I have been confused about what is the difference of the “conditional variance decomposition” commands, between in “estimation” and “stoch_simul”.
And recently I have read a paper that compute variance decomposition of every variable before 2007, after 2007, and the whole period.I would like to know how to do that.
stoch_simul will generate the variance decomposition at the specified parameter set. Bayesian estimation wil provide the average variance decomposition over the posterior distribution of parameters.
If you want to have it for different time periods, you need to estimate the model for these subsamples. See e.g. github.com/JohannesPfeifer/DSGE_mod/tree/master/Ireland_2004
Thanks a lot , it really helps