The variables INCOME and HSIZE are strongly positively correlated. What would happen to the coefficient of HSIZE if the variable INCOME were dropped from the regression? Would this estimator be biased? If a bias on the new coefficient of HSIZE exists, please derive a formal expression for this bias and state if the new coefficient is downward or upward biased