EViews.com

Hi

I am applying a meta-analysis method to an unbalanced panel dataset.
I need to calculate the average correlation between regression residuals.
so far I am doing the equation per each year, I create the residual series, I export it to excel and then I ll calculate manually the correlation between different years.
Does it exist a direct command in Eviews to do it?
given that is an unbalanced panel data (i.e. the observation among years are not the same) does the correlation of residuals is biased on the number of observation?
thank you

What I'd do is to create a pooled workfile and unstack the residuals into the pooled workfile. Then compute the correlation across the series in a group containing the unstacked series. You should be able to do it with a reshape command.

Are you sure that the residuals of a pooled dataset that contains all the years are the same as the residuals of each year separately?