Green Finance, Clean Energy, and AI Synergies in Global Shipping Markets

This paper aims to improve out‐of‐sample forecasting of global logistics risk, as reflected by the Baltic Dry Index (BDI) returns and volatility, using sustainability‐finance, energy‐market, and technology‐related indicators as candidate predictors. A secondary objective is to describe time–frequency co‐movements between BDI and these indicators. Using a hybrid framework that integrates Automated Machine Learning (H2O AutoML) with Quadruple Wavelet Coherence (QWC), the study evaluates nonlinear predictive performance under strict out‐of‐sample validation and captures both nonlinear predictive relationships and time‐frequency dependencies across financial, environmental, and technological dimensions. Model‐interpretability tools are used for predictive attribution, dedicated to explaining the model's predictions. The forecasting results indicate that crude oil prices (WTI), green bonds (GB), and the AI market index (AI) are among the most informative predictors for out‐of‐sample prediction of BDI volatility and returns, while clean energy (ICLN, SPCLEAN) and natural resource (GNR) indices provide additional incremental predictive information. The QWC analysis documents frequency‐specific coherence between BDI and sustainability finance, energy, and technology‐related indicators, consistent with time‐varying co‐movements across short, medium, and long‐horizon bands.