本文在传统网络模型中加入去杠杆一降价抛售机制,研究以下两类宏观经济冲击对银行体系系统性风险的影响。从房地产贷款违约压力测试看,房地产贷款违约引起的传染风险是系统性风险的重要来源;传染损失比重和去杠杆次数结果则表明,2007年我国银行面临的传染风险最高,之后呈现快速下降的趋势;参数敏感性结果表明,网络模型中去杠杆、降价抛售以及破产对传染风险的相对重要性依次递减。从地方政府融资平台贷款违约压力测试看,大型商业银行受平台贷款违约的影响小于股份制和城市商业银行。此外,平台贷款违约概率存在阈值,在阈值之上银行损失和倒闭急剧攀升。基于银行倒闭压力测试,量化出本文的网络模型相对于传统网络模型的优越性。本文还发现中国金融体系的系统重要性与系统脆弱性指标的“错配”对于维持金融体系稳定非常关键。
Based on the traditional financial network model, this paper introduces the mechanism of bank deleveraging and fire sale which can be used to analyze two macroeconomic shocks on the systemic risk of banks' industrial system. In the perspective of the stress tests about the default of real estate's loans, this paper finds that contagion risk arising from the default of real estate's loans is the crucial source of systemic risk. The contagion loss ratio and deleveraging times show that the contagion risk of China's banks was highest in 2007, then it rapidly decreased. The results of parameters' sensitivity analyses demonstrate that fire sale is less relatively important than deleveraging in the network model for the contagion risk, and bankruptcy is less than fire sale. The second test which is about the default of local government financing platform's (LGFP) loans shows that large commercial banks (LCBs) is less affected by LGFP's loans than joint-stock commercial banks (JSCBs) and city commercial banks (CCBs) . Additionally, this paper finds that there exists a probability default (PD) threshold. When LGFP's loans' PD is above the threshold, banks' losses and bank failure numbers increase sharply. The last test which is about the single bank failure highlights the advantages of this paper's network model compared with the traditional network model. In addition, the test shows that the mismatch between systemic importance and systemic vulnerability is very crucial for China's stable financial system.