The RATS Software Forum

Hi,

I was just wondering if it is possible to impose restrictions on the multivariate DCC GARCH model to ensure positive definiteness? I have tried adding asymmetric effects and robust errors and it doesn't solve the problem, is there anything else I could do?

At the moment some of my alpha1 and beta coefficients are negative in each specification.

Thanks

It doesn't sound like you have a problem with positive definiteness---instead, it sounds as if some of your series don't show well-behaved "GARCH" behavior. Have you tried fitting them with univariate GARCH models?