Returns from Cooperation in Criminal Activity: Estimating Crime Production Function and Returns to Scale

This study asks a simple question: how the damage produced in a criminal activity depends on the number of offenders cooperating? If offenders maximize damage as a proxy for utility of earnings acquired from given crime and they share the damage from criminal opportunity equally, we would expect that the total damage production function exhibits increasing or constant returns to scale. Using the rich-in-characteristics dataset of the Police of the Czech Republic covering the years 1995-2016, linear regression models of total damage, damage per person and probability of accusation are estimated while controlling for heterogeneity among criminal activities. The results reject both the possibility of non-decreasing returns to scale of crime production function and the possibility of compensation of decreasing damage per person by decreasing probability of accusation. Likely explanation is that the criminal opportunities for a single offender give the highest earnings because the earnings in pairs and groups are not shared equally; the profitability of these single-offender activities is then compensated by their skills requirements or generally their unavailability - otherwise everyone would prefer to work alone. These results are highly consistent with Levitt and Venkatesh (2000) who have also found that the criminals working alone are enjoying highest earnings comparing to those working in groups (gangs). This study also finds significant stylized facts and long-run trends of Czech criminals - their earnings are decreasing through years along with crime counts; they are also becoming older and lonelier - through their criminal career they tend to cooperate increasingly less.