Professional Survey Forecasts and Expectations in DSGE Models

This paper proposes a strategy to exploit timely information from survey data in DSGE models. We include the Survey of Professional Forecasters data on consumption, investment, output, and inflation expectations in the set of observable variables to discipline the dynamics of model-based predictions and evaluate alternative belief specifications. Due to surveys, we are able to improve identification of fundamental shocks and models predictive power by separating the sources of low and high frequency volatility. Models with an imperfectly-rational, time-varying expectation
formation mechanism based on Adaptive Learning can reduce limitations implied by the Rational Expectation hypothesis and process survey data more effectively.