This paper develops a theory of endogenously (non-)Ricardian beliefs. The foundation of the theory builds on the insights in Eusepi and Preston (2018, Fiscal Foundations of Inflation: Imperfect Knowledge, American Economic Review, forthcoming), who show that when the private sector has imperfect knowledge about long-run monetary and fiscal policy, in  particular whether the path of future taxes suffices to equate the government’s intertemporal budget constraint, Ricardian equivalence can fail even though the policy regime is  Ricardian.  The novelty here is a restricted perceptions viewpoint: agents, when facing a complex forecasting problem, prefer to forecast aggregate state variables with parsimonious models that are optimal within the restricted class, i.e., a restricted perceptions equilibrium (RPE).  A misspecification equilibrium is a refinement of an RPE where the choice of restricted model is endogenous.  We show that (1.) there can exist misspecification equilibria where beliefs are endogenously (non-)Ricardian, (2.) there are cases where multiple  equilibria exist, where the economy can coordinate on either a Ricardian or non-Ricardian equilibrium.  This latter result suggests an alternative interpretation of U.S. inflation data as  generated by endogenous belief-driven regime change.   We explore this possibility quantitatively and provide evidence of time-varying (non-)Ricardian beliefs.  Several counterfactual exercises illustrate a novel and nuanced trade-off in designing monetary policy rules.

Joint work with William Branch