Predicting ESG disclosure quality through board secretaries' characteristics: A machine learning approach

Abstract

This study employs a machine learning approach to explore the relationship between board secretaries' (BS') characteristics and the quality of Environmental, Social, and Governance (ESG) disclosure in enterprises. The results indicate that BS' competence characteristics are the primary drivers of ESG disclosure quality, with salary, functional experience, age, and tenure as key predictive features. BS' salary and functional experience show an upward trend with ESG disclosure quality. BS' age has a nonlinear predictive effect on ESG disclosure quality, exhibiting an approximately positive "U"-shaped relationship. BS' tenure also demonstrates a nonlinear predictive effect. Robustness checks confirm that BS' characteristics remain crucial factors even after controlling for CEO characteristics. Interaction effect analysis reveals that the impact of BS' salary, functional experience, and tenure on ESG disclosure quality is more significant among middle-aged board secretaries. Heterogeneity tests based on ownership type, institutional environment, and media attention show that the importance of BS' characteristics varies across different contexts. This study provides a new perspective and empirical evidence for predicting ESG disclosure quality, offering guidance for improving the appointment process and incentive mechanisms for board secretaries.