Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Abstract

A large literature expects that as protests unfold in electoral autocracies, voters who supported the ruling regime in the past will withdraw support and shift to supporting its opponents. Yet there are only a few empirical tests of how opposition protests influence voter defections in these regimes. To gain empirical traction on this question, I draw on evidence from Russia. Tying together evidence from a protest-event dataset and a panel survey of voters conducted prior to and during the 2011-2012 protest wave, I examine how voters who supported the ruling regime in the past respond to anti-regime mobilization. Results reveal differentiation in defections. While opposition protests dampen support for the ruling regime and depress engagement, they do not necessarily translate into greater support for the regime’s challengers. Findings, which have implications for debates on defection cascades in autocracies, speak to the literatures on authoritarian endurance and the legacies of (attempted) revolutions.

Keywords

voter defections protests public opinion electoral autocracies Russia

Introduction

An influential literature recognizes the threat that opposition protests pose to the stability of authoritarian rule. As far as the masses are concerned, existing research expects that when opposition activists take to the streets popular support for the ruling regime will erode (Hale & Colton, 2017; Hollyer, Peter Rosendorff, & Vreeland, 2015; Kuran, 1997; Lohmann, 1994). However, few studies provide empirical evidence of how unfolding anti-regime mobilization influences voter defections in electoral autocracies. As detailed datasets on how the electorate responds to opposition protests do not come easily for non-democratic regimes, the microfoundations of voter defections remain poorly understood (Hale & Colton, 2017, p. 323). In a similar vein, there are only a few empirical tests of the hypothesis that voters who withdraw support from the ruling regime as protests unfold will also shift to support for the opposition. Regime defectors, for example, may withdraw from politics and disengage from voting instead.

Understanding how protests affect voter defections in electoral autocracies, regimes that combine authoritarian practices with multiparty elections, is of utmost importance. Protesters’ ability to win the support of bystanders could increase participation in unfolding unrest (see, for example, Aytaç, Shiumerini, & Stokes, 2018; DeNardo, 1985; Onuch, 2015) and shift ruling elites toward the protesters’ preferences (McAdam & Su, 2002). When protests demobilize a large share of the population instead, including the bulk of regime supporters, the likelihood of political change declines. As existing research reminds us, citizen disengagement from politics often contributes to authoritarian resilience. Abstention from countrywide elections in Mexico, for example, helped the PRI to win elections and maintain its dominant position (Domínguez & McCann, 1996, p. 164). The deliberate disengagement of educated voters in Zimbabwe has also been found to contribute to the authoritarian regime’s resilience (Croke, Grossman, Larreguy, & Marshall, 2016).

Drawing on evidence from contemporary Russia, this article studies how ruling regime supporters respond to anti-regime protests in electoral autocracies. I propose that opposition protests provide information about the regime and its opponents and change voters’ opportunities to infer whether political change is likely or not (Lohmann, 1994; Magaloni, 2006; Meirowitz & Tucker, 2013). In line with existing research, I suggest that information about regime abuses and malpractices, made publicly available as opposition protests unfold, can dampen support for the ruling regime. Defections, however, could exhibit significant differentiation. I anticipate that while, in response to protests, some voters may withdraw support from the ruling regime and side with the opposition, others may withdraw support from the regime and disengage from politics. While disengagement also serves as a meaningful form of dissent in electoral autocracies, contexts where participation does not provide genuine input into the political process (see, for example, Croke et al., 2016; Karklins, 1986), it may be perceived as less damaging for the ruling regime than mass defections to the opposition (Hale & Colton, 2017, p. 333). I finally explore whether assessments of regime malpractices and confidence in protests’ ability to effect political change help to explain differentiation in defections.

To systematically assess these propositions, I leverage empirical evidence from Russian voters’ response to the 2011-2012 electoral protests in Russia. My research design ties together original, district-level protest-event data from the Lankina protest-event dataset (Lankina & Voznaya, 2015) with rich individual-level survey data from the panel component of the 2008 and 2012 Russian Election Study (RES) surveys (Colton & Hale, 2014; Colton, Hale, Kosolapov, Tonis, & Prudnikova, 2014; Hale & Colton, 2017). The combination of the two sets of data allows me to uniquely identify how respondents who voted for the ruling-regime candidate in the 2008 presidential election reported to vote in the 2012 presidential election, and to study whether protests taking place nearby influenced changes in support for the ruling regime. To preview the results, I find that opposition protests increased the likelihood that voters who voted for the ruling-regime candidate in 2008 would abstain from voting in 2012. Yet the analysis finds little support for the argument that protests also increase support for the regime’s challengers, the opposition parties, and activists who organized the protests and made information about regime abuses publicly observable. Contrary to conventional wisdom, findings show that the likelihood that regime supporters would defect to the opposition, relative to the likelihood that they would remain loyal to the regime, did not increase as a function of protests. Finally, I show that protests influence assessments of electoral fraud, which are in turn associated with both disengagement from politics and defections to the opposition. Disillusionment with the prospect of political change in Russia, little mitigated by local unrest, helps explain differentiation in defections.

Altogether, this work sheds light on how opposition protests affect support for illiberal regimes and their challengers, and contributes to broader research on defection cascades in electoral autocracies. To the best of my knowledge, this work constitutes one of the first attempts to empirically study protest effects on voter defections in a non-democracy. Findings highlight the importance of considering how, in addition to, or instead of, generating support for the opposition, opposition protests in electoral autocracies may also lead to disengagement (see also Croke et al., 2016; Lasnier, 2017, 2018). They also suggest that in trying to better understand the underpinnings of illiberal regime stability, as well as the legacies of (attempted) revolutions, we ought to pay greater attention to protest effects on public opinion and to important differentiation in voter defections in these regimes.

Protests in Electoral Autocracies

Illiberal governments live in the shadow of mass unrest (Hollyer et al., 2015; Svolik, 2012). Large crowds taking to the streets force incumbents to resort to riskier strategies to maintain their rule and make falsifications either impossible, or so obvious and outrageous that people will oppose them (Colton & Hale, 2014). By signaling that the incumbent is weak, protests could also trigger splits among the ruling elites, causing regime insiders to defect and the country’s ruling coalition to crumble (Bratton & Van de Walle, 1994; Hale & Colton, 2017; Kuran, 1997; Lohmann, 1994). Moreover, as cascade theories remind us, by revealing information that prompts regime supporters to update their evaluations of the regime (Lohmann, 1994), or by empowering individuals to overcome preference falsification (Kuran, 1991, 1995), opposition protests could undermine popular support for incumbents. By being directly or indirectly witnessed by bystanders—whether on one’s commute to work, or by friends and neighbors who participated in the events (see, for example Branton, Martinez-Ebers, Carey, & Matsubayashi, 2015; Onuch, 2015; Wallace, Zepeda-Millán, & Jones-Correa, 2014)—opposition protests construct a reality that cannot be easily dismissed by government propaganda.

Leveraging evidence from survey experiments and a combination of survey and protest data, recent studies have documented important effects of local protests on the political attitudes of bystanders, the spectators of the protests, in Western democracies (Branton et al., 2015; Madestam, Shoag, Veuger, & Yanagizawa-Drott, 2013; Wallace et al., 2014) and electoral autocracies (Frye & Borisova, forthcoming; Tertytchnaya & Lankina, 2020). These studies propose that nearby protests may influence attitudes through a number of psychological and informational mechanisms. Individuals, for example, have been found to assign great inferential weight to vivid events taking place in their localities, as such information is socially relevant and can be more easily recalled (Kahneman, Slovic, & Tversky, 1982; Weyland, 2012). Local networks of protesters also transmit information that could help bystanders update their political beliefs and could influence their propensity to protest (see, for example, Steinert-Threlkeld, 2017).

Yet, to date, existing studies do not provide direct evidence on how protests affect voter defections in electoral autocracies. Relatedly, there are only a few empirical tests of the hypothesis that voters who withdraw support from the ruling regime, specifically as opposition protests unfold, will also shift to supporting its opponents. This neglect is puzzling, as opposition parties’ and activists’ ability to generate support has important implications. As existing research reminds us, a reciprocal relationship between social movements, electoral and political change exists in democracies (Gamson, 2004; McAdam & Tarrow, 2010) and electoral autocracies alike (Lasnier, 2017, 2018; Smyth & Soboleva, 2016).

Opposition Protests and Differentiation in Defections

Seminal models of vote choice in autocracies present the electorate as facing a two-step dilemma. First, voters need to decide whether they support the ruling party or not. Evaluations of existing alternatives to the status quo, and assessments of the ruling party’s policy record, weigh heavily at this stage. Second, and only after voters have decided to defect, they need to decide which of the opposition parties to support (Domínguez & McCann, 1996; Magaloni, 2006, 2010). I build on these contributions to describe my expectations about protest effects on voter defections. Drawing on theories of informational cascades, I propose that opposition protests provide information about the regime and its opponents, and change voters’ opportunities to infer whether political change is likely or not (Lohmann, 1994; Magaloni, 2006; Meirowitz & Tucker, 2013).

To begin with, opposition protests reveal information about regime malpractices and abuses. Activists taking to the streets may inform voters about the extent of electoral fraud (Beissinger, 2007; Tucker, 2007) and reveal information about widespread corruption and economic mismanagement. Existing research shows that publicly available information about mismanagement and power abuses may be associated with widespread voter dissatisfaction and fluctuations in support for the regime (Hollyer et al., 2015). Therefore, I anticipate that when opposition protests unfold, voters who supported the ruling regime in the past may withdraw support. As Lohmann reminds us, when hitherto hidden information about the malign nature of the regime enters the public domain, its viability is undermined (Lohmann, 1994, p. 44). I trace opportunities for political learning and opinion change to the broader environment that surrounds opposition protests in illiberal regimes. As in democracies (see, for example, Atkeson & Maestas, 2012), opposition protests and the threat they represent in autocracies could heighten spectators’ anxiety and dampen the influence of partisan cues and regime attachments in the formation of political attitudes and vote choice (Branton et al., 2015).

Voters who decide to defect, however, also need to decide whether to disengage from politics or shift to supporting the opposition instead. Defectors’ decisions at this stage, I propose, could be influenced by their evaluations of the opposition and their beliefs about the likelihood of political change (see also Magaloni, 2006, 2010). It is reasonable to expect that in forming these assessments, voters will rely heavily on their observations of opposition parties’ and activists’ performance during street protests. This is because opposition parties and activists are highly ambiguous entities in electoral autocracies, contexts characterized by scarce information about the opponents of the ruling regime (Magaloni, 2006, 2010). Opportunities for voters to observe opposition parties and activists in action, to learn about their demands, and decide whether they support them or not do not come easily in these regimes. Using preemptive repression (Ritter & Conrad, 2016; Truex, 2019), propaganda and censorship (Guriev & Treisman, 2015; Spaiser, Chadefaux, Donnay, Russmann, & Helbing, 2017; Stockmann & Gallagher, 2011; Tertytchnaya & Lankina, 2020), autocrats can preempt opposition activists from taking to the streets and prevent support for the demands of the opposition from growing.

When the opposition is perceived as divided or unable to maintain a unified message during street protests, voters may conclude that existing parties are unable to credibly challenge the ruling regime and that political change is unlikely. As social movement theory reminds us, internal divisions compromise protesters’ ability to win the support of targeted populations and could trigger a sense of political alienation, even in advanced democracies (Wallace et al., 2014). When voters infer that costly political action, such as participation in protests and elections, fails to translate to genuine input into the political process, disengagement from politics may also trump defections to the opposition. Altogether, instead of encouraging mass defection to an opposition party or candidate, protests in electoral autocracies may also “quash the hope” and become associated with disengagement (see also Chong, De La O, Karlan, & Wantchekon, 2015). Defectors who conclude that political change is unlikely or that political participation is futile could withdraw from politics and disengage from voting.

In a context of uncompetitive elections, political disengagement is a meaningful expression of dissent (Croke et al., 2016; Karklins, 1986). Yet, considering differentiation among defections is important. As Hale and Colton (2017) remind us, distinguishing between defections to a nonvote and defections to the opposition “could be interpreted as a distinction between forms of defection that are more and less damaging” (p. 333) for the ruling regime. This is because illiberal governments’ ability to demobilize a large share of potential defectors could contribute to regime stability and resilience. Examples of how citizen disengagement has facilitated the survival of electoral autocracies abound. In Mexico, abstention from countrywide elections enabled the PRI to win elections and maintain its dominant position (Domínguez & McCann, 1996). In Zimbabwe, the deliberate disengagement of educated voters has been linked to the stability of the country’s authoritarian equilibrium (Croke et al., 2016). The Russian government has also relied for years on the demobilization of the urban educated voters whose support is otherwise difficult to mobilize (Gel’man, 2015). Similarly, public sector workers’ low rates of participation in the protests of 2011-2012 have been associated with the protest movement’s limited success (Rosenfeld, 2017). Conversely, protesters’ ability to win the support of bystanders could increase participation in unfolding unrest, bolster electoral support for opposition parties (Aytaç et al., 2018; DeNardo, 1985; Onuch, 2015; Smyth & Soboleva, 2016), and even shift ruling elites toward the protesters’ preferences (McAdam & Su, 2002).

Before turning to the case selection, it is worth briefly considering a number of potential qualifications to this argument. First, the effect of protests on defections may not be homogeneous across the electorate. Individual-level attributes such as risk acceptance could shape voters’ propensity to support an unknown opposition. Morgenstern and Zechmeister (2001), for example, have documented an important effect of risk aversion incentives on the probability of voters defecting from the PRI in Mexico. Given data availability, testing hypotheses related to risk aversion is unfortunately not possible in the context of this study. Second, considering regime responses to protests is important, as they could also influence differentiation in defections. Efforts to delegitimize the protest movement, for example, through the coverage of protests in state-controlled media, could direct defections toward disengagement, as voters dissatisfied with the regime and distrustful of the opposition withdraw from politics. Alternately, or additionally, autocrats can use targeted repression in an effort to associate protests with disruption and chaos and prevent moderates and risk-averse citizens from supporting the opposition (see, for example, Guriev & Treisman, 2015; Tertytchnaya & Lankina, 2020; Young, 2019). To prevent opposition coordination, autocrats could also strategically co-opt opposition parties represented in the national and regional legislatures (Armstrong, Reuter, & Robertson, 2017). Yet, no matter how sophisticated the regime’s propaganda or repression apparatus, it may be more challenging to discredit a protest movement that is neither fractious nor divided. Explicitly investigating how state responses to protests influence defections falls beyond the scope of this work. Yet control variables used in the empirical analysis, such as items that capture voters’ news-watching patterns and indicators of political competition, as well as additional tests with items of repression (Table C5 in the Supplemental Appendix), suggest that protests have a direct effect on defections, even when we take regime responses to the protests into account.

The Russian Case

Contemporary Russia shares many features in common with other electoral autocracies, making it a useful setting for research on protests and defections. Government control of the media, opposition harassment, and a consolidated “power-vertical” render political competition in the country unfair. Yet, in the aftermath of the December 2011 State Duma (parliamentary) election, tens of thousands of citizens took to the streets to protest against electoral fraud. The 2011-2012 electoral protest wave took place in a period of growing dissatisfaction with the government in Moscow, and with United Russia, the dominant party. United Russia’s vote share, for example, dropped from 64.3% of the electorate in the 2007 election to 49.3% in the December 2011 election. Between September and December 2011, that is, shortly prior to and just after the 2011 parliamentary election, Vladimir Putin’s approval rating also dropped by five percentage points, moving from 68% to 63% (Levada Analytical Centre, 2012).

Between the parliamentary election of December 2011 and Vladimir Putin’s inauguration in May 2012, electoral protests were widespread across Russia. Protesters raised a plethora of demands, ranging from a return to the Soviet Union to a repeat of the Duma election and the overthrow of the regime (Greene, 2013). Russia’s nonsystemic or extraparliamentary opposition played a leading role in organizing the events, while parties represented in the State Duma also joined the protesters (Sakwa, 2014). According to Reuter and Robertson (2015), this was the first time in many years that “parliamentary and extraparliamentary oppositions united in the streets.” Members of the Communist Party (KPRF) and of the Just Russia party were particularly active during the winter protests, before abandoning the protest movement in the spring. As Richard Sakwa (2014) notes, the leadership of Just Russia and that of the KPRF eventually “chose conformity over opposition” (pp. 11-14). Poor coordination among opposition parties and activists, both in the streets and in the context of the 2012 presidential election, may have played a key role in preventing electoral support for the opposition from growing.

The Kremlin’s initial response to the protests was rather tolerant. By the spring months, however, almost one in three protest events faced some sort of disruption (Tertytchnaya & Lankina, forthcoming). Coverage of the protests in state-run or -controlled media also sought to prevent the erosion of popular support for the regime. As Frye and Borisova (forthcoming) note, for example, in its coverage of the large protests of December 10, 2011, state television did not report protesters’ calls for Putin’s resignation. Progovernment users also tried to marginalize the opposition online, across social media platforms (Spaiser et al., 2017). Examined in existing literature are also questions regarding the demands of the opposition (Lankina & Voznaya, 2015; Robertson, 2013; Volkov, 2012), the profiles of the protesters themselves (Rosenfeld, 2017; Ross, 2015) and of their supporters (Chaisty & Whitefield, 2013). Focusing on protest effects on political attitudes, Frye and Borisova (forthcoming) have documented protest effects on trust in the government among Moscow residents, while Tertytchnaya and Lankina (2020) have examined the effect of regional protests on support for the demands of the protesters. I join these studies in revisiting the impact of protests on attitudes, and extend that agenda to empirically consider how protests influence voter defections in these regimes.

This study follows Hale and Colton (2017) and uses evidence from the authors’ RES surveys to shed greater light on the “mysterious” microfoundations of regime defection cascades in electoral autocracies. Yet my approach differs from theirs insofar as I am interested in studying how defections are motivated by opposition protests taking place in voters’ districts. Moreover, rather than using vote in the Duma election, which occurred prior to the onset of the protests, I use vote in presidential election as the dependent variable. While in 2008 the ruling-regime candidate was Dmitry Medvedev, in 2012 the regime’s candidate was Vladimir Putin himself. This is a limitation of the empirical analysis that needs to be acknowledged upfront, and one that is difficult for the study to assess. Yet it is possible to argue that this change in candidates would make it more difficult to establish protest effects on voter defections, as Putin has consistently enjoyed higher levels of support than Medvedev and United Russia. Moreover, in systems with high levels of personalist rule, such as Putin’s Russia, support for incumbents who embody the stability of the regime is often remarkably resilient to revelations of abuses and malpractices, even as support for other institutions and politicians declines. Existing research also shows significant personalist support for Vladimir Putin and finds that votes for the ruling party have traditionally increased on Putin’s coattails (see, for example, Smyth, 2014). Finally, focusing on a presidential election provides a tougher test for evaluating a disengagement scenario, as abstention in presidential elections in Russia has been traditionally lower than abstention in the Duma elections.

Data

The public opinion data I analyze come from the 2008 and 2012 RES surveys (Colton & Hale, 2014; Colton et al., 2014; Hale & Colton, 2017). The surveys, conducted face-to-face in respondents’ homes in March-May 2008 and April-May 2012, are nationally representative, each with a sample of over 1,000 respondents (1,130 respondents were interviewed in 2008 and 1,682 in 2012). Uniquely, the surveys included a panel component, as 661 of the respondents interviewed in 2008 were successfully interviewed in 2012. The empirical analysis presented in the sections that follow relies on this small sample of 661 respondents interviewed prior to and during the 2011-2012 electoral protest wave, as protests unfolded. The “March of the Millions,” for example, one of the largest opposition protests since the collapse of the Soviet Union, took place on May 6, 2012, while the RES survey of 2012 was in the field.

According to Hale and Colton (2017), of those respondents interviewed in 2008, men, people living in larger urban communities, and those with lower levels of education, that is, the more mobile social groups, were more likely to drop out from the panel. Attrition in panel data poses bias concerns, especially when selectively related to the outcome variables of interest. Similarly, as Hale and Colton (2017) also note (p. 328), attrition would possibly make it more difficult to find variables such as urban status, age, or education and gender significant, as we have fewer data points on them to analyze. While in the context of this study social mobility does not form part of the argument, I make sure to probe robustness using alternative data. Results obtained from the analysis of the RES panel surveys, namely, evidence that opposition protests dampened support for the ruling regime but did not bolster support for the opposition, remain consistent when we run the analysis using an entirely different, cross-sectional, and nationally representative survey that was in the field in January 2012 (White, 2015). I present my findings in Tables D4 and D5 in the Supplemental Appendix.

Dependent Variable

My main dependent variable captures how respondents who voted for the ruling regime in the presidential election of 2008 voted in the 2012 presidential election. It relies on the combination of two survey items from the 2008 and 2012 RES surveys, which ask respondents to state whether they voted in each presidential election and whom they voted for.¹ The wording of both questions is standard, as respondents who participated in the election are presented a list of candidates and asked to report for whom they voted. Respondents are also able to say whether they cast a spoiled ballot and to choose one of the “Hard to Say/ Refuse to Answer” responses.

Relying on these two items, I specify a three-category outcome variable. The first category of my main dependent variable consists of “loyal” regime supporters. These are respondents who indicated that they voted for the ruling-regime candidate in both 2008 and 2012. The second category consists of respondents who withdrew support from the regime yet did not vote for any of the opposition parties or candidates. These are respondents who voted for the ruling-regime candidate in 2008 but defected to a nonvote in 2012 by abstaining from the election or casting a spoiled ballot. This category also consists of respondents who did not give a definite answer, or who did not wish to comment. As Hale and Colton (2017, p. 332) suggest, all these responses constitute forms of defection and are consistent with the argument that protests make it costlier for people to publicly support the ruling regime. Yet, as I show in Table C1 in the Supplemental Appendix, omitting “Don’t Know” or “Difficult to Answer” responses from the analysis does not change the results. The third and final category of the outcome variable consists of respondents who voted for the ruling-regime candidate in 2008 but switched to supporting any of the opposition parties and candidates competing in the 2012 presidential election. Opposition defectors therefore are respondents who voted for Medvedev in 2008 and for the opposition in 2012. In the presidential election of March 2012, opposition defectors reported to vote for any of the following opposition parties or candidates: Vladimir Zhirinovsky, head of the nationalist Liberal Democratic Party, Gennady Zyuganov, leader of the Communist Party (KPRF), Sergey Mironov, leader of the Just Russia faction, and Mikhail Prokhorov, a prominent figure during the 2011-2012 electoral protest wave, who was also running as an independent candidate in the 2012 election.

We could of course ask whether in electoral autocracies citizens feel free to sincerely evaluate the regime and its opponents during “normal politics” in general and at times of protests in particular. While existing studies document that support for the regime reflects Russians’ actual opinions (Frye, Gehlbach, Marquardt, & Reuter, 2016), investigating whether support for the regime is sincere or not falls beyond the scope of this work. I argue that even when voters do not offer sincere statements of support for either the regime or the opposition to survey numerators, public statements of their preferences in opinion polls are politically meaningful (Hale & Colton, 2017). As individuals’ propensity to protest depends on their beliefs about the popularity of the regime and its opponents (Hollyer et al., 2015; Kuran, 1991), widely publicized public opinion surveys documenting that the opposition lacks support and that the incumbent is popular could hinder coordination in the political domain. In December 2011, for example, Russian TV channels and newspapers often reported that, in opinion polls conducted by the Levada Center, the most respectable public opinion firm in the country, the opposition enjoyed low levels of support and that Russians were deeply concerned about ensuing unrest in the new year.²

Independent Variables

To complement the individual-level survey data, I rely on detailed protest-event data on anti-regime political protests that took place in the aftermath of the December 4, 2011, Duma election. Protest data come from the Lankina protest-event dataset and cover the period from December 2011 to May 2012. The Lankina dataset, assembled from the liberal namarsh.ru website, covers political protests ranging from small-scale acts to large-scale demonstrations featuring tens of thousands of protesters. The dataset for this period has been cross-validated with newspaper and online protest archives sourced from additional opposition sources, such as the website of the left-leaning Institute of Collective Action (IKD). Events not initially reported by namarsh.ru were also added to the dataset.³ Political protests are those addressing issues of electoral fraud, calling for the resignation of elected or politically appointed officials at all levels of government, and dealing with political repression. The dataset does not include rallies organized by the regime or its supporters (see also Lankina & Voznaya, 2015; Tertytchnaya & Lankina, 2020).

The main independent variable used in the analysis, “protest count,” is a count of the number of opposition protests that took place in survey respondents’ districts between 4 December, 2011 (the day of the State Duma election), and up to the day of their interview (for a similar approach, see Branton et al., 2015; Wallace et al., 2014).⁴ In the 2012 RES, districts, or raions, are also defined as primary sampling units (Colton et al., 2014, p. 6). Districts are an administrative division of Russia’s oblasts. Oblasts, or regions, are the equivalent of states in the United States. The main protest variable, which ranges from 0 to 81 events, varies across districts and over time within districts, as respondents from the same district were interviewed over multiple days in April and May 2012. This item helps investigate whether unfolding protests influence voter defections. The analysis presented below also uses the natural logarithm of the continuous protest item. Finally, a third variable, “protests binary,” captures the occurrence of protests in respondents’ districts. This dummy variable is coded as one if any protest events took place in respondents’ districts up to the day of their interview, and zero if otherwise. This item makes no assumptions about a linear effect of protests on attitudes and does not distinguish between respondents in areas with just one, or with multiple events.

Finally, the empirical analysis controls a number of demographic and socioeconomic indicators. As existing research shows that politicized online media influenced attitudes toward fraud and the election during the 2011-2012 protest cycle, I control for online news consumption with a variable that captures whether and how often respondents read or watch news on the Internet (Enikolopov et al., 2018; Reuter & Szakonyi, 2015). This item also takes us a step closer to testing whether authoritarian techniques to control dissent online, which were prevalent during the 2011-2012 protests (Spaiser et al., 2017), were effective. Respondents who do not watch or read news online are assigned a value of 0, while respondents who watch news on the Internet almost every day are assigned a value of 4. The analysis also controls for respondents’ pocketbook and national economic assessments. Both items are measured on a 5-point scale, with higher values denoting greater economic pessimism.⁵ An item that captures whether respondents have protested in the past or not aims to distinguish protest participants from bystanders. Moreover, the surveys contain a number of demographic controls, including age, gender (male), education (measured on an 8-point scale with higher values denoting that respondents have higher education), and employment status (coded as one if respondents are currently employed and zero if they are not). A dummy variable that captures whether respondents live in an urban settlement or not helps to adjust for the fact that the majority of protest events during this time took place in large urban centers.

I complement the individual-level data with an indicator of regional democracy. This comes from the Petrov and Titkov (2013) index and helps address concerns related to the fact that more democratic regions could have more favorable attitudes toward the opposition to begin with. The item used in the analysis, measured on a 5-point scale, is only available at the regional (oblast) level and covers the period from 2006 and 2010. Higher values suggest that regional political competition is more robust; local, regional, and national elections are more contested; and manipulations and restrictions on active and passive electoral rights are limited. This indicator has additional advantages. As Frye and Borisova (forthcoming, p. 8) have argued, having seen fraud in past elections, Russians may have adjusted their attitudes to expect falsifications and may not be particularly responsive to protests. Tolerance of electoral falsifications could arguably be higher in Russia’s least competitive regions, where falsifications have been traditionally more prevalent. Altogether, the full set of controls introduced in the analysis allows me to estimate well-specified models and reduces threats to inference.

Descriptive Evidence: Differentiation in Defections

Relying on the panel component of the 2008 and 2012 RES surveys, Table 1 shows how all 386 voters who supported the ruling-regime candidate in the 2008 presidential election voted 4 years later, in the March 2012 presidential election. Results suggest that while around 73% of voters who voted for Medvedev in 2008 remained loyal, that is, voted for Putin in 2012, 27% of voters withdrew support by 2012. Fourteen percent of voters who supported the ruling-regime candidate in 2008 defected to a nonvote. These voters abstained from the election, reported casting a spoiled ballot or did not wish to report whom they voted for in the 2012 wave of the RES survey. Fourteen percent of all voters who supported the ruling regime in 2008 sided with the opposition in 2012. Gennady Zyuganov, the Communist Party leader and candidate in the 2012 election gained 5% of the regime voters who defected to the opposition. Mikhail Prokhorov gained a further 3.6% of regime voters, while Sergei Mironov and Vladimir Zhirinovsky received each around 2% of the vote.

Table 1.

Patterns of Ruling Regime Defections.

	Panel sample (2008 regime voters) (n = 386)
Voted for Vladimir Putin in 2012	73.32
Defected to nonopposition vote in 2012	13.74
Defected to opposition vote in 2012	12.95
Opposition vote breakdown:
Zhirinovsky (LDPR)	1.81
Zyuganov (KPRF)	5.44
Mironov (just Russia)	2.07
Prokhorov (independent candidate)	3.63

Source. 2008 and 2012 Russian Election Studies.

This table displays the share of ruling-regime voters in each category. Ruling-regime voters are respondents who in the 2008 presidential election voted for the regime candidate, Dmitry Medvedev.

The percentage of voters who remained loyal to the regime was higher in districts without opposition protests (76%) as opposed to districts with them (67%). The percentage of ruling-regime voters who defected to a nonvote, that is, who either chose not to vote or cast a spoiled ballot, was also higher in areas with protests (18% vs. 12% in districts without unrest). Finally, in districts with opposition protests, around 15% of all ruling-regime voters switched from supporting the ruling regime in 2008 to supporting the opposition in 2012. In places without protests, this figure was around 3 percentage points lower, as around 12% of voters who supported Medvedev in 2008 switched to supporting one of the opposition parties and candidates competing in the 2012 presidential election. Altogether, in districts with opposition protests the regime lost around 33% of all voters who voted for Medvedev in 2008. In districts without unrest the ruling regime lost around 24% of its 2008 voters.

Multivariate Analysis: Protests and Differentiation in Defections

I next turn to the multivariate analysis. As the main dependent variable employed in Table 2 consists of three unordered alternatives in the choice set, I model the defection process using a multinomial logit model (MLM). The specification of the dependent variable allows me to investigate how the risk of voting for the ruling-regime candidate in 2008 and 2012 compares to the risk of (a) 2008 regime voters defecting to a nonvote and (b) 2008 regime voters voting for another party. Coefficients reported in Table 2 are relative risk ratios from multinomial logistic models. The relative risk ratio of a coefficient shows how the risk of the outcome falling in the comparison group, compared with the risk of the outcome falling in the reference group, changes with the various predictors. Intuitively, a risk ratio of 1 indicates that the risk is comparable in the baseline and comparison group. A value greater than one indicates increased risk, while a value lower than 1 indicates decreased risk.

Table 2.

Protests and Ruling-Regime Defections.

	(1.1)	(1.2)	(2.1)	(2.2)	(3.1)	(3.2)
	Disengage/loyal	Opposition/loyal	Disengage/loyal	Opposition/loyal	Disengage/loyal	Opposition/loyal
	Protest count		Protest log		Protest binary
Protest item	1.020**	1.010	1.363*	1.061	2.130*	0.974
Protest item	(0.009)	(0.006)	(0.215)	(0.153)	(0.846)	(0.361)
Age	0.957***	1.013	0.958**	1.014	0.958**	1.014
Age	(0.016)	(0.015)	(0.016)	(0.015)	(0.016)	(0.015)
Male	1.536	1.436	1.548	1.414	1.528	1.398
Male	(0.494)	(0.428)	(0.508)	(0.413)	(0.488)	(0.405)
Education	0.647***	0.998	0.646***	1.003	0.650***	1.004
Education	(0.102)	(0.109)	(0.101)	(0.109)	(0.104)	(0.110)
Russian	1.229	2.955	1.202	3.182	1.252	3.354*
Russian	(0.931)	(2.141)	(0.904)	(2.311)	(0.915)	(2.408)
Employed	1.140	1.874	1.131	1.872	1.136	1.892
Employed	(0.404)	(0.995)	(0.403)	(0.989)	(0.406)	(1.006)
Online news	0.905	1.252**	0.893	1.252**	0.895	1.260**
Online news	(0.131)	(0.130)	(0.128)	(0.128)	(0.125)	(0.130)
Protested	0.317	0.495	0.276	0.481	0.237	0.493
Protested	(0.390)	(0.324)	(0.345)	(0.318)	(0.302)	(0.317)
Pocketbook	1.912***	1.714**	1.931***	1.719**	1.905***	1.707**
Pocketbook	(0.388)	(0.424)	(0.394)	(0.428)	(0.394)	(0.417)
Sociotropic	1.483*	1.472	1.511*	1.476	1.564**	1.485
Sociotropic	(0.334)	(0.427)	(0.335)	(0.426)	(0.344)	(0.429)
Urban	1.689	1.844	1.543	1.852	1.584	1.899
Urban	(0.813)	(0.892)	(0.766)	(0.909)	(0.751)	(0.916)
Democracy	1.700**	0.934	1.714**	0.896	1.680**	0.867
Democracy	(0.407)	(0.281)	(0.423)	(0.272)	(0.408)	(0.259)
Constant	0.054	0.001***	0.047*	0.001***	0.041*	0.001***
Constant	(0.099)	(0.002)	(0.086)	(0.002)	(0.074)	(0.002)
Observations	386	386	386	386	386	386

Source. RES 2008, 2012 and Lankina protest-event dataset.

Coefficients are relative risk ratios from multinomial logit models. Respondents staying loyal to the regime, that is, those voting for the ruling-regime candidate in the 2008 and 2012 presidential elections, are the baseline category. Robust standard errors, clustered by administrative units, in parenthesis.

p < .1. p** < .05. ***p < .01.

Models 1.1 and 1.2 use the continuous, nonlogarithmically transformed indicator of protest events. Model 1.1 suggests that every district-based protest increases the risk of disengaging from politics—relative to the risk of remaining loyal to the regime—by a factor of 1.02 given that the other variables in the model are held constant. The protest coefficient in Model 1.2 is also greater than 1, yet fails to reach statistical levels of significance. Contrary to expectations, results provide little evidence that opposition protests also increase defections to the opposition. Similar results are reported in the remaining models, which rely on a logarithmically transformed (Models 2.1 and 2.2) and a binary (Model 3.1 and 3.2) indicator of district protests. The risk of defecting to a nonvote, relative to the risk of remaining loyal to the regime, increases as district-level protests increase in Model 2.1 and is approximately 2 times greater in districts with, as opposed to districts without any protests in Model 3.1. As in Model 1.2, the protest items in Models 2.2 and 3.2, respectively, fail to reach statistical levels of significance. Across all three sets of models, the relative risk of defecting to an opposition party is very close to 1 and statistically indistinguishable from zero.

Turning to the controls, only a handful of significant variables emerge. Older and better educated respondents are more likely to remain loyal to the regime than they are to defect to a nonvote. The finding that better educated respondents are less likely to abandon the ruling regime appears counterintuitive, yet echoes earlier research on the correlates of defecting from Russia’s dominant party, United Russia. As Hale and Colton (2017) argue, this finding could be in line with informational cascade theories. Existing research suggest that individuals with higher education, as they have better information about national politics, may be less likely to update their evaluations of the regime in response to large-scale anti-regime mobilization than the rest of the electorate (Hale & Colton, 2017; Lohmann, 1994 p. 325).

In Model 3.2, the relative risk of defecting to the opposition is also greater by a factor of 3 for ethnic Russians, than it is for respondents of other ethnicities. This could be in part because nonethnic Russians are based in republics that rely more heavily on federal transfers, and where local leaders can more effectively deliver votes for the ruling regime (see also Hale & Colton, 2017, p. 334). Online news watching is correlated with defections to the opposition but not defections to a nonvote. The risk of defecting to an opposition party or candidate, relative to the risk of remaining loyal to the ruling regime, increases as the frequency of watching online news also increases. This finding echoes earlier research, on how politicized social media and online media sources may influence political attitudes in electoral autocracies (Enikolopov, Makarin, & Petrova, 2018; Reuter & Szakonyi, 2015). Across all models, the relative risk of disengaging from politics and defecting to the opposition is greater for respondents who report greater dissatisfaction with their pocketbook. This is in line with research that finds economic grievances to drive support for the 2011-2012 protest movement (Chaisty & Whitefield, 2013). Evaluations of national economic performance are also associated with a higher relative risk of disengaging from politics. Finally, across all three model specifications, the risk of defecting to a nonvote, relative to the risk of remaining loyal to the regime, is greater in more politically competitive regions. It is possible that observing electoral falsifications in the 2011 parliamentary election, voters in more competitive regions, less used to fraud than Russians elsewhere, concluded that participation in elections across the country does not provide meaningful input into the political process and as such withdrew from politics (see also Croke et al., 2016).

As I show in the Supplemental Appendix (Table C1), results remain consistent when we drop “Don’t Know/ No Answer” responses from the “disengagement” category, and when we re-run the analysis using multinomial probit models (Supplemental Appendix Table C2). Just like MLMs, multinomial probit models facilitate comparisons between pairs in a choice set, yet do not impose the independence of irrelevant alternatives assumption. The latter property implies that the removal of one alternative form of a choice set has no influence on individuals’ decisions among the remaining choices (Alvarez & Nagler, 1998). Results are also consistent when we use alternative clustering of the standard errors (Supplemental Appendix Table C3) and replicate the analysis using protest items coded at the regional (oblast), as opposed to the district, level (Supplemental Appendix Table C4). Finally, to partly account for government responses to protests, I rerun the analysis controlling for a variable that captures the use of police-led repression against regional protests (Supplemental Appendix Table C5). This is important, as existing literature reminds us that political violence may discourage risk-averse citizens from supporting opposition parties and activists (Magaloni, 2006, 2010). As the protest and repression items are highly correlated, models that include both indicators simultaneously should be interpreted with caution. Yet in neither model does the inclusion of the repression indicator change the coefficients on the protest variables of interest.

However, a concern with the observational evidence presented here is the potential endogeneity of protests and defections. It could be argued, for example, that the presence of protests is endogenous to voting patterns in the presidential election. Dissatisfaction with the ruling regime or prior support for opposition parties and candidates may be correlated with protests and defections. To address this possibility with the data available, I run several robustness tests, which are presented in the Supplemental Appendix. First, I show that support for the opposition, as captured in the 2008 wave of the RES survey, does not predict the occurrence of district and regional protests in 2011-2012. Next, drawing on the panel component of the 2008 and 2012 RES surveys, I use protests taking place during the 2011-2012 protest cycle to predict support for opposition candidates in the 2008 presidential election as a placebo test. The protest items fail to reach statistical significance. Third, using a nationally and regionally representative survey of public opinion conducted in February 2011, I show that support for United Russia, as captured just 9 months prior to the December 2011 election, does not predict protests between December 2011 and May 2012. Tables D1a-c in the Supplemental Appendix report these results. Altogether, robustness checks provide little support for the argument that the electoral protests of 2011-2012 necessarily occurred in districts or regions where prior dissatisfaction with the ruling regime was greater. Fourth, and as already highlighted, I replicate the analysis using evidence from a different, nationally representative survey of public opinion that was in the field in January 2012. Evidence suggests that as regional protests increased, the probability that voters who supported United Russia in the 2007 parliamentary election would report abstaining from voting in the 2012 election also increased. These results remain consistent when we instrument the frequency of regional protests using the deviation of the monthly average temperature in December 2011 and January 2012 from the long-term average temperature for this period. Tables D4 and D5 in the Supplemental Appendix report the results.

Discussion

There are many possible explanations for the link between protests and defections. While it is not possible, given the available data, to determine with certainty what precise process is at work, it is useful to consider and test two particular mechanisms, extensively discussed in existing literature. The first emphasizes perceptions of electoral injustice. According to a strand in the literature, voter responses to electoral protests depend largely on how they assess the ruling regime’s fraud-related strategy (see, for example, Magaloni, 2010, p. 755). While voters search out and interpret information on fraud in line with their political orientations (Robertson, 2017), it is also expected that when protests help voters to realize that their votes have been stolen, or that fraud has changed the outcome of an election, support for the ruling regime may erode (Beissinger, 2007; Reuter & Szakonyi, 2018; Tucker, 2007). Existing research also proposes that, compared with the rest of the electorate, regime supporters may be particularly responsive to information about fraud, as they have stronger prior beliefs about regime strength and electoral integrity. Drawing on a survey experiment from Russia, for example, Reuter and Szakonyi (2018) show that providing core United Russia supporters with information about fraud significantly reduces their propensity to support ruling party candidates. Yet current scholarship links evaluations of electoral injustice to both types of defection considered here—disengagement and defection to an opposition party. Several studies, for example, suggest that electoral falsifications and perceptions of electoral fraud may convince voters that participation in elections is futile (Birch, 2010; McCann & Domínguez, 1998; Nikolayenko, 2015). Another set of studies, however, links perceptions of electoral injustice to renewed political engagement and greater support for the opposition. In the post-communist colored revolutions, for example, widespread knowledge of electoral fraud fuelled protests and encouraged renewed support for opposition parties and activists (Beissinger, 2007; Tucker, 2007).

A second possible mechanism that links protests to defections emphasizes voter beliefs about the possibility of political reform. As Susanne Lohmann (1994) reminds us (pp. 51-55), people are more willing to send costly informational signals when they believe that their actions could be decisive in bringing about political change (see also Hale & Colton, 2017). Extending the logic of this argument, it is possible to expect that by creating feelings of political empowerment and inspiring confidence in the opposition’s ability to effect political change, protests may convince bystanders that mobilization is meaningful and political change is likely (see also Wallace et al., 2014). As already highlighted, when unfolding protests convince bystanders that participation in protests and elections can make a difference to the situation in the country, defections to the opposition may trump withdrawal from politics. Conversely, when protests fail to inspire hope in the opposition’s ability to deliver, or are unable to convince bystanders that participation is meaningful, withdrawal from politics may dominate defection to the opposition. Empirically testing the two mechanisms constitutes an important step in the study of voter defections in electoral autocracies.

Survey items from the RES allow me to investigate both mechanisms. The first item I rely on asks respondents to assess whether electoral violations changed the outcome of the 2011 Duma election. The question is asked as follows: “Imagine that the December 2011 elections were completely honest, without any kind of violations. What result, in your view, would have been most likely in that case?” I recode responses into a dummy variable that takes the value of one if respondents indicate that in the absence of violations United Russia would not have received a majority of seats in the Duma, and zero if otherwise. The second item asks respondents whether they feel that street protests can make a difference to the situation in the country.⁶ I recode responses into a 5-point scale, with higher values denoting greater confidence in protests’ ability to make a difference in national politics.

Model 1 in Table 3 reports coefficients from a probit model, with assessments of voter fraud in the Duma election as the dependent variable, and the protest-event count as the main independent variable of interest. The protest coefficient suggests that district-level protests increased the predicted probability of reporting that in the absence of electoral violations United Russia would have lost its majority in the Russian Duma. This finding is consistent when we run analysis using the logarithmically transformed and binary indicators of protests instead. Results are reported in Table C8 in the Supplemental Appendix.

Table 3.

Evaluating the Mechanisms.

	(1)	(2.1)	(2.2)	(3)	(4.1)	(4.2)
	Electoral violations (Probit)	Disengage/loyal	Opposition/loyal	Protests matter (OLS)	Disengage/loyal	Opposition/loyal
Protest count	0.010***	1.019**	1.004	0.003	1.020**	1.008
Protest count	(0.004)	(0.009)	(0.007)	(0.003)	(0.009)	(0.006)
Violations		2.814***	9.218***
Violations		(1.114)	(3.868)
Protest matter					1.114	1.289*
Protest matter					(0.112)	(0.172)
Age	0.005	0.956***	1.009	0.011**	0.956***	1.011
Age	(0.007)	(0.016)	(0.015)	(0.005)	(0.016)	(0.016)
Male	0.324*	1.453	1.123	−0.082	1.553	1.459
Male	(0.169)	(0.518)	(0.375)	(0.111)	(0.494)	(0.462)
Education	0.103	0.632***	0.941	−0.029	0.645***	1.005
Education	(0.069)	(0.103)	(0.133)	(0.046)	(0.103)	(0.114)
Russian	0.200	1.197	3.513	−0.341	1.268	3.185
Russian	(0.308)	(0.917)	(3.353)	(0.243)	(0.960)	(2.332)
Employed	0.522*	1.058	1.228	0.022	1.150	1.876
Employed	(0.271)	(0.389)	(0.693)	(0.176)	(0.401)	(1.005)
Online news	0.187***	0.856	1.112	0.043	0.903	1.233**
Online news	(0.070)	(0.120)	(0.118)	(0.069)	(0.131)	(0.129)
Protested past	0.596*	0.237	0.277**	0.628**	0.291	0.428
Protested past	(0.349)	(0.278)	(0.155)	(0.241)	(0.368)	(0.286)
Pocketbook	0.233*	1.863***	1.571*	−0.051	1.927***	1.753**
Pocketbook	(0.132)	(0.381)	(0.402)	(0.121)	(0.388)	(0.406)
Sociotropic	0.234*	1.425	1.294	−0.414***	1.559**	1.649*
Sociotropic	(0.130)	(0.324)	(0.388)	(0.116)	(0.349)	(0.490)
Urban	0.225	1.637	1.747	−0.101	1.728	1.850
Urban	(0.196)	(0.768)	(0.802)	(0.140)	(0.841)	(0.913)
Democracy	0.248**	1.629**	0.769	0.206*	1.671**	0.886
Democracy	(0.120)	(0.389)	(0.255)	(0.107)	(0.398)	(0.265)
Constant	−4.913***	0.090	0.004***	3.231***	0.038*	0.000***
Constant	(0.882)	(0.171)	(0.008)	(0.691)	(0.070)	(0.001)
Observations	386	386	386	386	386	386

Column 1 reports coefficients from a probit model, while Column 3 reports OLS coefficients. The coefficients in Columns 2.1 and 2.2., as well as 4.1 and 4.2, are relative risk ratios from multinomial logit models. Multinomial probit models are reported in Table C9 in the Supplemental Appendix. Respondents staying loyal to the regime, that is, those voting for the ruling-regime candidate in the 2008 and 2012 presidential elections, are the baseline category. Robust standard errors, clustered by primary sampling units in parenthesis. OLS = ordinary least squares.

p < .1. **p < .05. ***p < .01.

Men, employed respondents, and those who more frequently watch news online are also more likely to report that in the absence of fraud United Russia would not have won a parliamentary majority. Respondents who have protested in the past and those who report greater economic pessimism also more likely to report that fraud changed the outcome of the election in favor of United Russia. Finally, respondents in more politically competitive regions are more likely to report that fraud gave United Russia a majority in the 2011 Duma election. For respondents in these areas, not necessarily accustomed to the levels of electoral fraud seen during the Duma 2011 election, electoral violations and protests may have constituted particularly salient and important events.

Moving to Models 2.1 and 2.2, results suggest that perceptions of extensive electoral violations may be associated with both types of defection—to a nonvote and to the opposition. Yet the magnitude of the fraud indicator is considerably larger in Model 2.2. In Model 2.1, the relative risk of disengaging from politics is greater by a factor of 2.8 among respondents who believe that fraud gave United Russia its parliamentary majority. In Model 2.2, the relative risk of defecting to an opposition vote is greater by a factor of 9.2 for respondents who believed that electoral fraud was so extensive that in its absence United Russia would not have won a majority. What is more, controlling for perceptions of electoral violations in Models 2.1 and 2.2 in Table 3 appears to reduce the size of the protest coefficient, as initially reported in Models 2.1 and 2.2 in Table 2. Although of modest magnitude—the relative risk of defecting to a nonvote drops by just .001 and to an opposition vote drops by .006—the reduction of the protest coefficients implies that a modest part of the overall effect of protests on defections may be channeled through perceptions of electoral violations, here in the context of the 2011 election.

Moving on, Model 3 considers whether district-level protests increased support for the statement that opposition protests can make a difference to what happens in Russia. While positive, the protest coefficient fails to reach statistical levels of significance. As shown in Table C8 in the Supplemental Appendix, results do not change if we replicate the analysis using the logarithmically transformed or binary indicators of protest events. This implies that in the context of the 2011-2012 protest wave, opposition protests did not necessarily convince bystanders that protests could bring about political change. Reporting greater confidence in the statement that protests can make a difference are older respondents, those who protested in the past, and those who live in more politically competitive regions. Respondents who report more pessimistic assessments of the Russian economy are less likely to report that protests can make a different to what happens in Russia. In Models 4.1 and 4.2, however, beliefs that protests can make a difference appear to explain differentiation in defections. The risk of defecting to the opposition, relative to that of remaining loyal to the regime, is greater among respondents who report greater confidence in the statement that protests can bring about political change. The risk of disengaging from politics, relative to that of remaining loyal to the regime, is statistically indistinguishable for respondents who differ in their assessments of the effectiveness of protests. These findings are in line with evidence on the determinants of defections from Russia’s party of power, United Russia. As Hale and Colton (2017) have shown (p. 333), voter beliefs about how likely any other party was to come to power in the future were also correlated with defections to the opposition in the context of Russia’s 2011 parliamentary election. Altogether, results suggest that feelings of political empowerment and perceptions of protest effectiveness may be instrumental not only for expanding mobilization during periods of contention (Tarrow, 2011) but also for triggering and sustaining a wave of defections toward the opposition.

Robustness and Extensions

The preceding analysis has shown that opposition protests increase the likelihood that voters who once supported the ruling regime will withdraw support. Contrary to expectations, the results provide little evidence for the argument that, as opposition protests unfold, voters who once supported the ruling regime will also shift to supporting the opposition. Yet it is unclear whether opposition protests also dampened the ruling regime’s ability to attract new supporters, that is, voters who in the 2008 presidential election voted for an opposition candidate or did not vote at all. Investigating this pattern is important, as existing research suggests that prior to the onset of the 2011-2012 protest events, voters who abandoned United Russia were largely replaced by voters who had abstained from politics in the past (Hale & Colton, 2017, p. 334). When ruling regime defectors are replaced by new regime joiners, electoral support for the regime, as well as incumbents’ aggregate approval ratings, will be less affected by the protests. To examine protest effects on the regime’s ability to attract new voters, Table D2 in the Supplemental Appendix conducts an analysis of all respondents in the 2008-2012 panel who did not vote for the ruling-regime candidate in the 2008 presidential election. The outcome variable used in the analysis is a dummy which takes the value of one if respondents who did not vote for Medvedev in 2008 reported voting for Putin in 2012, and zero otherwise. Results suggest that protests did not bolster the ruling regime’s ability to attract new supporters.

Moreover, it is possible that across the electorate, protests increased support for opposition parties and activists who did and did not compete at elections, in ways not captured by the main dependent variable employed in the main part of the analysis. As already noted, Russia’s nonsystemic, or extraparliamentary opposition, played a leading role in organizing protest events throughout the 2011-2012 electoral protest cycle (Sakwa, 2014). Similarly, it could be that while bystanders concluded that participation in elections is futile and disengaged, they also learnt that opposition parties or activists are credible and trustworthy. If this were the case, we would expect opposition protest to be associated with both electoral disengagement and an increase in attitudinal support for the opposition. To explore this possibility, I present analysis that relies on an alternative set of outcome variables from the 2012 RES survey. Unfortunately, neither of these items asks respondents directly whether they approve of the opposition or not. Put together, however, they could serve as helpful proxies of voter sentiment toward the opposition.

The first set of items asks respondents to assess whether the time of several antisystemic, non-parliamentary protest leaders is “now or to come,” or whether their time has passed. I code responses to these questions as one if respondents report more hopeful assessments of the opposition and zero if otherwise. A second set of items asks respondents to identify which presidential candidate would do a better job at advancing Russian’s economic and international interests. The third and final set of items I use captures evaluations of opposition politicians, as either intelligent and knowledgeable, strong, honest and trustworthy, or caring about the interests of people like them. Results presented in Tables D3 (a-e) in the Supplemental Appendix provide no consistent evidence that protests increased attitudinal support for the opposition. To conclude, while not offering evidence of ruling-regime defections, which is the main question of this article, additional analysis offers a more rounded picture of protest effects on public opinion toward the regime and its opponents.

Conclusion

A burgeoning literature recognizes the threat that mass demonstrations pose to the stability of electoral autocracies. Yet the microlevel mechanisms by which protests influence voter defections have remained poorly understood. Detailed datasets on how political attitudes shift in response to unfolding unrest, like the ones presented in this study, do not come easily in many illiberal regimes. In a similar vein, there are only very few empirical tests of the hypothesis that voters who once supported the ruling regime will shift to supporting its opponents.

Leveraging evidence from new protest data and individual-level panel surveys of Russian voters, this article has sought to fill gaps in our understanding of the links between protests and voter defections in non-democracies. To the best of my knowledge, this constitutes one of the first works to provide direct evidence of protest effects on voter defections under illiberal rule. The principal contribution of this research has been to show politically important differentiation in the form voter defections can take in these regimes. Results suggest that opposition protests dampened support for the ruling regime and increased the likelihood that those who voted for the ruling-regime candidate in 2008 would abstain from voting in 2012. Yet the analysis finds little support for the argument that protests also increased support for the regime’s challengers, the opposition parties, and activists who participated in the protests and made information about regime abuses publicly available.

We could ask how generalizable the findings from the Russian case are likely to be. The type of learning that underpins the arguments presented here may be applicable on a range of cases where there is high general uncertainty about the government, the opposition, and the prospect of political change more broadly. In these contexts, information about the regime and its opponents, made publicly observable as protests unfold, is rare and could have important effects on voter attitudes and preferences. Opposition parties’ and activists’ ability to form coalitions at times of protests and elections also matters and could influence voter beliefs about the likelihood of political change. When voters infer that participation is futile and that political change is improbable, support for the opposition is unlikely to grow. Disillusionment with incumbents and the opposition has been associated with disengagement beyond Russia. For example, drawing on evidence from Zambia’s 1996 general election, Posner and Simon (2002) have demonstrated that voters who were dissatisfied with the economy were more likely to withdraw from politics than to support the opposition. Using field experiments in Mexico, Chong et al. (2015) have shown that unfavorable information about the incumbent may lead to disengagement. Altogether, when voters conclude that political change is unlikely and that participation is futile, disengagement could trump defections to the opposition. Implications follow for the literature on the legacies of attempted revolutions. Voter coordination against illiberal incumbents—just like voter coordination against abusive governments that come to power after democratization—will be more difficult to achieve when opposition parties and activists fail to convince the masses that they are able to credibly challenge the ruling regime, or to offer a better alternative (Meirowitz & Tucker, 2013; Tarrow, 2011).

Just as this analysis brings attention to significant differentiation in voter defections in electoral autocracies and provides empirical evidence of relevance to ongoing debates in the comparative democratization literature, it also raises questions that this article has not been able to answer. Future research should, for example, examine in greater detail whether or how state responses to protests may shape differentiation in defections. For example, could state-controlled media exacerbate feelings of fear or anxiety that could lead to disengagement? And could targeted repression direct defections toward withdrawal rather than toward actual opposition? Additional evidence is also needed to identify how costly yet unsuccessful attempts at regime change affect long-term support for the regime and its opponents across the electorate, and how experiences of “failed” demonstrations shape participation in ensuing unrest. Identifying the persistence of protest effects on political attitudes in electoral autocracies also constitutes a fruitful avenue for future research. Finally, scholarship on authoritarian politics would benefit from investigating how ruling-regime defectors, who have disengaged once, behave in subsequent elections. For example, existing research shows that individuals who deliberately disengage from politics in electoral autocracies are more likely to reengage during more competitive elections (Croke et al., 2016). Whether regime defectors will reengage in politics and side with the ruling regime, or switch to supporting the opposition instead, is of utmost importance for questions of authoritarian regime stability and resilience. Despite these limitations, this work constitutes one of the first attempts to directly study protest effects on voter defections in a non-democracy. Findings highlight the importance of considering how, in addition to, or instead of, generating support for the opposition, anti-regime protests may also demobilize a large share of the population. They also suggest that in trying to better understand the underpinnings of authoritarian stability and the legacies of (failed) revolutions, we should pay greater attention to important differentiation in the form defection cascades can take in non-democracies.

Supplemental Material

2012_data_no_labels – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, 2012_data_no_labels for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

Codebook – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, Codebook for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

fom_replication – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, fom_replication for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

january_replication – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, january_replication for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

Online_Appendix – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, Online_Appendix for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

protest_events – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, protest_events for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

readme – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, readme for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Supplemental Material

replication – Supplemental material for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia

Supplemental material, replication for Protests and Voter Defections in Electoral Autocracies: Evidence From Russia by Katerina Tertytchnaya in Comparative Political Studies

Footnotes

Acknowledgements

AcknowledgmentsI wish to thank Selim Erdem Aytaç, Tomila Lankina, Anja Neundorf, Grigo Pop-Eleches, Regina Smyth as well as seminar participants at the University of Oxford, LSE, Indiana University Europe Gateway, and the 2018 meetings of the European and American political science associations for insightful feedback on previous versions of this article. I am also grateful to the editors and three anonymous reviewers for their helpful comments and suggestions.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Katerina Tertytchnaya

Supplemental Material

Supplemental material for this article is available online at the CPS website

Notes

Author Biography

Katerina Tertytchnaya is a lecturer in Comparative Politics at University College London. Her research focuses on authoritarian politics and political behaviour.

References

Alvarez

R. M.

Nagler

(1998). When politics and models collide: Estimating models of multiparty elections. American Journal of Political Science, 42, 55-96.

Armstrong

Reuter

O. J.

Robertson

(2017). Getting the opposition together: Protest coordination in authoritarian regimes. Retrieved from http://ojreuter.com/wp-content/uploads/GOT_2017.pdf

Atkeson

L. R.

Maestas

C. D.

(2012). Catastrophic politics: How extraordinary events redefine perceptions of government. Cambridge, UK: Cambridge University Press.

Aytaç

S. E.

Shiumerini

Stokes

(2018). Why do people join backlash protests? Journal of Conflict Resolution, 62, 1205-1228.

Beissinger

M. R.

(2007). Structure and example in modular political phenomena: The diffusion of Bulldozer/Rose/Orange/Tulip Revolutions. Perspectives on Politics, 5, 259-276.

Birch

(2010). Perceptions of electoral fairness and voter turnout. Comparative Political Studies, 43, 1601-1622.

Branton

Martinez-Ebers

Carey

T. E.

Jr. Matsubayashi

(2015). Social protest and policy attitudes: The case of the 2006 immigrant rallies. American Journal of Political Science, 59, 390-412.

Bratton

Van de Walle

(1994). Neopatrimonial regimes and political transitions in Africa. World Politics, 46, 453-489.

Chaisty

Whitefield

(2013). Forward to democracy or back to authoritarianism? The attitudinal bases of mass support for the Russian election protests of 2011–2012. Post-Soviet Affairs, 29, 387-403.

10.

Chong

De La O

A. L.

Karlan

Wantchekon

(2015). Does corruption information inspire the fight or quash the hope? A field experiment in Mexico on voter turnout, choice and party identification. The Journal of Politics, 77, 55-71.

11.

Colton

T. J.

Hale

H. E.

(2014). Putin’s uneasy return and hybrid regime stability. Problems of Post-Communism, 61, 3-22.

12.

Colton

T. J.

Hale

H. E.

Kosolapov

Tonis

Prudnikova

(2014). RES technical report_2012_final.doc. In Russian Post-Presidential Election Survey in 2012. doi:10.7910/DVN/24202/I3WXOQ

13.

Croke

Grossman

Larreguy

H. A.

Marshall

(2016). Deliberate disengagement: How education can decrease political participation in electoral authoritarian regimes. American Political Science Review, 110, 579-600.

14.

DeNardo

(1985). Power in numbers: The political strategy of protest and rebellion. Princeton, NJ: Princeton University Press.

15.

Domínguez

McCann

(1996). Democratizing Mexico. Baltimore, MD: John Hopkins University Press.

16.

Enikolopov

Makarin

Petrova

(2018). Social media and protest participation: Evidence from Russia. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2696236

17.

Frye

Borisova

(forthcoming). Elections, protest and trust in government: A natural experiment from Russia. The Journal of Politics. Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2827551

18.

Frye

Gehlbach

Marquardt

K. L.

Reuter

O. J.

(2016). Is Putin’s popularity real? Post Soviet Affairs, 33, 1-15.

19.

Gamson

(2004). Bystanders, public opinion and the media. In Snow

Soule

Kriesi

(Eds.), Blackwell companion to social movements (pp. 242-261). Malden, MA: Blackwell.

20.

Gel’man

(2015). Authoritarian Russia: Analyzing post-soviet regime changes. Pittsburgh, PA: University of Pittsburgh Press.

21.

Greene

S. A.

(2013). Beyond Bolotnaia: Bridging old and new in Russia’s election protest movement. Problems of Post-Communism, 60(2), 40-52.

22.

Guriev

Treisman

(2015). How modern dictators survive: An informational theory of the new authoritarianism. Retrieved from http://www.nber.org/papers/w21136

23.

Hale

H. E.

Colton

T. J.

(2017). Who defects? Unpacking a defection cascade from Russia’s dominant party 2008–12. American Political Science Review, 111, 322-337.

24.

Hollyer

J. R.

Peter Rosendorff

Vreeland

J. R.

(2015). Transparency, protest, and autocratic instability. American Political Science Review, 109, 764-784.

25.

Kahneman

Slovic

Tversky

(1982). Judgment under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge University Press.

26.

Karklins

(1986). Soviet elections revisited: Voter abstention in noncompetitive voting. American Political Science Association, 80, 449-470.

27.

Kuran

(1991). Now out of never: The element of surprise in the East European Revolution of 1989. World Politics, 44, 7-48.

28.

Kuran

(1997). Private truths, public lies: The social consequences of preference falsification. Cambridge MA: Harvard University Press.

29.

Lankina

Voznaya

(2015). New data on protest trends in Russia’s regions. Europe-Asia Studies, 67, 327-342.

30.

Lasnier

(2017). Demobilisation and its consequences: After the Russian Movement Za chestnye vybory. Europe-Asia Studies, 67, 771-793.

31.

Lasnier

(2018). Russia’s opposition movement five years after Bolotnaia: The electoral trap? Problems of Post-Communism, 65, 359-371.

32.

Levada Analytical Centre. (2012). Russian public opinion: 2010-2011. Moscow, Russia: Author.

33.

Lohmann

(1994). The dynamics of informational cascades: The Monday demonstrations in Leipzig, East Germany, 1989–91. World Politics, 47, 42-101.

34.

Madestam

Shoag

Veuger

Yanagizawa-Drott

(2013). Do political protests matter? Evidence from the tea party movement. The Quarterly Journal of Economics, 128, 1633-1685.

35.

Magaloni

(2006). Voting for autocracy: Hegemonic party survival and its demise in Mexico. Cambridge, UK: Cambridge University Press.

36.

Magaloni

(2010). The game of electoral fraud and the ousting of authoritarian rule. American Journal of Political Science, 54, 751-765.

37.

McAdam

(2002). The war at home: Antiwar protests and congressional voting, 1965 to 1973. American Sociological Review, 67, 696-721.

38.

McAdam

Tarrow

(2010). Ballots and barricades: On the reciprocal relationship between elections and social movements. Perspectives on Politics, 8, 529-542.

39.

McCann

J. A.

Domínguez

J. I.

(1998). Mexicans react to electoral fraud and political corruption: An assessment of public opinion and voting behavior. Electoral Studies, 7, 483-503.

40.

Meirowitz

Tucker

J. A.

(2013). People power or a one-shot deal? A dynamic model of protest. American Journal of Political Science, 57, 478-490.

41.

Morgenstern

Zechmeister

(2001). Better the devil you know than the saint you don’t? Risk propensity and vote choice in Mexico. Journal of Politics, 63, 93-119.

42.

Nikolayenko

(2015). Do contentious elections depress turnout? In Norris

Frank

R. W.

Coma

F. M. i

(Eds.), Contentious elections: From ballots to barricades (pp. 25-44). New York, NY: Routledge.

43.

Onuch

(2015). EuroMaidan protests in Ukraine: Social media versus social networks. Problems of Post-communism, 62, 217-235.

44.

Petrov

Titkov

(2013). Reiting Demokratichnosti Regionov Moskovskogo Tsentra Karnegi: 10 let v stroyu [Democracy ratings of Russian regions by the Carnegie Moscow Center: 10 years in Service]. Moscow, Russia: Moscow Carnegie Center.

45.

Posner

D. N.

Simon

D. J.

(2002). Economic conditions and incumbent support in Africa’s new democracies: Evidence from Zambia. Comparative Political Studies, 35, 313-336.

46.

Reuter

O. J.

Robertson

(2015). Legislatures, cooptation, and social protest in contemporary authoritarian regimes. The Journal of Politics, 77, 235-248.

47.

Reuter

O. J.

Szakonyi

(2015). Online social media and political awareness in authoritarian regimes. British Journal of Political Science, 45, 29-51.

48.

Reuter

O. J.

Szakonyi

(2018, September). Electoral manipulation and regime support: Survey evidence from Russia. Paper presented at the annual meeting of the American Political Science Association. Boston, MA.

49.

Ritter

E. H.

Conrad

C. R.

(2016). Preventing and responding to dissent: The observational challenges of explaining strategic repression. American Political Science Review, 110, 85-99.

50.

Robertson

(2013). Protesting Putinism. Problems of Post-Communism, 60(2), 11-23.

51.

Robertson

(2017). Political orientation, information and perceptions of election fraud: Evidence from Russia. British Journal of Political Science, 47, 589-608.

52.

Rosenfeld

(2017). Reevaluating the middle-class protest paradigm: A case-control study of democratic protest coalitions in Russia. American Political Science Review, 111, 637-652.

53.

Ross

(2015). Systemic and non-systemic opposition in the Russian federation: Civil society awakens? Farnham, UK: Ashgate.

54.

Sakwa

(2014). Whatever happened to the Russian opposition? London, England: Chatham House.

55.

Smyth

(2014). The Putin Factor: Personalism, protest, and regime stability in Russia. Politics & Policy. 42, 567-592.

56.

Smyth

Soboleva

(2016). Navalny’s gamesters: Protest, opposition innovation, and authoritarian stability in Russia. Russian Politics, 1, 347-371.

57.

Spaiser

Chadefaux

Donnay

Russmann

Helbing

(2017). Communication power struggles on social media: A case study of the 2011–12 Russian protests. Journal of Information Technology and Politics, 14, 132-153.

58.

Steinert-Threlkeld

Z. C.

(2017). Spontaneous collective action: Peripheral mobilization during the Arab spring. American Political Science Review, 111, 379-403.

59.

Stockmann

Gallagher

M. E.

(2011). Remote control: How the media sustain authoritarian rule in China. Comparative Political Studies, 44, 436-467.

60.

Svolik

(2012). The politics of authoritarian rule. New York, NY: Cambridge University Press.

61.

Tarrow

S. G.

(2011). Power in movement: Social movements and contentious politics. Cambridge, UK: Cambridge University Press.

62.

Tertytchnaya

Lankina

(2020). Electoral protests and political attitudes under electoral authoritarianism. The Journal of Politics, 82(1), 285-299. Retrieved from http://eprints.lse.ac.uk/90262/

63.

Truex

(2019). Focal points, dissident calendars, and preemptive repression. Journal of Conflict Resolution, 63, 1032-1052.

64.

Tucker

J. A.

(2007). Enough! Electoral fraud, collective action problems, and post-communist colored revolutions. Perspectives on Politics, 5, 535-551.

65.

Volkov

(2012). The protesters and the public. Journal of Democracy, 23(3), 55-62.

66.

Wallace

S. J.

Zepeda-Millán

Jones-Correa

(2014). Spatial and temporal proximity: Examining the effects of protests on political attitudes. American Journal of Political Science, 58, 433-448.

67.

Weyland

(2012). The Arab spring: Why the surprising similarities with the revolutionary wave of 1848? Perspectives on Politics, 10, 917-934.

68.

White

(2015). Authoritarian elections: The Russian case, 2011-2012 [Data collection]. Colchester, UK: Data Archive.

69.

Young

(2019). The psychology of state repression: Fear and dissent decisions in Zimbabwe. American Political Science Review, 113, 140-155.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB

0.08 MB

0.43 MB

1.10 MB

0.07 MB

0.05 MB

0.02 MB

0.01 MB