Exercise 12.6b Consider a collection of r uncorrelated regression models Yk = Xk + ek , E(ek)

Exercise 12.6b Consider a collection of r uncorrelated regression models Yk = Xkβ + ek , E(ek) = 0, Cov(ek) = V, each with N observations. Find the least squares estimate ˆ β from all the data. Find an estimate of V. Find the sandwich estimator for Cov( ˆ β). Why is this a reasonable estimate? Hint: V = E[(Yk − Xkβ)(Yk − Xkβ)

].