Asymptotic robustness of tests of overidentification and predeterminedness
Asymptotic robustness of tests of overidentification and predeterminedness
复制标题
过度识别和预定性检验的渐近稳健性
DOI：
10.1016/0304-4076(94)90029-9
复制
发表时间：
1994
影响因子：
6.3
通讯作者：
N. Kunitomo
中科院分区：
文献类型：
--
作者：
T. W. Anderson;N. Kunitomo
已经提出了许多统计测试程序，用于识别一个或多个方程的限制以及结构方程组中一个或多个变量的计量经济学预定性。这项研究致力于以系统的方式统一许多检验程序，并在一组局部替代假设和非常一般的扰动条件下推导检验统计量的渐近分布。通过利用新的鞅中心极限定理和鞅收敛定理，我们证明检验统计量的极限分布是局部备择假设下的非中心 χ2 分布和原假设下的中心 χ2 分布。这些限制分布在适用于各种扰动分布和模型的意义上是稳健的。我们的结果表明，计量经济学家已知的许多测试可以在不做出通常的相对限制性假设的情况下进行。
Many statistical test procedures have been proposed for identification restrictions on one or several equations and the econometric predeterminedness of one or several variables in a system of structural equations. This study is devoted to unifying many test procedures in a systematic way and to deriving the asymptotic distributions of the test statistics under a set of local alternative hypotheses and very general conditions on the disturbances. By making use of a new martingale central limit theorem and a martingale convergence theorem, we show that the limiting distributions of test statistics are noncentral χ2-distributions under the local alternative hypotheses and central χ2-distributions under the null hypotheses. These limiting distributions are robust in the sense that they hold for a variety of disturbance distributions and models. Our results show that many tests already known among econometricians can be carried out without making the usual relatively restrictive assumptions.
DOI：
10.2307/2528652
发表时间：
1972-07
期刊：
--
影响因子：
--
作者：
T. Anderson
通讯作者：
T. Anderson