Seminario CEDE - Rachid Laajaj

This paper presents a new approach to decompose the effects of climate change on agriculture into the direct effect of a change in climate conditions and the effect of imperfect climate adaptation. We first provide a theoretical framework showing that to capture both effects, a regression should model agricultural outcomes as a function of both the change in expected weather conditions and its absolute value. The former captures the direct effect, which may be positive or negative, while the absolute value accounts for the fact that any deviation in expected weather, regardless of its direction, creates a need for adaptation, causing inefficiency until the farmer fully adjusts her farming decisions. We further derive econometrically how, in the presence of both a direct and adaptation effect, the omission of the absolute value of the change in expected weather, a common limitation in previous studies, compromises the interpretation of the results. Our analysis predicts that the coefficient(s) in front of the absolute value in weather change should be negative in the presence of a Climate Adaptation Gap (CAG). We take this prediction to US and Colombian data, using the long-differences methodology and variations in expected temperature and precipitation over long periods.  The results confirm the existence of a significant CAG in both countries, though substantially larger in Colombia.