Document Type

Dissertation

Degree

Doctor of Philosophy

Major

Political Science

Date of Defense

4-14-2026

Graduate Advisor

David Kimball, PhD (Political Science)

Committee

David Kimball, PhD

Anita Manion, PhD

Todd Swanstrom, PhD

Stephen Bagwell, PhD

Abstract

This dissertation examines how natural disasters shape electoral behavior in gubernatorial and presidential elections in the United States, with particular attention to geographic proximity, disaster severity, and partisan accountability. Using county-level gubernatorial and presidential election returns from 2016 to 2022, it analyzes the electoral effects of hurricanes, wildfires, and tornadoes through three frameworks: direct disaster impact, near-distance exposure, and disaster severity. Employing the difference-in-differences and fixed effects models across more than 31,000 county-year observations, the study finds that natural disasters can significantly reduce incumbent party vote share and voter turnout, particularly in gubernational elections. Results further demonstrate that electoral effects often extend beyond directly impacted counties and vary substantially by disaster type and election year. The dissertation’s most consistent finding is a partisan asymmetry in accountability where Republican incumbents generally experience larger electoral penalties than Democratic incumbents following comparable disaster exposure. These findings complicate traditional retrospective voting theory and highlight the geographic conditionality of democratic accountability in an era of increasing climate-related disasters.

Download

Share

COinS