统计研究 ›› 2013, Vol. 30 ›› Issue (4): 99-105.

• 论文 • 上一篇    下一篇

复杂环境下空间滞后模型的稳健LM检验

钱争鸣 刘立虎   

  • 出版日期:2013-04-15 发布日期:2013-04-15

The Robust LM Test of Spatial Lag Model with Non-normal Distribution and Spatial Layout

Qian Zhengming & Liu Lihu   

  • Online:2013-04-15 Published:2013-04-15

摘要: 针对误差项非正态分布和不同空间布局的复杂环境,本文提出一种新的空间滞后模型的稳健LM检验统计量及其检验方法,对比现有LM检验方法,它不仅消除了空间误差效应对检验的影响,而且克服了传统检验方法会受误差项非正态分布以及不同空间布局差异的影响。同时还具有在小样本情况下表现依然良好的优点。经过蒙特卡罗模拟的结果支持上述论断。

关键词: 非正态分布, 空间布局, 模型设定偏误, 矩估计, 稳健检验

Abstract: This paper proposes a robust LM statistic for diagnosing spatial lag effects, which is not sensitive to spatial error effect, spatial layout and distribution misspecification, while the existing test is not. Meanwhile, it has good performance under small sample. Monte Carlo results prove this.

Key words: Non-normal Distribution, Spatial Layout, Model Misspecification, GMM, Robust Test