中国通胀水平及其不确定性双长记忆统计特征研究——基于ARFIMA-HYGARCH-t模型  

On the Statistical Characteristics of Dual Long Memory in China's Inflation Level and Its Uncertainty:Based on ARFIMA-HYGARCH-t Model

作　　者：潘群星[1,2] PAN Qun-xing(School of Business, Nanjing University, Nanjing 210093 , China College of Finance, Nanjing Agricultural University, Nanjing 210095 , China)

机构地区：[1]南京大学商学院,江苏南京210093 [2]南京农业大学金融学院,江苏南京210095

出　　处：《统计与信息论坛》2017年第6期64-70,共7页Journal of Statistics and Information

基　　金：中国博士后科学基金项目<基于ARCH(∞)模型研究我国股票市场价格若干行为>(2015M571718);南京农业大学中央高校基本科研业务费人文社会科学研究基金项目<我国股票市场若干行为和价格波动性研究>(SK2015019)

摘　　要：基于ARFIMA-HYGARCH-t模型对1985年1月至2015年12月间中国月度通货膨胀的均值过程和波动过程进行统计检验,发现通货膨胀水平及其不确定性表现出"双长记忆"行为。在此行为下,利用VAR模型、ARFIMA-HYGARCH-M-t模型及ARFIMA-GJR-t模型检验通货膨胀水平与其不确定性之间的影响关系、影响方向与影响程度,结论支持Friedman-Ball假说;通货膨胀水平正向冲击引发的不确定性程度强于负向冲击引发的不确定性程度。经济政策操作时既要考虑维持通货膨胀的稳定性,也要考虑政策期限结构的长期性。The average process and the fluctuation process of monthly inflation rate from January 1985 to December 2015 in China is examined using the ARFIMA-HYGARCH-t model, and the ＂ dual long memory＂ behavior of inflation level and its uncertainty is found. In this behavior, the influence relationship, influence direction and influence degree between inflation level and its uncertainty are examined, respectively, by the VAR model, the ARFIMA-HYGARCH-M-t model and the ARFIMA-GJR- t model. The result supports the Friedman-Ball hypothesis. It is also found that the uncertainty degree of the positive impact of inflation level is stronger than that of negative impact. Economic policy operation should not only take into account the stability of inflation, but also consider the long-term nature of the policy term structure.

关 键 词：双长记忆特征 ARFIMA-HYGARCH模型 Friedman-Ball假说 Cukierman-Meltzer假说 

分 类 号：F222[经济管理—国民经济] O212[理学—概率论与数理统计]