Evaluating Leniency with Missing Information on Undetected Cartels: Exploring Time-Varying Policy Impacts on Cartel Duration

Abstract

This paper examines the effects of European Commission's (EC) new leniency program on the EC's capabilities in detecting and deterring cartels. As a supplementary analysis, the US leniency is studied. I discuss a dynamic model of cartel formation and dissolution to illustrate how changes in antitrust policies and economic conditions might affect cartel duration. Comparative statics results are then corroborated with empirical estimates of hazard functions adjusted to account for both the heterogeneity of cartels and the time-varying policy impacts suggested by theory. Contrary to earlier studies, my statistical tests are consistent with the theoretic predictions that following an efficacious leniency program, the average duration of discovered cartels rises in the short run and falls in the long run. The results shed light on the design of enforcement programs against cartels and other forms of conspiracy.

Keywords: evaluation of antitrust policies, time-varying policy effects, sample selection bias, leniency, semi-parametric estimation

JEL Classification: D43, K21, K42, L13