时间：2019年05月20日（星期一）14：00-15：00

地点：学院南路，图配楼 514

报告题目： Semi-parametric Dynamic Max-copulaModel for Multivariate Time Series

报告人：张正军，University of Wisconsin

报告摘要：This paper presents a novelnonlinear framework for the construction of flexible multivariate dependencestructure (i.e., copula) from existing copulas based on a straightforward ``pairwisemax'' rule. The newly constructed max-copula has a closed form and has strong interpretability.Compared to the classical ``linear symmetric'' mixture copula, the max-copulacan be viewed as a ``non-linear asymmetric'' framework. It is capable ofmodeling asymmetric dependence and joint tail behavior while also offering goodperformance in non-extremal behavior modeling. Max-copulas that are based onsingle-factor and block-factor models are developed to offer parsimoniousmodeling for structured dependence, especially in high-dimensionalapplications. Combined with semi-parametric time series models, the max-copulacan be used to develop flexible and accurate models for multivariate time series.A new semi-parametric composite maximum likelihood method is proposed forparameter estimation, where the consistency and asymptotic normality ofestimators are established. The flexibility of the max-copula and the accuracyof the proposed estimation procedure are illustrated through extensive numericalexperiments. Real data applications in Value at Risk estimation and portfoliooptimization for financial risk management demonstrate the max-copula'spromising ability to accurately capture joint movements of high-dimensionalmultivariate stock returns under both normal and crisis regime of the financialmarket. This is a joint work with Zifeng Zhao.

报告人简介：张正军教授现为美国威斯康辛大学统计系正教授、美国统计协会会士、国际数理统计协会财务总监、国际顶级期刊“商业和经济统计”副主编、“计量经济学期刊”“金融工程与风险管理特刊”共同主编、“泛华统计学报Statistica Sinica”副主编。张正军教授2002年毕业于北卡罗来纳大学教堂山分校，获统计学博士学位。主要研究方向包括：金融时间序列分析、极值理论、异常气候分析、稀有疾病（癌症、帕金森综合症、奥兹海默病等等）分析、金融风险的建模和评估、市场系统性风险评估等等。

本次活动受中央财经大学2019专题学术讲座项目资助。