The RATS Software Forum

Hi there,
I could not find any code that extend the Markov Switching GARCH model of Gray (JFE, 1996) to the MV-GARCH context. Could you please help with RATS coding this model given in pg.477-478 in the paper attached?

Thank you

I'm not a great fan of Gray's method; Dueker's filter seems to work quite a bit better. (Both of them require an approximation to replace the complicated history with a single result for each regime; Dueker's waits longer to do the collapsing).

Hi Tom,

thanks for this.

I was after a bivariate garch extension. Is this possible?

Thanks