The Electoral Consequences of Natural Disasters: A Dynamic Fixed-Effects Analysis

With the increasing frequency of major natural disasters, understanding their political consequences is of paramount importance for democratic accountability. The existing literature is deeply divided, with some studies finding that voters punish incumbents for disaster-related damages, while others find they reward them for relief efforts. This paper investigates the electoral consequences of natural disasters for incumbent mayors, broader electoral dynamics, and the long-term political ambition of officeholders. The study leverages a comprehensive panel dataset of over 10,000 candidate-election observations in U.S. mayoral races from 1989 to 2021, combining detailed election data with a global registry of disaster events. To identify causal effects, the analysis employs a robust dynamic two-way fixed-effects event-study design, validated by extensive pre-trend and placebo tests. The findings reveal that the electoral impact of disasters is highly conditional on their timing. A disaster that strikes in the same quarter as an election provides a significant electoral boost to incumbents, increasing their vote share by over 6 percentage points. However, disasters consistently suppress voter turnout, reducing it by an average of 1.4 percentage points. In a novel finding, the analysis demonstrates that the experience of managing a disaster significantly increases an incumbent's likelihood of seeking re-election in the subsequent cycle by as much as 12 percentage points. These findings help reconcile conflicting theories of retrospective voting by highlighting the critical role of voter myopia and salience. They also reveal a previously undocumented channel through which crises shape political careers, suggesting that disaster management is not only a test of governance but also a catalyst for political ambition. [The current version is a preprint.]