Efficient incentives in social networks : "gamification" and the coase theorem

This study explores mechanism design for networks of interpersonal relationships. Agents' social (i.e., altruistic or spiteful) preferences and private payoffs are all subject to asymmetric information; utility is (quasi-)linear, types are independent. I show that any network of at least three agents can resolve any allocation problem with a mechanism that is Bayesian incentive-compatible, ex-interim individually rational, and ex-post Pareto-efficient (also ex-post budget-balanced). By contrast, a generalized Myerson-Satterthwaite theorem is established for two agents. The central tool to exploit the asymmetry of information about agents' social preferences is "gamification": Resolve the agents' allocation problem with an efficient social-preference robust mechanism; ensure agents' participation with the help of a mediator, some network member, who complements that mechanism with an unrelated hawk-dove like game between the others, a game that effectively rewards (sanctions) strong (poor) cooperation at the expense (to the benefit) of the mediator. Ex interim, agents (and the mediator) desire this game to be played, for it provides them with a platform to live out their propensities to cooperate or compete. - A figurative example is a fund-raiser, hosted by the "mediator", complemented with awarding the best-dressed guest.