基于DSGE模型的中国经济周期波动成因分析及其宏观经济变量冲击效应检验

发布时间：2018-01-18 12:16

本文关键词：基于DSGE模型的中国经济周期波动成因分析及其宏观经济变量冲击效应检验　出处：《吉林大学》2014年硕士论文　论文类型：学位论文

【摘要】：经济增长和周期波动是宏观经济学研究的主要领域，许多学者对我国经济周期波动的特征和原因运用许多计量方法进行了分析和检验。经济周期波动问题不仅发生在西方发达国家，我国同样经历着频繁的经济周期波动。国民经济的各个方面都存在波动问题，总产出的波动、投资的波动、就业的波动等等。这些不仅影响经济的稳定与增长并且与每个公民的福利息息相关。众多学者对我国经济波动的原因展开广泛的研究，这也是本文研究的重点。经济周期波动的研究中，时间序列中波动成分的提取一直是研究的重点。随着RBC理论的出现与发展，滤波方法在时间序列的分解中得到广泛的应用。由于各种滤波对分离时间序列的波动成分的差别很小，，本文选取HP滤波处理我国经济时间序列，分离出各个宏观经济变量中的周期成分和波动成分。通过对这些成分进行统计分析，总结出我国经济波动的典型事实：固定资产投资、政府消费与净出口、工资水平、货币余额、价格水平的波动性波动性大于总产出，总消费、就业人数波动性小于总产出；固定资产投资、工资、货币余额、价格水平、总消费具有顺周期性，而就业人数、政府消费与净出口是逆周期波动。现代宏观经济学认为，外生随机冲击导致经济的周期波动。识别波动的来源是研究经济波动的关键。技术冲击、货币政策冲击、偏好冲击等等被认为是经济周期波动的根源，不同的理论有着不同的观点。本文在动态随机一般均衡模型的框架下研究我国经济周期波动的成因，模型参数使用贝叶斯方法进行估击，使用卡尔曼平滑算子估计模型冲击的实现值。通过将冲击的实现值带回原模型来考察经济系统的内生变量对各个冲击的反映情况。结果表明，技术冲击和投资冲击是我国经济周期波动的主要原因，政府支出和净出口对波动的影响较小。最后，本文利用DAG方法对我国宏观经济变量的因果关系进行检验。与传统的方法相比，DAG不仅可以同时识别多个变量间的因果关系并且可以限制VAR模型残差项的过度识别。通过PC算法可以识别出隐藏在残差项协方差矩阵中的变量同期因果关系。结果表明，投资与货币供给量是我国经济增长的同期原因，价格水平的变化主要受到货币供给量的影响。
[Abstract]:Economic growth and periodic fluctuations are the main areas of macroeconomic research. Many scholars have analyzed and tested the characteristics and causes of economic cycle fluctuations in China by using many econometric methods. The problem of business cycle fluctuations not only occurs in western developed countries. China is also experiencing frequent economic cycle fluctuations. There are fluctuations in all aspects of the national economy, total output fluctuations, investment fluctuations. The fluctuation of employment and so on. These not only affect the stability and growth of economy, but also are closely related to the welfare of every citizen. Many scholars have carried out extensive research on the causes of economic fluctuation in our country. This is also the focus of this paper. In the research of business cycle fluctuation, the extraction of fluctuation component from time series has been the focus of the research. With the emergence and development of RBC theory. The filtering method has been widely used in the decomposition of time series. Due to the small difference of the wave components in the separation of time series, HP filter is selected to deal with the economic time series in China. Through the statistical analysis of these components, the typical facts of China's economic fluctuations are summarized: fixed asset investment, government consumption and net exports. The volatility of wage level, monetary balance and price level is greater than that of total output, and the volatility of total consumption and employment is smaller than that of total output. Fixed asset investment, wages, monetary balance, price level, total consumption are pro-cyclical, while employment, government consumption and net exports are counter-cyclical fluctuations. In modern macroeconomics, exogenous random shocks lead to periodic fluctuations of the economy. Identifying the sources of fluctuations is the key to the study of economic fluctuations. Technology shocks, monetary policy shocks. Preference shocks and so on are considered to be the root of business cycle fluctuations, and different theories have different views. This paper studies the causes of business cycle fluctuations in China under the framework of dynamic stochastic general equilibrium model. The model parameters are evaluated by Bayesian method. Kalman smoothing operator is used to estimate the realized value of the model shock. By bringing the realized value of the shock back to the original model, the reflection of the endogenous variables of the economic system to each shock is investigated. The results show that. Technology shock and investment shock are the main reasons for the fluctuation of economic cycle in China. Government expenditure and net export have little effect on the fluctuation. Finally, this paper uses the DAG method to test the causality of macroeconomic variables in China, compared with the traditional method. DAG can not only identify causality among multiple variables at the same time, but also limit the over-recognition of residual terms in VAR model. By using PC algorithm, we can identify the covariance factors of variables hidden in the covariance matrix of residual terms. Fruit relationship. Results show. Investment and money supply are the causes of economic growth in China in the same period, and the change of price level is mainly affected by the amount of money supply.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F124.8

【参考文献】