Confounding Adaptation in Perennial Climate Damages: A Unified Statistical Approach for Brazilian Coffee

Climate change poses significant risks to agricultural production, especially for perennial crops, a crucial form of semi-durable capital in many developing countries. Misreported perennial yields and unobserved crop management decisions can undermine understanding of these risks. In the context of Brazilian coffee production, this study demonstrates that extreme temperatures not only reduce yields but also shrink reported harvest area due to plant death and farmers’ selective harvesting. The marginal damages from extreme temperatures on production are twice their effect on reported yields, as these are calculated using harvested area rather than bearing area. By merging the perennial supply and statistical yield literatures within a structural econometric framework, the effects of management decisions become distinguishable. The analysis reveals both a direct effect of extreme temperatures on biophysical yields and a multiyear plant death effect, and supports a more comprehensive understanding of how prices, weather, and adaptation interact across many perennial crops.