UNEMPLOYMENT,CRIME, AND IMPRISONMENT: A PANEL APPROACH*

Despite the widespreach acceptance of the notion that during periods of economic downturn, higher levels of unemployment lead to higher levels of crime and imprisonment, the research literature reveals very little consistent support for the existence of such a relationship. Studies that do suggest unemployment causes crime and imprisonment have used methodological techniques which, especially in over-time studies, could lead to the acceptance of spurious effects caused by trend and lag as evidence of a true relationship. Using data over a six-year period for the United States, a panel model is specified in which appropriate controls for intra-series trend and cross-series lagged effects are included. Although bivariate correlations are strongly suggestive of a relationship between unemployment and crime, results of the panel approach suggest that most of the apparent relationship is due to common trend effects. Little evidence is found for the existence of a relationship between crime/imprisonment and unemployment, regardless of the type of effect considered. In addition, results from stability tests indicate that the crime-unemployment relationship has been unstable. while the unemployment-imprisonment relationship has been relatively invariant.