The Democrat Disaster: Hurricane Exposure, Risk Aversion and Insurance Demand

81 Pages Posted: 21 Jan 2020

See all articles by Raluca L. Pahontu

Raluca L. Pahontu

Nuffield College, University of Oxford

Date Written: January 2020


How do voters respond to heightened risk? Dominant theories expect issues of accountability to surface or distributional conflict to intensify once threats become salient. Unsatisfactorily, these accounts rely on compound treatment effects of exposure not only to risk but also to direct losses or self-selection into unfortunate circumstances. To circumvent these issues, I use difference-in-differences estimates of nearly-hit disaster high-risk areas in the United States to study the effect of risk on vote choice. I identify significant electoral penalties for the Democratic Party whose vote share decreases following a near miss for both US House and Senate races between 2002 and 2014. Conventional explanations related to religiosity, authority, information, or competence fail to explain this effect. Instead, I propose that Republican gains are driven by voters' spending on private insurance and increased willingness to take risks when spared from disaster. I therefore advance an alternative theoretical explanation for vote choice under uncertainty. Relying on novel data on hurricane trajectories, longitudinal precinct electoral returns, risk aversion and private insurance inquiries, these results are politically meaningful not least because US general elections follow closely after the hurricane season.

Keywords: causal inference, voter behaviour, risk, natural disaster, insurance, uncertainty

JEL Classification: D72, D81, H20, Q54, P16, C23, C55, D60, P18

Suggested Citation

Pahontu, Raluca L., The Democrat Disaster: Hurricane Exposure, Risk Aversion and Insurance Demand (January 2020). Available at SSRN: or

Raluca L. Pahontu (Contact Author)

Nuffield College, University of Oxford ( email )

New Road
United Kingdom

Register to save articles to
your library


Paper statistics

Abstract Views
PlumX Metrics