The RATS Software Forum

Dear Tom,

I have 3 series (Y1,Y2,Y3)over past 100 years and I assume that each series is described by a mixture of two normal distributions but the transition to each distribution is common to all 3 series. In other words, Yt,i=Ust,i+et,i, i=1,2,3. I am wondering if I can use MSVAR by setting lag=0 or I should use MSSysregression? Thank you in advance

MSVAR is for a VAR (or autoregression) only. It's only a self-contained model like that which can allow for a switching "mean" (rather than intercept) the way that Hamilton handles it. Anything else should use the simpler (and more general) MSSysRegression.

TomDoan wrote:MSVAR is for a VAR (or autoregression) only. It's only a self-contained model like that which can allow for a switching "mean" (rather than intercept) the way that Hamilton handles it. Anything else should use the simpler (and more general) MSSysRegression.

Thank you for the quick reply. I ordered the course, but the course material has not sent to me yet. Is there any example you can kindly provide for me to study MSsysregression? Many Thanks

Ehrmann, Ellison, Valla. Balcilar, Gupta, Miller. Francis and Owyang.