Playing without a rulebook: Optimal enforcement when individuals learn the penalty only by committing the crime

This paper argues that enforcement can be designed not only to deter individuals, but also to inform them. Individuals may commit sanctionable acts whilf being imperfectly informed about the government's enforcement policy. By engaging in behavior repeatedly, and getting caught occasionally, individuals learn some of the information over time. In this setting, Becker's argument for imposing maximal sanctions along with a low probability of apprehension may not hold. Raising the probability of apprehension increases the number of occasions in which individuals get caught, giving them more opportunities to learn the enforcement policy. The acquired information improves behavior, a benefit that may exceed the added enforcement cost. © 1997 by Elsevier Science Inc.