Research Of Systemic Risk Contagion Of Banks Based On Complex Network Theory

The financial crisis outbreak in2008in the United States had caused significant economic losses to the world financial system. In the wake of the crisis, people began to learn from their mistakes, ponder the reasons for the financial crisis caused by the loss of a large area in the world. The study found that it is mainly the security of the world financial system, there are hidden dangers. Therefore, it requires people to study the risk from micro-level to the macro level, and each country own financial system has to strengthen macro-prudential management to prevent the next financial crisis spread widely. This is priority. From a macro point of view, we should take all financial institutions involved in the organization of economic activity as a whole to consider, in holistic perspective to study and prevent contagion in the system. To the overall thought there are a lot of methods to study the risk contagion in the system in the overall thought, but in recent years, using method of complex networks has become highly regarded. The European central bank said in a financial stability review once pointed out that modern society the relationship between factors is complex in the financial system. To research the complex relationships, you need to use the complex network as research methods.Complex network is made up of mathematical graph theory and in the light of the development of a new discipline, it covered contents is wide, mainly include the systematic science, management science, computer science, mathematics, statistics, and so on many aspects of theories and knowledge. Because of this, complex network is only recognized by scholars as one of the most effective tool in the study of complex systems. This mainly abstracts complex systems to be a kind of complex networks. In the financial system, there are a large number of financial institutions, financial organization, the organization plays a different role in the system, and often occur some business dealings between organizations. The business relationship, however, is not immutable, but as time goes on changing. From this perspective, we can say the financial system is a complex system and use the complex network theory to analysis and research.The financial institutions in the financial system are related mainly through the capital flow, and capital is mainly flowed through the payment system. So we can say that the payment system corresponding to the financial system will show the characteristics of the complex. This paper base on large-Value payment system to establish network model to estimated the transaction among banks. Then, use the complex network theory to build banking network model system to analyze the nature of the topology of the network and the network structure. Finally, the paper according to the network structural characteristics analyzed the systemic risk propagation in the network.Specific contents are as follows:The first part is a literature review. Firstly, tidied analyze the network structure of the bank. The introduction mainly from four aspects like systemic risk implications, specific measurement of systemic risk of scholars, research methods of systemic risk, the possibility of systemic risk.The second part of this paper describes the data sources. Data from the three main parts, payment system development report, the bank’s annual report, and a literature of Ouyang Weimin. Because of the data exposed not fully, this paper attempts to use regression analysis, the variable alternative methods to estimate the proportion of transactions of each bank amount the payment system. Then further estimate the total volume of transactions in the payment system between any two banks.The third part is a brief overview of the theory of complex networks. Then analyzes the complexity of the banking system and conclude we can use complex network to analyze, finally, according to the previous estimate matrix transaction to analyze the structure of transaction network mainly on network topology and network structure. The main parameter of the network topology is the average path length and degree distribution. Analysis is to analyze the network structure is mainly on network clustering coefficient and matching pattern. The results show that the network has the characteristics of a small world, that is, the network has a small average path length and larger clustering coefficient. The fourth part describes the specific methods to describe the inter-bank contagion. The most important is to net between the banks not bankrupt and bankrupt and then to compare with the assets of the bank to judge a bankruptcy, through an iterative process, to observe the effects of the crisis. Finally simulation by Matlab by assuming a bank, two banks, a number of bank failures try to compare the results. From the result we can see our banking system is very safe, without external factors there will no systemic risk.The fifth part is the conclusion and outlook. This part introduces the main conclusions and pointed out the shortcomings and limitations and then prospect further research.The main innovation of this paper there are two. The first point is to estimate data. As we know, the transaction between banks is not easy to collect, so this paper tries to find a relative factor to substitute transactions. There is also partly use the idea of maximum entropy estimate that is take all the information known into account. The second point is to analyze the process of contagion. First establish a network model of the system in the process of analyzing contagion, then according to the network abstract a corresponding matrix. Contagion process is a corresponding change in the matrix elements. That is the bank not bankruptcy now net the insolvent bank before, and then compared to the asset to see whether the bank will bankrupt. Through continuous iterative process observe the effects of the crisis.