The RATS Software Forum

Dear all:

I like to do a Markov Switching GJR-GARCH-M model which the regime-switching term is on the square-root of the conditional variance in mean equation. Unfortunately, I failed to complete it via modifying tsayp591, GrayGarch (1996) or SWARCH code. Please let me know if anyone has similar codes what I want.

Thank you.

Just to clarify, is the only thing that's different between regimes is the coefficient on the "M" term? Wouldn't you also at least need the intercept in the mean to change with it?