This paper considers estimation of a linear regression model using data where some covariate values are missing but imputations are available to fill-in the missing values. The availability of imputations generates a trade-off between bias and precision in the estimators of the regression...