Page 1 of 1

Cluster-robust standard errors

Posted: Wed Nov 21, 2012 10:02 am
by CopaLibertador
Hello all,

estimating a panel regression i would like to relax some assumptions concerning the errors.
Am I right that coef covariance method White period allows for correlation over periods for each individual?
In contrast to White period, what does White cross-section and White (diagonal) do?

Thanks for help.
CopaLibertador

Re: Cluster-robust standard errors

Posted: Wed Nov 21, 2012 11:17 am
by EViews Glenn
White cross-section allows for clustering by period (contemporaneous correlation). The White diagonal allows for general heteroskedasticity (but no cross-observation correlation).