网络结构与银行系统性风险

Network structure and systemic risk by banks

建立了完整的度量银行间违约传染及银行系统性风险的研究框架.在这个框架下,研究了不同银行间网络结构下银行系统性风险.在模型建立过程中,分析了现有研究广泛采用的违约算法中存在的问题并对其进行了修正.为了模拟不同的银行间网络,还提出了一种构造无标度网络的方法.通过仿真模拟,研究发现集中度越高的网络由于传染而倒闭的银行数量越多.但是,当基础违约的银行数量不多时,网络集中度越高,由于传染而倒闭的银行的总资产越少.此外,在集中度高的网络中大银行倒闭引发违约传染的可能性和影响力都会大于集中度低的网络.而小银行引发传染的可能性远低于大银行,但是小银行倒闭达到一定规模时,可以引发大银行传染倒闭.

In recent years, systemic risk by banks has become an important issue in finance research area. Inthis paper, we build a novel research framework to measure the inter-bank defauh contagion and systemic risk by banks. Under the framework, the problem of default algorithm in the present researches is pointed out and rectified. Moreover, the way of establishing scale-free network is put forward to simulate the inter-bank net- work. It is found that ： more banks are infected in a higher-centralized network ; the total assets of failed banks are less in higher-centralized networks when there are less banks suffering fundamental defaults. Furthermore, the influence of default contagion caused by large-bank failures in highly-centralized networks is bigger than that in low-centralized networks. Although small-bank failures do not easily bring about contagion, reaching a certain scale, they will also lead to large-bank failures contagiously.