The economic of that has emerged in recent decades concerns itself largely with (Becker, 1968; Stigler, 1970). In its most basic form, the first establishes that increases in the expected punishment for deter some criminal acts at the margin, and then asks how resources should be allocated to enforcement in order to minimize the social costs of crime, including enforcement costs. Refinements of the have gone on to explore such questions as, what is the optimal form of punishment, a fine or imprisonment (Polinsky and Shavell, 1984)? And what is the optimal trade-off between the probability and the magnitude of punishment (Shavell, 1987)? While of is surely an important social concern, in the treatment of accused offenders is equally important. Indeed, this concern is clearly reflected in the structure of the criminal justice system in the U.S. by the numerous safeguards, many constitutional, that have arisen to protect the rights of defendants, as well as the popularly held belief that punishments of convicted offenders should fit the crime (Adelstein, 1979; and Wittman, 1974). Yet, consideration of has been largely absent from economic models of crime, despite the fact that economic analysis has much to say about it. In particular, since unfairness arises in the criminal process mainly as a result of imperfect and costly information, economic analysis can be utilized to characterize the optimal trade-off between the various aspects of fairness. For instance, an important question of relates to the incidence of errors by the criminal process--namely, conviction of truly innocent defendants (type I errors), and acquittal of truly guilty defendants (type II errors). Presumably, both impose some costs on society, therefore how should resources be allocated to minimize these costs? The in this paper answers this question. Specifically, section II develops a of the criminal justice system based on which focuses solely on the treatment of offenders after a has been committed. The fairness model thus adopts an ex post view of the criminal. In contrast, the standard deterrence model views the criminal from an ex ante perspective--that is, what is the best means of deterring his or her crimes (if at all)? It thus relies on a behavioral of the rational criminal 's choice to commit a crime. A simple version of such a is developed in section III. Section IV then grafts the onto the model. The impor-