Natural Disasters and Electoral Consequences - Evidence for Voter Participation and Choice in U.S. Presidential Elections

This study investigates how natural disasters and federal disaster declarations influence electoral outcomes in U.S. presidential elections from 1996 to 2020. Using county-level panel data and a two-way fixed effects framework, we analyze both incumbent vote share and voter turnout to describe the mechanisms linking disasters to political accountability. We find that severe disasters significantly reduce the incumbent party's vote share, consistent with the retrospective voting hypothesis. However, Presidential Disaster Declarations (PDDs) substantially mitigate these electoral losses, particularly in politically aligned urban areas, supporting the attentive retrospection hypothesis. Disasters tend to suppress voter turnout moderately, and PDDs only partially offset this effect, suggesting that disasters primarily harm incumbents by prompting voters to switch to the opposition rather than by discouraging their supporters from participating. Categorical analyses further reveal that partisan alignment, rather than PDD issuance alone, consistently drives electoral responses. Overall, our findings highlight how institutional coordination and federal aid shape democratic accountability in the wake of natural disasters.