Environmental management practices for mitigating nutrient runo aect the productivity of agricultural land. Finland's agricultural policy oers a number of such practices, and we use Finnish grain farms as a case study of these productivity impacts. Productivity is endogenous with adoption when decisionmakers can choose from a menu of policy options, but few studies actually address this endogeneity. Our identication strategy thus involves a novel approach for correcting endogeneity, combining classic methods from stochastic frontier analysis and selection models. Using registry data from 2007-2013, we nd that the more ecient farms are also more likely to enroll in management practices. Standard estimates without the endogeneity correction understate productivity losses after adoption.