Estimation of unbalanced Panel VAR model?

[Atiq Ur Rehman]

I have data with N=8 and and three variables (two of which with T = 40 and one variable with T = 29). It means the panel data is unbalanced.

My question is how EViews handle unbalanced panel data when estimating VAR model?

Thanks in advance!

[EViews Glenn]

EViews doesn't do anything special with estimation of a VAR in a panel workfile. We simply take all of the lags and run on the stacked data. So the number of observations in different cross-sections will be different, but from the point of view of the estimator, T = sum_i T_i* where T_i* are the number of observations for cross-section T_i after excluding observations with missing data or lags.