通胀惯性、通胀预期与我国通货膨胀的空间特征——基于空间动态面板模型

Persistence,Expectation and Inflation Spatial Dynamics in China:A Spatial Dynamic Panel Model Study

本文将经济地理空间权重矩阵引入通胀惯性和通胀预期共存的新凯恩斯混合菲利普斯曲线框架,通过空间动态面板模型计量方法考察我国地区通胀在时间和空间上的特征。我们通过蒙特卡罗模拟发现,使用拟最大似然函数方法估计空间动态面板模型明显优于用于估计动态面板模型的向前正交离差广义矩方法。分析发现,消费者物价指数和零售商品物价指数空间传染性较强,GDP平减指数空间相关性较弱。新凯恩斯混合菲利普斯曲线较好地拟合了我国通胀的运行机理,其中,在影响当期通胀水平上,通胀惯性明显占优于通胀预期,治理通胀必须承受必要代价。

Including the economic-geographical spatial weights matrix into the Hybrid New Keynesian Phillips Curve,we study the property of China’s regional inflation in space and time dimensions.The Monte Carlo investigation suggests that,in estimating spatial dynamic panel model,the QMLE method outshines the forward-orthogonal deviations GMM method for dynamic panel model.The empirical results indicate some conclusions.First,CPI and RPI show a strong spatial infectiousness,while the GDP deflator doesn’t.Second,the HNKPC is consistent with the inflation dynamics in China,with the inflation persistence playing a dominant role：The coefficient of lagged inflation substantially exceeds the coefficient of expected future inflation.