This paper seeks to assess the robustness of the assumptions made by much of the theoretical literature on leniency programs, giving a glimpse of the uncertainties and complexities that apply in practice. First, sanctions are hard to estimate and the decision to form a cartel is not generally made by the firm as a rational monolith. Second, empirical evidence from the EU suggests an over-reliance on leniency, with only a weak threat of detection through investigations alone. Most leniency reporting may be occurring where a cartel has already ceased to operate or is very likely to be caught. Finally, the decision to come forward is not one that is taken lightly by the firm; it is fraught with uncertainties and dangers, including the challenges of ensuring cooperation from employees. The paper concludes with three recommendations for the strengthening of leniency policies.
