非奇次空间动态极值估计的模型与应用

The Model and Application of the Inhomogeneous Time Dynamic Extreme Value Theory

传统EVT方法是从静态的角度,研究超额数据的性质。然而,它没有同时考虑极端数据发生的时间所隐含的充分信息。本文首次在国内提出了非奇次空间动态极值理论(ITD-EVT)的概念,克服了EVT的上述缺陷,在极端数据的基础上考虑了时间因素,并引入多个解释变量,使极值分布的是三个参数为时变的,用二维泊松分布过程建立动态空间模型,是文中一大特色。把TD-EVT运用于极端情况下风险值的估计中,对金融风险管理、资产定价等问题有较大的理论和现实意义。

The number of financial crisis has been growing since the 1990s. A country＇s financial crisis can eventually lead to international financial crisis in the global economy. The abnormal fluctuations of asset prices are becoming an important financial risk management topic. Extreme Value Theory method （EVT） is the most suitable method used to calculate the tail risk of financial assets in the extreme and unusual circumstances. As can be seen from the domestic and international research on EVT used in risk assessment, most literature studies the nature of obtaining exceedances over the threshold from the whole sample.