1. The increase in the expected rate of return
on funds invested reflects both the increased amount of shareholders'
returns after the refinancing and the fact that all but a nominal
amount of the £7 million which had been invested by FPSL's
shareholders was repaid to them as part of the refinancing.
Source: National Audit Office from information
supplied by PricewaterhouseCoopers, FPSL's advisers
22. When the Prison Service awarded the Fazakerley
PFI prison contract to FPSL in 1995 the Service estimated that
the contract, worth £247 million, would only deliver marginal
financial savings of £1 million (less than one per cent).
By comparison, the Prison Service's contract for the Bridgend
PFI prison, awarded at the same time, was expected to generate
£53 million savings.
We asked the Prison Service why, given that it had estimated that
the Fazakerley prison contract was only expected to provide marginal
savings, it had not pressed for more than £1 million from
the refinancing. We were told that the Service and its advisers
had not forseen the prospect of a refinancing when the contract
was let in 1995 and there had been no Treasury guidance. The Service
said the advice it received at the time
was that if refinancing were to occur it was a matter entirely
for the contractor and not something to concern the Service. As
a result, the PFI contract had not addressed refinancing and the
sharing of benefits that might arise from refinancing.
23. The Service said there was an argument which
suggested that it did not need to approve the refinancing at all.
But it came to the conclusion that, because of the additional
termination liabilities that were forced upon it, it had a lever
to negotiate over £5.5 million of the refinancing benefits.
As the refinancing also gave FPSL greater confidence and made
the possibility of a contract termination more remote, the Service
considered that allowing the refinancing to proceed, with £1
million to be received by the Service, was a good deal, to the
taxpayers' advantage. The Service agreed, however, with our observation
that FPSL had benefited to a disproportionate extent from economic
stability and low interest rates and the taxpayer much less so.
24. In view of the high level of returns which would
accrue to FPSL's shareholders, we pressed the Home Office and
Prison Service as to whether the balance of benefit between taxpayers
and shareholders in the Fazakerley prison project is either reasonable
or justifiable. The Home Office said that it was unequivocal that
this had been an extremely good deal for the taxpayer and remained
a good deal. When letting the contract it had been in great need
of prison places at a time of rising prison population. It had
secured very high quality prison places at a cost per prison place
which was below that of the alternative means of providing them.
The Prison Service said that, while there had been some difficulties
when Fazakerley prison opened, it was an outstanding example of
a local prison and the best local prison it had. It considered
Fazakerley prison provided exceptional value for money in terms
of decent treatment of prisoners, education to prepare prisoners
for release and make them employable, security and treating prisoners
with dignity. In addition, FPSL had borne risk during the construction
phase and faced very significant penalties, at about £120
per prisoner place per night, for unavailable prisoner places.
25. The Home Office also observed that there was
a balance between the rewards for risk and the sharing of benefits
mutually that go with a long term, 25 year partnership. The partnerships
would not work if the two parties were intent on scoring off each
other. This balance had to be taken into account in negotiations.
The Home Office was satisfied that the negotiators approached
the refinancing deal professionally and thoroughly and was not
sure how much further the negotiations could have been pushed.
It might have been able to get a little more in the short term
but possibly at the expense of longer term liabilities. The Prison
Service did not consider its good partnership with FPSL had been
threatened by the refinancing and did not expect FPSL's performance
to be affected. It considered that, as FPSL was making more money
from the project, it was less likely to get into difficulties.
26. The Home Office said, however, that it would
be very unhappy if it was continuing to see a 39 per cent return
to private sector shareholders over the range of its contracts.
The Fazakerley contract had been the first PFI prison contract
which had established the market place. The Prison Service said
that, at the time, the Fazakerley prison deal had been good for
the public purse but it had subsequently got more competitive
deals with dramatically lower costs per prisoner place. Since
letting the first two PFI prison contracts
the present value of the average annual cost of each prisoner
place in five subsequent PFI prison contracts had been between
£9,850 and £12,100 compared with £16,467 for Fazakerley
prison. This average included a further PFI prison being opened
by Group 4 in 2001 (Rye Hill) where the cost per prisoner place
would be about 30 per cent lower.
27. The Home Office and the Prison Service acknowledged
that on another occasion they would seek more of the refinancing
benefits, as they were doing on current PFI contracts. The Prison
Service said that it intended that two PFI prison contracts it
was planning at Peterborough and Ashford would indicate that in
the event of any refinancing the Service should receive half of
any gains which are made by the private sector contractor. The
Home Office said that a PFI contract it had signed a few weeks
previously included a refinancing clause and it would expect every
single PFI programme with which it is associated to pick up on
28. At the time of concluding the refinancing negotiations,
the Prison Service also agreed to waive £500,000 which it
had previously withheld because of FPSL's under performance in
delivering the required service. This was part of an agreement
whereby the Service and FPSL agreed amendments to the contract
covering the payment deduction criteria, the service specification
and FPSL accepted some sharing of occupancy risk.
The Service acknowledged that it had given back £500,000
from the £1 million it had secured from the refinancing but
said it believed this was a very good deal for the taxpayer. It
would not have to pay for overcrowded places unless it was using
them. The Service suggested this would save millions of pounds
over the 25 year contract period.
29. Carillion told us that the increased rate of
shareholders'return after the refinancing was after tax in real
terms and reflected the full repayment of the shareholders' subordinated
debt which left only £100 invested. Carillion agreed with
us that as the prison had opened in December 1997 this meant that
FPSL's shareholders had achieved the payback of their investment
in two years. Carillion also acknowledged that the 39 per cent
rate of return excluded its profits on constructing the prison
and further profits which the shareholder businesses expected
to make on their trading operations with the prison. It further
confirmed that it expected higher construction profits on PFI
work compared to conventional building projects.
30. We were concerned whether the higher rate of
return which FPSL's shareholders would now receive could be justified
in terms of additional risk, as opposed to being a windfall or
extra payment simply for delivering the service FPSL had contracted
to provide. We therefore asked Carillion whether the level of
risk premium in the shareholders' expected returns was reasonable.
It said the 39 per cent return was technically accurate but was
distorted because the calculation was based on the nominal amount
of capital remaining after the refinancing. On the scale of the
rewards, Carillion and Group 4 said that as this was a privately
financed project they and the banks put finance at risk at the
outset and it was only at the end of the construction period that
they were receiving any monies back at all. They told us that
the perception of underlying risk on the part of sponsors and
financial institutions had been significantly higher when the
Fazakerley prison had been procured than would be the case today.
Never before had a prison been procured together with long term
external management services. Carillion and Group 4 said this
form of project was seen by some financial institutions to carry
political risk in addition to operational and commercial risk.
31. Carillion said that when it had negotiated the
Fazakerley prison contract the concept of a refinancing had been
very remote. As the first PFI prison the project had been very
difficult to finance and had to be financed entirely through foreign
banks. FPSL's shareholders had taken huge risks but the contract
had been successful in terms of the services it had delivered,
and had met its expectations in terms of the overall financial
terms. The prison construction had been completed five months
ahead of schedule which had benefited the Prison Service through
early access to prisoner places. The construction had been completed
in 41 months compared with 75 months for previous traditionally
procured prisons. There was also a stable regime in the prison.
Carillion and Group 4 believed the returns were not unreasonable
in relation to the way they had handled the risks and had performed.
32. Carillion had, however, previously told us that
it was seeking a 15 to 17 per cent rate of shareholders' return
on its initial PFI hospital investments. It said that it would
only get this level of return if it performed and this reflected
the high risks of hospital projects which were more risky than
prison projects. It considered hospitals were more risky because
they were highly complex and required the private sector consortium
to deal with different parties, including clinicians. By comparison,
with prison projects there was a fairly clear direction from the
33. We therefore asked whether Carillion had changed
its view on the relative risks of hospital and prisons projects
in the light of the 39 per cent shareholder return which it was
now expecting from the Fazakerley prison project. Carillion told
us that it had initially sought only a 12.8 per cent shareholder
return from the Fazakerley prison project which was considerably
less than the return it would expect on hospital projects. Even
as a result of performing the Fazakerley prison construction very
well, the return had only gone up to 16 per cent. It reasserted
that the 39 per cent shareholder return it was now expecting was
misleading because it was based on the nominal amount of share
capital remaining after the refinancing. Carillion also noted
that the refinancing benefits were not an immediate windfall and
would only be received over 25 years. It also said that, although
FPSL's shareholders were taking monies out of the project by way
of dividends, they would be using that to invest in future PFI
34. In the second half of 1999 FPSL was under great
time pressure to complete the refinancing. Tarmac wished to have
the subordinated debt it had lent to FPSL repaid by 31 December
1999 and because of uncertainty in the financial markets leading
up to the end of the millennium FPSL wished to complete the refinancing
by 30 November 1999.
The Prison Service acknowledged that with hindsight it had been
in a stronger negotiating position than it had realised.
35. Although the Prison Service had never before
been faced with a refinancing it did not ask its adviser Rothschild
to lead the negotiations or to attend any negotiation meetings.
Nor was Rothchild asked to provide any detailed briefing to the
Prison Service on how the negotiations should be handled to achieve
the best outcome.
We put it to the Prison Service that a better outcome could have
been achieved if the experience of Rothschild had been used in
the negotiations. The Prison Service agreed that there was a great
deal of merit in incentivising advisers and using them to negotiate
direct but considered that it had, nevertheless, achieved a very
satisfactory outcome. Its use of the best legal and financial
advice from the outset had contributed to this outcome. It said
it would, however, consider incentivising its advisers in future
negotiations. The Home Office said that incentivising advisers
should be considered on a case by case basis because there could
be situations where an adviser stood to benefit more by taking
a particular decision, or by deferring a decision.
36. The Service arranged for FPSL to pay the costs
of the Service's advisers.
In response to our concern that this arrangements could have impaired
the objectivity of the information provided by the advisers, the
Home Office assured us that FPSL had no say over the choice of
the Service's advisers or the extent of advice that would be offered.
The Prison Service was confident that the quality of Rothschild's
advice was not diminished by the payment arrangement which it
said Rothschild only knew about at the end of the negotiations.
37. When this contract was let in 1995, the Service
estimated that it would only deliver marginal savings of £1
million compared with conventional procurement. It was a much
less attractive deal in financial terms than the Bridgend prison
deal let at the same time to another consortium which was expected
to deliver savings of £53 million on a similar sized
contract. The refinancing appears, therefore, to give FPSL substantial
further benefits on a contract which at the outset did not give
the prospect of significant savings to the Prison Service.
38. The prison opened five months ahead of schedule
and is considered by the Service to be an outstanding local prison
which has paved the way for subsequent PFI contracts. But given
the scale of the improved benefits that have accrued to the consortium
from this refinancing the Service should have sought a more reasonable
balance of risk and rewards for both the Service and FPSL. The
gains should have been shared more equitably between the consortium
and the Service.
39. Carillion told the Committee that it would earn
a return of 15 to 17 per cent if it performed successfully on
hospital projects, and that it deemed these to be more risky than
prison projects. Carillion claimed that the investment return
of 39 per cent which it now expected to make from the Fazakerley
prison project was a misleading figure. In our opinion, however,
the 39 per cent figure correctly reflects the increased returns
that the FPSL shareholders will receive following the refinancing
on an investment which has been reduced to a nominal amount.
40. When assessing alternative PFI bids, departments
should take into account the various revenues which shareholders
of a consortium can earn from a PFI project, the likelihood of
a refinancing occurring and how this may affect the balance of
risk and reward, for both the procuring department and the service
41. Although this was the first major refinancing
of a PFI project, the Service chose not to make greater use of
its adviser, NM Rothschild & Sons (Rothschild), in determining
a negotiating strategy, and did not ask Rothschild to participate
in the negotiations. Given the complexities of PFI refinancing
and the potential financial consequences, departments should make
appropriate use of experienced advisers in developing, and participating
in, refinancing negotiations.
42. The Prison Service assured us that it has learned
lessons from the Fazakerley prison refinancing, that it is securing
rights to share in refinancing gains on current PFI prison contracts
and that it is achieving improvements of up to 30 per cent in
its PFI contract prices as a result of establishing a competitive
market following the successful opening of the Fazakerley prison.
The Need for
Further Central Guidance
43. When this early PFI contract was awarded in 1995,
there was no central guidance to show how gains from refinancing
could be shared equitably between private sector consortia and
departments. An awareness of what would be acceptable PFI contract
terms for the public and private sectors emerged over a number
of years as the first generation of PFI deals were concluded.
In July 1999, shortly before the Fazakerley prison refinancing
was finalised, the Treasury issued guidance on standard contract
terms which included some advice on how departments should address
refinancing issues. This guidance generally put forward a view
that refinancing benefits are gains which should accrue to the
private sector but stated that in limited circumstances it may
be appropriate for refinancing benefits to be shared with departments.
The Treasury considers that because of the complex issues raised
by refinancings, and in the light of market trends, more extensive
guidance is required.
44. There are likely to be similar opportunities
for refinancing other PFI contracts, particularly those signed
in the early stages of the development of this new form of procurement,
and where the required service has been successfully provided
by the private sector consortium. The National Audit Office identified
a number of principles which departments should keep in mind when
they assess issues relating to refinancing.
45. The National Audit Office's analysis of answers
to Parliamentary Questions tabled by the Rt Hon Alan Williams
MP (Swansea West, Labour) revealed, however, that only 24 per
cent of the 105 PFI projects covered in this survey included arrangements
whereby departments are entitled to share in refinancing gains.
The capital value of all contracts let under the PFI at 31 July
2000 was £17.3 billion.
46. The Prison Service said that when entering into
the Fazakerley prison contract, and subsequently considering the
refinancing, it had been in the dark about the benefits which
it would now hope to get from refinancing if it was in a similar
situation today. It said that had the Comptroller and Auditor
General's very helpful report been available it might have made
some different decisions. It had learnt from this experience and
had put in very different arrangements for future contracts.
47. We pressed the Treasury on why, bearing in mind
the large value of PFI contracts which had been signed up, it
had been unaware for so long of the possibilities of refinancing.
The Treasury said that as PFI was a new innovative process certain
things were not known when the Fazakerley prison contract was
entered into. When lessons were learned changes were made and
more changes would be made in the new guidance it planned to issue
in spring 2001. It agreed to draw the Committee's concerns to
the attention of the OGC which was taking forward the development
of the new guidance.
48. The Treasury expects the new guidance to continue
to recognise the private sector's rights to receive refinancing
benefits as a reward for the successful management of risks where
these are appropriately priced.
In considering whether this approach is sensible, we asked the
Prison Service why, if it agreed a contract expecting it to be
successful, its contractors should be further rewarded for carrying
out the contract successfully. The Service said that in the context
of taking forward the Fazakerley prison project, which had been
the first PFI prison, there had been considerable risk involved.
There was a chance that FPSL could have lost considerable sums
of money, although the Service now thought that unlikely and the
prison had opened successfully. Group 4 noted that, rather than
simply delivering the contract, the consortium had outperformed
the standard of service expected at the time the PFI contract
was let by virtue of the prison opening early. That increased
confidence amongst the capital markets had not only enabled the
refinancing to take place but had also fundamentally affected
the marketplace and moved it forward. There had been an increased
appetite from the capital markets for this type of project and
increased competitiveness which was reflected in terms available
on subsequent PFI projects.
49. We suggested that splitting refinancing benefits
50:50 between the public and private sectors was a more reasonable
outcome, in terms of the taxpayers' interests, than the Prison
Service had achieved in respect of the Fazakerley prison refinancing.
The Treasury told us that the current guidance was not specific
about whether PFI contracts should include a provision that the
public sector should receive half of any gains which are made
by the private sector contractor. It said it was considering whether
a mandatory provision of this sort should be included in its new
The Prison Service told us that it accepted the 50:50 rule for
The Home Office said it had draft guidance under discussion which
emphasised that PFI contracts should include an unambiguous refinancing
clause and should make it clear that the Home Office would expect
to share a proportion of any benefit from refinancing. The Home
Office noted that it would probably have to pay for a 50:50 split
of benefits from refinancing. It saw the best protection for the
taxpayer as being vigorous competition to ensure the soundest
arrangements and the tightest prices at the outset rather than
having to rely on what it considered could be an artificial clawback
arrangement which would require subsequent renegotiation.
50. There are lessons which can be learned from privatisations
on how clawback arrangements can be appropriate to allow departments
to share in windfall gains. We established from privatisations
the doctrine of sharing in the unexpected gain which was beyond
the expectation of either side. In our report "Getting Value
for Money in Privatisations" we noted that in a number of
cases investors had made much higher returns than they ever imagined.
We identified a key lesson that clawback provisions can protect
the interests of the taxpayer by providing for a share in value
not identified at the time of the sale. In these circumstances
high threshold clawback schemes would have had a minimal depressive
effect on the initial price.
51. Although PFI contracts with a capital value of
approximately £17 billion had been let by July 2000, there
was no central guidance on refinancing until July 1999, by which
time most of those contracts had been let or were under development.
The Treasury should aim to anticipate future issues where departments
may require guidance rather than simply producing guidance in
response to situations which have already developed. It should
consult external experts and the National Audit Office about emerging
issues where central guidance would be helpful.
52. Many PFI projects, particularly where contracts
were let in the early stages of the development of the PFI, are
likely to be refinanced. The National Audit Office's analysis
shows, however, that only 24 per cent of PFI projects surveyed
included arrangements whereby departments are entitled to share
in refinancing gains.
53. The Treasury and the PFI Policy Unit in the OGC
should therefore complete their planned updating of the central
guidance on refinancing as a matter of priority.
54. The opportunity for refinancing benefits appears,
in part, to arise from the successful delivery of a PFI project.
PFI deals should therefore reflect the benefit of the improved
financing terms that are likely to arise through the successful
delivery of the project. The benefit may be secured through the
pricing of the deal or through a share of subsequent refinancing
55. The experience of privatisations shows that in
some cases private sector investors have made much higher returns
than they ever imagined. We advocated that such unexpected gains
should be shared. Windfall refinancing benefits on PFI projects
which have not arisen through a higher than expected standard
of service from the private sector should similarly be shared
between departments and the private sector. Because deals will
not have been priced in anticipation of such gains arising, the
prospect of sharing the gains between the public and private sectors
will have no impact on the original pricing of the deals.
56. All departments must give careful consideration
to refinancing issues when they develop contractual arrangements
with PFI consortia, taking account of the lessons from the Fazakerley
prison refinancing and further guidance which the Treasury and
the OGC may develop.
57. We look to the National Audit Office to carry
out a further analysis at the end of 2001 of the extent to which
PFI contracts allow departments to share in refinancing gains
so that we can monitor progress on these important issues.
12 Evidence, Qs 70, 117-122 Back
report, Figure 2, p4 and para 3.17 Back
Qs 90-91, 101-111 Back
report, para 3.17 Back
para 2.11. Subordinated debt is a loan which will rank behind
the principal borrowings of a company for repayment on occurrence
of certain events (such as insolvency) Back
para 3.17 and the C&AG's report on the The PFI Contracts for
Bridgend and Fazakerley prisons (HC 253 Session 1997-98)
Figure 10, p46 Back
Lazard Brothers & Co. Limited (see C&AG's report, para
Qs 17, 50, 68, 131, 136 Back
Qs 1, 5, 51, 54-55, 58, 66-67, 129 Back
Qs 7, 29, 36-39, 72-74, 124 Back
Qs 69, 74, 76 Back
Fazakerley prison contract let in December 1995 and the Bridgend
prison (now known as HMP Parc) contract let in January 1996 Back
Qs 31-32, 46, 74-76, and Evidence, Appendix 1, pp 17-18 Back
Qs 2-3, 11, 17, 113 Back
report, paras 1.31-1.33 Back
Qs 57, 67 Back
Qs 23-28, 33, 64 Back
Qs 27, 65, 77-81, and Evidence, Appendix 2, pp 18-22 Back
Qs 8-10, 34-35, 58, and Evidence, Appendix 1, pp 17-18 Back
Report of the Committee of Public Accounts, Session 1999-2000
(HC 131 (1999-2000)) "The PFI Contract for the new Dartford
and Gravesham Hospital": Evidence, paras 55-58, 222-224 Back
Qs 43-46, 59-60, 65, 77 Back
report, para 3.34 Back
report paras 3.37-3.38 Back
6, 85-86, 135 Back
report, para 3.39 Back
Qs 83-84 Back
report, para 1.8 (The guidance on refinancings was set out in
section 14.6 of Standardisation of PFI Contracts (HM Treasury
July 1999) Back
report, para 1.12 Back
report, paras 2.1, 3.2 Back
Appendix 3, p23 et seq Back
para 13 (based on data from the Office of Government Commerce) Back
Qs 14-15 Back
report, para 1.12 Back
Qs 56, 61 Back
Qs 2-3 Back
Qs 13, 16 Back
Report of the Committee of Public Accounts, Session 1997-98 (HC
992 (97-98)), page ix Back