空间面板误差模型中随机效应的C(α)检验

C(α) Test for Random Effects in Panel Data Models with Spatial Error Correlation

由于空间面板误差模型中的空间误差项和随机效应之间存在相关性,传统的面板数据模型随机效应检验无法直接应用到空间面板误差模型。如果采用极大似然估计法估计空间面板误差模型然后构建随机效应检验,又会遇到计算困难的问题。文章构建了空间面板误差模型中用来检验随机效应的C(α)检验,该检验仅依赖于参数的一致估计且计算简单。蒙特卡罗模拟结果表明本文提出的C(α)检验有良好的有限样本性质。

Due to the correlation between the spatial error term and the random effects,the tests for random effects in the classical panel data model cannot be directly applied to the panel data model with spatial error correlation.However,heavy computation problem will arise when constructing random effects tests by maximum likelihood estimator.In this paper,a C（α）-type test for random effects is proposed,which relies on a consistent estimator only and is much easier to compute.The Monte-Carlo simulations show that the proposed test has good finite sample properties.