This paper investigates the relationships among criminal activity, police production and the allocation of police resources in a model using cross section data from the main Spanish cities in 1995. The purpose is to test a set of hypotheses concerning both the rational behaviour of potential offenders and the determinants of police activity. A three-equation simultaneous system is specified and estimated to link together both the crime and clearance rates and police manpower levels. The findings of the study provide empirical support for a number of the postulates of the economic theory of crime. Thus, the crime rate equation shows that the risk of arrest, educational attainment, the long-term unemployment rate and per cápita income all play an important role in explaining variations in crime. For its part, the police output equation indicates that the clearance rates depends mainly on non-controllable factors such as the population or the levels.of violent crime. Finally, the police resource allocation function confirms that the number of crimes has a strong positive effect on determining police presence.

Key words: economics of crime, police service, crime control, criminal justice system, simultaneous equation models.
JEL classification: C31, H56, K42.