固定效应模型基于个体LM的新可混合性检验  

A New Poolability Test for Fixed Effects Models Based on Individual Lagrange Multipliers

作　　者：徐凤[1] 黎实[2] 

机构地区：[1]西南财经大学经济数学学院 [2]西南财经大学统计学院

出　　处：《统计研究》2014年第9期91-96,共6页Statistical Research

摘　　要：对固定效应模型,本文基于拉格朗日乘数(LM)原理提出了一种新的可混合性检验。不同于已有的LM型可混合性检验,这里使用每个截面个体的LM统计量构建可混合性检验统计量。数理分析表明,本文所提出的方法有着渐进正态性,对于扰动项的异方差和非正态均稳健,且与PY检验(Pesaran&Yamagata,2008)渐近等价。Monte Carlo模拟实验表明,相对于PY检验及另外两种LM型的可混合性检验,对于不同大小的(N,T),本文提出的方法有着良好的水平表现和更优越的检验势。This paper proposes a new poolability test based on lagrange multiplier （LM） principle. Different from existing LM type of poolability test, we construct a new test statistic based on individual ]agrange multipliers. The new test is mathematically proved to be asymptotical normality and to be robust to the errors＇ Heteroscedasticity as well as non- normality. Meanwhile, it is also proved to be equivalent with the PY Test （Pesaran ＆Yamagata, 2008）. Through Monte Carlo simulation comparing with the PY test and the other two type of LM poolability tests, it is found that the proposed test performs well in size and better in power in various combinations of N and T.

关 键 词：拉格朗日乘数 可混合性检验 固定效应 

分 类 号：O212[理学—概率论与数理统计]