Deterrence, Work and Crime: Revisiting the Issues with Birth Cohort Data

In this paper we analyze criminal behavior of a cohort sample of young men over an eight year period using random effects probit and Tobit techniques. Our major advances relate to our careful conceptualization general deterrence, and our data. As far as we are aware, this work represents the first time that a richly specified model of criminal activity has been estimated using panel data for a general population group. We find very robust evidence for a general deterrent effect emanating from police resources. Our results regarding general deterrence are open to fewer questions than previous findings. We also find that working and going to school significantly decrease the probability of commmitting criminal acts and by virtually identical amounts. This similarity of effect when coupled with other findings suggests that crime does not serve mainly as a direct source of income and that incentive effects emanating from higher wages are not very strong. There is little empirical support for the crime as work' model that has dominated economic thought regarding crime over the last two decades. More fruitful models of work and crime may result if work is conceived as having its primary effects either through preferences or through information.