Hi everyone,
My question is sort of related to the following legacy post. I use the moments_varendo command to let Dynare compute the mean moments for a posterior sample size of 1200 (default). As I am interested in the standard deviations of my endogenous variables, I access the variance of the variables by using the following path.
oo_.PosteriorTheoreticalMoments.dsge.covariance.Mean.y_obs.y_obs

However, I find that for some of the variables the mean variance is negative which translates into a complex standard deviation.

Also, I have experimented with the posterior_max_subsample_draws = x (where I have set x = 500, 1000, 5000). However, the mean doesn’t change over the three samples.
Do you know how this is possible or am I doing something wrong?
I should add I’m using dynare version 4.5.6.
Best
Robert