多元Copula-ACD模型及其应用

发布时间：2018-02-04 21:19

本文关键词： 交易持续时间截面相关多元Copula-ACD模型 Gamma分布溢出效应　出处：《中国管理科学》2014年S1期 　论文类型：期刊论文

【摘要】：自回归条件持续时间(ACD)模型在金融经济中的作用日益明显,研究成果日益丰富。本文将Copula方法与ACD模型有机结合,运用Copula理论建立多元Copula-ACD模型,描述交易持续时间的时变条件相关关系,在一定程度上缓解了向量ACD模型参数难估计与非同步交易的问题。基于上证50指数中四只银行成分股的实证分析表明多元Copula-ACD模型能够较好捕捉中国证券市场上交易持续时间的聚集结构特征,给出交易持续时间的联合密度函数,估计各持续时间变量之间的自相关性与截面相关性,进一步描述和检验多元持续时间的溢出效应,为投资者提供决策参考。
[Abstract]:The role of autoregressive conditional duration (ACD) model in financial economy is becoming more and more obvious, and the research results are increasingly rich. In this paper, the Copula method and ACD model are combined organically. The multivariate Copula-ACD model is established by using Copula theory to describe the time-varying conditional correlation of transaction duration. To some extent, the problem of parameter estimation and asynchronous trading of vector ACD model is alleviated. Based on empirical analysis of four bank constituent stocks in Shanghai 50 Index, the multiple Copula-ACD model can be used. It is better to capture the aggregation structure characteristics of the trading duration in China's securities market. The joint density function of transaction duration is given, the autocorrelation and cross-section correlation between each duration variable are estimated, and the spillover effect of multiple durations is further described and tested, which provides a decision reference for investors.
【作者单位】：西南财经大学统计学院;
【基金】：国家自然科学基金资助项目(71101118) 教育部新世纪优秀人才支持计划资助项目(NCET-13-0961)
【分类号】：F830;F224
【正文快照】： ,3|_ 1 C I=1=1 交易过程中实时交易的金融数据通常被称为超高频金融数据或分笔交易数据。分笔交易数据处理的首要难题是交易在随机时刻发生,具有不等间隔特性,建立在相同时间间隔观测基础上的计量经济模型(如ARCH、SV模型)并不适用。目前,对于不等间隔交易数据的计量分析是从

【共引文献】