基于Markov切换模型的中国CPI指数波动特征分析  

Analysis on Chinese CPI Fluctuation Characteristics Based on Markov-Switching Model

作　　者：文林莉[1] 

机构地区：[1]广东科学技术职业学院,广东珠海519090

出　　处：《统计与决策》2017年第14期26-30,共5页Statistics & Decision

摘　　要：文章引入Markov切换模型探讨中国CPI指数波动的持续性、波动幅度以及波动频率等特征。通过ADF方法和JB统计量检验CPI时间序列的平稳性和非对称性,通过CUSUM方法检验时间序列的机制转移特征,采取极大似然估计方法确定Markov切换模型的相关参数。结果表明,CPI时间序列的低波动性持续的时间最长,并且三种波动状态之间的切换主要集中于低波动状态和高波动状态之间;CPI时间序列的波动主要稳定于低波动率,并且存在明显的"聚集波动"和"分段波动"现象。This paper introduces Markov-switching model to study the fluctuation persistence, fluctuation amplitude and fluctuation frequency of Chinese CPI. The paper firstly verifies the stability and asymmetry of CPI time series by using ADF meth- od and JB statistical magnitude, and then tests the regime switching features of time series through CUSUM method. Finally the maximum likelihood estimation method is adopted to determine the relevant parameters of Markov switching model. The analysis result shows： the low fluctuation of CPI time series lasts longest, and the three-state switching concentrates between the low fluctu- ation and high fluctuation; the fluctuation of CPI series is stable mainly at low state, and there are obvious ＂clustering fluctuation＂ and ＂stage fluctuation＂ phenomenon.

关 键 词：居民消费指数 Markov切换模型 波动特征 自回归分析 

分 类 号：F224.0[经济管理—国民经济]