John Hamman

Solving Coordination Failure with All-or-None Group-Level Incentives

Coordinating activity among members is an important problem faced by organizations. When firms, or units within firms, are stuck in bad equilibria, managers may turn to the temporary use of simple incentives - flat punishments or rewards - in an attempt to transition the firm or unit to a more efficient equilibrium. We investigate the use of incentives in the context of the minimum-effort, or weak-link, coordination game. We allow groups to reach the inefficient equilibrium and then implement temporary, flat, all-or-none incentives to encourage coordination on more efficient equilibria. We vary whether the incentives are positive (rewards) or negative (penalties), whether they have substantial or nominal monetary value, and whether they are targeted to a specific outcome (the efficient equilibrium) or untargeted (and apply to more than one outcome). Overall, incentives of all kinds are effective at improving coordination while they are in place, and we find that substantial and targeted incentives are most effective. However, there is little long-term persistent benefit of incentives - once the incentives are removed, groups tend to return to the inefficient outcome.

Self-Interest Through Agency: An Alternative Rationale for the Principal-Agent Relationship

Principal-agent relationships are typically motivated by efficiency gains from comparative advantage. However, such delegation may also arise because it allows principals the pursuit of selfish outcomes while avoiding explicitly selfish behavior. We report laboratory experiments in which principals repeatedly either decide how much money to share with a recipient or select agents to make sharing decisions on their behalf. Across several treatments, recipients receive significantly less when agents make allocation decisions. This results from principals seeking those agents willing to reliably share the least. We observe instances in which sharing is almost entirely extinguished when decisions are made through agents.

Fool Me Once: An Experiment on Credibility and Leadership

We study a situation in which a leader tries to induce followers to invest in a joint venture. This venture has three key properties: (1) returns are uncertain, (2) investments are complements, so investment is generally only profitable when the followers coordinate on investing, and (3) investment is often more beneficial for the leader than the followers. The leader has private information about the investment’s return and can facilitate successful coordination through cheap-talk recommendation. We study the extent to which leaders in such situations possess “social credibility�? defined as the ability to coordinate followers’ behaviors with their statements. In an experiment, we find that leaders manage social credibility by forgoing potentially profitable requests for investment in order to make it more likely that subsequent recommendations to invest are followed. We identify factors that affect the persistence of social credibility.