Overconfidence, Underconfidence, and Welfare

Using a simple framework due to Cooper and John (1988) and Cooper (1999), this paper derives the conditions under which overconfidence and underconfidence of agents lead to Pareto improvement. We show that an agent's overconfidence in a game exhibiting strategic complementarity and positive spillovers and an agent's underconfidence in a game exhibiting strategic complementarity and negative spillovers can lead to Pareto improvement.