基于GARCH-POT模型的中国外汇市场投资组合研究

发布时间：2018-01-23 17:47

本文关键词： 外汇投资风险度量外汇投资组合　出处：《哈尔滨工业大学》2015年硕士论文　论文类型：学位论文

【摘要】：2008年美国金融危机爆发后,各国资产开始重新配置,全球外汇市场波动的加剧。在后金融危机这一背景下研究处于汇改攻坚阶段的中国外汇市场的风险,并对其进行度量对于国内的投资者意义重大。通过对2008年1月2日到2015年3月9日美元、欧元、日元以及港币对人民币的汇率中间价的对数收益率进行统计检验研究,可以看出4组外汇数据收益率序列都存在“有偏尖峰且薄尾”的特征,均存在自相关与偏自相关效应以及异方差效应,基于以上特征建立的GARCH模型可以较好地反映出中国外汇市场收益率的波动性。通过实证检验可以发现,GARCH模型在描述数据波动性方面效果良好,但是对序列的尾部拟合不是很好。因此,将极值理论引入到描述模型的残差中来弥补GARCH模型的不足,以对残差尾部的点进行有效描述。通过Hill估计法确定POT模型的阈值u,进而估计出POT阈值模型的上下尾参数,将POT模型与Va R、CVa R风险度量方法联立,计算出单支外汇的风险值Va R和条件风险值CVa R。可以对30天的单支外汇风险值进行较准确地预测。在建立GATCH-POT模型的基础上,通过引入多元正态Copula模型、多元t-Copula模型以及多元时变Copula模型,联立美元、欧元、日元以及港币4组外汇残差序列的边缘分布进而形成一个统一的联合分布,可以计算出中国外汇投资组合的风险价值。并在风险最小原则下,度量出4种外汇的最优投资比例,进而计算出4组外汇投资的最佳比例,为投资者提供了投资决策的重要依据。
[Abstract]:In 2008, after the outbreak of the American financial crisis, the assets of various countries began to be redistributed, and the volatility of the global foreign exchange market intensified. Under the background of the post-financial crisis, the risk of China's foreign exchange market in the stage of foreign exchange reform was studied. And to measure it is significant for domestic investors. Through the January 2nd 2008 to March 9th 2015 dollar, the euro. The logarithmic rate of return of the exchange rate of yen and Hong Kong dollar to RMB is statistically tested, and it can be seen that the series of four groups of foreign exchange data rate of return all have the characteristics of "biased peak and thin tail". There are autocorrelation, partial autocorrelation and heteroscedasticity effects. The GARCH model based on the above characteristics can well reflect the volatility of China's foreign exchange market returns. The GARCH model is effective in describing the data volatility, but the tail fitting of the sequence is not very good. Therefore, the extreme value theory is introduced into the residual of the description model to compensate for the shortage of the GARCH model. In order to describe the point of residual tail effectively, the threshold value u of POT model is determined by Hill estimation method, and then the upper and lower tail parameters of POT threshold model are estimated. CVa R risk measurement method is simultaneous. The risk value V a R and the conditional risk value CVa R of a single foreign exchange can be calculated. The risk value of a single foreign exchange for 30 days can be accurately predicted. Based on the establishment of GATCH-POT model. By introducing the multivariate normal Copula model, the multivariate t-Copula model and the multivariate time-varying Copula model, the United States dollar and euro are established. The marginal distribution of the four groups of foreign exchange residuals in yen and Hong Kong dollars forms a unified joint distribution, which can calculate the risk value of China's foreign exchange portfolio, and under the principle of minimum risk. The optimal investment proportion of four kinds of foreign exchange is measured, and the best proportion of four groups of foreign exchange investment is calculated, which provides an important basis for investors to make investment decisions.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F832.6;F224

【参考文献】