I understand that the Newey-West estimator is an approach to estimate the covariance of the linear regression estimator (OLS) when there are heteroskedasticity and autocorrelation. Should it modify the estimator itself though? In other words, if $$\hat\beta_{\text{OLS}}=\left(X^\prime X\right)^{-1} X^\prime y,$$ then what should $\hat\beta_{\text{NW}}$ be?

  • 2
    $\begingroup$ The point estimator itself remains unaffected, NW is about, as you say, using another covariance estimator. $\endgroup$ Mar 29, 2016 at 11:22


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.