外汇市场收益波动与VaR风险度量研究

发布时间：2018-01-07 18:02

本文关键词：外汇市场收益波动与VaR风险度量研究　出处：《华中科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着固定汇率价格体系的崩溃及世界范围内兴起的金融自由化浪潮，由此引发的汇率波动导致作为外汇市场主体的金融机构和涉外企业所面临的汇率风险空前增大。将计量经济学模型中GARCH族模型及VaR计算方法应用于汇率收益波动特征及汇率波动风险的研究中，将有助于揭示人民币外汇市场汇率波动特征，同时可以为各大商业银行及监管层的决策提供理论及实证结果方面的支持。本文首先对汇率波动VaR风险度量模型主流方法进行了探讨，考虑到VaR的估计值与极值分布有着紧密联系，本文进一步探讨了VaR计算模型的极值方法。其次，在收益率分布方面,统计结果表明汇率收益率序列均具有“尖峰厚尾及”及非正态性的特征。从波动性建模角度考虑，考察了汇率收益率序列具有波动持续性、异方差及杠杆效应特征。实证分析方面，本文对GARCH过程适用前提进行探讨，随后给出了误差项基于正态分布、t分布及广义误差分布的GARCH族模型的最大似然估计结果。考虑到通常的信息准则并不能够反映GARCH过程拟合数据的优劣，本文对原有的信息准则提出了改进，并利用改进的信息准则挑选出了最优拟合模型。最后，为了有效的描述我国外汇市场汇率波动风险，本文使用了VaR的不同计算方法对其波动风险进行了度量，并利用回顾测试的方法检验了基于不同VaR计算方法的估计结果与实际收益率序列的契合度。在此基础之上，笔者探究了外汇投资收益与时变风险之间的关系，通过考察GARCH-M类模型，，笔者发现了针对不同收益率序列的时变风险项的最佳测度方式。
[Abstract]:With the collapse of the fixed exchange rate price system and the rising wave of financial liberalization in the world. The exchange rate fluctuation resulted in the unprecedented increase of exchange rate risk faced by financial institutions and foreign-related enterprises as the main body of foreign exchange market. The GARCH family model and VaR calculation method in econometrics model are applied in this paper. The characteristics of exchange rate volatility and the risk of exchange rate volatility. It will help to reveal the characteristics of exchange rate fluctuation in RMB foreign exchange market and provide theoretical and empirical support for the decisions of major commercial banks and regulators. In this paper, the mainstream method of VaR risk measurement model for exchange rate volatility is discussed, considering that the estimated value of VaR is closely related to the distribution of extreme value. In this paper, the extremum method of VaR model is further discussed. Secondly, in the aspect of return distribution, the statistical results show that the exchange rate return series has the characteristics of "peak and tail" and non-normal, considering from the perspective of volatility modeling. The exchange rate return series is characterized by volatility heteroscedasticity and leverage effect. In empirical analysis this paper discusses the applicable premise of GARCH process and then gives the error term based on normal distribution. The results of maximum likelihood estimation for GARCH family models with t distribution and generalized error distribution. Considering that the usual information criterion can not reflect the advantages and disadvantages of GARCH process fitting data. In this paper, the original information criterion is improved, and the optimal fitting model is selected by using the improved information criterion. Finally, in order to effectively describe the volatility risk of foreign exchange market in China, this paper uses different calculation methods of VaR to measure the volatility risk. The coincidence between the estimation results based on different VaR calculation methods and the real rate of return series is tested by the method of retrospective test. The author probes into the relationship between foreign exchange investment returns and time-varying risks. By investigating the GARCH-M model, the author finds out the best way to measure the time-varying risk terms for different return sequences.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F832.6

【相似文献】