基于贝叶斯TVP-VAR模型的货币政策价格效应

Price Effect Analysis of the Monetary Policy Based on Bayesian TVP-VAR Model

鉴于核心CPI是从货币政策角度定义的,为了探究我国货币政策的价格效应在不同时点上的差异性,本文基于八类价格指数采用小波分解和动态因子模型构建了核心CPI并将其作为价格变量,之后构建了贝叶斯框架下的时变参数向量自回归(TVP-VAR)模型,测度了我国货币政策在不同时点上的价格效应。实证结果表明:第一,相对于CPI而言,本文构建的核心CPI能更好地反映物价长期、稳定的变动趋势;第二,我国货币政策对整体物价的调控具有短期效应,且在不同时点上存在显著差异。鉴于此,建议中国人民银行编制并公布核心CPI,以便更有效地检测物价整体变动水平和测度货币政策的价格效应。

Because core CPI is defined from the perspective of monetary policy, in order to explore the differences of the price effect of monetary policies at different time points in China, this paper combined eight types of price indices, wavelet decomposition and dynamic factor models to construct the core CPI as a price variable. Then the paper further constructed the time-varying parameter vector autoregression(TVP-VAR) model under the Bayesian framework, and measured the price effect of monetary policies at different time points in China. The empirical results showed that: Firstly, compared with CPI, the core CPI constructed in this paper could better reflect the long-term and stable trend of price changes. Secondly, China’s monetary policy had a short-term effect on the regulation of overall prices, and there were significant differences at different time points. Therefore, it was recommended that the People’s Bank of China should compile and publish the core CPI in order to more effectively detect the overall level of price changes and measure the price effect of monetary policy.