一种估计DSGE模型脉冲响应函数的方法

An Approach of Estimation for Impulse Response Functions of DSGE Model

为了克服DSGE模型Bayesian分析方法的观测变量依赖性和避免状态空间模型表示的随机奇异性,在新凯恩斯DSGE模型的动态因子模型表示基础上,本文借助动态因子模型为DSGE模型的脉冲响应函数提供了一种估计途径。首先,本文讨论了这类特殊动态因子模型的识别性;其次,提出估计动态因子模型的二阶段估计方法,证明了估计量的一致性和渐近正态性;并且通过Monte Carlo模拟方法讨论了这种估计方法的有限样本性质。另外,实证分析发现,目前政府应持续实施诸如结构性减税、鼓励技术创新等促进技术进步的政策,积极规制市场垄断行为,以及引导行业发展促进消费偏好迁移的宏观调控政策,以刺激宏观经济持续稳定增长,平衡经济刺激政策对物价水平的影响。

In order to overcome observation variable dependence of the Bayesian analysis method of DSGE model and to avoid the stochastic singularity of the state space model repre- sentation, this paper suggests an estimation approach for impulse response functions of DSGE model, basing on the DFM representation of DSGE model. Firstly, the identifiability of this special DFM has been discussed; secondly, a two-stage estimation process of DFM has been proposed, the consistency and asymptotical normality of the estimators has been proved; and the finite sample property of this approach has been discussed by the Monte Carlo simulation. Moreover, the empirical analysis in this paper has discovered that the government should continue to implement the policies, such as cutting the structural tax, encouraging technology innovation and so on tion of the market monopoly and guiding transformation of consuming preference.