Data and code for "Mis-specified Forecasts and Myopia in an Estimated New Keynesian Model"

The paper considers a New Keynesian framework in which agents form expectations based on a combination of autoregressive mis-specified forecasts and myopia. The proposed expectations formation process is shown to be consistent with all three empirical facts on consensus inflation forecast errors: delayed over-shooting to shocks, under-reaction to ex-ante forecast revisions, and over-reaction to recent events. However, while mis-specified forecasts can be both sufficient and necessary to match all three facts, myopia alone is neither. The paper then derives the general equilibrium solution consistent with the proposed expectations formation process and estimates the model with likelihood-based Bayesian methods, yielding three results:(i) macroeconomic data strongly prefer a combination of autoregressive mis-specifiedforecasting rules and myopia over other alternatives; (ii) learning of mis-specified forecasting rules improves model fit relative to the consistent expectations equilibrium; and (iii) the proposed expectations generate internal persistence and amplification to exogenous shocks. The posterior distribution of the model-implied inflation expectations fits survey data well and reveals that learning of mis-specified forecasting rules is preferred in terms of replicating all three empirical facts on forecast errors.