Green finance and the socio-politico-economic factors’ impact on the future oil prices: Evidence from machine learning

This paper suggests an innovative method of estimating crude oil prices based on multiple socio-politico-economic factors in context of green finance using the Least Absolute Shrinkage and Selection Operator (LASSO) model. This work also examines the relevance of six factors (commodities market factors, geopolitical factors, supply, demand, and financial market factors), in addition to green finace, in evaluating several forecasting models and identifying statistically essential factors for predicting future oil prices. We implement state of the art LASSO model on a data set of the abovementioned factors (26 variables). The proposed model is evaluated against four bench mark models (traditional statistical models (OLS, GARCH), EIA and artificial neural networks) at different time steps (1 step, 3 steps, 6 steps, and 9 steps). Statistical analysis of outcomes shows that the LASSO technique produces better predictions than other benchmark models. The results also deliver detailed insights into the temporal association between numerous socio-politico-economic factors, green finance and crude oil. Our findings indicate that the global output of steel, the Kilian index, the Institute for Supply Management index, green finance index, the value of the dollar, and the frequency of terrorist strikes in Central East and Northern Africa are important demand drivers. These elements, taken as a whole, are more significant than supply and speculation. We also find that no variable from the supply factor is essential in determining the future oil value. Our results are significant for government and policy makers to gain insight into future oil prices in the context of various social, economic, and political factors.