THE BECKER PARADOX AND TYPE I VERSUS TYPE II ERRORS IN THE ECONOMICS OF CRIME∗

Two real-world observations are not easily replicated in models of crime. First, although capital punishment is optimal in the standard Becker model, it is rarely observed in the real world. Second, criminal procedure and the evaluation of evidence vary across societies and historical periods, the standard of proof being sometimes very high and sometimes quite low. In this article, we develop a general equilibrium model of judicial procedure allowing for innocent persons being convicted. We show that the median voter theorem applies to this model, making judicial procedure endogenous. So formulated, the model can replicate both empirical observations.