The role of ESG score in forecasting China corporate green bond issuance: Evidence from tree-based learning models

• Environmental, Social and Governance scores strongly predict green bond issuance volume in China. • ESG scores interact more with financial traits in green bond issuers than conventional bond issuers. • ExtraTrees model outperforms other machine learning methods in forecasting bond issuance volume. • SHAP analysis reveals ESG scores as the top driver for green bonds, unlike financial factors for conventional bonds. • Findings aid policymakers in designing incentives for sustainable finance in emerging markets. The increasing importance of Environmental, Social, and Governance (ESG) factors in financial decision-making has raised critical questions about their role in influencing green bond issuance. This study investigates the impact of Environmental, Social, and Governance (ESG) scores, corporate financial characteristics, and market sector attributes on the issuance volume of green bonds in China. To analyze the driving factors behind corporations' choices between issuing green bonds and conventional bonds, the study employs a comparative approach that distinguishes between an experimental group (green bond issuances) and a control group (conventional bond issuances). The methodology centers on tree-based ensemble learning models, including Random Forests, ExtraTrees, XGBoost, LGBM, and CatBoost, to predict bond issuance amounts with high accuracy. Among these models, the ExtraTrees algorithm emerges as the best-performing model, offering superior predictive power. A robustness evaluation of the ExtraTrees model is conducted under various real-world data disturbance scenarios, including additive noise, distributional shifts, adversarial perturbations, and stress-based rescaling. Additionally, the SHAP technique is applied to interpret the model’s predictions and validate hypotheses. The results reveal that ESG scores are highly significant in predicting the issuance volume of green bonds, outperforming their predictive value for conventional bonds. Based on these results, the study offers specific policy implications to enhance the effectiveness of green finance initiatives. Targeted support for sectors with high green financing potential can further amplify the adoption of green bonds and integrating ESG metrics into regulatory frameworks and investor decision-making processes can strengthen the alignment between corporate financing strategies and sustainability objectives.