Regulating risk culture in the insurance industry using machine learning
通过文本分析和机器学习研究保险业风险文化与监管的关系，发现风险文化受不确定策略、风险定义与报告约束、诉讼决策及风险管理实践影响，且大型保险公司风险文化恶化后更难逆转。
Abstract This study examines the relationship between risk culture and regulation in the insurance industry using textual analysis and machine learning. By analyzing 10‐K disclosures, we classify firms into distinct risk culture clusters and find that the risk culture of insurance firms is significantly shaped by their uncertain risk strategies, constraints in defining, implementing, and reporting risks, as well as litigious decisions and risk management practices. A temporal prediction analysis indicates that large insurers maintaining a poor risk culture trend are less likely to reverse it compared to those improving. Moreover, insurance firms show enhanced risk culture post‐Dodd–Frank Act. Our findings underscore the potential benefits of regulations aimed at monitoring and overseeing insurers' risk practices.