Reply
Thu 17 Oct, 2019 05:56 am
I have a question from econometrics, which is pretty much unclear for me. The question is:
If the errors are not homoscedastic, i:e: Var(ε|x) ≠ σ² for all x, then would the OLS estimator still be unbiased? What would be the consequence of ignoring that the errors are heteroscedastic?
Yep, that sure is unclear!