沪深股市极值相依关系测度研究——基于四类时变Copula-EVT模型  

Study on Extreme Value Dependence Measurement of Shanghai and Shenzhen Composite Index——Based on the Four Types of Time-Varying Copula Model

作　　者：于文华[1,2] 

机构地区：[1]西南交通大学经济管理学院,四川成都610031 [2]成都理工大学商学院,四川成都610059

出　　处：《西南交通大学学报（社会科学版）》2013年第4期90-96,共7页Journal of Southwest Jiaotong University(Social Sciences)

基　　金：国家自然科学基金项目(71071131;71090402);高等学校博士学科点专项科研基金资助课题(20120184110020);中央高校基本科研业务费专项资金资助项目(SWJTU11ZT30;SWJTU11CX137;SWJTU12CX120);成都理工大学研究基金项目(2011YR09);成都理工大学金融与投资科研创新团队项目(KYTD201303);四川省软科学研究计划项目(2013ZR0068)

摘　　要：随着金融市场的发展和理论研究的深入,金融资产尾部极值的相依关系已引起众多市场参与者和研究者的重视。以上证综指和深证综指为研究对象,结合EVT极值理论股指收益率构建边缘分布模型,以此为基础,运用四类时变Copula函数分别拟合股指间的联合分布,并通过AIC准则对各类模型的拟合效果进行比较,结果表明:不同模型的拟合度存在差异;沪深股市间具有较强的尾部极值相依程度,并呈现出明显的动态趋势。With the development of financial markets and theoretical researches, a lot of market participants and researchers pay attention to extreme tail dependencies of the financial assets. In this paper, models Shanghai and Shenzhen Composite Index are studied as the research object. Marginal distribution are constructed for the two indexes based on relevant theories and the Extreme Value Theory （EVT）. Then four types of time-varying Copula function are adopted to describe the joint distribution of the two index returns, and the goodness-of-fit of different models are compared through AIC. The experimental results show that every model has different fitting effect and there is strong extremal dependence with a dynamic trend between the two stock markets.

关 键 词：金融市场 金融相关性 风险管理 资产收益率 上证综指 深证综指 时变Copula函数 

分 类 号：F224[经济管理—国民经济] F832.51