基于多元Copula-SV-VaR模型的开放式基金投资组合风险测度

发布时间：2019-01-05 12:59

【摘要】：我国的开放式基金在风险管理中要求有更高的风险测度技术,而VaR是目前金融风险测度的主流指标,其计算方法有很多,由此形成的模型也有很多。考虑到计算的精确性和简便性,本文提出了多元Copula-SV-VaR模型。其中,Copula为连接函数,用于描述投资组合中金融资产之间的相关关系,这种相关关系是非线性的；SV为随机波动模型,用于描述投资组合中金融资产的边际分布,相比GARCH模型能更好地描述单个资产的尖峰厚尾性；VaR为风险测度的指标,用于描述投资组合的波动性风险的大小,通过蒙特卡罗模拟法计算出来。本文首先介绍Copula理论,着重介绍常用Copula函数的特点及适用范围,与Copula模型的参数估计方法及其适用范围,并指出概率积分变换在Copula理论中的重要意义,再介绍SV模型,着重介绍SV模型的分类与参数估计方法,然后介绍VaR,在详细总结VaR计算方法的基础上着重介绍均值—方差法和蒙特卡罗法两种计算方法,以便于理解和构建模型。在实证研究中,选择华夏大盘精选混合开放式基金为研究对象,对其前十大重仓的股票组成的投资组合进行风险测度,通过比较Copula-SV-VaR与Vari ance-Covariance-VaR的大小,以检验模型的准确性,通过比较Copula-SV-VaR与Copula-GARCH-VaR的大小,以检验模型的精确性。通过从理论和实证上详细介绍多元Copula-SV-VaR模型,希望能为我国的基金管理公司或基金监管部门对开放式基金投资组合进行风险测度提供参考依据。
[Abstract]:Open-end funds in our country require higher risk measurement technology in risk management, and VaR is the mainstream index of financial risk measurement at present. There are many calculation methods and many models. Considering the accuracy and simplicity of the calculation, a multivariate Copula-SV-VaR model is proposed in this paper. Among them, Copula is the connection function, which is used to describe the correlation between the financial assets in the portfolio, which is nonlinear. SV is a stochastic volatility model, which is used to describe the marginal distribution of financial assets in the portfolio. Compared with the GARCH model, it can better describe the peak and thick tail of a single asset. VaR is the index of risk measurement, which is used to describe the volatility risk of portfolio, which is calculated by Monte Carlo simulation. This paper first introduces the Copula theory, emphasizes on the characteristics and application range of the commonly used Copula function, and the parameter estimation method of the Copula model and its applicable scope, and points out the important significance of the probabilistic integral transformation in the Copula theory, and then introduces the SV model. This paper mainly introduces the classification and parameter estimation of SV model, and then introduces two calculation methods, mean variance method and Monte Carlo method, based on the detailed summary of VaR calculation methods, so as to facilitate the understanding and construction of the model. In the empirical study, the paper chooses China's large stock market as the research object, and measures the risk of its top ten stocks, and compares the size of Copula-SV-VaR and Vari ance-Covariance-VaR. The accuracy of the model is verified by comparing the size of Copula-SV-VaR and Copula-GARCH-VaR. By introducing the multivariate Copula-SV-VaR model theoretically and empirically, this paper hopes to provide a reference for fund management companies or fund regulators to measure the risk of open-end fund portfolio.
【学位授予单位】：五邑大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.48;F224

【参考文献】