Answers are grounded in this paper and cite their sources.
Internal control deficiencies remain a pervasive challenge for listed firms, significantly undermining the stability and growth of capital markets. Although regulatory inquiry letters (RILs) are widely recognized as a critical tool for non-financial information disclosure, their proactive efficacy in identifying and mitigating internal control deficiency risks (ICDR) remains underexplored. This study examines whether RILs can effectively identify latent ICDR and reduce their recurrence in subsequent periods. Using a sample of Chinese listed companies from 2015 to 2022, we employ optimized machine learning algorithms to quantify ICDR as continuous probability values. The SHapley Additive exPlanations (SHAP) framework is further applied to rank the relative importance of risk factors and interpret the model’s predictive mechanisms. Our findings demonstrate that the XGBoost model outperforms alternative machine learning algorithms in predicting ICDR, achieving optimal balance between precision and recall. Empirical evidence confirms that RILs not only serve as an effective mechanism for identifying ICDR in advance but also exert a substantive supervisory effect, reducing the likelihood of future deficiencies. These results offer novel insights for the early detection of internal control deficiencies, providing actionable implications for managers, regulators, and investors in evaluating corporate operational performance and mitigating systemic risks.
No takes yet. Share an insight, caveat, or question.
Cai et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: