Soaring crime rates and law enforcement costs have produced widespread disillusionment and disenchantment with our criminal justice system. Even those who operate its basic institutions, police officers, judges, and correctional officials, are frequent critics of its performance, although they often differ about what wrong. The mood of the country with regard to crime and law enforcement is one of frustration and bewilderment.1 This discontent has resulted in the questioning of our basic law enforcement institutions and our basic concepts of justice. While the system as a whole has been subjected to critical re-examination, the evaluation of performance has focused on the police. What we have been asking ourselves and our public officials is, can we stem the rising tide of crime, and if so, what will it cost? The attempt to obtain answers has raised additional questions. With regard to the police, the three critical questions are: (i) what do the police produce, or should they produce, and how can it be measured, (2) are the police producing at minimum cost, and (3) how much police service do we want to buy? As we shall demonstrate, the answers to these questions are a part of the overall answer to the same set of questions posed for the criminal justice system as a whole. A fundamental tenet of this paper that the activities of the police cannot be measured and evaluated without reference to the totality of criminal justice institutions and to the social environment in which these activities take place. This paper discusses the conceptual and practical difficulties of defining and measuring police output. The first section sketches the overall problem of criminal justice resource allocation as it relates to the measurement of police output. We argue that police services are not a final product, but rather an intermediate product in the overall production of justice or law enforcement. The effectiveness of the police with regard to overall law enforcement depends not only on their own operations but on the operations of the courts, the prosecutor's office, the correctional system, and the other participants in law enforcement. We then introduce the concept of the social cost of crime and demonstrate that the problem of defining justice can be circumvented-