4. [10 points] Suppose that you want to estimate the following regression: Yt = B1 + B 2 X2+ + B3 X3+ + ... + Bx Xkit + ut You have good reason to believe that the error term is correlated not only with the error one period before but also with the error two periods before. That is, you have second order autocorrelation. This relationship can be written as followed (with & being independent and identically distributed- iid). Ut = plut - 1 + put - 2 + Et a. (3 points) Yes or No: Will the OLS estimates be unbiased? Will the OLS estimates be efficient? Will the OLS standard errors be correct? b. (3 points) Propose a transformation of the data that will correct the autocorrelation problem. Write out the transformed regression explicitly. c. (2 points) Show that the transformed regression is homeskedastic. d. (2 points) Explain how you could estimate pl and p2 in two steps in STATA