【作者】 刘汪卉尧；

【导师】 聂高辉；

【作者基本信息】 江西财经大学 ， 数量经济学， 2012， 硕士

【摘要】 风险研究是一个内容极其广泛的领域,涵盖风险测度技术、相关性研究等。文章正是以相关关系为切入点,展开中国股票市场的组合风险研究。本文选取沪深港三市股票指数作为对象,具体内容包括：理论方面,本文进行了风险管理相关理论的概述,主要介绍了风险计量技术(波动理论、风险测度模型)和联结理论(Copula)。 Copula理论是一种新的构建联合分布的方法,同时也是一种适用于探索非线性复杂相关关系的工具,目前已被广泛运用于金融领域。文章在深入介绍Copula理论的同时,提出了运用该方法进行投资组合的VaR计算。在计算VaR的过程中,采用的是蒙特卡洛模拟技术。实证方面,文章分别针对沪深、沪港两市股票指数建立了Copula-GARCH模型,计算得到投资组合的市场风险,并采用蒙特卡洛模拟技术对模型的有效性进行验证。实证结果显示,三个市场的单市场波动情形适合采用具有尖峰厚尾假设分布的GARCH模型刻画；沪港和沪深两市之间体现出不同的相关关系：沪港之间适合使用Gumbel Copula,是一种上尾相关的关系,而沪深之间体现的则是一种上下尾对称的相关性,适合使用Frank Copula;沪港两市的等权重组合市场风险较小,优于沪深两市的等权重组合。更多还原

【Abstract】 The study of risk refers to so many fields, such as risk measurement techniques, correlation research and so on. This article is about the portfolio risk research of china stock market, which takes the relevance as the breakthrough point. The paper bases on the Shanghai、 Shenzhen and Hong Kong share index:Firstly, paper outlines theory of risk management:the risk measurement techniques (fluctuation theory, risk measure model) and coupling theory (Copula) in particular. Copula is an useful theory which can be implied to construct joint distribution and can explore non-linear relationship. The article puts forward to apply the theory to calculate VaR of portfolio, using the Monte Carlo simulation technology.Secondly, paper does the empirical research of the return series by Copula-GARCH model and gets the portfolio market risk. The models used are proved to be valid by Monte Carlo simulation. Besides, other conclusions are:GARCH model which has the assumption of t-distribution suits to sketch the fluctuation of single market. The relationships between them are different, Gumbel copula suits to shanghai-Hong Kong as the upper-tail correlation, while Frank copula for shanghai-Shenzhen as the symmetric correlation on upper and lower tail. Empirical research also suggests that the market risk combined by shanghai-Hong Kong is much smaller than combination of shanghai-Shenzhen.更多还原