8. Consider the usual regression model. The random vector Y is n x 1, with Y = X + 1 , where is a p 1 vector of (unknown) constants, X is an n xp matrix of known constants with rank (X) = p (so that (XTX)- exists), & is an n x 1vector of random variables with E() = 0 and vcv() = 0Inxn where Inxn is the n xn identity matrix. Let W = X(XTX)-XTY, where XT is the transpose of X. a. Find E(W). (10 points) b. Find the variance-covariance matrix of W, vcv(W). (40 points).