Block Bootstrap方法在季节时间序列单位根检验中的应用

Application of Block Bootstrap Method in Unit Root Test of Seasonal Time Series

传统的参数类检验方法存在参数极限分布非标准、参数解析式推导烦琐、差分平稳后序列均值难估计等问题。为了避免上述问题,文章改进季节时间序列单位根检验方法,首先基于非参数Block Bootstrap方法提出BB-HEGY检验统计量,并给出临界值;其次,通过比较BB-HEGY统计量和传统的HEGY统计量的有限样本性质,发现BB-HEGY统计量具有更好功效和检验水平,尤其在季节频率上有更明显的优势;最后,通过对中国宏观经济指标的实证检验,进一步证明BB-HEGY检验统计量的优越性。

The traditional parameter class inspection methods have some problems,such as non-standard parameter limit distribution,cumbersome deviation of parametric analytical formula,difficult estimation of sequence mean after the difference is stable,etc.In order to avoid the above problems,this paper improves the unit root test method of seasonal time series.Firstly,BB-HEGY test statistics is proposed based on the non-parametric Block Bootstrap method,with the critical value offered.Secondly,by comparing the limited sample properties of BB-HEGY statistics and traditional HEGY statistics,it is found that BB-HEGY statistics has better efficacy and inspection level,these obvious advantages shown especially in terms of seasonal frequency.Finally,through the empirical test of China’s macroeconomic indicators,the paper further proves the superiority of BB-HEGY test statistics.