The investment decision making for new nuclear power stations is a long and arduous road for would be investors. The gigantic upfront costs of construction are amplified by financing costs, and these are driven up by the asymmetric risk profile associated with new build, where the upside potential is outweighed by the downside risk of delay, politics and regulatory interventions in a market characterised by price uncertainty.
But the complications of the design and construction are not the only issues facing investment panels and analysts. Energy production has another equally complicated and asymmetric partner – the decommissioning and environmental clean-up. These represent future uncertainties a century away: hard to grasp, difficult to estimate and which do not fit into the corporate finance techniques of investment bankers and policy makers with their twenty year event horizons.
The decommissioning and environmental clean-up costs are taking a greater role in the investment decision making than ever before. These costs were previously ignored, perhaps a deliberate ignorance. Too far away in the future, not thought to be mission critical to the need to generate power and material – something that could literally be put off for two or three generations. And anyway, the argument went, after more than 100 years, better technologies will certainly exist to more easily deal with the clean-up tasks. The situation is akin to ensuring that pension liabilities are correctly represented and properly funded – a task that has taken policy makers and accountants decades, and which still has many imperfections.
It is not surprising that decommissioning and clean-up costs were originally treated as a contingent liability instead of an actual one. They were simply invisible against the night sky, and played no part in investment decision making.
But the experience of decommissioning the first generation of nuclear plant has been a significant lesson on the gravity of the decommissioning and clean-up tasks. Now that the UK are engaged in the decommissioning and clean-up of the early nuclear plant it has become clear that these costs are heavy and dense and even though still far away they have significant influence.
At the UK’s NDA, in 2006 acting as NDA’s CFO responsible for the estimate, I struck the NDA’s figure at £30.6Bn discounted (£53.3Bn undiscounted). In the ten years since then, the discounted liability has increased at an average compound rate of 8.6% and now stands at £69.9Bn discounted (£117.8Bn undiscounted). The NDA describe this improved understanding of the liability as a ‘journey of discovery’. A journey illustrated in the charts.
The increase in the mass of these liabilities estimates seems to bend the laws of physics given that the NDA has spent over £30 billion between 2005 and 2015. The NDA in its 2015 accounts that estimates that ~80% of its spend is on the legacy. The other side of the decommissioning event horizon is a strange and unfamiliar place, where the relativity of time and cost is distorted, with negative discount rates applied for the first ten years by the NDA. In my experience, one reason for the costs refusing to come back down to earth is that the NDA in its short term budgetary cycle is forced to address the higher hazards rather than the best value for money elements of its programme.
Back in 1987, whilst working on an early attempt to evaluate the nuclear liability for BNFL, I created some Monte-Carlo simulations to give a probabilistic estimate, and when these were computed, it was a very substantial increase. The figure set out in 1987 increased from £72 million undiscounted to £135 million discounted. This was at a sensitive time, as consideration was then being given to privatising BNFL.
Lessons have undoubtedly been learned, particularly about the need to address decommissioning and clean-up tasks early and not leave plant to dilapidate, and also about ensuring an independent review of the costings to counter potential conflicts of interest.
Ten years of this high level of investment in decommissioning should be yielding innovation, learning through experience and deliver more effective and more efficient decommissioning. New investors will need to see that with investment in innovation and with more experience, these costs can stop growing and roll back down, particularly for relevant plant.
Policy makers and investors are now understandably keen to ensure that decommissioning and clean-up costs are properly considered in advance and appropriately funded. The UK’s Energy Act 2008 requires new nuclear build projects to create an external and segregated fund that will pay for the costs of decommissioning and environmental clean-up.
But in order to satisfy these new requirements, much more effort required to really understand what is happening the other side of the end of production event horizon. There is very little hard evidence to support key investment decisions, whilst the cash flows of creating the segregated funding now impacted hugely on the investment decision making. The decommissioning and clean-up cost estimates are now critical to the investment decision making, and they need firmer foundations to support them than are currently available.
The International Atomic Energy Agency say that almost 200 of the world’s 434 reactors in operation will be retired by 2040, but the IAEA lack cost information. Despite the growing number of closed stations in decommissioning, actual cost information is still almost impossible to come by, with no consistent accounting or approach.
There are enormous variations caused by differences in decommissioning strategies, estimates of future inflation, discount rates, employment practises, radiation levels, changes in scope and political and regulatory frameworks. Costs will be greatly affected if any particular plant has had accidents or contamination. Timescales chosen affect the discounted figures. The biggest variations are often caused by the costing approach to spent nuclear fuels, waste materials and the options for disposal, storage or treatment.
The recent March 2016 paper by the OECD recommends far greater cost transparency. Their report included the results of a questionnaire based on the International Structure for Decommissioning Costing (ISDC) structure, aimed to collect information on unit-specific cost estimates. They say that their review of the adequacy of projects is currently hampered by the limited amount of reliable and comparable information on decommissioning costs. All the data collected was from future planned activity, rather than costs of completed projects. Their report recommends ‘enhancing transparency around such costs and putting in place better methods to collect and share information.’
The UK is cited by the EU has being amongst the most transparent, where the NDA publishes extensive information about planned costs, with detailed lifetime plans available for all of their sites, and with extensive analysis in the annual report and accounts. However, the NDA estate contains the complications of research, reprocessing and military programmes, and the first generation of nuclear plant which were not designed with easy clean-up in mind. Their reprocessing facilities are unique, and their first generation gas cooled Magnox reactors do not provide a straightforward read across for potential investors in the new generation, and their experience of ever increasing costs is not likely to be repeated for newer plant designed with decommissioning in mind.
As for the UK’s AGR fleet, EDF has not published any recent detailed plans on how it expects to carry out the decommissioning and environmental clean-up. The most recent publicly available plans were published in September 2000 and a critical review of these plans by the British Government’s Nuclear Installations Inspectorate was published in June 2001. Each year the Nuclear Liabilities Fund publishes a liability estimate, the most recent being £8.96bn discounted (£19.5bn undiscounted) at 31 March 2015.
The EU Commission services published a report in 2007, “Comparison among different decommissioning funds methodologies for nuclear installations”. The European Commission estimates that approximately one third of the 145 power reactors currently operating in the European Union will need to be shut down by 2025.
The EU have a concern about ensuring that the funding of these decommissioning activities will be adequate and available when needed in order to avoid negatively affecting the safety of EU citizens. An updated report is expected in the very near future, with recommendations for more transparency, more consistent accounting for provisions, and the use of a common framework to allow comparison of decommissioning costs, which could be an estimated €268.3 billion for the European Union.
Policy makers demanding that investors commit funds to decommissioning is a step forward, as is their insistence that the costing be done independently. However, this is only useful if more accurate predictions of decommissioning and clean-up costs can be created to allow for competitive financing of new plant. If policy makers do not insist on a common approach and greater transparency on actual costs for projects ongoing, they are inflicting further uncertainties on investment decision-makers, which have to be reflected in pricing and guarantees, and make it difficult for Board’s to sign off on investments of such magnitude. It is hard to disagree with the EU Commission Services, IAEA, OECD and NEA on the need for much greater effort to improve the transparency and reporting of decommissioning costs.
For would be investors, the need to ensure funds are available for decommissioning and clean-up means that reliable benchmarks on cost estimates are both essential and overdue. We might not really know what is on the other side of the decommissioning event horizon, and that should spur greater efforts to go and find out. Greater study on actual cost experience will bring more certainty over cost estimates. and more confidence that decommissioning and clean-up costs improve with experience is what is called for.
 NDA Annual Report and Accounts 2005/06
 NDA Annual Report and Accounts 2014/15
 Nuclear Provision – explaining the cost of cleaning up Britain’s nuclear legacy: NDA Feb 2015
 Evaluation of nuclear decommissioning and waste management: Professor Gordon MacKerron
SPRU University of Sussex March 2012
 OECD – Costs of decommissioning nuclear power plants, NEA No. 7201, Professor G Rothwell March 2016
 NLF Annual Report and Accounts 2015
 TREN/05/NUCL/S07.55436 2609_EUDecommFunds_FinalReport.pdf 2007