It is suggested that an alternative interpretation of Ehrlich's 1975 study may be made from his findings. Little attention has been given to the identification of Ehrlich's 'murder supply function,' an equation representing the response of murder behavior to the punishments meted out by society. A structural model provides estimates of a murder supply function and assumes that these estimates may actually represent the response of society's punishments to murder behavior. Because the identification of a murder supply function requires an exact knowledge of each equation in the complete structural model, the present study provides a structural model that includes equations representing the response of society to murder behavior. First, the structural model is described and the a priori identification of its equations is discussed. Second, estimates of the model's coefficients and the results of empirical tests of validity are presented. Third, the alternative interpretation of the causality underlying Ehrlich's estimates is provided. Results obtained from applying the model suggest that perhaps the true murder supply function is one in which the effect of the probability of apprehension is larger than that found by Ehrlich. In addition, the effects of the conditional probability of conviction given apprehension and execution given conviction, are zero. Ehrlich found that the coefficients for the deterrence variables were negative, larger than their standard errors, and in the ascending order predicted by his theory. It is suggested that these impressive results could have been caused by the response of the criminal justice system to orders. Thus, Ehrlich's murder supply function, if not identified, could represent criminal justice system behavior and not murder behavior. Four tables, numerous equations, 27 footnotes, and approximately 47 references are provided.