Abstract
This paper presents a method for estimating treatment effects of local climate shocks when regions trade with each other. Because of spillovers induced by trade flows, comparing the evolution of outcomes between pre-shock and post-shock periods in regions exposed versus unexposed to local shocks leads to a biased estimate of treatment effect. We model these across-region dependencies using standard assumptions from international trade theory. We use our model-consistent estimation strategy to revisit the literature on the evaluation of impacts from climate change onto country-level gross output using year-to-year variation in temperature.