基于动态Probit模型的通胀周期预测研究

Abstract

使用1990年以后的CPI序列建立了中国通胀周期图，将通胀周期划分为通胀期和通缩期，周期的分析与预测转化为二元时间序列的分析和预测问题。实证结果表明：1991年以来中国通胀共经历了4个完整的周期，当前处于第5个周期之中；存款利率的期限结构、通胀预期、货币增速对通胀周期有显著影响；GDP增速和汽油价格增速对通胀周期无显著影响；动态Probit模型具有较强的通胀趋势预测能力。On the basis of establishment of Chinese inflation cycle with CPI series since 1990, inflation cycle is classified as expansion period and recession period. The cycle analysis and prediction conversion are the analysis and prediction of bivariate time series. Empirical studies show that China has experienced four inflation cycles since 1991, currently in the fifth cycle. The term structure of de-posit interest rate, inflation expectation and speed of currency increase have significant effect on China’s inflation cycle; GDP increment rate, controlled gasoline price have no significant effect on the cycle. The dynamic Probit model has strong predicting abilities.