Job Losses and Criminal Gains: Analyzing the Effect of Unemployment on Criminal Activity

In this paper I analyze the relationship between unemployment and criminal activity. Using a unique panel data set on German counties covering the years 2003 to 2009, I estimate the effect of unemployment on crime rates for different kind of offenses. I extensively control for potential confounders as economic and demographic factors as well as time and county fixed effects. To circumvent the endogeneity of the unemployment rate in the structural equation of interest I interact two sources of variation which are exogenous to the change in crime within each county to construct an instrumental variable for unemployment. In addition to mean regression I also estimate quantile regressions in order to shed more light on the complex relationship between unemployment and criminal activity. The results based on mean regression show insignificant effects in the most general specification, while the results based on quantile regression are mixed.