基于DCC-Copula方法的中国QDⅡ产品风险管理

发布时间：2018-05-08 08:51

本文选题：风险管理 + CVaR　；参考：《湖南大学》2012年硕士论文

【摘要】：自2006年QDⅡ制度推出以来，QDⅡ产品的整体跌幅远超海外各指数，其业绩惨淡除了因为全球经济低迷，另一主要原因是QDⅡ产品主要投资于海外和香港上市的中国境内公司，资产配置缺乏国际性。采用CVaR度量QDⅡ产品风险，并从投资组合优化的角度对我国QDⅡ产品的资产配置进行分析，对我国QDⅡ产品的管理和发展具有实际意义。 CVaR作为一致性风险度量方法，在投资组合风险度量和优化中得到广泛应用，但CVaR作为一种尾部风险度量方法，受收益率尾部分布影响很大。因此，应用EVT对资产收益率尾部分布建模，能比较准确地刻画收益率尾部分布，从而更精确度量投资组合CVaR。Copula函数能够有效解决投资组合收益率多元正态分布假设存在的问题，而传统静态Copula方法在实际应用中存在局限性。因此，采用Engle提出的DCC方法将条件Gaussian Copula和t Copula的相关结构动态化，对资产收益率新息过程之间的相依结构进行度量，构造投资组合收益率的多元联合分布函数。然后，结合Monte Carlo方法，应用均值-CVaR模型进行QDⅡ产品组合优化和风险分析。最后利用来自MSCI公司的四个指数进行实证研究。实证结果表明DCC-Gaussian Copula和DCC-t Copula模型下的CVaR值均小于相应常态Copula计算的风险值，静态Copula在一定程度上夸大了投资组合的风险；收益率一定时，DCC-t Copula度量的QDⅡ产品的风险值总小于相应DCC-Gaussian Copula度量的风险值；风险规避型投资者的资产配置应以发达市场为主，在发达市场的资产配置高达45%左右；基于投资组合的优化理论，为了获得更高的收益，QDⅡ产品应降低在发达国家和新兴市场的资产配置比例，将更多的资金投向中国A股和在海外上市的中资股票。综合来看，QDⅡ产品管理人应该采取“自上而下”资产配置策略，现阶段的投资方式应以“被动型”投资策略为主，，精选个股策略为辅。
[Abstract]:Since the introduction of the QD II system in 2006, the overall decline of the QD II product is far beyond the overseas index, with a dismal performance in addition to the global economic downturn, another major reason is that the QD II products are mainly invested in overseas and Hongkong listed Chinese domestic companies, and the asset allocation is not international. The risk of QD II products is measured by CVaR and from the investment portfolio. The analysis of asset allocation of QD II products in China from the perspective of optimization has practical significance for the management and development of QD II products in China.
As a method of consistency risk measurement, CVaR is widely used in portfolio risk measurement and optimization, but as a tail risk measurement method, CVaR is greatly influenced by the tail distribution of return rate. Therefore, using EVT to model the tail distribution of asset yield rate can depict the tail distribution of yield more accurately and thus more accurate. The portfolio CVaR.Copula function can effectively solve the problem of the multivariate normal distribution hypothesis of the portfolio return rate, while the traditional static Copula method has limitations in the practical application. Therefore, the DCC method proposed by Engle is used to dynamically change the related structure of the conditional Gaussian Copula and t Copula. The dependent structure of the interdependence is measured and the multiple joint distribution function of investment portfolio returns is constructed. Then, combined with the Monte Carlo method, the combination optimization and risk analysis of QD II products are carried out by means of the mean -CVaR model. Finally, the four indices from MSCI company are used to carry out an empirical study.
The empirical results show that the CVaR value under the DCC-Gaussian Copula and DCC-t Copula model is less than the corresponding normal Copula calculation, and the static Copula exaggerates the risk of the portfolio to a certain extent; when the return rate is certain, the risk value of the QD II product of DCC-t Copula measure is less than the risk value of the corresponding DCC-Gaussian Copula measure. The asset allocation of risk averse investors should be based on the developed market, and the asset allocation in the developed market is up to 45%. Based on the portfolio optimization theory, in order to obtain higher income, QD II products should reduce the proportion of asset allocation in developed and emerging markets and invest more funds in China's A shares and overseas listing. In a comprehensive view, the QD II product manager should adopt the "top-down" asset allocation strategy. The current investment mode should be based on the "passive" investment strategy and select the strategy as a supplement.

【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】