Page 1 of 1

Bayesian IRF

PostPosted: Fri Apr 27, 2012 7:56 am
by viking76
Dear Dynare users,

retrieving the posterior IRFs, through the command "oo_.PosteriorIRF.dsge", you can get the mean, median, the variance, the confidence interval for the mean, and the a number of distribution on which they are computed. In particular, by writing "oo_.PosteriorIRF.dsge.Distribution" you can get directly these different IRFs, obtained from different draws of the parameters.
My question is the following: Why are these IRFs always 9? Why not a higher number, which seems to me more appropriate? Thanks

Adrian

Re: Bayesian IRF

PostPosted: Tue May 01, 2012 3:14 pm
by jpfeifer
What do you mean with
Why are these IRFs always 9? Why not a higher number, which seems to me more appropriate?

Also, could you post the mod and data-file?

Re: Bayesian IRF

PostPosted: Mon May 14, 2012 3:59 pm
by viking76
Thanks jpfeifer for your reply,

it's not a problem of a specific model. After the estimation, if you write "oo_.PosteriorIRF.dsge.Distribution", you will always get nine series, that I think are nine posterior IRFs obtained by drawing nine times from the posterior distribution. It's a fact that emerges whatever is the model estimated. Thanks for your attention

Adrian