Dynamic Network Formation with Foresighted Agents

What networks can form and persist when agents are self-interested? Can such networks be efficient? A substantial theoretical literature predicts that the only networks that can form and persist must have very special shapes and that such networks cannot be efficient, but these predictions are in stark contrast to empirical findings. In this paper, we present a new model of network formation. In contrast to the existing literature, our model is dynamic (rather than static), we model agents as foresighted (rather than myopic) and we allow for the possibility that agents are heterogeneous (rather than homogeneous). We show that a very wide variety of networks can form and persist; in particular, efficient networks can form and persist if they provide every agent a strictly positive payoff. For the widely-studied connections model, we provide a full characterization of the set of efficient networks that can form and persist. Our predictions are consistent with empirical findings.