A function is developed for the supply of offenses which relates the number of offenses committed by a person to variables that affect the expected utility derived from the offense, such as its probable value, the probability of being punished, and the severity of punishment. In addition, an examination is made of how society helps to determine the level of safety, seen as a function relating the probability of conviction and police expenditures. Since enormous expenditures cannot be made by most municipalities, decisions must be made by society to determine the amount of safety to be purchased. A study of 181 New Jersey municipalities is used to implement the model. All of the major empirical results were consistent with the assumptions of the model. The rate of conviction was shown to have a significant positive elasticity on the safety rate. Those environmental variables which can be interpreted as indicating an above average net benefit from crime produced negative elasticities, while those variables that are believed to reduce the incentive of offenders displayed positive effects on the demand for safety. It was shown that an increase in marginal cost produced a negative impact on the quantity of safety demanded. Increases in exogenous variables such as income and wealth displayed positive elasticities on the demand for safety. The model is of value to social planners seeking to estimate gains from criminal justice programs and measures. Five tables, six appendices showing measures and other variables' impacts, and a bibliography of 26 citations are included.