Relation Credit Risk Modeling And Simulation Based On Complex Network

Credit risk refers to the risk that the debtor may default, it is always troubling financial institutions, especially in the banking industry. Until the late 1990 s, academic research on credit risk mainly focus on the credit risk assessment model of single subiect,such as the KMV model and the intense model. But the Asian financial crisis let the world know the consequence of credit risk contagion, so credit contagion caused great concern in the financial institutions、supervisory department and the researchers.In the late 1990 s, Davis & Lo built the first credit contagion model, through 10 years’ development, many researchers did a lot of improvements to credit contagion model. But the financial crisis in 2008 let us know that the model still have a lot of space to improve. In order to measure credit risk correctly, we must study the effect of credit contagion, so this paper is meaningful in theory and practice.As the securities market is a complex system, as we know more about the complex network we found than the structure of the network has much effect on the transfer of information and subjects, so in order to study the credit risk contagion we must consider its network structure. Previous studies on the network structure used to assume it obeys a certain network, such as the small world networks or the scale-free network. But these network structures may not be necessarily associated with the true credit risk. So considering the process of the evolution of credit risk contagion depends heavily on its network structure, it is necessary to determine the characteristics of associative networks of the real credit risk.For this reason this paper choosing the study of the credit risk network structure as a starting point, investigates the evolution mechanism evolution mechanism of credit risk contagion in the complex system. The main research work is as follows:Firstly, we take a review of credit risk model and credit contagion model, then we introduce some knowledge of complex network.Secondly, by using the KMV model and the credit rating data,this paper calculate the changes of listed companies’ credit risk. Then by using the time series association rules algorithm to study the listed companies’ credit risk associated network and its network structure.Thirdly, under the framework of factor model and on the assumption that the industries’ credit conditions can change in a heartbeat with the gathering of industries’ credit risk, this paper establishes the credit risk contagion model based on trade association, and take a simulation by using matlab.Finally, on the basis of the second and the third part, this paper did a simulation under the situation of complex network and studied the impact of different network structure to credit risk contagion.The outcome shown that,with the rise of connection probability the contagion is more serious.And the size of network can provide some kind of risk dispersing effect.To sum up, by using the time series association rules algorithm this paper firstly determined the network structure of credit risk network, then built a credit risk contagion model on the basis of factor model, finally took a simulation of credit contagion under the situation of complex network,by using the simulation we studied the effect of changes of network structure on the process of credit contagion. The conclusions can provide theory evidence and empirical reference for regulators and financial institutions to implement risk controls...