Do Election Results Shape Legitimacy Perceptions in Autocracy?

Elections remain a central feature of many authoritarian regimes despite widespread manipulation and limited political competition. Using a survey experiment with a nationally representative sample of Russian voters, this study examines whether improving perceptions of legitimacy can help explain why autocrats hold elections. The results show that information about high turnout increases trust in government, while information about low turnout reduces it, with effects driven by government supporters and individuals who believe in election integrity. This suggests that authoritarian leaders may use elections and reported electoral outcomes strategically to reinforce legitimacy among their support base and manage public perceptions over time.

Puzzle of Autocratic Elections

In recent decades, many authoritarian regimes have increasingly adopted institutions resembling those of democracies, particularly elections (Guriev and Treisman, 2019). Autocrats often organize multiparty elections and invite international observers, even as they manipulate outcomes through widespread fraud. This combination raises an important puzzle: if elections do not truly determine political power, why do authoritarian leaders hold them?

A large body of research has examined how authoritarian elections are organized (Gandhi and Lust-Okar, 2009; Gehlbach et al., 2016; Egorov and Sonin, 2020), including strategies such as limiting competition (Gandhi and Przeworski, 2007; Egorov and Sonin, 2021), managing media and information (Egorov et al., 2009; Edmond, 2013), and using elections to signal regime strength or monitor elites (Gehlbach and Simpser, 2015). However, less is known about whether these elections actually shape voters’ perceptions of government legitimacy (Dukalskis and Gerschewski, 2017).

One seemingly straightforward way to approach this question is to look at the relationship between electoral participation and trust in government, a key measure of political legitimacy. For example, across OECD countries, higher turnout is strongly correlated with greater trust in national governments (Figure 1). However, this descriptive pattern does not establish causality. Economic conditions may simultaneously shape both trust and electoral outcomes, creating omitted variable bias, while legitimacy itself may influence participation, leading to reverse causality.

These limitations point to the need for causal evidence on whether election results influence perceptions of legitimacy, particularly in non-democratic settings. The importance of such evidence is underscored by recent policy interest, including a commissioned report for the European Parliament on authoritarian legitimation through elections (Demmelhuber and Youngs, 2023). This policy brief presents findings from a recent study that addresses this issue, using a Russian election as a case study.

Figure 1. Trust in government and voter turnout in parliamentary elections in OECD countries (2017-2020)

Note: Trust is measured as a percentage of the population over 15 years old who answered ”Yes” to the following question in a nationally representative survey: “In this country, do you have confidence in the national government?” (Source: OECD). Turnout is the percentage of the registered voting population who voted in the last parliamentary election (Source: IDEA). Countries with compulsory voting are excluded.

Survey Experiment

To causally assess whether reported election outcomes influence perceptions of government legitimacy, the study implemented a survey experiment using a nationally representative sample of 1,603 Russian voters. The central feature of the design was a randomized information treatment that generated exogenous variation in respondents’ exposure to election results.

After completing the initial socio-demographic questions, respondents reported their prior political participation as well as their recollections of how past elections were conducted and their outcomes. Respondents were then randomly assigned to one of five treatment arms and asked to evaluate a hypothetical government formed after an upcoming election, with information about the election outcome randomly varied across treatment arms.

A control group received no information about hypothetical election results. Two groups were informed only about hypothetical voter turnout, which was presented as either low (38%) or high (66%). Two additional groups received information about turnout, either low or high, combined with a high vote share for the leading party (72%).

Following the information treatment, respondents reported their levels of trust in government, perceptions of whether the government represented national and personal interests, and their approval of and willingness to comply with hypothetical laws. These outcomes served as proxies for different dimensions of political legitimacy.

Election Outcomes Shape Trust in Government – but Only for Incumbent Supporters

By comparing responses across treatment groups, the experiment isolated the causal impact of election outcomes on legitimacy perceptions while holding constant respondents’ other characteristics. Respondents exposed to information about low turnout express significantly lower trust in government compared to those who received no information. On average, low turnout reduces trust by approximately 0.77 points on a ten-point scale, equivalent to about 0.25 standard deviations. In contrast, exposure to high turnout increases trust by around 0.68 points, or 0.22 standard deviations.

Providing additional information about the ruling party’s vote share does not significantly alter these effects. When high vote share information was combined with low turnout, trust increased slightly by 0.07 points, while adding vote share information to high turnout reduced trust by about 0.40 points; neither difference is statistically significant.

The impact of turnout information is highly heterogeneous. The observed effects are driven by individuals who expressed support for the ruling party, United Russia. Among these respondents, low turnout substantially lowers trust, while high turnout leads to a significant increase in trust. In contrast, opposition supporters do not update their perceptions in response to any of the information treatments: their trust levels remain statistically indistinguishable from those of the control group.

Moreover, the study examines heterogeneity based on baseline perceptions of electoral fraud. Before administering the information treatments, respondents were asked how frequently they believed irregularities in vote counting occur in Russian elections. Individuals who reported frequent violations are likely to view election outcomes as non-transparent and therefore to distrust official results, suggesting that information about turnout and vote share should have limited impact on their perceptions. Consistent with this expectation, no significant effect of turnout information on trust in government is observed among respondents who report a higher frequency of such violations.

Figure 2. Effect of information on trust relative to the control group

Note: This plot shows the effects of information treatments on trust in government relative to receiving no information (control group). Black circles are coefficient estimates for each group, with horizontal lines showing 95% confidence intervals.

Mechanisms: Expectation Shock and Anchoring

To examine how election information affects perceived legitimacy, the study relies on respondents’ reported recollections of past election results, including turnout and leading party performance. These prior beliefs provide a baseline against which new information is interpreted, as respondents tend to anchor their expectations about future elections to what they remember from previous ones.

When information about hypothetical election outcomes is presented, it generates exogenous shocks to these expectations. The magnitude and direction of each shock is defined as the difference between a respondent’s prior belief and the reported hypothetical outcome. By varying turnout between low and high values and combining turnout with high ruling party results, the experiment produced both positive and negative expectation shocks.

The results indicate that positive shocks, when the reported turnout exceeds prior beliefs, increase legitimacy across treatment groups, while negative shocks, when reported turnout is below the expected one, decrease legitimacy regardless of the treatment arm.

These findings suggest that election outcomes shape legitimacy by generating expectation shocks, and that respondents anchor beliefs about future elections to their perceptions of past results; in the absence of such anchoring, deviations between reported outcomes and respondents’ priors would have had little effect.

However, in the case of the leading party’s vote share, the resulting shock was rather small: an average respondent reported recalling the past vote share as 65%, while the value used in the information treatments was 72%. If respondents indeed anchor expectations about future election outcomes to past results, this may explain the absence of an additional effect of the vote share information, as the treatment did not generate a sufficiently strong expectation shock.

Conclusion

Do election results affect an autocrat’s perceived legitimacy? Using a survey experiment with a nationally representative sample of Russian voters, this study provides evidence that information about election outcomes can shape trust in government in an authoritarian setting. The results show that exposure to high (low) voter turnout increases (decreases) trust in government, with these effects concentrated among government supporters and individuals who believe elections are generally fair. This pattern suggests that autocrats may have limited ability to influence opposition supporters and instead rely on reinforcing legitimacy within their existing support base.

In addition, because voters anchor their expectations to past results, autocrats may be incentivized to generate higher outcomes while exercising caution in revealing lower ones in future elections. This underscores the role of autocratic elections as a tool to manage public perceptions over time.

The results of this study show that information about election outcomes holds strategic significance in non-democracies, as it can shape perceptions of government legitimacy. Policymakers should therefore prioritize support for independent media that provide credible information about election outcomes, even when results in authoritarian contexts appear predictable.

References

• Demmelhuber, T. and R. Youngs (2023). Strengthening the right to participate: legitimacy and resilience of electoral processes in illiberal political systems and authoritarian regimes. Technical Report PE702.581, European Parliament.
• Dukalskis, A. and J. Gerschewski (2017). What autocracies say (and what citizens hear): proposing four mechanisms of autocratic legitimation. Contemporary Politics 23(3), 251–268.
• Edmond, C. (2013, October). Information Manipulation, Coordination, and Regime Change. Review of Economic Studies 80(4), 1422–1458.
• Egorov, G., S. Guriev, and K. Sonin (2009). Why Resource-poor Dictators Allow Freer Media: A Theory and Evidence from Panel Data. American Political Science Review 103(4), 645–668.
• Egorov, G. and K. Sonin (2021). Elections in Non-Democracies. The Economic Journal 131(636), 1682–1716.
• Gandhi, J. and E. Lust-Okar (2009). Elections Under Authoritarianism. Annual Review of Political Science 12(1), 403–422.
• Gandhi, J. and A. Przeworski (2007). Authoritarian Institutions and the Survival of Autocrats. Comparative Political Studies 40(11), 1279–1301.
• Gehlbach, S. and A. Simpser (2015). Electoral Manipulation as Bureaucratic Control. American Journal of Political Science 59(1), 212–224.
• Gehlbach, S., K. Sonin, and M. W. Svolik (2016). Formal Models of Nondemocratic Politics. Annual Review of Political Science 19(1), 565–584.
• Guriev, S. and D. Treisman (2019). Informational Autocrats. Journal of Economic Perspectives 33(4), 100–127.

Disclaimer: Opinions expressed in policy briefs and other publications are those of the authors; they do not necessarily reflect those of the FREE Network and its research institutes.