Unemployment and Crime: Is There a Connection?

Abstract

A panel of Swedish counties over the years 1988-1999 is used to study the effects of unemployment on property crime rates. The period under study is characterized by turbulence in the labor market-the variation in unemployment rates was unprecedented in the latter part of the century. Hence, the data provide a unique opportunity to examine unemployment effects. According to the theory of economics of crime, increased unemployment rates lead to higher property crime rates. A fixed-effects model is estimated to investigate this hypothesis. The model includes time-and county-specific effects and a number of economic and socio-demographic variables to control for unobservables and covariates. The results show that unemployment had a positive and significant effect on some property crimes (burglary, car theft and bike theft).

Common Trends in the US State-Level Crime. What Do Panel Data Say?

“…Our crime model is inspired by the principles in Becker (1968) and follows the formulation in Freeman (1999), Edmark (2005), and Wu and Wu (2012). 4 The model describes the choice of the individuals between work and crime, as their source of income during one period.…”

“…Accordingly, we indicate with W the wage from honest work, whereas W b is the income from illegal activities. Like Edmark (2005), the presence of a idiosyncratic psychological cost (c) of committing a crime is also considered. This cost, that can be positive or negative, is assumed to be independent and continuously distributed over the population.…”

“…c) is lower than the quantity on the right-hand side. Moreover, it helps to elicit the e↵ect of the parameters of the model on the supply of crime at more aggregate level (see Edmark, 2005). In fact, the higher is the right-hand side of (4), the higher is the probability for individuals to commit crimes, with an impact on the aggregate supply of crime.…”

“…Hypothesis 1 (Edmark, 2005): Assume that W > A and u < 1 (both realistic assumptions). This implies that the right-hand side of Eq.…”

“…We first test for unit root in the data using the PANIC approach by Bai and Ng (2004), and then we check for cointegration using the Durbin-Hausman type panel co-integration by Westerlund (2008). Lastly, we estimate the crime equation, which we derive from the theoretical approach by Edmark (2005) and Wu and Wu (2012), using the continuously-updated (CUP) estimator developed by Bai and Kao (2006), that controls for serial correlation and endogeneity. We use several measures of crime rates (property, violent, robbery, burglary, larceny and auto theft), inequality (top 10% and top 5% of income earners, and Gini index), and two measures of crime prevention (prison admissions per crime and state expenditure in police defense).…”