Prediction of market value of firms with corporate sustainability performance data using machine learning models

This study attempts to build models for prediction of market value of firms with Corporate Sustainability Performance data using machine learning models. We analyze a comprehensive global dataset of 5,450 firms operating in 10 sectors. Machine learning models of Random Forest, XGBoost, SVM, and Nearest Neighbor models were constructed with E,S,G,C scores (Environmental, Social, Governance, and ESG Controversies) and financial ratios obtained from the Refinitiv (LSEG) Database. The most suitable model (Random Forest Model) built for Market Capitalization prediction shows that Environmental (E) and ESG Controversies (C) scores stand out as important predictors of market value. The findings of the study emphasize the importance of integrating ESGC factors into market value prediction models. Moreover, our findings suggest that the importance of corporate sustainability performance factors (E, S, G, C) is more pronounced in Europe and America compared to other regions. This study may provide insights for companies, investors, and analysts to achieve a more sophisticated assessment of market value.