Research On Measurement Of Default Correlation Based On Copula

From the end of last century, with the rapid development of finance integration, the financial risks of different forms become more difficult to measure and prevent. It is seriously influencing the health and security of financial industry. With China's accession to the WTO and to honor its commitments, China's commercial banks, especially the state-owned commercial banks and foreign banks will have to compete under the same conditions. In an attempt to remain competitive position, effective risk management is important. However, risk management has chosen to be the Achilles heel of China's banking industry. Due to the high proportion of non-performing assets of the risk management techniques and backward restrictions in the face of increasingly complex financial risk appears to be inadequate.Credit risk is one of the main risks to commercial banks. Comparing with other financial risks such as market risk, his theory is much more complicated and more difficult to measure. So the current domestic and foreign researches focus on financial risk management. Credit risk and default risk as the main form naturally become the top priority of the study. Correlated default, which is derived mainly from the assets correlation between enterprises, is one of the typical forms among financial risks. Therefore, how to measure the default correlation becomes one of the centers of research about correlated default. Due to the foreign risk of default database is better, so it focus on finding perfect model to solve the problem and does not attend on how to estimate the default correlation. Domestic database construction about default risk is very backward and there is no directly available data about default risk. Therefore, default correlation is difficult to effectively measure and it increases the risk of domestic commercial banks management.Therefore, in this article we carried on special research on the measurement of default correlation. We got around the obstacle which is due to scarce data on default risk and, from this starting point, chose Copula approach to study in the paper. We chose the tail dependence coefficient (TDC) as an indicator of default correlation and sampled the stock returns of Domestic Corporation to estimate it. Thus we found a way that we measure default correlation with market risk data under the present condition.In this article, we firstly summarized the relevant researches on default correlation and appraised them combining the domestic reality. In the second chapter, we systematically introduced the base of theory about Copula which included its definition, character and common used Copula function. Then, we discussed the limitations of traditional Pearson correlation coefficient and described the advantage about copula function. Finally we led to the TDC which was used to describe default correlation. It demonstrated the use of the financial risk modeling methods Copula the general formula.In the chapter three, we started this paper from default risk and introduced its definition, classification and statistical features. Then, we proposed a definition and discussion of the reasons of its formation. This paper focused on default correlation model. Generally speaking, default correlation model is that a single model is embedded into a related structure, so one-dimensional model become two-dimensional even higher dimensional model. Therefore, this paper firstly introduced one-dimensional model then embedding a related structure, thus made model from the one-dimensional to two-dimensional. As the Copula function had unique advantage on describing the correlation, we commonly chose Copula function as the embedding related structure. Finally, because of the scarcity of data on default risk, we would not directly start our work from the modeling. Instead, we would find a index to measure default correlation and broke away from. This is not the plight of the scarcity of data from the data do not, but use data to measure the market risk of default correlation. We pointed out the shortcomings of traditional measurement methods, and the assessment method proposed in this paper. The correlation coefficient is used as a default related to the tail of metrics the scarcity of data. This method also needed data, but it would use data of market risk to measure default risk. We pointed out the shortcomings of traditional measurement methods and proposed the method that we used TDC to describe default correlation in this paper. Tail dependence was that correlation between the tail data in two-dimensional distributing. It has some unique advantages when describing correlation between financial market variables. For this reason, this article would use it to describe correlation between enterprises. Specifically, it is mainly based on the following three points: (1) Default is essentially a kind of tail dependence. (2) The TDC could break away from scarce data on default risk. (3) The use of TDC could make the problem incorporate into the system to analyze Copula function.In chapter four, according to chapter two and chapter three, we knew that the method that the two previous chapters mentioned led a stable estimate on TDC. Then we verified their simulation, and used the domestic-listed companies with different credit ratings for Empirical Analysis of the samples, then comparing the difference about default correlation between different credit ratings. Eventually we came to the conclusion that because of scarce data on default risk in our country, we could use the data of market risk especially proceeds of stock as a replacement. We could get estimate of TDC between different companies and this estimate in a certain extent reflect default correlation.