Platform selection in the lab

Emerging literature explores experimental platform selection games. These games converge rapidly on the superior platform under a wide range of conditions. We replicate the remarkable results of Hossain and Morgan (2009) in which such a game tips almost perfectly to the superior platform. Next, we show that platform coordination fails when seemingly innocent increases in out-of-equilibrium payoffs are introduced. The inflated payoffs keep the best reply structure unchanged and do not influence players’ security levels. Our design allows control for the explanatory force of risk dominance. We find that equilibrium selection theory is unable to account for coordination failure while observed behavior is consistent with non-rational learning. Furthermore, and contrary to the literature, we find that efficiency suffers when an inferior platform is granted initial monopoly.