A Novel Complex Network based Credit Risk Management Strategy

摘要

Financial system exhibits a high degree of interdependence. Different types of interconnections exist among financial agents, such as loanguarantees and cross share-holding. Financial contagion could diffuse along the interconnections and lead to wide spread of risk. Thus how to effectively maintain the stability of financial system and prevent financial contagions from propagating is a crucial problem in financial risk management. In this paper, we present a novel risk management strategy based on complex network theory with regard to the guarantee bank loans in China. First, we transform the records of guarantee loans in to directed guarantee networks, with nodes representing companies and edges representing the guarantee relations. With regard to the guarantee networks, we define two types of risk measurements for each company. The first one is the risk spreading ability of each company when it defaults. The second is the exposure risk each company is faced with. These two measurements together decide the credit risk of companies. Second, to evaluate these two types of measurements, a novel directed k-shell decomposition method is presented. It is an effective way to measure the node centrality in terms of both in-degrees and out-degrees. The spreading abilities of companies can be obtained when the out-degrees of the companies are considered in the novel directed k-shell decomposition method, while the exposure risk of companies can be obtained when thein-degrees are considered. Experimental analysis shows that the directed k-shell decomposition method could identify meaningful companies in guarantee networks. Companies with high exposurerisk are more likely to be infected and thus default. Meanwhile companies with great spreading abilities could lead to wide spread of financial risk. Thus with our strategy, the financial regulators are able to monitor and immunize the targeted companies to maintain the stability of financial system.