ESG disclosure, audit quality, and corporate performance: an analysis through a machine learning approach

This study investigates the relationship between environmental, social, and governance (ESG) disclosure and corporate performance (CP), as well as the moderating effect of audit quality (AuditQ). The empirical data comprise a balanced panel of 71 non-financial companies traded on the Pakistan Stock Exchange (PSX) during 2015–2024, yielding 710 firm-year observations. ESG disclosure is analysed empirically through an overall disclosure index, and other covariates are employed to explore direct and moderating effects on corporate performance, such as gender diversity (GD), board size (BS), firm size (FS), firm age (FA), leverage, liquidity, and macroeconomic indicators. This comparative study uses dynamic panel models, System Generalised Method of Moments (System GMM), support vector regression (SVR), and artificial neural networks (ANNs) to detect possible non-linearity. The results show that ESG disclosure has a statistically significant and positive impact on ROA and ROE. Audit quality reinforces the link between ESG performance and gender diversity, board size, and firm size, further supporting the impact of these factors. The findings confirm the hypotheses that ESG disclosure is embedded in corporate strategy and highlight the importance of high-quality auditing in improving corporate performance. Thus, the findings have implications for corporate managers, auditors, investors, and policymakers. Managers should embed ESG disclosure into their decision-making processes to enhance long-term corporate performance, and auditors should strengthen the credibility of sustainability reporting.