The effects of extreme weather events on self-protection investments

Abstract

We use a unique data set from a tornado shelter (safe-room) rebate program in Arkansas from 2006 through 2010 to examine the role of risk perceptions in stimulating homeowner investments in self-protection. Using empirical models that explore both extensive non-parametric specifications and streamlined parametric forms for lagged responses, we find that the decision to self-protect depends clearly upon both the recency and the proximity of tornado events, as well as on average education and income levels in the county in question. The pulse in self-protection investment after a tornado is relatively large yet short-lived and relatively local, and there is some evidence that short-run supply constraints limit the expression of peak demand over time and across space. Our findings suggest the potential importance of self-protection investments as an adaptive response to changes in the severity and spatial extent of extreme weather risks that may result from climate change. We also highlight the role that rebate policy design may play if the goal is to encourage self-protection investments at least cost.