The impacts of climate change on agriculture in sub-Saharan Africa: A spatial panel data approach

This paper reports estimates of the economic impact of changes in weather variables on sub-Saharan African pearl millet yield based on panel data for 1970–2016. We control for spatial effects in all the components of our exposure–response function, plus a lag in time of the covariates through spatio-temporal econometrics techniques. Our results indicate own-location weather variables have significant contemporaneous impacts on millet yield. Specifically, we find that vapor pressure deficit, wet day frequency and temperature are important determinants of millet yield. In addition, accounting for spatial and temporal spillovers exacerbates and attenuates wet day cumulative effect, respectively, and local crop production is affected by neighboring countries’ production. The results are robust to several sensitivity checks, including accounting for adaptation using long-term averages, and are consistent across country-income groups. We also use our estimates to forecast how crop production would respond to climate change in the mid-future.