Application Of Copula-VaR Model In Financial Risk Management

From the bankruptcy of the British Bahrain Bank in 1995,Lehman Brothers in 2008,and then to the bankruptcy of the Greek government in 2015,the financial crisis has brought tremendous destructive power from its emergence to its outbreak.In this environment,risk aversion and secure investment environment become the urgent expectation of every investor and enterprise.Financial risk management refers to timely and accurate detection of investors' risks,and on this basis to take certain professional measures to reduce the risk exposure identified by investors,to help investors avoid risks and obtain reasonable income.Financial risk management is of great significance to individuals,enterprises and even state institutions.Financial risk management is a systematic process of risk identification,risk measurement,risk hedging and risk transfer.As the basis and core of the whole risk management system,risk measurement is becoming more and more important for the reliability of its measurement results.Value at Risk(VaR)is one of the most widely used risk measurement methods.It is defined as the maximum possible loss of financial assets under a given holding period and confidence level.It is easy to understand and can obtain the probability of occurrence of related risks and the magnitude of loss value at the same time.In 1996,the Basel Accord recommended the VaR method as a common risk management method to financial institutions to calculate their internal capital adequacy requirements.Thus,the practicability and extensiveness of VaR method can be seen.The accuracy of financial risk measurement results is limited by real variables.A large number of studies have found that financial time series show the characteristics of peak,thick tail and volatility clustering,which makes the traditional normal distribution assumption no longer accurate.And the trend of economic development in various countries is getting closer and closer.The correlation between different financial markets is becoming higher and higher.The traditional linear correlation measurement method shows obvious limitations in its use.How to determine the dependency structure of multiple financial assets has become an unprecedented challenge.Solving these problems has become the focus of this paper.Aiming at the non-normality of financial time series,this paper introduces the ARMA-GARCH model.GARCH model can capture the peak and volatility clustering characteristics of financial time series well,so as to achieve a good description of sample series,and lay a foundation for the next construction of marginal distribution.Copula theory has its unique advantages in correlation analysis.Firstly,it can be independent of the marginal distribution of variables and model the correlation structure among variables.Secondly,the correlation measure derived from the Copula function will not change because of the non-linear monotone incremental transformation of variables,so as to ensure the correct description of the correlation structure between random variables.Finally,the Copula function can well capture the non-normal among variables.The tail information of asymmetric distribution is very useful for tail correlation analysis.This paper will introduce Copula function and construct joint multivariate distribution function combined with GARCH model.Based on different types of Copula function,the risk value of portfolio composed of Shanghai Composite Index and Shenzhen Composite Index is calculated.According to the empirical results,effective financial risk management advice is given.The first chapter is the introduction,which introduces the background and significance of this paper.Then it elaborates the research results of VaR risk value,CVaR conditional risk value and Copula function by domestic and foreign scholars,and finally gives the research structure of this paper.Chapter 2 introduces the definition and calculation method of VaR risk value.According to the merits and demerits of VaR value,CVaR conditional risk value is introduced,and the advantages of CVaR value relative to VaR value in consistency risk measurement index are discussed.This paper expounds the concept of Copula function,analyses Copula function and its Sklar theorem,introduces their characteristics of distribution function,and gives the relationship between rank correlation coefficient and tail correlation coefficient and Copula function.Finally,it introduces the parameter estimation method of Copula function.Chapter 3 calculates the VaR value of a single stock based on GARCH model.I chose four representative stock samples from different industries to study their VaR and CVaR values one day ahead under different confidence levels.Through empirical analysis,it is proved that CVaR value can cover tail risk more comprehensively and estimate value is more conservative than VaR value.Chapter 4 takes portfolio as the research object,calculates VaR value based on Copula model.The portfolio is constructed by weights of Shanghai stock index and Shenzhen stock index,and the joint distribution function is constructed by GARCH model and Copula function.The VaR risk value of the portfolio is predicted one day ahead of time under different confidence levels and tested by retest.The fifth chapter is a summary.It discusses the achievements and shortcomings of this study,and puts forward relevant financial risk management opinions.