Endogenous Barriers to Learning

Building on the idea that lack of experience is a source of errors but that experience should reduce them, we model agents' behavior using a stochastic choice model (logit quantal response), leaving endogenous the accuracy of their choices. In some games, higher accuracy leads to unstable logit-response dynamics. Starting from the lowest possible accuracy, we define the barrier to learning as the maximum accuracy which keeps the logit-response dynamic stable (for all lower accuracies). This defines a limit quantal response equilibrium. We apply the concept to centipede, travelers' dilemma, and 11-20 money-request games and to first-price and all-pay auctions, and discuss the role of strategy restrictions in reducing or amplifying barriers to learning.