladimir Ia., Grigorii V. Golosov, and E. Iu. Meleshkina, eds. 2002. Vtoroi Elektoral’nyi Tsikl v Rossii (1999–2000 gg.) (Moscow: Ves’ Mir). Gel’man, Vladimir, and Tomila Lankina. 2008. “Authoritarian Versus Democratic Diffusions: Explaining Institutional Choices in Russia’s Local Government,” Post-Soviet Affairs 24, no. 1, 40–62. Gel’man, Vladimir, and Otar Marganiya, eds. 2010. Resource Curse and Post-Soviet Eurasia: Oil, Gas and Modernization (Lanham, MD: Lexington Books). Gel’man, Vladimir, and Sergei Ryzhenkov. 2011. “Local Regimes, Sub-national Governance and the вЂPower Vertical’ in Contemporary Russia,” Europe-Asia Studies 63, no. 3 (May), 449–65. Gessen, Masha. 2013. The Man Without a Face: The Unlikely Rise of Vladimir Putin, paperback ed. (New York: Riverhead Books). Geys, Benny. 2006. “Explaining Voter Turnout: A Review of Aggregate-Level Research,” Electoral Studies 25, no. 4 (December), 637–63. Gharibvand, Lida, and Lei Liu. 2009. “Analysis of Survival Data with Clustered Events,” paper presented at the SAS Global Forum 2009, Washington, DC, March 22–25. Gibson, Edward L. 2005. “Boundary Control: Subnational Authoritarianism in Democratic Countries,” World Politics 58, no. 1 (October), 101–32. Gibson, Edward L. 2012. Boundary Control: Subnational Authoritarianism in Federal Democracies (New York: Cambridge University Press). Gibson, James L. 2001. “Social Networks, Civil Society, and the Prospects for Consolidating Russia’s Democratic Transition,” American Journal of Political Science 45, no. 1 (January), 51–68. Gill, Graeme. 2012. “The Decline of a Dominant Party and the Destabilization of Electoral Authoritarianism? ” Post-Soviet Affairs 28, no. 4 (October-December), 449–71. Gitelman, Zvi. 2005. “The Democratization of Russia in Comparative Perspective,” in Stephen White et al., eds., Developments in Russian Politics, 241–56 (Durham, NC: Duke University Press). Page 245 →Giuliano, Elise. 2011. Constructing Grievance: Ethnic Nationalism in Russia’s Republics (Ithaca: Cornell University Press). Gleason, Gregory. 1990. Federalism and Nationalism (Boulder, CO: Westview). Gleditsch, Kristian Skrede, and Michael D. Ward. 2006. “Diffusion and the International Context of Democratization,” International Organization 60, no. 4 (October), 911–33. Glinski, Dmitri, and Peter Reddaway. 1999. “The Ravages of вЂMarket Bolshevism,’” Journal of Democracy 10, no. 2, 19–34.

GГ¶bel, Christian. 2011. “Semiauthoritarianism,” in John T. Ishiyama and Marijke Breuning, eds., 21st Century Political Science: A Reference Handbook, 258–67 (Thousand Oaks, CA: Sage). Goldman, Marshall I. [2008] 2010. Petrostate: Putin, Power and the New Russia, paperback ed. (New York: Oxford University Press). Golosov, Grigorii V. 1999. “From Adygeya to Yaroslavl: Factors of Party Development in the Regions of Russia, 1995–1998,” Europe-Asia Studies 51, no. 8, 1333–65. Golosov, Grigorii V. 2003. “Electoral Systems and Party Formation in Russia: A Cross-Regional Analysis, ” Comparative Political Studies 36, no. 8 (October), 912–35. Golosov, Grigorii V. 2004. Political Parties in the Regions of Russia: Democracy Unclaimed (Boulder, CO: Lynne Rienner). Golosov, Grigorii V. 2006. “Disproportionality by Proportional Design: Seats and Votes in Russia’s Regional Legislative Elections, December 2003–March 2005,” Europe-Asia Studies 58, no. 1 (January), 25–55. Golosov, Grigorii V. 2008. Baza Dannykh вЂRossiiskaia Elektoral’naia Statistika’ (Moscow: Interregional Electoral Network of Assistance), accessed at http://db.irena.org.ru [currently at http://db.geliks.org], on June 2, 2008. Golosov, Grigorii V. 2009. “Building a New Political Machine,” Russian Analytical Digest, no. 57 (March 17), 2–6. Golosov, Grigorii V. 2010. “The Effective Number of Parties. A New Approach,” Party Politics 16, no. 2 (March), 171–92. Golosov, Grigorii V. 2011a. “The Regional Roots of Electoral Authoritarianism in Russia,” Europe-Asia Studies 63, no. 4 (June), 623–39. Golosov, Grigorii V. 2011b. “Russia’s Regional Legislative Elections, 2003–2007: Authoritarianism Incorporated,” Europe-Asia Studies 63, no. 3 (May), 397–414. Golosov, Grigorii V. 2014a. “Co-optation in the Process of Dominant Party System Building: The Case of Russia,” East European Politics 30, no. 2 (April), 271–85. Golosov, Grigorii V. 2014b. “The Last Resort of Political Regionalism: Electoral Blocs in Russia’s Regional Legislative Elections, 2003–2005,” East European Politics 30, no. 1 (January), 71–85. Golosov, Grigorii V. 2014c. “The Territorial Genealogies of Russia’s Political Parties and the Transferability of Political Machines,” Post-Soviet Affairs 30, no. 6, 464–80. Golub, Jonathan. 2008. “Survival Analysis,” in Janet M. Box-Steffensmeier et al., eds., The Oxford Handbook of Political Methodology, 530–46 (Oxford: Oxford University Press). Goode, J. Paul. 2007. “The Puzzle of Putin’s Gubernatorial Appointments,” Europe-Asia Studies 59, no. 3 (May), 365–99. Page 246 →Goodin, Robert E., and Charles Tilly, eds. 2006. The Oxford Handbook of Contextual Political Analysis (Oxford: Oxford University Press). http://dx.doi.org/10.1093/oxfordhb/9780199270439.001.0001 Grzymala-Busse, Anna, and Pauline Jones Luong. 2006. “Democratization: Post-Communist Implications,” in Barry R. Weingast and Donald Wittman, eds., The Oxford Handbook of Political Economy, 656–69 (New York: Oxford University Press).

Gubernatory.ru. 2012. “Dmitrii Medvedev Podpisal Zakon o Vyborakh Gubernatorov,” Gubernatory.ru (May 2), http://governors.ru/?regmode=regions&statja=381758, accessed May 3, 2012. Haber, Stephen, and Victor Menaldo. 2011. “Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse,” American Political Science Review 105, no. 1 (February), 1–26. Haerpfer, Christian, et al. 2009. “Introduction,” in Christian Haerpfer et al., eds., Democratization, 1–7 (New York: Oxford University Press). Hahn, Gordon. 2012. “Perestroyka 2.0: Toward Non-Revolutionary Regime Transformation in Russia?” Post-Soviet Affairs 28, no. 4 (October-December), 472–15. Hahn, Jeffrey W. 1997. “Regional Elections and Political Stability in Russia,” Post-Soviet Affairs 38, no. 5 (May), 251–63. Hale, Henry E. 2003. “Explaining Machine Politics in Russia’s Regions: Economy, Ethnicity, and Legacy,” Post-Soviet Affairs 19, no. 3, 226–61. Hale, Henry E. 2004. “Yabloko and the Challenge of Building a Liberal Party in Russia,” Europe-Asia Studies 56, no. 7 (November), 993–1020. Hale, Henry E. 2005. “Why Not Parties? Electoral Markets, Party Substitutes, and Stalled Democratization in Russia,” Comparative Politics 37, no. 2 (January), 147–66. Hale, Henry E. 2006. Why Not Parties in Russia? (New York: Cambridge University Press). Hale, Henry E. 2007. “Correlates of Clientelism: Political Economy, Politicized Ethnicity and PostCommunist Transition,” in Herbert Kitschelt and Steven I. Wilkinson, eds., Patrons, Clients and Policies: Patterns of Democratic Accountability and Political Competition, 227–50 (New York: Cambridge University Press). Hale, Henry E. 2010. “Eurasian Polities as Hybrid Regimes: The Case of Putin’s Russia,” Journal of Eurasian Studies 1, no. 1 (January), 33–41. Hale, Henry E. 2011. “Hybrid Regimes: When Democracy and Autocracy Mix,” in Nathan J. Brown, ed., The Dynamics of Democratization: Dictatorship, Development and Diffusion, 23–45 (Baltimore: Johns Hopkins University Press). Hale, Henry E. 2015. Patronal Politics: Eurasian Regime Dynamics in Comparative Perspective (New York: Cambridge University Press). Hannan, Michael T. 1991. Aggregation and Disaggregation in the Social Sciences, rev. ed. (Lexington, MA: Lexington Books). Hanson, Stephen. 2003. “Instrumental Democracy: The End of Ideology and the Decline of Russian Political Parties,” in Vicki L. Hesli and William M. Reisinger, eds., The 1999–2000 Elections in Russia: Their Impact and Legacy, 163–85 (New York: Cambridge University Press). Hanson, Stephen E. 2010. Post-Imperial Democracies: Ideology and Party Formation in Page 247 →Third Republic France, Weimar Germany, and Post-Soviet Russia (New York: Cambridge University Press). Harasymiw, Bohdan. 1969. “Nomenklatura: The Soviet Communist Party’s Leadership Recruitment System,” Canadian Journal of Political Science 2, no. 4 (December), 493–512. Harding, Sue-Ann. 2012. Beslan: Six Stories of the Siege (Manchester, UK: Manchester University Press).

Hedlund, Stefan. 2008. “Rents, Rights, and Service: Boyar Economics and the Putin Transition,” Problems of Post-Communism 55, no. 4 (July–August), 29–41. Hendley, Kathryn. 2009. “вЂTelephone Law’ and the вЂRule of Law’: The Russian Case,” Hague Journal on the Rule of Law 1, no. 2 (September), 241–62. Herb, Michael. 2005. “No Representation without Taxation? Rents, Development, and Democracy,” Comparative Politics 37, no. 3 (April), 297–316. Herrera, Yoshiko M. 2005. Imagined Economies: The Sources of Russian Regionalism (New York: Cambridge University Press). Herszenhorn, David M. 2013. “Putin Orders Change in Election Rules,” New York Times (January 3), A10. Hesli, Vicki L., and William M. Reisinger, eds. 2003. The 1999–2000 Elections in Russia: Their Impact and Legacy (New York: Cambridge University Press). Heston, Alan, Robert Summers, and Bettina Aten. 2012. Penn World Table Version 7.1 (dataset available from Center for International Comparisons of Production, Income and Prices at the University of Pennsylvania, https://pwt.sas.upenn.edu/php_site/pwt_index.php), accessed November 1, 2012. Hill, Fiona. 2004. Energy Empire: Oil, Gas and Russia’s Revival (London: The Foreign Policy Centre). Hill, Ronald J. 1985. Soviet Union: Politics, Economy and Society (Boulder, CO: Lynne Rienner). Hill, Ronald J., and Peter Frank. 1986. The Soviet Communist Party, 3rd ed. (London: Allen and Unwin). Holmes, Leslie. 2008. “Corruption and Organised Crime in Putin’s Russia,” Europe-Asia Studies 60, no. 6 (August), 1011–31. Horvath, Robert. 2011. “Putin’s вЂPreventive Counter-Revolution’: Post-Soviet Authoritarianism and the Spectre of Velvet Revolution,” Europe-Asia Studies 63, no. 1 (January), 1–25. Hough, Jerry. 1969. The Soviet Prefects: The Local Party Organs in Industrial Decision-Making (Cambridge, MA: Harvard University Press). Hough, Jerry F. 1998. “The Failure of Party Formation and the Future of Russian Democracy,” in Timothy J. Colton and Jerry F. Hough, eds., Growing Pains: Russian Democracy and the Election of 1993, 669–711 (Washington, DC: Brookings Institution Press). Houle, Christian. 2009. “Inequality and Democracy: Why Inequality Harms Consolidation but Does Not Affect Democratization,” World Politics 61, no. 4 (October), 589–622. Howard, Marc Morje, and Philip G. Roessler. 2006. “Liberalizing Electoral Outcomes in Competitive Authoritarian Regimes,” American Journal of Political Science 50, no. 2 (April), 365–81. Page 248 →Huckfeldt, Robert, and John Sprague. 1992. “Political Parties and Electoral Mobilization: Political Structure, Social Structure, and the Party Canvass,” American Political Science Review 86, no. 1 (March), 70–86. Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Century (Norman: University of Oklahoma Press). Huntington, Samuel P. 1996. The Clash of Civilizations and the Remaking of World Order (New York: Simon & Schuster).

Huskey, Eugene. 2005. “Putin as Patron: Cadres Policy in the Russian Transition,” in Alex Pravda, ed., Leading Russia: Putin in Perspective, Essays in Honour of Archie Brown, 161–78 (New York: Oxford University Press). Hutcheson, Derek S. 2004. “Disengaged or Disenchanted? The Vote вЂAgainst All’ in Post-Communist Russia,” Journal of Communist Studies and Transition Politics 20, no. 1 (March), 98–121. Hyde, Matthew. 2001. “Putin’s Federal Reforms and Their Implications for Presidential Power in Russia, ” Europe-Asia Studies 53, no. 5, 719–43. Hyde, Susan D., and Nikolay Marinov. 2012. “Which Elections Can Be Lost?” Political Analysis 20, no. 2 (Spring), 191–210. Inglehart, Ronald. 1977. The Silent Revolution: Changing Values and Political Styles Among Western Publics (Princeton: Princeton University Press). Inglehart, Ronald. 1988. “The Renaissance of Political Culture,” American Political Science Review 82, no. 4 (December), 1203–30. Inglehart, Ronald. 1990. Culture Shift in Advanced Industrial Society (Princeton: Princeton University Press). Inglehart, Ronald, and Christian Welzel. 2010. “Changing Mass Priorities: The Link between Modernization and Democracy,” Perspectives on Politics 8, no, 2 (June), 551–67. Inkeles, Alex, and David H. Smith. 1974. Becoming Modern: Individual Change in Six Developing Countries (Cambridge, MA: Harvard University Press). Isaacs, Rico, and Sarah Whitmore. 2014. “The Limited Agency and Life-Cycles of Personalized Dominant Parties in the Post-Soviet Space: The Cases of United Russia and Nur Otan,” Democratization 21, no. 4, 699–721. Ishiyama, John T., and Ryan Kennedy. 2001. “Superpresidentialism and Political Party Development in Russia, Ukraine, Armenia and Kyrgyzstan,” Europe-Asia Studies 53, no. 8, 1177–91. Ivanchenko, A. V., and A. E. Liubarev. 2007. Rossiiskie Vybory: Ot Perestroiki do Suverennoi Demokratii (Moscow: Aspekt Press). Jackman, Robert W. 1987. “Political Institutions and Voter Turnout in the Industrial Democracies,” American Political Science Review 81, no. 2 (June), 405–23. Jaquette, Jane S. 2001. “Women and Democracy: Regional Differences and Contrasting Views,” Journal of Democracy 12, no. 3 (July), 111–25. Javeline, Debra. 2003. Protest and the Politics of Blame: The Russian Response to Unpaid Wages (Ann Arbor: University of Michigan Press). Jepsen, Eric M. 2011. “Processes of Democratization,” in John T. Ishiyama and Marijke Breuning, eds., 21st Century Political Science: A Reference Handbook, 275–83 (Thousand Oaks, CA: Sage). Jewell, Malcolm E., and David M. Olson. 1988. Political Parties and Elections in American States, 3rd ed. (Homewood, IL: Dorsey Press). Page 249 →Johnson, Gerald W. 1971. “Research Note on Political Correlates of Voter Participation: A Deviant Case Analysis,” American Political Science Review 65, no. 3 (September), 768–76. Jones Luong, Pauline. 2002. Institutional Change and Political Continuity in Post-Soviet Central Asia: Power,

Perceptions, and Pacts (New York: Cambridge University Press). Jones Luong, Pauline, and Erika Weinthal. 2010. Oil Is Not a Curse: Ownership Structure and Institutions in Soviet Successor States (New York: Cambridge University Press). Kahn, Jeff. 2001. “What Is the New Russian Federalism?” in Archie Brown, ed., Contemporary Russian Politics: A Reader, 374–83 (New York: Oxford University Press). Kahn, Jeffrey. 2002. Federalism, Democratization and the Rule of Law in Russia (New York: Oxford University Press). Kailitz, Steffen. 2013. “Classifying Political Regimes Revisited: Legitimation and Durability,” Democratization 20, no. 1 (January), 39–60. Karklins, Rasma. 1986. “Soviet Elections Revisited: Voter Abstention in Noncompetitive Voting,” American Political Science Review 80, no. 2 (June), 449–69. Karl, Terry Lynn. 1986. “Imposing Consent? Electoralism versus Democratization in El Salvador,” in Paul Drake and Eduardo Silva, eds., Elections and Democratization in Latin America, 9–36 (San Diego: Center for Iberian and Latin American Studies). Karp, Jeffrey A., Susan A. Banducci, and Shaun Bowler. 2008. “Getting Out the Vote: Party Mobilization in a Comparative Perspective,” British Journal of Political Science 38, no. 1 (January), 91–112. Kartsev, Vladimir. 1995. !Zhirinovsky! (New York: Columbia University Press). Kaya, Ruchan, and Michael Bernhard. 2013. “Are Elections Mechanisms of Authoritarian Stability or Democratization? Evidence from Postcommunist Eurasia,” Perspectives on Politics 11, no. 3 (September), 734–52. Kayser, Mark Andreas. 2005. “Who Surfs, Who Manipulates? The Determinants of Opportunistic Election Timing and Electorally Motivated Economic Intervention,” American Political Science Review 99, no. 1, 17–27. Keiser, Richard. 2001. “Political Machine,” in Joel Krieger, ed., The Oxford Companion to Politics of the World, 2nd ed., 666–67 (New York: Oxford University Press). Kennedy, Ryan. 2010. “The Contradiction of Modernization: A Conditional Model of Endogenous Democratization,” Journal of Politics 72, no. 3 (July), 785–98. Key, V. O. [1934] 1936. The Techniques of Political Graft in the United States (Chicago: University of Chicago Libraries). Kirkow, Peter. 1998. Russia’s Provinces: Authoritarian Transformation versus Local Autonomy? (New York: St. Martin’s Press). Kitschelt, Herbert. 1995. “Formation of Party Cleavages in Post-Communist Democracies,” Party Politics 1, no. 4 (October), 447–72. Kitschelt, Herbert, et al. 1999. Post-Communist Party Systems: Competition, Representation, and Inter-Party Cooperation (New York: Cambridge University Press). Kline, Rex B. 2005. Principles and Practice of Structural Equation Modeling, 2nd ed. (New York: Guilford Press).

Kolsto, Pal. 2000. Political Construction Sites: National-Building in Russia and the Post-Soviet States (Boulder, CO: Westview Press). Page 250 →Konitzer, Andrew. 2005. Voting for Russia’s Governors: Regional Elections and Accountability Under Yeltsin and Putin (Washington, DC: Woodrow Wilson Press). Kozlov, Nikolai D. 2008. “Politicheskie Kul’tury Regionov Rossii: Uravnenie so Mnogimi Neizvestnymi [Political Cultures of the Regions of Russia: An Equation with Many Unknowns],” POLIS: Politicheskie Issledovaniia, no. 4, 8–26. Kozlov, V. N., and D. B. Oreshkin. 1996. “Regional’nyi Analiz Rezul’tatov Vyborov i Tipologiia Regionov po Politicheskim Predpochteniiam Izbiratelei,” in Tsental’naia Izbiratel’naia Komissiia Rossiiskoi Federatsii, ed., Vybory Prezidenta Rossiiskoi Federatsii 1996: Elektoral’naia Statistika, 172–97 (Moscow: Ves’ Mir). Kramer, David J., and Susan Corke. 2013. “Contending with Putin’s Russia: Proposals for a New Approach,” Freedom House (March 12, 2013), http://www.freedomhouse.org/report/contending-putins-russia /essay Kuenzi, Michelle, and Gina M. S. Lambright. 2011. “Who Votes in Africa? An Examination of Electoral Participation in 10 African Countries,” Party Politics 17, no. 6 (November), 767–99. Kuzio, Taras. 2001. “Transition in Post-Communist States: Triple or Quadruple?” Politics 21, no. 3 (September), 168–77. Kynev, A. V., and A. E. Liubarev. 2011. Partii i Vybory v Sovremennoi Rossii: Evoliutsiia i Devoliutsiia (Moscow: Novoe Literaturnoe Obozrenie). Laakso, Markku, and Rein Taagepera. 1979. “вЂEffective’ Number of Parties: A Measure with Application to West Europe,” Comparative Political Studies 12, no. 1 (April), 3–27. Laitin, David D. 2000. “Post-Soviet Politics,” Annual Review of Political Science 3 (June), 117–48. Langston, Joy. 2006. “Elite Ruptures: When Do Ruling Parties Split?” in Andreas Schedler, ed., Electoral Authoritarianism: The Dynamics of Unfree Competition, 57–76 (Boulder, CO: Lynne Rienner). Lankina, Tomila. 2003. “Federal, Regional Interests Shape Local Reforms,” Russian Regional Report 8, no. 18 (September 29), 4–9. Lankina, Tomila. 2004. Governing the Locals: Local Self-Government and Ethnic Mobilization in Russia (Lanham, MD: Rowman & Littlefield). Lankina, Tomila. 2012. “Unbroken Links? From Imperial Human Capital to Post-Communist Modernisation, ” Europe-Asia Studies 64, no. 4 (May), 623–43. Lankina, Tomila V., and Lullit Getachew. 2006. “A Geographic Incremental Theory of Democratization: Territory, Aid, and Democracy in Postcommunist Regions,” World Politics 58, no. 4 (July), 536–82. Lankina, Tomila, and Lullit Getachew. 2008. “The Virtuous Circles of Western Exposure in Russian Regions: A Case for Micro-Polity Analysis of Democratic Change,” Journal of Communist Studies and Transition Politics 24, no. 3 (September), 338–64. Lapidus, Gail W. 1999. “Asymmetrical Federalism and State Breakdown in Russia,” Post-Soviet Affairs 15, no. 1, 74–82. Ledeneva, Alena V. 2006. How Russia Really Works: The Informal Practices that Shaped Post-Soviet Politics and

Business (Ithaca: Cornell University Press). Lehmann, Susan Goodrich. 1997. “Islam and Ethnicity in the Republics of Russia,” Post-Soviet Affairs 13, no. 1 (January–March), 78–103. Leip, Dave. 2012. Atlas of U.S. Election Results, http://uselectionatlas.org/RESULTS/data.php? year=2012&datatype=national&def=1&f=0&off=0&elect=0,Page 251 → accessed December 30, 2012. Lerner, Daniel. 1958. The Passing of Traditional Society: Modernizing the Middle East (Glencoe, IL: Free Press). Levada Analytical Center. 2012. President’s Performance in Office—Trends, Russia Votes (Centre for the Study of Public Policy, University of Strathclyde, Scotland), http://www.russiavotes.org/xlfiles/file190.xls, accessed March 24, 2012. Levitsky, Steven, and Lucan Way. 2002. “The Rise of Competitive Authoritarianism,” Journal of Democracy 13, no. 2 (April), 51–65. Levitsky, Steven, and Lucan Way. 2010. Competitive Authoritarianism: Hybrid Regimes After the Cold War (New York: Cambridge University Press). Levitsky, Steven R., and Lucan A. Way. 2012. “Beyond Patronage: Violent Struggle, Ruling Party Cohesion, and Authoritarian Durability,” Perspectives on Politics 10, no. 4 (December), 869–89. Lewis-Beck, Michael. 1998. Economics and Elections: The Major Western Democracies (Ann Arbor: University of Michigan Press). Lijphart, Arend. 2012. Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries, 2nd ed. (New Haven: Yale University Press). Lindberg, Staffan I. 2006. Democracy and Elections in Africa (Baltimore: Johns Hopkins University Press). Lindberg, Staffan I., ed. 2009. Democratization by Elections: A New Mode of Transition (Baltimore: Johns Hopkins University Press). Linz, Juan. 1975. “Totalitarian and Authoritarian Regimes,” in Fred I. Greenstein and Nelson W. Polsby, eds., Handbook of Political Science, vol. 3, 175–411 (Reading, MA: Addison-Wesley). Linz, Juan J., and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation: Southern Europe, South America and Post-Communist Europe (Baltimore: Johns Hopkins University Press). Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic Development and Political Legitimacy,” American Political Science Review 53, no. 1, 69–105. Lipset, Seymour Martin. 1960. Political Man: The Social Bases of Politics (Baltimore: Johns Hopkins University Press). Liubarev, A. E. 2003. “Golosovanie вЂProtiv Vsekh’: Motivy i Tendentsii,” POLIS: Politicheskie Issledovaniia, no. 6, 104–13. Liubarev, A. E. 2004. Konstitutsionno-Pravovye Osnovy Mnogoobraziia Izbiratel’nykh Sistem v Rossiisskoi Federatsii. A thesis prepared in partial fulfillment of the requirements for the Candidate of Juridical Sciences degree (Moscow: Moscow Humanities University), http://lyubarev.narod.ru/elect/diss/diss_ogl.htm, accessed July 2, 2015. Lo, Bobo, and Lilia Shevtsova. 2012. A 21st Century Myth: Authoritarian Modernization in Russia and China

(Moscow: Carnegie Endowment for International Peace). Lowenhardt, John. 1997. “The 1996 Presidential Elections in Tatarstan,” Journal of Communist Studies and Transition Politics 13, no. 1 (March), 132–44. Lukinova, Evgeniya, Mikhail Myagkov, and Peter C. Ordeshook. 2011. “Metastasized Fraud in Russia’s 2008 Presidential Election,” Europe-Asia Studies 63, no. 4 (June), 603–21. Page 252 →Lynch, Allen C. 2005. How Russia Is Not Ruled: Reflections of Russian Political Development (New York: Cambridge University Press). Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico (New York: Cambridge University Press). Magaloni, Beatriz, and Ruth Kricheli. 2010. “Political Order and One-Party Rule,” Annual Review of Political Science 13, no. 1 (May), 123–43. Mahdavy, Hussein. 1970. “The Patterns and Problems of Economic Development in Rentier States: The Case of Iran,” in M. A. Cook, ed., Studies in the Economic History of the Middle East: From the Rise of Islam to the Present Day, 428–57 (London: Oxford University Press). Mainwaring, Scott, and Anibal Perez-Linan. 2013. Democracies and Dictatorships in Latin America: Emergence, Survival, and Fall (New York: Cambridge University Press). Makarenko, Boris I. 2012. “The Post-Soviet Party of Power: United Russia in Comparative Context,” Russian Politics and Law 50, no. 1 (January-February), 54–83. Maksimov, Andrei. 1999. “Chistie” i “Griaznie” Tekhnologii Vyborov (Rossiiskii Opyt) (Moscow: Mordodey & Imagemakers). Malinova, O. Iu. 2001. “Partiinye Ideologii v Rossii: Atribut ili Anturazh?” POLIS: Politicheskie Issledovaniia, no. 5 (October), 75–96. Malone, Mary Fran T. 2011. “Explaining Democratization,” in Mary Fran T. Malone, ed., Achieving Democracy: Democratization in Theory and Practice, 61–84 (New York: Continuum). March, Luke. 2002. The Communist Party in Post-Soviet Russia (New York: Palgrave). Marsh, Christopher. 2002. Russia at the Polls: Voters, Elections, and Democratization (Washington, DC: CQ Press). Marshall, Monty G., Ted Robert Gurr, and Keith Jaggers. 2015. Polity IV Project: Political Regime Characteristics and Transitions, 1800–2014 (Vienna, VA, dataset available from Center for Systemic Peace, http://www.systemicpeace.org/inscrdata.html), accessed August 14, 2015. Maslov, Dmitry. 2012. “Internal Contradictions in Russian Reforms,” Russian Analytical Digest, no. 118 (October), 11–13. Matsuzato, Kimitaka. 2004. “Regional Politics and Municipal Building: The Reshuffling of Local Chief Administrators in Russia, 1990–1996,” in Alfred B Evans Jr. and Vladimir Gel’man, eds., The Politics of Local Government in Russia, 169–202 (Lanham, MD: Rowman & Littlefield). McAllister, Ian, and Stephen White. 1998. “To Vote or Not To Vote: Election Turnout in Postcommunist Russia,” in Matthew Wyman et al., eds., Elections and Voters in Postcommunist Russia, 15–39 (London: Cass).

McAllister, Ian, and Stephen White. 2008. “Voting вЂAgainst All’ in Postcommunist Russia,” Europe-Asia Studies 60, no. 1 (January), 67–87. McAuley, Mary. 1997. Russia’s Politics of Uncertainty (New York: Cambridge University Press). McDonald, Michael. 2011. 2004 General Election Turnout Rates (United States Election Project), http://elections.gmu.edu/Turnout_2004G.html, accessed July 29, 2012. McDonald, Michael. 2012. 2012 General Election Turnout Rates (United States Page 253 →Election Project), http://elections.gmu.edu/Turnout_2012G.html, accessed December 30, 2012. McFaul, Michael, and Nikolai Petrov, eds. 1995. Previewing Russia’s 1995 Parliamentary Elections (Moscow: Carnegie Endowment for International Peace). McFaul, Michael, and Nikolai Petrov, eds. 1998. Politicheskii Al’manakh Rossii, 1989–1997 gg. (Washington, DC: Carnegie Endowment for International Peace). McFaul, Michael, Nikolai Petrov, and Andrei Ryabov, eds. 1999. Primer on Russia’s 1999 Duma Elections (Moscow: Carnegie Endowment for International Peace). McFaul, Michael, and Kathryn Stoner-Weiss. 2008. “The Myth of the Authoritarian Model: How Putin’s Crackdown Holds Russia Back,” Foreign Affairs 87, no. 1 (January–February), 68–84. Medvedev, Dmitrii. 2010. Stenograficheskii Otchet o Zasedanii Soveta po Sodeistviiu Razvitiiu Institutov Grazhdanskogo Obshchestva i Pravam Cheloveka (Moscow: President of Russia Official Web Portal), http://news.kremlin.ru/transcripts/7792, accessed May 24, 2010. Melville, Andrei, and Mikhail Mironyuk. 2016. “вЂBad Enough Governance’: State Capacity and Quality of Institutions in Post-Soviet Autocracies,” Post-Soviet Affairs 32, no. 2, 132–51. Merkel, Wolfgang. 2004. “Embedded and Defective Democracies,” Democratization 11, no. 1 (December), 33–58. Metcalf, Lee Kendall. 1996. “Presidential Power in the Russian Constitution,” Journal of Transnational Law and Policy 6, no. 1, 125–42. Mickiewicz, Ellen. 2008. Television, Power, and the Public in Russia (New York: Cambridge University Press). Mikhailov, Valentin. 2004. “Regional Elections and Democratization in Russia,” in Cameron Ross, ed., Russian Politics Under Putin, 198–220 (Manchester, UK: Manchester University Press). Milbrath, Lester. 1965. Political Participation (Chicago: Rand McNally). Miller, Arthur H., et al. 2000. “Emerging Party Systems in Post-Soviet Societies: Fact or Fiction?” Journal of Politics 62, no. 2 (May), 455–90. Miller, William L., Stephen White, and Paul Heywood. 1998. Values and Political Change in Postcommunist Europe (New York: St. Martin’s Press). Mills, Melinda. 2011. Introducing Survival and Event History Analysis (Thousand Oaks, CA: Sage). Minchenko, Evgenii N. 2001. Kak Stat’ i Ostat’siia Gubernatorom (Cheliabinsk: Ural LTD). Miroshnichenko, A. A. 2003. Vybory: Ot Zamysli do Pobedy (Predvybornaia Kampaniia v Rossiiskom Regione) [Elections: From Conception to Victory (Campaigning in a Russian Region)] (Moscow: Tsentr Publishing). Moeller, Joergen, and Svend-Erik Skaaning. 2013. “Autocracies, Democracies, and the Violation of Civil

Liberties,” Democratization 20, no. 1 (January), 82–106. Moeller, Jorgen, and Svend-Erik Skaaning. 2010. “Beyond the Radial Delusion: Conceptualizing and Measuring Democracy and Non-Democracy,” International Political Science Review 31, no. 3 (June), 261–84. Molinar, Juan. 1991. “Counting the Number of Parties: An Alternative Index,” American Political Science Review 85, no. 4 (December), 1383–91. Monroe, Alan D. 1977. “Urbanism and Voter Turnout: A Note on Some UnexpectedPage 254 → Findings, ” American Journal of Political Science 21, no. 1 (February), 71–78. Moraski, Bryon J. 2003. “Electoral System Design in Russian Oblasti and Republics: A Four Case Comparison,” Europe-Asia Studies 55, no. 3 (May), 437–68. Moraski, Bryon J. 2006a. Elections by Design: Parties and Patronage in Russia’s Regions (DeKalb: University of Northern Illinois Press). Moraski, Bryon J. 2006b. “Mandating Party Development in the Russian Federation: Effects of the 2001 Party Law,” Journal of Elections, Public Opinion and Parties 16, no. 3 (October), 199–219. Moraski, Bryon J. 2006c. “Prospects for Professional Parliaments in Russia’s Regions,” Journal of Communist Studies and Transition Politics 22, no. 2 (June), 135–61. Moraski, Bryon J., and William M. Reisinger. 2003. “Explaining Electoral Competition Across Russia’s Regions,” Slavic Review 62, no. 2 (Summer), 278–301. Moraski, Bryon J., and William M. Reisinger. 2007. “Eroding Democracy—Federal Intervention in Russia’s Gubernatorial Elections,” Democratization 14, no. 4 (August), 603–21. Moser, Robert G. 1999. “Independents and Party Formation: Elite Partisanship as an Intervening Variable in Russian Politics,” Comparative Politics 31, no. 2 (January), 147–65. Moser, Robert G. 2001. Unexpected Outcomes: Electoral Systems, Political Parties, and Representation in Russia (Pittsburgh: University of Pittsburgh Press). Moses, Joel C. 1974. Regional Party Leadership and Policy-Making in the USSR (New York: Praeger). Moses, Joel C. 1985. “Regionalism in Soviet Politics: Continuity as a Source of Change, 1953–1982,” Soviet Studies 37, no. 2 (April), 184–211. Moses, Joel C. 1992. “Soviet Provincial Politics in an Era of Transition and Revolution, 1989–91,” Soviet Studies 44, no. 3, 479–509. Moses, Joel C. 2002. “Political-Economic Elites and Russian Regional Elections, 1999–2000,” EuropeAsia Studies 54, no. 6 (September), 905–31. Moses, Joel C. 2003. “Voting, Regional Legislatures and Electoral Reform in Russia,” Europe-Asia Studies 55, no. 7 (November), 1049–75. Mozaffar, S., and R. Vengroff. 2002. “A вЂWhole System’ Approach to the Choice of Electoral Rules in Democratizing Countries: Senegal in Comparative Perspective,” Electoral Studies 21, no. 4 (December), 601–16. Munck, Gerardo L. 2004. “Tools for Qualitative Research,” in Henry E. Brady and David Collier, eds.,

Rethinking Social Inquiry: Diverse Tools, Shared Standards, 105–21 (Lanham, MD: Rowman & Littlefield). Myagkov, Mikhail, Peter Ordeshook, and Dimitri Shakin. 2009. The Forensics of Election Fraud: With Applications to Russia and Ukraine (New York: Cambridge University Press). Natsional’naia Sluzhba Novostei. 1995. Itogi Vyborov 17 Dekabria 1995 goda po Federal’nym Spiskam (Moscow: Natsional’naia Sluzhba Novostei), http://www.nns.ru/elects/izbobyed/i95f02al.html, accessed February 12, 1997. Neto, Octavio Amorim, and Gary W. Cox. 1997. “Electoral Institutions, Cleavage Structures and the Number of Parties,” American Journal of Political Science 41, no. 1 (January), 149–74. Page 255 →Nie, Norman H., G. Bingham Powell, Jr., and Kenneth Prewitt. 1969. “Social Structure and Political Participation: Developmental Relationships, Part I,” American Political Science Review 63, no. 2 (June), 361–78. Niemi, Richard G., and John Fuh-sheng Hsieh. 2002. “Counting Candidates: An Alternative to the Effective N: With an Application to the M + 1 Rule in Japan,” Party Politics 8, no. 1 (January), 75–99. Nikolaev, A. N. 2000. “Administrativnyi Resurs v Regional’nykh Izbiratel’nykh Kampaniiakh,” POLIS: Politicheskie Issledovaniia, no. 4, 171–74. Oates, Sarah. 2003. “Television, Voters, and the Development of the вЂBroadcast Party,’” in Vicki L. Hesli and William M. Reisinger, eds., The 1999–2000 Elections in Russia: Their Impact and Legacy, 29–50 (New York: Cambridge University Press). Oates, Sarah. 2006. Television, Democracy and Elections in Russia (New York: Routledge). O’Donnell, Guillermo, and Philippe C. Schmitter. 1986. “Tentative Conclusions About Uncertain Democracies,” in Guillermo O’Donnell et al., eds., Transitions From Authoritarian Rule: Prospects for Democracy, Part IV, 1–72 (Baltimore: Johns Hopkins University Press). O’Loughlin, John. 2000. “Geography as Space and Geography as Place: The Divide Between Political Science and Political Geography Continues,” Geopolitics 5, no. 3 (Winter), 126–37. O’Loughlin, John. 2003. “Spatial Analysis in Political Geography,” in John Agnew et al., eds., A Companion to Political Geography, 30–46 (Oxford: Basil Blackwell). O’Loughlin, John, et al. 1998. “The Diffusion of Democracy, 1946–1994,” Annals, Association of American Geographers 88, no. 4, 545–74. Olson, Mancur. 1993. “Dictatorship, Democracy and Development,” American Political Science Review 87, no. 3 (September), 567–76. Orenburg Oblast. 1995. “O Vyborakh Glavy Ispolnitel’noi Vlasti Orenburgskoi Oblasti,” http://www.vybory.ru/zakonregion/orenburg/zakonregion.php3?55.TXT, accessed July 2, 2015. Oreshkin, D. B. 2001. “Geografiia Elektoral’noi Kult’tury i Tsel’nost’ Rossii,” POLIS: Politicheskie Issledovaniia, no. 1, 73–93. Orlov Oblast’. 1997. “O Vyborakh Glavy Administratsii Orlovskoi Oblasti,” http://docs.cntd.ru /document/974201332, accessed July 2, 2015. Orttung, Robert W. 2000. The Republics and Regions of the Russian Federation: A Guide to Politics, Policies and Leaders (Armonk, NY: M. E. Sharpe).

Orttung, Robert W. 2004. “Key Issues in the Evolution of the Federal Okrugs and Center-Periphery Relations Under Putin,” in Peter Reddaway and Robert W. Orttung, eds., The Dynamics of Russian Politics: Putin’s Reform of Federal-Regional Relations, vol. 1, 19–52 (Lanham, MD: Rowman & Littlefield). Ostrovskaia, Natal’ia. 2008. “вЂBabai’ na Amure: Chto Stoit za Cheredoi Gromkikh Skandalov v Okruzhenii Glavy Amurskoi Oblasti,” Komsomol’skaia Pravda (May 20), http://www.kp.ru/daily/24100 /327105/, accessed February 12, 2012. Oversloot, Hans, Joop van Holsteyn, and Ger P. van den Berg. 2002. “Against All: Exploring the Vote вЂAgainst All’ in the Russian Federation’s Electoral System,” Journal of Communist Studies and Transition Politics 18, no. 4 (December), 31–50. Pacek, Alexander C. 1994. “Macroeconomic Conditions and Electoral Politics in Page 256 →East Central Europe,” American Journal of Political Science 38, no. 3 (August), 723–44. Pacheco, Julianna. 2012. “The Social Contagion Model: Exploring the Role of Public Opinion on the Diffusion of Anti-Smoking Legislation across the American States,” Journal of Politics 74, no. 1 (January), 187–202. Pain, Emil. 2011. “The Political Regime in Russia in the 2000s: Special Features, Inherited and Acquired,” Russian Politics and Law 49, no. 3 (May-June), 7–28. Palmer, Harvey D., and Guy D. Whitten. 2000. “Government Competence, Economic Performance and Endogenous Election Dates,” Electoral Studies 19, no. 2/3, 413–26. Panorama Informational-Expert Group. 2013. Izbiratel’noe Zakonodatel’stvo Rossiiskikh Regionov, http://www.panorama.ru/works/izbir, accessed July 16, 2013. Panov, Petr, and Cameron Ross. 2013. “Patterns of Electoral Contestation in Russian Regional Assemblies: Between вЂCompetitive’ and вЂHegemonic’ Authoritarianism,” unpublished manuscript (Dundee: University of Dundee, March). Paramonov, D. O., and V. V. Kirichenko. 2007. Metody Fal’sifikatsii Vyborov [Methods of Falsifying Election Results] (Moscow: Fictionbook.ru). Pateman, Carole. 2012. “Participatory Democracy Revisited,” Perspectives on Politics 10, no. 1 (March), 7–19. Pavlov, Oleg. 2010. “A Lesson for Luzhkov? Tatarstan’s Shaimiev Shows How to Cling on to Power,” o-DRussia (October 8), http://www.opendemocracy.net/od-russia/oleg-pavlov/lesson-for-luzhkov-tatarstan%E2% 80%99s-shaimiev-shows-how-to-cling-on-to-power, accessed June 26, 2011. Paxton, Pamela, Melanie M. Hughes, and Matthew A. Painter. 2010. “Growth in Women’s Political Representation: A Longitudinal Exploration of Democracy, Electoral System and Gender Quotas,” European Journal of Political Research 49, no. 1 (January), 25–52. Perez-Linan, Anibal. 2001. “Neoinstitutional Accounts of Voter Turnout: Moving Beyond Industrial Democracies,” Electoral Studies 20, no. 2 (June), 281–97. Permskii Krai. 2007. “Current Events” (December 3), http://www.perm.ru/eng/events/event/?id=344, accessed February 14, 2008. Petrov, Nikolai. 2005. “How Have the Presidential Envoys Changed the Administrative-Political Balance of Putin’s Regime?” in Peter Reddaway and Robert W. Orttung, eds., The Dynamics of Russian Politics: Putin’s Reform of Federal-Regional Relations, vol. 2, 33–64 (Lanham, MD: Rowman & Littlefield).

Petrov, Nikolai V., and Aleksei Titkov. 2008. Ekspertnaia Otsenka (Reiting) Demokratichnosti Regionov za 1991–2004 gg. (Moscow: Independent Institute for Social Policy), http://atlas.socpol.ru/indexes /index_democr.shtml, accessed August 27, 2008. Petrov, Nikolay. 2008. “Moscow Losing Clout in the Regions,” Moscow Times (October 28), http://carnegieendowment.org/2008/10/28/moscow-losing-clout-in-regions/1bie, accessed October 28, 2008. Pevehouse, Jon C. 2002. “Democracy from the Outside-In? International Organizations and Democratization “ International Organization 56, no. 3 (August), 515–49. Phillips, Timothy. 2008. Beslan: The Tragedy of School No. 1 (London: Granta Books). Pierson, Paul. 2004. Politics in Time: History, Institutions, and Social Analysis (Princeton: Princeton University Press). Page 257 →Pomeranz, William E. 2010. “Legal Reform Through the Eyes of Russia’s Leading Jurists: The Vlast Debate on the Russian Judiciary,” Problems of Post-Communism 57, no. 3 (May-June), 3–10. Powell, G. Bingham. 1982. Contemporary Democracies: Participation, Stability and Violence (Cambridge, MA: Harvard University Press). Powell, G. Bingham, Jr. 2000. Elections as Instruments of Democracy: Majoritarian and Proportional Visions (New Haven: Yale University Press). Power, Timothy J., and J. Timmons Roberts. 1995. “Compulsory Voting, Invalid Ballots, and Abstention in Brazil,” Political Research Quarterly 48, no. 4 (December), 795–826. Przeworski, Adam. 1999. “Minimalist Conception of Democracy: A Defense,” in Ian Shapiro and Casiano Hacker-Cordon, eds., Democracy’s Value, 23–55 (New York: Cambridge University Press). Przeworski, Adam, et al. 2000. Democracy and Development: Political Institutions and Well-Being in the World, 1950–1990 (New York: Cambridge University Press). Przeworski, Adam, and Fernando Limongi. 1997. “Modernization: Theories and Facts,” World Politics 49, no. 2 (January), 155–83. Przeworski, Adam, and Henry Teune. 1970. The Logic of Comparative Social Inquiry (New York: Wiley). Putnam, Robert D. 1993. Making Democracy Work: Civic Traditions in Modern Italy (Princeton: Princeton University Press). Radcliff, Benjamin. 1992. “The Welfare State, Turnout, and the Economy: A Comparative Analysis,” American Political Science Review 86, no. 2 (June), 444–54. Radio Free Europe/Radio Liberty. 2001. “2000 Gubernatorial Races,” Russian Federation Report 3, no. 1 (January 3), http://www.rferl.org/content/article/1344476.html, accessed April 26, 2016. Radio Free Europe/Radio Liberty. 2003. “Putin to Governors: Take as Many Terms as You Can Swallow,” Russian Federation Report 5, no. 1 (January 3), http://www.rferl.org/content/article/1344334.html, accessed April 26, 2016. Rae, Douglas W. 1967. The Political Consequences of Electoral Laws (New Haven: Yale University Press). Regnum News Agency. 2007. “Voting Ends in Kamchatka” (December 2), http://www.regnum.ru/english /925008.html, accessed February 14, 2008.

Reisinger, William M. 1997. “Establishing and Strengthening Democracy,” in Robert D. Grey, ed., Democratic Theory and Post-Communist Change, 52–78 (New York: Prentice-Hall). Reisinger, William M. 2003. “The 1999–2000 Elections and Russia’s Prospects for Democracy,” in Vicki L. Hesli and William M. Reisinger, eds., The 1999–2000 Elections in Russia: Their Impact and Legacy, 261–74 (New York: Cambridge University Press). Reisinger, William M. 2013. “Studying Russia’s Regions to Advance Comparative Political Science,” in William M. Reisinger, ed., Russia’s Regions and Comparative Subnational Politics, Routledge Research in Comparative Politics, 1–24 (New York: Routledge). Reisinger, William M., et al. 1994. “Political Values in Russia, Ukraine and Lithuania: Sources and Implications for Democracy,” British Journal of Political Science 45, no. 2 (April), 183–223. Page 258 →Reisinger, William M., and Bryon J. Moraski. 2010. “Regional Changes and Changing Regional Relations with the Centre,” in Vladimir Gel’man and Cameron Ross, eds., The Politics of Sub-National Authoritarianism in Russia, 67–84 (Aldershot, UK: Ashgate). Reiter, Dan, and Allan C. Stam. 2002. Democracies at War (Princeton: Princeton University Press). Remington, Thomas F. 2003a. “Majorities without Mandates: The Russian Federation Council since 2000,” Europe-Asia Studies 55, no. 5 (July), 667–91. Remington, Thomas F. 2003b. “Putin, the Duma, and Political Parties,” in Dale R. Herspring, ed., Putin’s Russia: Past Imperfect, Future Uncertain, 39–59 (Lanham, MD: Rowman & Littlefield). Remington, Thomas F. 2008. Politics in Russia, 5th ed. (New York: Pearson Longman). Remington, Thomas F. 2011. The Politics of Inequality in Russia (New York: Cambridge University Press). Remnick, David. 1997. Resurrection: The Struggle for a New Russia (New York: Random House). Reuter, Ora John. 2010. “The Politics of Dominant Party Formation: United Russia and Russia’s Governors,” Europe-Asia Studies 62, no. 2 (March), 293–327. Reuter, Ora John, and Thomas F. Remington. 2009. “Dominant Party Regimes and the Commitment Problem: The Case of United Russia,” Comparative Political Studies 42, no. 4 (April), 501–26. Reuter, Ora John, and Graeme Robertson. 2012. “Subnational Appointments in Authoritarian Regimes: Evidence from Russian Gubernatorial Appointments,” Journal of Politics 74, no, 4 (October), 1023–37. Richardson, Paul E. 2009. “100 Things Everyone Should Know About Russia,” Russian Life 52, no. 3 (May–June), 29–44. Roberts, Sean P. 2012a. Putin’s United Russia Party (New York: Routledge). Roberts, Sean P. 2012b. “United Russia and the Dominant-Party Framework: Understanding the Russian Party of Power in Comparative Perspective,” East European Politics 28, no. 3 (September), 1–16. Robinson, Neil. 2007. “The Political Is Personal: Corruption, Clientelism, Patronage, Informal Practices and the Dynamics of Post-Communism,” Europe-Asia Studies 59, no. 7 (November), 1217–24. Robinson, Neil. 2012. “Institutional Factors and Russian Political Parties: The Changing Needs of Regime Consolidation in a Neo-patrimonial System,” East European Politics 28, no. 3 (September), 298–309. Roeder, Philip G. 1989. “Electoral Avoidance in the Soviet Union,” Soviet Studies 41, no. 3 (July),

462–83. Roeder, Philip G. 1994. “Varieties of Post-Soviet Authoritarian Regimes,” Post-Soviet Affairs 10, no. 1, 61–101. Roeder, Philip G. 2007. Where Nation-States Come From: Institutional Change in the Age of Nationalism (Princeton: Princeton University Press). Rose, Richard. 2000. “How Floating Parties Frustrate Democratic Accountability: A Supply-Side View of Russia’s Elections,” East European Constitutional Review 9, nos. 1–2 (Winter-Spring), 53–59. Rose, Richard, William Mishler, and Neil Munro. 2006. Russia Transformed: Developing Popular Support for a New Regime (New York: Cambridge University Press). Page 259 →Rose, Richard, William Mishler, and Neil Munro. 2011. Popular Support for an Undemocratic Regime: The Changing Views of Russians (New York: Cambridge University Press). Rosenstone, Steven, and John Hansen. 1993. Mobilization, Participation and Democracy in America (New York: Macmillan). Rosenstone, Steven J. 1982. “Economic Adversity and Voter Turnout,” American Journal of Political Science 26, no. 1 (February), 25–46. Ross, Cameron. 2000. “Federalism and Democratization in Russia,” Communist and Post-Communist Studies 33, 403–20. Ross, Cameron. 2002. Federalism and Democratisation in Russia (Manchester, UK: Manchester University Press). Ross, Cameron. 2005. “Federalism and Electoral Authoritarianism under Putin,” Demokratizatsiya 13, no. 3 (Summer), 347–71. Ross, Cameron. 2010. “Federalism and Inter-governmental Relations in Russia,” Journal of Communist Studies and Transition Politics 26, no. 2 (June), 165–87. Ross, Cameron. 2011. “Regional Elections and Electoral Authoritarianism in Russia,” Europe-Asia Studies 63, no. 4 (June), 641–61. Ross, Michael L. 2001. “Does Oil Hinder Democracy?” World Politics 53, no. 3 (April), 325–61. Rossiiskaia Federatsiia. 2002 [2014]. Federal’nyi Zakon ob Osnovnykh Garantiiakh Izbiratel’nykh Prav i Prava na Uchastie v Referendume Grazhdan Rossiiskoi Federatsii (Moscow: Gosudarstvennaia Sistema Pravovoi Informatsii, December 1, 2014), accessed at http://pravo.gov.ru/proxy/ips/?docbody=&nd=102076507. Rossiiskaia Federatsiia. 2014. Federal’nyi Zakon o Vyborakh Deputatov Gosudarstvennoi Dumi Federal’nogo Sobraniia Rossiiskoi Federatsii (Moscow: Gosudarstvennaia Sistema Pravovoi Informatsii, February 22), accessed at http://pravo.gov.ru:8080/page.aspx?88303 Rothstein, Bo. 2011. The Quality of Government: Corruption, Social Trust, and Inequality in International Perspective (Chicago: University of Chicago Press). Royston, Patrick. 1991. “Sg3.5: Comment on sg3.4 and an Improved D’Agostino Test,” Stata Technical Bulletin, no. 3, 23–24, http://www.stata.com/products/stb/journals/stb3.pdf, accessed June 21, 2013. Ruger, J. P. 2005. “Democracy and Health,” QJM 98, no. 4 (April), 299–304.

Rulers.org. 2013. Russian Administrative Divisions, http://rulers.org/russdiv.html, accessed January 21, 2013. Rummel, R. J. 1998. Statistics of Democide: Genocide and Mass Murder Since 1900 (Muenster, Germany: LIT Verlag). Russett, Bruce M. 1964. World Handbook of Political and Social Indicators (New Haven: Yale University Press). Russian Federal State Statistics Service. 1996. Rossiiskii Statisticheskii Ezhegodnik: Statisticheskii Sbornik (Moscow: Goskomstat Rossii). Russian Federal State Statistics Service. 2005. Osnovnye Itogi Vserossiiskoi Perepisi Naselenia 2002 g., vol. 14 (Moscow: Rosstat). Russian Federal State Statistics Service. 2007. Regiony Rossii, Sotsial’no-Ekonomicheskie Pokazateli 2006: Statisticheskii Sbornik (Moscow: Rosstat). Russian Federal State Statistics Service. 2012. Vserossiiskaia Perepis’ Naseleniia 2010 g., (Moscow: Rosstat), http://www.perepis-2010.ru, accessed August 2, 2012. Page 260 →Russian Federal State Statistics Service. 2014. Regiony Rossii, Sotsial’no-Ekonomicheskie Pokazateli 2011: Statisticheskii Sbornik (Moscow: Rosstat). Russian Federal State Statistics Service. various years. Regiony Rossii: Informatsionno-Statisticheskii Sbornik (Moscow: Rosstat). Rustow, Dankwart A. 1970. “Transitions to Democracy: Toward a Dynamic Model,” Comparative Politics 2, no. 3 (April), 337–63. Rutland, Peter. 2005. “Putin’s Economic Record,” in Stephen White et al., eds., Developments in Russian Politics 6, 186–203 (Durham, NC: Duke University Press). Rutland, Peter. 2008a. “Putin’s Economic Record: Is the Oil Boom Sustainable?” Europe-Asia Studies 60, no. 6 (August), 1051–72. Rutland, Peter. 2008b. “Russia as an Energy Superpower,” New Political Economy 13, no. 2 (June), 203–10. Saikkonen, Inga A. L. 2016. “Variation in Subnational Electoral Authoritarianism: Evidence from the Russian Federation,” Democratization 23, no. 3, 437–58. Sakwa, Richard. 1995. “The Russian Elections of December 1993,” Europe-Asia Studies 47, no. 2 (March), 195. Sakwa, Richard. 2002. “Federalism, Sovereignty and Democracy,” in Cameron Ross, ed., Regional Politics in Russia, 1–22 (Manchester, UK: Manchester University Press). Sakwa, Richard. 2005. “The 2003–2004 Russian Elections and Prospects for Democracy,” Europe-Asia Studies 57, no. 3 (May), 369–98. Sakwa, Richard. 2011. The Crisis of Russian Democracy: The Dual State, Factionalism and the Medvedev Succession (New York: Cambridge University Press). Salmore, Stephen A., and Barbara Salmore. 1996. “The Transformation of State Electoral Politics,” in Carl E. Van Horn, ed., The State of the States, 3rd ed., 51–76 (Washington, DC: CQ Press). Schedler, Andreas. 2006. Electoral Authoritarianism: The Dynamics of Unfree Competition (Boulder, CO: Lynne

Rienner). Schedler, Andreas. 2013. The Politics of Uncertainty: Sustaining and Subverting Electoral Authoritarianism (New York: Oxford University Press). Scheuch, Erwin K. 1969. “Social Context and Individual Behavior,” in Mattei Dogan and Stein Rokkan, eds., Quantitative Ecological Analysis in the Social Sciences, 133–55 (Cambridge, MA: MIT Press). Schulz, Ann. 1979. Local Politics and Nation-States: Case Studies in Politics and Policy (Santa Barbara, CA: Clio Books). Schumpeter, Joseph A. 1942 [2003]. Capitalism, Socialism and Democracy, electronic ed. (New York: Routledge). Scott, James C. 1972. Comparative Political Corruption (Englewood Cliffs, NJ: Prentice-Hall). Sharafutdinova, Gulnaz. 2007. “Why Was Democracy Lost in Russia’s Regions? Lessons from Nizhnii Novgorod,” Communist and Post-Communist Studies 40, no. 3 (September), 363–82. Sharafutdinova, Gulnaz. 2010a. “Redistributing Sovereignty and Property under Putin: A View from Resource-Rich Republics of the Russian Federation,” in Vladimir Gel’man and Cameron Ross, eds., The Politics of Sub-National Authoritarianism in Russia, 191–210 (Aldershot, UK: Ashgate), 191–210. Page 261 →Sharafutdinova, Gulnaz. 2010b. “Subnational Governance in Russia: How Putin Changed the Contract with His Agents and the Problems It Created for Medvedev,” Publius: The Journal of Federalism 40, no. 4 (October), 672–96. Sharafutdinova, Gulnaz. 2013. “Gestalt Switch in Russian Federalism: The Decline in Regional Power Under Putin,” Comparative Politics 45, no. 3 (April), 357–76. Sharavina, A. A., A. A. Khramchikhina, and I. S. Snegirevoi. 1996. Vybory Prezidenta Rossiiskoi Federatsii 1996 Goda: Itogi i Vyvody (Moscow: Institut Politicheskogo i Voennogo Analiza). Sharlet, Robert. 2003. “Constitutional Law and Politics in Russia: Surviving the First Decade,” Demokratizatsiya 11, no. 1 (Winter), 122–28. Shevtsova, Lilia. 2007. Russia: Lost in Transition (Washington, DC: Carnegie Endowment for International Peace). Shevtsova, Lilia. 2012. “The Next Russian Revolution,” Current History 111, no. 747 (October), 251–57. Shi, Tianjian. 1999. “Voting and Nonvoting in China: Voting Behavior in Plebiscitary and Limited-Choice Elections,” Journal of Politics 61, no. 4 (November), 1115–39. Shin, Doh C. 1989. “Political Democracy and the Quality of Citizens’ Lives: A Cross-National Study,” Journal of Developing Societies 5, 30–41. Shipan, Charles R., and Craig Volden. 2008. “The Mechanisms of Policy Diffusion,” American Journal of Political Science 52, no. 4 (October), 840–57. Shlapentokh, Vladimir, Roman Levita, and Mikhail Loiberg. 1997. From Submission to Rebellion: The Provinces Versus the Center in Russia (Boulder, CO: Westview Press). Shugart, Matthew Soberg, and John M. Carey. 1992. Presidents and Assemblies: Constitutional Design and

Electoral Dynamics (New York: Cambridge University Press). Shvetsova, Olga. 2003. “Resolving the Problem of Pre-Election Coordination: The Parliamentary Election as an Elite Presidential вЂPrimary,’” in Vicki L. Hesli and William M. Reisinger, eds., The 1999–2000 Elections in Russia: Their Impact and Legacy, 213–31 (New York: Cambridge University Press). Slater, Dan. 2010. Ordering Power: Contentious Politics and Authoritarian Leviathans in Southeast Asia (New York: Cambridge University Press). Slider, Darrell. 1994. “Federalism, Discord and Accommodation: Intergovernmental Relations in Post-Soviet Russia,” in Theodore H. Friedgut and Jeffrey Hahn, eds., Local Power and Post-Soviet Politics, 239–69 (Armonk, NY: M. E. Sharpe). Slider, Darrell. 1996. “Elections to Russia’s Regional Assemblies,” Post-Soviet Affairs 12, no. 3 (JulySeptember), 243–64. Slider, Darrell. 2001. “Russia’s Governors and Party Formation,” in Archie Brown, ed., Contemporary Russian Politics: A Reader, 224–34 (New York: Oxford University Press). Slider, Darrell. 2010. “How United Is United Russia? Regional Sources of Intra-Party Conflict,” Journal of Communist Studies and Transition Politics 26, no. 2 (June), 257–75. Slider, Darrell, Vladimir Gimpel’son, and Sergei Chugrov. 1994. “Political Tendencies in Russia’s Regions: Evidence from the 1993 Parliamentary Elections,” Slavic Review 53, no. 3 (Fall), 711–32. Page 262 →Smets, Kaat, and Carolien van Ham. 2013. “The Embarrassment of Riches? A Meta-analysis of Individual-Level Research on Voter Turnout,” Electoral Studies 32, no. 2 (June), 344–59. Smirnov, Valerii. 2008. Afery na Vyborakh (Moscow: Algoritm). Smirnov, William V. 2001. “Democratization in Russia: Achievements and Problems,” in Archie Brown, ed., Contemporary Russian Politics: A Reader, 517–29 (New York: Oxford University Press). Smith, Benjamin. 2007. Hard Times in the Lands of Plenty: Oil Politics in Iran and Indonesia (Ithaca: Cornell University Press). Smyth, Regina. 2006. Candidate Strategies and Electoral Competition in the Russian Federation: Democracy without Foundation (New York: Cambridge University Press). Smyth, Regina, Anna Lowry, and Brandon Wilkening. 2007. “Engineering Victory: Institutional Reform, Informal Institutions, and the Formation of a Hegemonic Party Regime in the Russian Federation,” Post-Soviet Affairs 23, no. 2 (April-June), 118–37. Sniderman, Paul M. 1975. Personality and Democratic Politics (Berkeley: University of California Press). Snyder, Richard. 2001. “Scaling Down: The Subnational Comparative Method,” Studies in Comparative International Development 36, no. 1 (Spring), 93–110. Sobianin, Aleksandr, Eduard Gel’man, and Ol’ga Kaiunov. 1994. “The Political Climate of Russia’s Regions: Voters and Deputies, 1991–1993,” The Soviet and Post-Soviet Review 21, no. 1, 63–84. Solnick, Steven L. 1995. “Federal Bargaining in Russia,” East European Constitutional Review 4, no. 4 (Fall), 52–58.

Solnick, Steven L. 1998. “Gubernatorial Elections in Russia, 1996–1997,” Post-Soviet Affairs 14, no. 1 (January-February), 48–80. Solnick, Steven L. 2010. “Authoritarian Legality and Informal Practices: Judges, lawyers and the State in Russia and China,” Communist and Post-Communist Studies 43, no. 4 (December), 351–62. Solomon, Peter H., Jr. 2008. “Law in Public Administration: How Russia Differs,” Journal of Communist Studies and Transition Politics 24, no. 1 (March), 115–35. Solov’ev, Vladimir, and Elena Klepikova. 1995. Zhirinovsky : Russian Fascism and the Making of a Dictator (Reading, MA: Addison-Wesley). Southwell, Priscilla L. 1988. “The Mobilization Hypothesis and Voter Turnout in Congressional Elections, 1974–1982,” Political Research Quarterly 41, no. 2 (June), 273–87. StataCorp. 2011. Stata Base Reference Manual, 12th ed. (College Station, TX: StataCorp LP). Stepan, Alfred. 2000. “Russian Federalism in Comparative Perspective,” Post-Soviet Affairs 16, no. 2 (April-June), 133–76. Stokes, Susan C., et al. 2013. Brokers, Voters, and Clientelism, the Puzzle of Distributive Politics (New York: Cambridge University Press). Stoliarov, Mikhail. 2003. Federalism and the Dictatorship of Power in Russia (New York: Routledge). Stoner-Weiss, Kathryn. 1999. “Central Weakness and Provincial Autonomy: Observations on the Devolution Process in Russia,” Post-Soviet Affairs 15, no. 1, 87–106. Page 263 →Stoner-Weiss, Kathryn. 2006. Resisting the State: Reform and Retrenchment in Post-Soviet Russia (New York: Cambridge University Press). Stoner-Weiss, Kathryn. 2010. “Comparing Oranges and Apples: The Internal and External Dimensions of Russia’s Turn Away from Democracy,” in Valerie Bunce et al., eds., Democracy and Authoritarianism in the Postcommunist World, 253–73 (New York: Cambridge University Press). Sugiyama, Natasha Borges. 2008. “Ideology and Networks: The Politics of Social Policy Diffusion in Brazil, ” Latin American Research Review 43, no. 3, 82–108. Sutela, Pekka. 2012. The Political Economy of Putin’s Russia (New York: Routledge). Swearer, Howard R. 1961. “The Functions of Soviet Local Elections,” Midwest Journal of Political Science 5, no. 2, 129–49. Szporluk, Roman. 2000. Russia, Ukraine, and the Breakup of the Soviet Union (Stanford, CA: Hoover Institution Press). Taagepera, Rein, and Matthew Soberg Shugart. 1989. Seats and Votes: The Effects and Determinants of Electoral Systems (New Haven: Yale University Press). Taylor, Brian D. 2011. State Building in Putin’s Russia: Policing and Coercion After Communism (New York: Cambridge University Press). Teorell, Jan. 2010. Determinants of Democratization: Explaining Regime Change in the World, 1972–2006 (New York: Cambridge University Press). Thorson, Carla. 2012. Politics, Judicial Review and the Russian Constitutional Court (New York: Palgrave

Macmillan). Tolz, Vera, and Irina Busygina. 1997. “Regional Governors and the Kremlin: The Ongoing Battle for Power, ” Communist and Post-Communist Studies 30, no. 4, 401–26. Treisman, Daniel S. 1997. “Russia’s вЂEthnic Revival’: The Separatist Activism of Regional Leaders in a Postcommunist Order,” World Politics 49, no. 2, 212–49. Treisman, Daniel. 1999. After the Deluge: Regional Crises and Political Consolidation in Russia (Ann Arbor: University of Michigan Press). Treisman, Daniel. 2007. The Architecture of Government: Rethinking Political Decentralization (New York: Cambridge University Press). Treisman, Daniel. 2011. The Return: Russia’s Journey from Gorbachev to Medvedev (New York: Free Press). Trenin, Dmitri. 2010. “Russia’s Conservative Modernization: A Mission Impossible?” SAIS Review 30, no. 1 (Winter-Spring), 27–37. Trenin, Dmitri, et al. 2012. “The Russian Awakening,” Carnegie Papers (Moscow: Carnegie Endowment for International Peace, November), http://carnegieendowment.org/files/russian_awakening.pdf, accessed November 20, 2012. Tucker, Joshua A. 2002. “The First Decade of Post-Communist Elections and Voting: What Have We Studied and How Have We Studied It?” Annual Review of Political Science 5, 271–304. Tucker, Joshua A. 2007. “Enough! Electoral Fraud, Collective Action Problems, and Post-Communist Colored Revolutions,” Perspectives on Politics 5, no. 3 (September), 535–51. Turovskii, Rostislav. 2009. “The Representation of Business Elites in Regional Politics: Etatism, Elitism and Clientelism,” in Cameron Ross and Adrian Campbell, eds., Federalism and Local Politics in Russia, 184–206 (New York: Routledge). Turovskii, Rostislav. 2010. “How Russian Governors Are Appointed: Inertia and Page 264 →Radicalism in Central Policy,” Russian Politics and Law 48, no. 1 (January-February), 58–79. Turovsky, Rostislav. 2007. “The Mechanism of Representation of Regional Interests at the Federal Level in Russia: Problems and Solutions,” Perspectives on European Politics and Society 8, no. 1 (April), 73–97. Urban, Joan Barth, and Valerii D. Solovei. 1997. Russia’s Communists at the Crossroads (Boulder, CO: Westview). US Energy Information Agency. 2013. “Russia,” Analysis Briefs (Washington, DC: November 26), accessed at http://www.eia.gov/countries/cab.cfm?fips=RS. USSR Central Statistical Administration. 1983. SSSR v Tsifrakh v 1982 godu (Moscow: Finansy i Statistika). USSR State Committee for Statistics. 1989. Narodnoe Khoziastvo SSSR v 1988 g. (Moscow: Finansy i Statistika). Van der Walle, Nicolas. 2006. “Tipping Games: When Do Opposition Parties Coalesce?” In Andreas Schedler, ed., Electoral Authoritarianism: The Dynamics of Unfree Competition, 77–92 (Boulder, CO: Lynne Rienner). van Ham, Carolien. 2015. “Getting Elections Right? Measuring Electoral Integrity,” Democratization 22, no. 4, 714–37.

Van Zon, Hans. 2008. Russia’s Development Problem: The Cult of Power (New York: Palgrave Macmillan). Vanhanen, Tatu. 1984. The Emergence of Democracy: A Comparative Study of 119 States, 1850–1979 (Helsinki: Finnish Society of Arts and Letters). Vanhanen, Tatu. 1990. The Process of Democratization: A Comparative Study of 147 States 1980–1988 (Bristol, PA: Crane and Rusak). Vanhanen, Tatu. 1997. Prospects of Democracy: A Study of 172 Countries (New York: Routledge). Varshney, Ashutosh. 2001. “Ethnic Conflict and Civil Society: India and Beyond,” World Politics 53, no. 3 (April), 362–98. Verba, Sydney, and Norman H. Nie. 1972. Participation in America (New York: Harper and Row). Veshniakov, A. A., and O. K. Zastrozhnaia, eds. 2004. Vybory Prezidenta Rossiiskoi Federatsii 2004: Elektoral’naia Statistika (Moscow: Ves’ Mir). Vorontsova, A. B. and V. B. Zvonovskii. 2003. “Administrativnii Resurs Kak Fenomen Rossiiskogo Izbiratel’nogo Protsessa,” POLIS: Politicheskie Issledovaniia, no. 6, 114–24. Wahman, Michael, Jan Teorell, and Axel Hadenius. 2013. “Authoritarian Regime Types Revisited: Updated Data in Comparative Perspective,” Contemporary Politics 19, no. 1 (March), 19–34. Walker, Edward W. 1996. “The Dog that Didn’t Bark: Tatarstan and Asymmetrical Federalism in Russia, ” The Harriman Review 9, no. 4 (Winter), 11–18. Wantchekon, Leonard. 2002. “Why Do Resource Dependent Countries Have Authoritarian Governments?” Journal of African Finance and Economic Development 5, no. 2, 57–77. Ware, Robert Bruce. 2005. “Recent Russian Federal Elections in Dagestan: Implications for Proposed Electoral Reform,” Europe-Asia Studies 57, no. 4 (June), 583–600. Way, Lucan. 2010. “Resistance to Contagion: Sources of Authoritarian Stability in Page 265 →the Former Soviet Union,” in Valerie Bunce et al., eds., Democracy and Authoritarianism in the Postcommunist World, 229–52 (New York: Cambridge University Press). Webb, Edward. 2011. “Totalitarianism and Authoritarianism,” in John T. Ishiyama and Marijke Breuning, eds., 21st Century Political Science: A Reference Handbook, 249–58 (Thousand Oaks, CA: Sage). Wedel, Janine R. 2005. “Flex Organizing and the Clan State: Perspectives on Crime and Corruption in the New Russia,” in William Alex Pridemore, ed., Ruling Russia: Law, Crime and Justice in a Changing Society, 101–16 (Lanham, MD: Rowman & Littlefield). Weiner, Myron, and Joseph LaPalombara. 1966. “The Impact of Parties on Political Development,” in Joseph LaPalombara and Myron Weiner, eds., Political Parties and Political Development, 399–435 (Princeton: Princeton University Press). Weinthal, Erika, and Paula Jones Luong. 2006. “The Paradox of Energy Sector Reform,” in Timothy J. Colton and Stephen Holmes, eds., The State After Communism: Governance in the New Russia, 225–60 (Lanham, MD: Rowman & Littlefield). White, Stephen. 2007. “Russia’s Client Party System,” in Paul Webb and Stephen White, eds., Party Politics in New Democracies, 21–51 (New York: Oxford University Press). White, Stephen. 2011a. “Elections Russian-Style,” Europe-Asia Studies 63, no. 4 (June), 531–56.

White, Stephen. 2011b. Understanding Russian Politics (New York: Cambridge University Press). White, Stephen, and Ian McAllister. 2004. “Dimensions of Disengagement in Post-Communist Russia,” Journal of Communist Studies and Transition Politics 20, no. 1 (March), 81–97. White, Stephen, Ian McAllister, and Yeongmi Yun. 2002. “Electoral Behavior in the Russian Far East, 1995–2000: Contextual and Compositional Effects,” Eurasian Geography and Economics 43, no. 2 (March), 143–59. White, Stephen, and Sarah Oates. 2003. “Politics and the Media in Postcommunist Russia,” Politics 23, no. 1 (February), 31–37. White, Stephen, Richard Rose, and Ian McAllister. 1997. How Russia Votes (Chatham, NJ: Chatham House Publishers). White, Stephen, Matthew Wyman, and Sarah Oates. 1997. “Parties and Voters in the 1995 Russian Duma Election,” Europe-Asia Studies 49, no. 5, 767–98. Whitefield, Stephen. 2005. “Putin’s Popularity and Its Implications for Democracy in Russia,” in Alex Pravda, ed., Leading Russia: Putin in Perspective, Essays in Honour of Archie Brown, 139–60 (New York: Oxford University Press). Whitmore, Sarah. 2010. “Parliamentary Oversight in Putin’s Neo-patrimonial State: Watchdogs or Showdogs?” Europe-Asia Studies 62, no. 6 (August), 999–1025. Wielhouwer, Peter W., and Brad Lockerbie. 1994. “Party Contacting and Political Participation, 1952–90, ” American Journal of Political Science 38, no. 1 (February), 211–29. Wilhelmsen, Julie. 2005. “Between a Rock and a Hard Place: The Islamisation of the Chechen Separatist Movement,” Europe-Asia Studies 57, no. 1, 35–59. Page 266 →Willerton, John P. 1990. “Reform, the Elite and Soviet Center-Periphery Relations,” Soviet Union/Union Sovietique 17, nos. 1–2, 55–94. Willerton, John P. 1992. Patronage and Politics in the USSR (New York: Cambridge University Press). Willerton, John P. 1998. “Post-Soviet Clientelist Norms at the Russian Federal Level,” in Graeme Gill, ed., Elites and Leadership in Russian Politics: Selected Papers from the Fifth World Congress of Central and East European Studies, Warsaw, 1995, 52–80 (London: Macmillan). Willerton, John P., and William M. Reisinger. 1991. “Troubleshooters, Political Machines and Moscow’s Regional Control,” Slavic Review 50, no. 2 (Summer), 347–58. Wilson, Andrew. 2005. Virtual Politics: Faking Democracy in the Post-Soviet World (New Haven: Yale University Press). Wintrobe, Ronald. 2007. “Dictatorship: Analytical Approaches,” in Carles Boix and Susan C. Stokes, eds., The Oxford Handbook of Comparative Politics, 363–94 (New York: Oxford University Press). World Bank. 2014. World Development Indicators (Washington, DC: International Bank for Reconstruction and Development), http://data.worldbank.org/products/wdi, accessed July 23, 2014. Wyman, Matthew, et al. 1995. “Public Opinion, Parties and Voters in the December 1993 Russian Elections, ” Europe-Asia Studies 47, no. 4 (June), 591–614. Yargomskaia, Nataliia B. 1999. “Izbiratel’naia Sistema i Uroven’ Partiinoi Fragmentatsii v Rossii

[The Electoral System and the Level of Party Fragmentation in Russia],” POLIS: Politicheskie Issledovaniia, no. 4, 122–29. Zadorin, Igor. 2000. “Sredstva Massovoi Informatsii i Elektoral’noe Povedenie Rossiian,” in Michael McFaul et al., eds., Rossiia na Dumskikh i Prezidentskikh Vyborakh (Moscow: Carnegie Endowment for International Peace), http://pubs.carnegie.ru/elections/president2000/book/03zadorin.asp, accessed June 10, 2013. Zaslavsky, Victor, and Robert J. Brim. 1978. “The Functions of Elections in the USSR,” Soviet Studies 30, no. 3 (July), 362–71. Zimmerman, William. 1987. “Mobilized Participation and the Nature of the Soviet Dictatorship,” in James R. Millar, ed., Politics, Work and Daily Life in the USSR: A Survey of Former Soviet Citizens, 332–53 (New York: Cambridge University Press). Zubarevich, Natalia. 2015. “Krizis v Regionakh: Nastalo Vremia Zaplatit’ za вЂKrymnash,’” Slon.ru (February 24), http://slon.ru/insights/1218531/, accessed March 1, 2015. Zvonovskii, Vladimir B. 2004. “вЂElektoral’nyi Tranzit’: K Dvum Stranam v Granitsakh Odnoi? ” POLIS: Politicheskie Issledovaniia, no. 3, 33–43.

Page 267 →

Index administrative resources, 38, 131, 148–49, 157, 201. See also elections: fraud in authoritarianism, 2–12, 15–16, 19, 50, 99, 118, 174, 180, 200, 225 competitive/electoral, 1, 4, 7, 11, 19–20, 50, 96, 99, 118, 178, 185–86, 188, 203, 207, 225 subnational, 10–11, 16, 62, 81, 131, 147, 154, 176, 195, 205, 207–8 Chechnya, Republic of, 22, 35, 37–38, 41, 45, 60, 63, 154, 166, 220, 222–23 clientelism. See patron-client relations Communist Party of the Russian Federation, 35, 37–38, 42–43, 45, 47, 129, 134–35, 144, 163 Communist Party of the Soviet Union, 130, 225 competitive authoritarianism. See authoritarianism: competitive/electoral constitution (1993), 14, 23, 27, 32, 36–37, 42, 62–63, 184 deference (regional electoral support for the Kremlin), 19, 20, 114, 147, 156–63, 173–97, 201–7 democracy, 1, 3–13, 16–17, 19–21, 32–33, 37, 39–45, 49–52, 55, 57–62, 82–88, 95, 118–19, 129, 146, 152–53, 163, 165, 174, 176, 180, 196, 199–201, 206, 207–8, 211–12, 225 diffusion, political, 19, 176, 178–80, 185, 191–92, 197, 202, 205 dominant party. See party of power; United Russia economic development. See modernization elections Federation Council, 36, 225 fraud in, 7, 30, 40, 47, 50, 104, 131, 144 gubernatorial, 11–12, 17–21, 29–30, 40, 45, 47, 54–56, 61–71, 78–97, 101, 115, 123–24, 127–28, 137–39, 144, 177, 184, 187–88, 191, 198–200, 203–4, 206–7, 209, 215, 222–24 presidential, 4, 19, 25, 27–28, 42, 50, 87, 93, 95, 99, 104, 119–22, 133, 139, 144, 152, 154–55, 159, 171–73, 176, 183–84, 186–88, 189, 192, 194, 197, 202, 207 State Duma, 20, 27–28, 37–38, 41, 45–47, 63, 77, 80, 119, 121, 124–25, 150, 159, 162, 176, 181–82, 209, 218 turnout in, 18–19, 27, 36, 46, 87–88, 113, 117–52, 157–58, 162, 201–2, 215, 217–19 electoral authoritarianism. See authoritarianism: competitive/electoral ethnicity. See nationality

Fatherland/All Russia (political party), 41, 43, 80, 150, 159, 181–82, 193, 209 federal administrative districts, 44, 141, 171–72, 174, 179, 183, 196, 202 federalism, 1, 11, 35, 39, 40, 44–45, 61, 76, 79, 81, 100, 102, 105, 115–16, 123, 131, 174–76, 202–3, 206, 209 gubernatorial appointments, 18–20, 29, 46, 82, 99, 101–3, 106, 108, 110, 116, 187, 201, 209 Page 268 →gubernatorial elections. See elections: gubernatorial Ingushetia, Republic of, 22, 122, 150–51, 154, 156, 159, 166, 182, 192, 220, 223 Krasnodar Krai, 212, 223 legislatures federal (see State Duma) regional, 28–30, 45–46, 130, 135 Liberal Democratic Party of Russia, 32, 37–38, 43, 174 Luzhkov, Yurii, 41, 103 machines, political, 18–19, 40, 62, 65, 70, 79, 101, 105, 108, 115, 130–31, 149–50, 153, 177–78, 184, 188, 195, 200–201, 206, 209 Medvedev, Dmitrii, 8, 11–12, 27, 40, 44, 46–47, 103, 106, 108, 110, 112, 115–16, 156, 159, 203 modernization, 13, 17, 42, 47–48, 51–52, 58–59, 82, 94, 162–63, 183, 199, 205, 211–213 Mordovia, Republic of, 63, 122, 159 Moscow (city), 66–67, 143, 154, 161 Moscow (oblast), 22, 143, 162 nationalism, 31 nationality, 1, 17–18, 22–23, 30–35, 57, 61–63, 67, 69–71, 78, 80, 84, 88, 91, 93, 99, 101, 104–5, 109, 112, 114, 116, 123, 126–27, 130–35, 138–41, 145, 147, 154, 158–59, 162–63, 165–71, 175, 183, 186, 188, 201–2, 221 Our Home is Russia (political party), 37–38 party of power, 14, 28, 38, 41, 43, 45–46, 80, 150, 152, 165, 185 party system, 43, 57, 60, 64, 129 patron-client relations, 15, 43, 60, 101, 130–32, 208 political parties. See party system; names of individual parties power vertical. See federalism Primorye Krai, 96, 223

Putin, Vladimir, 1–5, 8–21, 25, 27, 30, 32, 40–52, 58, 66, 71, 80–100, 104–6, 110–16, 122, 123, 129, 134–35, 137, 147, 149–51, 156–57, 159, 162–63, 167, 169, 173–83, 185, 187–89, 193–94, 198, 201–12, 223, 225 regime, Russia’s political, 2, 4, 14–16, 19, 21, 39, 43, 47–50, 100–101, 115, 174, 178, 186, 188–89, 196, 198, 206–8, 211 republics, 11, 17, 22, 29–37, 54–55, 60–80, 93–96, 101, 104, 108–9, 122, 124, 126, 132, 135, 139–42, 147, 154, 159, 166–67, 169–70, 173, 187, 200, 206, 209, 216, 225 Russia’s Choice (political party), 36–38, 153 Shaimiev, Mintimer, 79, 103, 105, 108, 193 St. Petersburg, 22, 41, 59, 63, 96, 123, 147, 162, 173, 202, 212, 220, 224 State Duma, 23, 27, 40–41, 45, 128 Tatarstan, Republic of, 10, 22, 29, 35–37, 41, 63, 79, 103, 108, 119–22, 135, 154, 157, 179, 192–94, 197, 202 turnout. See elections: turnout in United Russia (political party), 15, 28, 30, 41–47, 50, 99, 105, 115, 122, 129, 133, 149, 150, 156, 159, 162, 167, 177, 180–82, 185, 188, 209 vertical of power. See federalism voting. See elections Yabloko (political party), 37 Yeltsin, Boris, 2, 4–5, 8, 11–12, 17, 20–44, 48, 50, 54, 56, 60–63, 66–67, 71, 76, 78–79, 81–86, 93–96, 122–23, 133, 137, 152–55, 159, 163, 168, 173–74, 176, 181, 186, 187, 189–90, 201–9, 216 Zhirinovskii, Vladimir, 32 Zyuganov, Gennadii, 35, 38, 42, 45, 47, 133, 163, 174, 186