Risk is the Reward: Extreme Weather Experience and Responses to Climate Change

Abstract

By using panel data from the American Panel Survey conducted from 2012-2017, I test whether individuals who experience an extreme weather event are more likely to agree that the federal government should take action to address climate change. While significant work has been performed on similar questions, I distinguish my approach from the existing literature in several ways. First, most previous research on the connections between climate change opinion and extreme weather has focused on climate change belief as the outcome of interest. By contrast, I am more concerned with better understanding the relationship, if any, between extreme weather experience and federal policy support. Secondly, by employing panel data and objective weather measures drawn from the NOAA Storm Events Database, I am able to strengthen causal inference by controlling for unobserved heterogeneity between individuals and identifying affected respondents with a high degree of precision, thereby responding to recent calls from researchers to center causal inference in climate change opinion studies. Thirdly, I examine whether extreme weather in the aggregate and particular kinds of weather events (e.g. floods, wildfires, heatwaves, etc.) affect individuals' climate change opinions, as opposed to many papers which only do one or the other. Finally, I test additional hypotheses less frequently tested in the literature, such as whether experiencing multiple instances of extreme weather makes a difference in climate change policy support and the way in which an individual's interest in politics does or does not mediate the mooted influence of partisan identity. I show that experience with extreme weather can associated with policy support, conditional on the kind of event, partisan identity, and an individual's interest in politics. However, bearing in mind the at times doubtful effects of public opinion on public policy and policymakers, as well as the potential disconnect between fleeting statements in a survey and actual behavior, I turn to examining the hypothesized effects of extreme weather on a group of individuals perhaps likelier to influence politicians: individual campaign donors. The literature on this question is quite limited, and my analysis consequently fills several gaps in the literature. First, and perhaps most importantly, I seek to add to the growing literature that has examined politically relevant revealed behavior in the wake of an extreme weather event, as opposed to stated opinions in a survey or poll (Liao and Junco 2022). Secondly, this analysis is performed using data thus far not employed in the literature: campaign contributions to gubernatorial candidates. This is highly relevant, as the states – as noted above – can be important laboratories for both climate-friendly and climate-skeptical policies. However, in contrast to the public opinion panel data, the campaign finance data used in this analysis are aggregated at the ZIP code level, which means I do not attempt to draw causal conclusions about the relationship between extreme weather and individuals' contribution behavior. Thirdly, I further expand on the available literature on revealed behavior in the wake of an extreme weather event by examining whether extreme weather events are associated with individual donors, again aggregated at the ZIP code level, giving to more or less extreme candidates, as measured by Bonica (2014)'s CFscores. I show that donations in gubernatorial races – as well as the ideology of candidates to whom donors give – appear to be either unaffected or slightly dampened by extreme weather events in a given ZIP code, contrary to previous findings in the research with regard to federal races.