汪玲玲,白仲林.估计MS-DSGE模型脉冲响应函数的EM算法及其应用[J].数量经济技术经济研究,2017,(2):121-138 |
估计MS-DSGE模型脉冲响应函数的EM算法及其应用 |
EM Algorithm for Estimation of MS-DSGE Model's IRF and Its Application |
|
DOI: |
中文关键词: 新凯恩斯 MS-DSGE 脉冲响应函数 动态因子模型 EM算法 |
英文关键词: New Keynesian MS-DSGE Impulsive Response Function Dynamic Factor Model EM Algorithm |
基金项目:本文获得国家自然基金“具有Markov区制转换的动态因子模型建模方法及其应用研究”(71271142)、天津财经大学研究生创新基金“DSGE模型的估计方法研究——基于动态因子模型的视角”(2014TCB04)的资助。 |
|
中文摘要: |
研究目标:考察不同区制下外生冲击对中国宏观经济的非对称性效应。研究方法:引入两状态的Markov区制转换过程建立MS-DSGE模型,并基于MS-DSGE模型的Markov区制转换动态因子模型的表示提出了估计MS-DSGE模型脉冲响应函数的极大似然估计EM算法。研究发现:本文提出的估计方法具有良好的有限样本性质和收敛性,参数估计量具有渐近正态分布。实证分析发现,应持续施行扩张性政策以刺激经济稳定增长,对冲挤占效应以及稳定物价水平。尤其,当经济处于“衰退”区制时,政府应实施及时有效的调控政策刺激经济运行区制的转移。研究创新:与Bayesian分析方法比较,本文提出的估计方法避免了对数线性化MS-DSGE模型的随机奇异性以及对先验分布的设定和观测变量选取的非稳健性。研究价值:提出了一种估计MS-DSGE模型脉冲响应函数的方法。 |
英文摘要: |
Research Objectives:Inspecting the asymmetric influence of external shocks on China's macro-economy in different regimes. Research Methods:Introducing Markov-switching process with two regimes to establish MS-DSGE model and proposing the ML estimation of MS-DSGE models' IRF through EM algorithm based on the representation of DFMs' with Markov-switching process of MS-DSGE models. Research Findings:This paper's estimation method has the good finite sample property and convergence property, and parameter estimator has asymptotic normal distribution. Through the empirical study we conclude that Chinese government must stick to expansive policies so as to stimulate steady growth which can offset these policies' squeeze-out effect and stabilize prices. Especially when the economy is in a recession system, Chinese government should carry out timely and effective macro-control policy to stimulate the transfer of the economic operation system. Research Innovations:Comparing with Bayesian analysis method, this paper's method has avoided the stochastic singularity of logarithmic linearization MS-DSGE models, the non-robustness of the setting of the prior distribution and the selection of observed variables. Research Value:Proposing an estimation method of MS-DSGE models' IRF |
查看全文 |