Growth in natural gas capacity is ‘substantially’ less affected by environmental policies, such as carbon pricing, than coal, concludes a new University of Oxford study published in iScience.

The findings are cause for concern because, although gas can play a useful role in balancing grids which use intermittent renewable energy sources, an expansion in gas capacity increases the risk of carbon lock-in (a state of inertia in which fossil fuels continue to be used long into the future). It also exposes companies to the risk of stranded assets – resources which cease to have value because of changes in policy or technology.

The study employed a machine learning model – the first of its kind to be used within energy transition research, examining 10 factors affecting technology investment patterns by utility and independent power producer sectors across 33 OECD and emerging economies over 20 years. It uncovers the impact of electricity market characteristics like competitiveness, and environmental policy instruments, including renewable energy subsidies and carbon pricing policies.

Dr Ben Caldecott

The results suggest that power markets with a smaller utility sector’s market share tend to reduce growth in utilities’ coal portfolios, but not in gas. This signals that utilities in more liberalised markets might be shifting their portfolios from coal to gas.

“An important conclusion of our study is that the relationship between environmental policy and the technology choices made by electric companies is not linear. We see the effect of some policies on certain technologies level off as the policies become harsher. This is, for example, the case for emission standards on utility companies’ coal capacity,” said Galina Alova, study lead author and researcher at the Oxford Sustainable Finance Programme.

These results point to policies being factored into decision making within the energy market, either indefinitely or until they become even tougher. “This is a useful consideration for forward-looking climate and energy scenario modelling aiming to inform policy makers, investors and corporates in their efforts to decarbonise the global economy,” added the researcher. The study also finds that four out of the top five factors affecting growth in renewables among independent power producers are policy instruments. But for utilities, electricity market characteristics tend to matter more.

Utility-dominated markets do not necessarily impede, and could contribute to, growth in renewables. This is the case in Scandinavian countries, where utilities tend to lead renewables-based power generation. These findings suggest that renewable energy growth and the shift away from fossil fuels might not always go hand-in-hand, and require a careful and nuanced design of electricity markets and environmental policies.

The machine-learning model also uncovers a positive link between investor ownership of utility companies and growth in utilities’ solar capacity. This finding serves as an indicator of the increasing pressure from investors and lenders for utilities to disclose and mitigate exposure to climate-related transition risks associated with carbon-intensive assets.

“Future research could focus on tracking and examining the effects active ownership and engagement activities could have on both clean technology investments and reductions in fossil fuel capacity,” added Alova.

A machine learning model to investigate factors contributing to the energy transition of utility and independent power producer sectors internationally was written by Galina Alova and Dr Ben Caldecott as part of the Sustainable Finance Programme at the Smith School, University of Oxford. Led by Dr Ben Caldecott, the centre has pioneered research on, among other things, stranded assets and spatial finance, and works across many of the key areas of sustainable finance.