摘要
我们使用我国1996年1月至2008年6月期间的银行同业拆借利率,对我国利率均值过程及其波动过程的长期记忆性进行测度和检验。利用ARFIMA模型和FIGARCH模型的检验结果说明,我国利率序列的一阶矩中不存在长期记忆性,而二阶矩中存在显著的长期记忆性;进一步运用ARFIMA-FIGARCH模型对利率均值过程及其波动过程的双长期记忆性进行检验时发现,我国利率序列均值过程中不存在明显的长期记忆性,但波动率序列中存在非常显著且较强的长期记忆性特征;通过考虑Student-t分布进一步说明,我国利率序列中明显存在"尖峰厚尾"分布特征。
This paper studies the dual long memory of interest rate and volatility in China, by using the monthly interest rate of China from January 1996 to June 2008. On the basis of the newly proposed methods to describe the long memory, including the ARFIMA model, FIGARCH model and ARFIMA-FIGARCH model, the results reveal that there exists no long memory property in the first moment of the interest rate, but the long memory properties are significant to the second moment of the interest rate. Moreover, using the ARFIMA-FIGARCH model, we find that there exists no long memory property in interest rate, but the long memory properties are significant to the volatility of the interest rate. In consideration of the Student-t distribution, the results further prove that the interest rate has the characteristics of high peak and fat tail.
出处
《技术经济与管理研究》
2008年第6期3-7,共5页
Journal of Technical Economics & Management
基金
吉林大学"985工程"
"经济分析与预测哲学社会科学创新基地"
吉林大学"985工程"研究生创新基金重点项目(20081101)资助
