基于非平稳时间序列的VAR模型实证分析

VAR Model Empirical Analysis based on nonstationary time serie

通过中国统计年鉴及互联网相关数据平台收集我国1952年至2008年国内生产总值(GDP),从业人员(E)及国内资本形成总额(GCF)的官方数据,通过使用非平稳时间序列建立向量自回归模型(VAR模型).数据处理过程主要包括两个重要分析结果:(1)根据Johanson协整检验分析变量的长期均衡关系;(2)利用误差修正模型VEC对变量间的短期关系进行修正.最终通过VAR模型的实证分析对国内生产总值,从业人员和资本形成总额间的关系做出预测.

This paper collects official data of China's GROSS Domestic Product(GDP),Employee(E)and Gross Domestic Capital Formation(GCF)from 1952 to 2008 through China Statistical Yearbook and Internet platform,vector autoregression models is established by using non-stationary time series.Two important analysis results are included in the data processing procedure:one is to analyze the long-term equilibrium relationship of variables according to the Johanson co-integration test;the other is to use the vector error correction model VECM to modify the short-term relationship between variables.Finally,the relationship between GDP,E and GCF could be predicted through empirical analysis of VAR model.