Leading by example? Investment decisions in a mixed sequential-simultaneous public bad experiment

This paper investigates the effect of having a leader in a laboratory public bad experiment with five subjects in each group. The control treatment is a standard public bad experiment, while in the leader treatments the design is such that in each group the leader decides first on his or her investment in the public bad. After being informed about the leader’s decision, the four ‘followers’ in each group make their investment decision. Two treatments of the leadership game are played with each group. In the same-leader-costs treatment, all subjects are confronted with the same costs, while in the no-leader-costs treatment the leader faces no direct costs of acting socially. It is found that followers invest significantly less in the public bad when there is a leader compared with a situation when there is no leader. Comparing the two treatments, we find, moreover, that the leadership effect is somewhat stronger when leaders face the same costs as followers compared with the situation in which leaders bear no costs. Randomly chosen leaders set an example by investing less than average players in a standard public bad game, and leader investments are lowest when the costs of leading are low.