The Optimal Standard of Proof in Criminal Law When Both Fairness and Deterrence Matter

This paper addresses the issue of the optimal standard of proof in criminal law.It is assumed that people in society care about both fairness and deterrence.It is important to punish those who are guilty and only those.However, error is unavoidable and hence a trade-o¤ emerges between the three aims of punishing the guilty, not punishing the innocent and deterring potential criminals.It is shown that when only deterrence matters the optimal standard of proof is a preponderance-of-the-evidence standard (given some other assumptions) while if fairness is an issue the standard will generally be stricter and involve Bayesian up-dating.When both fairness and deterrence matter the standard of proof will (generally) lie in between the two standards.An example illustrates how the model might be applied in practice to determine the optimal standard of proof for a given crime.