March 2016 issue contents
A forecasting metric for evaluating DSGE models for policy analysis

by Abhishek Gupta


This paper evaluates the strengths and weaknesses of a dynamic stochastic general equilibrium (DSGE) model from the standpoint of its usefulness in doing monetary policy analysis. The paper isolates cross-correlations among one-step-ahead forecast errors as the most relevant feature for practical monetary policymaking and uses the diagnostic tools of posterior predictive analysis to evaluate them. The paper accounts for the observed flaws in the model with regards to these features using the correlation structure among the estimated shocks. This corresponds to testing and rejecting the over-identifying restriction of no correlation among the structural shocks in the model. The paper attributes this correlation among the estimated structural shocks to model misspecification.

JEL Codes: C11, C52, E1, E58.

Full article (PDF, 33 pages, 928 kb)