EViews.com

Hi all,

does someone has any idea how to get robust errors in ml estimation?

I am trying to estimate a GARCH model and need to calculate Bollerlev - Wooldrige robust errors.

Any help or hint?

Thanks

Dav

I am trying to estimate a GARCH model [in ML object] and need to calculate Bollerlev - Wooldrige robust errors.

I don't think you can do this in the ML object. BW-standard errors are calculated after ML estimation and in a particular way that's only applicable to GARCH models. This is not how the ML object works where the SEs are estimated from the gradients of the ML model. It's not design for user to tinker with this last step. I can't think of a way tricking EViews to calculate the BW SE.