PPPs: A policy in search of a rationale?
Private finance and 'value for money' in Britain's public hospitals
- The private finance initiative (PFI) brings no new capital investment into public services and is a debt which has to be serviced by future generations.
- The government's case for using the PFI rests on a value for money assessment skewed in favour of private finance.
- The higher costs of the PFI are due to financing costs which would not be incurred under public financing.
- Many hospital PFI schemes show value for money only after risk transfer, but the large risks said to be transferred are not justified.
- The PFI more than doubles the cost of capital as a percentage of trusts' annual operating income.
Since 1992 the British government has favoured paying for capital works in the public service through the private finance initiative (PFI) - that is, through loans raised by the private sector. For public hospitals, this means that a private sector consortium designs, builds, finances and operates the hospital. In return, the National Health Service (NHS) trust pays an annual fee to cover both the capital cost, including the cost of borrowing, and the maintenance of the hospital and any non-clinical services provided over the 25-35 year life of the contract. The policy has been controversial because of its high cost and the impact on clinical budgets.1
When first introduced in 1992, proponents claimed that the PFI would lead to more investment without increasing the public sector borrowing requirement. The UK budget surpluses of recent years (Â£23 billion - $A60 billion - for 2000-01 alone) have, however, been much greater than the total of Â£14 billion ($A37 billion) in private investment deals signed in 1997-2001. The present generation of taxpayers could have funded considerably more capital investment out of existing revenue, instead of displacing the cost onto future generations.2
Furthermore, there is no evidence that the PFI has increased overall levels of service. On the contrary, the policy's use in the NHS has had two main effects. Firstly, it has displaced the burden of debt from central government to NHS trusts, and with it the responsibility for managing spending controls and planning services, thereby hindering a coherent national strategy.3 Secondly, the high cost of the PFI schemes has presented NHS trusts with an affordability gap. This has been closed by external subsidies, the diversion of funds from clinical budgets, sales of assets, appeals for charitable donations, and, crucially, by 30 per cent cuts in bed capacity and 20 per cent reductions in staff in hospitals financed through PFI.4 Though NHS funds have increased since 1999, there is no evidence that much has flowed through to baseline services.5
Thus, not only are the macroeconomic arguments in favour of the PFI illusory; the PFI has also had a negative impact on levels of service. Largely as a result of this, the case for the PFI now rests on the "value for money" argument. The government's claim is that the PFI delivers value for money through lowering costs over the life of the project, because of greater private sector efficiency, and because the private sector assumes the risks that the public sector normally carries. Here we examine the extra costs to NHS trusts of private finance compared with public finance, and evaluate the value for money argument with respect to the risks transferred.6
Capital charging regime
Until 1991 all major capital expenditure in Britain's NHS was funded by central government from tax or government borrowing. The NHS did not have to pay interest or repay capital, so, from the perspective of the NHS trusts, in effect new equipment and buildings came 'free'. The 1990 NHS and Community Care Act established hospitals as independent business units in the public sector, and required them to pay for their use of capital through 'capital charges'. Capital charges are included in the prices charged to purchasers and comprise depreciation, interest, and 'dividends' based on the current replacement value of the assets. To pay their interest charges and dividends to the Treasury, trusts must make a surplus - after paying their operating costs and making a charge for depreciation - equal to 6 per cent of the value of their land, buildings, and equipment.
Value for money methodology and risk transfer
The government's procedure for demonstrating value for money under the PFI policy is based on an economic appraisal that compares the economic costs and benefits of alternative investment decisions. Using this appraisal, the annual costs of a scheme financed by the PFI are summed and compared with those of a notional publicly financed scheme, called the 'public sector comparator'. The methods used contain at least two disputable components: discounting and the costing of risk transfer.
Discounting: The government's preferred value for money method states that important economic costs arise from public expenditure and its timing. It is argued that unless public expenditure reflects the market cost of capital it could crowd out more beneficial private investment. This is the cost of capital argument. Secondly, it is argued that the timing of payments is economically significant because people value consumption today over consumption in a year's time or later. This is the time preference argument. Both these economic costs, the cost of capital and time preference, are expressed in a single rate known as the discount rate. By applying a discount rate to future payments, a net present cost is obtained. Thus, net present cost is not the actual cash cost but a way of expressing as a single value the effect of two hypothetical economic costs.7 The net present value (or net present cost) is derived by discounting future annual cash costs to reflect the time value of money - the fact that pounds spent in the future are worth less than pounds spent today. This method has implications for the relative costs of the two methods of financing the project. Under conventional public procurement, the capital costs are met and accounted for during the construction period, rarely more than three years, and so have relatively higher net 'present' value. Under the PFI, the costs are spread over 30 years and the more distant payments have lower net present value. The discount rate adopted therefore has a crucial impact on whether the PFI offers better value than the traditional grant system.8 The Treasury's discount rate is 6 per cent for NHS PFI projects, and welfare economists have repeatedly criticised it as being too high.9
Risk transfer: The second element of the value for money methodology is risk transfer. This requires the identification of the future pattern of risks and costs over the life of a project for a privately financed hospital, compared with a publicly financed hospital. The government claims that the apparently lower cost of publicly financed investment is due to the failure to take proper account of the extra costs incurred when things go wrong. Thus, a key component of the value for money case is to estimate the cost of the risks transferred to the private sector, and to add these costs to the public sector comparator.
The next part of this article identifies the extra costs of using private finance, and then we examine the two central justifications for these extra costs by evaluating the evidence for, and the impact of, discounting and risk transfer.
Comparing cash costs of PFI and public funding values
The cash costs and net present costs of individual PFI hospital schemes and their risk valuations were derived from published data in the House of Commons Health Select Committee Public Expenditure Memorandum (2000, 2001), and from full business cases for individual hospitals.10
We could not obtain comparative data on the total cash costs and annual cash flows of the PFI and the public sector comparators before the value for money analysis was made, because these data are not available. To understand the costs of the PFI we examined the structure of costs for three PFI schemes: North Durham, Carlisle, and Worcester.
We then examined the impact of the new capital investment on the annual capital costs to NHS trusts, comparing the present capital finance regime (capital charges) with the projected capital charges and payments to the consortia under the PFI (known as the annual availability fee). We also tried to estimate what the cost of the new investment would have been if the scheme had been funded out of public capital, by applying the 6 per cent capital charge that the Treasury currently requires (see background above) to the total construction costs of the new asset and adding in capital charges on retained estate.
For further schemes, we then show the effect on trust operating budgets of new investment financed using the PFI compared with current capital charges.11
Value for money analysis
We next examined the value for money case and show the impact of discounting before and after risk transfer on the PFI and public sector comparator for several schemes. We searched unsuccessfully for the methods and evidence base underpinning risk transfer calculations in the hospitals' full business cases and government guidance.
Structure of costs of PFI and public funding
For the three selected schemes, the financing costs - that is, the costs of raising the finance - account for 29 per cent of the total project costs under the PFI. Publicly financed capital does not incur these costs. [Click for Table 1]
Annual cost of capital: We have shown the effect of new investment on the capital charges trusts have to pay from their revenue budgets using the PFI, and an estimate of the costs that would be incurred using public finance for a scheme with similar construction costs. In all cases the annual cost of capital rises steeply but is more marked for the PFI. In the case of North Durham the PFI costs are almost double the estimated costs of a similar scheme funded by public finance.
Annual revenue costs: We have shown for eight trusts the capital costs under the PFI as a percentage of each hospital's current revenue in the first year of its operation when private finance is used, compared with the capital charges currently using public finance. In all cases, the annual cost of capital is higher under the PFI. [Table 2]
Risk and the value for money analysis
The impact of risk and discounting on the value for money analysis: We have shown for six hospitals that the net present value of the public sector comparator was lower than that of the PFI option, even after applying a 6 per cent discount rate. Only after risk transfer was included was the net present value of the PFI less than the public sector comparator. Also, after risk transfer, the difference between the PFI and the public sector comparator in all cases is marginal - for example, 0.05 per cent at Swindon and Marlborough.[Table 3]
Risk as a proportion of total capital costs: We have shown that the private sector's risk as a proportion of the total capital costs under the PFI varies enormously between projects, from 17.4 per cent in Swindon and Marlborough to 50.4 per cent at South Tees.
The contribution of risk to costs: We have shown that the value of risk transferred to the private sector is remarkably close to the amount needed to close the gap between the public sector comparator and the PFI. [Table 4]
The two most commonly asked questions about the use of the PFI are, firstly, how the costs of private finance and public funding of capital projects compare and, secondly, what would be the annual revenue cost to the NHS trust if the scheme were publicly funded. The data required to answer these questions have not been made publicly available, but our best estimates are that the costs of private finance are higher and that trusts pay much more than they would if the new buildings had been publicly funded.
The higher costs of PFI
The high costs of using the PFI are due in part to financing costs that a public sector alternative would not incur. The costs of raising finance at North Durham, Carlisle, and Worcester added an average of 29 per cent to the total capital costs of the schemes. There are several reasons for this. Firstly, private debt always costs more than public debt. Secondly, the amount of capital to be raised through loans or equity under the PFI is inflated by financing charges, such as professional fees and the "rolled up interest" due during the construction period (when the PFI consortium is not yet receiving any payments from the NHS trust). In addition there are fees for preparing the PFI bid and contract negotiations, which are not always identified in advance. For example, NHS Estates showed that the Carlisle proposal did not identify any costs "prior to the date of the signature of the agreement, unlike the [public sector comparator] where the trust has identified a cost associated with the preparation of the business case."12
Although new PFI hospitals replace two or three hospitals with one, are sited on less expensive out of town sites, have fewer beds, and use the proceeds of land sales and Treasury subsidies, they are still not revenue neutral. The cost of private capital as a percentage of trusts' annual revenue expenditure rises from an average of 8 per cent to up to 37.7 per cent. Without a concomitant increase in revenue, local services will struggle. In school PFIs local authorities receive an annual PFI credit from central government over and above their standard spending assessment to pay for the capital costs of PFI.13 In contrast, NHS trusts are expected to find the extra money from their own resources. Treasury policy is that there are still efficiency savings to be made in the NHS.
But, as we have shown, the switch in 1991 from government grant to debt finance means that all new investment, whether publicly or privately financed, increases the cost of capital to NHS trusts and translates into new revenue pressures. This explains why scarce NHS capital budgets are underspent, the backlog in maintenance and repairs has been rising, trusts have merged to dispose of estate, and 13 000 NHS beds have closed since 1997.14
Justifying the higher costs of private finance
The value for money analysis seems to be no more than a mechanism that has been created to make the case for using private finance. Even with a high discount rate (which favours the PFI), the PFI costs are still higher than those of the public sector comparator. So the value for money claims rest on risk transfer.
As we show, in all schemes risk transfer is the critical element in proving the value for money case. There is considerable variation between schemes in the absolute and relative value of risk transferred. What is striking, however, is that in all cases risk transfer almost equals the amount required to bridge the gap between the public sector comparator and the PFI. This suggests that the function of risk transfer is to disguise the true costs of the PFI and to close the difference between private finance and the much lower costs of conventional public procurement and private finance.
Even after this manipulation, however, the difference between the public sector comparator and the PFI is marginal, in many cases less than 0.1 per cent. Not only does this raise questions about the reliability and validity of the methods used, but it also raises serious questions about why the government is using an unevaluated method of procurement for critically important services.
The evidence for the risk assessment method
Risk is the most difficult and contentious part of the value for money methodology. The argument is that by getting the private sector consortium to bear some of the risks associated with the construction of the hospital and its subsequent management, a trust enjoys greater value for money than under a publicly financed alternative, where the trust would bear all the risks. There are three points to note.
Firstly, the Treasury's policy on risk transfer is that risk should be held by the party best able to control it. Contract theory, however, holds that risk is best managed when held by the party best able to bear it. The state is better able to bear the risks than the private sector.15
Secondly, risk transfer requires the ability to quantify the probability of things going wrong. There is no standard method for identifying and measuring the values of risk, and the government has not published the methods it uses. The business cases we examined do not reveal how the risks were identified and costed. Our findings are supported by a Treasury commissioned report which found that, in over two thirds of the business cases for hospital PFI schemes, the risk could not be identified. In the other cases risk transfer was largely attributed to construction cost risks, which would be dealt with by penalty clauses under traditional procurement contracts.16
Thirdly, risks can be transferred only through a contract that identifies them. Yet there is reason to cast doubt on the claim that contracts offer a means of transferring financial risk.17 Where a trust wishes to terminate a contract, either because of poor performance or insolvency of the private consortium, it still has to pay the consortium's financing costs, even though the latter is in default. It would otherwise have to take over the consortium's debts and liabilities, given that the lending institutions make their loans to the consortiums conditional on NHS guarantees. In such cases "the attempt to shift financial responsibility from the public to the private sector fails. De facto, a risk-sharing arrangement results from force majeure," as the Railtrack collapse has shown.18
And risk transfer can never cover all risks. For example, at Darent Valley Hospital in Dartford and Gravesham, nurses complained that the design was not conducive to effective care, and equipment was not working properly when the hospital first opened.19 At the new Princess Margaret Hospital in Swindon the recovery room is located 80 metres from the operating theatre. It is unclear how trusts can be compensated for such poor design.
The high value of risk transfer - up to 50 per cent of the total capital cost to the private sector - indicates the high levels of compensation being paid to the private sector for risk transfer. Yet external evidence questions the basis for such high valuations of risk. In several PFI projects the consortiums have refinanced their loans at a lower cost because the risks turned out to be less than expected, but the public sector is still paying the cost of the initial financing. PFI consortia have even advertised that their projects contain "little inherent risk", and have been able to issue bonds with a triple A rating, which indicates low risk. Finally, at least one construction company (Jarvis) has sold off its construction arm in order to concentrate on PFI projects, which it says are less risky than conventional construction projects.
What really happens to the risk
Two failed PFI schemes in Australia contain important evidence on risk transfer.20 The Victorian government had to buy back La Trobe Hospital from Australian Hospital Care in October 2000, because "the losses incurred by AHP on the contract meant it could no longer guarantee the hospital's standard of care". At Modbury Hospital the South Australian government had to come to the rescue of the contractor and increase its contractual payments or the contractor would have defaulted. Closer to home, the Benefits Agency and Passport Office fiascos, and other failed private finance schemes, show that ultimately the risk is not transferred - the taxpayer ends up paying for private sector risks.21
But, irrespective of whether and how much financial risk is actually transferred and to whom, the main risks are those that arise from technical obsolescence, changing regulation, and unmet patient needs; risks which ultimately the NHS, local communities, and patients will have to bear. Should conditions change during a 30 year contract, rendering the facilities unsuitable, the NHS will find itself locked into long term contracts and payments, and patients may find they have to go without care.
Allyson Pollock is Professor and Head of the Health Policy and Health Services Research Unit in the School of Public Policy, University College London (UCL), and the Director of Research and Development, University College Hospitals Trust. Jean Shaoul is Senior Lecturer in the School of Accounting and Finance, University of Manchester. Neil Vickers is Senior Research Fellow in the Health Policy and Health Services Research Unit, UCL. This article is reproduced from, and with the permission of, the British Medical Journal (BMJ) (vol. 324, 18 May 2002, pp. 1205-09, which may be read at http://bmj.com/cgi/content/full/324/7347/1205; the original also explains the notes attached to the tables, and includes extra tables). Note that this version of the article incorporates the later correction. The article sparked controversy in the British House of Commons, and the story can be read in the Guardian of 20 May by Roy Hattersley: "The perils of plain speaking: A privatisation critic is the victim of a government smear campaign". You may send correspondence on the article to Allyson Pollock at: email@example.com.
1. Gaffney, D., Pollock, A. M., Price, D., Shaoul, J., "NHS capital expenditure and the private finance initiative: expansion or contraction?", BMJ 1999a; 319: 48-51; Gaffney, D., Pollock, A.M., Price, D., Shaoul, J., "PFI in the NHS: is there an economic case?", BMJ 1999b; 319: 116-119; Pollock, A.M., Dunnigan, M., Gaffney, D., Price, D., Shaoul, J., "Planning the new NHS: downsizing for the 21st century", BMJ 1999c; 319: 179-184; Gaffney, D., Pollock, A.M., Price, D., Shaoul, J., "The private finance initiative: The politics of the private finance initiative and the new NHS", BMJ 1999d; 319: 249-253; Hawksworth, J., "Implications of the public sector financial control framework for PPPs", in: The private finance initiative: saviour, villain or irrelevance?, London: Institute of Public Policy Research, 2000; Heald, D., Scott, D., "Lessons from capital charging in the UK national health service", International Association of Management Journal, 1996; 8: 29-45.
2. Accounting Standards Board, Amendment to FRS5: reporting the substance of transactions: Private Finance Initiative and similar contracts, London: Accounting Standards Board, 1998; Broadbent, J., Haslam, C., Laughlin, R., "The origins and operation of the private finance initiative", in: Robinson, P. (ed.), The PFI: saviour, villain or irrelevance?, London: IPPR, 2000:26-47.
3. Pollock, A.M., Dunnigan, M., Gaffney, D., Macfarlane, A., Majeed, F.A., "What happens when the private sector plans hospital services for the NHS: three case studies under the private finance initiative", BMJ 1997; 314: 1266-1271.
4. For subsidies, diversions, asset sales and chariable appeals, see: Norfolk and Norwich Healthcare Trust, Final business case, Norwich: NNHT, 1999; University College London Hospitals, Final business case, London: UCLH, 2000. For bed capacity and staff, see: Gaffney, D., Pollock, A.M., Price, D., Shaoul, J.(1999b; 1999c), op. cit.
6. Sussex, J., The economics of the private finance initiative in the NHS, London: Office of Health Economics, 2001
7. Sussex, J., op. cit.; HM Treasury, Appraisal and evaluation in central government [Green book], 2nd ed., London: HMSO, 1997, chap 4, appendix G.
8. Gaffney, D., Pollock, A.M., Price, D., Shaoul, J.(1999b), op. cit.
9. Pearce, D.W., Ulph, D., A social discount rate for the United Kingdom Norwich: Centre for Social and Economic Research on the Global Environment, 1995.
10. Department of Health, Expenditure questionnaire 2000: Memorandum to the Health Committee, NHS resources and activity, London: Stationery Office, 2000; Department of Health, Memorandum to the Health Committee: Public Expenditure Questionnaire 2001, London: Stationery Office, 2001.
12. Carlisle Hospitals FBC, Appendices Volume 1, Appendix G, paragraph 8.1.e.
13. Rowland, D., Pollock, A.M., Price, D., The school governors' essential guide to PFI, London: Unison, 2002.
15. Ulph, D., Contract theory and the public private partnership proposals for the London Underground railway system, London: Industrial Society, 2000.
17. Pollock, A.M., "Will primary care trusts lead to US-style health care?", BMJ, 2001; 322: 964-967.
19. Hellowell, M., "Problems at Darent Valley Hospital", The PFI Report 2000; Sept:45:9.
20. Senate Community Affairs References Committee, Healing our hospitals, Canberra: Commonwealth of Australia, 2000.
21. Pollock, A.M., Vickers, N., "Private pie in the sky", Public Finance, 14 April 2000; 22-3; Pollock, A.M., Price, D., Dunnigan, M., Deficits before patients, London: UCL, 2000; Gaffney, D., Pollock, A., Downsizing for 21st century, London: Unison, 1999; Price, D., Gaffney, D., Pollock, A., The only game in town? A report of the Cumberland Infirmary, London: Unison, 1999.
Also on the Evatt site:
- The trouble with PPPs: An un-holy alliance, by Christopher Sheil
- There are other ways: PPPs & public policy, by Sharan Burrow
- News of the world: PPPs are a disaster, by Kenneth Davidson
- Public fraud initiative, by George Monbiot
- PPPs: Beneath the rhetoric, by John Quiggin & Christopher Sheil
Read Letters in response to our PPP articles:
- The next great privatisation debate, writes Geoff Lazarus
- A misleading and coercive reform, writes Athena Vongalis