Estimating the Impact of Regulatory Program Enforcement

Abstract

The most common model for estimating the impact of regulatory program enforcement on intended outcomes is based on a standard version of deterrence theory. The corresponding estimating equation regards the amount of noncompliance by a firm as a linear function of enforcement by the regulatory agency, holding a variety of other variables constant. This article shows, based on recent theories in formal political science about how regulatory programs are likely to be implemented, that an equation of this form is likely to produce biased estimates. Specifically, the literature implies that the function should probably be a second-degree polynomial; that it should be two simultaneous equations, with one for compliance and one for enforcement; and that relevant equations should include variables that characterize the political coloration of the local area in which enforcement and compliance decisions are made. The general conclusion is that theory and method cannot and need not be separated. Evaluators of regulatory programs who overlook formal models of regulatory program implementation and who rely on outmoded models of deterrence will produce incorrect findings about the effective ness of those programs