Research

We present a methodology for decoupling taste-based versus statistical discrimination in political behavior. We combine a flexible empirical model of voting, featuring vertical and horizontal candidate differentiation in gender, ability, and policy positions, with a large-scale micro-targeted electoral experiment aimed at increasing female candidate vote shares. Our structural econometric approach allows us to separately identify preference parameters that drive taste-based discrimination and belief parameters that drive statistical discrimination via expectations about the ability and policy positions of female politicians. Our application to Brazilian municipal elections uncovers substantial levels of both taste-based and statistical discrimination. Counterfactual political campaigns show promise in reducing both.

The underrepresentation of minorities in key government bodies persists across all democratic institutions. For the U.S. House, scholars have identified two leading causes: voter discrimination and election aversion (lower political ambition), which are difficult to isolate from one another. To address this, the paper structurally estimates a model of political entry, voter discrimination, and campaign spending to separate the role of voter discrimination from that of election aversion in explaining underrepresentation. The framework also differentiates discrimination in primaries from that in general elections by modeling both stages. The model identifies election aversion by comparing general election outcomes (campaign spending and voting) between districts with only majority race (or gender) candidates and those with only minority candidates. When candidates on the ballot share the same identity, voters cannot discriminate based on identity. General election voter discrimination is then identified by comparing equilibrium outcomes between same-identity and mixed-identity districts. The dynamic structure of primaries followed by the general election allows me to account for candidate incentives and general election voter discrimination while recovering primary voter discrimination. I find that primary voter discrimination is the main driver of underrepresentation in the U.S. House. Although underrepresented groups show lower political ambition and face general election voter bias, these factors contribute minimally to underrepresentation. Policy counterfactuals show that a $150,000 campaign support subsidy during the primaries for underrepresented groups increases representation by 30% for Democrats and 177% for Republicans, while the same support in general elections has a negligible impact.

Political rallies constitute a large part of electoral campaigning in the U.S. and in modern democracies since the 19th century and remain a salient politico-economic phenomenon today. This paper accounts for candidates’ strategic decisions to rally as a finite-horizon dynamic game of electoral competition and applies it to structurally estimate rally spatial and temporal choices by candidates. For the 2012 and 2016 U.S. presidential elections, we show that rallies substantially increase poll margin leads in targeted constituencies over non-rallying opponents and trigger systematic dynamic responses by opponents. In terms of magnitudes, rallies by presidential candidates are more persuasive than television ads, and estimates of the gross effect show that President Trump’s rallies were in fact electorally pivotal. Instead, rallies by all other candidates did not change their win probability. Counterfactual policy experiments reveal that the effects of short-term campaign silences (i.e., electoral blackouts) are limited since candidates can time their rallies and gain sufficient support from the electorate before campaign silences begin.

While studying government formation in parliamentary democracies, researchers have always assumed that political parties possess identical preferences over government portfolios. This has led us to rely on games that work under the assumption that there is “one pie,” and every coalition member wants it. In this paper, I show that this is not the case in the context of the Western European Parliaments from 1965–2018. I document novel patterns showing that right party politicians were more likely to be Ministers of Defense, Agriculture, Justice, and Prime Ministers, while the left was more likely to be allocated to Labor, Environment, Health, Science and Technology, Education, and Transport departments. These patterns indicate horizontal differentiation of government portfolios by political parties. Traditional bargaining models lack flexibility for heterogeneous preferences; hence, I model this as a Colonel Blotto game, allowing identification of party preferences as a function of ideology. Counterfactual experiments reveal how heterogeneous preferences and strategic interactions explain observed allocations.

The paper constructs a model of dynamic electoral competition where politicians compete against each other to stay popular on election day. The model possesses a finite time horizon with a perfect information structure which results in a unique equilibrium—a contribution in itself. An extensive simulation study is conducted to understand the model's comparative statics, which provides essential intuition that can be used to explain the dynamics of campaign strategy used in practice. The model is applied to the 2019 Indian General Election to test how fairly it performs. This paper also provides one of the first examinations of Modi rallies that helps us to get a sense of how effective Modi rallies were in the 2019 Indian General Elections.

We study identity-based reactions to economic disparities in the context of anti-immigration attitudes in the United States. We develop a structural model where rising aspirations can both inspire and frustrate. If an individual’s economic status exceeds their social aspirations, they may invest more in monetary terms, benefiting future generations. However, if aspirations rise beyond feasible levels, frustration may lead to lower investment and redirect resources toward identity-based pursuits. Using this framework, we quantify how frustrated aspirations from inequality contribute to anti-immigrant sentiments and assess redistribution policies to mitigate these effects.

We analyze a 13-year dataset from the São Paulo Court of Justice, observing judges’ promotion outcomes, case closures, and seniority rankings. We find that higher cumulative case closures predict promotion likelihood, but marginal productivity decreases with seniority. We structurally estimate judges’ case-closing decisions to test whether merit-based promotions increase productivity or incentivize premature case closure to inflate perceived merit.