CVaR-EVT和BMM在极端金融风险管理中的应用研究

发布时间：2018-08-25 10:35

【摘要】：随着风险度量一致性原则的提出,研究发现金融机构广泛采用的VaR模型存在严重不足,尤其针对分布具有厚尾特征的极端金融风险无法有效度量。本文采用极值理论(EVT)解决VaR方法的尾部度量不足问题,利用CVaR-EVT和BMM模型分析美国、香港股票市场和我国沪深两市指数18年的日收益数据,研究发现:(1)在95%置信区间及点估计中,分位数为99%的CVaR-EVT所揭示的极端风险优于VaR的估计值,且BMM方法为实施长期极端风险管理提供了有力的决策依据,其回报率受分段时区的影响,期间越长,风险估计值越高;(2)模型采用ML和BS方法统计估值显示,我国股票市场极端风险尾部估计值高于香港和美国市场,但是,国内市场逐步稳定,并呈现出跟进国际市场且差距缩小的发展趋势。
[Abstract]:With the development of the consistency principle of risk measurement, it is found that the VaR model widely used by financial institutions has serious shortcomings, especially for the extreme financial risk with thick tail distribution can not be effectively measured. In this paper, extreme value theory (EVT) is used to solve the problem of insufficient tail metric of VaR method. Using CVaR-EVT and BMM models to analyze the daily income data of the stock markets in the United States, Hong Kong and Shanghai and Shenzhen stock markets for 18 years. The results show that: (1) in 95% confidence interval and point estimation, The CVaR-EVT with 99% of the quantiles reveals more extreme risks than the VaR estimates, and the BMM method provides a strong basis for decision making for the implementation of long-term extreme risk management, and the longer the period is, the longer the return rate is affected by the segmented time zone. The higher the risk estimate is, the higher the risk estimate is. (2) the ML and BS methods are used to estimate the tail value of extreme risk in Chinese stock market, but the domestic market is gradually stable. And showing a follow-up to the international market and narrowing the gap in the development trend.
【作者单位】：复旦大学金融研究院;复旦大学计算机学院;复旦大学经济学院;
【基金】：国家自然科学基金项目课题“基于消费者行为分析的网上支付风险管理与监管研究”(项目批准号:70702028,主持人杨青) “上海浦江人才计划”(主持人杨青)资助
【分类号】：F830

【相似文献】