Reasonable doubt and the optimal magnitude of fines: should the penalty fit the crime?

Models of the enforcement-compliance relationship generally treat both the probability and magnitude of fines as independent choice variables of the government. This approach has led Becker (1968) and others1 to conclude that it is optimal to set uniformly maximal penalties for all crimes, and to set the probability of conviction at the minimum level necessary to enforce compliance with the law. While this normative prescription is natural and intuitive, there still remains a larger positive question of why in most nations of the world penalties are not uniformly high, but rather rise with the severity of the crime.