基于LSTAR模型的中国经济周期非线性特征分析

Nonlinear Characteristics of Chinese Business Cycle-Based on LSTAR Model

已有关于我国经济周期非线性特征研究中所选均为GDP等单变量数据,而宏观经济景气一致指数(CI)能够更全面地反映经济周期的特征。所以,本文针对我国1991年1月至2012年12月的月度宏观景气一致指数,通过检验,分别建立了两机制和三机制LSTAR模型,探讨了我国经济周期波动的非对称性和持续性以及经济在各个波动阶段之间转换的内在演化机理。实证研究表明,把经济周期阶段划分为紧缩、恢复和扩张三个机制。与划分为紧缩和扩张的两体制相比,在整体拟合效果和对经济增长结构的解释能力方面都有显著提高。因此,本文为我国经济周期的划分和非线性特征分析提供了更科学的方法。

Usually, single variable is used in analysis of nonlinear characteristic of Chinese business cycle, but Coincident Index（CI） can more fully reflect the characteristics of the business cycle. So, this paper, using Chinese Coincident Index between January 1991 and December 1991, respectively, sets up two and three mechanism of LSTAR model, discussing the asymmetry of Chinese business cycle fluctuation and sustainability, and economic in the internal evolution mechanism of the transformation between the fluctuation phase. Empirical studies show that the economic cycle phase is divided into three： contraction, recovery and expansion mechanism. Compared with the division into contraction and recession regimes, the overall fitting effect and the explanation power structure on economic growth have improved significantly. Therefore, the paper provides a more comprehensive and more scientific method for the division of Chinese business cycle and nonlinear characteristics.