摘要
在金融风险管理中,金融风险的事先判断具有极其重要的意义,然而金融机构金融决策事前支持技术的缺陷常常被忽略,在金融投资收益率概率分布估计方法尚未建立以前,将样本数据特征纳入风险度量的计算则不失为一种改进风险判断的有效途径。本文选择度量金融风险的主流方法—VaR技术来讨论概率分布设定风险,探讨依据数据特征改进和扩展VaR计算方法,通过对Delta-正态方法与Delta-Gamma-Cornish-Fisher扩展方法估计VaR值的比较,从实证分析角度论证了扩展方法在VaR估计中的有效性与稳健性。
Prejudgment of financial risk plays very important role in financial risk management.However,defects of beforehand support technologies in financial decision-making often are ignored by financial institutions.Because estimation of probability distribution for financial investment return have not been established,placing characteristics of the sample data into algorithm for risk measurement could be an effective way for improving risk judgment.Based on VaR technology,this paper explores the risk in assumption of probability distribution and researches how to improve and expand the method.The study compares between the Delta-normal method and the Delta-Gamma-Cornish-Fisher extension method,and then proves validity and stability by Empirical Analysis when the extension method is estimated to VaR.
出处
《统计研究》
CSSCI
北大核心
2010年第10期40-46,共7页
Statistical Research
基金
国家社科基金项目“不确定性、概率分布设定错误与风险管理方法研究”(项目批准号09BTJ008)的资助
