“Partial Monitoring and Message Trading” by Tao ZHU
Hong Kong University of Science and Technology
Federal Communications Commission
In a random-matching risk-sharing model, the role of public messages is explored when in each pairwise meeting, risk-sharing actions are only monitored by the pair in the meeting (partial monitoring). A risk-sharing outcome and the message about the outcome are determined simultaneously, allowing the message and outcome to be traded with each other (message trading). The folk theorem holds if agents can commit to not renegotiate a Pareto-dominated trade. Absent of such commitment, the folk theorem holds in a special case but for many utility functions requires a higher greatest lower bound on the discount factor than the commitment counterpart; in general the folk theorem does not hold, with the loss not vanishing as the discount factor approaches unity, but there is some continuity from special to general cases. Message trading is related to monetary exchange where money generates messages that need not be public and have more restricted information content.