投资组合动态VaR风险度量  

Dynamic Value at Risk Measure of Portfolio

作　　者：许启发[1] 

机构地区：[1]天津大学管理学院

出　　处：《统计与信息论坛》2008年第6期40-45,50,共7页Journal of Statistics and Information

基　　金：国家自然科学基金项目《动态Copula模型的构建及其在金融领域的应用研究》(70671074);中国博士后科学基金项目《高阶矩风险条件下动态组合投资理论、方法与应用》(20060400192);全国统计科研计划重点项目《动态高阶矩风险对金融投资决策的影响》(2006B07);教育部人文社会科学青年基金项目《金融市场微观结构噪声研究》(07JC790046)、《我国统计管理体制与企业统计改革的研究》(06JD910006)

摘　　要：投资组合的VaR风险度量依赖于投资组合中金融资产间联合分布函数的确定,随着投资组合规模的扩大,其VaR的计算难度也不断加大。利用ICA可以将多元联合概率分布函数转化为一元概率分布函数乘积实现简化计算的特点,基于ICA的投资组合动态VaR风险度量方法和计算步骤,克服了多元非正态条件下VaR测算上的困难。实证研究表明,与EWMA模型法、MGARCH模型法相比,ICA法能够准确地度量投资组合动态VaR。VaR measure of portfolio is depended on the joint distribution function of assets. With the size of the portfolio enlarged, it is difficult to calculate the VaR. According to the facts that the ICA method can be used to transform the joint distribution function to probability distribution function, we propose the measure methodology and process for dynamic VaR of portfolio through ICA. The trouble in measuring VaR under the circumstance of multivariate abnormal has been overcome by the methods proposed in the paper. In empirical analysis, compared to the method of EWMA model and MGARCH model, method of ICA can estimate the dynamic VaR of portfolio exactly.

关 键 词：VAR 独立成分分析 动态 投资组合 

分 类 号：F830[经济管理—金融学]