Criminal Victims, Victimized Criminals, or Both? an Econometric Analysis of the Victim‐offender Overlap

Offenders are more likely than non‐offenders to be victims, and victims are more likely than non‐victims to be offenders. So far, the overlap between offenders and victims is not well understood in criminology, and in the economics of crime this stylized empirical fact is even widely ignored. One of the major limiting problems seems to be the lack of adequate individual data. This paper tries to fill this gap by, a) summarizing previous theoretical arguments underlying the overlap found in the criminological and economic literature, and, b) by analyzing a unique survey data set, originally collected as control group to the German Inmate Survey. Bivariate Probit results show that both criminal and victimization risks are partly driven by common observed and unobserved factors. Moreover, exogeneity tests are in line with the hypothesis that the joint bivariate process can be characterized by exogenous variation of offending and an endogenous role of victimization.