中国核心通货膨胀的估计——基于贝叶斯Gibbs Sampler状态空间模型  

The Estimation of China's Core Inflation:Based on the Bayesian Gibbs Sampler State-Space Model

作　　者：汤丹[1] 赵昕东[1] 

机构地区：[1]华侨大学数量经济研究院,邮政编码362021

出　　处：《经济评论》2011年第5期39-46,共8页Economic Review

基　　金：国家社科基金重点项目"转变经济增长方式的理论与政策研究"(项目编号:10AJL004);教育部人文社科青年基金项目"我国通货膨胀持久性及其宏观经济效应研究"(项目编号:10YJC790221);福建省自然科学基金"供给冲击对我国价格水平的动态影响研究"(项目编号:2009J01312);华侨大学"中央高校基本科研业务费"国家自然科学基金培育计划专项项目(项目编号:JB-ZR1135)的资助

摘　　要：观测到的通货膨胀可以分解为趋势成分和暂时成分,其中的趋势成分即核心通货膨胀。准确度量核心通货膨胀对宏观经济政策的制定和经济形势的判断有重要意义。本文首先建立了估计核心通货膨胀的状态空间模型,然后将贝叶斯Gibbs Sampler方法应用于估计该状态空间模型的参数,以克服卡尔曼滤波的缺陷。本文估计了1991-2010年的核心CPI,结果表明估计的核心CPI很好地反映了货币政策的变化,同CPI相比,核心CPI有较小的波动性并且与货币供给增长率具有更强的相关性。由此得出结论,当前只要食品价格不出现持续的大幅上涨,同时继续保持稳健的货币政策,中国就不会出现严重的通货膨胀,即使个别食品价格上涨造成CPI上涨也必然是不可持续的。The observed inflation can be decomposed into trend component and temporary component, with the trend being the core inflation. Measuring core inflation precisely is very important to the judgment of economic situation and adpotion of the macroeconomic policies. In this paper we firstly establish a state space model of core inflation, and then apply the Bayesian Gibbs Sampler to the estimation of the state space model to overcome the drawback of kalman filter. We estimate China＇ s core inflation from 1991 to 2010 and the results show that the estimated core inflation reflects the fluctuation of money supply very well. Compared with the CPl,the core CPI has lower volatility and has much stronger correlation with the growth rate of money supply. Finally we analyzed the trend of future price. As long as the food price does not rise continually and substantially, and meanwhile the monetary policy keeps moderate, the increase of CP] caused by the increase of one or two commodity price will not be persistent.

关 键 词：贝叶斯Gibbs SAMPLER 状态空间模型 核心通货膨胀 

分 类 号：F822.5[经济管理—财政学] F224