The impact of uncertainty, financial market dynamics and commodity indices on green asset price forecastability: evidence from advanced data-driven methods

Purpose This paper investigates the predictive relationships between four major green assets and several global factors. It aims to identify the most accurate forecasting model for each green asset studied and to quantify the predictive contribution of each influencing factor. Design/methodology/approach This study uses daily data from January 2018 to December 2024 for four green assets: the S&P Green Bond Index, the S&P Global Clean Energy Index, the S&P GSCI Water Index and the MSCI Global Environment Index and several global factors, including the MSCI World Index, the VIX, the S&P U.S. Aggregate Bond Index and key commodity indices. For the methodology approach, we use the Shapley Additive Explanations (SHAP) analysis is used to identify the contribution of each global factor to the forecastability of each green asset. Furthermore, machine learning methods (Random Forest, XGBoost, LightGBM) and deep learning models (RNN, LSTM, GRU) are used for prediction purposes. Findings Results indicate that deep learning models frequently outperform traditional machine learning methods. However, the best-performing model differs across green assets, reflecting the heterogeneous dynamics of green financial markets. Furthermore, SHAP analysis shows that aggregate bonds and developed equity markets emerge as the main drivers of green asset movements, whereas commodity markets show limited influence. Research limitations/implications These findings can guide investors and policymakers in promoting portfolio stability and the development of green finance. Since aggregate bonds and developed markets are the main drivers of green asset prices, investors can use them as useful signals when choosing green investments, while policymakers should work on building stronger connections between traditional and green financial markets. Originality/value This study offers a novel integration of deep learning forecasting and SHAP-based explainability applied simultaneously to multiple green asset classes based on different global influencing factors. It contributes to the green finance literature by providing a comprehensive, data-driven framework for understanding the dynamics of green asset prices across a recent, policy-relevant period.