Hi everyone!

When we have panel data, we have to take into account that the observations are not independent (that is, we observe the same individual during a different periods of time). When we compute the standars errors we have to take that into account.

From my knowledge, there are three different ways to compute consistently the standard errors in panel data (assuming a Random Effect model with strict exogenaity):

1) Pooled OLS with cluster
2) GLS using the fitted residuals of the OLS to compute the matrix Omega
2) GLS using the specific case of the standard error component model.

To obtain 1), I just do with command reg and option vce with cluster. To obtain 3), with command xtreg is enough because by default the standard errors are calculate in that way.

My question is, how can I compute number 2? is anyway to do it using command xtreg but putting some specification in option vce?

Thank you,

Diego.