Testing for sources of electoral competition under authoritarianism: an analysis of Russia’s gubernatorial elections

Abstract

What drives electoral competition in competitive authoritarian regimes? Most scholarship has assumed that the outcome of these elections is decided by regime manipulation alone. Using three rounds of newly reinstated gubernatorial elections in Russia’s regions, I test this assumption. I identify three distinct measures of competition calibrated to authoritarian elections and assess whether voter preferences or regime manipulation best explain the degree of electoral competition. Relying on new data on protests across Russia’s regions, I find that regions with high protest activity have more contested elections with narrower margins of victory. The results also confirm recent scholarship highlighting the importance of voter turnout for delivering pro-regime victories.

Keywords:

• Electoral authoritarianism
• elections
• protest
• Russia
• voting

Notes

1. Gubernatorial elections in Russia’s regions were formally instituted in 1991 but did not become common in all regions until 1995–1997. During the 1990s, President Boris Yel’tsin’s practice of piecemeal deal-making with governors in order to shore up his political support facilitated the development of strong regional leaders with extensive and independent patronage networks. Gubernatorial elections were abolished in 2005 as part of an effort to reconnect the regions to the center, thereby ending regional fiefdoms. From 2005 to 2012, all regional executives were appointed by the president. Direct election of governors was reintroduced in June 2012 in part as a response to the demands of participants in the 2011–2012 anti-electoral fraud protests, but also as a way for the Kremlin to install popularly vetted regional leaders who would be better able to deliver pro-Kremlin electoral results in the regions.

2. I express my gratitude to Dr Tomila Lankina for sharing her valuable data on protest frequency and patterns in Russia from 2007 to 2012.

3. See the online Appendix 1 for information on data sources used to construct the variables.

4. Golos compiles reports from observers in each region. The reports can be found on the organization’s website golosinfo.org.

5. Originally used in economics, this measure has been adapted in the political science literature and used both in the Russian context and in studies of regional elections elsewhere (see Moraski and Reisinger 2003; Afzal 2014).

6. The formula for Golosov’s effective number of parties is N_p= $\sum \frac{1}{1+({\text{S}}_1^2/{\text{S}}_{\text{i}})-S_i}$, where S₁ is the winning vote share and S_i are other vote shares.

7. The coefficient for Protest remained at least three times the size of the standard error and statistically significant at the 0.01 level when other predictors and control variables were dropped from the model (see Table A1 in the online Appendix 1). Variance inflation factors for each predictor were <10 (see Table A3 in the online Appendix 1).

8. The coefficient for turnout remained at least three times the size of the standard error and statistically significant at the 0.01 level when other predictors and control variables were dropped from the model (see Table A2 in the online Appendix 1). Variance inflation factors for each predictor were <10 (see Table A4 in the online Appendix 1).