Employment, Dynamic Deterrence and Crime

Using monthly panel data we solve and estimate, using maximum likelihood techniques, an explicitly dynamic model of criminal behavior where current criminal activity adversely affects future employment outcomes. This acts as "dynamic deterrence" to crime: the threat of future adverse effects on employment payoffs when caught committing crimes reduces the incentive to commit them. We show that this dynamic deterrence effect is strong in the data. Hence, policies which weaken dynamic deterrence will be less effective in fighting crime. This suggests that prevention is more powerful than redemption since the latter weakens dynamic deterrence as anticipated future redemption allows criminals to look forward to negating the consequences of their crimes. Static models of criminal behavior neglect this and hence sole reliance on them can result in misleading policy analysis