【作者】 李超；

【导师】 胡国恒；

【作者基本信息】 河南师范大学 ， 产业经济学， 2017， 硕士

【摘要】 自Copula理论提出之后,便得到迅速的发展和广泛的应用,尤其是在金融领域的相关分析、衍生品定价、风险管理中。Copula函数具有明显的优势:一是能准确地描述多个变量之间的相依性,二是灵活地构造多元分布函数,三是对随机变量的边缘分布的选取不做限制,四是可以捕捉到变量间非对称、非线性和尖峰厚尾的特性。它弥补了传统度量风险的技术缺陷,提高了风险模型测度的准确性和可靠性,为风险的防范和管理提供了理论依据。由于我国股市和债市发展历程比较短,市场效率明显低于国外,所以对两市场之间的相依结构及风险的研究不能简单的趋同于国外。虽然近年来国内在该方面取得了一定的成果,但并未结合Copula函数考虑在不同行情下两市场之间的相依性及组合风险。因此,本文为了研究资产组合间的相依性,且更好的度量金融市场组合间的风险,主要工作体现在:1、本文首先阐述了研究的背景及意义,指出了引入Copula理论的必要性和重要性,对国内外有关Copula的文献做了综述,并详细地介绍了Copula相关理论。在风险分析中,比较全面的分析了风险的Va R和CVa R。2、本文选用沪深300指数和中债总指数为样本,选取的时间段是从2014年7月到2017年3月。在K线图中,以60日均线为基准,对样本按照股市行情(牛市、熊市、反弹、震荡)进行划分,分别对各个行情进行实证分析:在对数据的检验、分析中使用了单位根检验、自相关性检验、ARCH效应检验及K-S检验等;在对模型参数的估计中利用了Eveiws和Matlab软件,在对一定置信水平下的两市场风险的Va R和CVa R及组合风险的计算中使用了Copula函数和蒙特卡洛模拟法,并对分段行情下的组合风险进行了简单的优化。通过实证发现:首先,在牛市和反弹行情中,股市和债市之间呈现出负相关关系,在熊市和震荡行情中,股市和债市之间呈正相关关系,且在反弹行情中两市间的相关性最强;其次,GARCH-(1,1)-t模型能够很好地描述金融资产时间序列的波动性,组合风险在熊市行情中最大,牛市行情中最小,牛市行情和反弹行情下,债市的风险相对而言比较高,而在熊市行情下的债市风险最低;再次,在不将股市分段时,整体的相关性掩盖了局部的相关性,整体的风险高估或者低估了局部的风险;最后,改变组合的投资比例确实可以降低组合的风险,投资者可以根据股市和债市间相关系数的大小及正负性去判别两市场处在什么行情中,再根据不同行情中风险的大小进行权衡投资。更多还原

【Abstract】 When the theory of Copula was proposed since 1959,then it got rapid development and extensive application,especially in the financial sector of correlation analysis,derivative pricing and risk management.Copula function has some obvious advantages:one,it can accurately describe the dependence between variables,two,it is flexible to construct the distribution of random variables,three,it doesn’t limit the selection of the edge of the random variable distribution,four,it can capture the asymmetry,the characteristics of nonlinear and rush thick tail between variables.It makes up for the traditional risk measure of technical shortcomings and improves the accuracy and reliability of the risk model,as well as provides the theoretical basis for risk prevention and management.Because the development of our country stock market and bond market is shorter,and the market efficiency is significantly lower than the foreign countries,so the structure of the dependence between the two markets and the study of risk must not simply belong to foreign countries.Although the research in this respect has been made some achievements in recent years,not combine Copula functions to take into account the dependencies between the two markets in different market and portfolio risk.Therefore,in order to study the dependency between the portfolio and measure the risk of financial market combinations,the main work of this paper is reflected in:One,this paper describes the background and significance of the research,and points out the importance and necessity of introducing the theory of Copula and does a simple review on the literature of Copula connect both at home and abroad.It makes a comprehensive comparison between Va RwithCVa Rin the theory of portfolio risk analysis.Two,this paper uses CSI300 index and total bond index as the sample,which is from July 2014 toMarch 2017.To divide the sample in accordance with the market(bull market,bear market,rebound in the market,volatile market)is on the basis of 60 per line in K line graph,and conduct empirical analysis for each market separately:It uses the unit root test,the correlation test,the ARCH effect inspection and K-S test,etc in the inspection and analysis of the data;Uses Econometric Views and Matrix Laboratory software to estimate the parameters in the model;Connects Copula functions with the Monte Carlo simulation method to estimate theVa RandCVa Rand portfolio risk of stock and bond markets under a certain degree of confidence level,and optimizes the combination of risk under the section of the market.Through the study we find that:Firstly,it is a negative correlation relationship between stock and bond markets in a bull market and the rally,it is a positively correlated relationship between stock and bond markets in a bear market and volatile market,and the strongest correlation between the two markets is in the market rally.Secondly,GARCH-(1,1)-t model can well describe the time sequence of volatility of financial assets.The portfolio risk is the largest in a bear market and smallest in the bull market.The risk of bond market is relatively high in bull market and rebound market,but it is the lowest in the bear market of the bond market risk.Once again,when the stock market is not divided,the overall relevance obscures the local relevance and the overall risk is overestimated or underestimates the local risk;Finally,investors can according to the size and the positive and negative of the correlation coefficient between stock market and bond market to distinguish what market the two markets is,and to judge the market risk and portfolio risk of two types of asset size,then make a investment balance.更多还原