模型不确定下我国商业银行系统性风险影响因素分析

Analysis of Systemic Risk Determinants of China's Commercial Banks——Based on the Model Uncertainty

本文采用成分期望损失CES方法,基于公开市场数据,对我国16家上市商业银行的系统性风险进行度量。基于CES的方法无论从时间维度还是横截面维度上,都与我国银行业的实际情况有着较好的契合。本文还采用贝叶斯模型平均BMA方法,广泛纳入现有相关文献中选取的影响因子作为解释变量,解决传统回归中的模型不确定性。研究结果表明,对于我国上市商业银行而言,银行规模、股权市账比及是否处于系统重要性地位与银行系统性风险呈现出显著的正相关关系,而非利息收入的提高能够有效地分散系统性风险;在行业层面,银行系统的波动率越高,单个机构面临的系统性风险也越大。以上结论可以为银行监管部门政策制定提供较为明确的启示及实证支持。

This paper measures the systemic risk of China＇s listed commercial banks totally based on open market trading data with Component Expected Shortfall model. The CES model corresponds to the actual changes of systemic risk in China＇s Banking Sector both in the time series and from the cross sectional perspective. To solve the uncertainty problem of regression model in the most possible risk factor analysis, we adopt Bayesian Model Average method and choose the risk factors from highly related literature as independent variables. The results show that bank size, market to book value and whether it＇s systemically important financial institution are positively related with systemic risk; the increase of non-interest income, however, can effectively mitigate systemic risk. In the banking sector, individual banks have higher systemic risk when the volatility of the banking system increases. These results provide policy implications and empirical evidences for China＇s banking sector regulators.