Climate Change and the Geography of the U.S. Economy

Abstract

This paper examines how the spatial distribution of people and jobs in the United States has been and will be impacted by climate change. Using novel county-level weather data from 1951 to 2020, we estimate the longer-run effects of weather on local population, employment, wages, and house prices using a panel distributed lag model. The historical results point to long-lasting negative effects of extreme temperatures on each of these outcomes. We highlight that a long lag structure is necessary to appropriately capture the longer-run effects of climate change, as short-run effects are often small and imprecisely estimated. Using county-level weather projections based on alternative greenhouse gas emissions scenarios, we use the estimated models to project the spatial distribution of these local economic outcomes out to 2050. The results point to substantial reallocations of people and jobs across the country over the next three decades, with mobility increasing by between 33 and 100 percent depending on the scenario. Population and employment are projected to shift away from the Sunbelt and toward the North and Mountain West. We document that this would, in fact, be a continuation of a historical pattern: Over the past four decades the relationship between population growth and hot climates across the United States has turned from strongly positive to slightly negative. We present a spatial equilibrium model to interpret the results, highlighting the impacts of climate change on amenities and productivity, and find significant roles for both channels in accounting for our empirical findings.