bandpass filter stoch_simul

[DW916]

Dear Johannes,

If I use bandpass filtered data to do Bayesian estimation with 4.4.3 version dynare ( I know you do not recommend bandpass filter since it is two sided, but I have to try this now....), then I calibrate the model parameters with posterior means and to see theoretical moments, like unconditional variance decomposition and standard deviation... How should I write the stoch_simul demand? ( I mean that, if use hp-filter,then stoch_simul command should include hp_filter=1600, then How about bandpass filter?)

Just

Code: Select all

stoch_simul(order=1);

would be wrong?

Could you please give me some hint?

Many thanks in advance,
Catherine

[jpfeifer]

If you know what your are doing, estimating with bandpass-filtered data is ok.
Regarding generating bandpass-filtered theoretical moments from the stoch_simul-command: that is not possible with Dynare 4.4.3. In the unstable version, to be released as 4.5, you have a bandpass_filter option. See https://github.com/DynareTeam/dynare/blob/master/tests/moments/example1_bp_test.mod

[DW916]

If you know what your are doing, estimating with bandpass-filtered data is ok.
Regarding generating bandpass-filtered theoretical moments from the stoch_simul-command: that is not possible with Dynare 4.4.3. In the unstable version, to be released as 4.5, you have a bandpass_filter option. See https://github.com/DynareTeam/dynare/blob/master/tests/moments/example1_bp_test.mod

Many thanks Johannes. I have tried to use unstable version with bandpass_filter option and got the bandpass filtered unconditional variance decomposition.

Code: Select all

stoch_simul(order=1,irf=0,bandpass_filter=[6 32], conditional_variance_decomposition=[1,4,8,16,24])

. However , it seems that dynare does NOT report bandpass filterd conditional variance decomposition.

Could you please tell me how to see bandpass filtered conditional variance decomposition in the unstable version？

Kind regards,
Catherine

[jpfeifer]

Dear Catherine,
this functionality is not supported and I am not sure there is an easy way to get a theoretical filtered conditional variance decomposition (I have never seen anything like this). Dynare also does not provide this for the hp_filter option.