We investigate a rural household's decision to migrate part or all of its labor to urban areas. Labor migrates only when the expected return passes a hurdle rate that is affected by migration networks. We develop a dynamic model of incomplete information, and characterize the unique pure strategy perfect Bayesian equilibrium. Based on the predictions of the equilibrium, we develop a set of econometric models to estimate the migration decisions. Our results show that more current information helps migration, while the expectation of more information in the future may higher migration.