基于MRS Copula-ARJI-GARCH模型的投资组合VaR估计与优化

发布时间：2018-03-22 11:12

本文选题：MRS　切入点：Copula-ARJI-GARCH模型　出处：《湖南大学》2013年硕士论文　论文类型：学位论文

【摘要】：多样化投资一直是规避风险的主要手段，在国际金融格局变迁、潜在金融风险加剧的现实背景下，投资组合的风险管理和构建受到越来越广泛的关注，成为目前的一个热点问题。而投资组合的风险度量、资产联合分布的构建以及资产间相关性的刻画又是该热点问题的核心要素。针对目前已有研究在这三者设定中的不足之处，本文在综合考虑资产间非对称相关结构和资产收益率跳跃特征的基础上，构建一个MRS Copula-ARJI-GARCH模型，进行资产组合动态VaR估计和投资组合优化的研究。本文首先进行相关理论分析，定性论证MRS Copula-ARJI-GARCH模型能够提高投资组合VaR估计的准确性和投资组合优化的有效性。接着构建MRS Copula-ARJI-GARCH模型，采用该模型并结合Monte Carlo模拟估计得到行业股指组合的VaR值，通过风险控制图、失败率检验方法以及区间预测检验法将其与MRSCopula-GARCH-t模型、MRS Copula-GARCH-n模型、动态Copula-GARCH-t模型以及动态Copula-GARCH-n模型的VaR估计效果进行比较，从而定量说明本文构建的模型能够提高投资组合VaR估计的准确性。最后，分别在风险最小化策略、收益最大化策略和效用最大化策略下构建Mean-VaR投资组合模型，基于MRSCopula-ARJI-GARCH模型进行投资组合优化。研究结果表明，MRS Copula-ARJI-GARCH模型在VaR的估计中能更全面地反映资产组合极端收益的可能性，可以有效提高VaR估计的准确性，能够帮助投资者制定更郑重的投资组合决策，包括在收益一定的情况下实现投资组合风险最小，在风险一定的情况下实现投资组合收益最大以及在同时考虑收益和风险时实现效用最大化。
[Abstract]:Diversification investment has always been the main means to avoid risks. Under the background of the change of international financial structure and the intensification of potential financial risks, the risk management and construction of portfolio has been paid more and more attention. The risk measurement of portfolio, the construction of joint distribution of assets and the characterization of the correlation between assets are the core elements of this hot issue. On the basis of considering the asymmetric correlation structure between assets and the characteristics of asset return jump, this paper constructs a MRS Copula-ARJI-GARCH model to study the dynamic VaR estimation and portfolio optimization of asset portfolio. In this paper, we first analyze the relevant theories, and qualitatively prove that MRS Copula-ARJI-GARCH model can improve the accuracy of portfolio VaR estimation and the efficiency of portfolio optimization, and then construct MRS Copula-ARJI-GARCH model. Using the model and Monte Carlo simulation to estimate the VaR value of the industry stock index portfolio, the paper uses the risk control chart, the failure rate test method and the interval prediction test method to combine it with the MRSCopula-GARCH-t model and Mrs Copula-GARCH-n model. The results of VaR estimation of dynamic Copula-GARCH-t model and dynamic Copula-GARCH-n model are compared to quantitatively demonstrate that the proposed model can improve the accuracy of portfolio VaR estimation. The Mean-VaR portfolio model is constructed under the profit maximization strategy and the utility maximization strategy, and the portfolio optimization is carried out based on the MRSCopula-ARJI-GARCH model. The results show that Mrs Copula-ARJI-GARCH model can more comprehensively reflect the possibility of extreme return of portfolio in the estimation of VaR, can effectively improve the accuracy of VaR estimation, and can help investors to make more solemn portfolio decisions. It includes realizing the minimum portfolio risk in the case of a certain return, maximizing the portfolio return in the case of a certain risk, and maximizing the utility when both the income and the risk are considered.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.59;F224

【参考文献】