Discussion of “Can Corporate Governance Information Facilitate Accounting Fraud Detection? Machine Learning Evidence for Chinese Listed Firms”

Dong et al. (2025) investigated the incremental value of corporate governance information, as a form of nonfinancial information, in enhancing accounting fraud detection in China using machine learning techniques. They focused on Chinese listed firms from 2007 to 2019, employing the RUSBoost- DT algorithm to analyze the predictive power of raw financial data, corporate governance data, and a combined dataset, hypothesizing that incorporating corporate governance information alongside financial data improves fraud detection accuracy. Their study found that the combined dataset yielded a higherAUCscore (0.737) compared to using only financial data (0.718) or only corporate governance data (0.664). Feature importance analysis indicated that chairperson age, performance-based compensation, and ownership concentration are significant predictors. While the study offers valuable insights into the potential of nonfinancial data in fraud detection, there are spaces for improvement when considering the statistical and practical significance of the AUC increase. A more nuanced analysis of corporate governance’s relationship with accounting fraud within China’s evolving regulatory environment would also be interesting for the reader.