摘要
本文基于面板数据空间误差分量模型,提出空间Hausman检验,并构造出辅助回归模型的空间Hausman检验,进而通过Monte Carlo模拟实验,研究空间Hausman检验,以及辅助回归空间Hausman检验的有限样本性质。研究结果表明,空间Hausman检验能有效矫正空间面板数据下经典Hausman检验的水平扭曲,但随着空间相关性和样本量增大,其水平扭曲偏离理想值;辅助回归空间Hausman检验始终保持理想的水平扭曲。此外,二者均具有优越的检验功效。
We propose spatial Hausman test based on panel data spatial error components model,and construct spatial Hausman test based on auxiliary regression model through mathematical derivation,then research the finite sample performance of spatial Hausman test and spatial Hausman test based on auxiliary regression model through Monte Carlo simulation experiments.The results indicate that,spatial Hausman test can decrease the size distortion caused by classic Hausman test,but with spatial correlation or sample increasing,size distortion diverge ideal value;Spatial Hausman test based on auxiliary regression model always keeps ideal size distortion;What’s more,the two spatial Hausman tests have ideal power.
出处
《统计研究》
CSSCI
北大核心
2013年第5期90-97,共8页
Statistical Research
基金
国家自然科学基金项目"面板数据空间误差分量模型估计方法研究"(71071060)资助
