Distributional and temporal heterogeneity in the climate change effects on U.S. agriculture

• Published in: Journal of environmental economics and management 2020-11, Vol.104, p.102386, Article 102386
• Main Authors: Malikov, Emir, Miao, Ruiqing, Zhang, Jingfang
• Format: Article
• Language: English
• Subjects: Adaptation; Agricultural industry; Agriculture; Climate change; Crop yields; Discrimination against disabled persons; Environmental aspects; Global temperature changes; Heterogeneity; Quantile regression; Soybean; Weather
• ISSN: 0095-0696; 1096-0449
• DOI: 10.1016/j.jeem.2020.102386

Existing studies on climate change effects on crop yields mainly focus on the average climate–yield relationship that is typically assumed to be time-invariant. We apply a flexible panel-data quantile regression with time-varying coefficients to examine distributional heterogeneity and temporal variation in this relationship. We find that U.S. corn and soybean yields have gradually become less sensitive to temperature and precipitation over 1948–2010, which is especially the case for upper yield quantiles. Consequently, the negative impacts of future climate change are of larger magnitudes at the lower yield quantiles. Failure to accommodate temporal changes in the climate–yield relationship leads to significantly overestimated responsiveness of crop yields to weather variation and, therefore, overestimated negative impacts of future climate change. On many occasions, the corn yield decline projections from such time-invariant specifications are about twice as large as (and sometimes triple) the predictions from our time-varying-coefficient model.