EViews.com

Dear friends,

I want to calculate the (system) log likelihood value for a vecm myself. I can calculate the determinant of the residual covariance (degree of freedom adjusted) but fail to calculate the log likelihood (cf. Eviews User Guide II p. 461f). I attached an excel file with the calculations. Where is my mistake?

Can someone please help me?

Best regards,
Dada

Anyone?

in VECM models, the log-determinant used in computing the likelihood is the restricted one, not the unrestricted one. These are the same ones you get when doing the cointegration tests. We don't report the restricted determinant so it's going to be hard for you to compute without doing all of the computations done in a Johansen testing.

Ok, thank you!