基于点过程的极值VaR研究

Research on Extreme VaR based on Point Process

风险无时不有、无处不在,风险本身并不可怕,金融机构就是通过承担风险、管理风险来获得收益的。真正可怕的是极值风险,即稀少或极端事件的发生给经济主体带来巨额损失的风险。因此对极值风险的建模就成为风险管理的重中之重。极值理论为极端事件的统计建模和极值风险测度的计算提供了坚实的理论基础,故有必要通过介绍和比较传统极值事件的建模,基于点过程构建极值风险的一般模型,并利用外汇市场的日数据和VaR的估计与检验进行实证分析。

Extreme risk, which is the risk brought by rare or extreme events, is one of a focus of risk management. Extreme theory supplies with substantial fundamentals to statistical modeling and risk measure of extreme events. This paper introduces and compares traditional modeling of extreme events, then extends POT model basing on point process and analyses the simulation effects using the daily data of foreign exchange market in China.