General Equilibrium and Dynamic Inconsistency

We study the role of expectations of naive agents in a general equilibrium version of the Ramsey model with quasi-hyperbolic discounting. When agents recognize others' naivete, as strongly suggested by empirical evidence, they revise consumption paths, correctly anticipating prices in a resulting sliding equilibrium (perfect foresight). When agents are unaware of others' naivete, as is typically assumed in the literature, they revise both consumption paths and price expectations (quasi-perfect foresight). We prove the existence of sliding equilibrium under perfect foresight for the class of isoelastic utility functions. We show that generically quasi-hyperbolic discounting matters for saving behavior: sliding equilibrium under perfect foresight is observationally equivalent to some optimal path in the standard Ramsey model if and only if utility is logarithmic. We compare sliding equilibria under different types of foresight and show that perfect foresight implies a higher saving rate, long-run capital stock, and consumption level than quasi-perfect foresight.