Page 1 of 1

posterior theoretical moments

PostPosted: Tue Feb 12, 2013 12:29 pm
by tess1704
Dear all,

I used the estimation option "moments_varendo" in Dynare to get the estimated posterior theoretical moments. Then, I find the results in the oo_.PosteriorTheoreticalMoments.
I am interested in the correlations with output (y). So I open: oo_.PosteriorTheoreticalMoments.dsge.correlation.mean.y.

Are these the correlations of each variable with output? Because if I look at the correlation of y with y, I would expect to find a "1", instead there is 0.54. Is this right?

I also checked the correlation of two variables, x and y for example and I found that corr(y,x) is different from corr(x,y). Is there any explanation?

Finally, I need also the estimated standard deviations and autorcorrelations. Where can I find them? Is the estimated variance of y, for example, stored in oo_.PosteriorTheoreticalMoments.dsge.covariance.mean.y(1,1) if y is the first variable?

Thank you in advance for your help

Alice

Re: posterior theoretical moments

PostPosted: Fri Mar 08, 2013 2:40 pm
by SébastienVillemot
oo_.PosteriorTheoreticalMoments.dsge.correlation actually contains autocorrelation functions. I will fix the reference manual.

Note that this field is also affected by a bug: in some cases, only the first element of the autocorrelation function is computed. See https://github.com/DynareTeam/dynare/issues/290