A machine learning model to investigate factors contributing to the energy transition of utility and independent power producer sectors internationally

There is evidence of independent power producers dominating the electricity sector's uptake of renewable energy, with utilities lagging behind. Here, we build a machine-learning-based model with multiple dependent variables to simultaneously explore environmental policy and market structure contributions to investment patterns in different technologies by utility and independent producer sectors across 33 countries over 20 years. With the analysis enabling the capture of non-linear relationships, our findings suggest substantial resistance of gas capacity to even strict carbon pricing policies, while coal appears more responsive. There is also an indication of policy pricing in effects. The positive link of renewables subsidies and fossil fuel disincentives to renewables expansion, particularly wind, is more prominent for independent power producers than utilities. Regarding market structures, different characteristics tend to matter for renewables growth compared to fossil fuel reductions. The results also suggest considerable differences in policy and market factor contributions to technology choices of Organisation for Economic Co-operation and Development vis-à-vis emerging economies.