风险管理的CVaR方法及其简化模型

The CVaR method of risk management and its simplified model

VaR模型度量的是在某一置信水平下,资产损失的最高期望值,但是它没有指明一旦超过了这个期望值,资产的损失究竟是多少。1997年出现的CVaR模型弥补了这个缺陷。如果以f(x,y)表示投资组合的损益函数,其中x为资产的比例向量,y表示市场随机因素,则CVaR考察的是在f(x,y)超过了VaR值时,f(x,y)的量度问题。这种方法的一个优越之处在于,最终的求解可以转化为线性规划问题,从而具有良好的操作性,而且问题的结论不仅包含了CVaR的大小,同时也可以求出资产的VaR值以及资产组合的最佳比例。

The scale of VaR model is under certain confidence level,to compute the capital′s,but it isn′t pointed out that once the expected value of maximum loss is exceeded,how much the loss of capital will be.In 1997 the emerged CVaR model covered the defection.If f(x,y) represent the portfolio′s profit and loss function,among of which x is defined capital′s ratio,y is defined market′s random factor,so CVaR is mainly used to metric f(x,y) scale′s problem when f(x,y) exceeds VaR.The superiority of this way is the final solution may be transferred to linear program problem,and it doesn′t only include CVaR′s value but also the VaR and the portfolio′s best ratio can be computed.