中国不同金融市场价格变化的相依性和风险控制

发布时间：2018-04-03 05:33

本文选题：Copula　切入点：函数　出处：《长春工业大学》2012年硕士论文

【摘要】：金融市场间的相依关系越来越密切,准确地刻画高频率变化的金融市场之间的相依结构是金融研究领域重要的课题,这对于金融市场运行机制的建立和金融决策都具有非常重要的理论和现实意义。随着金融的全球一体化,金融市场的风险管理越来越被投资者和管理者所重视。关于金融市场相依性方面,国内目前的研究大多局限在分析沪深两市之间的日周期的相关性。对于我国股指期货市场和股票市场的相依性方面的研究较少。因此,本文从高频数据极值这样一个全新的视角来研究金融市场的相依性和风险控制。由于金融市场的收益率大多数呈现尖峰厚尾分布,所以不能用传统的正态分布或t分布来描述变量间的关系。Copula函数是研究随机变量间相依性的一个新的统计工具,它连接了随机变量的联合分布和边缘分布,不同族Copula函数有不同的研究应用范围。本文介绍了Copula函数基本理论及Copula函数的参数估计和检验方法。在金融市场的风险控制中介绍了VaR理论的研究方法,VaR的计算方法等。本文以我国股指期货指数和上证指数2011年10月24到2011年11月18日的5分钟极值序列为研究对象,对股指期货IF1112指数和上证000001指数5分钟极大值和极小值的收益序列的相依性进行了研究。对对数收益率序列进行正态性检验,结果都拒绝正态假设。接着用Gumbel、Clayton、Frank和GS Copula函数对数据进行拟合,通过用Q-Q图、K-S检验发现,用Gumbel函数描述股指期货IF1112指数和上证000001指数5分钟极大值和极小值的收益序列相依性最为合适,并对股指期货IF1112指数和上证000001指数5分钟极大值和极小值的收益序列的尾部相关性进行研究,发现上股指期货IF1112指数和上证000001指数5分钟极大值和极小值的收益序列之间呈现明显的上尾相关。基于不同方法分析了股指期货指数极大值收益率序列的VaR,得到了VaR序列,给投资者或管理者从定量的角度提供了对风险控制的参考。
[Abstract]:The dependence of financial markets is becoming more and more close. It is an important subject in the field of financial research to accurately depict the dependence structure of financial markets with high frequency changes.This is of great theoretical and practical significance to the establishment of financial market operation mechanism and financial decision-making.With the global integration of finance, the risk management of financial market is paid more and more attention by investors and managers.Regarding the dependence of financial markets, most of the current studies are confined to the analysis of the correlation of the daily cycle between Shanghai and Shenzhen stock markets.There are few researches on the dependence of stock index futures market and stock market in China.Therefore, this paper studies the dependence and risk control of financial markets from the perspective of high frequency data extremum.Because most of the returns of financial markets show peak and thick tail distribution, the traditional normal distribution or t distribution can not be used to describe the relationship between variables. Copula function is a new statistical tool to study the dependence of random variables.It connects the joint distribution and edge distribution of random variables. Different families of Copula functions have different applications.In this paper, the basic theory of Copula function and the parameter estimation and test method of Copula function are introduced.In the risk control of financial market, this paper introduces the research method of VaR theory and the calculation method of VaR.In this paper, we study the 5-minute extreme value series of China's stock index futures index and Shanghai stock index from October 24, 2011 to November 18, 2011.This paper studies the dependence of the return sequence of the IF1112 index and the 000001 index of Shanghai stock index futures.The normal test of logarithmic return series is carried out, and the results reject the normal hypothesis.Then the data are fitted by Gumbelton Clayton Frank and GS Copula function. By using Q-Q chart K-S test, it is found that the Gumbel function is the most suitable to describe the dependence of the 5-minute maximum and minimum return sequence of IF1112 index and 000001 index in Shanghai stock index futures.The tail correlation of the five minute maximum and minimum value of IF1112 index and 000001 index of stock index futures is studied.It is found that the IF1112 index and the 000001 index of Shanghai stock index show a significant correlation between the return sequence of the maximum value and the minimum value in 5 minutes.Based on different methods, this paper analyzes the VaR of the maximum return rate of stock index futures index, and obtains the VaR sequence, which provides a reference to risk control for investors or managers from a quantitative point of view.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.5

【相似文献】