Dynamic models for clean energy technology diffusion and improvement

Thesis

Mitigating global warming is one of the most significant structural transformations under- taken by society. The energy sector is central to this endeavour since it currently accounts for over 73% of global greenhouse gas emissions and trillions of USD in fossil fuel assets. Increased adoption of clean energy technologies, such as solar photovoltaics, onshore wind, and battery-electric vehicles plays an central role in the energy transition by displacing incumbent fossil fuel technologies.

In this thesis, I study dynamic models of future clean energy technology diffusion and cost in the face of uncertainty. I answer key questions about the energy transition whilst advancing statistical models of technological change.

In my first paper, I critically examine recent models which have forecasted limits to the diffusion of solar photovoltaics and onshore wind, finding that many results are not robust due to underlying uncertainty and the lack of statistical rigour in the assessments. Global solar and wind deployment have not significantly departed from their exponential growth trajectory so far, and there are no reliable signs when they will do so.

For my second paper, I apply more reliable statistical methods to develop forecasts for national solar and wind costs. Based on their decomposition, solar costs will likely decline further, but wind may soon reach its floor cost. Solar levelised costs of electricity in different countries will most likely converge in the future due to a high degree of correlation between countries. Wind costs, however, will likely continue to display cross-sectional variations of almost an order of magnitude due to differences in natural resources.

My third paper uses similar forecasts to develop dynamic investment strategies in technology portfolios under uncertainty and learning. For instance, in the case of battery- electric passenger vehicles, assuming low discount rates and negligible switching costs, we are now at a tipping point where it is no longer cost-optimal to invest in traditional combustion-based vehicles. However, with an discount rates above 2%, continued policy support is required if the transport sector is to reach this tipping point.

Authors: Baumgärtner, CL

• DOI: 10.5287/ora-6q6gmmrqd
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: energy transition; technological change; forecasting; multi-armed bandits; S-curves; economics; learning curves
• Subjects: Dynamic control; Sustainability; Economics; Technological complexity; Business forecasting