机构地区: 太原理工大学经济管理学院
出 处: 《商业研究》 2011年第3期212-216,共5页
摘 要: 目前VaR模型是国际上通行的综合衡量和管理风险的计量技术,但当分布是非正态分布或是不连续的时候VaR就没有稳定性。在应用VaR以及CVaR的基本原理,利用具有尖峰后尾特征的对数正态分布并结合贝叶斯风险,对参数通过先验分布和后验分布进行修正得到市场风险资产CVaR以及VaR的计算公式,对其进行比较分析,将其与pareto分布下的市场风险VaR值进行比较。研究表明:对数正态分布下市场风险资产的损失率比损失率呈pareto分布的风险损失显现出更好的拟合特征,能够描述市场风险资产的随机性最大损失,以及CvaR计量下的市场风险资产损失比VaR计量下的有更好的尾部特征,更加符合风险的充分性。 Recently,VAR model is the popular calculation method on measuring and managing risk in the whole world.But VaR is unstable when the distribution is non-normal or discontinuous.In this paper,based on the theory of VaR and CVaR principles,meanwhile making a good use of the Log-normal distribution′s peak tail characteristics with Bayesian methods to calculate the amendment parameters through the posterior distribution and the a priori distribution theory,then we can get the formula of the VaR and CvaR′s calculating method.Then we compare VaR and CvaR based on their characteristic.The results of examination show that VAR under the lognormal distribution has a better risk characteristic than VAR under pareto distribution,has a better description on the random market risk,and the market risk,based on CVaR,which is extremly suitable to describe the sufficiency of the risk,has a better tail characteristic and sufficiency of the risk than VaR.