分位数回归模型和金融风险尾部相关性的实证分析

发布时间：2018-03-03 13:07

本文选题：分位数回归　切入点：条件VaR　出处：《南昌大学》2012年硕士论文　论文类型：学位论文

【摘要】：本文主要是应用分位数回归方法对条件VaR估计展开实证研究；并对Copula分位数回归以及Copula函数在金融风险的尾部相关性分析中的应用进行研究。本文应用分位数回归方法给出条件VaR的估计方法,直接得到收益率在某置信水平下分位数的值,即在一定的条件下的VaR值。这样就避免了分布是正态等假设,而且计算相对容易。在含虚拟变量分位数回归模型之下,实证分析上证指数和沪深指数在日内波幅条件下的VaR,并与无条件模型及线性分位数回归模型进行了比较。结果表明：含虚拟变量分位数回归模型比无条件模型及线性分位数回归模型能更好的度量风险。本文分析研究Copula模型及Copula分位数回归,推导了几种常见的阿基米德Copula的分位数曲线。Copula分位数回归把Copula理论和分位数回归理论结合起来,能更好的度量变量之间的关系,特别是尾部的相关关系。因此本文利用Copula模型实证分析了沪深300和深证成指的尾部相关性,用尾部相关系数将上尾相关性量化,发现沪深300和深证成指有明显的尾部相关性。并且用同种方法对创业板指数和中小板指数的相关性进行了分析,发现创业板指数和中小板指数也有明显的尾部相关性。
[Abstract]:In this paper, the quantile regression method is applied to the empirical study of conditional VaR estimation, and the application of Copula quantile regression and Copula function in the tail correlation analysis of financial risk is studied. In this paper, the quantile regression method is used to estimate the conditional VaR, and the quantile value of the return rate at a certain confidence level is obtained directly, that is, the VaR value under certain conditions, thus avoiding the assumption that the distribution is normal. And it's relatively easy to calculate. In a regression model with virtual variables, This paper analyzes the VaR of Shanghai Stock Exchange Index and Shanghai and Shenzhen Index under the condition of intraday fluctuation, and compares them with unconditional model and linear quantile regression model. The results show that the quantile regression model with virtual variables is better than the unconditional model. Linear quantile regression model can better measure risk. In this paper, Copula model and Copula quartile regression are analyzed and studied. Several common Archimedes Copula quartile curve. Copula quartile regression combines Copula theory with quantile regression theory, which can better measure the relationship between variables. Especially the tail correlation. So we use Copula model to analyze the tail correlation between CSI 300 and Shenzhen Stock Exchange, and use the tail correlation coefficient to quantify the tail correlation. It is found that there is an obvious tail correlation between Shanghai and Shenzhen 300 and Shenzhen Stock Exchange, and the same method is used to analyze the correlation between the gem index and the small and medium-sized board index, and it is found that there is a significant tail correlation between the gem index and the small and medium-sized board index.
【学位授予单位】：南昌大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】