Premium
Kantorovich inequalities and efficiency comparisons for several classes of estimators in linear models
Author(s) -
Liu S.,
Neudecker H.
Publication year - 1997
Publication title -
statistica neerlandica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.52
H-Index - 39
eISSN - 1467-9574
pISSN - 0039-0402
DOI - 10.1111/1467-9574.00058
Subject(s) - mathematics , estimator , trace (psycholinguistics) , generalized method of moments , matrix (chemical analysis) , variance (accounting) , generalized linear model , generalized least squares , least squares function approximation , statistics , philosophy , linguistics , materials science , accounting , business , composite material
New matrix, determinant and trace versions of the Kantorovich inequality (KI) involving two positive definite matrices are presented. Some of these are used to study the efficiencies of minimum‐distance (MD) estimators, generalized method‐of‐moments (GMM) estimators and several estimators specific to longitudinal or panel‐data analysis. They are also used to give upper bounds for the determinant and trace of the asymptotic variance matrix of a weighted least‐squares (WLS) estimator in the generalized linear model.