非正态假设下投资组合的风险分解

为提高风险分解结果的准确性，本文给出组合收益分布服从非正态假设下，组合VaR及ES分解的两个新方法：局部线性加权平均；用分段线性函数拟合极端损失情形下，资产与组合回报之间的关系。针对实际金融数据的尖峰厚尾性与杠杆效应，风险度量采用EGARCH-GED模型。实证分析及模型检验的结果表明，新方法能准确、快捷地分解组合风险。

Decomposing portfolio risk under non-normal distributions

Two new methods to improve the accuracy of risk decomposition have been proposed, based on different non-normal distributions-linear local weighted average and piecewise—linear curve-fitting. In the light of the high kurtosis, fat tail and lever effect of actual financial returns, the EGARCH-GED model was employedin order to estimate the risk of a portfolio. The Value-at-Risk (VaR) and Expected Shortfall (ES) of an example portfolio were estimated and decomposed successfully, confirming that the proposed methods are able to decompose portfolio risk both accurately and rapidly.