The Study Of Financial Market Risk Measurement Based On Copula Model

As economic globalization goes deeper, the interdependence of different countries’ financial market has increasingly expanded. The outbreak of global financial crisis in 2007 increasingly highlighted the "infectious" impact on the financial risk. The greater impact and damage on the economy poses new challenges to the risk measurement of financial market. The correlationship between the current financial market is becoming more and more complex to show the structure characteristics such as asymmetric, nonlinear and tail dependence. To a certain extent Copula function can capture the complex correlation between random variables to simplifies the modeling of financial market with variables correlation. In this context, the research of correlation and risk measurement of financial market based on the Copula theory has great significance both in theory sense and application value.Based on the Copula theory framework and its application in the finance field, the dissertation summarizes the current at home and abroad situations of this study. Then, it systematically introduces the basic theory of traditional Copula and Vine Copula, and examines the dual parameter Copula function and Vine Copula function based on the financial market perspective in details. From a theoretical point of view, the dissertation summarizes the application methods of measuring financial risk using Copula model.In the empirical study, the dissertation firstly uses GJR-GARCH Model to filter the SH Index’s rate of return and turnover sequence, and then combines the Generalized Pareto Distribution and Smooth Kernel together to modify the marginal distribution’s fitting effect. On this basis, the structure is estimated by applying the Copula function, and calculates VaR by Monte Carlo simulation method, and does the regression testing furthermore. Meanwhile, the method’s advantages summarize by comparing with other methods.On the base of the marginal distribution estimation by using EGARCH, it uses the single and dual parameter Copula function to estimate the correlation structure of SH stock index and trading volume series. The empirical results show that according to the character of the data the dual parameter Copula function describe the correlation structure more accurately. It uses Vine Copula function to measure the correlation of six SH stocks. The correlation of 6 SH stocks and SZ stocks by using Vine Copula function, the results show that comparing with the traditional Copula, the measurement of Vine Copula to VaR is more accurate.This dissertation has 3 unique features. First of all, it summarizes the theory framework of Copula and explores Copula model’s setting, estimation, test and evaluation comprehensively. On the other hand, it also brings new explanations to Copula’s function and gave it more valuable economical meaning. Second, the dissertation uses semi-parametric method to estimate marginal distribution, lays the foundation for the realization of the model. Meanwhile, integrated the GARCH theory, extreme value theory and VaR theory to estimate and test, and measure the risk using the Monte Carlo simulation. So that it can reduce the probability of superposed risk of the model, and provides a more practical method to the financial risk measure research.At last, according to the data’s character, this dissertation analyzes the correlation of financial time series by using single parameter and dual parameter Copula, three kinds of Vine Copula, and summarizes the advantages of different models. It also provides a tool to further study of related problems.