Contracting Outcomes with Communication and Learning

We show that allowing communication can increase optimal choices and efficiency in a multi-task, incomplete contracting, principal-agent setting. We study two simple communication protocols, one allowing for one or more requests on non-contractible choices, and the other allowing for a request, promise and ex post payment. The protocol where principals are asked to communicate requests to the agent regarding non-contractible choices promotes better learning of optimal strategies on the part of the principals but shows no tendency for coordination to superior outcomes. The benefits accrue mainly due to changes in the choices of principals, who issue communication, rather than those of agents. Coordination is promoted, and learning subdued, when the protocol permits promises and ex post payments, in addition to allowing a request. This protocol also increases efficiency, with the efficiency gains equal across the protocols