固定效应检验是面板数据模型研究的重要课题。文章将固定效应检验的范畴扩展至空间面板数据模型,采用LM检验方法,研究空间面板数据模型的固定效应事前检验,数理推导了固定效应检验的边际检验和条件检验统计量,并采用MonteCarlo模拟实验,研究检验统计量的有限样本性质。结果显示,文章建议的检验统计量具有较小的水平扭曲和理想的检验功效,尤其是条件检验统计量,是实证研究中有效的检验统计量。
Fixed effect test is one of the important areas in panel data model. This paper extends fixed effect test to spatial panel models and studies the fixed tests before model estimation of spatial panel models using LM test method. We make mathematical derivations of fixed effect marginal test and conditional test, and through Monte Carlo simulation experiments, study finite sample performance of these test statistics. The result shows that, the tests we proposed have idle size tortuosity and power, especially conditional test, which is effective test statistics in empirical research.