Coordination via delay: Theory and experiment

Abstract

This paper studies the effect of introducing an option of delay in coordination games—that is, of allowing players to wait and then choose between the risk-dominant and payoff-dominant actions. The delay option enables forward-induction reasoning to operate, whereby a player's waiting and not choosing the risk-dominant action right away signals an intention to choose the payoff-dominant action later. If players have ϵ-social preferences—they help others if they can do so at no cost to themselves—then iterated weak dominance yields a unique outcome in which everyone waits and then chooses the payoff-dominant action if everyone else waited. Thus, efficient coordination results. Experimental evidence from a binary-action minimum-effort game confirms that adding a delay option can significantly increase the occurrence of efficient outcomes. Moreover, consistent with our theory, the clear majority of subjects in our experiment take the unique iteratedly undominated strategy and not other strategies that are implied by equilibrium analysis.