On the efficiency of the Cochrane-Orcutt estimator in the serially correlated error components regression model for panel data

We consider the panel data regression model when the disturbances follow a serially correlated one-way error structure. For the usual linear regression model with AR(I) disturbances, the Cochrane-Orcutt(CO) estimator is a well known alternative to OLS-and GLS-estimator. In this paper we show in which situations the GO-estimator is worse or better than its competitors in terms of efficiency which is here defined as the ratio of total variances.