Abstract

Recent years have seen a surge in the econometric analysis of dynamic stochastic general equilibrium (DSGE) models using an MCMC Bayesian method. Since DSGE models are structured from micro-foundation theories, they can identify various shocks in a theoretically consistent way. It is, however, important to evaluate the DSGE model before applying it to the policy analysis because the results may depend on the model. This article develops a new method for the evaluation of DSGE models. One of the methods used for the evaluation of DSGE models is the approach using the DSGE-VAR model. In this approach, a DSGE model is used as a prior for a VAR model and the fit is compared by relaxing the cross-equation restrictions implied by the DSGE model systematically. The problem with this approach is that a VAR model is only an approximation of a DSGE model because a DSGE model cannot be represented by a VAR model. Since DSGE model can be represented by a state-space (SS) model, we use a SS model and select its prior using a DSGE model. The resulting DSGE-SS model is compared with the DSGE-VAR model by applying them to a simple New Keynesian DSGE model.