This estimator was generalized by Lee to account for a nonscalar covariance matrix and yielded the LMT-G2SLS estimator. Lee demonstrated that in a simultaneous-equation Tobit model the LMT-G2SLS estimator is more efficient than the Amemiya GLS estimator. However, Amemiya's GLS estimator is merely a member of the class of Amemiya GLS estimators containing members which beat the LMT-G2SLS, as well as one which is asymptotically equivalent to the LMT-G2SLS. Equations, notes, and six references are supplied.
Comparison of the Amemiya GLS and the Lee-Maddala-Trost G2SLS in a Simultaneous-Equations Tobit Model
NCJ Number
88422
Date Published
1982
Length
10 pages
Annotation
This paper compares Amemiya's method of estimating structural parameters in a simultaneous-equations Tobit model with an alternative method proposed by Lee, Maddala, and Trost (1980) termed the LMT-2SLS estimator.
Abstract