Agricultural policy evaluation with large-scale observational farm data: Environmental impacts of agri-environmental schemes

Abstract

Agri-environmental schemes (AES) target at improving environmental status of cultivated land by remunerating farmers willing to commit to higher environmental standards. Thus far, no consensus exists whether AES incentivize adoption of pro-environmental production or simply offer windfall profits for those already operating at lower intensities. Using farm-level data, evaluation typically rests on comparing farms with and without AES. For differencing out unobservables that drive farmers into AES participation and therefore confound impact measurement, DID-matching methods are widespread, yet critical reflection remains sparse. We target at closing this gap by shedding light on the implicit assumptions about cause and effect paths linking participation and treatment outcome. We discuss challenges for identification of causal effects in presence of unobservable confounders over a broad range of methods and illustrate DID methods to estimate AES effects on land-use in West Germany.