Essays on the Political Aftermath of Natural Disasters

Abstract

In a series of essays, this dissertation examines three political phenomena in the context of natural disasters: retrospective voting, civic engagement, and social capital. In “Voting after the Storm: Hurricanes, Electoral Retrospection, and the 2008 US Presidential Election,” I investigate the ability of retrospective voters to hold incumbent politicians accountable for their performance in office. Using county-level data on severe weather events and presidential election returns, I find that counties affected by election-year hurricanes increased their support for incumbent-party candidate John McCain by at least three percentage points in the 2008 presidential election. This essay’s main findings provide evidence that supports the existence of an attentive electorate, one in which retrospective voters evaluate politicians’ responsibility for outcomes, rather than blindly punishing them for events beyond their control. In “Civic Engagement in the Aftermath of Natural Disasters,” I explore whether American citizens become more or less civic when their communities are struck by natural disasters. Using individual-level survey data on civic engagement and county-level data on natural hazard events, I find that experiencing a natural disaster in the previous 12 months decreases the probability of an individual engaging in non-electoral forms of political participation—by as much as five percentage points—but increases the probability of an individual volunteering and giving to charity—by up to six percentage points. Finally, in “Pulling Together or Pulling Apart? Social Capital in the Aftermath of the 2010 Benin Floods,” I analyze the impact of the historic 2010 floods in the Republic of Benin on stocks of social capital among affected individuals. Using original household survey data collected in flood-affected communes, I show that individuals’ experiences during the floods—and in their immediate aftermath—are strongly associated with indicators of social capital observed years later. In particular, the occurrence of grassroots disaster-relief efforts and the formation of new voluntary associations following the floods are positively related to present-day social capital, while disaster-induced hardship is negatively associated with subsequent social capital levels. These results suggest that natural disasters can trigger social dynamics that produce lasting change in affected communities’ social capital stocks.