中国期货市场价格久期波动聚类特征研究  

Research on China futures market price duration volatility cluster effect

作　　者：刘向丽[1] 程刚[2] 成思危[3] 汪寿阳[3] 洪永淼 

机构地区：[1]中央财经大学金融学院,北京100081 [2]香港城市大学经济与金融学系 [3]中国科学院数学与系统科学研究院,北京100190 [4]美国康乃尔大学经济学系与统计科学系,纽约14850

出　　处：《管理科学学报》2010年第5期72-81,共10页Journal of Management Sciences in China

基　　金：中央财经大学211工程三期重点学科建设资助项目

摘　　要：分别在四种残差分布假设下对四种ACD模型进行参数估计,通过检验模型的性能,分析适合我国期货市场的ACD模型及残差的分布.并以此为基础,在价格久期模型中,分别加入久期内平均交易量、久期内平均绝对收益率和久期时点处的持仓量这三个微观结构因子,据此分析交易强度、价格波动和市场深度对价格久期的影响.实证结果表明:复杂的ACD模型并不能显著提高模型拟合能力,但各残差分布间差异较大.久期内的平均交易量、平均绝对收益率对价格久期都有显著的负向作用,而持仓量对其有微弱的影响.引入微观结构变量的扩展模型比封闭的模型表现更好.By introducing four ACD models and four types of distribution of residuals, the paper aims to find the appropriate ACD model of China futures market and the appropriate distribution of residuals. Furthermore, average volume in duration, average absolute yield in duration and open interest of the time point are added in the previous ACD models to analyze the impact of trade intensity, price volatility and market depth on price duration. The empirical study shows that the complicated models can not outperform the simple ones to describe the durations, but different choices of distribution of residuals have a significant impact on the deseription ability. The average volumeand average absolute yield in durations have negative impacts on price durations, while open interest has weak impacts. The extended ACD models perform better than the classical ones.

关 键 词：久期 ACD模型 微观结构 聚类效应 

分 类 号：F830.9[经济管理—金融学]