The Committee has recommended that sound decisions on whether to proceed with PFI deals should be based on a systematic examination of comparisons between each proposed deal and realistic alternative options [42nd Report, Session 1997-98 (HC 348): Skye Bridge]. In the case of the National Savings deal with Siemens Business Services the C&AG's report [HC 493, Session 1999-00] states that the Department constructed a public sector comparator against which to assess the proposed PFI deal. The Department's analysis showed that the PFI deal was expected to lead to a net present cost of £635 million, £158 million lower than the Public Sector Comparator net present cost of £793 million.

  In evidence to the Committee, the Managing Director of Siemens Business Services said that one of the reasons why his company's bid was so much cheaper than National Savings' Public Sector Comparator was that it would not be subject to the constraints on capital spending to which National Savings would be subject. The question therefore arose as to whether it would be legitimate in comparing the PFI deal with the Public Sector Comparator to include the impact on the Public Sector Comparator of presumed constraints on the public funds available to the department.

  There are currently two prime sources of guidance from the Treasury on the use and construction of comparators in the appraisal of Public Finance Initiative projects:

—  Treasury Task Force Policy Statement No 2: Public Sector Comparators and Value for Money.

—  Treasury Task Force Technical Note No 5: How to construct a Public Sector Comparator.

  These documents refer to a further source of guidance, "The Green Book" (Appraisal and Evaluation in Central Government, HM Treasury 1997).

  The question of capital constraints is addressed in Policy Statement No 2, which distinguishes two cases:

(a)  where there is a genuine choice between procuring a service following PFI principles and procuring the same service using conventional funding;

(b)  where the realistic alternative to PFI procurement will be the pursuit of the same policy objective with no or much lower investment in new assets.

  The Policy Statement recommends that the Public Sector Comparator be based on the full cost of providing the services specified to bidders including the full cost of the assets that the public sector would have constructed. In the latter case, the Statement says that there is a choice to be made between preparing a comparator on the basis of what is affordable, within the existing public expenditure allocation, or preparing a notional calculation based on the assumption that additional public funds are available to construct new assets. The Statement advises procurers to consider two points in making this choice:

—  the likelihood of conventional funding being made available and whether the creation of a comparator which is unaffordable would be misleading; and

—  the difficulties of making an objective comparison between a bid and a constrained comparator which is necessarily based on different levels of service.

  Where the arguments are very finely balanced the Statement says that it may be appropriate to examine both the constrained and unconstrained options.

  The issue is also examined in Technical Note 5, which states:

"Sometimes Public Sector Comparators are constructed on the assumption that major construction work will be delayed due to constraints on the availability of public capital. This approach is not recommended as any assumptions made re inherently non-verifiable and recent history, particularly in the health sector, has shown that levels of available public capital can be quite volatile even over relatively short periods. If there is any doubt regarding the availability of public capital, sensitivity analysis should be undertaken to quantify the effect of delayed construction work. However, assumptions about the start, completion, and if applicable, the phasing of construction work should reflect what could be realistic to expect in the public sector and will not necessarily correspond to bidders' proposals."

  The Green Book contains guidance on investment appraisal in general. It addresses the issue of expenditure constraints in the following terms:

"Expenditure may be constrained in particular years. An expenditure constraint does not reduce the importance of appraisal, but rather enhances the importance and technical difficulty of choosing those uses of limited funds, which will provide the greatest benefits. It may influence the choice of options, if options, which would otherwise be preferred, require higher expenditure in the period during which the constraints are expected to be most severe.

However, if the constraint is expected to be no less severe in the longer term than the shorter term then this should not generally lead to capital rationing of investments which reduce public costs or increase public revenues. Making room for such investments within the overall expenditure ceiling may cause great difficulty; but the savings they generate will bring equally great relief."

  National Savings followed all of this guidance. Because National Savings senior management considered that there would be doubt about the availability of public sector capital, in accordance with the guidance, they examined the sensitivity of the comparison between the PFI bid and the public Sector Comparator to changes in the assumption about availability of public sector capital. The outcome of this sensitivity analysis was that, in the absence of any constraints on the availability of public sector capital, the Public Sector Comparator would have had a net present cost of £748 million. This figure was still higher than the £635 million cost of the PFI deal, but by a smaller margin than the £793 million cost of the Public Sector Comparator without access to additional capital funding.

National Audit Office

4 October 2000