EViews.com

What is the interpretation of the AIC in the vector error correction?
I have 3 AIC. The first 2, are the ones from each regression.
But what is the third one? I see it is the sum of the other two, is it kind of the "global" AIC for the VEC?

I am asking becouse I would like to compare the VEC with an AR(1) model with MA(1) errors, and I dont know which AIC of the VEC should I use to compare the models.

Thanks!!

It is the AIC based on the overall likelihood.

So should I use the overall AIC then to compare the model with an uniequetional dynamic equation? (I.e with an adaptative expectations model, , AR(1) and with MA(1) residuals)????