Page 1 of 1

MS-BSVAR

PostPosted: Mon Apr 27, 2015 3:47 pm
by bercerder
Hello!

How to choose optimal number of regimes for MS-BSVAR?

Thank you in advance.

Re: MS-BSVAR

PostPosted: Tue Apr 28, 2015 5:42 am
by jpfeifer
There are no general guidelines. You could work from marginal data densities or similar criteria to determine the optimal number. But in practice people often have a strong prior and go for a small number of regimes as the data is often not sufficiently informative to provide insights into a bigger number of regimes.