The Effect of Arrests on Crime: A Multivariate Panel Analysis

We estimate multiwave panel models for the effect of clearance rates and a vector of socioeconomic control variables on index crimes, using a sample of 98 U.S. cities for the years 1964-70. No consistent evidence of a substantial effect is found. The effect of police practices on rates of law violation is a matter that has a practical bearing on crime prevention strategies. It can also enhance our understanding of the contributions made by formal methods of social control to conformity. Despite the importance of these concerns, surprisingly little is known about the effects of criminal justice institutions on crime rates. Indeed, only in recent years have researchers begun to study the degree to which crime can be reduced by marginal changes in police employment, expenditure, patrolling strategies, or efficiency in solving crimes. I Statistical investigations of the relationship between aggregated rates of crimes known to the police and various indicators of police activity have employed various analytic strategies. Some researchers have analyzed cross-sectional data (Brown; Geerken and Gove; Sjoquist; Wilson and Boland, a, b); others have analyzed time series (Cloninger and Sartorius; Phillips and Votey) or panel data (Greenberg et al., a; Logan). The present analysis has been designed to meet methodological criticism of this body of research. Critics of research dealing with the effect of law enforcement on