[5 points] In this exercise we investigate what happens to the distribution of Bo in linear regression when the homoscedasticity assumption is not satisfied. Assume that, given x1, . .., In, Yi = Bot Bititei, where ci ~N(0, 2). Assume that c are independent. This is similar to the simple linear model we have used in class with the exception that the assumption of homoscedasticity is not satisfied. Consider the estimator for the slope Bo = Y - Bit, where 31 =di-thi Find the distribution of Bo. (Keep in mind that with heteroschedasticity you loose independence between Y and B1.)