Comparing expert elicitation and model-based probabilistic technology cost forecasts for the energy transition

Abstract

We conduct a systematic comparison of technology cost forecasts produced by expert elicitation methods and model-based methods. Our focus is on energy technologies due to their importance for energy and climate policy. We assess the performance of several forecasting methods by generating probabilistic technology cost forecasts rooted at various years in the past and then comparing these with observed costs in 2019. We do this for six technologies for which both observed and elicited data are available. The model-based methods use either deployment (Wright’s law) or time (Moore’s law) to forecast costs. We show that, overall, model-based forecasting methods outperformed elicitation methods. Their 2019 cost forecast ranges contained the observed values much more often than elicitations, and their forecast medians were closer to observed costs. However, all methods underestimated technological progress in almost all technologies, likely as a result of structural change across the energy sector due to widespread policies and social and market forces. We also produce forecasts of 2030 costs using the two types of methods for 10 energy technologies. We find that elicitations generally yield narrower uncertainty ranges than model-based methods. Model-based 2030 forecasts are lower for more modular technologies and higher for less modular ones. Future research should focus on further method development and validation to better reflect structural changes in the market and correlations across technologies.

Probabilistic feasibility space of scaling up green hydrogen supply

“…Second, the growth rate is an inherently uncertain function of policy support 46 , technological characteristics 47 and possible cost reductions 44 , of which the latter are notoriously difficult to predict 48 . As the annual growth rate gradually decreases due to market saturation, we parameterize the emergence growth rate 26 , which is the maximum annual growth rate that is realized after the formative phase, related to the steepness parameter in the logistic function (Methods).…”

The quantity-quality transition in the value of expanding wind and solar power generation

“…Lastly, we use technology costs taken from the U.S. Energy Information Administration (U.S. Energy Information Administration (EIA), 2020) and Lazard's levelized cost of storage report (Lazard, 2019). We use the Energy Information Administration as an objective source for cost estimates; note, however, that a growing body of research is finding that the cost of clean energy is falling faster than has been forecasted by experts (Meng et al, 2021). As these costs decline and the ratio of costs between the different clean technologies changes, the exact quantity of each technology that would be deployed in a least-cost system would change.…”

Understanding the complexity of existing fossil fuel power plant decarbonization

“…Such a lack of bottom-up experience accumulation accounts for the main reason why certain key modeling parameters like capture cost diverges across different studies ( Anadón et al., 2017 ). As a result, some future technology projections in top-down energy system models come from expert elicitation with bias, thus undermining the credibility of energy system models ( Meng et al., 2021 ). Moreover, capital investment in power plant retrofitting is usually very large.…”