摘要
以2002年1月至2017年12月中国银行间国债到期收益率数据为样本,采用半参数局部Whittle估计的方法对国债收益率和收益率利差的长记忆性进行研究。结果表明,国债收益率和收益率利差均存在一定程度的长记忆性。大多数国债收益率不是协方差平稳的但存在均值回归特性。收益率利差基本上是协方差平稳的但相比I(0)模型存在一定的持续性。本文建议在进行投资决策和收益率预测时应当考虑国债收益率和收益率利差的长记忆性。
This paper analyzes the long memory of the Chinese yield rates and yield spreads using the semiparametric local Whittle approach. Monthly yield data from January 2002 to December 2017 are used. This paper finds that both the yield rates and yield spreads exhibit a certain degree of persistence. Most of the yield rates are not covariance stationary but still mean-reverting. The yield spreads are basically covariance stationary but more persistent than the I (0) model. This paper suggests that the long memory feature should he accounted for in investing decisions and forecasting models.
作者
冯攀
FENG Pan(Antai College of Economics and Management,Shanghai Jiao Tong University,Shanghai 200240,Chin)
出处
《系统工程》
CSSCI
北大核心
2018年第1期39-46,共8页
Systems Engineering
基金
国家自然科学基金资助项目(71673183)