Robust Estimation of the SUR Model

Keywords: Seemingly unrelated regression; robustness; S-estimators

Abstract: This paper proposes robust regression to solve the problem of outliers in seemingly unrelated regression (SUR) models. We present an adaptation of S-estimators to SUR models. S-estimators are robust, with high breakdown point, and are much more efficient than other robust regression estimators commonly used in practice. Furthermore, modifications to Ruppert's algorithm allow a fast evaluation of it in this context. The classical example on U.S. corporations is revisited, and it appears that the procedure gives an interesting insight of the problem.