We introduce behavioral learning equilibria (BLE) into DSGE models with boundedly rational agents using simple but optimal first order autoregressive forecasting rules. The Smets-Wouters DSGE model with BLE is estimated and fits well with inflation survey expectations. As a policy application, we show that learning requires a lower degree of interest rate smoothing.
When countries are hit by supply shocks, central banks often face the dilemma of either looking through such shocks or reacting to them to ensure that inflation expectations remain anchored. In this paper, we propose a tractable framework to capture this dilemma and then explore optimal policy under a range of assumptions about how expectations are formed.