House of Commons - Private Finance Initiative

Private Finance Initiative - Treasury Contents

3 Value for money

Cost and availability of finance

26. Private finance is invariably more expensive than direct government borrowing and therefore we explored the difference in the availability and cost of private and government debt in our evidence. Balfour Beatty told us in written evidence that the "financing costs of PFI are typically 3-4% over that of government debt."[33] All witnesses agreed that the differential between government debt and private project finance was significant and that it had increased since the financial crisis. James Wardlaw of Goldman Sachs told us "In terms of the cost of that finance, it is definitely higher [...] the levels were 60 [basis points][34] over swaps at the peak, and now we are talking 250 and more."[35] When we asked him whether or not the upward shift in the cost of capital was locked in for the foreseeable future Mr Wardlaw agreed, telling us "Yes, and also that the willingness of people to finance remains quite short".[36] Richard Abadie explained that the difference between government debt and private debt "will never come down [...] to the levels pre-credit crunch" adding "I do think we are in a world of more expensive debt".[37] Steve Allen of Transport to London also told us "there is a significant premium for the cost of finance through a PFI".[38]

27. A National Audit Office report of 2010 which examined financing PFI in the credit crisis "found that the part of the cost relating to loan margins on PFI deals, which had been 1 per cent or less, widened significantly to around 2.5 per cent on average" and that some "will rise to more than 3 per cent in stages over the project life". The NAO explained that this "resulted in substantial increases to the cost of finance".[39] As well as loan margins increasing the NAO report also showed that since the credit crisis arrangement fees, commitment fees and the 'swap credit spread' had all increased.[40] These extra fees, which were projected to be more than 3% of the total value of the debt drawn down, further increase the difference between the cost of government debt and private finance. The NAO report also noted that "PFI is less likely to be value for money unless there are substantial and credible savings to offset higher financing costs."[41]

28. Some of the reasons for the increased costs and lack of availability of finance were explained in the International Handbook on Public-Private Partnerships published in 2010:

The global financial crisis has led to a significant increase in the cost of private finance—in particular the senior debt component. Commercial bond finance—hitherto the cheapest form of senior debt for PFI projects—has been unavailable since mid-2008,when many of the big US 'monoline' insurers such as Ambac and MBIA lost their 'triple-A' credit rating in the midst of the 'subprime' mortgage crisis. These institutions had played a key role in the provision of senior debt for large projects, by guaranteeing ('wrapping') repayments to bondholders in return for a fee and thereby reducing overall financing costs. The withdrawal of the monolines' ability to provide a triple-A guarantee has removed commercial bond financing as a low-cost option for the foreseeable future.

At the same time, banking sector liquidity has reduced dramatically as the financial crisis has developed.

The paper also noted that "the size of the increase in margins [...] contains a substantial premium that is unrelated to default risk, and is associated with credit constraints and the oligopolistic nature of the senior debt market."[42]

As well as the increased cost of debt finance it is important not to forget that a PFI is also partly funded by equity. This means that the cost of capital (which includes a return for equity holders) is higher than just the cost of the private debt. The Weighted Average Cost of Capital for a conventional availability-based PFI project in the accommodation sector[43] is now in excess of 8.5%.[44] This compares to the current long term government gilt rate of just over 4%.[45]

Box 1: Private finance comparison with public finance - worked example

Analysis by Mark Hellowell - Specialist Adviser to the Committee
For a PFI to cost less than a conventional procurement, it must deliver savings in construction, maintenance and/or service provision that are, relative to the risk-adjusted costs of a conventionally procured alternative, sufficient to offset the higher financial cost.[46] Therefore, it is important to consider the scale of the difference in financial costs between public and private finance. Here, we examine this by looking at the cost projections relating to the Royal Liverpool and Broadgreen University Hospital NHS Trust's PFI project, which is currently in the procurement phase. These projections are contained in spreadsheets associated with the Trust's Outline Business Case, which was approved by successive governments in 2009 and 2010 respectively.
In the version of the spreadsheet used in this analysis, the project is assumed to involve initial capital expenditure of £244 million and the contract is expected to run for 34 years, including a four-year construction period and a 30-year management phase in which the private partner will deliver maintenance services. During the management phase, the Trust will pay to the private partner a periodic unitary charge. This provides the private partner with a revenue stream from which to meet operational costs (primarily maintenance and lifecycle costs, along with the costs of running an office and paying insurance), and financial costs (primarily the costs of making principal and interest payments to "senior" and "junior" debt providers and a return to the owners of equity). The cash-flow to all these investors is called the Project Cash-Flow, and this is the data source for this analysis.
This cash-flow takes the form of a series of expenditure cash-flows (relating to the four year construction period) followed by a series of revenue cash-flows (in which income from the public sector significantly exceeds the private sector's operational costs, thereby providing revenue for distribution to investors). The additional financial cost of PFI can be derived by discounting the stream of cash-flows at the relevant discount rate—which is here taken to be the "gross redemption yield" on government "gilts" of the approximately the same maturity as the PFI loans (i.e. 30 year gilts). The current yield is approximately 4.2%.
Discounting the Project Cash-Flow stream at 4.2% produces an NPV of £175 million. This figure represents the additional financial cost of using private, rather than public finance, to deliver that amount of capital expenditure. If we assume that the outturn costs of construction, maintenance and services will be the same between the PFI and conventional procurement options, the government could have spent £175 million less, in NPV terms, by borrowing directly from the capital markets, rather than through an SPV intermediary.
A different way to examine this is to discount the expenditure cash-flow and the revenue cash-flow separately at 4.2%, and then compare the present values of each. On this basis, the present value of the revenue cash-flow is £421 million and the present value of the expenditure cash-flow is £246 million, a ratio of 1.7/1. Had the financing been provided at the gilt rate, rather than at the private finance rate, the ratio would be 1/1.
There are two different ways of interpreting the results of this analysis. The first, as noted, is that the public sector is paying £175 million more than it needs to in order to secure the amount of capital expenditure required.[47] This is the NPV of the higher cost of private finance—the cost that the PFI model needs to offset, in terms of efficiencies in construction, maintenance and/or services compared with conventional procurement, to represent a cost-efficient solution. An alternative way to view this is in terms of foregone opportunities for additional capital investment. Assuming that PFI does not deliver efficiencies in construction, maintenance and/or services then, for the same present value of finance-related payments, the government could have secured 71% more investment by borrowing on its own account.
Source data: Royal Liverpool and Broadgreen University Hospital Outline Business Case, 2010

Table 2: Summary of worked example - Financial cost of capital expenditure

Cost - Present Value
(@ 4.2% discount rate)

Private finance £421 million

Government finance £246 million

Savings and benefits PFI needs to deliver in other areas to offset the extra cost of private finance £175 million

Potential increase in investment possible if using government financing, assuming no offsetting efficiencies from PFI (%) 71% increase

Potential saving from using government financing, assuming no offsetting efficiencies from PFI (%) 42% saving

Source: Committee Specialist Adviser analysis of RLBUH Business Case - see Box 1 for details

29. To understand better the cost difference between private finance and public finance over the life of a project we asked our Specialist Adviser to perform an analysis (Box 1 and Table 2). As the differential between the cost of government gilts and private sector debt has increased notably since the financial crisis we considered it was important to look at a contemporary example. The analysis undertaken used figures from a 2010 Outline Business Case for a new hospital and showed that there was significant extra cost of using private finance rather than public finance. The higher cost of capital for the PFI option compared to government gilts meant that, without any offsetting efficiencies, the cost of the PFI option would be 70% higher over the life of the project. One other way of looking at the difference in cost is to consider how long it takes government to pay off outstanding debt. If government borrowed directly and followed the same repayment schedule as the PFI charges the government debt could be fully repaid many years before the equivalent PFI liability could be paid off.

30. Government has always been able to obtain cheaper funding than private providers of project finance, but the difference between direct government funding and the cost of this finance has increased significantly since the financial crisis. The substantial increase in private finance costs means that the PFI financing method is now extremely inefficient. Recent data suggests that the Weighted Average Cost of Capital of a PFI is double that of government gilts. PFI will only provide value for money if this differential in the cost of finance, which has significantly increased, is outweighed by savings and efficiencies during the life of a PFI project.

31. Analysis undertaken by the Committee's Specialist Adviser suggest that, all else being equal, paying off a PFI debt of £1bn may cost the same as paying off government debt of £1.7bn. This would mean that a 70 percent increase in investment could be achieved for the same long term cost if government funding were used instead of private finance. An alternative way of expressing this is that the cost of paying off a PFI debt would be over 40 percent cheaper if government funding were used. The current higher cost of finance means there may be a significant opportunity cost from using PFI.

32. As part of our inquiry we considered if PFI had resulted in a better risk allocation and whether or not this allocation had resulted in savings and other benefits for the public sector which could offset the higher costs of financing. We consider these points in the following sections.

Risk allocation

33. We asked the witnesses if they believed that PFI had resulted in risk being transferred to the private sector efficiently. Professor Helm told us: "In terms of inefficiency, it is quite hard to think of many other aspects of the British economy that are more inefficient than that risk allocation."[48] Andy Friend explained that in the past "PFI theology said you would transfer any risk you could identify". He felt that this had led to inappropriate risks being transferred such as energy risk."How a private sector provider of a capital asset is in a better position to manage energy tariff risk than a public authority with its potential buying power, I don't know."[49] He also highlighted some other specific risks such as insurance and the management of derivatives that he considered could have been better managed by the public sector.

34. Witnesses did however point out that some risks such as construction risk had been transferred successfully. Richard Abadie told us:

One of the clear benefits of contracting out to the private sector is the transfer of construction risk. Let them build it, let them give you a fixed price for it [...][50]

Andy Friend highlighted a number of examples where construction risk had been transferred, including one which he had direct experience of:

I was Chief Executive of John Laing Plc when a project that had been entered into in 1998 went badly wrong, the National Physical Laboratory. We booked £68 million of losses on that.[51]

Mr Friend considered that other procurement methods "have involved much greater additional cost in terms of getting those projects operational" although he did concede that "there has been improvement in many of the mechanisms".[52] Mr Friend suggested that once the construction stage had been completed and the operational stage had started there was a case for allowing the public sector to refinance the debt.[53] This would allow the private sector to bear the risk during the construction phase but transfer the risk back to the public sector in the operational phase. The benefits of private finance projects transferring construction risk were made in other submissions to the Committee:

Construction risk is transferred in PPP [Public Private Partnership] projects from the public sector to the private sector. Fixed price, date certain contracts are the norm with no facility to make new claims on the public sector purse if unforeseen difficulties arise.[54]

Other evidence however noted that PFI and PPP were not the only ways of ensuring that construction risk was transferred:

The Treasury has acknowledged that on-time and on budget performance can be secured through conventional procurement, so long as the design and build services are procured through a fixed-price, "turn-key"[55] contract.[56]

35. In its written submission Balfour Beatty discussed risk transfer, telling us that "clients, often encouraged by their external advisory teams, are tempted to incrementally increase the risk transferred to the private sector". However it considered that often this was inappropriate, explaining that "these increases in risk transfer are not properly evaluated in terms of the potential impact on value for money". In particular, it pointed out that financial penalties that were used to transfer risk led to higher prices and a deterioration in value for money.

The payment mechanism is the authorities' main commercial tool to incentivise performance against the expected standard. However, our experience is that over time, increasingly aggressive payment mechanism arrangements result in poor value-for-money as PFI operators build-in risk to avoid the consequences of disproportionate penalties.[57]

The infrastructure company also highlighted four areas where they considered it better value for money for the public sector to bear the risk. These were: insurance; energy; pensions; and demand risk. In terms of energy it believed:

The public sector should resist the temptation to attempt to transfer risk on tariff which the private sector cannot manage any better than the public sector. Procurement of energy must be more effectively managed by the public sector, which can achieve significant economies of scale compared to the private sector.

It also provided detail about the limited circumstances where it believed demand risk should be transferred to the private sector:

Except where the private sector is genuinely responsible for generating customers/users, the transfer of demand risk (eg traffic counters on highways projects) should be avoided. Demand risk tends to increase the cost of lending and result in a sub-optimal project structure which leads to a reduction in value for money for the public sector.[58]

36. Professor David Heald explained in his submission that "it should not be an objective of PFI to transfer risk to the private sector but only to transfer those risks which the private sector is better equipped to handle". He also told us:

Attempting to transfer inappropriate types of risk will instead lead to excess costs and to potential default, with the materialising costs falling on the public sector. This echoes an important lesson from outsourcing in the petroleum industry: if the responsibility—legal and reputational—remains with the 'principal', the loss of operational knowledge and control may offset the apparent cost savings. Especially in an institutionally fragmented public sector, it is difficult to be an intelligent client and to sustain that through a 30-year PFI.[59]

37. Transport for London had some insights regarding risk transfer. It explained that "risk can be fully transferred only if the procuring authority could abandon a failing PFI concession, which is unlikely ever to be the case", adding that "TfL's experience is that the general public have little appetite for a blame game—clearly to the extent TfL can control its own assets, it can control its performance." It added:

TfL's view is that the private sector is willing to bear significant risk but only if it is paid enough. The question should be which party is best placed to manage each risk [...] where the private sector can manage risk better than the public sector, it should do so. However, this decision does not necessarily lead to using PFI—turnkey construction or maintenance contracts can be effective in risk transfer.[60]

38. Allocating risk to the private sector is only worthwhile if it is better able to manage the risk and can pass on any subsequent savings to the client. The main benefit highlighted to us by PFI providers was the transfer of construction risk. However a PFI contract which lasts for 30 years is not necessary to transfer this risk. There are also other methods such as turnkey contracts which can be used for the same ends. We have seen evidence that PFI has not provided good value from risk transfer—in some cases inappropriate risks have been given to the private sector to manage. This has resulted in higher prices and has been inefficient.

39. Some of the claimed risk transfer may also be illusory—the government is ultimately accountable for the delivery of public services. Therefore it would not be able to allow a number of services provided under a PFI contract to cease for any length of time.

Whole life cost and innovation

40. In PFI, the SPV is responsible for both the construction and operation of the asset, and the cost of both (along with the cost of finance) is included in a single price provided to the public authority. Supporters of PFI say that bundling in this way encourages up-front investments that will contribute to cost reduction over the asset's life cycle—i.e. spending more on construction might make sense if this will result in lower maintenance spending in the long term. The Treasury points out that this aspect of PFI distinguishes it from other forms of procurement:

Unlike other forms of procurement, PFI projects benefit from whole-life costing over 30 years, involving both construction and service delivery [...][61]

However the NAO noted in a paper published in 2009 that there are other methods that can be used to ensure that whole-life costs are considered:

Private finance is not, however, the only way to ring-fence maintenance funding or consider whole-life costs. The London Borough of Lewisham, for example, has established a sinking fund to ensure its non-PFI schools are maintained to the same standard as its PFI schools.[62]

The Treasury believes that, owing to the benefit of whole life costing, operating costs of PFI projects cannot be bettered by the services tendered as part of a non-PFI procurement or provided in-house. If they do cost less this will be done by "compromising the quality of service" with "sub-optimal investment".[63] They advise public bodies to adjust the PFI cost according to sector experience. This has resulted in the Department of Health recommending an assumption that annual 'life cycle costs' will be 15% cheaper for PFI deals for Trusts considering the different procurement options for a new build hospital.[64]

41. Many of the PFI contractors, investors and advisers that submitted evidence to the committee highlighted the consideration of 'whole life cost' as a major benefit of PFI. PricewaterhouseCoopers told us PFI resulted in "Focusing procurers on the whole-life cost and performance of infrastructure rather than making short term decisions based on short term budgets".[65] Canmore Partnership Ltd explained that "one of the main benefits of the PPP-type provision of public use infrastructure has been the whole-life integration of design, building, maintenance and life cycle costs." It went on to explain:

This correctly incentivises developers to invest in quality facilities at the outset, thus also increasing the availability of those facilities. At the end of a typical PPP concession the public sector will inherit assets which have been properly maintained.[66]

Barclays Infrastructure also pointed out that "PFI procurement encourages whole life costing, whereas traditional procurement focuses mainly on the initial construction costs."[67]

42. If bidders know they can achieve lower whole-life costs and the procurement process is sufficiently competitive, then this should result in lower prices for the public sector. If the benefits of whole life costing are working it would be reasonable to expect that building design would make use of innovations in order to provide higher quality buildings that will last longer in good condition. A previous Treasury Committee's report on PFI in 2000 made this very point and recommended that PFI projects should be monitored for "innovative approaches" that could be "transferred effectively to publicly-funded projects"[68]. Professor James Barlow has done research on innovative design in the health sector and we explored the issue of innovation with him. He was clear that PFI had hindered rather than encouraged design innovations: "I think the way risk was devolved and transferred has meant that it has made very difficult to stimulate any kind of innovative thinking about the design of the buildings".[69] We asked him how this compared to hospital building programmes of the past and he told us: "there was more design innovation in the 1960s and 1970s."[70] However he had not done research on the quality of the buildings and considered that PFI "should drive up quality".[71]

43. Although PFI theory states that the process should drive up building quality to keep long term costs down we received evidence which directly contradicted this. The Royal Institute of Architects told us that "the quality of the buildings delivered through PFI schemes remained poor in many cases". It explained that: "The poor quality of the buildings' design lead to a number of issues, such as rising maintenance costs over the lifetime of the building". One of the reasons it pointed to was "value-engineering by contractors", telling us that there was strong anecdotal evidence that contractors were withholding information from clients. This resulted in "essentially reducing the intended quality and cost of the project compared to that specified by the architect, to the detriment of the finished building, without the knowledge of an unaware client." The reason this was done was to "maintain the contractor's preferred levels of profitability".[72]

44. Where possible it is useful to compare PFI buildings to non-PFI buildings to see if benefits are being realised. The Audit Commission did a report on PFI schools in 2003. Although it found no difference between the construction costs of PFI and non-PFI schools[73] it did find that the quality of PFI schools was significantly worse than that of the traditionally funded schools. The average score given by the Building Research Establishment (BRE) for the PFI schools was lower than the non-PFI schools in all of the areas tested such as architectural design, user productivity and ownership costs. The report also noted: "The best examples of the type of innovation that can improve fitness for purpose and minimise running costs over a school's lifetime came in traditional schools".[74] In its inquiry on PFI of 2009-10, the Lords Committee on Economic Affairs received a written submission on design quality from academics at the University of Edinburgh:

The NAO commissioned the Building Research Establishment to compare design quality between a group of PFI and a group of non-PFI hospitals. It found that there were "no meaningful differences" in build quality between the two groups. However, it also noted that the average age of the non-PFI hospitals was much older.[75]

There are also other comparisons that have been done between PFI and non-PFI hospitals. A recent Committee of Public Accounts report said that:

One of the stated benefits of PFI is that it should ensure buildings are maintained to a high standard through the contracts' lives, yet 20% of Trusts were not satisfied with the maintenance service provided within their PFI contracts. In addition, unlike support services, the costs of maintenance cannot be revisited and are not subject to regular benchmarking.[76]

45. The National Audit Office's report The performance and management of hospital PFI contracts gave some examples of problems regarding the maintenance element of PFI contracts:

King's College Hospital was dissatisfied with lift maintenance. Broken lifts meant patients often share lifts with visitors to get to operating theatre. This is an ongoing issue yet to be resolved.

Hull and East Yorkshire experienced poor performance on some maintenance work. A high level of involvement from matrons has since ensured that clinical and maintenance services run smoothly together.[77]

The NAO report found when comparing PFI and non-PFI hospitals that there was no significant difference in the assessment of the environment and little difference in costs charged for services.[78]

46. It is difficult to establish clear cut evidence in the area of whole life costing. In theory whole life costing should encourage the use of innovative designs in PFI to deliver buildings of better quality. These should in turn provide cost savings over the life of the building that can, to some extent, offset the higher financing costs inherent in a privately financed deal. The long term nature of a PFI contract should also incentivise providers to maintain buildings to a high quality thus reducing costs in later life. However we have not been provided with clear evidence to suggest that PFI performs better in this area. Indeed in the area of design innovation and building quality we have seen some evidence to suggest that PFI performs less well than traditionally procured buildings.

To 'time and budget'

47. We received a number of written submissions which presented one of the key benefits of PFI as being the method more likely to deliver to time and budget than conventional procurement methods. The CBI made this point in its written submission:

[...] transferring financial risk to the private sector partner has contributed to improved performance during the construction phase, with a larger proportion of projects being delivered on time and within budget.[79]

So did PricewaterhouseCoopers:

At the outset financiers perform detailed due diligence on assets, costs and contracts using technical advisors to ensure the project will be delivered on time and to budget.[80]

The NAO, in a report for the Lords Economic Affairs Committee, noted that: "Most private finance projects are built close to the agreed time, price and specification." However they noted that of their sample of PFI projects over 31% had been completed late and 35% had not been delivered for the contracted price. They explained that "using PFI is not a panacea for solving construction problems."[81]

48. Any improved performance in terms of time and budget is only an achievement if the benefit outweighs any extra cost involved. The BMA considered that there was a 'risk premium' which meant overall the benefit of being on time and budget was not good value:

[...] research which found that hospital trusts were paying a 'risk premium'—conservatively estimated at 30% of the total construction costs—to ensure projects are running to time and budget. So while it is true that the private sector absorbs the cost of overruns etc, additional charges are written into the contracts to account for this.[82]

A report published by the European Investment Bank estimated that the contracted price was 24% higher for PPP roads than conventionally procured roads. The authors considered that the difference was largely due to cost overruns in traditional procurement meaning that there was little difference in the overall out-turn cost of both methods.[83]

49. If the budget is already 20% higher in a PFI procurement then a budget overrun of less than 20% in a conventional procurement would mean it was still cheaper. It is therefore important to consider how much projects which do not meet their budget exceed it. A National Audit Office report which considered a group of public sector projects that went over budget in 2003 and 2004, reported that the average level of overspend was 4.1%.[84] Any improvement in delivery to time also needs to be seen in the context of the procurement process. Submissions to the Committee recognised that for PFI this process was complex and lengthy.[85] The UK Contractors group told us that "even now the procurement process for a new hospital project in the UK can take over two years before any construction work is undertaken."[86] The NAO reported in 2007 that on average the overall tendering process took over two years for schools and over three years for hospitals.[87] HM Treasury in its document Meeting the Investment Challenge recognised this as an issue: "Procuring through PFI can be complex and can involve lengthy negotiations before contracts are signed." It added "Long lead times are a result of a number of factors, some common to all procurement, and some associated with PFI".[88]

50. There are also other reasons why to focus on the baseline of 'time and budget' may be misleading. The price of construction in conventional procurement is agreed at a stage of project development that is equivalent to a much less advanced stage than in a PFI. The risk control mechanisms built into the PFI model are factored into the price before contracts are signed. It is known that contract costs can increase during the preferred bidder stage of procurement, an exclusive stage of bidding in which competitive tension is absent and the public authority is in a weak negotiating position.[89]

51. The fixed nature of PFI contracts means they are likely to provide more certainty regarding price and time. However there is no convincing evidence to suggest that PFI projects are delivered more quickly and at a lower out-turn cost than projects using conventional procurement methods. On the contrary, the lengthy procurement process makes it likely that a PFI building will take longer to deliver, if the length of the whole process is considered. Proposing that post-contractual price certainty can be taken as a good measure of overall cost efficiency is to use a comparison already likely to favour PFI. This is because the PFI contract price is set at a much more advanced stage in the process. It is evident that a project delivered "to time and to budget" (in post-contractual terms) may nonetheless represent poor value for money if the price paid for the risk transfer was too high.

Flexibility

52. One issue that was prominent in both the written evidence we read and from our witnesses was the inherent inflexibility of PFI. Transport for London were clear that its "experience is that PFIs are the least flexible form of contract". It told us that PFI bound "both client and contractor to a series of outputs that have diminishing desirability and/or affordability, with much less scope to negotiate change than under other forms of contract". However this inflexibility did have both pros and cons: "This can be a strength—as client changes are often a significant cause of cost overruns—but is also a major constraint."[90] Steve Allen, the Managing Director—Finance at TfL told us that PFI was "therefore only suitable for procurements where you don't need to change what it is you require over the life of the contract". Mr Allen explained that the financing of a PFI made any changes much more difficult:

The involvement of the finance in the PFI makes it more inflexible, because it is not just a question of negotiating with the contract who built the asset. Particularly if the change is going to require some significant amount of funding, they are going to negotiate with the equity investors and with the debt holders as well.[91]

Professor Helm agreed that the structure of a PFI made it more inflexible. He told us that a PFI acted to "bundle the finance and fossilise the contract and put in the inflexibility that costs us so much both in terms of the efficiency of the project and in terms of the cost of capital".[92] Barclays Infrastructure also noted that the financial structure was a reason for inflexibility:

These problems are accentuated by the capital structure used in most PFI transactions, where leverage is = 90% and hence all variations require multi-party involvement and consent. Such leverage results in a low cost of capital but is restrictive to future change as there is little incentive on lenders (who are the dominant capital providers) to facilitate change.[93]

53. Anthony Rabin, the Deputy Chief Executive of Balfour Beatty, told us that the best way to allow for future changes in requirements "would be to have that discussion at the start of the contract to allow the public sector sufficient flexibility".[94] However Mr Allen told us that although some flexibility could be built into a contract there would always be limitations particularly as some issues would only emerge once the work had started:

You can build certain amounts of flexibility into the contract that you let if you can foresee what flexibilities you need, but there will always be limits around that, and it will affect the appetite of people to bid and the price that they will bid for that.

He explained that in some cases the only way to resolve problems was to bring a PFI back in house:

[...] if I go back to the Croydon [Tramlink] example, essentially what we had to do was buy the SPV from the shareholders because there wasn't the flexibility to renegotiate the terms of the contract.[95]

He emphasised that it was the ability of TfL to borrow directly that gave them the flexibility to opt out of the inflexible PFI contracts:

[...] having the ability to borrow ourselves [...] gives us some flexibility in renegotiating existing contracts in that we can buy the debt back and refinance it on our own terms, and we have had examples of that.[96]

54. We explored the effect of the inflexibility in PFI projects in some detail particularly with regard to the health sector. Professor James Barlow told us "I think one problem is that we have large, highly-specified buildings that are inflexible". He went onto explain that "my concern really is about the inflexibility of these buildings and the impossibility of, over a 30 or 40-year period, predicting what the demand is going to be like for the bed spaces in those buildings". He believed that the "way risk was devolved and transferred" in a PFI meant it was "very difficult to stimulate [...] any real sort of attempt to think about future flexibility".[97] Jo Webber of the NHS Confederation expressed similar concerns about being fixed into long term contracts:

[...] the most recent sort of direction of travel for care is to have it much closer to home, much more around people in their own communities. It is very difficult to change a very high-value environment like a ward environment into something that is affordable.[98]

55. The British Medical Association considered that the fixed nature of the unitary payments agreed in PFI contracts would mean that efficiency savings would be more difficult to achieve. "The NHS is being tasked to find efficiency savings of £20 billion by 2014-15". It went onto explain:

[...] at the same time (during the next spending review period from 2011 to 2014) repayments for NHS PFI projects will reach £4.18 billion, an increase of almost £1 billion from current levels. As a legal contract PFI removes discretion in capital spending and it is likely that hospitals will be forced to make cuts to health care services to make their ongoing PFI repayments.[99]

Jo Webber told us that meeting PFI payments in the light of other pressures meant that "there will be a big affordability challenge over a long period"[100] adding that it "will become more of a challenge for people over the next few years".[101]

56. PFI contracts are inherently inflexible. Specifications for a 30 year contract must be agreed in detail at the start of a project. The PFI financing structure also requires negotiation with the equity and debt holders before any substantial changes are made during the life of a contract. Debt and equity holders have little to gain from changing profitable contracts so will be unlikely to agree to changes unless they significantly enhance profitability. We have received little evidence of the benefits of these arrangements, but much evidence about the drawbacks, especially for NHS projects. The inflexibility of PFI means that any emergent problems or new demands on an asset cannot be efficiently resolved. In the case of Transport for London its only option was to buy out the SPV, but most PFI procurers cannot afford to do this.

PFI and competition

57. If there is healthy competition in the PFI market this should drive down costs and result in better value for taxpayers. We received written submissions that pointed to a lack of competition in PFI. The Royal Institute of British Architects explained that competition was reduced as many architects were unable to bid for work "due to the limited entry routes to the market—the lack of design frameworks, or open competitions". It also pointed out that "the fact that contractors are required to have a design team on-board before bidding for the work, meaning they frequently use their own in-house design teams and a small number of practices that they have worked with previously."[102] Martin Blaiklock considered that one of the disadvantages of PFI was that it "reduces competition: high costs and complexity means only major companies can afford to bid for such concessions"[103]. However Dr James Robertson noted that one of the potential benefits of PFI was that it "may open up domestic markets to overseas competition".[104]

58. One of the issues we explored with witnesses was the procurement process and cost and whether or not this affected competition. Mr Rabin agreed that "relative to other forms of procurement it probably is expensive. It is more complex; there is a whole machinery about PFI that you need to get right, otherwise it doesn't work"[105]. Regarding the competitiveness of the market he added "I would perceive from our side of the table that there is a reasonable amount of competition".[106] When we asked Mr Rabin about how many new school PFIs Balfour Beatty had bid for he replied: "I would guess that one in three possibly we would have bid for, something like that."[107] When we challenged him about the fact that in some areas his company had been the only serious bidder, he told us: "That is not something we were aware of at the time".[108]

59. If costs are too high to bid this will act as a barrier to entry in the PFI market. Mr Friend concurred with the view expressed in a written submission that failed bids cost approximately £2m per school and £12m per hospital.[109] Mr Friend told us "We at Laing thought we were doing well if we won 40% of what we were shortlisted for. So, you are writing off those".[110]

60. In 2007 the National Audit Office noted that "there is evidence that PFI projects are receiving fewer developed bids than previously". A third of the PFI projects they surveyed (between April 2004 and May 2006) had attracted only two detailed bids. In the same period only 20% of the PFI projects received four or more bids—this compared unfavourably to an earlier survey (2003 and before) which showed that 50% of the projects received four or more bids.[111] It noted that the reason for fewer competing bids was "in part due to the cumulative impact of lengthy tendering periods and high bid costs". For example the overall tendering period lasted on average 34 months.[112]

61. The nature of PFI means that competition is likely to be less intense compared to other forms of procurement. We believe the barriers to entry to be too high, resulting in an uncompetitive market. The long, complex and costly procurement process limits the appetite for consortia to bid for projects and also means that only companies who can afford to lose millions of pounds in failed bids can be involved. The fact that consortia are formed to bid for projects also limits choice and competition. For example an architects' firm may have the best design or there may be one contractor that has produced the best proposal, but unless these designs and proposals are part of the chosen consortium's bid they will not be used. The long term nature and inherent complexity of the contracts also make comparison more difficult for clients, further undermining competitive pressure.

Assessment bias

62. All PFI projects have to complete a Value for Money (VfM) assessment of the PFI option compared to an conventional procurement option with funding provided by central government known as the PSC (Public Sector Comparator). The recent decision to use a PFI to redevelop the Royal Liverpool and Broadgreen University Hospital is a helpful example by which to consider the value for money case for PFI. The Outline Business Case shows that the 'VfM assessment' calculated that PFI could provide a value for money benefit of 0.03% compared to conventional procurement.[113] We are surprised about the supposed precision of this comparison given the inherent uncertainty in any long term investment decisions, and we are also concerned about some of the assumptions behind the VfM assessment. Many of these assumptions act to make the choice of PFI more likely—see Box 2 for details.

Box 2: Examples of assumptions in the 'VfM assessment' which made the choice of PFI more likely (Royal Liverpool and Broadgreen Hospital: Outline Business Case)

The Internal Rate of Return: The assessment tested the VFM case for investors who needed a IRR of between 13% and 15%. It was assumed that investors would only demand 13%—the lowest rate of return, which has rarely been achieved in similar projects. At the 14% and 15% level the PFI route was assessed as poor value for money.

The discount rate: "In accordance with Treasury guidance, a real discount rate of 3.5% has been used for the first 30 years of the project"[114]—this is much higher than the real rate on 30 year index linked gilts which is currently less than 1%.[115]

Whole life costs: Life cycle costs were adjusted down by 15% for the PFI option.[116]

Optimism bias: There is an assumption that the costing of the conventional procurement route will always be over optimistic. This resulted in an upward revision of the PSC option of 19% for the capital expenditure and of 15% for the operational expenditure.[117]

Tax: "An adjustment of 6% has been made" to increase the PSC option - to take corporation tax receipts from the PFI option into account. This adjustment suggests that approximately a quarter of all revenues paid to the PFI provider will be profits subject to corporation tax at the full rate.

Risk transfer: "The discounted value of transferred risk is assessed at 9.78%". This adjustment for 'risk transfer' acts to reduce the PFI cost.[118]

Transaction costs: The same value for transaction costs are used for both the PFI and PSC option. This goes against both the evidence we have taken[119] and the Treasury guidance which recognises that a PFI procurement involves "significant transaction costs"[120] which are greater than those of a PSC procurement.

Third party income: The trust estimates that income of £50,000 will be generated under the PFI option but not under the PSC option.[121] If this income had not materialised under the PFI option, the option would not have been assessed as best value.

63. The vast majority of the costs of this PFI project are related to the capital expenditure and its financing. Analysis (see Box 1 & Table 2) of the financing costs shows that the costs of financing the building of the new hospital were significantly higher (71%) than if the same finance had been raised by the government. It is therefore clear that the discounting of cash flows and other adjustments were significant as they resulted in the PFI option being assessed as better value. Coincidently, around the same time the go-ahead was given for this PFI hospital, a plan from the North Tees and Hartlepool Trust for a publicly funded[122] hospital was cancelled. The publicly funded plan had been originally chosen as it was judged as providing best value for money. Since the cancellation the trust have produced another VfM assessment which indicates that PFI is now best value for money and so it is now the "preferred option".[123]

64. We received submissions about the VfM assessment system. Martin Blaiklock noted in his submission that in 2003 government had reduced the discount rate. Reducing the discount rate meant that PFI projects would be less likely to be assessed as value for money compared to a public sector comparator. He explained:

To counterbalance this abrupt change, HM Treasury introduced the concept of 'Optimism Bias' to reflect, as they thought, the inherent under-estimation of costs that Government departments had demonstrated over past decades.

He went onto point out "no other government has formalised the over-runs into an 'across the board' regulation as has the UK through the application of Optimism Bias".[124] A written submission from Greg Dropkin and Sam Semoff also raised similar issues:

"Optimism Bias" is applied to conventional procurement but not to PFI, giving an inbuilt advantage to PFI in the comparison. Yet the Treasury has acknowledged that on-time and on-budget performance can be secured through conventional procurement, so long as the design and build services are procured through a fixed-price, "turn-key" contract.[125]

We received a written submission from JP Heawood, a resident of York, who provided an account of how a York Schools PFI project had come to be approved:

In the York Schools PFI Project, the executive summary of the Outline Business Case (OBC) gave the PFI cost as £11.1 million, with the projected Public Sector Comparator (PSC) better value at £10.3 million. But of course a bid with those figures wouldn't get public funding [...]

So, as was customary, an "estimated risk" figure of £1.4 million was added to the PSC, which made PFI look better value; York's bid was then accepted [...][126]

This account was consistent with an Audit Commission report on PFI schools. 9 of the 11 PFI schools that the report considered had relied on a risk adjustment to show they were better value for money than the PSC. It also explained that "where the PSC estimate of construction and running costs was much below the PFI cost, the cost of risk transfer added on was on average higher".[127]

65. We are concerned that the VfM appraisal system is biased to favour PFI. Assuming that there will always be significant cost over-runs within the non-PFI option is one example of this bias. There is an incentive for both HM Treasury and public bodies to present PFI as the best value for money option as it is often the only avenue for investment in the face of limited departmental capital budgets.

PFI—Value for Money?

66. HM Treasury has consistently said that PFI should only be used if it is the best value for money route of procurement.[128] We are concerned however that if public sector organisations do not have alternatives to PFI to complete capital projects there is a danger that they will use PFI even if it is not value for money. The British Medical Association felt that this had been the case in the NHS:

In theory, projects are value tested against what the project would cost under public finance. If this process concludes that private finance does not represent value for money, a public procurement method is supposed to be chosen. In a context where PFI is the only funding available and many NHS hospitals are in need of capital works, managers have faced 'perverse' incentives to 'manipulate' their assessments and subsequently we have seen a proliferation of PFI projects.[129]

One of our submissions quoted the leader of Liverpool City Council being interviewed on BBC radio in November 2010 about the recently approved Outline Business Case for redevelopment of the Royal Liverpool and Broadgreen University Hospitals. He explained in the interview: "I know it doesn't provide Value for Money now or in the future, but its the only game in town".[130] Professor Ron Hodges shared a similar concern: "PFI was seen as the only likely route to obtain funding [...] its result was to promote a view that managers needed to support the PFI route if a project was to be completed".[131]

67. The lack of alternatives to PFI was pointed out in many pieces of evidence. Andy Friend said that his own observations when involved in the market in the past concurred with a "repeated phrase in the evidence before you: it was the only game in town, therefore we went for it".[132] The Foundation Trust Network told us "there should be other alternatives" and that "historically there has been insufficient support for capital investment and maintenance".[133] The British Medical Association made the point that "governments' preference for PFI means it has been viewed as 'the only game in town' for the last decade".[134] The NHS Confederation also told us in their written evidence that "it is important the debate acknowledges that without PFI there would have been few alternative sources of capital funding for large projects".[135] Jo Webber from the Confederation told us that "101 of the 135 new NHS hospitals have been through PFI between 1997 and 2009".[136] A recent Committee of Public Accounts report also noted that "In many cases local authorities and Trusts chose the PFI route because the Departments offered no realistic funding alternative."[137]

68. The most straightforward alternative to the use of PFI is capital spending direct from a capital budget. The annual budget allocated to every government department and public body is split out between the current (resource) budget and the capital budget. If a Department does not have a capital budget to meet its investment needs the only alternative is to turn to some form of private financing and use the current budget to meet the annual payments. This issue is likely to become more acute in the coming years as the capital budgets of Departments are cut significantly whereas current budgets are reduced to a lesser extent. The October 2010 Spending Review detailed cuts of 29% in real terms to the total capital budgets of departments, compared to a 8.3% cut to their current spending over the same 5 year period.[138]

69. For too long PFI has been the 'only game in town' in some sectors which have not been provided with adequate capital budgets for their investment needs. This problem is likely to get worse in the future with capital budgets cut significantly at the Spending Review. If PFI is the only option for necessary capital expenditure then it will be used even if it is not value for money. A much-needed reappraisal of PFI needs to be accompanied by a similar reassessment of its effects on overall capital spending in the public sector.

70. We received evidence from one organisation that did have access to alternative forms of finance—Transport for London. It explained that this "focuses the decision on Value for Money (VfM) rather than being skewed by a desire to access either 'free money' or guarantees of long-term funding to support the PFI payments". It said regarding the East London Line Extension "TfL, on inheriting the project from the Strategic Rail Authority, switched it from being financed privately to being financed by TfL."[139] Steve Allen from TfL explained: "most of the PFI contracts that we have were let, there was no alternative source of finance for the sort of predecessor entities, so there was no valid comparison". However now TfL "have the ability to borrow directly, we do have that comparator, and so you can, in a very real sense, assess the value for money of the PFI solution."[140] Mr Allen told us that although TfL's "cost of borrowing is probably something to the order of between 0.5% and 1% above gilt rates" it was lower than the additional cost of financing from PFI.[141] He concluded "I think it is hard to say that if you look across all the projects, overall PFI is value for money against that additional cost of finance."[142]

71. The price of finance is significantly higher with a PFI. The financial cost of repaying the capital investment of PFI investors is therefore considerably greater than the equivalent repayment of direct government investment. We have not seen evidence to suggest that this inefficient method of financing has been offset by the perceived benefits of PFI from increased risk transfer. On the contrary there is evidence of the opposite. Organisations which have the option of other funding routes have increasingly opted against using PFI and have even brought PFIs back in-house. TfL's cost of borrowing is higher than government's, and yet it still considers this is overall better value for money than PFI. The incentive for government departments to use PFI to leverage up their budgets, and to some extent for the Treasury to use PFI to conceal debt, has resulted in neglecting the long term value for money implications. We do not believe that PFI can be relied upon to provide good value for money without substantial reform.

33 Ev 33 Back

34 A basis point is 0.01%(1/100^th of one percent). For example 250 basis points is 2.5% Back

35 Q 9 Back

36 Q 12 Back

37 Q 14 Back

38 Q 122 Back

39 C&AG report, Financing PFI projects in the credit crisis and the Treasury's response, HC 287, 2010-11, p9, para 18 Back

40 C&AG report, HC 287, 2010-11, p31, Appendix 2, Figure 31 Back

41 C&AG report, HC 287, 2010-11, p12, para 30 Back

42 International Handbook on Public-Private Partnerships, Chapter 14: The UK's Private Finance Initiative: history, evaluation, prospects. Mark Hellowell, 2010, p326-328 Back

43 An availability based project is of lower risk as the SPV only has to ensure that accommodation is available. If demand risk is also taken on by the SPV the cost of capital will be higher. Back

44 Source: Royal Liverpool and Broadgreen University Hospital Trust (2010), A New Health Service for Liverpool - World Class Hospitals, World Class Services, Volume 1 - Outline Business Case Back