20) Consider a linear regression model Yi = Bo + 31Xlit . . .+ 3x Xki tui. Please evaluate the following five statements. i) A biased estimator could have smaller variance than the OLS estimator under the OLS assumptions and homoskedasticity. ii) A nonlinear estimator could have smaller variance than the OLS estimator under the OLS assumptions and homoskedasticity. iii) OLS estimator is robust to large outliers. iv) OLS estimator is the best among all the estimators. v) We could always assume E(uz ) = 0 even if the intercept term Bo is zero. The number of correct statements is A) 2 B) 3 C 4 D) 5