Voter Turnout in a Dominant

Taiwan Journal of Democracy, Volume 11, No. 2: 25-52
Voter Turnout in a Dominant-Party System
An Analysis of Electoral Participation in Russia’s Regions
Joseph W. Robbins and Mikhail Rybalko
Abstract
This study employs an original dataset of Russia’s titular regions (also known
as ethnic homelands) to understand better how Russia’s party of power
influences the inclinations of individual voters to cast their ballots. The essay
argues that the presence of Edinaya Rossiya (ER, or United Russia) depresses
voter turnout in subnational races due to its dominance, which affects the
polity in numerous ways. Specifically, dominant parties may scare off quality
challengers, result in lower “stakes” for an election, and distort the final
electoral outcomes, which together discourage individuals from voting. The
study’s pooled cross-sectional analysis finds support for this argument. The
essay concludes with a discussion of the study’s implications for Russia’s
fragmented opposition parties.
Keywords: Dominant-party systems, voter turnout, regional politics, Russia.
Although
voter turnout in democracies has been extensively studied,
research on voter turnout in nondemocracies or authoritarian regimes is quite
a different matter. As earlier works have shown, leaders of nondemocracies
increasingly are embracing the façade of electoral politics to extend regime
longevity and for a host of other reasons.1 While “electoral authoritarian”
regimes2 have garnered considerable scholarly attention, examinations of
Joseph W. Robbins is Assistant Professor in the Department of Political Science at Shepherd
University, Shepherdstown, West Virginia. <[email protected]>
Mikhail Rybalko is Assistant Professor in the Department of Oriental Languages at Moscow
Linguistic State University, Irkutsk. <[email protected]>
1Jennifer Gandhi and Adam Przeworski, “Cooperation, Cooptation, and Rebellion under
Dictatorships,” Economics and Politics 18, no. 1 (2006): 1-26; Ellen Lust-Okar, “Elections under
Authoritarianism: Preliminary Lessons from Jordan,” Democratization 13, no. 3 (2006): 456471; and Michael K. Miller, “Electoral Authoritarianism and Democracy: A Formal Model of
Regime Transitions,” Journal of Theoretical Politics 25, no. 2 (2013): 153-181.
2Andreas Schedler, Electoral Authoritarianism: The Dynamics of Unfree Competition (Boulder,
CO: Lynne Rienner Publishers, 2006).
December 2015 | 25
voter turnout patterns in these systems have been scant. This is particularly
troubling since electoral politics in these regimes differ in various ways from
those in democratic societies. Studying the contests in nondemocracies also
can offer insight into how dominant-party regimes survive, as existing work
has focused on other important topics related to dominant parties. Accordingly,
this study examines subnational voter turnout rates in Russia for theoretical
and empirical reasons.
From a theoretical perspective, understanding what accounts for the
variations in voter turnout is particularly important, as the factors underlying
turnout are likely to be different in electoral authoritarian systems from those
in democracies. Research on democratic polities often finds that electoral
systems, federalist systems, and party systems frequently shape variations
in turnout. Yet, because elections in electoral authoritarian systems often are
influenced by dominant parties, applying the lessons learned from turnout in
democracies may provide limited insight. Existing work on voter turnout in
democracies finds that party fragmentation and stability can affect aggregate
turnout levels, but neither finding helps us to understand the variations in the
Russian context (as shown below). Because dominant parties play an important
role in electoral authoritarian systems like Russia’s, more attention is needed
to understand how such parties affect participation. Not only will this improve
our understanding of turnout in these systems but also it will provide insight as
to how voters respond to contests featuring dominant-party candidates.
Furthermore, there are various empirical motivations for a study such as
this. Central to this study is the fact that, in a nondemocracy such as Russia,
there is a surprising amount of variation in turnout rates in the titular republics
(also known as ethnic homelands). As shown in figure 1, the mean voter
turnout for the years and republics under review is just over 50 percent (51.63
percent); meanwhile, the Jewish Oblast (just over 39 percent) and Tatarstan
(approximately 34 percent) reported the lowest participation, while turnout
was highest in Bashkortostan (65.60 percent), Karachay-Cherkessia (67.68
percent), and Mordoviya (87.63 percent). This evidence demonstrates that
variations are present among these regions, even though our understanding
of what has produced these variations has been overlooked. Although we
must proceed with some degree of caution given the litany of electoral abuse
accusations leveled against Russia, it is intriguing that considerable variation
in turnout marks the country’s subnational elections.
Why do such variations exist? And, why do so many regions have
considerably lower turnout levels, if it is in the best interest of the regional
leadership to report above-average turnout rates? If high turnout rates indicate
that the party enjoys widespread support,3 the lower turnout rates in some
3Alan
Arian and Samuel H. Barnes, “The Dominant Party System: A Neglected Model of
Democratic Stability,” Journal of Politics 36, no. 3 (1974): 592-614.
26 | Taiwan Journal of Democracy, Volume 11, No. 2
Russian regions potentially could generate some concern, especially since a
number of Russia’s regions appear to deviate from this pattern.4 Furthermore,
significant regional differences are present among Russia’s regions with respect
to the ubiquity of Edinaya Rossiya (ER) candidates and voter turnout levels.
For instance, the spacious Arctic Nenets region regularly has high turnout
rates but few ER candidates competing in its elections. A similar pattern can
be found in the Republic of Karachay-Cherkessia, which is an agriculturally
dependent, underdeveloped region in the Caucuses. Conversely, other areas
have patterns that are quite different. The Republic of Chuvashiya, which
is located in Western Russia and contains much of the Volga River valley,
generally has many more ER candidates competing in its electoral races that
customarily yield low turnout levels. Likewise, Adygeya Republic, another
Caucasian republic that is agriculturally dependent, has patterns similar to
Chuvashiya’s. However, even though both Adygeya and Karachay-Cherkessia
are in the Caucuses region, the two areas exhibit noticeably different electoral
patterns. This discrepancy suggests that region alone cannot account for the
patterns, and that domestic politics, as explained below, help shape variations
in voter turnout.
Figure 1. Mean Voter Turnout by Republic
The broader relevance of this study is how dominant parties sustain
themselves. As noted, the vote share for Russia’s dominant party, Edinaya
Rossiya, has declined steadily, despite scholarly expectations of continued
4Jorge
I. Dominguez, “The Transformation of Mexico’s Electoral and Party Systems, 19881997,” in Toward Mexico’s Democratization, ed. Jorge I. Dominguez and Alejandro Poire
(London: Routledge, 1999), and David White, “Dominant Party Systems: A Framework for
Conceptualizing Opposition Strategies in Russia,” Democratization 18, no. 3 (2011): 655-681.
December 2015 | 27
dominance.5 Thus, understanding the dynamics at the national and subnational
levels is increasingly valuable. The other pressing reason to examine
subnational voter turnout is that, despite the wide variation in turnout levels
spatially and temporally, researchers often overlook regional turnout patterns
not only in Russia but also in many other countries. To be sure, there are
numerous studies of turnout in the context of subnational contests in the United
States,6 but far fewer beyond this country. Therefore, this study uses lessons
learned from the dominant-party literature to gain theoretical leverage on the
variations in Russian subnational turnout. Given our theoretical motivation, we
view Russia as an intriguing test case for learning more about voting behavior
in a dominant-party system.
To account for the variations in the subnational elections, we posit
related explanations that build on existing knowledge in the voter-turnout
and dominant-party literature. Our arguments also are based on the recent
developments regarding ER and Russian gubernatorial positions. Regarding
the former, we contend that the level of voter turnout falls in the presence of an
ER candidate. The reasons for this have to do with the advantages associated
with dominant parties-which are explained in the following pages. We argue
that Russia’s party of power has had a resounding effect on voter turnoutnot to mention many other forms of political participation. For reasons
detailed below, Russia’s dominant-party system has played and continues to
play an important role in politics at virtually all levels. However, this is not
to suggest that the party is ubiquitous; in compiling our unique dataset, we
have discovered some important trends that, we believe, help account for the
corresponding variation in voter turnout. More to the point, dominant-party
systems often stack the electoral deck in their favor and this, we argue, in all
probability has a discouraging effect on the electorate. That is, the presence
of an ER candidate likely signals that a race is marginally competitive due to
the benefits associated with this dominant party, which leads some citizens to
abstain from voting. Dominant parties generally benefit in electoral contests
in which the opposition candidates or parties have little realistic chance of
victory. Given this, when voters are faced with unsatisfactory options on the
ballot, they may choose to vote for the candidate or party that has the best
likelihood of success. However, they may choose another option: abstention
from the voting process altogether. This option should not be overlooked, for
it can pose problems to the ruling government. The results from a series of
statistical analyses affirm that the presence of an ER candidate depresses voter
turnout, even when controlling for a host of ancillary factors.
5Regina
Smyth, Anna Lowry, and Brandon Wilkening, “Engineering Victory: Institutional
Reform, Informal Institutions, and the Formation of a Hegemonic Party Regime in the Russian
Federation,” Post-Soviet Affairs 23, no. 3 (2007): 118-137.
6For an overview, see Benny Geys, “Explaining Voter Turnout,” Electoral Studies 25, no. 4
(2006): 637-663.
28 | Taiwan Journal of Democracy, Volume 11, No. 2
Regarding the effect of Russian governors on turnout, we anticipate a
different relationship. Since Russian governors are an extension of Putin’s
leadership, we anticipate that governors affiliated with the ER will work to
secure high turnout rates in their regions because their increased mobilization
efforts can boost their standing in the eyes of the populace or central
leadership.7 As previously mentioned, high turnout rates can signal greater
legitimacy to some interested observers, thus governors have an incentive to
report high turnout rates. Our final series of tests analyzes and interprets the
complex dynamic when both conditions are present. Specifically, we test for
an interactive effect in the final series of tests, by means of which we examine
the influence on voter turnout when ER candidates are on the ballot for races
in regions controlled by ER governors. The results of this test indicate that
turnout is particularly low, which we attribute to individuals’ being even more
likely to abstain from voting when the competitiveness of races is affected
by the ER in multiple ways. This finding is explored in the essay’s section on
results.
The study’s findings speak to the effects of dominant-party systems on
voter turnout, a relationship that has been underexplored in the past. The
research also contributes two additional important findings: (1) dominant
parties tend to prefer high turnout levels for the signals such levels send, and
(2) high turnout levels have fallen, even in the wake of some mobilization
efforts throughout Russia.8 Thus, the findings of this study represent a notable
departure from what one might expect, given some of the existing research on
dominant-party systems and the mobilization efforts exercised by ER forces.
Voter Turnout Explanations
The research on voter turnout is extensive and has examined several topics over
a long period of time. Culture, economics, and various political-institutional
factors are among the leading influences on voter turnout.9 Other work has
7Timothy
Frye, Ora John Reuter, and David Szakonyi, “Political Machines at Work: Voter
Mobilization and Electoral Subversion in the Workplace,” World Politics 66, no. 2 (2014): 195228.
8Smyth et al., “Engineering Victory.”
9Carolina Fornos, Timothy J. Power, and James C. Garand, “Explaining Voter Turnout in
Latin America, 1980-2000,” Comparative Political Studies 37, no. 8 (2004): 909-940; Robert
Jackman, “Political Institutions and Voter Turnout in the Industrial Democracies,” American
Political Science Review 81, no. 2 (1987): 405-424; Robert Jackman and Ross A. Miller, “Voter
Turnout in the Industrial Democracies during the 1980s,” Comparative Political Studies 27,
no. 4 (1995): 467-492; Tatiana Kostadinova, “Do Mixed Electoral Systems Matter?: A CrossNational Analysis of Their Effects in Eastern Europe,” Electoral Studies 21, no. 1 (2002): 2334; Tatiana Kostadinova and Timothy Power, “Does Democratization Depress Participation?:
Voter Turnout in the Latin American and Eastern European Democratic Transitions,” Political
Research Quarterly 60, no. 3 (2007): 363-377; Alexander C. Pacek and Benjamin Radcliff,
“Economic Voting and the Welfare State: A Cross-National Analysis,” Journal of Politics 57,
no. 1 (1995): 44-61; Guy Bingham Powell Jr., “Voting Turnout in Thirty Democracies: Partisan,
December 2015 | 29
found that party systems impact voter turnout. A number of studies have found
that well-organized parties generate high turnout levels,10 while party systems
in developing democracies-due to their weaker organizations-often correlate
with low turnout rates. Some studies have further tested, and affirmed, the
relationship between turnout and party systems in other contexts.11 Each of
these general research strands is elaborated in the remainder of this section.
Cultural Impacts
Cultural traits-such as education levels, ethnic cleavages, and urbanizationhave been linked with voter turnout in earlier works. Ethnicity often is seen as
a source of identification that, in some cases, serves as a short-cut in candidate
selection by voters. That is, voters with certain identifying characteristics
often are inclined to support political candidates with similar backgrounds,
particularly if little else is known about the candidates. In fact, cohesive ethnic
cleavages are expected to boost voter turnout, as politicians rely on this type of
grouping as a mobilized voting bloc.
As for other cultural indicators, such as urbanization, the literature often
finds that there are various reasons why turnout declines in densely populated
areas. One reason is that citizens in urban areas tend to be individualistic and
less subject to social pressure to vote, leading to lower turnout rates than in
nonurban regions.12 Other reasons for this relationship are that politics in
more sparsely populated areas may be more personal, or voters may know
more about the candidates in rural areas; thus, lower information costs lead
to higher turnout rates in various instances.13 This is particularly noteworthy
in the Russian context, as significant regional differences exist with respect to
political preferences.14
Legal, and Socioeconomic Influences,” in Electoral Participation: A Comparative Analysis,
ed. Richard Rose (Beverly Hills, CA: Sage, 1980), 5-34; id., “American Voter Turnout in
Comparative Perspective,” American Political Science Review 80, no. 1 (1986): 17-43; and
Steven J. Rosenstone, “Economic Adversity and Voter Turnout,” American Journal of Political
Science 26, no. 1 (1982): 397-430.
10David M. Farrell and Paul Webb, “Political Parties as Campaign Organizations,” in Parties
without Partisans: Political Change in Advanced Industrial Democracies, ed. Russell J. Dalton
and Martin P. Wattenberg (Oxford: Oxford University Press, 2002), 102-128.
11Michael Bratton, “Political Participation in a New Democracy: Institutional Considerations
from Zambia,” Comparative Political Studies 32, no. 5 (1999): 549-588; Michelle Kuenzi
and Gina Lambright, “Voter Turnout in Africa’s Multiparty Regimes,” Comparative Political
Studies 40, no. 6 (2007): 665-690; James H. Polhemus, “Botswana Votes: Parties and Elections
in an African Democracy,” Journal of Modern African Studies 21, no. 3 (1983): 397-430; and
Joseph W. Robbins and Lance Y. Hunter, “Impact of Electoral Volatility and Party Replacement
on Voter Turnout Levels,” Party Politics 18, no. 6 (2012): 919-939.
12Geys, “Explaining Voter Turnout,” and William H. Riker and Peter C. Ordeshook, “A Theory of
the Calculus of Voting,” American Political Science Review 62, no. 1 (1968): 25-42.
13Robert H. Blank, “Socio-economic Determinism of Voting Turnout: A Challenge,” Journal of
Politics 36, no. 3 (1974): 731-752.
14Ralph S. Clem and Peter R. Craumer, “Regional Patterns of Political Preference in Russia: The
December 1999 Duma Elections,” Post-Soviet Geography and Economics 41, no. 1 (2000):
1-29.
30 | Taiwan Journal of Democracy, Volume 11, No. 2
Economic Conditions
Although economic considerations arguably have an important impact on
political participation, they have been neglected in several studies on voter
turnout.15 In earlier work on political participation, when included, economic
conditions have been approached from several different perspectives.
Scholars have examined turnout levels in conjunction with unemployment
rates, GDP rates, GDP growth rates, and income growth levels. While a
number of measures have been used, no consistent link has been established
between underlying economic conditions and voter turnout. One exception to
this is Rosenstone’s well-known work,16 which finds that subpar economic
conditions depress voter turnout because they lead voters to focus on more
pressing matters (i.e., survival, financial sustainability, and so on) than political
participation. However, other scholars have found that economics often have
little discernible effect on voter turnout.17 Even though the significance of
economics is inconsistent, we tested for such effects in a couple of different
ways (described below) and anticipated that favorable economic conditions
should boost turnout levels.
Electoral Systems and Other Political Institutions
Numerous studies have examined how the context of political institutions, such
as size of the legislature, concurrent elections, compulsory laws, federalism,
presidentialism, and electoral system design, affect voter turnout.18 Earlier
work has surmised that the size of a municipality is inversely related to voter
turnout.19 Other studies have found that executive structures affect turnout as
well. In the United States, for instance, presidential elections engender higher
turnout levels than other types of elections. Furthermore, work by Siaroff and
Merer20 finds that turnout rates tend to be lower in states with directly elected
presidents and powerful regional executives.
Much of the work on the institutional causes of voter turnout has gone to
great lengths to study the relationship between electoral systems and turnout.
Likewise, our focus in this section is on electoral systems and their effects on
15André
Blais, To Vote or Not to Vote? The Merits and Limits of Rational Choice (Pittsburgh, PA:
University of Pittsburgh Press, 2000).
16Rosenstone, “Economic Adversity.”
17André Blais and Agnieszka Dobrzynska, “Turnout in Electoral Democracies,” European
Journal of Political Research 33, no. 2 (1998): 239-261; Blais, To Vote or Not to Vote?; Tatiana
Kostadinova, “Voter Turnout Dynamics in Post-Communist Europe,” European Journal of
Political Research 42, no. 6 (2003): 741-759; and Fornos et al., “Explaining Voter Turnout in
Latin America, 1980-2000.”
18See Geys’s work for an overview.
19Robert A. Dahl and Edward R. Tufte, Size and Democracy (Stanford, CA: Stanford University
Press, 1973).
20Alan Siaroff and John W. A. Merer, “Parliamentary Election Turnout in Europe since 1990,”
Political Studies 50, no. 5 (2002): 916-927.
December 2015 | 31
voters and party systems, as they are integral to understanding variations in
turnout. Such electoral system research generally has been split into one of
two camps. The first camp finds that proportional representation (PR) systems
boost turnout levels. This conclusion is rooted in both theory21 and empirics.22
At the aggregate-level, the idea is that PR systems lead to higher turnout rates
than in single member districts (SMDs) because large electoral districts tend to
have more prominent contests with greater nationwide appeal. Consequently,
turnout is higher because the perception is that there is more at stake. Compared
with elections in PR districts, those in SMDs tend to be less competitive, with
fewer people voting in those contests which are more one-sided.23
Multimember PR electoral systems regularly have been linked with higher
turnout numbers than in SMD systems for various reasons. Some scholars argue
that this is because political parties mobilize more fervently in all electoral
districts. In order to secure representation, parties competing in multimember
PR contests do not need to win as many votes as in a plurality system.
Accordingly, campaigning everywhere is an effective strategy, especially
when other electoral rules permit pooling a party’s votes across candidates
or districts. Other observers suggest that turnout numbers are higher in PR
races because these contests often engender greater competition throughout the
electoral system. The extant work has held that there is a higher probability of
voting in close or more competitive races, and, thus, it follows that PR races
should produce higher turnout rates since the threshold for electoral success is
lower in these contests. Finally, because party-group linkages are stronger in
proportional electoral systems, we should expect higher turnout rates. Where
parties have close ties to various social, political, and/or economic groups,
and where they can rely on these groups to mobilize group members and other
voters, turnout should respond accordingly. For these reasons, party strategies
should be a function of electoral system design; hence, voter turnout also
should be a function of these rules.
While there are a number of researchers who argue that PR systems yield
high turnout rates, other scholars reach very different conclusions. Indeed,
others find that PR systems depress turnout levels.24 Two reasons are given for
21Gary
W. Cox, Making Votes Count: Strategic Coordination in the World’s Electoral Systems
(Cambridge, UK: Cambridge University Press, 1997).
22Blais and Dobrzynska, “Turnout in Electoral Democracies”; Andre Blais and R. Kenneth Carty,
“Does Proportional Representation Foster Voter Turnout?” European Journal of Political
Research 18, no. 2 (2006): 167-181; Mark Franklin, “Electoral Engineering and Cross-National
Turnout Differences: What Role for Compulsory Voting?” British Journal of Political Science
29, no. 1 (1999): 205-224; Jackman, “Political Institutions and Voter Turnout in the Industrial
Democracies”; Jackman and Miller, “Voter Turnout in the Industrial Democracies during the
1980s”; and Powell, “American Voter Turnout in Comparative Perspective.”
23Powell, “Voting Turnout in Thirty Democracies.”
24David Brockington, “The Paradox of Proportional Representation: The Effect of Party Systems
and Coalitions on Individuals’ Electoral Participation,” Political Studies 52, no. 3 (2004): 469-
32 | Taiwan Journal of Democracy, Volume 11, No. 2
why PR systems may actually stymie voter turnout. One is that, since PR races
often are the precursor to coalition governments in parliamentary systems,
some individuals may be deterred from casting a ballot as it is unlikely that
their vote will significantly affect the government’s governing coalition.25
The second reason is that voters face high information costs to learn about
the often numerous competitors running for office in a PR contest. With more
parties competing in these races-unlike the two-party systems found in many
SMD systems-the cost of being informed about the various competitors rises,
deterring some individuals from voting.26
For our dataset, the vast majority of regional governments employ either
SMD or mixed electoral systems (i.e., the elections for regional offices employ
both PR and SMD races). As a result, we have tested for mixed electoral systems
rather than for PR or SMD races alone, given the minimal variance with broad
categories. We anticipate that mixed systems should engender higher turnout
levels than nonmixed systems, since the former provide to voters additional
or multiple opportunities to express their opinions and to participate in the
electoral process.
Party Systems
This section discusses the effect party systems have on voter turnout, then
transitions into our contribution regarding systems that are anchored by
a dominant party. One school of thought is that turnout should be higher in
multiparty regimes than in those with a dominant party for multiple reasons.
For starters, if voters indeed are affected by party mobilization strategies,
turnout is expected to be higher in multiparty regimes, as several parties likely
are employing these strategies. This notion builds on the studies that contend
that parties in PR races generally push for increased turnout everywhere to
improve their chances of electoral success.27 Another reason for higher turnout
in multiparty systems is that the various parties in operation make it easier
for voters to sincerely vote, increasing their inclination to vote. It is this line
of thinking that led Powell28 to argue that higher turnout is more likely in
490, and Jeffrey Karp and Susan A. Banducci, “Political Efficacy and Participation in TwentySeven Democracies: How Electoral Systems Shape Political Behavior,” British Journal of
Political Science 38, no. 2 (2008): 311-334.
25Jackman, “Political Institutions and Voter Turnout in the Industrial Democracies.”
26Fornos et al. “Explaining Voter Turnout in Latin America, 1980-2000”; Jackman, “Political
Institutions and Voter Turnout in the Industrial Democracies”; Jackman and Miller, “Voter
Turnout in the Industrial Democracies during the 1980s”; Kostadinova, “Voter Turnout
Dynamics in Post-Communist Europe”; and Kostadinova and Power, “Does Democratization
Depress Participation?”
27Harold F. Gosnell, Why Europe Votes (Chicago: University of Chicago Press, 1930); Powell,
“Voting Turnout in Thirty Democracies”; and Herbert Tingsten, Political Behavior: Studies in
Election Statistics (London: P. S. King and Son, 1937).
28Powell, “Voting Turnout in Thirty Democracies.”
December 2015 | 33
PR systems than in SMD systems, given the ties between parties and various
social groups. Not only this, but also competitive party systems have been
linked to voter turnout in earlier works. While this relationship may be evident
in many polities across the world, it is unlikely to have the same effects in
systems with dominant parties. Thus, dominant parties, by their nature, result
in uncompetitive races; yet, the effects of these parties on turnout levels
have not been studied in a systematic fashion. The next section explores this
relationship in greater detail.
Dominant Parties and Political Participation
The connection between party systems and political participation, such
as voter turnout, is well-established, but the operationalization of party
systems generally has relied on the notion that these institutions, in fact, are
systems often consisting of multiple organizations vying for support among
the electorate. The recent emergence of the dominant-party or party-system
literature has revealed a number of key implications and consequences,
although the relevance for voter turnout has been unexplored. Before turning
to the hypothesized link between dominant parties and voter turnout, we offer
a summary of some of the main findings in the burgeoning canon regarding
dominant parties.
As noted at the beginning of this essay, very little research has examined
voter turnout at the subnational level, and this is particularly the case in the
Russian context. While there are convincing reasons for this, there is still much
to be learned from closer inspection of regional politics. In this vein, we bridge
together the literatures on voter turnout and dominant-party systems. Such an
approach, we contend, is appropriate and quite useful for the study of Russia
and its regions, as the ER is unarguably the party of power throughout much of
the country. This approach also is useful because voter turnout in the republics
appears to wax and wane based on the variations in the ER’s presence. In
addition, the body of work on turnout cross-nationally has largely overlooked
how dominant-party systems may shape voters’ incentive to participate, which
potentially could affect a myriad of other processes.
One of the seminal works in the arena of dominant-party systems, written
by Reuter and Gandhi, labels these systems as “regimes in which dictators
rule with the aid of one dominant party...while still holding regular multiparty elections.” They note that these systems aim to control policy making,
patronage, and most of the prominent or powerful political offices.29 By having
control over these resources, the dominant party possesses the upper hand
29Ora
John Reuter and Jennifer Gandhi, “Economic Performance and Elite Defection from
Dominant Parties,” British Journal of Political Science 41, no. 1 (2010): 83-110, esp. 87.
34 | Taiwan Journal of Democracy, Volume 11, No. 2
during elections.30
Aside from such revelations, the difficult question of measuring
or operationalizing dominant-party systems has generated contrasting
viewpoints,31 although those who have written on the topic have agreed
on a few basic elements. First, dominant-party systems are found in both
democratic and nondemocratic settings (that hold elections). These elections
lead us to the second characteristic: members of the legislature are selected
through contests that pit the dominant party against other parties. Third, the
ruling party controls an absolute majority of the seats in the primary legislative
organ. This condition is of the utmost importance for two reasons. As Reuter
and Gandhi explain, this control enables the party not only to exercise or
parade its dominance publicly, but also to have carte blanche to pursue and
implement its policies with little (if any) interference.32 Without a majority,
the party would be compelled to form partnerships (or coalitions, depending
on the institutional milieu) with other parties that could hamstring its ability to
adopt its desired policies. The fourth characteristic of these systems, longevity,
has fostered the most debate or discussion (aside from the cut-off point for
control). Both Magaloni and Greene assert that durability is a key component
when classifying such systems.33 Along these lines, scholars often have relied
on a threshold of twenty consecutive years of control to mark a party or system
as dominant. For the present study, we agree with Reuter and Gandhi, who
argue that the twenty-year requirement or threshold is overly restrictive and
unnecessarily excludes numerous other observations, such as Russia’s ER, that
otherwise satisfy the key criteria.34 In the end, Sakwa,35 Reuter and Gandhi,36
Reuter and Remington,37 and White,38 among others, have labeled the ER
a dominant party, even though some have raised questions about the ER’s
precise label.39
30Kenneth
F. Greene, Why Dominant Parties Lose: Mexico’s Democratization in Comparative
Perspective (Cambridge, UK: Cambridge University Press, 2007), and Beatriz Magaloni,
Voting for Autocracy: Dominant Party Survival and Its Demise in Mexico (Cambridge, UK:
Cambridge University Press, 2006).
31Matthijs Bogaards, “Counting Parties and Identifying Dominant Party Systems in Africa,”
European Journal of Political Research 43, no. 2 (2004): 173-197; Greene, Why Dominant
Parties Lose; Magaloni, Voting for Autocracy; T. J. Pempel, Uncommon Democracies (New
York: Cornell University Press, 1980); and Reuter and Gandhi, “Economic Performance and
Elite Defection from Dominant Parties.”
32Reuter and Gandhi, “Economic Performance and Elite Defection from Dominant Parties.”
33Greene, Why Dominant Parties Lose, and Magaloni, Voting for Autocracy.
34Reuter and Gandhi, “Economic Performance and Elite Defection from Dominant Parties.”
35Richard Sakwa, “Party and Power: Between Representation and Mobilisation in Contemporary
Russia,” East European Politics 28, no. 3 (2012): 310-327.
36Reuter and Gandhi, “Economic Performance and Elite Defection from Dominant Parties.”
37Ora John Reuter and Thomas F. Remington, “Dominant Party Regimes and the Commitment
Problem: The Case of United Russia,” Comparative Political Studies 42, no. 4 (2009): 501-526.
38White, “Dominant Party Systems.”
39Smyth et al., “Engineering Victory.”
December 2015 | 35
The ER and the Implications for Voter Turnout
We now turn to a discussion of how the traits of dominant-party institutions
affect electoral politics and, eventually, voter turnout. There are several major
ways in which dominant-party systems impact politics and policy making.
Research on dominant parties has found that the institutions of these systems
have flexible ideologies, access to state resources, and control of the media.40
Numerous scholars have affirmed that Russia’s ER generally possesses these
traits, all of which are likely to affect voter participation levels. Each trait’s
relevance to voter turnout is discussed, in turn, and then we elaborate on other
ways in which dominant parties are apt to correspond with low turnout levels.
It is generally well-established that the ER lacks a cohesive ideology.
In this manner, the party is not unlike other dominant parties. As Pempel41
explains, the absence of a coherent ideology enables the party of power to
adapt to changing political conditions, while also making it more difficult for
opposition forces to launch ideological attacks against it. The result of this
dearth of ideological coherence is that voters looking for ideological cues
may come up short. Struggling to find a party that matches with a would-be
voter’s own ideology means that the individual may abstain from voting if
there is no clear ideological match. Indeed, the ideological coherence of the
ER leaves much to be desired and, therefore, voters may abstain from voting
as information on the candidates, parties, or issues may prove scarce. The
absence of reliable cues on the main players is likely to further diminish an
individual’s desire to vote.42
Access to state resources and media control presumably affects voter
mobilization as well. Monopolization of state resources means that the party
could strategically target certain groups in the population-rather than the
broader populace-leading to lower turnout rates. Controlling this pool of
resources also means that the need to forge ties with grass-roots organizations
or voters in a broader sense may be curtailed. Alternatively, even though most
students of Russian politics concur that state resources are controlled by the ER,
the ability of legislative candidates to employ these resources is questionable.43
Thus, the disbursement of “goods” may not be easily accessible for legislative
candidates, which could result in low turnout numbers. However, possession
or control of vital resources may be more visible with governors who are ER
members. Media control, meanwhile, enables the dominant party to broadcast
its actions and accomplishments, often without scrutiny. In turn, this means
40White,
“Dominant Party Systems ”
Uncommon Democracies.
42Stephen E. Hanson, “Post-Imperial Democracies: Ideology and Party Formation in Third
Republic France, Weimar Germany, and Post-Soviet Russia,” East European Politics and
Societies 20, no. 3 (2006): 343-372, and Smyth et al., “Engineering Victory.”
43Smyth et al., “Engineering Victory.”
41Pempel,
36 | Taiwan Journal of Democracy, Volume 11, No. 2
that voters hoping to learn more about rival parties are unable to rely on the
media to learn about other partisan options. Thus, control over state resources
is likely to hinder turnout because of weakened connections between the party
and voters, while the dominant party’s control over the media means that rival
parties are apt to face an uphill battle in appealing to the populace.
Another reason why Russia’s party of power is prone to hinder turnout has
to do with the nature of political power. Research on the ER has concluded that
Russia’s powerbrokers have intentionally avoided investing too much power in
the hands of the party. Instead, the ER has emerged as a vehicle or organizing
tool for those in power, with the real power resting with Putin.44 As many other
studies have noted, the absence of ideologically based parties typically results
in parties dependent on personalities, which certainly describes Russia’s ER.
The importance of this is that there has been little investment in grass-roots
mobilization, making it more difficult for the party apparatus to boost turnout
throughout the country. One of the benefits of this study is that we are better
able to assess this possibility by examining politics at the subnational level.
The last aspect of our argument concerns the ability of rival political parties
to coax individuals to the polls. As mentioned, much of the previous work
contends that PR electoral systems lead multiparty systems to mobilize voters
throughout a region, with the end result being higher turnout. The threshold to
secure representation is generally lower in PR races than in SMD races, and,
therefore, parties in PR systems-collectively-work harder to mobilize voters.
However, dominant-party systems tend to deviate from this pattern, regardless
of the electoral system. This probably is due to the dominance of these parties,
which adversely affects the ability of other parties to form, organize, and reach
out to voters.
Given the above circumstances, dominant parties, by their nature, are
adept at controlling (in some cases co-opting) opposition forces. Expressed
differently, the outcome of elections in dominant-party systems may be known
a priori, which tends to have a dampening effect on the probability of voting.
In fact, Powell45 notes that turnout is often lower in SMD electoral systems
because they lack competitiveness that is more commonly found in PR races.
More broadly, Cox also points out that competitiveness significantly affects
voter turnout.46 Consequently, dominant parties-given their very natureexist in uncompetitive electoral environments that are apt to discourage
citizens from casting a vote at an election whose outcome is known in advance.
Hypotheses
As many studies have noted, Russian politics today is greatly affected by its
dominant party, doubtlessly affecting political participation throughout Russia.
44Sakwa,
“Party and Power,” and White, “Dominant Party Systems.”
“Voting Turnout in Thirty Democracies.”
46Cox, Making Votes Count.
45Powell,
December 2015 | 37
Earlier works indicate that voters often can assess the competitiveness of an
election as well as whether its outcome is a foregone conclusion.47 Indeed,
as stated above, competitive races tend to yield higher turnout rates than
noncompetitive ones for a variety of reasons, including mobilization efforts,
interest in a race, and the increased likelihood that an individual vote could
have more impact on the electoral outcome than in a lopsided contest.48
Therefore, it is plausible that, once voters assess the electoral scenario in their
district, they will evaluate the costs and benefits of voting: if voting costs
are exacerbated by a race that voters view as seriously one-sided, we would
expect them to abstain. Voters in post-communist states such as Russia have
been found to have a keen sense of competitiveness and of what is at stake
in various elections, which has been linked to voter turnout in other work.49
Given the ER’s dominance, name recognition, and other political advantages
that undermine electoral competitiveness, it is reasonable to conclude that
the presence of an ER candidate discourages many would-be voters from
participation at the polls. Based on the observations above, we anticipate that
the presence of the ER should dampen incentive to vote. In other words,
Hypothesis 1: The presence of an ER candidate should
correspond with low turnout rates.
As noted earlier in this essay as well as in numerous other works, Russia’s
governors or regional leaders have played an enormous role in the political
system for many years, even if Putin’s reforms have curtailed the powers of
these individuals. Russia’s governors continue to exercise major influence in
a variety of ways,50 so it is necessary to control for these actors. Much of the
governors’ lasting influence stems from their ability to command an impressive
array of resources that have paid large electoral and political dividends.51
These administrative resources provide great discretion to the governors
regarding local politics and afford them more resources than otherwise would
47Alexander
C. Pacek, Grigore Pop-Eleches, and Joshua A. Tucker, “Disenchanted or Discerning:
Voter Turnout in Post-Communist Countries,” Journal of Politics 71, no. 2 (2009): 473-491.
48Gary W. Cox and Michael C. Munger, “Closeness, Expenditures, and Turnout in the 1982
U. S. House Elections,” American Political Science Review 83, no. 1 (1989): 217-231; Geys,
“Explaining Voter Turnout”; and John G. Matsusaka and Filip Palda, “The Downsian Voter
Meets the Ecological Fallacy,” Public Choice 77, no. 4 (1993): 855-878.
49Pacek et al., “Disenchanted or Discerning.”
50For an overview of this influence, see Henry E. Hale, Why Not Parties in Russia? Democracy,
Federalism, and the State (Cambridge, UK: Cambridge University Press, 2006).
51Georgiy Rumyantsev, “ ‘A Normal Reaction on the Part of Normal People’-Pollsters Have
Discovered the Main Cause of United Russia’s Success at the Elections,” Izvestiya, November 2,
2006; Darrell Slider, “United Russia and Russia’s Governors: The Path to a One-Party System,”
paper presented at the annual meeting for the American Association for the Advancement of
Slavic Studies, New Orleans, Louisiana, November 17-20, 2006; and Smyth et al., “Engineering
Victory.”
38 | Taiwan Journal of Democracy, Volume 11, No. 2
be available. These resources can be used to foster high turnout rates for
certain areas, especially since governors frequently have incentive to generate
high turnout as members of the dominant party.52 Thus, we expect a positive
relationship between the ER governors and voter turnout-an expectation that
clearly contrasts with our first hypothesis. This second hypothesis builds on
other work that finds that influential regional leaders can affect voter turnout.
How does such influence of governors affect a voter’s calculus as to
whether to vote? The anticipated relationship concerns the need of governors
to favorably influence the electoral process in order to maintain their positions;
ever since Putin’s reforms were implemented, governors must appease the
central government to retain their posts. This means that governors must boost
turnout levels to prevent their embarrassment and, by extension, the humiliation
of Putin’s government. Yet, how does this happen? Recent work suggests that
this occurs through applying political pressure. Frye and his co-authors53 have
found that firms dependent upon regional governments have a higher incidence
of cajoling their workers to turnout to vote than nondependent firms-a
relationship that alludes to the influence wielded by regional leadership. A
number of means are used by governors to prompt voters to go to the polls,
such as (thinly) veiled threats, the distribution of resources, and various
mobilization efforts. These dynamics produce effects similar to the party or
elite mobilization efforts described by Cox and Munger that have been linked
with high turnout levels.54 Thus, when ER gubernatorial candidates are on the
ballot, efforts are taken to boost turnout levels among voters in a variety of
ways, even though the primary tactic may be through mobilization efforts. To
test for this possibility, we include a second, related metric for ER affiliation,
but this time we include an indicator to capture the effects of governors with
ER affiliation. This leads us to our second primary hypothesis:
Hypothesis 2: The presence of an ER governor should be
positively related to voter turnout rates.
Finally, we test for an interactive effect between our two primary
independent variables. Here, we study how voter turnout levels are affected
when a race takes place in a region controlled by an ER-affiliated governor
and also includes an ER candidate. Due to the compounded effect of the ER’s
presence, we expect that this situation leads to a significant drop in turnout
level. Thus, we anticipate that elections meeting both criteria will be heavily
tilted in favor of the ER, leading to lower turnout levels. Formally stated, this
leads to our third hypothesis:
52Arian
and Barnes, “The Dominant Party System.”
et al., “Political Machines at Work.”
54Cox and Munger, “Closeness, Expenditures, and Turnout in the 1982 U. S. House Elections.”
53Frye
December 2015 | 39
Hypothesis 3: Voter turnout is expected to be lower than
otherwise when an electoral race includes
an ER candidate and takes place in a region
controlled by an ER governor.
Research Design
Dataset
To test our hypotheses, we have created a unique dataset from many of Russia’s
regions. In many of these republics, the mean number of elections is two or
three, limiting our study longitudinally (the time span ranges from 2001 to
2011). It is generally understood that the ER was the product of the merger of
the Unity and Fatherland All-Russia parties; however, we have focused our
analysis on those candidates who have had a formal affiliation with the ER and
not with either of the party’s partisan predecessors. Despite these omissions,
we are left with more than twenty regions, spanning more than a decade. The
models below regress voter turnout with a series of standard control variables,
while our unit of analysis is electoral district-year (referred to below as okrug).
We have logged nearly 1,500 observations from the various electoral districts
in the more than twenty regions studied.
Dependent Variable: Voter Turnout
In this study, we utilize one dependent variable most commonly associated
with voter turnout. This measure-voter turnout-is operationalized as the
percentage of eligible registered voters who cast a ballot in a legislative election
(lower chamber only). This measure of political participation commonly has
been employed in previous works.55 The data were collected from the website
of Vladimir Golosov.56
A number of reports of electoral misconduct and/or fraud have been
reported in recent national Duma elections;57 however, such concerns need
not completely deter research on the country. Similar concerns with electoral
and survey data from countries such as Mexico, Venezuela, Zimbabwe, and
Belarus, to name only a few, have not prevented scholarly work in these areas.58
55Blais
and Dobrzynska, “Turnout in Electoral Democracies”; Franklin, “Electoral Engineering
and Cross-National Turnout Differences”; and Kostadinova, “Voter Turnout Dynamics in PostCommunist Europe.”
56“Russian Electoral Statistics Database,” Vladimir Golosov, http://db.geliks.org/ (accessed
December 1, 2012).
57Mikhail Myagkov, Peter C. Ordeshook, and Dimitri Shakin, The Forensics of Election Fraud
(Cambridge, UK: Cambridge University Press, 2009).
58See works by Derek S. Hutcheson and Elena Korosteleva, “Patterns of Participation in PostSoviet Politics,” Comparative European Politics 4, no. 1 (2006): 23-46; Masipula Sithole
and John Makumbe, “Elections in Zimbabwe: The ZANU (PF) Hegemony and Its Incipient
Decline,” African Journal of Political Science 2, no. 1 (1997): 122-139; and Stephen White
and Ian McAllister, “Turnout and Representation: Bias in Post-Communist Europe,” Political
Studies 55, no. 3 (2007): 586-606.
40 | Taiwan Journal of Democracy, Volume 11, No. 2
Researchers should understand the potential hazards when analyzing the data,
however, and employ methods to mitigate potential problems emanating from
them. We do this in the models below. Notably, we look to other scholars
who have utilized alternative approaches to decrease the distortional effects
of including worrisome cases. Most notably, we follow the advice of others
who have argued that turnout levels that are two standard deviations above
the sample mean (which amounts to around 96 percent) could indicate flagrant
electoral misconduct. Following this lead, we have identified such cases and
years in Russia (e.g., Karachay-Cherkessia in 2004, both Tatarstan and Tyva in
2006, Bashkortostan in 2008, and Mordoviya in 2009) and have run analyses
with and without these suspect cases.59 The primary findings are unaltered by
excluding these cases.60
Independent Variable: The Dominant Party
To assess the influence of the ER-Russia’s dominant party-we utilize two
separate dichotomous measures. The first is clearly an electoral indicator
but one that exists at the republic’s legislative level. Specifically, we include
a dichotomous indicator, ER Candidate, for those races in which an ER
candidate competed. An ER candidate was present in approximately 68
percent of the districts, with the party’s strongest showings in Adygeya, Altay,
Chuvashiya, and Ust Orda. Meanwhile, the party was scarce in republics such
as Bashkortostan, Chukotka, the Jewish Oblast, and Karachay-Cherkessia.
The second primary independent variable, ER Governor, is a second
dichotomous indicator for the dominant party. In this case, though, we control
for those republics that are governed by an ER member. The number of ER
governors is considerably lower than the number of ER candidates competing
in elections. In fact, republic governors were members of the ER in just
over one-third (or 38 percent) of the republics under review. Bashkortostan,
Mordoviya, and Udmurtiya were controlled by ER members, while republics
such as Aginsk, Chukotka, Karachay-Cherkessiya, and Komi were not (in the
years under review). There also are some notable temporal variations, as ER
gubernatorial control has risen in Altay, Buryatiya, and Chuvashiya, although
stabilized in other areas (e.g., Adygeya, Taymyr, Tyva, and so on). As discussed
above, we expect regions with ER-affiliated governors to have higher turnout
rates than those without.
59Kramer
also reports that turnout or the number of votes cast actually exceeded the number of
eligible voters in at least one precinct in Chechnya. Since Chechen data are not included in our
analysis, no additional steps are needed with respect to this region. See Andrew E. Kramer, “At
Chechnya Polling Station, Votes for Putin Exceed the Rolls,” New York Times, March 5, 2012.
60These results are available upon request.
December 2015 | 41
Control Measures
In addition to the two primary dominant-party variables discussed above, we
include several additional control measures that are common in voter turnout
studies. We provide a brief discussion of each of these in this section.
We offer a number of institutional controls, as earlier work on voter
turnout has found that certain institutional arrangements have important
implications for voter participation. Given the variation in legislature size
in Russia’s republics, we include a control for this measure; the smallest
legislature is found in Chukotka (six seats), while Tatarstan has the largest (130
seats). The second institutional measure used is a dichotomous measure for
mixed electoral systems, mixed elect. sys., found in places such as Chukotka,
Khakasiya, Udmurtiya, and Yamal Nenets. Aside from Kostadinova,61 the
voter turnout literature has focused little on the relevance of mixed electoral
systems, although we expect these systems to result in high turnout levels, as
they afford voters multiple opportunities to influence the political system.
As for the remaining controls, we include multiple measures for economic
and social indicators. Economic measures have long been linked with turnout
levels; woeful economic conditions generally trigger high turnout as voters
head to the polls to enact political change in the hope of better economic
prospects. We hold a similar expectation regarding our first economic variable,
unemployment. The second economic control is income growth, and, again, we
expect that trends in income growth (or negative values) correlate with higher
than normal turnout rates. Aside from these measures, controls for urbanization,
Russian population, and titular population also are employed. We anticipate
these measures to be positively related to voter turnout as well, although this
is likely to be contingent upon intraregion dynamics. Due to collinearity with
these measures (i.e., Russian population and titular population), we have
estimated the models below separately to avoid potential contamination by
means of the analysis. Finally, the data were obtained from Goskomstat’s
website.62 Summary statistics of all the variables used in the analysis are
found in table 1, while a regional breakdown of the main dependent variable
is available in figure 1.
Findings
To test our hypotheses, we have compiled a unique dataset that examines voter
turnout variations at the subnational level in Russia’s various republics. We
estimate our models with an ordinary least squares regression (although the
61Tatiana
Kostadinova, “Do Mixed Electoral Systems Matter? A Cross-National Analysis of Their
Effects in Eastern Europe,” Electoral Studies 21, no. 1 (2002): 23-34.
62“Russian Federal Statistics Service,” Goskomstat, http://www.gks.ru (accessed March 10,
2011).
42 | Taiwan Journal of Democracy, Volume 11, No. 2
Table 1. Summary Statistics
Variable
Voter Turnout
Edinaya Rossiya (ER) Cand.
Gov. ER Cand.
Unemployment
Income Growth
Legislature Size
Mixed Electoral System
Urban Population
Titular Population
Russian Population
Mean
51.63
.70
.38
10.23
110.23
57.92
.18
63.10
39.15
48.51
Stand. Dev.
24.08
.48
.49
3.99
5.22
34.53
.39
12.88
18.99
17.61
Minimum
7.68
0
0
3.60
94.30
6.00
0
20.30
1.22
20.10
Maximum
99.25
1.00
1.00
25.60
125.20
130.00
1.00
84.80
77.00
89.90
results from a tobit analysis-which produced virtually identical results-are
available upon request). Due to a number of factors, there are salient differences
that must be accounted for, and, therefore, we have attempted to control for this
by clustering our standard errors by okrug. Last, we present multiple models in
each of the following tables due to the high collinearity between our two social
variables (Russian population and Titular population).
Tables 2 and 3 present the results of our Ordinary Least Squares (OLS)
analysis of voter turnout and reveal a number of interesting findings. For
the primary variable in table 2, ER Candidate, the results in model 1 were
statistically significant and provide support for Hypothesis 1. Specifically, the
results indicate that the presence of an ER candidate hinders turnout, and the
substantive effects are quite large (nearly 2.5 percent change in turnout rates).
ER Candidate effects were not statistically significant in model 2, however.
While the variable returned a negative coefficient, ER Candidate fell just shy
of statistical significance. This indicates that the influence of a dominant-party
candidate may be limited; based on the two models, this may have something
to do with ethnic group control, as model 1 included a control for Russian
population size, while model 2 controlled for the region’s main titular group.
Furthermore, it is notable that larger Russian populations (e.g., in Buryatiya,
Khakasiya, and Kareliya) corresponded with a statistically significant decrease
in voter turnout (model 1), while titular population size was positively signed
and statistically significant. Therefore, turnout levels tend to be higher in
regions such as Alaniya, Chuvashiya, and Karachay-Cherkessia that are home
to larger titular populations.
Aside from the above variables, a number of other controls were statistically
significant. Both institutional measures-legislative size and mixed systemswere significant. The results suggest that turnout was comparatively high in
republics with small legislatures and in areas that utilized mixed electoral
systems. The latter is likely the case because voters have multiple opportunities
December 2015 | 43
Table 2. Regression of ER Candidate on Voter Turnout in Russia’s Regions
Variables
ER Candidate
Legislative Size
Mixed System
Unemployment
Income Growth
Urban Population
Russian Population
Titular Population
Constant
r2
F
N
Model 1
Model 2
(DV: Turnout)
(DV: Turnout)
111.17
(8.79)***
.37
174.25***
1,456
.12
(.03)***
81.68
(9.08)***
.34
160.19***
1,456
-2.29
(.96)***
-.29
(.01)***
18.10
(1.16)***
-1.20
(.16)***
.13
(.09)
-.56
(.05)***
-.31
(.02)***
*p<.10; **p<.05; ***p<.01; Standard errors clustered by okrug.
-1.60
(.99)
-.24
(.01)***
17.21
(1.10)***
-1.18
(.16)***
.21
(.09)**
-.58
(.06)***
to influence races, as they are permitted to cast two votes. Regarding the
former, small legislatures are likely to amplify the competitiveness for the
seats that are at stake, which doubtlessly goads more voters to the polls. In
contrast, the economic controls presented a mixed finding. Unemployment
rates were statistically significant, corroborating existing work on the subject.
Based on this result, we see that high unemployment rates depress turnout
levels-a finding that is out of line with many other studies. This finding
appears to be robust across our models, however, which suggests that more
work is needed on this topic. The second economic control, income growth,
rarely reaches significance in our models (model 1 being the lone exception),
revealing that this variable matters little to voter turnout levels among these
cases. Finally, we see that urban population is a robust predictor of voter
turnout levels, although the relationship is a bit surprising. Other studies that
have examined or included urban population controls generally have found
a positive connection between the measure and political participation; this,
however, is not the case in this project. Instead, we find that turnout levels are
higher in Altay, Karachay-Cherkessia, and Ust Orda-three republics with the
smallest urban populations in our dataset. This negative relationship between
44 | Taiwan Journal of Democracy, Volume 11, No. 2
urban population and voter turnout is not without precedent, however.63 In
part, the finding may have to do with the fact that the parties that have relied on
urban supporters (e.g., Yabloko) have been less successful than those that have
relied on rural dwellers (e.g., KPRF). As a result, fewer voters may turn out in
urban areas because the parties that ordinarily would mobilize these denizens
are struggling politically and organizationally.
Table 3 tests the second dominant-party-related variable, gubernatorial ER
affiliation. As previously discussed, Russia’s governors for some time have
been prominent players in the country’s politics, and although their powers
have changed in recent years, they continue to be influential in contemporary
politics. That said, we use a dichotomous measure of gubernatorial ER
association in models 3 and 4 to assess the governors’ salience with respect
to voting. The results in table 3 reveal that turnout is higher in republics led
by ER members than others. This significant relationship supports our second
hypothesis that ER governors may be either more supportive of, or more
Table 3. Regression of ER Governor on Voter Turnout in Russia’s Regions
Variables
Gov. ER Cand.
Legislative Size
Mixed System
Unemployment
Income Growth
Urban Population
Russian Population
Titular Population
Constant
r2
F
N
Model 3
(DV: Turnout)
(DV: Turnout)
100.75
(8.66)***
.43
132.77***
1,456
.13
(.02)***
72.47
(8.63)***
.41
127.48***
1,456
13.97
(1.19)***
-.29
(.01)***
9.89
(.97)***
-.63
(.13)***
.04
(.08)
-.42
(.04)***
-.27
(.02)***
*p<.10; **p<.05; ***p<.01; Standard errors clustered by okrug.
63Kuenzi
Model 4
14.98
(1.22)***
-.26
(.01)***
8.66
(.93)***
-.61
(.13)***
.12
(.08)
-.43
(.04)***
and Lambright, “Voter Turnout in Africa’s Multiparty Regimes,” 676.
December 2015 | 45
inclined to generate, high turnout levels. We suspect that this is due to the
administrative resources possessed by these individuals, which enables them to
target certain sectors of the population to encourage them to vote. Reuter found
that governors who possessed limited resources were among the most likely to
join the ER, and as a result, it could be that ER-affiliated governors are able to
use their newly acquired resources to mobilize voters.64 Moreover, this power
may be more easily dispatched at the gubernatorial level than at the legislativecandidate level (accounting for the latter’s adverse relationship with turnout).
Alternatively, there may be incentive for ER-affiliated governors to oversee
high turnout rates to boost their legitimacy. In contrast, governors without ER
affiliation are likely to see lower turnout rates, which could be ascribed to these
officials’ personal connections and administrative resources that diminish the
need for party affiliation (or formation) that could be used to mobilize broader
response from voters. Patronage networks associated with governors may be
more substantive than mere partisan affiliation. Nevertheless, with respect to
turnout and Gov. ER Cand., the results are robust across each model, and we
see that more voters cast a ballot in republics controlled by ER members. In
addition, the substantive effects of this relationship are large-much larger than
the relationship between voter turnout and the presence of an ER legislative
candidate. Elections taking place in republics with ER executives witness
an increase in turnout of approximately 14 percent (13.97 percent and 14.98
percent in models 3 and 4, respectively).
The findings from the remaining variables are quite consistent with those
shown in table 2. The remaining findings were nearly identical in terms of the
directional relationship presented. For example, legislature size, unemployment
rates, the percent of urban population, and the Russian population variables,
again, were significant and negative, and the substantive effects were more
or less similar to those seen in table 2. Likewise, mixed systems and titular
population both returned positive coefficients that were statistically significant.
Aside from the results in tables 2 and 3, we also present a second set
of statistical tests with these same variables, but this time we include both
variables in the same models. The initial separation (from models 1-4) was
necessary to test the effects of each variable in isolation, as the correlation
between the indicators is moderately strong (p = .30). In table 4, however, ER
Candidate and Gov. ER Cand. are included in both models 5 and 6. The results
of this additional set of findings generally are consistent with what already has
been presented, and, in particular, the direction and statistical significance of
the two key variables remains intact. One notable distinction, however, stems
from the ER Candidate results. In both of the models in table 4, the variable
is statistically significant and negatively signed-as expected. The size of the
64Ora
John Reuter, “The Politics of Dominant Party Formation: United Russia and Russia’s
Governors,” Europe-Asia Studies 62, no. 2 (2010): 293-327.
46 | Taiwan Journal of Democracy, Volume 11, No. 2
Table 4. Regression of ER Candidate and ER Governor
on Voter Turnout in Russia’s Regions
Variables
ER Candidate
Gov. ER Cand.
Legislative Size
Mixed System
Unemployment
Income Growth
Urban Population
Russian Population
Titular Population
Constant
r2
F
N
Model 5
Model 6
(DV: Turnout)
(DV: Turnout)
106.52
(8.32)***
.45
122.30***
1,456
.15
(.02)***
75.69
(8.30)***
.42
121.46***
1,456
-7.04
(.95)***
15.93
(1.33)***
-.32
(.01)***
10.26
(.97)***
-.62
(.12)***
.01
(.07)
-.39
(.03)***
-.29
(.02)***
*p<.10; **p<.05; ***p<.01; Standard errors clustered by okrug.
-6.88
(.96)***
16.97
(1.36)***
-.29
(.01)***
8.93
(.92)***
-.62
(.13)***
.09
(.07)
-.40
(.04)***
coefficient is considerably larger (roughly three times the size) than what was
presented in table 2. The findings for executive ER affiliation are congruent
with what has been discussed previously: the variable remains significant and
positively signed, signifying that turnout rates rise in republics controlled
by ER members. In contrast, voter turnout figures are considerably lower in
those republics that are controlled by an independent or rival party member.
Looking at the broader picture, in terms of comparing the results from tables
2 and 3 to those in table 4, we find that the effects of the two dominant-partysystem measures are significant and robust, regardless whether they are tested
simultaneously or separately. The findings for the control measures echo the
general conclusions obtained from the first four statistical models.
The final set of results continues to examine the peculiar pictures that
emerge in this analysis. Given the success and reach of the ER, it is worthwhile
to examine the interactive effects of these measures. In other words, what does
voter turnout look like in okrug-level elections in which both an ER candidate
December 2015 | 47
is competing and the republic is run by an ER member? The hypotheses and
argument offered above would expect a decline in turnout levels as this would
represent a scenario in which the dominant party’s influence would be even
more evident than when only one of these conditions was met. To that end, we
have created an interactive term to gauge this possibility.
We test the possible scenario with the aforementioned interactive measure
(ER Cand. X Gov. ER Cand.) in table 5 (models 7 and 8). Although the ER’s
success manifests itself in a variety of ways, the occurrences in which the
party controls a regional executive and has a candidate compete in the same
race are surprisingly rare: only 16 percent of the okrugs under review meet
this description. But what happens in those races in which the ER’s presence
is doubly felt? The results are statistically significant beyond the 99 percent
level, and the substantive effect remains quite large (10.11 percent and 12.45
Table 5. Regression of Interactive Effect on Voter Turnout in Russia’s
Regions
Variables
ER Candidate
Gov. ER Cand.
ER Cand. X Gov. ER Cand.
Legislative Size
Mixed System
Unemployment
Income Growth
Urban Population
Russian Population
Titular Population
Constant
r2
F
N
Model 7
Model 8
(DV: Turnout)
(DV: Turnout)
110.51
(8.75)***
.45
116.85***
1,456
.16
(.02)***
80.33
(8.95)***
.43
132.18***
1,456
-1.56
(1.25)
18.51
(1.57)***
-10.11
(1.78)***
-.32
(.01)***
11.64
(1.01)***
-.59
(.12)***
-.05
(.08)
-.39
(.03)***
-.28
(.02)***
*p<.10; **p<.05; ***p<.01; Standard errors clustered by okrug.
48 | Taiwan Journal of Democracy, Volume 11, No. 2
-.23
(1.33)
20.17
(1.62)***
-12.45
(1.87)***
-.29
(.01)***
10.66
(.96)***
-.60
(.13)***
.03
(.08)
-.40
(.04)***
Figure 2. Conditional Effects of Gov. ER Affiliation on Voter Turnout
percent in models 7 and 8, respectively). The direction of the relationship
may speak more to the cynical nature of politics in dominant-party systems,
as voter turnout decreases by more than 10 percent in those cases where
an ER candidate competes in okrug-level elections that occur in republics
controlled by ER governors. This relationship also is presented graphically in
figure 2, following the guidance outlined in Brambor, Clark, and Golder.65 The
information available in table 5 and figure 2 reveals a rather straightforward
conclusion: in those settings where the ER’s presence is most strongly felt,
voters are significantly less likely to turn out to vote. This result probably is
tied to the mechanisms outlined above. In those circumstances where the ER’s
presence is strongly felt, multiparty competition is apt to be lacking, outcomes
are seemingly predetermined, and mobilization efforts are more limited in
scope and less effective-that is, if present at all.
Robustness Checks
In addition to the battery of tests already discussed, we have run additional
analyses to gauge the robustness of our findings. Of immediate interest is
the prevalence of electoral misconduct. Systematic measures of electoral
65Thomas
Brambor, William Clark, and Matt Golder, “Are African Party Systems Different?”
Electoral Studies 26, no. 2 (2007): 315-323.
December 2015 | 49
misconduct are difficult to find for most countries in the world and even harder
to find for subnational governance. Thus, we rely on other work by Myagkov et
al.,66 White,67 and others who employ alternative approaches to assess electoral
misconduct. Specifically, these studies conclude that extremely high levels of
voter turnout (e.g., two standard deviations above pooled mean turnout levels)
serve as useful approximations of electoral misconduct. Accordingly, we, too,
follow this approach and have re-estimated our analyses with a dichotomous
variable for those republics with curiously high voter turnout levels for
contests. The main findings from the re-estimated results are consistent with
the main findings discussed previously-both statistically and substantively.68
We also considered how other controls affect voter turnout. In terms of
economics, we examined how the income inequality (GINI coefficient) and
the percentage of the population below the poverty line affected turnout. The
results of these additional tests again affirm our primary findings with respect
to the ER’s effect on turnout. However, neither variable itself was statistically
significant and has been omitted from the analysis. As for political controls,
we estimated our analyses with additional party controls, but the findings
held. In addition, we replaced the ER control with alternative party indicators
(e.g., LDPR and the KPRF), yet the findings were not consistently statistically
significant-indicating that it is the presence of Russia’s party of power rather
than any party organ that influences voter turnout.
Nowhere to Go But Up?
Interpretations of the results in the preceding section are likely to produce one
of two projected outcomes. The first is that the picture is grim for Russia’s
democratic prospects as well as for the viability of its marginalized opposition
parties. This view conforms to the research by Hanson,69 Smyth et al.,70 and
others who predict the long-term dominance of the ER. As Russia’s dominant
party, the ER’s control or access to resources (e.g., patronage), its influence
over the media, and its flexible ideology could very well extend the party’s
dominance far into the future. The ER’s level of control, as seen by the findings
presented here, undermines political participation, making it less likely that
serious opposition forces can overcome the ER electorally.
A second interpretation, however, presents an alternative outcome.
66Myagkov
et al., Forensics of Election Fraud.
C. White, “The Determinants of Dominant Party Strength and Opposition Party
Weakness in Russia: A Comparison of 1995 and 2011 Elections to the State Duma,” paper
presented at the annual meeting for the Midwest Political Science Association, Chicago, Illinois,
April 11-14, 2013.
68These results are available upon request.
69Hanson, “Post-Imperial Democracies.”
70Smyth et al., “Engineering Victory.”
67Allison
50 | Taiwan Journal of Democracy, Volume 11, No. 2
Turnout appears quite low in many areas, especially when a race features an
ER member. In this scenario, we often see limited or feckless opposition party
members challenging an ER member only to lose in many instances. Yet, as
White71 explains, high voter turnout levels could start to chip away at the
foundation of the ER’s dominance. Because turnout levels remain low in many
areas, the threshold for electoral victory is also low, making it more feasible
for opposition forces to regain seats in certain regions. And, since the evidence
above does not report a significant number of races that had an ER legislative
candidate and also were in a republic controlled by an ER governor, there is
perhaps more reason for optimism than seen at first glance.
Furthermore, public demonstrations (e.g., in Kaliningrad in 2010 and
in Moscow in 2011) suggest that a restive population may provide fertile
ground for an organized and effective opposition movement. This, however,
would require some strategizing by certain opposition parties and presumably
would require some parties to swap ideological dogma for pragmatic, broadbased appeals for members (a development that is particularly necessary for
the ideologically rigid Yabloko). An effort to reach out to (potential) voters
could pay tremendous electoral dividends for opposition parties, especially
given their limited access to, and coverage by, the media. Even though recent
changes in electoral law appear to have stacked the deck against opposition
parties (e.g., Yabloko, CPRF, and Solidarity), White points out that these
changes could ultimately backfire against the ER.
Conclusion
Longstanding research on voter turnout has determined that voters are
affected by numerous factors. Economic conditions, ethnic differences, the
type of electoral system, and the nature of a party system all have been linked
with an individual’s decision whether to cast a ballot. The research on party
systems generally has found that those that present a dizzying array of parties
or inconsistent informational cues to voters tend to depress voter turnout.
However, there are other ways that party effects can manifest themselves
during electoral cycles, and they can vary broadly in fashions that have not
been captured in existing work.
The guiding argument of this project is that dominant parties, such as
Russia’s ER, influence politics in many ways. This project looks solely at the
ER’s effect on voter turnout, as the presence of the party-for many reasonsmay dampen the incentive to vote. This is what we find in the results presented
above. However, the relationship is more complicated than this and some of
the other discoveries have raised more questions than they have answered. For
this reason, voter turnout in Russia’s subnational elections warrants additional
71White,
“Dominant Party Systems.”
December 2015 | 51
scrutiny. In future work, we hope to improve upon this project in a few key
ways. One is that we currently are expanding the scope of our analysis to
incorporate additional regions that will offer a more exhaustive test of the
theory posited here. Also, given the political changes surrounding gubernatorial
appointments,72 it may prove fruitful to more closely scrutinize the ER
gubernatorial members, as the way in which these leaders are replaced and
appointed has changed over the past ten years. In addition, alternative controls
are needed to better understand the complex dynamics at work between voter
turnout and institutional and economic measures. Finally, devising additional
measures of dominant party systems would afford greater theoretical leverage.
The findings here are relevant empirically and theoretically to other
contexts. In multiple ways, it would be interesting to consider how these
findings compare with other former Soviet states or with countries that have
similarities. For instance, comparing the Russian and Mexican cases could be
particularly valuable, given the countries’ questionable democratic practices,
federalist trappings, and one-party rule. Empirically, the findings add to
research on subnational politics in Russia, a burgeoning research topic. As
seen in the growing number of studies in this area, there is much to learn from
Russia’s regions, despite Putin’s reforms that have weakened gubernatorial
leadership during the past several years. Also, the theoretic importance of this
research stresses the need to move beyond existing work on party systems to
gain understanding of voter turnout. Earlier work concluded that the number
of parties or party fractionalization offered valuable insight into predicting
voter turnout patterns; yet, extending these findings to Russia-especially at
the subnational level-can offer little insight into the variations seen across the
country’s regions. Instead, other measures of party system effects are needed to
understand how voters respond to parties with differing levels of strength and
influence. In the Russian context, this has meant looking at the ER’s influence
on the number of voters going to the polls. Once we factor in the presence
of ER candidates, we find that turnout is decreased considerably, a finding
that we attribute to races with ER candidates being unlikely to have surprising
outcomes-thus, leading to low turnout rates. The theoretical contribution
also adds to the dominant-party systems literature in that it examines not just
policy making and electoral outcomes-two important topics-but also how
individuals participate in light of the regime’s presence. In some contexts,
such as Russia, participation recedes, but in other contexts this may not be the
case. Focus on voter participation in electoral contests in which the regime’s
dominant party is present could result in intriguing research in the future.
72For
a thorough discussion on the subject, see Rostislav F. Turovskii, “How Russian Governors
Are Appointed: Inertia and Radicalism in Central Policy,” Russian Politics and Law 48, no. 1
(2010): 58-79.
52 | Taiwan Journal of Democracy, Volume 11, No. 2