This dissertation addresses how to assess police performance and how to use performance assessments to strengthen incentives for the police. First, an econometric analysis provides evidence of negative effects of police on property crime, violent crime and public disorder. The police funding formula is modeled to identify the endogenous variation in police levels. The remaining variation is used to identify police effectiveness. By using victimization data, this study provides evidence on the effects of police on public disorder rather than crime only, circumvents measurement error common to police statistics, and controls for both individual and municipality characteristics. Second, an analysis of incentive design suggests a tradeoff between two alternative approaches to holding the police accountable to results. A performance reward can be based on a definition of “good performance” ex ante. Alternatively, subjective performance assessment can be used to close the gap between performance data and judgment of performance ex post. Given the multidimensional nature of police work and the prevalence of non-discrete outcomes, rule-based assessment proves to be difficult. The government faces an unfavorable tradeoff between power of incentives versus distortion of effort and discretionary power. Subjective assessment alleviates distortion of effort and leaves more room to respond to (changing) regional conditions.
