We consider a model of evolution and local interactions in a circular city as in Ellison (1993) [Ellison, G., 1993. Learning, Local Interaction, and Coordination. Econometrica 61, 1047-1071], where agents follow imitation rules rather than myopic best-response. If interaction is limited to the two immediate neighbors, in the short run the process has a large number of absorbing states where both equilibria coexist, but risk-dominant equilibria are selected in the long run. Contrary to Ellison's model, though, payoff-efficient equilibria, and not risk-dominant ones, might be uniquely selected in the long run when players interact with neighbors further away.
- Risk dominance
Available at: http://works.bepress.com/alos/10/