Double shrinkage estimation of common coefficients in two regression equations with hetersocedasticity (Q1275411)

The admissibility of feasible generalized least squares estimators of the common regression coefficients in two linear regression equations with different unknown error variances is considered. It is proved that the FGLS estimators are inadmissible in light of minimizing the covariance matrices if the dimension of the common regression coefficients is greater than or equal to three. Double shrinkage unbiased estimators are proposed as possible candidates of improved procedures.