MSE performance of some shrinkage estimators in a regression model with non-normal errors

"In this paper, we consider a linear regression model when the error term has chi-square dustribution. We compare some shrinkage estimators (the Stein-rule estimator, the positive-part Stein-rule estimator, the minimum mean squared error estimator and the adjusted minimum mean squared error estimator) and the OLS estimator under MSE criteria. By the Monte Carlo experiments, it is shown that the risk performances of the estimators depend not only on the numbers of regression coefficients but also on the degrees of freedom of the chi-square distribution."--Page 1.