This study aims to identify the key determinants of carbon emissions disclosure from an environmental, social and governance (ESG) perspective, offering insights into how these factors influence corporate transparency and sustainability practices.
This study uses H2O Automated machine learning (AutoML), a sophisticated machine learning framework, to analyze CO2 emissions disclosure among 77 French nonfinancial companies listed on the SBF 120 index between 2017 and 2021. This investigation robustly evaluates CO2 emission disclosures based on the Carbon Disclosure Project Index criteria. This approach enhances the accuracy of the findings and pioneers a new path in ESG research, blending sophisticated computational tools with traditional environmental reporting metrics.
The study shows an optimal balance between model complexity and accuracy, with social factors and the book market being more influential in CO2 disclosure than direct environmental factors. The heatmap analysis revealed the significance of these variables in predicting CO2 disclosures.
This research provides insights for firms and policymakers to improve environmental transparency and reporting, emphasizing the importance of considering ESG aspects. Carbon emissions disclosure is crucial for sustainability, ensuring regulatory compliance, attracting investors and improving risk management.
This research introduces a cutting-edge methodology for analyzing CO2 emissions disclosure, applying the H2O AutoML framework specifically to French nonfinancial companies listed on the SBF 120 index. This unique application within the French regulatory context, combined with a focus on ESG factors, sets this study apart from previous research. By emphasizing model diversity and the integration of multiple advanced algorithms, the approach provides a more nuanced understanding of environmental disclosure, offering novel insights that can guide policymakers and businesses in enhancing transparency and sustainability practices.
