[Gretl-users] panel model news 2

An addendum to http://lists.wfu.edu/pipermail/gretl-users/2016-March/011726.html On studying the Cameron and Miller piece on clustered variance estimation (JHR, 2015), I see that the Arellano estimator performed on quasi-demeaned data is reckoned to be a valid cluster-robust method for the (one way) random effects model. So this is now available in gretl, use panel <regspec> --random-effects --robust And given those options, the Hausman test is now based on a cluster-robust auxiliary regression. Allin