Metropolitan fragmentation, law enforcement effort and urban crime

Abstract

This paper investigates how a system of local law enforcement agencies operates within a metropolitan area. In a cross section of 236 US MSA the paper finds that greater agency fragmentation leads to less law enforcement effort (resources), but also to less crime! This seemingly contradictory result is robust to many alternative specifications. To explain the result a model is developed in which a unilateral increase in local law enforcement effort has the effect of “spatially displacing” criminals as well as incarcerating them. In this model, greater agency fragmentation by itself leads to lower spending but higher crime. The model is expanded to include an “X-efficiency” advantage by smaller agencies, which is shown to be necessary to explain the empirical results. The exact source of this advantage however, is not uncovered.