NAO Report on PFI – Does it Justify the Headlines?

A National Audit Office (NAO) Report on PFI and PF2 published this month generated a raft of negative headlines, most of which questioned the value of these projects.

However, what does the NAO report actually say and does it justify the headlines?

Firstly, this is not the first time the NAO has reported on PFI. In fact, there have been a series of reports, some on individual projects, some on sectors (such as its 2010 report on PFI in Housing) and others of more general application, including:

  • Its 2003 and 2009 reports on the construction performance of PFI projects which concluded that generally PFI will deliver better price and time certainty than traditional public sector procurements.
  • Its October 2013 review of the VFM Assessment Process for PFI, which questioned how the Treasury carried out its cost-effectiveness analysis of PFI projects. Although the analysis was consistent with the Green Book, it assumed that the government had already set its expected levels of borrowing and spending limits for individual procuring authorities as part of the annual spending reviews and therefore did not compare the value for money of PFI projects against the cost of general government borrowing.

The purpose of the NAO report

The new report considers the rationale, costs and benefit of PFI, the use and impact of PFI, the ability to make savings from operational contracts and the introduction of PF2.  The report stresses that it is presenting information on the programme as a whole and is not seeking to express a view on the model or individual projects.  In particular, the NAO states that it has not formed a view on the value for money of PFI and PF2.

In terms of scale, there are currently more than 700 operational PFI and PF2 deals.  Capital investment using PFI and PF2 has averaged around £3 billion per year, in comparison to publicly financed government capital investment of around £50 billion per year, so   privately financed deals comprise less than 10% of total capital investment.

Findings beyond the headlines

Against that background, the report contains many interesting findings and in spite of the press coverage, not all of these are negative towards the use of PFI.  These include:

  • As widely reported in the press, the NAO noted that the expected “overall cash spending” for “one group of schools” would be 40% higher for PF2 than it would be for a project financed by government borrowing[1]. The report appears to indicate that the main reason[2] for this difference is the higher cost of finance[3] that private sector borrowers have to pay.  This comparison of PF2 cost with the general cost of Government borrowing is therefore picking up the theme of its earlier VFM Assessment report.  Its analysis assumes that Government borrowing could have been used for projects in question.  However, the report also notes that five of the six departments that answered the NAO’s survey reported that their capital budgets at the relevant time would not have been sufficient to cover the new investment that was funded by PFI.  It was the only game in town.
  • Cost certainty was still seen as a benefit of PFI by the majority of departments that responded to the NAO’s survey. It is worth noting that in the NAO’s 2003 report, 73% of traditionally procured projects had cost overruns compared to 22% in PFIs and their 2009 report reported 46% of traditionally procured projects had cost overruns, compared with 35% of PFIs.
  • Although the Treasury Committee’s 2011 review found that some PFI projects charge higher prices for construction to cover unforeseen costs, the NAO report notes that recent research by the Department for Education has, so far, found that the financing route (private/public) has little or no effect on the construction costs of schools.
  • Previous studies by the NAO have found no evidence of operational efficiency in PFI: the costs of services were similar to traditional contracts. More recent studies however, have suggested that the costs of services may be higher in PFI contracts, but this may be because of the greater risk transfer involved in PFI and the differing standards between PFI contracts (which tend to have very demanding standards) and traditional contracts.
  • Most of the departments that responded to the NAO’s survey considered maintenance standards were higher under PFI than its non-PFI assets. This may be explained by current pressures on public sector budgets leading to significant reductions in maintenance spending on non-PFI assets.  In 2015-16 health trusts reported a critical infrastructure maintenance backlog on their non-PFI assets of more than £2.3 billion.
  • Operational inflexibility remains a major drawback of PFI according to five out of the six departments that took part in the NAO’s survey. The PFI structure means that changes can be more expensive with lenders and investors charging administration and management fees.  The report notes that the latest PF2 guidance does contain a variation mechanism aimed at reducing the cost and administration of changes.
  • The NAO repeats its previous criticism of the way public sector comparators have been prepared to ascertain whether the private finance option offers better value for money. In particular, it questions the discount rate used to determine the net present value of the PFI option. Presently the discount rate used is the Social Time Preference Rate which is considerably higher than the “actual cost” of government borrowing (yield on a 20-year gilt).  It also criticises the use of adjustments in the models, such as for “optimum bias” and “risk transfer”. It notes that whilst the Treasury formally withdrew its previous guidance for value for money assessments in 2012, it has yet to publish a replacement model.
  • Most contracts have benchmarking/market testing clauses for services such as cleaning to ensure that the price being paid by the public sector is fair. These mechanisms work both ways so can result in the public sector paying more. The NAO’s analysis has shown that on PFI hospital and schools projects, such value testing tended to result in above inflation rises in costs.  The NAO notes that often the public sector does not have its own data to challenge benchmarking figures provided by the private sector.
  • Whilst the launch of PF2 introduced some changes, the fundamental characteristics of PFI remained unchanged in the PF2 model.
  • The NAO notes that to ensure that PF2 projects remain off-balance sheet under the new Eurostat rules, the Treasury will be making some changes to the PF2 structure, even though these changes may reduce value for money. For example, existing guidance provides that the public sector will receive 70% of refinancing gains over £3 million.  However, because of the Eurostat rules, the refinancing provisions will now limit the amount that can be received under the refinancing gain-share mechanism to 33%.

By partner Robert Turner, who specialises in advising on major construction and infrastructure projects

For further information, please contact Robert Turner, Paul Buckland, Kris Kelliher, Caroline Mostowfi, or Jamie Leonard from our Infrastructure and Projects Group.

[1] For these purposes the report assumes government borrowing using a 2.5% amortising loan with an average life of the project debt of less than 20 years
[2] The report does identify certain other additional costs that would apply in PF2 but not traditional procurement such as insurance costs, the holding of cash reserves, external adviser costs, lender arrangement fees and SPV administration fees.
[3] The report does not actually identify the interest rate that was used for the PF2 comparator, but elsewhere  it notes that the indicative cost of capital for PFI projects estimated in the National Infrastructure Plan is 2 – 3.75% above the cost of Government gilts.



Share this post

Related categories



The latest news from Devonshires, sent to you direct.

Join our mailing list and find out what we’re up to and what we think about recent events and future possibilities.

Join our Mailing List