DSGE-VAR: Posterior distribution of PHI and SIGMA

[Aldo]

Greetings dear all,
I'm trying to use dynare codes for analyzing an DSGE-VAR model as Del Negro & Schorfheide(2003): "PRIORS FROM GENERAL EQUILIBRIUM MODELS FOR VARS" https://www.ecb.europa.eu/events/pdf/conferences/schorfheide.pdf.

In equations (30) and (31), about posterior distribution of PHI and SIGMA, I see that PHI and SIGMA are generated from an Inverted Wishart-Normal form. My question are:

1. In dynare codes ¿Where is done the extraction from Inverted Wishart-Normal form?. I ask this because I reviewed in http://www.dynare.org/dynare-matlab-m2html/matlab/dsge_var_likelihood.html and I didn't find one generating from an Inverted Wishart-Normal form.

2. I want to use this code: http://www.dynare.org/dynare-matlab-m2html/matlab/distributions/rand_inverse_wishart.html for generating a m-by-m matrix from an inverse Wishart distribution with parameters Tau and df, that is with df degrees of freedom and where Tau is a symmetric and positive definite matrix. Is it okay to use this syntax:

Code: Select all

rand_inverse_wishart(m, df, Chol(Tau))

for my purpose?

I'd appreciate it if you could reply to my question.

[jpfeifer]

1.) That function only evaluates the posterior density. You do not need to generate anything from the Inverse Wishart normal. For IRFs, you need this.
2.) Please have a look at the call to rand_inverse_wishart.m in http://www.dynare.org/dynare-matlab-m2html/matlab/PosteriorIRF_core1.html

[Aldo]

Thank you veru much.

Cheers

Aldo