Unveiling internal drivers of corporate greenwashing: a machine learning approach

Environmental, social, and governance (ESG) greenwashing has emerged as a critical issue, yet research on its prediction remains underdeveloped. This study builds a model to predict corporate greenwashing using machine learning. Using data on Chinese listed companies from 2009 to 2023, we examine features across four dimensions: finance, governance, chief executive officer (CEO), and chairman. The results indicate that organizational variables (finance and governance) outperform managerial individual traits (CEO and chairman) in both explanatory power and predictive accuracy. Non-linear models perform better than linear regression, especially eXtreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM). We further use the SHapley Additive exPlanations (SHAP) algorithm to evaluate feature importance, finding that total assets play a key role in predicting greenwashing. This study advances the literature on corporate greenwashing from a machine learning perspective, offering novel evidence that can inform regulatory oversight and promote corporate sustainability.