Page 1 of 1

Problem with conditional_variance_decomposition 4.2.1

PostPosted: Mon Jul 04, 2011 11:01 pm
by balax
Hi All,

I am using dynare 4.2.1 and trying to run a conditional variance decomposition after estimating a DSGE model and I get the following error message:

?? Reference to non-existent field 'sigma_e_is_diagonal'.

Error in ==> dsge_simulated_theoretical_conditional_variance_decomposition at 116
StateSpaceModel.sigma_e_is_diagonal = M_.sigma_e_is_diagonal;

I have checked the path and there is no problem with it, dynare is calling the 4.2.1 version. When calling the M_ I actually realize that there isn´t a sigma_e_is_diagonal field. Is this a bug?

This is the line for the estimation command that I am using:

estimation(order=1,datafile=DATAUSAEUROLONG,mh_replic=2000,mh_nblocks=1,mh_jscale=0.15,mh_drop=0.2,prior_trunc=1e-32,nograph,moments_varendo,conditional_variance_decomposition =1)

Hope you can help me.

Best,

Re: Problem with conditional_variance_decomposition 4.2.1

PostPosted: Thu Aug 25, 2011 4:10 pm
by SĂ©bastienVillemot
This is a bug in Dynare. Thanks for reporting it. We will hopefully fix it in the next release.

Re: Problem with conditional_variance_decomposition 4.2.1

PostPosted: Tue Apr 10, 2012 10:06 am
by beeimp
I am using the latest version 4.2.5 and there is the same problem.

Is the bug fixed in dynare 4.2.5?

Re: Problem with conditional_variance_decomposition 4.2.1

PostPosted: Wed Apr 18, 2012 8:23 pm
by jpfeifer
When I use Dynare 4.2.5, conditional_variance_decomposition in the estimation command works. But you must also specify moments_varendo (as you do). However, there is no explicit output. All results are stored in the structure
Code: Select all
oo_.PosteriorTheoreticalMoments.dsge.ConditionalVarianceDecomposition