3 thoughts on “Least Squares as an unbiased estimator – matrix formulation”

  1. When you take expectation of (X'X)^-1 X'u, I am thinking it should have been E[(X'X)^-1 X']  E[u] (due to independence), not (X'X)^-1 X' E[u] as you have written, although you get zero from both cases.

Leave a Reply