So far, we have discussed the reasons why the Labour Government has promoted the PFI. We have looked at the Andersen model and from that it is clear that hospitals financed under the PFI are likely to have a before-risk capital cost which is at least 60% greater than a publicly-financed alternative. And even if an allowance of about 13% is made for cost overruns ('optimism bias') on publicly-financed hospitals, PFI contracts will end up being at least 40% more expensive.
Why, then, do most of the value-for-money studies carried out not reach this conclusion? As we have seen, the answer is; public finance has not generally been available for large NHS investments and as a result, there has been considerable pressure to show that PFI projects are superior. This is what has been done - either by exaggerating the PSC in the first place and/or adding on an excessively large allowance for 'optimism bias'.
Is this the end of the story? Not for the PFI advocates. They argue that there is another advantage of PFI contracts, namely that such contracts allow for the integration of maintenance with construction. Jon Sussex suggests a number of reasons why PFI hospitals may be maintained better, most notably that the maintenance funds are ring-fenced by the PFI contract or that the design of the hospital is undertaken with an eye to facilitating maintenance. However he also points out that these advantages could be "captured by competitively tendering hospital design, building and maintenance with long-term contracts that include penalties for poor performance" (Sussex, 2003, 68). As he says "These would be DBO (design, build and operate) contracts" but they do not need to be DBOF contracts - that is they would not need the F word - here short for finance.
In any case there can be problems associated with the private finance company (the Special Purpose Vehicle or SPV) being responsible for maintenance as well as construction. As Edwards P et al 2004 put it;
"Thus the SPV finds itself carrying risks associated with the use of the building but does not have direct control over the users of the building, including the employees of other syndicate members, trust staff, patients and the general public. One outcome of this is that when repairs are required, especially those associated with damage to the fabric of the building, there may be negotiations about where responsibility lies" (page 196)
Not only are there confusions about responsibilities but ironing out these confusions turn out to be very expensive. For example even when the PFI company (the Special Purpose Vehicle or SPV) is not itself carrying out a change to the project, it commonly charges a fee for simply acting as a conduit. As a recent Public Accounts Committee report pointed out, these fees have ranged from 2% to 25%. In 2006, such fees added £6 mn. to the cost of changes made (PAC, September 2008, 3)
And so the claimed advantage of the integration of maintenance with construction turns out to be a probable disadvantage. In addition there is another disadvantage of PFI contracts, namely the high transactions costs associated with negotiating PFI contracts.
On the high transactions costs associated with PFI contracts, Jon Sussex writes;
"The PFI selection/bidding and contract negotiation process takes months, or even years longer than for Exchequer financed schemes" and "this offsets any advantages of quicker completion for PFI contracts once they get under way" (Sussex 2001, 46).
Sussex also points out that;
"the NHS Confederation argues that the PFI is unpopular with health service managers because of the larger transactions costs that are entailed compared with conventional procurement" (Sussex 2003, 64).
Sussex points out that; "an average of nearly £3 million each was spent on external legal, financial and other professional advice during the procurement process by the NHS Trusts involved in the first 18 major PFI hospital schemes to be signed-off" (Sussex 2001, 47). This average of £3 million is the money spent by the Trusts. Shaw reports that details obtained through Parliamentary Questions revealed that the advisors' costs for the first 15 NHS PFI hospitals represented between 2.4% and 8.7% of the capital cost of those projects, Shaw also argues that in the early stages of PFI there was a substantial increase in transactions costs over the level of pre-PFI schemes (Shaw 2004, 70, 71)22
This is supported by the Department of Health which has also pointed out that transactions costs were particularly high for the early PFI schemes in the NHS (see Department of Health, February 2004, question 11). An example of one of the earliest PFI schemes is the Norfolk and Norwich University Hospital. It is also an example of a scheme which paid very heavily for being one of the earliest. Thus the 'PFI set-up costs' incurred by the NNHCT/NNUHT totalled just under £13 million23 between 1995/96 and 2001/02, but in addition to the costs incurred by the Trust directly, there were those incurred by the private finance company and recovered in the contract. At the Norfolk and Norwich University Hospital, the 'development costs and fees' incurred by Octagon Healthcare (the SPV) and included in the capital cost totalled £36.3 mn. It seems that some of this £36 mn was for items which would have applied equally to a publicly-financed scheme. Thus £5 mn was for highway costs, £8 mn for IT hardware, and £4 mn for catering and other service equipment (UK Parliament July 1999, para 226). This is a total of £17 mn for items which were not 'transactions costs' but we are still left with something like £19 mn of transactions costs including the consortium's tender cost of £6.6 mn and the financing cost of £5.7 mn (see UK Parliament July 1999, para 227). Thus if we add the 'PFI set-up costs' of the Trust (£13 mn) to the £19 mn of the SPV (Octagon Healthcare), we have a total of something like £32 mn. This is more than 20 per cent of the £159 mn. building construction cost (as given on page 18 of NAO, June 2005).
There are two ways in which these transactions costs can be reduced.
One is to reduce the competition. The Treasury Select Committee has acknowledged that "there is a trade-off between competition and the length and cost of [PFI] negotiations" (quoted in Shaw 2004, 70) but the Public Accounts Committee has bemoaned the fact that competitive tension is not maintained if there is only one bidder for a contract (see Flinders 2005, 224). It is likely of course that, with less competition, the successful tender will be more expensive than otherwise.
Clearly then, there is a contradiction between the urge to increase competition and the desire to cut transactions costs. A BBC 'File on Four' radio programme broadcast on June 12 2007 reported that it was increasingly difficult to get a number of bids from companies for PFI projects. The problem is so acute that in one case (Southmead hospital in Bristol), the Hospital Trust is considering reimbursing some of the millions of pounds in bidding costs for the companies losing out in the bidding process (page 8 of the transcript of the programme accessed in December 2007 on bbc.co.uk/fileon4). Thus to reduce transactions costs, it may be necessary to reduce competition in bidding.
The second way of reducing the transactions costs is to standardise the procedure. This has been done. In 1999 a PFI guidance manual was introduced together with a standard contract for major PFI schemes (Department of Health December 15, 1999, 1). However there is a danger of a loss of flexibility in such a process of standardisation. At a PPP forum in 2003, Malcom Stamp, the ex-chief executive of the Norfolk and Norwich University hospital, stated that; "I do not believe standardising contracts is the way forward…….More weight has to be given to innovation, change and flexibility" (quoted in Edwards P et al 2004, 203). And at a seminar in 2002 on Building Long-term Partnerships, it was stated that; "the general perception expressed was of remarkable similarity between different projects' design outcomes in practice, rather than of notable innovation" (Sussex 2002, 4). Furthermore at the same seminar, there was reported to be a consensus that the PFI approach to procurement was perpetuating the dominance of the District General Hospital (DGH) model in acute secondary care provision and that the PFI process has not encouraged major innovations such as considering replacing the 1962 DGH model with a radically different concept built around care pathways (Sussex 2002, 3). The NHS Confederation has also been reported as saying that the PFI scheme is too inflexible for the needs of the NHS (BBC News, 2004). When Patricia Hewitt was the Secretary of State for Health, she was reported as saying that she wants to create a counterweight to the power of hospitals (Guardian 29 June 2005). It is ironic that this statement has come soon after very large investments have been made through the PFI in new District General Hospitals.
So, on balance, in addition to the higher construction/capital cost associated with PFI contracts, there seem if anything to be further disadvantages. There seem to have been very high transactions costs associated with the PFI schemes and it is hard to believe that these would have been as high for publicly-financed schemes. That is, the costs associated with the financing and legal aspects of the PFI scheme seem to add at least 10% to the construction cost - over and above the design costs which would of course have to be paid under a publicly-financed scheme. Thus it seems likely that the transactions costs associated with PFI schemes are high enough (at more than 10% of the 'core construction costs') to cancel out most of (if not all) the higher costs associated with possible 'optimism bias' associated with publicly-financed schemes. If so, we are back to an excess capital+finance cost of PFI schemes of more than 60% even after allowing for optimism bias.
Of course the counter of pro-PFI lobbyists to the claim that capital costs are so much higher might be to say; "Yes that's all very well, but the increase in capital costs may well have led to lower operating costs". However it is hard to see why non-clinical recurrent costs should be any different when associated with a DBOF contract as opposed to a DBO contract.
So far, we have looked (through the Andersen model) at the likelihood of the PFI capital+finance costs being larger. We have also looked at the with-and-without comparisons of PFI and PSC costs.
Clearly another way to assess the value for money from the PFI is a before-and-after comparison which looks at the changes in costs over time. As David Price put it in a File on 4 radio programme; "To find out the truth about value for money, you have to go and look at schemes that are actually up and running…." (BBC Tuesday, July 6, 2004, 5).
This is what I do in the next two chapters but there are two obvious problems with looking at schemes which are up and running. Firstly most of the PFI projects have been implemented since 1997 and have not been running for very long so that there is not much history. Not surprisingly, as Edwards P et al say; "there has been little empirical financial research into the cost and effectiveness of private finance in hospitals after implementation" (2004, page 173). The second problem with a before-and-after comparison is that conditions may have changed so much that the before environment is very different from the after environment.
As far as the first is concerned, I look here at the Norfolk and Norwich University Hospital (the NNUH) which is one of the oldest and largest of the PFI schemes in the health sector and I attempt to deal with the second problem by making allowances for changing conditions as you will see in chapter 9. But first, in chapter 8, I give a brief introduction to the NNUH.
22 . Audit Scotland/Accounts Commission found that the transactions costs associated with six PFI contracts for schools averaged about 10% of 'core construction costs' (2002, page 10)
23 . This consist of the sum of £1.2mn in 1995/96 and £1.9 mn in 1996/97 (Annual Accounts 1996/97, page 11); £2.6 mn in 1997/98 (Annual Accounts 1997/98, page 14); £1.5 mn in 1998/99 (Annual Accounts 1998/99, page 17); £2.4 mn in 1999/2000 (Annual Accounts 1999/2000, page 16); £2.2 mn in 2000/2001 (Annual Accounts 2000/01, page 17); and £0.9 mn in 2001/02 (Annual Accounts 2001/02, page 18).