Causal estimation of the determinants of crime is a key focus for policymakers – for example, Arnold and Carnes (2012) suggest that crime affects the approval rates of mayors. Understanding whether incentives drive crime is also a central concern for economists, as evidenced by the central role of incentives in seminal economic theories of crime, such as Becker’s (1968). Despite a substantial body of literature which finds that unemployment impacts crime at the aggregate level, there remains disagreement on whether job loss causes criminal activity.
Becker (1968) argues that a core driver of criminal behaviour is an individual comparison of the benefits and costs of committing crime. Area-wide estimates of the impact of unemployment rates on aggregate crime rates capture direct effects of individual unemployment on crime as well as spillover effects; an important challenge is to identify what, if any, of this aggregate relationship between unemployment and crime rates is due to the direct impact of individual unemployment on crime. We estimate the impact of individual job separations on the probability of committing crime and detail both the type of individuals and environmental factors which determine whether job loss leads to increases in crime.
New evidence on unemployment and crime
There is substantial evidence from the unemployment and crime literature which documents significant and modest impacts of unemployment on total and property crimes in the US (Gould et al. 2002, Raphael & Winter-Ebmer 2001), Sweden (Öster & Agell 2007), France (Fougère et al. 2009), and for right-wing extremist crime in Germany (Falk et al. 2011). These studies instrument unemployment rates to identify a causal impact of unemployment on crime. Similarly, previous studies have also documented a relationship between other economic conditions and crime, such as wages (Grogger 1998, Machin & Meghir 2004), time spent in unemployment (Bindler 2015), and the impact of graduating during a recession (Bell et al 2014). While this extensive literature documents a relationship between unemployment (and other economic conditions) and crime, a common feature of such studies is that they relate aggregate (i.e. mostly county-level) unemployment and crime. Given the range of estimates, Levitt (2004) suggests that the state of the economy cannot explain the evolution of crime rates during the 1990s.
In a recent paper, we use individual level records to provide new evidence on the impact of job displacement on crime (Bennett and Ouazad 2016). Using administrative, population-wide Danish Register data, we construct a matched employer-employee data set of prime-aged Danish males from 1985-2000 and match individuals to their police records – arrests or citations, charges, and convictions.
The richness of the data is of great help, but simply correlating individual unemployment with individual crime does not answer the causal question. Rather, we focus on job displacement – that is, job loss that occurs during sudden and unexpected mass-layoffs of high tenured employees – as an arguably exogenous source of job separations to estimate the impact of unemployment on earnings (Jacobson et al. 1993), health (Sullivan & von Wachter 2009, Black et al. 2012) and children outcomes (Rege et al. 2011).
Job displacement provides a credibly exogenous source of job separation, overcoming the endogeneity of job separations due to individuals selectively choosing to quit or change jobs. The criminal behaviour of displaced and non-displaced individuals both after job loss is informative about the causal impact of job loss on crime, as the pre-displacement crime records of future displaced individuals does not differ systematically from those who will not be displaced.
Our baseline estimates, depicted in Figure 1 below, suggest a strong relationship between job displacement and crime. Displaced individuals are significantly more likely to commit a crime which leads to a conviction post-displacement, but not in the time leading up to displacement. This impact is predominantly driven by an impact of job displacement on property crimes, which increase by about 26% of the population-wide average, and is economically and statistically significant up to seven years following displacement. Consistent with literature on education and crime, the impact on crime is driven by those with low education – high school or less and, to a lesser extent, vocational education – where those with university education experience no increases in crime post-displacement.
Figure 1 Impact of displacement on the probability to commit crime (total), property crime, driving under influence (DUI), and violent crime
Note: The vertical axis is in percentages relative to the pre-displacement year. The probability to commit property crime increases by 0.4 percentage points in the year of displacement.
Recent work in the US by the Council of Economic Advisers (2016) suggests that high rates and lengths of incarceration may have large opportunity costs; indeed, the same paper suggests that investments in policies to improve labour market opportunities are likely to be more beneficial than policies that lead to greater incarceration. Consistent with this, we find that earnings losses of displaced individuals convicted to prison are well below the predicted earnings losses for these individuals, and well below earnings losses of displaced individuals whose conviction does not lead to incarceration. This suggests that prison terms are either correlated with unobservable traits that cause lower earnings or that prison terms have a direct impact over and above the impact predicted by incapacitation, the inability of the individual to be in employment.
Impacts on family dynamics
Matching detailed family information to displaced individuals, we analyse how displacement events interact within an individual’s family. We show that the impact of job displacement on crime is around three times higher in magnitude for single male adults, a finding consistent with the hypothesis that intra-family income pooling may offer partial consumption insurance post-displacement. We find that displacement has a large and long-lasting impact on the probability of being married, but there is little evidence that the timing of marriage dissolution coincides with the timing of crime. By linking displaced fathers to their younger family members, we estimate the impact of parental job displacement on the criminal behaviour of younger household members, finding some potential but small impacts of father’s displacement on son’s crime.
Local income distribution and displacement
Existing studies have analysed the relationship between inequality and crime as well as poverty and crime (Kelly 2000, Choe 2008). We combine our identification strategy using displaced workers with local measures of income inequality and poverty concentration to assess whether the impact of job displacement on crime is larger or smaller in areas of inequality or poverty. Using income data, we construct measures of local inequality (measured by gini coefficients) and poverty (defined as the fraction of families living below 50% of the national median equivalent family income) for each Danish municipality from 1985-2000. We find that greater inequality and poverty amplify the impact of job displacement on property crime immediately following displacement, where the impact of displacement on crime in a municipality in the upper quartile of the inequality distribution is twice as large as the impact in a municipality in the lower quartile of the inequality distribution.
Our work finds evidence of an impact of job separations on crime. Post-separation crime impacts society over and above what affects the employer and the employee. Such results in a country with relatively low crime rates and, in the period of analysis, with generous unemployment benefits, suggest substantial overall impacts of job loss on crime.
Becker, G S (1968) “Crime and punishment: An economic approach”, The Journal of Political Economy, 76(2): 169–217.
Bennett, P and A Ouazad (2016) “Job displacement and crime: Evidence from Danish microdata”, Copenhagen Business School, Department of Economics, Working Paper.
Council of Economic Advisers (2016) Economic perspectives on incarceration and the criminal justice system, The White House.
Gould, E D, B A Weinberg and D B Mustard (2002) “Crime rates and local labor market opportunities in the United States: 1979–1997”, Review of Economics and Statistics, 84(1): 45–61.
Jacobson, L S, R J LaLonde and D G Sullivan (1993) “Earnings losses of displaced workers”, The American Economic Review, 83(4): 685–709.
Kelly, M (2000) “Inequality and crime”, Review of Economics and Statistics, 82(4): 530–539.
Levitt, S D (2004) “Understanding why crime fell in the 1990s: Four factors that explain the decline and six that do not”, Journal of Economic Perspectives, 18(1): 163–190.
Raphael, S and R Winter-Ebmer (2001) “Identifying the effect of unemployment on crime”, Journal of Law and Economics, 44(1): 259–283