投资组合风险价值分解方法研究

提出了两种分别对组合损失和极端损失进行线性近似的投资组合风险价值(VaR)分解的新方法——非对称响应模型估计法和局部线性近似估计法，两种方法均可用于分解参数法与非参数法计算出的投资组合风险价值，弥补了现有分解方法只能分解部分参数法计算出的投资组合风险价值的不足，实证研究结果表明：非对称响应模型法比局部线性近似估计法更为简单、快捷，适用于当组合回报不对称现象明显时；而当组合回报具有显著的厚尾现象时，后者更为准确，两者都具有一定的实际应用价值。

Methods of decomposing portfolio value at risk (VaR)

The asymmetric response and linear local approximation methods were presented to decompose portfolio VaR by the linear approximation of the loss and extreme loss of portfolio return. These methods can be used to decompose portfolio VaR calculated by the parametric and non parametric methods, compensating the disadvantage of current decomposing methods. The empirical research result shows that the linear local approximation method is more accurate and complicate, the asymmetric response method is simple and quick. Both the methods are valuable.