Public-private partnerships in hospital development: lessons from the UK's `private finance initiative'.

Since 1994, the UK government's Private Finance Initiative (PFI) has required all public sector agencies seeking capital investment to try and obtain it via partnership with the private sector. This policy applies to the UK National Health Service (NHS). The partnership takes the specific form of a long-term contract, typically 30 years, with a private consortium to design, build, finance and operate a hospital or other facility for the NHS. Before 1994, direct financing by the private sector of investments in public assets was negligible. The Treasury's rules for appraising investment proposals had made it practically impossible to show a privately financed scheme to be more cost-effective than an Exchequer financed alternative. But a political change of heart reversed that position, so that today the government's presumption is that major NHS investments will all be privately financed unless there are strong reasons otherwise.

Although the PFI was introduced by a Conservative government, it was subsequently taken up and pursued enthusiastically by Labour. The first PFI hospital opened to patients in mid-2000. Many more are planned to follow. Since July 1997, when the first PFI hospital contract was signed, 85 percent of the funds for major NHS capital investment projects have been borrowed from the private sector and only 15 percent from the UK government. Before 1997, effectively 100 percent of NHS investment was financed by government borrowing. Exhibit 1 shows how much of the government's planned growth in NHS hospital investment is expected to be privately financed. The continuing large amounts of publicly financed investment (including that funded by sales of surplus assets) are largely devoted to small schemes, including repairs and maintenance. PFI finance is concentrated on larger new-build (rather than refurbishment) investments.

[EXHIBIT 1 OMITTED]

Much effort has been put by politicians, civil servants and public sector managers into sustaining the PFI policy. Although it is also applied to the provision of schools, roads, bridges, prisons, government offices and many other public sector activities, it is in the area of health care that the PFI policy has proved to be most controversial. Appraisal and discussion of the policy has tended to become over-simplified and polarized into `pro' and `anti' camps. This paper attempts a balanced assessment of the main lessons, positive and negative, that can be learned from the PFI experience in the UK NHS.

The following features of the NHS financial regime are particularly important to note:

1. The NHS is very largely tax-funded. Ninety-eight percent of NHS operating costs plus depreciation and interest charges in 1999 were met from general taxation and just two percent from charges on patients (mainly for prescription medicines and dental treatment).

2. NHS expenditure in total is subject to an annual budget limit, which is strictly enforced. Expenditure overruns in aggregate at the end of each fiscal year are generally a matter of a fraction of one percent of the total budget. NHS hospitals cannot go bankrupt; ultimately the government is responsible for paying a hospital's debts.

3. There is another cash limit, equally strictly enforced, on the amount of publicly financed capital investment that the NHS may undertake each year. But this does not apply if investment is financed by borrowing from the private sector under the PFI. This creates a strong incentive for NHS hospital managers to seek PFI finance for their planned investment, regardless of its cost-effectiveness compared with a publicly financed alternative.

4. In the prices they charge the various NHS health care purchasing bodies, NHS hospitals have to recover the costs of depreciation plus a six percent dividend charge on the current depreciated value of their publicly-funded assets. They also have to pay the capital servicing costs that are part of any PFI contracts they may have. Thus, capital servicing costs have to be met whatever the source of borrowing.

5. Although hospital services are labor- rather than capital intensive, the scale of assets tied up in the UK hospital sector, and the importance attached to them, are great. Staff account for 61 percent of hospital costs; depreciation and interest charges only eight percent. But the current replacement cost of fixed assets in the NHS is over 70 billion [pounds sterling]. Hospital investment decisions, and any associated hospital closures, are highly politically sensitive events.

6. Capital expenditure in the NHS is, in principle, planned to take account of the geographical distribution of health care needs, and of the scale, type and condition of the existing hospital stock. The social solidarity principle remains strong, including not favoring any one part of the country unduly. Decisions to permit major hospital building schemes remain in the hands of government ministers and their civil servants, regardless of whether the investment is funded by borrowing from the government or from private banks and capital markets.

For NHS capital investments financed by borrowing from the government, the design and construction work is undertaken by private contractors, selected by competitive tender. Subsequent provision of non-clinical services to operate the hospital (all services other than those provided by medical or nursing staff or by professions allied to medicine) is subject to competitive tender. The contracts in this case may be of short duration (e.g. three years); there may be numerous separate contractors or hospital directly-employed staff providing different services; and there is no necessary link between any of these contractors and the private firms who designed and built the hospital. Capital costs are paid as they are incurred. The key differences from this of the PFI approach to capital investment are that:

* design, build and operation of the hospital are all undertaken by a single private consortium, often a special purpose vehicle company created for the project in question;

* the PFI contract is very long-term, typically with a life of 30 years or more;

* the consortium borrows capital privately to fund the initial construction of the hospital.

Thus, PFI schemes are `design, build, finance and operate' (DBFO) arrangements.

BENEFITS AND COSTS: THEORY AND EVIDENCE

A problem encountered when researching the PFI in the NHS is the dissonance between publicly stated and privately confided views. Many NHS managers, (private sector) advisers to them and civil servants will readily admit off the record that conventionally financed schemes are more cost-effective than PFI alternatives. But they are less forthcoming in public. They have been told for years that if an affordable private sector bid cannot be attracted to provide their hospital development, then they will either have to do without, wait much longer, or make do with only a part of the desired investment. A further problem is the scarcity of hard data. It is impossible to be sure of the cost of capital within PFI consortia's bids; the true magnitudes and likelihoods of the risks involved in any project; and what the costs and benefits of a hypothetical, conventionally financed alternative would be. These analytical difficulties mean that much of the discussion of the benefits and costs of the PFI in the NHS inevitably relies on circumstantial evidence.

The following discussion focuses on the differences between the effects that PFI and conventional Exchequer funded routes to NHS capital investment can be expected to have. Assessing the net benefit or otherwise of the PFI implies determining its impact on the equity and efficiency with which resources are used.

There are numerous types of equity that could be considered relevant, but the main impact of investments in hospitals will be on the geographical equity of access to such facilities. The main mechanism by which greater geographical equality of access to NHS facilities is intended to be achieved is via the formula-based funding of the over 100 geographically defined health authorities/boards. England, Wales, Scotland and Northern Ireland each employ different formulae, but the underlying forms are similar. In each case, the intention is to take account of relative need for NHS services by the population in each area of the country. The formula-allocated funds cover all operating and capital servicing costs. A hospital can only invest in more assets if it can afford to within the revenues it is paid from these formula-allocated funds. This remains the case whether investment is funded conventionally or via the PFI. The existence of the PFI will therefore make little or no difference to the geographical distribution of NHS capital assets around the UK.

Efficiency may be considered under two broad headings:

* allocative efficiency--whether the choice of PFI or conventional financing affects the balance of different types of health services that the NHS produces, e.g. hospital versus community services;

* productive efficiency--to achieve a given mix of health care services, is the PFI a more or less cost-effective way of spending the NHS's budget than conventionally financed procurement of assets and services?

Concerning allocative efficiency, several authors have argued that the PFI distorts investment priorities in the NHS by concentrating resources on large new acute hospitals. Smaller schemes more suitable for primary, community and mental health care services, and refurbishment-based schemes that make use of existing assets, are argued to be relatively neglected. The PFI is seen by numerous commentators as replacing health service planning by doing what the private sector is most willing to finance (Dawson & Maynard, 1996; Boyle, 1997; Gaffney & Pollock, 1997; Pollock, Dunnigan, Gaffney, Price & Shaoul, 1999; Boyle & Harrison, 2000a, 2000b). The Audit Commission in its 1998 guidance Taking the Initiative a Framework for Purchasing under the Private Finance Initiative, warned NHS managers:

   It is also important to be aware that some schemes may be more
   attractive to the private sector than others--for example, because
   of their larger size. These schemes may not, however, be the highest
   priority for the public sector. (Audit Commission 1998, para. 19)

Most PFI investment is in large projects. Most small projects continue to be financed directly by the Exchequer. A survey of English Primary Care Groups (Paxton & Lissauer, 2000) finds that "partnerships with the private sector are being widely considered by Primary Care Groups" but that "significant reservations remain." The English NHS Plan announced that: "The NHS will enter into a new public private partnership within a new equity stake company--the NHS Local Improvement Finance Trust (`NHS Lift')--to improve primary care premises in England" (Department of Health 2000c, para. 4.11). The details of how NHS Lift will work and how cost-effective it will be relative to conventional financing of primary care capital investment projects remain to be revealed. The Scottish and Welsh NHS plans make no specific mention of using PFI to support investment in primary care facilities (Scottish Executive, 2000; National Assembly for Wales, 2001).

The private sector's lack of interest in smaller schemes is presumably due to lower expected profits there at any given level of risk. The costs of bidding for a PFI contract, including legal and other fees, are high and are only partly related to the scale of the contract on offer. The Health Service Journal published in September 2000 a review of progress with the PFI, which quoted the property investment director at Norwich Union Investment Management as saying "I suspect we will be looking principally at projects with a value in excess of 5 million [pounds sterling]. The very small ones perhaps involve more work than is viable ..." (Ward, 2000).

The PFI selection/bidding and contract negotiation process takes months, or even years, longer than for Exchequer financed schemes. Meara (1997) found that, for five major London hospital schemes, going through the PFI procurement process had added in each case up to two years. This offsets any advantages of quicker completion for PFI construction projects once they do get under way (Dawson & Maynard, 1996). The NHS Confederation argues that the PFI is unpopular with health service managers because of the larger transaction costs that are entailed compared with conventional procurement:

   ... the PFI process ... is at best a hindrance to the way we plan
   our capital developments. PFI is slow, it is bureaucratic, it
   requires us to put a vast amount of management time and consultancy
   fees at risk without the certainty of success. (House of Commons
   Health Committee, 1999, para. 142)

An average of nearly 3 million [pounds sterling] 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 (House of Commons, 2000).

No-one has yet suggested, let alone demonstrated, that the PFI will actually improve the allocative efficiency of NHS investment. At best it can be hoped to do no worse than conventional procurement of assets and services. At worst the PFI may have skewed investment towards high cost hospital schemes and away from smaller primary and community health care facilities.

The main debate about the net benefits or costs of the PFI turns on whether it increases productive efficiency. To what extent does the PFI lead to lower cost and/or higher quality services than conventional procurement? NHS PFI projects combine the following private sector inputs:

* design;

* construction and equipping of the new facility;

* provision of non-clinical services for the duration of the contract, typically 30 years or more;

* finance for the capital investment.

A necessary condition for a PFI scheme to be of net benefit relative to an equivalent conventionally funded scheme (the `public sector comparator'), is that the PFI project must yield lower costs or greater benefits in at least one of these areas.

DESIGN

Official statements about the PFI are optimistic about its ability to harness for the NHS and its users greater benefits of private sector ingenuity and creativity than are captured by conventional publicly financed procurement. Conventional procurement is characterized as specifying inputs: the number, size, type and location of rooms required, the frequency with which floors are to be cleaned, and so on. PFI contracts, by contrast, are supposed to specify outcomes: the types and scales of services that the hospital has to be capable of delivering, and the quality and standards of buildings etc. that need to be available to enable it to do that. How to achieve the outcomes is then left to the skill and ingenuity of the private consortium that wins the job.

It is a matter of speculation how the design of any particular hospital built under the PFI might have differed if the scheme had followed a conventional procurement route. To the outside observer there is no distinguishable difference between the two design processes. Under the PFI, the NHS is supposed to specify the outputs it wants rather than the inputs it expects to need. However, in practice, it is the same process as ever of NHS doctors, nurses, technicians, therapists and managers discussing with externally-appointed consultants, architects, surveyors and engineers how much of what size and type of facility to put where.

The one area of obvious difference is that with PFI schemes the hospital's designers are required by the private consortium they work with to give greater thought to the maintenance of the hospital over the 30 years after it is built than was the case when working directly for the NHS. Given its tight annual, government-imposed, cash limits, the NHS emphasizes avoiding up-front capital expenditure rather than trading-off higher spend now for lower maintenance and repair costs later. This is an important potential source of benefit under the PFI as it could result in lower lifetime costs.

BUILDING AND EQUIPPING

NHS hospitals have long been built and equipped by the private sector, the successful contractors being selected via competitive tender. In that respect the PFI is nothing new. Conventional construction contracts contain penalty clauses for poor performance or late completion. Improving the efficiency of the procurement of equipment and buildings has been the focus of much attention by NHS managers and civil servants for many years, which has led to progressively tighter control over project costs during the 1980s and 1990s (NHS Executive, 1994, 1999). However, there is a common belief that NHS capital projects invariably overrun on time and cost, and by large amounts. This belief has been fed by one or two high profile disasters but, in fact, the average performance has not been bad. In a recent memorandum to the Public Accounts Committee of the House of Commons, the NHS Executive stated:

   ... the overall performance of the NHS [on major capital projects]
   has shown a long-term improvement. The three year moving average for
   1988-91 showed a time over-run of 14 per cent and a cost over-run of
   13 per cent. The figures have steadily declined to about 8 per cent
   and 7 per cent respectively. (House of Commons Committee of Public
   Accounts, 1999)

All of the projects included in these figures were conventionally, not PFI, funded. Furthermore, measuring cost and time overruns on a capital project is not straightforward. Many claims of large cost increases are based on comparing the outturn cost with initial estimates made many years previously. Such comparisons take no account of the intervening general price inflation rate or of subsequent modifications to the scheme up to the moment that design and construction is put out to tender. If the same approach were used to evaluate PFI projects then they too would appear to overrun greatly on cost and time (Sussex, 2001, p. 50-55).

One of the main causes of the cost and time overruns that do occur in some publicly financed hospital construction schemes is design variations instigated by NHS managers or medical staff. A PFI project might avoid this element of cost escalation by preventing the NHS client from making design changes. But if the changes are necessary then they need to happen even with a PFI project, and the terms of the PFI contract would also have to be renegotiated to accommodate them. Even if the design changes were not essential they were presumably of some value to the NHS client, otherwise they would not have been allowed to increase costs and project duration. To focus solely on the extra cost is to ignore the extra value of the revised project.

Arguments in favor of PFI point out the incentives it provides for the contractor to avoid construction cost or time overruns. But similar penalties can be, and are, built into conventional procurement contracts. Arguably, the PFI has made NHS managers more aware of risk management and better at it than they were with pre-PFI, conventionally financed projects. If so, this is an important benefit. But the risk management lessons learned from PFI could now be applied equally well to publicly as to privately financed procurement. Also, writing penalty clauses into contracts does not guarantee that no cost overruns, delays in delivery or other performance shortfalls will have to be borne by the NHS. There is still plenty of room for dispute over who caused the overrun or below par performance and, therefore, who should bear the cost. Further, if to enforce the terms of a contract would risk driving a private contractor into bankruptcy, it may well be preferable to the NHS to agree to renegotiate the contract and pay more, or demand less.

OPERATING THE HOSPITAL: NON-CLINICAL SERVICES

Services delivered by doctors, nurses and the professions allied to medicine have so far been excluded by the government from PFI deals in the face of powerful opposition from those staff groups. The government is also keen not to appear to be privatizing the NHS--something that would be opposed by a majority of the British public. But provision of a wide range of non-clinical services by the private sector is an essential part of all PFI deals. Exhibit 2 lists the scope of non-clinical services that can be included.

EXHIBIT 2
SERVICES THAT MAY BE PROVIDED BY A PFI CONSORTIUM

** Accommodation

** * Building maintenance

** * Car parking

** * Catering

** Courier and postal services

** * Domestic services (i.e. cleaning) window cleaning and pest control

** Energy and utilities management

** Equipment maintenance

** Financial services

** Grounds and gardens maintenance

** Information management and technology

** * Laundry and linen

** * Pottering

** Reception

** Residential accommodation

** * Security

** Sterile supplies

** Stores

** * Switchboard and telecommunications

** * Transport (non-emergency)

* Waste disposal (including incineration)

Note: * A PFI contract may cover some or all of the services
listed. For example, only those marked with an asterisk were
included in Dartford and Gravesham NHS Trust's contract with
the Pentland consortium (National Audit Office, 1999a).

Source: Adapted from Smith (1999), p. 26.

However, the NHS has been required to competitively tender for catering, domestic services, laundry and linen, since 1983, long before the advent of the PFI. In-house teams have often won these contests, so that competitive tendering does not necessarily mean contracting out to the private sector. The other services listed in Exhibit 2 are also commonly procured by competitive tendering. Non-clinical services may be put out to tender singly or bundled together in broader facilities management contracts. Therefore, the cost-effectiveness of PFI-style procurement of non-clinical services relative to conventional procurement rests on whether signing a long-term contract with a PFI consortium that is also responsible for the design, financing and construction of the hospital, yields greater benefits than contracting with service providers directly.

A recent twist in government policy throws doubt on whether provision of some services by external contractors, including those in PFI consortia, is cost-effective. The English NHS Plan published in 2000 singles out the need for hospitals to be cleaner. Funds are now earmarked for the purpose of improving hospital cleanliness and NHS managers are made explicitly answerable for it. There are unannounced visits by "a specialist inspection team including patients" (Department of Health 2000c, paras. 4.14-4.15). All of this because, in the words of the Secretary of State for Health:

   All too often Compulsory Competitive Tendering lowered standards of
   cleanliness in our hospitals. Of course we should measure and compare
   how well cleaning and other services perform--and change them ff they
   do not give the best value for patients. But Compulsory Competitive
   Tendering has not improved care for patients. It has damaged the
   NHS for far too long. It will now go. (Milburn, 2000)

The Scottish and Welsh NHS Plans also take up the hospital cleanliness issue. Where this leaves the requirement hitherto to include non-clinical services in PFI tenders for hospital construction is unstated and unclear. However that conflict is to be resolved, this statement does not suggest great confidence in the likelihood of future efficiency gains from tendering hospital ancillary services.

An area still seen as a potential source of significant benefit from the PFI's bundling together of services with the provision of buildings, is the improved maintenance of those buildings. Whether these benefits are material will remain unanswered until the first few PFI-built hospitals have been in use for 10-20 years. However, there is certainly scope for improvement over existing building maintenance standards in the NHS. The backlog of overdue maintenance work totals 3.1 billion [pounds sterling] in England alone (Department of Health, 2000c, para. 4.9). To NHS hospitals facing tight financial constraints, it has long appeared acceptable to defer maintenance in order to release funds for new developments or to support existing staff commitments. The resulting shabby state of some NHS hospital buildings is apparent to all who visit them.

If in the future PFI buildings are indeed maintained better, this may be for one or more of the following reasons:

* more funds being committed to maintenance under the PFI contract than would otherwise be the case, because the PFI contract effectively ring-fences the resources. The NHS hospital is bound by contract for 30 years to keep paying for that maintenance and the private contractor is bound to ensure that certain standards are maintained. The NHS cannot divert the money elsewhere;

* a given amount of funding for maintenance having been used more effectively by the private PFI contractor than would have been achieved by NHS staff;

* synergies resulting from design, construction and maintenance of the hospital all being the responsibility of the same consortium;

* synergies, if any, resulting from the design, construction and maintenance consortium also borrowing the funds to finance the initial capital investment, rather than being paid up-front for it by the NHS client borrowing directly from the Exchequer.

However, only the last of these hypothetical sources of benefit requires DBFO procurement. The others could all be captured by competitively tendering hospital design, building and maintenance with long-term contracts that include penalties for poor performance. These would be DBO (design, build and operate) contracts and would represent a new form of public-private partnership. They would require a long-term relationship between the private contractor and the NHS Trust but, unlike the PFI, no private finance.

RISK AND THE COST OF CAPITAL

The fourth element provided by the private sector under PFI-based partnerships with the NHS, is capital. The funds for the initial investment in new hospital buildings and equipment are borrowed directly from private banks, bond- and equity markets, rather than from the Exchequer. The cost-effectiveness of the PFI as an alternative to conventional, publicly financed, procurement depends on the balance of risks and rewards that are transferred to the private sector. The cost of capital contained within the annual charges paid to PFI consortia is higher than the rate of interest that would have to be paid by the NHS on Treasury debt if Exchequer financing were used. The question is whether the greater borrowing cost is fully justified by the risks transferred to the private sector under the terms of PFI contracts. Much of the debate about this aspect of the PFI has been confused. Space limitations preclude a full discussion of the intricacies of this debate but the key points are as follows (Sussex, 2001 provides more detail).

Tax is not the issue, although the UK tax system can certainly open up a tax wedge between pre- and post-tax rates of return to a capital investment. The post-tax rate of return to a private corporate investor is typically one or two percentage points lower than the pre-tax rate. However, if tax were the sole explanation for higher private than public sector costs of capital, then the extra cost of PFI to the NHS would be balanced by an increased flow of tax funds to the Treasury. This need not worry the country's citizens. Hence the following discussion is about the post-tax cost of private finance compared with the cost of Exchequer finance.

It is helpful to distinguish between the cost of borrowing to finance a portfolio of investments covering a diverse range of risk characteristics, and borrowing for one specific scheme with a specific set of project risks attached to it. Arguments about whether public borrowing costs less than private for a broad portfolio of activities centre on views about the relative transaction costs involved and whether the perceived risk of default is, other things being equal, greater when lending to the private sector than when lending to the government. The relative private and public costs of capital for investment in a specific project will necessarily reflect any such underlying source of difference. Before the launch of the PFI, official guidance to the NHS and other public bodies stressed that:

   The government is generally able to finance activities more cheaply
   than individual agents in the private sector mainly because it is,
   from the perspective of the financial markets, a good debtor. It
   has the unique power to tax and its risks are spread over a wide
   range of activity. The transaction costs of government financing
   are also low and the market in government debt is liquid and
   efficient. The government is therefore able to borrow at
   very attractive terms. (Treasury, 1991, Annex F, para. 5)

Spackman (1991) argued that "bond finance for large private sector bodies typically costs up to a percentage point more than public borrowing." He referred to analysis which showed that over the period 1970-1985 private sector debentures yielded 0.9 percentage points more than government bonds (Melliss, 1991). This result suggests there may be an underlying and persistent cost disadvantage for private capital relative to government lending. This may in part be due to greater private sector transaction costs per pound invested, but presumably also derives from investors' fears, however small, of an unrecoverable default by even the largest of private corporations. Investors retain a (slightly) greater fear of someone disappearing in the night with their money when they lend to a private body than when they lend to the government.

With respect to project finance and project-specific risk, the current economic orthodoxy recognizes that the rate of interest to be paid on government borrowing is lower than that on private borrowing for similar projects, but argues that this is because the public sector cost of capital hides the true cost of risk. Klein (1997), Grout (1997) and Flemming & Mayer (1997) put the case clearly that public borrowing only appears cheaper than private because the government is able to coerce future taxpayers into meeting the cost of downside risks should they materialize.

Private consortia bidding for NHS PFI schemes typically combine a large building contractor, a facilities management company, and often other specialized parties such as equipment suppliers or energy management companies. The pattern hitherto has been for these companies to form special purpose vehicles with non-recourse financing, predominantly either bank loans or bonds linked to the contract with the NHS hospital, and a small amount of equity. The lenders' returns depend solely on the financial success of the specific project and the contract signed with the NHS hospital. Thus in a competitive capital market, the argument goes, the cost of capital implied by the terms of the PFI deal represents the true cost of the risks involved in the project. If public finance is available at a lower cost of capital for the same project, then the Exchequer is effectively subsidizing the scheme at the expense of taxpayers in future years, who would have to pay for any downside risks that materialized.

The cost of borrowing from the private sector that is implied within the overall annual charge made by a private consortium under a PFI contract is not identifiable from publicly available information. Heald (1997) commented,

   The Treasury cannot or will not quantify the additional financing
   costs consequent upon financing ... by private finance rather than
   by government borrowing, or quantify the interest rate differential
   ... Given the confidentiality which attaches to loan arrangements,
   systematic evidence about the additional cost of private finance
   can only be produced by the Treasury or, with a considerable time
   lag, the National Audit Office. (Heald, 1997)

A study commissioned by the Treasury found that across a sample of non-NHS PFI schemes the weighted average cost of private sector capital was "1-3 percentage points higher than public sector borrowing as measured by current gilts rates" and that the gap was narrowing over time (Arthur Andersen & Enterprise LSE, 2000, paragraph 2.9). Gilts, or `gilt-edged stocks', are government securities, the purchase of which represents near-riskless lending. It is unclear, however, how much in this study the private cost of capital had been adjusted for risk relative to government gilts in reaching this range of numbers.

As circumstantial evidence about the relative costs of public and private capital, the findings of Gaffney and Pollock are interesting. In a series of articles they have noted that NHS PFI schemes assume exceptionally large reductions in the required bed capacities of the hospitals concerned (Gaffney & Pollock, 1997, 1999; Pollock, Dunnigan, Gaffney, Macfarlane & Azeem Majeed 1997, 1999). There may be good reasons for some of these planned reductions, such as day cases replacing inpatient episodes, shorter lengths of inpatient stay and higher bed occupancy rates. The NHS had downward trends in bed numbers and upward trends in day cases long before the PFI arrived (Department of Health, 2000d; see especially para. 18). However, Pollock et al. (1997) noted that PFI schemes were based on lower projections of demand growth, and even greater reductions in lengths of stay and higher levels of occupancy, than national fiends would imply. Length of patient stay and the level of demand for such stays are independent of whether a hospital is PFI or conventionally financed. So one interpretation of Gaffney and Pollock's findings is that the severe downsizing of NHS hospitals in PFI schemes is partly in order to leave room for a higher private sector cost of capital. This interpretation is supported by unattributable conversations the author has had with NHS managers and their management consultant advisers involved in PFI schemes. A recurring view is that a conventionally financed scheme would be preferable but that no Exchequer funds are available for large investments in the NHS. Faced with a choice between continuing with poor facilities or going for a new but smaller PFI funded hospital, it is easy to select the latter with a clear conscience. Demonstrating on paper the cost-effectiveness of this choice relative to a public sector comparator, in order to satisfy official appraisal and audit requirements, then requires only a little ingenuity.

The crux of the argument about whether private finance costs more than public finance for an NHS capital investment, rests on whether the undoubtedly higher rate of interest paid to private lenders accurately reflects the costs of the risks they assume. Exhibit 3 lists the type of risks that might be involved in a public sector project for which a PFI deal would be considered. For brevity the description of the risks is limited to the downside, but there are usually matching upsides too.

EXHIBIT 3
CATEGORIES OF RISK

Design and Construction Risks

* Surveys and investigations fail to identify problems

* Construction period overruns, alternative service provision needed
during the delay

* Construction costs overrun

* Facilities not provided to the required specification

Commissioning and Operating Risks

* Services fall short of specified performance standards

* Assets unavailable for use--`availability risk'

* Operating costs higher than expected

* Inadequate maintenance of assets

* Costs of maintenance higher than expected

Demand Risk

* Use of assets/services falls short of expectations--also
called `volume risk'

Residual Value Risk

* Value of assets at end of contract falls short of expectations

Technology/Obsolescence Risks

* Quality of services adversely affected by assets becoming
technologically obsolete

* Asset renewal costs higher than expected

Regulation Risk

* Changes to tax system, planning regulations, environmental
standards, health and safety regulations, other legal requirements
including NHS-specific

Disposal Risk

* Where project includes disposal of surplus assets, risk that the sale
price is lower than expected

Source: National Audit Office (1999b), Appendix 2.

In NHS PFI deals the private consortium takes on no demand risk. Whatever the future demand for the services of the hospital, the consortium's profits will be unaffected. This distinguishes PFI in the health service from transport infrastructure schemes such as roads, bridges and tunnels. In those cases the PFI consortium's costs are largely invariant but their income and hence profits vary directly with the volume of use of the facilities they provide. The availability payment element of an NHS PFI contract--i.e. the charge for the contracted buildings, plant and equipment being available in working order throughout the specified period--will be paid whether the hospital is used or not.

Many of the risks listed in Exhibit 3 could also be transferred to the private sector with conventional, Exchequer financed, procurement. For example, construction can be paid for in a fixed price contract to prevent cost overruns falling on the NHS, whether Exchequer or PFI financed.. Equally, the same penalty clauses for underperformance can be built into contracts for provision of services whether they are separate contracts for individual services obtained by competitive tender in the usual way or are bundled up with numerous other services in a PFI contract. The following list sets out the public to private risk transfers possible under PFI that have not occurred with conventional NHS procurement in practice in the past:

1. Construction time overruns are deterred because the contractor is not paid anything until the hospital is finished;

2. Construction time overruns are deterred because the duration of the total contract is fixed, so that every month's delay in opening the new facility means one month less in which the contractor can earn availability payments from the NHS Trust;

3. Payments to the contractor are reduced if areas of the hospital are not used because the contractor has failed to make them available, and (smaller) reductions to payments if areas which are substandard are nevertheless used by the Trust for want of a better alternative;

4. The contractor is responsible for maintaining the hospital and the NHS Trust can reduce its payments to the contractor if they fail to do this; the contractor also bears the risk of maintenance costs or any related insurance costs increasing unexpectedly;

5. If the private consortium retains ownership of the hospital at the end of the contract period, they bear the residual value risk if the NHS Trust decides that it no longer wishes to use part or all of the hospital.

However, the first three types of risk transfer could in principle be achieved with publicly financed design build and operate contracts. There is no need for private borrowing.

It is unlikely that the residual value risk referred to in item 5 can in practice be transferred to the private sector in PFI deals. The government requires that PFI contracts either return assets to NHS ownership at the end of the contract life or that they include an option for the NHS to take the hospital back into public ownership then. The private sector does not write blank checks for the NHS, however, and so will only accept this if the financial terms under which return of assets to the NHS will take place are agreed in advance.

Transfer of maintenance risk (item 4) is the clearest area where the PFI could provide significant benefit beyond the best kind of publicly financed procurement of capital assets and competitive tendering of non-clinical services. The question is whether this justifies the higher cost of private borrowing. Private capital will not cost less than public, even after allowing for risks transferred. How much more it costs is uncertain from existing evidence.

It is sometimes argued that to suggest that the cost of public capital for an investment is less than the cost of private capital implies a belief that the efficiency of the economy would be improved if the government were to finance all investment. But this debating point overlooks the question of allocative efficiency. Private capital markets are part of the mechanism by which funds are allocated according to the relative expected risks and returns of investment opportunities. The government cannot fulfill that market function well as it has no fear of bankruptcy when it makes a mistake. When considering NHS and other public sector investments, however, the private market mechanism has already been explicitly set aside because it would fail to make socially desirable investments in those activities. Political allocation of resources has by popular vote been explicitly substituted for market allocation in those special cases only, which in the UK include health care.

SUMMARY AND CONCLUSIONS

The preceding pages provide an assessment of the individual elements of the PFI package to try to determine the scope for it to yield benefits relative to the publicly financed alternatives. The balance of pros and cons of PFI compared with conventional financing and procurement is unclear. There are no decisive arguments either way. PFI is not a priori better or worse value for money than conventional procurement when taking all elements into account: design, construction, provision of non-clinical services, financing. Relative to well-managed publicly financed procurement, the PFI:

* may or may not offer design improvements and lower construction costs;

* probably does not lead to more cost-effective support services;

* may involve higher costs of borrowing, even after accounting properly for risk; but

* will probably lead to more projects being completed on time; and

* will probably yield better maintained hospitals.

The PFI has taught the NH$, and other parts of the public sector, some important lessons. It has highlighted the need for NHS managers to think more carefully about the outcomes they require, rather than the inputs they think they need, and to take full account of all the risks that a project involves. Explicit estimation of risks, their scale and likelihood of occurrence, permits proper management of them, including designing the project so that it will maximize its benefits net of costs over its entire lifetime.

In aggregate, the cost advantages claimed in the official business cases for NHS PFI schemes, relative to their public sector comparators are small. Boyle & Harrison (2000b) report that for the first 11 major NHS PFI schemes in England, the average cost saving officially claimed for the PFI option relative to the public sector comparator was just 1.6 percent (range 0%-4.2%). These are the savings after adding to the public sector comparator's cost the estimated value of the risks that are transferred to the private sector in the PFI option, and have been calculated using the government's required 6 percent real annual discount rate. At a discount rate of 5 percent or less these calculated savings would disappear. Of course, such net present value calculations can never be precise and they may fail to take full account of possible qualitative differences between a privately financed scheme and an Exchequer financed alternative.

The PFI process has gradually been refined over time. Further gains could be achieved now by learning the lessons of PFI procurement and applying them in conventionally financed projects. If newly built PFI hospitals can be guaranteed a better maintained lifetime by signing long-term DBO contracts with private companies, so too can new hospitals whose construction is Exchequer financed. The result would be public-private partnerships without private capital; taking private banks out of the equation but leaving the private architects, engineers, builders, equipment and service suppliers in.

REFERENCES

Arthur Andersen, & Enterprise LSE. (2000, January 17). Value for money drivers in the private finance initiative, London: Treasury Taskforce Private Finance.

Audit Commission (1998). Taking the initiative: a framework for purchasing under the private finance initiative, London: Audit Commission.

Boyle, S. (1997, April, 26). The private finance initiative undermines rational planning of health services, BMJ, 314 (7089), 1214-1215.

Boyle, S., & Harrison, A. (2000a). PFI in health: the story so far, In G. Kelly, & P. Robinson (Eds.) A Healthy Partnership: the Future of Public Private Partnerships in the Health Service, Institute for Public Policy Research Commission on Public Private Partnerships, London: IPPR.

Boyle, S., & Harrison, A. (2000b, Autumn). Private Finance and Service Development, In J. Appleby, & A. Harrison (Eds.) Health Care UK, London: The King's Fund.

Dawson, D., & Maynard, A. (1996, August, 10). Private Finance for the Public Good? BMJ. 313,312.

Department of Health (1997). The Government's Expenditure Plans 1997/98 to 1999-2000. Department of Health Departmental Report, Cm 3612, London: The Stationery Office.

Department of Health (1998). The Government's Expenditure Plans 1998-1999. Departmental Report: the Health and Personal Social Services Programme, Cm 3912, London: The Stationery Office.

Department of Health (1999). The Government's Expenditure Plans 1999-2000. Departmental Report: the Health and Personal Social Services Programme, Cm 4203, London: The Stationery Office.

Department of Health (2000a). The Government's Expenditure Plans 2000-2001. Department of Health Departmental Report, Cm 4603, London: The Stationery Office.

Department of Health (2000b). Departmental Investment Strategy, London: Department of Health.

Department of Health (2000c). The NHS Plan: a Plan for Investment, a Plan for Reform, Cm 48181, London: The Stationery Office.

Department of Health (2000d). Shaping the Future NHS: Long Term Planning for Hospitals and Related Services. Consultation Document on the Findings of the National Bed Inquiry, London: Department of Health.

Flemming, J., & Mayer, C. (1997, Winter). The Assessment: Public-sector Investment, Oxford Review of Economic Policy, 13(4), 1-11.

Gaffney, D., & Pollock, A.M. (1997). Can the NHS Afford the Private Finance Initiative? London: BMA Health Policy and Economic Research Unit.

Gaffney, D., & Pollock, A.M. (1999, January-March). Pump-priming the PFI: why are privately financed hospital schemes being subsidized? Public Money and Management, 55-62.

Grout, P.A. (1997, Winter). The economics of the private finance initiative, Oxford Review of Economic Policy, 13(4),53-66.

Heald, D. (1997, December). Privately financed capital in public services, The Manchester School of Economic and Social Studies, LXV(5).

House of Commons (2000, July, 17). Hansard, col. 71w.

House of Commons Committee of Public Accounts (1999). Session 1998-99, Twenty-eighth Report. Department of Health: Cost Over-runs, Funding Problems and Delays on Guy's Hospital Phase III Development, HC 289, London: The Stationery Office.

House of Commons Health Committee (1999). Session 1998-99, Third Report. Future NHS Staffing Requirements, HC 38, Volume 1, London: The Stationery Office.

Klein, M. (1997, Winter). The risk premium for evaluating public projects, Oxford Review of Economic Policy. 13(4),29-42.

Meara, R. (1997). A Capital Conundrum: the Effect of the Private Finance Initiative on Strategic Change in London's Health Care, London: King's Fund.

Melliss, C. (1991). On a Puzzle of Returns to Physical and Financial Assets in the UK Company Sector 1965-1987, Government Economic Service Working Paper No. 114, London: HM Treasury.

Milburn, A. (2000, September, 27). Speech to the Labour Party Conference in Brighton by the Secretary of State for Health, Department of Health, press release 2000/0536.

National Assembly for Wales (2001, January). Improving Health in Wales. A Plan for the NHS with its Partners, Cardiff: Health Service Strategy Team, National Assembly for Wales.

National Audit Office (1999a). The PFI Contract for the New Dartford and Gravesham Hospital, HC 423, Session 1998--99, London: The Stationery Office.

National Audit Office (1999b). Examining the Value for Money of Deals under the Private Finance Initiative, HC 739, Session 1998-99, National Audit Office, London: The Stationery Office.

NHS Executive (1994). Capital Investment Manual, London: Her Majesty's Stationary Office.

NHS Executive (1999). Public Private Partnerships in the National Health Service: the Private Finance Initiative--Good Practice, London: The Stationery Office.

Paxton, W., & Lissauer, R. (2000). Partnerships in primary care, In G. Kelly, & P. Robinson (Eds.) A Healthy Partnership: the Future of Public Private Partnerships in the Health Service, Institute for Public Policy Research, Commission on Public Private Partnerships, London: Institute for Public Policy Research.

Pollock, A., Dunnigan, M., Gaffney, D., Macfarlane, A., & Azeem Majeed, F. (1997, April, 26). What happens when the private sector plans hospital services for the NHS: three case studies under the private finance initiative, BMJ, 314(7089), 1266-1271.

Pollock, A., Dunnigan, M., Gaffney, D., Price, D., & Shaoul, J. (1999, July). planning the "new" NHS: downsizing for the 21st century, BM J, 319,179-184.

Scottish Executive (2000, December). Our National Health. A Plan for Action, a Plan for Change, NHS Scotland, Edinburgh: The Stationery Office.

Smith, C. (1999). Making Sense of the Private Finance Initiative: Developing Public-private Partnerships, Abingdon: Radcliffe Medical Press.

Spackman, M. (1991). Discount Rates and Rates of Return in the Public Sector: Economic Issues, Government Economic Service Working Paper No. 113 (Treasury Working Paper No. 58), London: HM Treasury.

Sussex, J. (2001). The Economics of the Private Finance Initiative in the NHS, London: Office of Health Economics.

Treasury (1991). Economic Appraisal in Central Government. A Technical Guide for Government Departments, London: Her Majesty's Stationary Office.

Treasury (2000). Budget 2000. Prudent for a Purpose, London: The Stationery Office. Ward, S. (2000, September, 7). Not just acute idea, Health Service Journal, Special Report: Private Finance Initiative, 1-3.

JON SUSSEX Office of Health Economics (UK)

Address for correspondence: Jon Sussex, Office of Health Economics, 12 Whitehall, London SW1A 2DY UK, [e-mail].