Page 1 of 1

Connectedness in Bayesian TVC VAR ?

Posted: Thu Mar 20, 2025 2:03 pm
by Tom&Spike
Hello.
I am new here and I have a little problem. I create a Baysian TVC VAR, using 3 time series with about 1500 records.
Everything works fine (the name of the model is "bayesian_tvc_var").
Now I want to compute connectedness on this model, in Eviews.
As far as I kmow, it requires extracting the time-varying Forecast Error Variance Decomposition (TV-FEVD) manually because EViews does not have a built-in Diebold & Yılmaz connectedness function for Bayesian TVC-VAR model.
I go to estimated Bayesian TVP-VAR Model Output and I click on VIEW, looking for "Variance Decomposition" (like an ordinary standard VAR). But there is not such an option there.

I tried the following commands:
smpl @first @last
show bayesian_tvc_var.fevd

I received the following error message: "FEDV is not a valid view for BAYESIAN_TVC_VAR" .

Can anyone explain me what to do, step by step ?
Thank you.

Re: Connectedness in Bayesian TVC VAR ?

Posted: Fri Apr 04, 2025 12:00 pm
by EViews Kai
Hi,

EViews does not currently support variance decomposition for the Bayesian TVCVAR; please see the forum post viewtopic.php?p=71090.