金融机构风险关联与行业差异研究——基于股市波动视角

Research on Risk Correlation and Industry Difference of Financial Institutions——Based on the Perspective of Stock Market Volatility

2021年8月17日,中央财经委员会第十次会议指出,目前防范化解重大金融风险攻坚战取得重要阶段性成果,要增强系统观念,统筹做好重大金融风险防范化解工作。网络关联性是系统性风险的重要组成部分,金融机构间错综复杂的关联结构使风险易在系统中传染,增加负面冲击的影响和范围。各机构和行业间股价波动溢出效应是网络关联性的直接体现,本文从股市波动视角结合随机森林模型构建中国2011年1月至2021年4月59家上市金融机构关联网络,研究结果表明,网络存在行业差异和分层,银行业是核心层,证券业起中介层作用,多元服务业处最外层。银行业是证券业和多元服务业溢出变动的格兰杰原因,证券业和多元服务业溢出变动互为格兰杰因果,股市震荡时期网络关联性更强。结合微观层面数据发现,资产负债率和机构波动溢出关系不显著,但高资产负债率使机构易受外部溢出影响,国家控股能够抑制机构波动溢出。建议根据机构和行业在关联网络中的地位,实行动态化差异监管,积极探索适合我国金融体系的风险监管框架。

Summray: While initial results have been achieved in preventing and resolving major financial risks, network relevance remains an important part of systemic risk. The intricate structure of relevance among financial institutions makes risk easily contagious within the system, with the impact and scope of negative shock significantly increased. Network relevance is directly reflected in the linkage mechanism of stock price fluctuation between and among various institutions and industries. From the perspective of stock market volatility spillover effect, this article adopts the random forest model to construct a network of 59 listed financial institutions in China from January 2011 to April 2021. It’s found that industry differences and stratification exist in the network, with the banking industry as the most important core layer, the security industry playing the role of the intermediary layer, and the multi-service industry being the outermost layer. The banking industry is the Granger causality for the fluctuation of the security industry and the multi-service industry, while the volatility of the security industry and the multiservice industry are Granger causality for each other, and the network connection is stronger during the period of stock market turbulence. Combined with micro-level indicators, it is found that the relationship between asset-liability ratio and institutional volatility spillover is not significant, but high asset-liability ratio makes institution vulnerable to external spillover, while state holdings can curb institutional volatility spillovers. The research conclusions suggest that more flexible and dynamic differentiated supervision should be implemented according to the status of various institutions and industries in the associated network, and the risk supervision framework suitable for China’s financial system should be actively explored.