导　　师： 张卫国

学科专业： 1201

授予学位： 硕士

作　　者： ;

机构地区： 华南理工大学

摘　　要： 随着我国汇率在两次汇改之后由盯住美元制度走向更为市场化的浮动外汇管理制度，对其市场风险的研究显得尤为必要。本文以美元兑人民币汇率作为研究对象，内容包括以下四个方面。 首先，利用长记忆性参数估计方法发现人民币美元汇率存在显著的长记忆特征。从结构突变、频率结构和时间聚合三个角度考察该市场长记忆特征的根源，得到以下结论：（1）结构突变是引起人民币汇率市场长记忆特征的一大原因；（2）低频序列是支撑长记忆特征的根源，高频意味着随机扰动；（3）长记忆特征存在期限结构特征，即不同周期的数据呈现不同的长记忆动态特征。 其次，为了寻找最优的人民币汇率风险预测模型，本文以回测检验方法作为评价方法，介绍6种刻画波动聚集性、非对称性或长记忆性的风险模型，同时引入FHS技术、EVT技术和EVT-SKT等进行全面分析。结果表明：FIGARCH类模型通过检验且检验值处于低位，但回测方法无法有效区别波动模型的优劣性以及优劣程度。 再次，基于传统回测方法的不足，本文综合监管成本和超额亏损冲击风险的概念，构造一个新的风险评价指标Indic。应用结果表明：Indic指标可以提供更为详细的局部动态信息，且结合SPA检验方法能有效区别精选长记忆波动模型的优劣性和优劣程度，对风险度量模型择优具有一定的指导意义。 最后，目前的风险度量算法缺乏灵活与可控性，模型之间的选择非此即彼。本文基于Boorstraping算法与模型样本重构的思想，将FHS技术和EVT技术相结合，获得关于VaR的稳健统计分布，同时设置分位数水平为滑点，构造一类改进免疫遗传算法--Indic-SPA检验方法寻找最优的风险度量模型。结果表明：改进的算法能以最低的Indic成本获得风险规避效用。 After two exchange rate reform, China's exchange rate head for the more market-oriented floating foreign exchange management system from dollar peg system, therefore it isparticularly necessary to study the China's RMB exchange rate market risk. In this paper theRMB exchange rate against the US dollar is studied for the following four parts: First of all, with the test of long memory characteristic, we find that there is a significantlong memory in the RMB exchange rate market.We investigate the source of market longmemory characteristics through three aspects about structural mutation frequency structureand Time aggregation, and conclude that:1/) there is a significant structural break point in theRMB exchange rate against the US. What is more, to a certain extent the long memorycharacteristics of the foreign exchange market results from structural mutations.2/)Longmemory characteristic is supported by the low-frequency signal, that is to say low frequencymeans the trend.3/)there exist the term structure characteristics for long memory parameterbecause the different cycles of data show different dynamic characteristics of long memory. Secondly, back-test recommended as evaluation method, we introduce six risk predictionmodels, which describe the volatility clustering, asymmetry or long memory characteristic, tofind the optimal risk prediction model, furthermore, we introduce SKT、FHS and EVT formore comprehensive analysis. The results show that the FIGARCH Class model pass theback-test and its test valuev are relatively low, but back-test methods can not effectivelydistinguish the pros and cons of volatility models. Thirdly, in light of the deficiencies of the traditional back-tested method, This paperconsolidated The concept of cost of market risk and excess loss impact of the risk and makesup new risk evaluation index-Indic Index, which effectively distinguish the pros and cons oflong memory volatility models combining with SPA test methods. It provide certainsignificance for the choice of opti

关 键 词： 人民币汇率 波动率长记忆特征 风险评价指标 风险度量算法

领　　域： [自动化与计算机技术] [自动化与计算机技术] [经济管理]