非平稳面板数据离散选择模型的Wald统计量研究

Wald Statistic in Discrete Choice Model with Nonstationary Panels

随着大数据时代的来临和统计制度的完善,宏观金融领域越来越倾向于使用大维面板数据进行经验性研究,而大维面板数据模型理论研究已成为现代计量经济学理论研究的一个热点。本文主要进行非平稳大维面板数据离散选择模型的渐近理论研究。主要研究发现,在真实回归参数值为0假设前提下,极大似然估计量具有一致性并且渐近服从正态分布;传统显著性检验Wald统计量渐近服从卡方分布。

With the advent of the big data era and the improvement of the statistical system, the macroeconomic and financial field increasingly tend to use large-dimensional panel data to do empirical research and the theory of large- dimensional panel data model has become a research focus in modern econometric theory. This paper studies the limit theory for discrete choice nonstationary panels. The results show that when the true parameter vector of explanatory variables is zero, the maximum likelihood estimator is consistent and has a normal limit distribution. Also the classical Wald statistic is still useful and has a Chi-squared limit distribution.