We study testable implications for the dynamics of consumption and income of models in which first best allocations are not achieved because of a moral hazard problem with hidden saving. We show that in this environment agents typically achieve more insurance than that obtained under self...