Hi, when estimating dsge-var: when and where do you put the real data? The file simul_hybrid.mod generates artificial data, from (I guess) the posterior means of parameters, their values are given at the beginning(?). In the dsgevar_forward_estimated_lambda.mod are used simulated data. The posteriors for DSGEVars (N-IW) depend on the real data. So where should they be used.
Best regards
Renata
Hi, In these files we use artificial data to illustrate how the DSGE-VAR estimation works. You can safely replace the artificial sample by ``true’’ data.This has nothing to do with the dummy variable approach of the Bayesian VAR estimation. Actually the DSGE-VAR estimation is not implemented with dummy variables, which would introduce noise in the estimation, but with the population moments appearing in the expression of the ML estimator of the VAR model, as explained in the paper by Schorfheide and Del Negro (Priors from General Equilibrium Models for VARS, in International Economic Review, 2004).
Best,
Stéphane.