Averting economic collapse and the solipsism bias

We study the behavior of experimental subjects who have to make a sequence of risky investment decisions in the presence of network externalities. Subjects follow a simple heuristic-investing after positive experiences and reducing their propensity to invest after a failure. This result contrasts with the theoretical findings of Jeitschko and Taylor [Jeitschko, T.D., Taylor, C., 2001. Local discouragement and global collapse: A theory of coordination avalanches. Amer. Econ. Rev. 91 (1), 208-224] in which even agents who have only good experiences eventually stop investing because they account for the fact that others with worse experiences will quit. This can trigger sudden economic collapse-a coordination avalanche-even in the most efficient Bayesian equilibrium. In the experiment, subjects follow their own experiences and disregard the possible bad experiences of others-thus exhibiting behavior that we term “solipsism bias.” Solipsism results in sustained investment activity and thus averts complete collapse.