Use of Matching to Evaluate Environmental Regulatory Enforcement

Abstract

In this paper, we measure the treatment difference between administrative and criminal enforcement of environmental violations. Our aim is to make a comparison between similar offenses and evaluate how they would have been treated in both enforcement tracks. This paper can help to evaluate the compliance and welfare outcomes in a complementary criminal/administrative enforcement system with an enforcement system relying on criminal prosecution only. Given the discussions at European level on the most appropriate enforcement system and the different practices at national level, answering our research question is quite relevant. We apply statistical matching techniques on a unique dataset of environmental enforcement cases to control for sample selection bias and approximate the setup of a randomized experiment. We do the matching on case characteristics and estimate the average treatment effect for similar cases. Overall, we find that the marginal penalty is slightly lower in administrative enforcement compared to criminal enforcement, but that this difference would be widely overestimated without the matching procedure. Our methodology can be more generally applicable than the environmental law context that we use: it can help to control for caseload difference, which is a typical problem in law enforcement evaluation, and allow for correct causal inferences in a wide variety of law enforcement evaluation settings.