A SIMPLE CHARACTERIZATION OF SEEMINGLY UNRELATED REGRESSIONS MODELS IN WHICH OLS IS BLUE

被引：22

作者：

BARTELS, R

FIEBIG, DG

机构：

来源：

AMERICAN STATISTICIAN | 1991年 / 45卷 / 02期

关键词：

EQUIVALENCE CLASSES; NECESSARY AND SUFFICIENT CONDITIONS; TRANSLOG SYSTEM;

D O I：

10.2307/2684378

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Recent contributions to the discussion about the conditions under which ordinary least squares in the seemingly unrelated regressions (SUR) model is the best linear unbiased estimator suggest a characterization of SUR models that is convenient for checking whether these conditions are satisfied. Standard pedagogic examples for SUR's, as well as more complex cases, are easily derived from our characterization.

引用

页码：137 / 140

页数：4

共 10 条

[1]

Baltagi B. N., 1989, ECONOMET REV, V7, P165

[2]

Baltagi BH., 1988, ECONOMET THEOR, V4, P536

[3]

GRIFFIN JM, 1976, AM ECON REV, V66, P845

[4]

Judge GeorgeG., 1985, THEORY PRACTICE ECON, V2nd ed.