In spite of this, the reports of the National Audit Office (NAO) on PFI projects have generally shown them to provide value for money as compared with the Public Sector Comparator. Why is this?
This is the question analysed in chapters 5 and 6.
In just about every case that the NAO has looked at, the PFI alternative is superior once an allowance for 'optimism bias' is included. Generally in the reports which have compared the PFI costs with those of the PSCs, the PFI costs were found to be marginally lower and therefore were found to give value for money.
However these comparisons were usually faulty for two reasons. Firstly their estimates for overruns ('optimism bias') on publicly-funded projects were overstated and secondly the comparisons were not always done at the same stage.
Let's look at optimism bias first. If you look at the website of the NNUH, you will find under their 'publications', a section headed 'Private Finance Initiative - Norfolk and Norwich University Hospital'. That document admits that the government can borrow more cheaply than the private sector, but then argues that "any additional costs of borrowing are offset by the private sector … making more efficient use of resources". The document goes on to give the publicly-funded Scottish Parliament as an example of public sector overruns. It was, the website tells you, some three years behind schedule and ten times over budget when completed. It then says that NNUH was delivered 20 weeks ahead of schedule and on budget.
However this is not a valid comparison. Hospitals are more routine projects than one-off Scottish Parliaments. The NNUH website refers to a report in 2002 by Mott McDonald (consulting engineers) but does not give the details of that report. The Mott MacDonald report pointed out that for standard buildings (such as District General Hospitals), the 'optimism bias' was between 2 per cent and 24 per cent, the average being 13 per cent.
This is not enough to offset the PFI's cost-of-capital disadvantage and in any case it seems to be an overstatement of 'optimism bias' as shown in a study by Pollock, Price and Player in 2007. The latter was highly critical of the Mott MacDonald study, arguing that it was deficient in its sampling techniques and that it contained serious measurement bias (Pollock et al 2007, pages 6, 7). In addition, in 1999, the House of Commons Public Accounts Committee had given lower cost overruns stating that they were averaging about 7 per cent. As a 2001 report by Jon Sussex, a health economist, put it; "The myth of continuing rampant cost and time overruns on conventionally procured major hospital projects is just that; a myth" (Sussex 2001, 50)
As we have seen from the re-run of the Andersen model, the allowance of optimism bias has to be well above the 13 per cent average given in the Mott MacDonald report, and as we have seen the 13 per cent is itself very much on the high side.
The second criticism of the NAO's comparisons is that they have frequently assumed a different stage in the project cycle for the PFI as compared with the PSC. As the Mott MacDonald report admitted, the PSC is prepared usually at a different point in the project cycle from the PFI estimate so that like-for-like comparisons are not performed (Mott MacDonald 2002, 26).
This is a significant criticism since cost figures generally escalate over the life of the project. Thus the average cost increases for 43 'prioritised' PFI schemes between the Outline Business Case stage (when project plans are initially approved by the Department of Health) and the stage at which contracts were signed was 74 per cent. The increase for the NNUH was 28.5 per cent (Hellowell and Pollock 2007, table 6). Allowing for inflation, the increases in real terms would be less but almost certainly well above the 13 per cent average overrun given in the Mott MacDonald report for publicly-financed projects.
A reasonable conclusion is that the PSC and optimism bias figures have been manipulated so that the PFI cost comes out at just below that of the PSC. This was certainly the case for the NNUH in the Full Business Case of 1996. In costing its Public Service Comparator, an overrun of 34.22 per cent was assumed. Note, not 34%. Nor 35%. But 34.22% (see Appendix 1.5). The comparison in the Full Business Case of the NNUH was a shambles. As the chair of the Select Committee on Health put it; "In other words, the full business case does not tell us the full business Case" (UK Parliament, May 1999, paragraph 20).
The reason for this manipulation is that few hospital trusts have had much choice about whether to get their new hospital from private or public finance. They have been told that if they want approval for a new hospital, they will have to go down the PFI route. This was the prospect facing the NNUH Trust. In 1997 there had been a strong endorsement of PFI at the national level by the then Health Minister (Alan Milburn)3, and "deputations from Norwich were warned that if the Colney deal failed to go ahead, there would be no realistic hope of a hospital in the near future" (Greenaway et al 2007, 726). As a result, the figures had to be fiddled to show the superiority of the PFI alternative. And fiddled they were.
3 After Alan Milburn finished being Secretary of State for Health in 2003, he became a £30,000 pa adviser to Bridgepoint, a venture capital firm heavily involved in financing private health care firms moving into the NHS (see Pollock et al 2004, 6 )