Robust estimation in the errors variables model via weighted likelihood estimating equations

Parameters estimates for the errors-in-variables model are obtained by solving weighted likelihood estimating equations. They are consistent, asymptotically normal and asymptotically fully efficient, and exhibit robustness properties similar to the minimum disparity estimators (Basu and Sarkar 1994a) but are immensely simpler to compute and have some theoretical advantages over the latter. We illustrate the robustness properties through some numerical studies similar to those of Zamar (1989).