Reputation and Efficiency in Social Interactions: An Example of Network Effects

Abstract

Reputations emerge if an actor's future partners are informed on his present behavior. Reputations depend on the "embeddedness" of interactions in structures or networks of social relations. They illustrate the effects of such embeddedness on the outcomes of interactions.This article presents simple game-theoretic models of reputation effects on efficiency (in the Pareto sense) in interactions. In a comparative perspective, the authors start with a baseline model of a social system in which reputation effects (of a specific kind) are excluded: actors do not receive information on their partners' behavior in interactions with third parties. Such a system of "atomized interactions" is compared to a system with interactions that are "perfectly embedded": actors are immediately informed on all interactions of their partners with third parties.Efficiency is more easily attained as a result of individually rational behavior in perfectly embedded systems. In a final step, the comparative perspective is broadened, and the extreme assumptions of either an atomized or a perfectly embedded social system replaced. Intermediate cases arise in the consideration of "imperfect embededness," that is, a situation in which actors are informed only after some time lag on the behavior of their partners vis-a-vis third parties. It is shown that the conditions for efficiency become more restrictive as the information time lag lengthens.