A comparison of simple action-based and outcome-based policies for emergency-like situations

This paper explores a class of social dilemmas in which the participation of a given number of individuals is required to achieve a social end in the absence of a coordinating authority (e.g., rescuing a person in peril or preventing an imminent crime). After describing the first- and second-best outcomes, we examine whether simple policy instruments such as punishments and rewards can induce the second-best outcome, distinguishing between policies based on an individual's actions (i.e., action-based policies) and policies based on the outcome (i.e., outcome-based policies). For the domain of simple policies considered, we establish that action-based policies enjoy a crucial advantage over outcome-based ones: namely, outcome-based policies always feature an equilibrium with no participation, whereas action-based policies exclude this equilibrium.