Consider the following population model

inc,- = Bleduci + e,- where e,- is such that E[i|educi] = 0. For convenience, define inc,- = 31,- and educi = xi. Again, we have omitted the intercept for convenience. Now suppose that, unlike lecture, we have Var[i2|xi] =E[,-2|x,-] = azxi (in class we assumed Var[,-2|xi] = Var[i2] = E[i2] = 02). 1 n E i=1xiYI' . . m. Is b1 an unbiased estimator for [91? a) Define the estimator bl = n i=1xi 1 E 151:1)\" T n i=1 b) Define the estimator bi\" = . ls bf\" an unbiased estimator for El? xi c) What is Var(b{VIxI-)