# Are HAC estimators used for estimation of regression coefficients?

The references I can find on HAC procedures (like Newey-West) in regression focus on the standard error of the estimated regression coefficients and hypothesis testing involving the same. I cannot find any references where the HAC estimate of error covariance is used in a generalized least squares procedure to get 'better' estimates of the regression coefficient. Is this not done? Or are the hypothesis tests just more interesting?

-
 I interpret the brief Wikipedia article as saying that this is done. Have you looked at its references? – whuber♦ Nov 14 '11 at 18:57 @whuber: I started from Andrews and went forward. Perhaps I should work backwards. – shabbychef Nov 14 '11 at 19:11