中国通货膨胀率的非线性动态行为研究——来自MS—ARFIMA模型的新证据

使用允许长记忆参数d服从区制转换的MS—ARFIMA模型对中国月度通货膨胀路径的动态行为进行新的实证研究，结果显示：中国通货膨胀不仅均值水平和不确定性存在着“低通胀”区制和“高通胀”区制，而且更为重要的是，通货膨胀序列的平稳性也表现出显著的区制转换动态。“低通胀”区制下，长记忆参数d1=0．361，说明通货膨胀是协方差平稳序列，“高通胀”区制下，长记忆参数d2=1．145，说明通货膨胀是非平稳序列。这一新的研究结论意味着中国通货膨胀冲击的持久性效应也存在相应的区制转移变化。这要求央行在管控通货膨胀过程中，既要考虑均值和不确定性的区制变化，又要兼顾平稳性和持久性的区制变化。

Nonlinear Dynamic of Inflation in China:New Evidence from MS-ARFIMA Model

This paper performs new empirical study on dynamic of monthly inflation path of China by adopting MS-ARFIMA model which allows long memory parameter follow regime switching. The results show that both mean and uncertainty of China＇s inflation have low inflation regime and high inflation regime, and more importantly, the stability of inflation series also follows significant regime switching dynamic. Under low inflation regime, long memory parameter , which means inflation series is covariance stationary. Under high inflation regime, long memory parameter, meaning that inflation series is nonstationary. This new result implies the persistence effect of shocks to China＇s inflation also has regime switching changes, which requesting the Central Bank to not only take regime changes of mean and uncertainty into consideration, but also take regime switching of stability and persistence into account when controlling inflation.