期望损失的后验分析

发布时间：2018-02-13 11:55

本文关键词： 风险管理期望损失后验分析尾部风险在险价值　出处：《财经研究》2017年12期 　论文类型：期刊论文

【摘要】：巴塞尔委员会已经批准用期望损失(Expected Shortfall,ES)作为市场风险指标对银行业进行监管,以替代现有的在险价值指标(Value-at-Risk,VaR)。这主要是因为期望损失满足风险度量应该满足的性质,而在险价值则不满足。在这个转变过程中,金融机构面临的主要困难是没有工具可以用来评估期望损失模型,即后验分析,而本文则提出了一套简单的ES后验分析工具。具体而言,我们基于累积碰撞序列(Cumulative Violations)构造了一套ES模型的检验方法,该方法是对VaR后验分析的自然推广。我们推导出了该检验的渐进分布性质,并且通过蒙特卡洛模拟研究了它们的有限样本表现。通过将该方法应用于全球三大股票指数,我们发现,在2007年金融危机期间,VaR模型无法识别出金融危机时期与正常时期市场风险模型的差异,而ES模型则可以。
[Abstract]:The Basel Committee has approved the use of expected loss expected shortest fallout ESs as a market risk indicator to replace the existing value at risk indicator, Value-at-Riskfaire VaR.This is mainly due to the nature of risk measurement to be satisfied by the expected loss. The main difficulty faced by financial institutions in this transition is that they do not have the tools to evaluate the expected loss model, a posteriori analysis. In this paper, a simple es posteriori analysis tool is proposed. In particular, we construct a test method of es model based on cumulative collision sequence. This method is a natural generalization of the VaR posteriori analysis. We derive the asymptotic distribution properties of the test, and study their finite sample performance by Monte Carlo simulation. We find that in 2007, the VaR model can not identify the difference between the financial crisis period and the normal period market risk model, but es model can.
【作者单位】：西南财经大学经济与管理研究院;美国印第安纳大学文理学院经济系;
【基金】：国家自然科学基金(71401140) the Spanish Plan Nacional de I+D+I(ECO2014-55858-P)
【分类号】：F831.51

【相似文献】