A Comment on 'Bayesian Learning in Social Networks'

Abstract

Gale and Kariv (2003) study Bayesian learning in social networks and the resulting properties of long run behavior of agents. This note points out a mistake in the proof of Corollary 1 and a counter-example to the statements of Corollary 1 and Corollary 2. The proof of the main result, Theorem 2, makes use of Corollary 2. Hence Theorem 2 remains unproven. However, the recent papers Rosenberg, Solan, and Vieille (2009) and Mueller-Frank (201''7 provide results that are very similar to Theorem 2 considering more general settings.