【作者】 杨夫立；

【导师】 周爱民；

【作者基本信息】 南开大学 ， 金融学， 2012， 博士

【副题名】基于GARCH模型、VaR及CVaR方法与Copula函数的视角

【摘要】 在金融全球化和自由化的背景下，随着证券投资基金市场的发展，基金的市场风险越来越成为金融实务界、学术界和监管机构共同关注的对象。作为现代风险管理理论的核心与基础，VaR风险价值理论及在其基础上发展形成的CVaR条件风险价值方法已成为风险测度研究的前沿课题。本篇论文是围绕着基金的市场风险测度展开讨论的，研究的领域涉及到基金收益率时间序列模型的各个方面。特别是针对基金收益率时间序列的尖峰厚尾特征与异方差现象，在风险价值VaR与条件风险价值CVaR的研究中，先后引入了GARCH模型与Copula连接函数进行探讨。试图在前人工作的基础上给出自己的研究思路，并拓展一些理论分析和实证检验方法。本文的第一章为绪论。主要阐述了本文的研究背景及选题意义,研究的基本思路与主要内容,综述了国内外关于VaR风险价值量、CVaR条件风险价值与Copula连接函数的研究现状。并简要介绍了本文的基本框架。本文的第二章为证券投资基金的传统市场风险度量方法评析。本章系统的研究了基金的传统市场风险度量方法,主要包括名义量法、均值方差分析、灵敏度方法、波动性方法。对基金市场风险度量方法的发展历程进行了简要回顾，对传统的各种风险度量方法进行了点评。对各方法的优缺点进行了分析。本文的第三章对VaR的风险度量方法及在其基础上改进产生的CVaR方法方法进行了详细介绍与实证研究。第四章讨论并建立了基金收益率的波动模型，详细介绍了基金收益率的波动模型——GARCH模型族，以及GARCH模型建模程序，并结合上证指数收益率进行了实证分析。第五章为基于GARCH模型的VaR计算方法及应用，分别讨论了VaR计算模型、VaR计算的基本模块、传统的VaR计算方法、基于GARCH模型的VaR计算方法与计算步骤，分别构建GARCH-N分布模型、GARCH-t分布模型和GARCH-GED分布模型，进行样本基金的VaR值计算和实证分析；第六章为基于GARCH模型的CVaR风险度量，介绍了CVaR模型的计算方法、基于GARCH模型的CVaR计算设计、基于CVaR的基金风险度量实证研究。本章用GARCH族模型进行基金市场的波动性模拟,分别在传统正态分布假设、刻画尖峰重尾特性的学生t分布及GED分布假设下,结合不同基金公司的9只三种类型的开放式基金来进行实证研究。本文第七章详细介绍了基于Copula函数的GARCH模型优化方法——Copula-GARCH，并利用其进行沪深股市的日收益率相关性分析，同时尝试基于蒙特卡洛模拟算法进行基金投资组合的Clayton Copula-VaR实证计算。。本文第八章为全文总结，并对论文的不足和未来的研究方向进行了展望。更多还原

【Abstract】 In the background of the financial globalization and liberalization, with thedevelopment of market of securities investment funds, Market risks management is ahot topic in financial institutions, academia, and financial supervisors. As the coreand Foundation of the modern theory of risk management, theory of VaR and itsformation based on CVaR method has become the current topics of research on riskmeasurement. This essay is discussions focus on the Fund market risk measures,research areas related to yield time-series model, the varios aspects of the Fund. Inparticular, against the income of the Fund rate time series of Spike characteristics ofheavy-tailed and heteroscedasticity, VaR and CVaR study, has introduced the GARCHmodel to explore connected with the Copula function. Attempting to predecessors onthe basis of the work to their own research, and expand theoretical analysis andempirical methods.The first chapter is the introduction of this article. It mainly elaborates theresearch on the research topic selection of backgrounds and significance of this article,basic ideas and main contents of the study, a summary of the domestic value of VaRrisk and CVaR research status of conditional value at risk connected with the Copulafunction. And it outlines the basic framework of this article.Chapter II of this article is about the traditional market risk measurement methodof securities investment funds. Systems in this chapter on the Fund’s traditionalmarket risk measurement methods, mainly by including name, mean-VaRianceanalysis, sensitivity, volatility methods. To fund development of market riskmeasurement method for a brief review, the traditions of the various comments on therisk measurement methods. Analysis of the advantages and disadvantages of eachmethod.This third chapter is about risk measurement method of VaR and itsimprovement on the basis of CVaR method.The fourth chapter discussed and established a fund yield fluctuation model,details the funds rate fluctuation model--GARCH model and GARCH model modeling programs, and combined with yield on the index of empirical analyses.Fifth is based on calculation of VaR method and its application of GARCHmodel, discusses the calculation of VaR models, VaR calculation of basic modules,traditional methods of calculation of VaR, VaR calculation method and algorithmbased on GARCH models, respectively, the construction of GARCH-N distributionmodel, GARCH-t model and GARCH-GED model, sample VaR value calculationand analysis of the Fund;Chapter VI is on CVaR risk measurement based on GARCH models, introducedthe CVaR calculation method, based on the model of CVaR computational design of aGARCH model and the empirical study based on CVaR risk measurement. Thischapter fund market volatility with GARCH model simulation, respectively, in thetraditional normal distribution assumptions, depicting spike characteristics ofheavy-tailed Student’s t distribution and distribution assumptions of GED, with9different fund companies to conduct empirical research on three types of open-endfunds.Chapter VII is about Copula methods of empirical research on investmentportfolio risk measurement. This chapter at Southern high growth Fund’s top holdingsconstitutes the investment portfolio as a sample, using the Copula function to studythe risks of the investment portfolio, calculate the VaR of the portfolio and CVaR.The last chapte is the full-text summary and the research direction in the futureprospects.更多还原