Climate and Crop Yields in Australia, Brazil, China, Europe and the United States

Abstract

We examine whether nonlinear relationships between climatic conditions and crop yields are globally generalizable at subnational levels. We match subnational longitudinal administrative data on crop yields for barley, maize, oats, rice, soybeans, and wheat with growing season temperature and rainfall measures across Australia (n=406), Brazil (n=94,299), China (n=38,480), the European Union (n=15,505), and the United States (n=177,069). Using a flexible nonparametric approach that accounts for unobserved differences between locations, we estimate nineteen separate nonlinear climate-yield response surfaces. We verify that high temperatures and low rainfall substantially and significantly reduce crop yields across essentially all contexts and crops, with the exception of rice which tends to display more muted effects. Our results demonstrate a large and generalizable sensitivity of crop yields to climatic conditions across major crops in many of the world’s most productive agricultural regions.