Re-evaluating the Cantor and Land (1985) model of unemployment and crime: A multilevel analysis of multiple economic measures

The complex relationship between crime and economic change has had a long pedigree in criminological research. Much of the recent research is premised on, or critical of the Cantor and Land (1985) model of unemployment and crime, which considers the unemployment rate as representative of the state of the economy. The theoretical assumption behind the model rests on the notion that to accurately assess the unemployment-crime relationship, the impact of criminal motivation and criminal opportunity need be considered in a common framework. Accordingly, Cantor and Land (1985) developed a structural approach that synthesized the counteracting effects of motivation and opportunity into a single working model, finding that opportunity dominates motivation. Although the theory behind the empirical model is not often questioned, both methodological and empirical concerns have arisen with regard to the procedures employed by Cantor and Land (1985) and subsequent studies that rely on the Cantor and Land approach. Methodologically, researchers have questioned whether flaws in specification have contributed to mixed and inconsistent results. Empirically, testing the model using unemployment as the sole indicator of economic performance has been widely contested in the literature. This paper considers both issues as the Cantor and Land (1985) model will be evaluated using distinguished measures of unemployment and a multilevel methodological approach, with the Canadian provinces from 1981-2013 as the units of analysis. The inclusion of multiple economic measures, in addition to the comparative utility allows for a comprehensive representation of economic performance. Furthermore, multilevel regression enables greater precision in the estimates, allowing researchers to accurately analyze hierarchical data at two distinct levels. Overall, by extending the seminal work of Cantor and Land (1985) the intent is to bridge the empirical gaps in the crime-economy literature and, in doing so, provide an instructive example for the operationalization of the model in future studies. In examining the effect of multiple economic measures on eight separate crime types, support was found for the Cantor and Land (1985) model in both property and violent crimes. The results are robust to the inclusion of time as a random effect, controls for simultaneity using contextual, deterrent and variables controlling for inequality and demographics.