'Show that when the covariance matrix of the errors is S 2 diagf1=W2 1 ; ... ; 1=W2 n g [ 2

'Show that when the covariance matrix of the errors is S ¼ σ2

ε · diagf1=W2 1 ; ... ; 1=W2 n g [ σ2

εW"1 the weighted-least-squares estimator flb ¼ ðX0 WXÞ

"1 X0 Wy

¼ My is the minimum-variance linear unbiased estimator of fl (Hint: Adapt the proof of the GaussMarkov theorem for OLS estimation given in Section 9.3.2.)