节点文献

证券投资风险值VaR的度量与组合优化研究

The Measurement of the Stock-jobbing Value at Risk and Its Study of the Combining Optimization

【作者】 刘艳春

【导师】 高立群;

【作者基本信息】 东北大学 , 模式识别与智能系统, 2005, 博士

【摘要】 伴随着全球经济一体化、投资自由化以及金融创新的不断深入,金融市场的波动性和风险也在不断加剧,对于金融风险的管理已经成为金融机构和投资者所面临的最重要问题。VaR(Value at Risk)即风险值,作为金融风险分析、测度与防范的重要工具,是近年来国际上兴起的一种定量度量金融风险的管理方法。它是将众多不可测的主观因素转化为运用数理统计方法和计量技术的客观概率数值,使隐性风险显性化。VaR的概念虽简单,然而对它的度量却是一个具有挑战性的统计问题。 本文就金融风险管理研究的起源、金融风险的定义及分类、金融风险管理存在的理论基础以及现代金融风险管理的主流模型和方法技术做出了综述。在此基础上,重点研究了证券投资组合选择与组合评价的理论,并提出了金融风险管理研究未来可能的发展方向。主要研究有: (1) 对目前国内外学术界对风险的定义进行了归纳和综述。归纳出七种学说观点,并对各种学说观点进行了评述和数量刻画。目前学术界对风险的内涵没有统一的定义,针对不同的风险源、风险管理的不同目标,不同的学者对风险的理解和认识程度与研究角度的不同,产生了不同的风险度量方法。通过对各种学说的归纳和总结,反映了人们对风险的认识不断深化、不断完善的递进过程。从现有的风险定义出发,重点分析了证券投资领域投资风险的特殊性,探讨了证券投资风险的本质特征。指出用VaR来描述金融市场风险将更为全面合理。 (2) 对中国股票市场收益率序列基本性质的研究。运用SAS和S-plus统计软件对上海股票市场收益率分布的正态性、自相关性、异方差性、独立同分布性等进行了相应的统计检验。实证分析与检验结果表明,在中国股票市场中,收益率序列具有明显的非正态特征,收益率序列独立同分布的假设被拒绝。收益率序列呈“尖峰厚尾”分布,且序列之间存在着明显的自相关性。 (3) 建立了风险值VaR的基本概念和理论分析框架。给出了服从几何布朗运动的股票和期权的VaR。紧密跟踪学科发展的前沿,对VaR度量的新方法进行了研究。针对模型中要求收益分布服从正态分布的假定,对VaR进行了拓展研究,引入了最坏条件VaR(worst-case VaR,WCVaR)的概念,并给出了它的计算公式。为了检验度量VaR方法的有效性,引入了Kupirc所给出的后验检验统计量,并采用似然比检验的方法,给出了模型有效的拒绝域。 (4) 首次提出了循环修正的组合评价方法。针对目前多指标综合评价方法很多,各种方法的出发点不同,解决问题的思路不同,适用的对象不同等,存在着

【Abstract】 With the deepening of global economics integration, investment free port and financial innovation, the financial magnate’s volatility and risk are also increased. The financial risk management has become the most important question which is the financial policy and the investor are facing. VaR, that is Value at Risk, as the main way to analyzing, measuring and preventing the financial risk. It is a kind of management way which has just come out to measure the financial risk in the world. It offered us a modern management theoretical conception and thinking. VaR is the method of converting many unmeasurable subjective factors into objective probability numeral value using mathematical statistic and measuring technology. This made the risk obvious and convenient to be managed and controlled. The conception of the VaR is simple, but its measurement is a challenging statistic question.The article describes the original, definition, types and the basic theory of the financial risk management. And also gives an illustration about modern risk management, it contains main model, method and technology. It makes further discussion about the theory of portfolio selection and combination evaluation. It offered the possible developing tendency of the future financial management. The main studies contain:(1) The article summarized the definition of the risk from international different research fields at present. It induced seven theory opinions and criticized the different theories and depicted the amount. Now the research has no integrated definition to the contents of the risk. For the different risk source and different risk management goal, different research scholars have different risk measuring methods from the understanding level and studying angle to the risk. The summarizing to the kinds of theories has reflected the people’s understanding to the risk incessantly deepen and to be perfect. From the existing risk definition, we emphasize the character of investment risk in the stock investment field. And probe the inherent quality of the risk. It points out that VaR will be more comprehensive and reasonable.(2) It studies the basic quality of the profit order of stock market in China. It uses SAS and S-plus statistic software to test the relative profit distribution of normal, autocorrelation, heteroskedasticity and independence in common distribution Shanghai stock market. The analysis to the example and result expresses that the profit order has obvious normal character in China stock market. And independence in common distribution hypothesis has been rejected. And the profit order distributes like sophisticated top and thick bottom and has obvious autocorrelation among the order.(3) It establishes the basic conception and theory analysis frame of VaR. It gives

  • 【网络出版投稿人】 东北大学
  • 【网络出版年期】2006年 12期
  • 【分类号】F830.91;F224
  • 【被引频次】14
  • 【下载频次】2122
  • 攻读期成果
节点文献中: