统计研究

• 论文 • 上一篇    下一篇

中国市场国债利率期限结构的动态特征研究

丁志国等   

  • 出版日期:2016-01-15 发布日期:2016-01-22

Dynamic characteristics of interest term structure of China’s national bonds market: An empirical research based on the Vasicek-AR(1) process model

Ding Zhiguo et.al   

  • Online:2016-01-15 Published:2016-01-22

摘要: 本文在分析中国市场国债收益率的统计特征,并采用Vasicek模型对其进行动态检验的基础上,将状态变量的变动过程分别设定为AR(1)、AR(2)、ARMA(1,1)以及随机游走四种方式,比较上述四种变动方式的拟合效果与Vasicek模型的拟合效果。实证结果表明,将消费者信心指数变化率、M2增长率的自然对数、生产者价格指数变化率(PPI)以及中国成镇居民失业率等四个状态变量的变动方式设定为AR(1)过程的修正Vasicek模型,无论是对国债收益率数据的拟合还是预测效果都显著优于Vasicek模型,说明基于AR(1)过程的修正Vasicek模型能够更加准确地刻画中国市场国债利率期限结构的动态特征。

关键词: 仿射模型, 利率期限结构, Vasicek-AR(1)模型, Vasicek模型

Abstract: On the basis of statistical analysis of interest rate on China’s national bonds market return rate and the dynamic test results of Vasicek model, this paper sets the changing process of state variables as AR(1),AR(2),ARMA(1,1) and Random Walk types and mainly compares the imitative effects with the Vasicek model. It is found that the Vasicek model, with the changing process of CCI, Natural logarithm of M2, PPI and China's urban unemployment rate setting as AR(1) , has a better performance on both predicting and fitting the return interest rate of national bonds than the Vasicek model. It means Vasicek-AR(1)model could more precisely describes the dynamic characteristics of interest term structure of China’s national bonds market.

Key words: Affine Mmodel, Tterm Sstructure of Iinterest Rrate, Vasicek-AR(1) Mmodel, Vasicek Mmodel