风险度量工具CDaR在股票组合选择中的应用

发布时间：2018-03-26 22:35

本文选题：投资组合优化　切入点：VaR　出处：《华中师范大学》2014年硕士论文

【摘要】：金融风险研究的重要组成之一是Markowitz的投资组合理论。投资组合理论的发展过程主要包括了均值-方差模型,以及均值-VaR和均值-CVaR方法。目前,国际上对VaR(风险价值)的研究已经有很多系统化的理论,然而不可忽视的是,绝大部分投资者是风险厌恶的,他们在投资的过程中会更关注损失,而绝非收益,因此急需建立一个与资产损失直接相关的损失函数,使得在控制了损失的情况下取得最大收益,这样的方法更简单直观,也更切合投资者的心理。弗罗里达大学的三位学者因此提出了一个新的风险度量工具—DaR。首先它的概念很简单,又同时满足不少很好的性质,吸引了众多研究者的关注,在国外已经是金融风险管理的研究前沿,关于它的研究发展非常迅速。本文通过介绍VaR, CVaR, CDaR的定义和特性,罗列了它们的优缺点,并比较了将条件风险价值和条件风险跌幅两种理论应用于组合选择的问题的处理技巧,主要介绍和研究单参数风险度量模型中的CDaR (conditional drawdown-at-disk)方法,这一风险度量工具量化的是,在一定时期内,投资组合在特定样本路径上的下跌频率和幅度。CDaR (?)通过将一个投资组合优化模型表述为一个线性规划(LP)的问题,实现了高效和稳健的投资组合分配算法,使得处理成千上万的优化问题工具和方案成为了可能。
[Abstract]:One of the important components of financial risk research is the portfolio theory of Markowitz. The development process of portfolio theory mainly includes mean-variance model, mean-VaR and mean-CVaR methods. There have been many systematic theories on VaR (risk value) in the world. However, it can not be ignored that most investors are risk-averse, and they will pay more attention to the loss, not the return, in the process of investment. Therefore, it is urgent to establish a loss function directly related to the loss of assets, so that the maximum benefit can be obtained when the loss is controlled. This method is simpler and more intuitive. Three scholars at the University of Florida have therefore proposed a new risk measurement tool, DaR.First, its concept is very simple, and at the same time it satisfies a lot of good properties, and it has attracted the attention of many researchers. It has been the frontier of financial risk management research abroad, and the research on it is developing very quickly. This paper introduces the definitions and characteristics of VaR, Cvar, and CDaR, lists their advantages and disadvantages, and compares the techniques of applying the theory of conditional risk value and conditional risk decline to the problem of combination selection. This paper mainly introduces and studies the CDaR conditional drawdown-at-disk method in a single parameter risk measurement model. The risk measurement tool quantifies the fall frequency and amplitude of the portfolio on a given sample path over a certain period of time. By describing a portfolio optimization model as a linear programming (LP) problem, an efficient and robust portfolio allocation algorithm is implemented, which makes it possible to deal with thousands of optimization tools and schemes.
【学位授予单位】：华中师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F830.91

【参考文献】