Research

I study how economic expectation shocks induced by elections contribute to democratic discontent in polarized societies, focusing on the 2022 Brazilian presidential election. Using newly collected survey data, I investigate how respondents' expectations and attitudes toward democracy evolved after the opposition candidate's narrow victory. By resurveying respondents post-election, I show that losing polarized voters who overestimate their candidate's chances of winning experience a large negative economic expectation shock, as well as an increase in their anti-democratic sentiment. In an additional survey experiment, I provide causal evidence that this shock is a driver of democratic discontent by showing that positively updating respondents' economic expectations reduces their anti-democratic sentiment.

Using new survey and experimental data, this paper investigates how salience and partisanship shape voters' willingness to support corrupt politicians. Drawing on salience theory, which models how decision-makers shift attention toward attributes that stand out most relative to the contextual average, I conduct a large-scale online survey experiment in the United States in which respondents evaluate hypothetical candidates varying in competence, corruption, and political affiliation. When party cues are absent and both candidates are highly corrupt, the gap in honesty between them appears small relative to the contextual average, making honesty less distinctive and causing voters to shift attention toward competence, ultimately increasing support for the more competent but also more corrupt candidate. However, once political affiliation is introduced, it strongly overrides other traits: voters overwhelmingly prefer co-partisans regardless of competence or honesty. A treatment mimicking increased media exposure to corruption scandals reinforces these patterns. These findings highlight salience-based attention and partisan loyalty as distinct mechanisms sustaining corruption.

Several recent studies examine the ease of voting across states using a “cost of voting index” (COVI) based on state election administration laws in a given election year (e.g., Pomante II et al., 2023). However, the original COVI is based on different sets of state laws across elections and normalized to have a mean of zero in each election. Consequently, the original COVI does not afford straightforward and meaningful comparisons of changes in the cost of voting over time. We gently revise the original COVI to generate a time-consistent “dynamic cost of voting index” (DCOVI) that is better suited to comparing within state changes in the cost of voting over time. We demonstrate that DCOVI is more strongly associated with state-level voter turnout than COVI, as well as a significant determinant of self-reported individual-level turnout and several measures of perceived frequency of illegal voting (albeit not perceived errors in vote counting). In general, higher costs of voting within a state are associated with lower voter turnout and improved public perceptions of the integrity of elections, although all of these effects are modest.

We analyze original survey data from Colombia, Italy, and Pakistan to study determinants of legislator preferences for redistribution. Large majorities of legislators in all three express pro-redistributive preferences, closely tracking preferences of citizens. Legislators’ preferences are no closer to those of affluent citizens than to the poor, and most legislators and citizens favor redistributive policies. Among legislators, partisan affiliation is the strongest predictor of redistributive preferences: legislators affiliated with leftwing or challenger parties are supportive of redistribution even in Colombia and Punjab, Pakistan — settings where party systems are considered weak and not organized on the traditional left-right economic dimension. Pro-social motivations for entering politics also independently predict support for redistribution, while material self-interest plays a limited role. These findings suggest that partisanship structures elite economic preferences even where conventional accounts would not predict it.

To understand the paucity of forceful anti-corruption policies, we study legislator and citizen beliefs and preferences about corruption in three countries. Deploying parallel surveys in Colombia, Italy, and Pakistan, we investigate support for a political agency theory that sees politicians as rent-seekers and for an information theory under which politicians misperceive voter preferences. We find limited support for either. Using vignettes that invoke tradeoffs that could result in corruption, we find that citizens and politicians in all three countries perceive corruption as common and are equally likely to condemn it. Politicians’ understanding of citizens’ concerns are largely accurate, and an information treatment informing legislators of citizens’ preferences leads legislators to believe citizens are less concerned about corruption, since the treatment primarily impacts those who initially overestimated citizen concerns. Our evidence suggests that feeble anti-corruption policy agendas may persist because established political parties lack electoral incentives to prioritize fighting corruption.

This paper studies how beliefs about racial inequalities and their causes vary and shape support for race-targeted and redistribution policies among Black and white Americans, including both adults and teenagers. We collect original large-scale survey data to provide new evidence on perceptions, attitudes, and policy views on racial issues and study the causal impact of information on policy views. We highlight significant heterogeneity along racial and partisan lines in perceived racial gaps in income and mobility. Yet, the biggest discrepancies are in how people explain the existence of these gaps, i.e., their perceived causes. White Democratic and Black respondents are much more likely to attribute racial inequities to systemic factors, such as adverse past and present circumstances, and want to act on them with race-targeted and general redistribution policies. White Republicans are more likely to attribute racial gaps to individual-based factors, such as individual effort or actions. These views are already deeply entrenched in teenagers, based on their race and their parents’ political affiliation. A policy decomposition shows that the perceived causes of racial inequities are the strongest predictors of support for race-targeted or redistribution policies, a finding confirmed by the experimental results.