摘要
本文从结构突变的视角对金融危机前后我国CPI涨跌(π)序列进行了内生结构变动的单位根检验,证明了其数据生成过程(DGP)为两次结构突变的趋势平稳过程而非单位根过程。并运用考虑结构突变的时序模型对π序列进行了拟合,拟合优度达到了97.28%,克服了简单地用差分序列进行建模而造成数据信息大量损失的缺陷,同时,对模型的残差序列进行ARCH效应检验,结果显示,金融危机前后我国CPI涨跌的波动并不存在ARCH效应。
In this Paper, before and after the financial crisis whose China's CPI change (π) time series is carried out the endogenous structure change's unit root test, which proves that the π variable's data generation process (DGP) is the trend stationary process of structural breaking. And the article has applied the time series' model which considerate structure change to carry on the fitting for the π series, the goodness of fit had achieved 97.28%, overcame the weakness of the data message massive losses by carried on the modeling with the difference sequence, simultaneously, carried on the ARCH effect test to the model residual error sequence, finally demonstrated that around the financial crisis our country CPI change's undulation did not have the ARCH effect.
出处
《产经评论》
2010年第1期106-113,共8页
Industrial Economic Review