The Role of Nonfarm Influences in Ricardian Estimates of Climate Change Impacts on US Agriculture

Abstract

The Ricardian approach is a popular hedonic method for analyzing climate change impacts on agriculture. The approach typically relies on a cross‐sectional regression of farmland asset prices on fixed climate variables, making it particularly vulnerable to omitted variables. I conduct a long‐spanning Ricardian analysis of farmland prices in the eastern United States (1950–2012) and find a convergence of evidence indicating that large estimates of climate change damages for recent cross‐sections (>1970s), also found in the literature, can be explained by the growing influence of omitted factors extraneous to the agricultural sector. I propose and evaluate a simple strategy to circumvent such nonfarm influences in the form of a Ricardian model based on cash rents (2009–2016), which better reflect agricultural profitability and do not capitalize expected land use changes. The new damage estimates on nonirrigated cropland and pasture rents are more optimistic and cannot be distinguished from zero. However, estimates remain imprecise under extreme climate change scenarios pointing to a cautionary long‐term outlook for United States agriculture. The findings are robust to multiple checks and alternative explanations.