The RATS Software Forum

Thank you Tom,

You're right but, by using @MSRegStdResiduals I can not get regime-specific standardized residuals. So, is there any suggestion?

Regime-specific standardized residuals don't make sense either. The "standardized" residuals (computed with @MSRegStdResiduals) use the full model and weight based upon the probabilities of the regimes. Regime-specific residuals don't distinguish between periods where that particular regime is likely and those where it isn't (and possibly has effectively a zero probability).