Machine Learning Models for Optimal Debt Capital Structuring in Renewable Energy Firms

Abstract

The growing complexity of financing renewable energy projects has intensified the need for data-driven approaches to optimize capital structures, particularly in determining the ideal debt-to-equity mix. This study applies four machine learning (ML) models—Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting Machine (GBM), and Artificial Neural Network (ANN)—to predict and optimize the debt ratio of renewable energy firms using quantitative financial and macroeconomic data. A dataset comprising financial records from 120 renewable energy firms across Europe, Asia, and Africa between 2013 and 2023 was analyzed. Eight key variables, including profitability, asset tangibility, firm size, growth opportunities, tax shield, interest rate spread, policy stability, and volatility, were used as predictors, with the debt ratio (total debt/total assets) serving as the dependent variable. The dataset was partitioned into training (80%) and testing (20%) subsets, and model performance was assessed using R², RMSE, and MAE metrics. Results showed that the ANN model achieved the highest predictive accuracy with an R² value of 0.93, RMSE of 0.042, and MAE of 0.038, outperforming the GBM (R² = 0.88), Random Forest (R² = 0.86), and SVM (R² = 0.79) models. Feature importance analysis revealed that profitability accounted for 27.4% of the total model variance, firm size for 21.8%, and interest rate spread for 18.6%, while tax shield and policy stability contributed 12.3% and 10.7%, respectively. The results indicate that profitability and firm size have the strongest positive influence on optimal leverage, whereas rising interest rate spreads and unstable policy environments negatively affect debt structuring decisions. The study concludes that ML-driven approaches, especially artificial neural networks, provide a powerful and accurate framework for optimizing debt capital structure in renewable energy firms. By capturing nonlinear relationships among financial variables, these models enable more precise and adaptive financial decision-making, ultimately supporting cost efficiency, investment stability, and sustainable growth in the renewable energy sector.