估计VaR的一种新方法及实证分析  

A New Method of Computing VaR and the Relevant Empirical Test

作　　者：鹿长余[1] 高禹[2] 李夫明[3] 

机构地区：[1]上海金融学院,上海201209 [2]浙江海洋学院,浙江舟山316004 [3]山东理工大学,山东淄博255049

出　　处：《上海金融学院学报》2005年第3期31-35,共5页Journal of Shanhai Finance University

基　　金：国家自然科学基金资助项目(10471043)

摘　　要：在VaR计算的众多方法中,历史模拟法以其概念直观、计算简单、容易实施,被越来越广泛地应用到市场风险的管理中。然而采用历史模拟法计算VaR时,会出现VaR高估现象,另外,历史模拟法在计算VaR时没有很好的反映金融数据的聚集现象和厚尾现象。有经济学家在1999年曾提出过一种改进方法。本文提出了一种新的估计方法,实证分析显示新的估计方法能更好地揭示资产的风险水平。Historical simulation method for computing VaR has grown increasing popular in practice because it is very easy to implement and also very intuitive to understand. However ,there are two very obvious disadvantages in this method, the first one is that all the returns within the window are assumed to have the same distribution thus the probability density function of the return sery doesn't change though the time; the second one is that only static dataes in the past are used when computing VaR without considering the dynamic volatility of the market at present. To overcome these disadvantage, an improvement was introduced by some economists in 1999. In this paper, we come up with a new improvement, empirical tests show that our method can get a better result than Boudoukh's method.

关 键 词：VAR 历史模拟法 指数加权移动平均法 返回检验法 

分 类 号：F830.39[经济管理—金融学]