Implications of uncertainty for optimal policies

We study the implications of ambiguity for optimal fiscal policy in macro public finance environments with heterogeneous agents and private idiosyncratic shocks. We describe conditions under which ambiguity implies that it is optimal to periodically reform policies. Periodic reforms lead to simplified optimal policies that are not fully contingent on future shocks; at times they also lose dependence on the full history of past shocks. These simplified policies can be characterized without complete backward induction when the time horizon is finite. However, linear policies can be far from optimal. We also show that equilibria in decentralized versions of these economies are not generally efficient, implying a meaningful role for government provision of insurance, unlike in conventional environments with a narrower view of uncertainty.