4. Question 1 (a): How should the cost and benefits of PFI schemes be assessed?
The evidence of high costs, and high projected profits, in several of the PFI schemes we have examined, (see reference 3), indicates that the existing techniques designed to ensure value for money and affordability in PFI schemes have not worked well. However, on a more positive note, our research also indicates that more careful examination, by the public sector client, of the kind of financial projections we have been studying, could in itself reveal many potential problems in prospective PFI deals. (The projections we have been examining are the financial projections produced by the operating consortia at the time of the signing of the PFI contract: these are available to the public sector side).
Among the key indicators which we suggest the public sector side should be looking at are the following:
(a) The financial projections can be used to split out of the projected stream of unitary charge payments those costs associated with services (like operations, management, maintenance, and lifecycle costs), leaving what we have called the non-service element of the unitary charge: this non-service element essentially represents the cost to the public sector client of the provision of the original capital asset. The net present value of this non-service element can then be calculated, using a discount rate equal to the current National Loan Fund interest rate. This net present value indicates how much the public sector client could have borrowed from the National Loan Fund, for the same cost as the stream of non-service element payments. If this net present value is high relative to the cost of the original capital asset, then this is prima facie evidence that the project may involve excessive costs to the public sector. For several of the schemes we have examined, the ratio of NPV to capital was indeed very high, being close to or greater than two.
(b) The financial projections can also be used to calculate the projected internal rates of return (IRR) on the different sources of capital finance: and also, for each different source, the average outstanding debt on which this return is earned over the lifetime of the project. For example, in all of the eight cases we have examined the projected IRR on broad sense equity (that is, the aggregate of subordinate debt and equity proper) is 15% or more, and in two cases is over 20%: moreover, these returns were being earned on an outstanding debt which in five cases averaged, over the lifetime of the project, more than 200% of the amount of capital raised. Figures like these are potentially indicative of excess profits, which the public sector side in the negotiations should probe.
(c) The financial projections can also be used to probe for a mismatch between the profile of non-service element payments, (as projected on the basis of the original inflation assumptions), and the profile of projected senior debt charges. A common feature of the financial projections we examined was that the non-service element payments were basically projected to increase through the life of the project: while senior debt charges were projected to decline, and would commonly terminate completely several years before the end of the project. This left an increasing wedge between the non-service element and senior debt charges, which was largely available to be taken as profit. The existence of a marked mismatch between the profiles of the non-service element and senior debt charges is therefore another indicator of potential excess profits, which the public sector side should probe.
(d) Another aspect which the public sector should probe is how sensible the arrangements are for adjusting the unitary charge payments in the light of future variations in inflation rates away from their initially assumed values. In our wider study of the business cases and contracts for PFI schemes, we have found many examples where these arrangements appear highly questionable. For example, we have found several examples where the whole of the unitary charge is subject to indexation at whatever the future rate of inflation will be-even though major elements of the consortium's costs, (namely, senior and subordinate debt charges), are pre-determined, and will not increase in line with inflation. Such inappropriate indexation potentially hands the operating consortium a large future windfall profit if inflation increases above its initially assumed levels.
5. Question 1(b): What Discount Rate Should Be Used in Comparing Private Finance with Conventional Public Procurement?
Since public procurement would involve borrowing from the National Loan Fund, a fair discount rate for comparison with the public sector is the current NLF rate. Prior to 2003, the discount rate which the Treasury specified, of 6% real, had the effect of burdening the public sector comparator with an additional cost equal to the government's capital charge, even though, according to the Treasury guidance, the capital charge should not have come into the value for money comparison. The effect was to significantly bias the value for money comparison in favour of the PFI option. (See reference 5 for analysis showing how the Treasury discount rate had this effect.)
6. Question 1(b): Are Current Procurement Procedures Satisfactory?
A number of aspects of current procedures are problematic:
(a) The practice of selecting a preferred bidder, may, in effect, limit competition: there are indications of significant cost increases occurring between the selection of the preferred bidder and the final signing of the contract.
(b) PFI projects are typically large, bundling together into one contract both the provision of the capital asset, and services. This bundling is essential if the project is to be classed as an operating lease, and hence kept "off the books". There is, however, no doubt that large, bundled, contracts restrict competition. As evidence of this, witness the following quotation from a presentation by a major PFI provider, where it was stated that an advantage of PFI for them was "Tender costs and complexity reduce competition".
(c) Our examination of PFI contracts indicates that arrangements used to examine value for money and affordability are often inadequate. (See in particular reference 4 for a detailed analysis of a number of problems identified in the case of the new Royal Infirmary of Edinburgh.) In some cases the assessment of affordability is given in the business case for only one year: in some others, affordability is assessed by calculating an aggregate over the lifetime of the project with negative balances for many of the early years being "cancelled out" by positive ones in the later years-even though this procedure makes little sense.
(d) There is a downside to the narrow concentration on value for money in current procurement procedures. The scale of PFI is such that it has important effects on the wider economy. Passing an entire large project, possibly the biggest construction and services project an area is likely to have for many years, to a firm based outside that area, can have important adverse effects. Such effects could impact, for example, on the viability of local industry, on the depth of the local skills base, and on the amount of research and development carried out in the area. Some account should be taken of these wider effects in PFI procurement.
7. Question 1(d): Is enough information disclosed on private finance projects fully to assess whether the taxpayer is getting value for money?
No.
(a) Our own research, using Freedom of Information, has indicated the benefits of the wider information on PFI which is now starting to become available under FoI: however, FoI itself is by no means an easy or ideal way of accessing such information.
(b) As noted at 6 (d) above, the broader implications of PFI for the wider economy are very important. Information should be made available to enable these implications to be assessed: for example, on where procurement and research involved in PFI is actually sourced.
8. Question 3: Is there significant risk transfer to the private sector or is it more apparent than real?
The very high levels of costs to the public sector and projected profits observed in some of the PFI schemes we have examined have implications for the true extent of risk transfer. If costs and profits are high relative to the levels of risk transfer conventionally assumed, then meaningful risk transfer cannot really be said to have taken place. In such circumstances, instead of meaningful risk transfer, what is really happening is that the private sector could be said to be having a flutter at the public sector's expense.
9. Question 5: Are post procurement alteration arrangements problematic?
We have no first hand evidence on actual re-negotiations. But our scrutiny of PFI contracts has identified a potentially significant weakness in the arrangements for re-negotiation. PFI contracts sometimes specify that, when re-negotiations take place, the guiding principle should be that the internal rate of return on equity should be maintained. But the IRR on its own is an inaccurate indicator of the value of a sequence of returns: the net present value of a sequence of returns depends both on the IRR and on the profile of debt outstanding. It would be perfectly feasible to structure a renegotiation so that the IRR was unaltered, but the profile of outstanding debt was changed so that the net present value of the return to the investor was greatly increased. Rules on renegotiation need to be much more tightly specified to avert this possibility.
10. Question 6: Treatment of PFI contracts in national accounts?
The evidence on PFI schemes which we have obtained under FoI raises two important implications for the way in which ONS handles PFI schemes in the national accounts.
(a) First of all, there is the question of how ONS treats existing "on book" PFI schemes in the national accounts.
When a PFI scheme is classified as "on book", ONS makes adjustments to the public sector capital stock and to the public sector net debt: the same adjustment, equal to the capital value of the asset involved, is made to both.
However, for the eight PFI schemes whose financial projections we have analysed, the ratios of the net present value of the non-service element of the unitary charge to the capital value of the asset were 2.04, 1.97, 1.97, 1.82, 1.68, 1.60, 1.49 and 1.28. (In calculating the net present values, the payment streams were discounted at the National Loan Fund interest rate.)
This means that the stream of payments which the public sector has contracted to pay for the capital asset, (leaving aside payments for services and lifecycle costs during the project), has a current value which is in each case much greater than the value of the capital asset. Given this, ONS's practice for "on book" schemes of including an amount equal to the value of the capital asset in the public sector net debt significantly understates the cost of the actual liability being taken on by the public sector.
(b) Secondly, there are implications for the way ONS has operated the existing ESA95 test of whether a scheme should come on to the books. As argued in paragraph 8 above, the high levels of projected profits associated with several of the schemes we have examined means that meaningful risk transfer has not actually taken place. This invalidates the standard ESA95 test of whether a scheme should be on or off the books, which is based on risk transfer.
For both these reasons, ONS's current approach is likely to grossly underestimate the cost of PFI schemes to the public finances. Too few schemes are "on book": and for those that are, ONS's addition to net debt understates the true level of public sector commitment. Note that under proposed changes to accounting rules, which may bring all schemes "on book", the second of these problems will still remain. (See reference 1 and 2 for an account of our debate with ONS on this issue.)
It is of vital importance that ONS move towards more accurate accounting treatment for PFI schemes. It is not just that the current approach grossly underestimates the true cost of PFI to the public finances. The misguided desire to classify PFI schemes as operating leases, (and hence to keep them "off book"), has fundamentally distorted the way PFI as a whole has operated: this has led to the very large bundled PFI contracts which have restricted competition, and which have almost certainly contributed to excessive costs.
11. Question 7; Would public sector investment have been lower without PFI?
It is not just the level of investment which is important, but its nature. We have no direct evidence on whether levels of investment would have been lower without PFI. But there is plenty of anecdotal evidence that the nature of investment would have been different without PFI, with much more refurbishment as compared to new build. A bias towards new build rather than refurbishment is potentially a sub-optimal allocation of resources.
12. Question 9: Are there realistic alternatives to PFI?
It would potentially greatly increase competition, and hence reduce cost to the public sector, if PFI projects were to be unbundled into much smaller constituent parts. If PFI schemes are automatically "on book" in future, then there is no longer any need to attempt to meet the condition that they should be capable of being classed as operating leases: and hence there is no need for such large bundled contracts in the future.
18 September 2009