\f(6) (4 pts) One consequence of the least square assumptions is that the covariance of X; and u is zero, that is, Cov( X,, w;) = 0. When Cov(X,, w;) * 0, the OLS approach will be biased. Now suppose we have another variable Z, satisfying Cov(X,, Z;) #0, Cov(Zi, w/) = 0. (1) (i) (2 pts) Show that B1 = Cov(14, Zi) Cov(X1, Zi) (ii) (2 pts) Propose an estimator of . (Hint: the sample covariance of any two random variables W, V is given by swv = Et (W -V)(W -W)/(n -1), where W = ) 2 W, V = EL, W.)