CAJ | 학술논문

现有测度通胀惯性的常用模型是仅含自身滞后因子的自回归模型，该模型不能有效反映通胀预期和波动性对通胀惯性的影响。本文在自回归模型基础上，构建了一个包含通胀惯性、学习型预期和波动性特征的通胀动态模型。该动态模型从均值和波动项两个方面反映了我国通胀水平的动态变化趋势和特征。针对上述模型不能有效反映通胀状态体制变化的缺点，本文还引入Markov机制转换模型来测度我国通胀水平的状态转移特征。利用分位数回归方法下的自回归模型、本文构建的通胀动态模型以及Markov机制转化模型实证研究了我国通胀的动态特征。分析表明：我国通胀水平具有较强的惯性特征；通胀惯性的形成机理比较复杂，学习型预期只能部分解释通胀惯性；通胀水平的状态转移时间比较长；通胀水平与其波动性有着正向的关系。
현유측도통창관성적상용모형시부함자신체후인자적자회귀모형，해모형불능유효반영통창예기화파동성대통창관성적영향。본문재자회귀모형기출상，구건료일개포함통창관성、학습형예기화파동성특정적통창동태모형。해동태모형종균치화파동항량개방면반영료아국통창수평적동태변화추세화특정。침대상술모형불능유효반영통창상태체제변화적결점，본문환인입Markov궤제전환모형래측도아국통창수평적상태전이특정。이용분위수회귀방법하적자회귀모형、본문구건적통창동태모형이급Markov궤제전화모형실증연구료아국통창적동태특정。분석표명：아국통창수평구유교강적관성특정；통창관성적형성궤리비교복잡，학습형예기지능부분해석통창관성；통창수평적상태전이시간비교장；통창수평여기파동성유착정향적관계。
Autoregressive model only containing own lag factor is commonly used to measure inflation persistence, but the model can not effectively reflect the influences of the inflation expectation and volatility on the inflation persistence. Based on an autoregressive model, a dynamic model of inflation containing inflation persistence, learning expectation and fluctuating characteristic is built. The dynamic model reflects the changing trends and dynamic characteristics of the inflation level from the mean and volatility. For the above models can not effectively reflect the institutional change of inflation, this paper also introduces markov switching model to measure the state transition characteristics of the inflation level. Empirical studies of the dynamics of inflation in China are given respectively based on the autoregressive model under quantile regression, the dynamic model built in the paper and the markov transformation mechanism model. Analysis shows that： Chinese inflation has strong persistence characteristics; The formatting mechanism of inflation persistence is complex in China, and learning expectations can only partially explain inflation persistence; The state switching time for the inflation level is long; Chinese inflation and its volatility has a positive relationship. Finally, some specific recommendations are given.