1.On the basis of a report by the Comptroller and Auditor General, we took evidence from the Nuclear Decommissioning Authority (NDA), the Department of Business, Energy & Industrial Strategy (the Department), and UK Government Investments (UKGI) about the failure of the NDA’s Magnox contract. We also took evidence from Cavendish Fluor Partnership (CFP), the successful bidder and current contractor undertaking decommissioning work on the 12 Magnox sites until the contract comes to an end in September 2019.
2.With a value of up to £6.2 billion, the Magnox contract is among the largest put out to tender by Government. The NDA’s failure–in both procurement and managing the contract after it was let—is one of the most significant we have seen in public contracting, costing the taxpayer upwards of £122 million and disrupting vital decommissioning work at 12 nuclear and research sites. The latter comprise power stations that are at, or nearing, the end of their operational life.
3.After running a competitive procurement exercise for services to decommission the 12 sites, the NDA awarded the contract to Cavendish Fluor Partnership (CFP) in April 2014. In response to a legal case brought forward by one of the losing bidders, Energy Solutions, the High Court found that the NDA had actively ‘fudged’ the evaluation and awarded the contract to the wrong bidder. The NDA agreed to settle legal claims worth £97 million with Energy Solutions and its consortium partner, Bechtel.
4.The NDA also failed to manage the contract effectively after it was let, and announced in March 2017 that it had decided to terminate the contract 9 years early. This was because of a “significant mismatch” between the work tendered in the contract and the actual work that needed to be carried out on the sites. The NDA will now have to find a suitable alternative to managing the sites after the contract comes to an end in September 2019.
5.The Department, which had devolved oversight of the NDA to UKGI, approved key decisions made by the NDA throughout the entire process of procuring, evaluating and managing the Magnox contract. The government has launched an independent inquiry into the events surrounding the contract’s failure to establish the root causes and identify lessons. It is expected to report to the Secretary of State in early 2018.
6.In 2012, the NDA launched a competitive procurement exercise for services to decommission 12 nuclear sites. Four consortia submitted bids and each bid was then evaluated against over 700 criteria which included pass/fail criteria known as ‘threshold requirements.’ These requirements exclude bidders from the competition if they fail to include specific details (such as graphics) in their bid. The Tender Evaluation Report, including these evaluation critiera, was approved by the Department and HM Treasury.
7.We heard from CFP, the successful bidder, that the evaluation was very complex, and that the high number of evaluation criteria added a significant amount of complication to the bidding and evaluation process. The NDA told us that it now accepts that with hindsight, the competition was more complex than it needed to be. The NDA was not able to confirm however whether it had fully followed the advice provided to it by its legal advisors, Burges Salmon, during the evaluation process. In written evidence, the NDA’s former chief executive clarified that while he is unable to confirm with certainty the extent to which every piece of advice received from Burges Salmon through the procurement, the subsequent litigation and the consolidation process, was followed as he had not conducted an analysis of this. He added that the NDA’s decision to defend the claims brought against it by Energy Solutions was with the support of the legal advice the NDA had received at the time.
8.In July 2016, the High Court found that the NDA had wrongly awarded the contract to CFP. Had the NDA’s evaluators applied the evaluation criteria correctly, CFP would have been excluded from the competition altogether. The High Court also found that NDA evaluators had incorrectly applied the pass/fail criteria (threshold requirements) to avoid eliminating CFP from the competition. We pressed the NDA on whether it purposefully manipulated the evaluation process and why. The NDA told us that their evaluators believed they had the discretion to lawfully change these scores. It added that it was probable that evaluators had changed some of the scores after recognising that with 300 pass/fail criteria in place, a strict application of that would have resulted in all bidders being excluded from the competition.
9.These procurement failures occurred despite six internal and external assurance reviews, all of which rated the process as ‘green’ and ‘amber green. The Department told us that the NDA had deliberately designed the competition to limit access to the evaluation proceedings to a small number of individuals to help ensure that bidders remained anonymous. But this meant that scrutiny and oversight of the evaluation process may have been compromised.
10.In designing its commercial strategy for the Magnox contract, the NDA aimed to ‘do the same for less’, by giving potential contractors a stronger incentive to deliver savings for the taxpayer, through the use of a target cost incentive fee contract (TCIF). In this type of contract, the contractor is incentivised to bring down the overall costs of decommissioning because their fees go up if they are able to do so. Conversely, the contractor’s fee goes down if the costs of decommissioning increase. Under the previous contract for the Magnox sites, the contractor was reimbursed for all its allowed costs, and paid a fee regardless of how large the costs of decommissioning were. Crucially, for a TCIF contract to be successful, the NDA needed to have a very good understanding of the work that needed to be carried out on the 12 sites.
11.The NDA told us that when it started the competition in 2012, it believed it had a good grasp of the state of the sites at that time and of the work that needed to be undertaken as part of the contract; a good plan in place for delivering that work; and that performance on the sites was on track. However, when they let the contract in 2014, six of the 12 sites were behind schedule, and there was a significant gap between the amount of work the NDA expected the incoming contractor, CFP, to do and what was originally specified in the contract. The NDA concedes that their understanding of the state of the sites was, in retrospect, lacking.
12.The NDA said it knew that the state of the sites would differ from the time it began the competition in 2012 to the time it let the contract in 2014. For this reason, it designed the contract to include a process it calls “consolidation”—a truing up between what the contractor was told to expect and what it actually found on taking over responsibility for the sites. Contractually, this process had to conclude in 12 months, by September 2015. However, due to the significant differences between the NDA’s understanding of the sites in 2012 and the actual state of the sites in 2014, this consolidation exercise continued unresolved into 2017. During this time, the expected costs of decommissioning the Magnox sites increased from £3.8 billion in 2012 to £6.0 billion by March 2017.
13.Aside from the substantial increase in public funds required to decommission the sites, this increase in the contract cost was significant for two reasons. First, the significant cost increase and the changes to the required work were so significant that they materially varied from the contract the NDA awarded in 2014. This would have left the NDA vulnerable to legal challenge from parties who could claim that they would have bid, or bid differently, had they realised the larger scope and cost of the contract. This prompted the NDA to end the contract 9 years early. Second, the NDA is not yet able to fully account for £0.5 billion of the £2.2 billion cost increase. This implies either that the NDA’s information about the state of the sites was incorrect before it tendered the contract in 2012, or the previous contractor underperformed while the competition was underway between 2012 and 2014, or a combination of the two.
14.In March 2017, the NDA’s internal audit function reported that there is a possibility that the NDA may have paid its previous contractor for work that was not completed on the sites. The NDA confirmed that it has launched an investigation into whether it may have paid for work that was not undertaken.
1 Report by the Comptroller and Auditor General, , Session 2017–19, HC 408, 11 October 2017
2 , paras 1 and 12
3 para 1.10, 1.11 and 1.15
4 para 7
5 Qq 49–51, para 3.1 to 3.3
6 , para 4
7 para 1.7 and 1.11, Qq 5 and 62
8 Figure 12
9 Qq 5–6
10 Q 56
11 Qq 67–72
12 Correspondence with Nuclear Decommissioning Authority, dated 15 December 2017
13 Q62 and , para 1.10–1.12
14 Qq 62–66
15 Q162 and para 8
16 Qq162 and 168, para 3.9–3.10
17 para 1.5–1.6, and Q 53
18 Q 53
19 Qq 54 and 82, and para 2.8
20 Qq 78, 82
21 para 4 and 5, Qq 128, 152–153
22 Qq 33, 46, 83–84, 94, 127–129
23 para 4
24 Qq 130–137
27 February 2018