This paper analyzes the determinants of public law enforcement policies when citizens vote for the timing and level of fines. We consider situations where citizens and politicians disagree on the value of the expected social harm associated with some activities. We find that citizens vote for act-based (harm-based) sanctions when they expect that social harm is low (resp high). On the other hand, we show that the equilibrium fines may be higher or lower than the optimal one, depending on the difference between politicians and citizens' expectations.