On the GDP effects of severe physical hazards

Abstract

We use local projection methods with instrumental variables to analyze a panel dataset that links monetary damages and geophysical hazard strength (which serve as our instruments), associated with a wide range of severe weather events. This approach allows us to understand the GDP impact of these events at the country level. The estimated impulse response functions indicate a persistent GDP decline lasting several years after an increase in disaster-related monetary damages. More severe disasters leave a disproportionately larger negative effect on the economy. For the top 10 percent of disasters, GDP remains approximately 2 percent lower in the medium term (5-7 years) and does not fully recover over the 10-year horizon of our analysis. When disaggregating by disaster type, we find similar results across categories, with storms emerging as the primary driver. High-income countries experience significantly smaller effects than middle- and low-income countries. Our findings are robust to alternative impact measures, such as disaster-related deaths, the number of people affected, or a simple disaster occurrence indicator.