Rethinking the crime reducing effect of education? Mechanisms and evidence from regional divides

We estimate the contemporaneous effect of education on adolescent crime by exploiting the variation in crime rates between different cohorts and at different ages that followed a reform that raised the school-leaving age in Italy. A 1 percentage-point increase of the enrollment rate reduces adolescent crime by 1.3 per cent in the North of Italy but increases it by 3.9 per cent in the South. The crime-reducing effect depends mainly on incapacitation (i.e. adolescents stay in school instead of on the street); the crime-increasing effect is consistent with a channel of criminal capital accumulation, operating through social interactions and organized-crime networks.