季节调整对单位根检验的影响——基于蒙特卡罗模拟的研究  

The Effect of Seasonal Adjustment on Unit Root Tests:A Study Based on Monte Carlo Simulation

作　　者：栾惠德[1] 

机构地区：[1]中国人民银行营业管理部,北京100045

出　　处：《数理统计与管理》2009年第3期457-467,共11页Journal of Applied Statistics and Management

基　　金：国家自然科学基金(No.70571039)资助

摘　　要：季节调整是使用季节数据进行经济分析的前提.但是季节调整是否会对单位根检验的结论造成影响却是一个存在争议的问题.本文扩展了Ghysels和Perron的分析框架,通过蒙特卡罗模拟考察了X-12-ARIMA季节调整对ADF和DF-GLS单位根检验的影响.结果表明,对于季节时间序列,使用季节调整后数据,只要滞后阶数选取合适,单位根检验的功效就不会发生严重损失.Seasonal adjustment is the premise to economic analysis using seasonal data. But it is still under dispute whether seasonal adjustment affects the conclusion of unit root tests or not. Based on the extended framework of Ghysels and Perron[1], this paper presents the simulation study of both the size and power of the usual ADF and DF-GLS statistics. The results show that for seasonal time series, there will not be serious loss in power with seasonally adjusted data if the appropriate truncation lag is selected.

关 键 词：季节调整 单位根检验 蒙特卡罗模拟 

分 类 号：F224[经济管理—国民经济]