时间序列聚类在投资组合风险管理中的应用

发布时间：2018-01-27 17:29

本文关键词： 风险度量投资组合相似度度量时间序列聚类　出处：《首都经济贸易大学》2014年硕士论文　论文类型：学位论文

【摘要】：2007年金融危机之后，国际经济金融市场发生了很大的变化，比较突出的方面是金融市场的波动性增加，金融市场上出现持续动荡。而且我国自加入世贸组织以来，，不断面临着资本市场开放的压力，国际资本市场的运动对国内的资本市场影响越来越大，导致国内金融市场的风险增加。最后由于近几年金融行业竞争的加剧，也导致金融市场风险的增加。由此，人们更加注重金融风险的管理，越来越关注投资面临的风险。在投资理论界和实务界，Markowitz的均值-方差模型为投资决策提供了一个基本和完整的框架，成为金融分析的主流方法之一。在均值-方差模型中，最核心的部分是对风险进行度量，但是该模型通过方差来对风险进行度量，存在一个显著的缺陷，这便要求探索更为准确的风险度量方法。而且现代的金融定价理论和投资理论也是建立在风险度量的基础之上，由此可见，风险度量在金融领域占据重要的位置。因此，在过去的60多年中，从Markowitz提出的方差到后来的风险价值，人们一直在追求更为合理的风险度量方法。本文首先对金融风险和风险管理进行介绍，在金融风险度量方面主要分析了VaR和CVaR两种风险度量方法。然后文章介绍了基于Copula尾部相关系数的时间序列聚类方法，最终把这种方法运用于投资组合的构建。实证分析结果表明通过时间序列聚类方法构建的投资组合的风险要小于随机构建的投资组合的风险，对实际中的投资具有指导意义。
[Abstract]:After the financial crisis in 2007, great changes have taken place in the international economic and financial market, and the more prominent aspect is the increase of the volatility of the financial market. Since China's entry into WTO, China has been facing the pressure of opening up the capital market. The movement of international capital market has more and more influence on the domestic capital market. It leads to the increase of the risk in domestic financial market. Finally, because of the aggravation of the competition in the financial industry in recent years, it also leads to the increase of the risk in the financial market. As a result, people pay more attention to the management of the financial risk. More and more attention is being paid to the risks to investment. In the field of investment theory and practice, Markowitz's mean-variance model provides a basic and complete framework for investment decision-making. In the mean-variance model, the core part is to measure the risk, but the model measures the risk through variance, which has a significant defect. This requires the exploration of more accurate risk measurement methods. And the modern financial pricing theory and investment theory is also based on risk measurement, so it can be seen. Risk measurement plays an important role in the field of finance. Therefore, in the past 60 years, from the variance proposed by Markowitz to the later value of risk. People have been pursuing more reasonable risk measurement methods. This paper first introduces financial risk and risk management. In the aspect of financial risk measurement, two risk measurement methods, VaR and CVaR, are analyzed, and then the time series clustering method based on Copula tail correlation coefficient is introduced. Finally, this method is applied to portfolio construction. The empirical results show that the risk of portfolio constructed by time series clustering method is smaller than that of randomly constructed portfolio. To the actual investment has the guiding significance.
【学位授予单位】：首都经济贸易大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F830.91

【参考文献】