DSGE下的中国货币政策非线性特征研究

发布时间：2018-04-09 18:09

本文选题：泰勒规则　切入点：Markov机制转换DSGE模型　出处：《华中科技大学》2012年硕士论文

【摘要】：央行进行货币调控遵循一定的规则，而不是相机抉择，既有利于抑制物价的过快上涨，同时有利于经济的平稳快速增长。目前各国央行普遍采用的是利率规则和货币供应量规则，也就是利率与货币供应量随着通胀缺口与产出缺口而调整。比较典型的有泰勒规则与麦克勒姆规则。本文以泰勒规则为例，认为传统的货币规则暗含的假设是央行的利率调整是一个线性的函数，这与人们的经济观察不相符合。相反，央行根据通胀缺口与产出缺口进行调整并不是稳定不变的，而是呈现一定的非线性特征和结构性变化。通过非线性的货币政策方程描述中国的货币政策，有利于市场主体形成正确的政策预期，也有利于央行更好的发挥宏观经济调控的作用。本文假设两种不同的新凯恩斯DSGE模型。一种是将货币政策设为线性的泰勒规则，另一种则将其设为Markov机制转换的。在估计后一种模型时，我们将转移概率嵌入到机制转换模型当中，从而巧妙的将非线性的理性预期方程转化为线性的理性预期方程，并运用基于随机游走MH抽样的MCMC方法进行参数估计。同时，，利用边缘似然函数证明后者有更优的数据拟合度，说明我国的货币政策存在两种不同的机制，即存在“积极”与“消极”两种状态。
[Abstract]:The central bank's monetary regulation follows certain rules, not the discretion. It is not only conducive to curbing the excessive price rise, but also conducive to steady and rapid economic growth.At present, central banks generally adopt interest rate rules and money supply rules, that is, interest rates and money supply adjust with inflation gap and output gap.The typical ones are Taylor rule and McLum rule.Taking the Taylor rule as an example, this paper holds that the implicit assumption of the traditional monetary rules is that the interest rate adjustment of the central bank is a linear function, which is inconsistent with the economic observation.On the contrary, the central bank's adjustment based on the inflation gap and output gap is not stable, but shows certain nonlinear characteristics and structural changes.Describing China's monetary policy through nonlinear monetary policy equation is conducive to the formation of correct policy expectations by the market main body, and also to the central bank's role in macroeconomic regulation and control.This paper assumes two different new Keynesian DSGE models.One is to set monetary policy as linear Taylor rule, the other is to set it as Markov mechanism conversion.In estimating the latter model, we embed the transition probability into the mechanism transformation model, so that the nonlinear rational expectation equation can be transformed into the linear rational expectation equation.The MCMC method based on random walk MH sampling is used to estimate the parameters.At the same time, the edge likelihood function is used to prove that the latter has better data fit degree, which shows that there are two different mechanisms of monetary policy in China, that is, "positive" and "negative".
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F822.0

【参考文献】