Estimating the Consequences of Climate Change from Variation in Weather

I formally demonstrate how to use the effects of short-run weather shocks to learn about the consequences of long-run climate change. I show that short-run adaptive responses to weather shocks differ from long-run responses to climate change when payoffs depend on a capital or resource stock. I derive a new indirect least squares estimator that bounds long-run climate impacts from short-run responses to weather. In an application to U.S. counties’ economic output, I show that conventional methods project end-of-century losses of 8–12%, depending on the region of the country, whereas my theoretically grounded estimator projects larger losses of at least 12–20%.