Other-regarding behavior under collective action

In many collective action settings, such as decisions on public education or climate change mitigation, actions of a group have welfare consequences for themselves as well as their followers. We conduct laboratory experiments with two-stage predecessor-follower prisoners' dilemma and coordination games with dynamic externalities to study whether concerns for the followers' welfare affect the predecessors' behavior. We find that predecessors often give up own payoffs to avoid imposing negative externalities on the followers, but not to generate positive externalities for the followers. A concern for the followers aligned with own group payoff maximization motive helps to resolve social dilemma and coordination problems; yet, a conflict in motives greatly exacerbates both free-riding and coordination on the payoff-inferior equilibrium. We also find strong evidence of social learning: the followers tend to blindly mimic their own predecessor, but act opposite to their match's predecessor, no matter whether these actions are welfare-improving or not.