【作者】 秦璇；

【导师】 田新时；

【作者基本信息】 华中科技大学 ， 金融学， 2006， 硕士

【摘要】 近年来,金融风险成为国内外金融实务界、理论界和监管机构共同关注的对象。风险测度是风险管理中首要而核心的部分,在金融自由化的国际背景下研究风险测度对于风险的有效管理、我国风险管理研究的发展,乃至我国金融体系的建设都具有十分重要的意义。本文以“CDaR模型的有效前沿研究”为题。文章首先介绍了论题研究的背景和意义,接着按照时间的顺序,概述了在金融风险管理理论发展过程中具有里程碑意义的重要模型。具体包括均值-方差(MV)模型、均值-风险价值(VaR)模型、均值-条件风险价值(CVaR)模型和均值-条件风险跌幅量(CDaR)模型。本文的重点是均值-条件风险跌幅量(CDaR)模型:采用R.T.Rockafellar和S.Uryasev的一种优化算法[1],构造了一个以条件风险跌幅量(CDaR)来度量金融风险的投资组合优化模型。在实证分析方面,本文选择了沪市的10种股票构成了一个投资组合,用Matlab科学计算软件进行了优化计算,得到了该组合的投资权重和有效前沿。同时,本文还对所得到的有效前沿进行了横向和纵向的比较分析。最后得出结论:从整体上看,基于条件风险跌幅量(CDaR)的优化模型无论从广度还是精度上,在对投资组合的风险度量和风险控制方面都比均值-条件风险价值(CVaR)模型具有更强的适应性。更多还原

【Abstract】 Financial risks management is a hot topic in financial instituations, academic, and financial supervisors for recent years. Risks measurement is the core of effective risks management. Therefore, it is significantly important to study risks measurement in the background of financial globalization. Reasonable risks measurement is fundamental for China’s risk management study, as well as China’s finance development and financial markets construction.The main topic of this article is“The Research of CDaR-based efficient frontier”. Firstly, we explain the reason why we chose this thesis and the background of it. Secondly, we discuss some very important theory about risk measurement and management ,such as Mean-Variance(MV)model, Mean-Value-at-risk(VaR)model, Mean-Conditional Value-at -risk(CVaR) model and Mean-Conditional Drawdown-at-risk(CDaR) model .The emphasis of this article is Mean-CDaR model: we adopted an optimizing arithmetic method advanced by R.T.Rockafellar and S.Uryasev. Then we are able to construct an optimizing model which made use of conditional drawdown-at-risk to measure the financial risks. Empiricially, we constitute a portfolio of 10 significant stocks in Chinese stock market. And by use of the famous calculating software-Matlab, we optimize the portfolio, and got the efficient frontier. Besides,we can also determine the share of the every piece of stock. Finally, we give out the comparing analysis for the frontier we have got ,breadthways and lengthways. Then we can conclude that the optimizing model based on CDaR is superior and more suitable than the model on other risk measurement.更多还原