More crime in cities? On the scaling laws of crime and the inadequacy of per capita rankings – a cross-country study
share
Objectives: To evaluate the relationship between population size and number of crimes in cities across twelve countries and assess the impact of per capita measurements on crime analyses, depending on offense type. Methods: We use data on burglaries and thefts at the city level and evaluate the relationship between crime numbers and population size using probabilistic scaling analysis. We estimate the growth exponent of each offense type and use Kendall rank correlation to assess the impact of a linear growth assumption (i.e., per-capita analysis) on cities rankings. Result: In nine out of eleven countries, theft increases superlinearly with population size; in two of them, it increases linearly. In eight out of ten countries, burglary increases linearly with population size; in two of them, it increases superlinearly. In nonlinear scenarios, using per capita rates to rank cities produces substantially different rankings from rankings adjusted for population size. Conclusions: Comparing cities using per capita crime rates (e.g., crime per 100,000 people per year) assumes that crime increases linearly with population size. Our findings indicate, however, that this assumption is unfounded, implying that one should be cautious when using per capita rankings. When crime increases nonlinearly with population, per capita rates do not remove population effects. The contrasting crime growth of burglary and theft also suggests that different crime dynamics at the local level lead to different macro-level features in cities.
READ FULL TEXT