Many believe that nuclear power could become an important low-carbon source of energy as efforts to mitigate climate change accelerate. But it's hard to predict the long-term costs as the technology evolves. With this in mind, several studies have garnered together expert views to inform policy-making, in what's known as expert elicitation. Now a team in the US and Italy has found that an expert's estimate of the cost of nuclear power in 2030 depends on whether he or she is an academic, in the public sector or industry.

The average expected cost of nuclear technologies in 2030 was around $4800 per kW, with estimates ranging from $506 to $14,156 per kW. Experts in the public sector expected costs to be 14% higher than academics did, whilst industry experts were less optimistic still, predicting costs 32% above those anticipated by their university colleagues.

Location was a factor too. Experts in the US gave median expected costs 22% lower than their European counterparts, and greater uncertainty ranges. The team believes that Europeans could have anticipated higher costs because the EU has more recent experience building nuclear power plants than the US, and these projects have suffered cost over-runs.

"Expert selection – particularly the sectoral background and country – and elicitation design, particularly the level of technology disaggregation, have a statistically significant impact on the outcome of elicitations in nuclear technologies that is robust to different empirical specifications," Laura Díaz Anadón of Harvard University told environmentalresearchweb. "Not considering elicitation design characteristics leads to a 25% lower estimate of public nuclear RD&D [research, development and demonstration] returns."

On average, experts reckoned that a doubling of public RD&D would reduce the cost of nuclear technology in 2030 by 8%, although the uncertainty was large. Public investment was not expected to reduce the uncertainty regarding future costs, this first study of the relationship between public RD&D investment and the future costs of nuclear in expert elicitations found.

Uncertainty in future costs was generally high: experts gave a range of estimates that on average was 58% of their median estimate.

"Current estimates of the cost of meeting climate change do not appropriately represent uncertainty and may be underestimating the future costs of nuclear and overestimating the impact of RD&D," said Díaz Anadón. "This may be particularly true if academic or public institution experts provide estimates."

Along with Gregory Nemet of the University of Wisconsin-Madison and Elena Verdolini of the Fondazione Eni Enrico Mattei (FEEM) and CMCC, Italy, Díaz Anadón analysed data from three separate elicitations by groups at Carnegie Mellon University, US, Harvard University and FEEM. These studies saw 67 experts give their views on the costs in 2030 of large-scale Generation III/III+ nuclear power reactor systems, large-scale Generation IV systems, and small modular reactors.

This was the first use of meta-analysis techniques on individual participant data from expert elicitations on energy technology. "Previous work [looking at the returns to public RD&D] mainly relied on patents, while here we use cost estimates from experts," said Díaz Anadón. "We use a new method for extracting conclusions from multiple elicitations with an easier applicability for policy analysis."

Now the team, who reported their results in Environmental Research Letters (ERL), is working on additional meta-analysis research, including elicitations for other energy technologies.

Related links

Related stories