Hello,
I have a question about HPD intervals for impulse responses when B-VAR is estimated with data in first differences. I understood that we have to cumulate IRFs ourselves, as there is not that option in Dynare but what about HPD intervals? How should they be adjusted for in that case?
Thank you in advance.
That is tricky. The cumulated HPDI is generally not equal to the HPDI of the cumulated IRFs. I am pretty sure that Dynare does not provide the necessary output to generate the desired HPDIs. So you might have to look at a different toolbox or spend quite some time modifying the Dynare routines to suit your needs.