基于结构变化与共同因子串行效应的面板协整检验

Testing Panel Cointegration with Serial Effect Between Structural Breaks and Common Factors

面板数据中的结构变化和共同因子存在并行和串行两种作用方式。首先,对串行效应和并行效应的定义、后果进行说明,提出效应存在性的检验方法。然后,从模型设定、参数估计、检验统计量等方面提出串行效应协整检验理论框架,并推导了协整检验统计量的极限性质。最后,应用中国经济数据进行实证分析,验证了串行效应的存在,说明区分面板数据内在效应存在方式的重要性。

There are parallel and serial ways of structure breaks and common factors in panel data. In this paper, the definition and consequence of serial effect and parallel effect are described and the test method of existence of effect is proposed. The theoretical framework of cointegration test with serial effect is proposed from the aspects of model setting, parameter estimation and test statistic, and the limit properties of cointegration test statistics are derived. Finally, the empirical analysis of China＇s economic data verifies the existence of serial effect, indicating the importance of distinguishing the existence of the intrinsic effect of panel data.