Publications
-
Strategic Replacements and Popular Support: Political Consequences of Authoritarian Replacement of Elected Officials (JOP, 2025)
This paper presents a game-theoretic model to analyze a hybrid form of governance, combining competitive local elections with the central government's unilateral power to intervene and replace elected officials. The equilibrium analysis highlights the implications of the central government's trade-off between local officials' competence and their partisan affiliation. I show that the key consequences of this trade-off include 1) strategic retentions of underperforming opposition incumbents and 2) replacements of pro-regime incumbents, even when their expected competence exceeds that of their electoral replacements. I demonstrate, then, that this institutional framework induces a higher proportion of replaced opposition officials compared to co-partisans of the regime and encourages the population to electorally support candidates affiliated with the central government party in prior open-seat elections. The observed local electoral performance of the regime, under the influence of this hybrid institution, consistently surpasses its true popularity, even in the presence of competitive elections.
-
Advice and Competence (AJPS, 2024)
We develop a theory of policy advice that focuses on the relationship between the competence of the advisor (e.g., an expert bureaucracy) and the quality of advice that the leader may expect. We describe important tensions between these features present in a wide class of substantively important circumstances. These tensions point to the presence of a trade-off between receiving advice more often and receiving more informative advice. The optimal realization of this trade-off for the leader sometimes induces her to prefer advisors of limited competence -- a preference that, we show, is robust under different informational assumptions. We consider how institutional tools available to leaders affect preferences for advisor competence and the quality of advice they may expect to receive in equilibrium.
Working Papers
-
Voter Resentment and the Strategic Dynamics of Enfranchisement
Political parties in electoral democracies face a strategic dilemma regarding franchise extension: immediate resistance offers short-term electoral insulation but risks long-term viability due to voter resentment. We develop a dynamic formal model to analyze why parties might resist enfranchising new voter groups, despite the potential for long-term electoral consequences. A core mechanism in our model is voter resentment: parties that obstruct enfranchisement accumulate negative sentiment among the affected group, leading to an electoral penalty if enfranchisement eventually occurs. Our analysis demonstrates that while resisting enfranchisement can offer short-term protection against electoral loss, it risks the party's long-run electoral viability. There are three key findings: (1) Increased voter sensitivity to past disenfranchisement generally incentivizes incumbent parties to enfranchise sooner. However, (2) this effect is moderated by the party's existing partisan advantage. Counterintuitively, parties with a strong initial partisan base may become less likely to enfranchise as voter sensitivity to past obstruction rises, opting to rely on their core supporters rather than appeal to a resentful new group. (3) The precision of information regarding the new group's latent support also interacts with partisan strength; higher precision can encourage enfranchisement for electorally strong parties but deter it for weaker ones. We contextualize the theoretical results within key historical movements of (dis)enfranchisement, including that of Black Americans and of women in the Netherlands and Italy.
-
The Politics of Anti-Corruption: Bureaucratic Sabotage and Collusion
Why does the level of corruption sometimes react paradoxically to anti-corruption efforts? While classic models explain graft through the industrial organization of state agencies, this paper develops a dynamic political model to show how electoral pressures shape bureaucratic corruption. The model centers on a rent-seeking bureaucrat whose choice of corruption level serves as a public signal of an incumbent's unobserved commitment to enforcement. A tough, anti-corruption incumbent can inadvertently trigger strategic sabotage, where the bureaucrat inflates graft to provoke the incumbent's electoral removal, hoping for a more lenient replacement. Conversely, a bureaucrat may engage in strategic collusion under a lenient incumbent, deliberately moderating their bribe-taking to protect a mutually beneficial corrupt equilibrium from electoral backlash. These opposing incentives create a non-monotonic relationship between enforcement and graft: the lowest level of corruption emerges under an intermediate, rather than a maximal, anti-corruption regime. Moreover, strong state capacity can limit the scope for sabotage while encouraging collusive arrangements by raising the expected toughness of any potential challenger. By endogenizing the political environment, the analysis provides a new lens for understanding the persistence of corruption and the often-unintended consequences of anti-corruption reforms.
-
Non-monotonic Disclosure in Policy Advice
The strategic context of bureaucratic advice to policymakers often takes the form of a disclosure game in which the relevant bureaucracy has an ideal policy interior to the policymaker's action space. We characterize conditions under which this game has sequential equilibria in which the sender adopts a non-monotonic disclosure strategy, implying partial disclosure. Further, multiple sequential equilibria exist under some conditions, including a fully revealing equilibrium and multiple partially revealing equilibria that vary in extent of disclosure. We show that these equilibria are strictly rankable both by actors' welfare and by their robustness to belief perturbations. Further, for sender preferences that are sufficiently close to the expected value of the state, (1) the most robust equilibrium is partially revealing and (2) set of states that the sender discloses becomes larger as the divergence in sender's and receiver's ex ante preferences increases.