本文对我国不同层次的货币供应动态过程进行建模研究,通过贝叶斯Gibbs抽样方法估计t分布先验设定下动态线性模型参数和状态变量后验均值,以甄别模型中观测和状态过程均可能存在的异常值和结构突变特征。结果表明研究区间内流通中现金M0和狭义货币供应量M1序列均发生结构突变,而广义货币供应量M2受2008年全球金融危机影响也出现异常变动。最后在结构变化的原因分析基础上提出了相关政策建议。
We modeled dynamic process of different levels of money supply in China, through the Bayesian Gibbs sampling method to estimate the posterior mean of of dynamic linear model parameters and state variables under a priori student t-distribution set, outlier points and structure change could be screened in observation process and state process, results showed that structural breaks of cash in circulation and M0 and narrow money supply M1 within study interval were found and outlier of broad money supply M2 because of 2008 global financial crisis. The relevant policy advice was proposed on the basis of analyzing causes of structural changes.