人民币短期利率行为研究方法的一个改进——双指数Jump-GARCH-Vasicek模型的构建与应用  

An Improvement of Research Methods on the Behavior of RMB Short-term Interest Rates: Construction and Application of Double Exponential Jump-GARCH-Vasicek Model

作　　者：谢赤[1,2] 张娇艳[1] 王纲金[1,2] 余聪[1] 

机构地区：[1]湖南大学工商管理学院,湖南长沙410082 [2]湖南大学金融与投资管理研究中心,湖南长沙410082

出　　处：《运筹与管理》2014年第5期198-204,共7页Operations Research and Management Science

基　　金：国家自然科学基金创新研究群体基金项目(71221001);国家软科学研究计划项目(2010GXS5B141);教育部创新群体项目(IRT0916);教育部人文社会科学规划项目(09YJC630063);湖南省自然科学基金创新群体项目(09JJ7002)

摘　　要：受货币政策调控频率提升及大型新股申购等因素的影响,近年来人民币短期利率表现出明显的跳跃行为。为了更准确地描述利率跳跃行为,本文通过假设跳跃幅度服从双指数分布构建一个能刻画短期利率波动聚类、均值回复和跳跃行为的双指数Jump-GARCH-Vasicek模型。利用人民币短期利率数据,将双指数JumpGARCH-Vasicek模型与Vasicek模型、GARCH-Vasicek模型、正态Jump-Vasicek模型、双指数Jump-Vasicek模型、正态Jump-GARCH-Vasicek模型进行实证对比分析。研究结果表明,人民币短期利率确实存在GARCH效应、均值回复和跳跃行为,且双指数Jump-GARCH-Vasicek模型较其它模型能更好地刻画人民币短期利率的跳跃行为。Affected by factors such as monetary policies made by the people＆#39;s bank of China and large-scale new shares subscriptions , the RMB short-term interest rates showed obvious jump behavior in recent years .In order to better describe the jump behavior of interest rates , we construct a Jump-GARCH-Vasicek model which can de-scribe the volatility clustering , mean reversion and jump behavior of Chinese short-term interest rates by assuming the distribution of jump size to be double exponential .Based on the datum of Chinese short-term interest rates, we compare double exponential Jump-GARCH-Vasicek model with the following five models： Vasicek model , GARCH-Vasicek model , normal Jump-Vasicek model , double exponential Jump-Vasicek model and normal Jump-GARCH-Vasicek model .We find that short-term interest rates show obvious GARCH effect , mean rever-sion and jump behavior , and the double exponential Jump-GARCH-Vasicek model is the best of the six models in describing the jump behavior of short-term interest rates .

关 键 词：金融工程 双指数Jump-GARCH-Vasicek模型 极大似然估计 人民币短期利率 跳跃行为 

分 类 号：F830.9[经济管理—金融学]