基于DMA方法的通货膨胀预测研究  

The Study of Inflation Forecasting Using Dynamic Model Averaging

作　　者：朱培金[1] 

机构地区：[1]中国人民银行杭州中心支行,杭州310001

出　　处：《统计与决策》2017年第2期9-13,共5页Statistics & Decision

摘　　要：文章从非线性时变建模角度入手,利用DMA模型方法对通货膨胀的动态预测进行了深入细致的实证分析。研究发现:DMA模型方法估计的时变参数充分反映出了近二十年我国经济结构的变迁历程;基于模型预测,我国未来面临通缩压力;虽然预测期限不同,各类指标变量对通胀的贡献率不同,但是货币供应量增长率对通胀的解释力度有所下降,而利率在预测通胀作用方面有所增强。From the perspective of Non-linear Time-variant Modeling, this paper has a depth and detailed empirical analysis on the prediction of inflation based on DMA Model. Three conclusions are drawn from the study. Firstly, the DMA Model fully reeflects the courses of inflation in nearly two decades of China＇s economic structure changes. Secondly, the DMA Model forecasts fu- ture deflationary pressures. Thirdly, the growth rate of money supply has a declined influence on inflation, while the interest rate has became an enhanced role in inflation prediction.

关 键 词：通货膨胀预测 菲利普斯曲线 货币政策 时变性 DMA模型 

分 类 号：F820.5[经济管理—财政学]