Evaluating voluntary environmental programs with spillover effects

Abstract

Voluntary environmental programs have the potential to impact not only participants but also nonparticipants, through, for example, the dissemination of information about abatement technologies, thereby creating spillover effects. Standard program evaluation techniques do not accurately estimate program impacts in such cases. This paper presents a theoretical and empirical analysis of spillovers in a voluntary environmental program (VEP). The theoretical model identifies the key distinction between the impact of participation and the impact of the program, as well as the nature of the bias that results from failure to incorporate spillover effects. It also motivates the inclusion of spillovers in both the participation and outcome equations of the empirical model. Using data from EPA’s 33/50 Program, we show that failure to account for spillovers can lead to significant underestimation of program impacts.