Revisiting the Crime-Unemployment Nexus

Abstract

While the potential for bi-directionality in the crime-unemployment relationship has been noted as a cause for concern by many researchers, no prior studies have empirically established the presence of bi-directionality. To address such concerns, we jointly estimate both directions of the crime-unemployment relationship using dynamic panel methods in a simultaneous equations framework in conjunction with a 26-year, county-level panel from California. We find evidence that not only does higher unemployment lead to higher crime as is well known, but that crime may also affect unemployment with higher levels of offending leading to unemployment reductions.