PFI versus conventional public
procurement: assessing Value for Money
78. In its memorandum the Department underlined the
fact that value for money (VFM) had to be proved before a project
could proceed down the PFI route. The only way to prove this with
absolute certainty would be to build two identical hospitals,
one using conventional, Treasury funding and the other PFI.[128]
In the absence of this, the PFI has to be tested against a hypothetical
model, the Public Sector Comparator (PSC). This value for money
test compares the full life cost of public provision (the PSC)
with that of the PFI alternative, and assesses the value of the
risk retained by the public sector in both options.[129]
Therefore, the PSC is basically a pass or fail test: if the net
present value of the PFI bid is below that of the PSC then the
deal is considered to be good value for money.
79. The value of the risk and its transfer from the
public sector to the private sector is central to the value for
money equation. As the Department explained: "PFI transfers
the risk of time-and cost-overruns to the private sector, who
are only paid once the facility is operating to the required standard.
Publicly funded projects were often subject to delays and increased
costs, and required extra capital to repair defects." Furthermore,
"under the PFI contract the annual payments to the private
sector partner are linked to performance and quality standards
which is not possible under the conventional public capital funding
route".[130] The
Secretary of State made this point to us when he suggested that
the only things which were privatised through the PFI were the
cost overruns and the time overruns.[131]
80. Certainly the performance of conventionally procured
projects has not always been impressive. Rectifying faults at
St Mary's Isle of Wight is costing the NHS £20 million, and
the cost of Guy's Hospital Phase 3 has risen 300 per cent and
been delayed by over three years.[132]
The Secretary of State believed that there were fundamental problems
with publicly procured projects:
"The point about the way that we procure
traditionally through the public sector regime is there is no
real incentive on the contractor to come in on time or on cost.
By and large what happens in the real world is they know a new
National Health Service hospital is a precious thing, precious
to the trust and precious to the government and they assume that
we will bail them out. The truth is that is what has happened.
Chelsea & Westminster is a great example of that, we bailed
them out, we are still bailing them out today as a consequence
of that."[133]
However, major construction contracts include substantial
penalty clauses in relation to time overruns and can limit the
passing on of costs, although not to the same extent as is possible
in PFI contracts. In the event of a PFI contractor walking away
from a project the DoH would ultimately have to meet the additional
costs of replacing the contractor or bailing out the project,
as it would with a conventionally procured project. In reaching
this position it is acknowledged that the PFI contractor will
have incurred substantial losses.
81. As the risk is being taken on by the private
sector, it has to be reflected in the VFM comparison between the
PFI and PSC models. This is a complex calculation and as many
people have commented, risk transfer is an art not a science.
Risk valuation is not conducted to a standard procedure, but carried
out on a trust by trust basis.[134]
The Department explained that this was done because the individual
trusts were best placed to understand local circumstances. The
Department did, however, provide trusts with advice on how to
quantify these risks and its PFI guidance outlines 22 typical
construction and development risks. It also provided guidance
on how individual risks could be allocated between the NHS and
the private sector. [135]
82. The Institute for Public Policy Research suggested
that some risks could be clearly assigned but others could not.
In the view of the IPPR, design and construction risks (including
time and cost overruns) and operating costs should be borne by
the private sector partner, while the 'political' risks involved
in changes to health policy should be borne by the public sector.
However, risks associated with the demand levels for the service,
and the obsolescence of technology in the project were far harder
to assign.[136] Appropriate
risk allocation between the public and private sector is a key
requirement to the achievement of value for money on PFI projects.
In its report on VFM in PFI deals the National Audit Office explained
the importance of this:
"Without risk transfer the private sector
receive the benefit of a very secure income stream, similar to
a gilt-edged security, but may set their charges at a level which
earns them a return far higher than is available on such a security.
However, if a Department seeks to transfer a risk which the private
sector cannot manage, then the value for money will reduce as
the private sector seeks to charge a premium for accepting such
risks. [A department therefore should seek to achieve] not the
maximum but rather the optimum transfer of risk, which allocated
individual risks to those best placed to manage them."[137]
The IPPR agreed with this assessment: "Departments
should strive for optimal risk allocation, not maximum risk transfer;
whether this results in the project being off or on the public
sector's balance sheet should be irrelevant to whether the project
goes ahead".[138]
83. Several of our witnesses questioned the validity
of risk transfers. The NHS Confederation believed that while any
assessment of the true value of risk transferred could only be
calculated once a contract had run its course, there remained
a concern that some of the transfer of risk had not been particularly
valid.[139] The Confederation
also drew attention to recent work by the Office for Health Economics
which suggested that the original claim that PFI procurement reduced
the risk of cost overruns was open to question.[140]
Professor Pollock argued that at the point when the contract is
drawn up risk valuation was theoretical rather than real.[141]
As an example of how this could unravel, she cited the Passport
Agency's PFI deal with Siemens. Siemens Business Services were
contracted to develop IT systems for the Passport Agency. Part
of the contract included the transfer of risks of late delivery
or system failure. When failure occurred it was valued at £12.6
million. Yet according to Professor Pollock, only £2.44 million
was being paid in compensation by Siemens.[142]
She further asserted that it was not possible to identify and
cost risks which might arise over the course of a 25-35 year contract
and therefore that it had to be a subjective judgement.[143]
Professor David Mayston from York University also suggested it
was difficult to assess the risks of PFI contracts in their present
form. In order correctly to assess risk, he proposed that PFI
projects should be unbundled into their component parts (that
is, Design, Build, Finance, Maintenance and Operation), with separate
tendering for each. Doing this "would provide the opportunity
to purchase the constituent elements from the most efficient sources,
with a much closer association between risk and reward than PFI
projects at present provide".[144]
This, he maintained, would provide a level of transparency that
could "overcome the suspicion that PFI is driven mainly by
... political factors that are extraneous to the long-term needs
of the NHS".[145]
84. Professor Pollock in a supplementary memorandum
cited the research of Jon Sussex, Associate Director of the Office
of Health Economics, who argued that, in Professor Pollock's words,
" risk transfer is liable to exaggeration in PFI business
cases" and that this "arises because of trusts' perception
that there is no alternative to PFI when public capital is subject
to tight cash limits. Trusts are therefore inclined to treat VFM
as a hurdle they have to surmount rather than as an objective
test".[146]
85. Valuation of 'risk' is the key determinant
of value for money as between the PFI and Public Sector Comparator.
Yet risk valuation is as much of an art as a science. It must,
however, be clearly understood that saying that risk is difficult
to value is not the same as implying that risk is somehow cost-free.
It is not in the interest of the taxpayer to transfer as much
risk as possible to the private sector since risk attracts cost.
What is essential is that an optimal transfer of risk takes place,
with the private sector partner taking only the risks it is best
equipped to manage. Again, more transparency would be beneficial,
so that the partner best able to manage the risk is identified.
86. Once the risk transfer has been assessed and
apportioned, a public sector discount rate is applied to anticipated
future cash flows to allow the 'present cost' of a project to
be assessed. The IPPR explains:
"payments from the public purse for the
capital element of a PFI scheme will be made at a later date than
is the case under conventional procurement. A payment made later
effectively costs less so these future payments have to be discounted."[147]
87. The discount rate chosen for VFM purposes has
for many years been set by the Treasury at 6% and is intended
to represent the pre-tax long-term cost of capital for low risk
purposes in the private sector.[148]
88. The VFM margin between PFI projects and the PSC
is relatively slim, and according to the Department averages out
at 1.7%.[149] Therefore,
if the discount rate is revised downwards by a couple of points
it could make the PFI route the more expensive option - all other
things being equal, a lower net discount rate favours public procurement
over PFI. Professor Pollock gave us an example of how a change
in the percentage rate could dramatically affect the value for
money of a PFI. She contended that the Carlisle PFI scheme showed
a £1.7 million margin in favour of the PFI scheme against
the PSC at a discount rate of 6%, but a margin against of £900,000
at a rate of 5.5% and of £13.6 million at a rate of 3%. The
IPPR also stated that changes in the net discount rate could have
a significant impact on the PSC. It therefore recommended that
all PPP/PFI proposals should be subjected to a "sensitivity
analysis to see whether different assumptions, for example about
different forms of risk allocation or a different discount rate,
would significantly alter the value for money assessment".[150]
89. We explored this point with Mr Nicholas Macpherson,
Managing Director of the Public Services Directorate at the Treasury.
Mr Macpherson explained that the net discount rate was currently
under review as part of the review of the Treasury Green Book.[151]
On the specific point of the influence of the level of the rate
he agreed that if the rate was lowered, in isolation, that would
indeed make conventional procurement more attractive than the
PFI. However, he argued that the net discount rate was in fact
being reviewed in a wider context which would also take into account
issues such as the treatment of risk and uncertainty, both in
terms of time and cost overruns, tax, and possible flow backs
to the Exchequer from private operators. Looking ahead to the
outcome, he said that he certainly would not conclude that the
review would necessarily change the balance between the public
sector comparator and PFI projects.[152]
90. Given the current discount rate was set when
rates were higher, a lower rate may now be more appropriate. We
recognise that other factors need to be considered in the current
review but we would want to be assured that the fact that the
calculations to establish the PSC are so complex is not being
used as an excuse to manipulate the PSC to produce whatever result
is needed. To stop such a view gaining credence we recommend
that the National Audit Office should assess the PSC process as
a matter of urgency in the light of any revision of Treasury accounting
rules. It is essential that the calculations underlying the determination
of the PSC are clear, and that the means by which VFM is established
are transparent and in the public domain.
82