Evolution, Coordination, and Banking Panics

Abstract

I study equilibrium selection by an evolutionary process in an environment with multiple equilibria, one of which involves a banking panic. The analysis is built on a repeated version of the Diamond-Dybvig (1983) model. The optimal (run free) equilibrium is uniquely selected if it is also "risk dominant." Furthermore, the probability of observing a panic increases as the size of the banks decreases. I discuss local interaction and contagion effects that allow for a bank run to spread first among banks in the same geographic location and then throughout the entire population.

JEL Classification: G21, C73