The RATS Software Forum

For MGARCH models, the time series textbooks state that we need to impose restrictions on the parameters to ensure the non-negativity of the conditional variances of the individual series.

Do we need to impose these restrictions manually in WinRATS? Or, the WinRATS automatically impose the restrictions to ensure positive definite matrix Ht?

P.S. I am a self GARCH learner and I have not attended any GARCH lecture or class. I apologize if my question is too simple or stupid. I already read extensively to avoid stupid questions. Thank a lot.

"Need to" is overstating the situation. There are plenty of data sets where the simple DVEC model (the default MV=STANDARD) fits fine. However, there are also many where it doesn't. Instead of trying to somehow impose this manually (which would be very hard to do), just use one of the other options for MV which choose a parameterization that enforces positive-definiteness.