Strategic Information Acquisition in Networked Groups with “Informational Spillovers”

This paper develops a model of costly information acquisition by agents who are connected through a network. For a exogenously given network, each agent decides first on information acquisition from his neighbors and then, after processing the information acquired, takes an action. Each agent is concerned about the extent to which other agents align their actions with the underlying state. A new equilibrium notion, which is in the spirit of perfect Bayesian equilibrium, is proposed to analyze information acquisition decisions within networked groups. This equilibrium notion allows each agent to compute, when deciding about information acquisition, the extent to which changes in his information acquisition decision will affect his own perception of future expected payoffs. Agents anticipate and incorporate such changes in their information acquisition decisions. Both the efficient and the equilibrium information acquisition profiles are characterized and the compatibility between them is related to the density of the network.