David Lobell
Stanford University
“Are satellite data ready for agricultural economists? ”
Abstract: Satellites have been observing the Earth for many decades, with monitoring of agriculture a key application since the beginning. Yet the generation of quantitative datasets has been lacking, hampering the ability of econometricians to achieve their divine purpose of causal inference, especially in developing countries where existing agricultural datasets are limited. Recent advances both in sensors and algorithms have been improving the scope for tracking key agricultural outcomes. This talk will review these advances, with particular emphasis on (i) how yield estimates from satellite compare to traditional ground measures, and (ii) examples of how these estimates can be applied to quantify associations and assess causality in both developed and developing country settings. It will also review remaining obstacles to make satellite-based datsasets truly mainstream for use by econometricians.