Are Public Private Partnerships Value For Money PDF

Accounting Forum 29 (2005) 345–378
Are Public Private Partnerships value for money?

Evaluating alternative approaches and comparing
academic and practitioner views
Darrin Grimsey 1 , Mervyn K. Lewis ∗
a PricewaterhouseCoopers, 215 Spring Street, Melbourne, Vic. 3000, Australia
b Banking and Finance, School of Commerce, Division of Business, Way Lee Building, University of South
Australia, City West Campus, G.P.O. Box 2471, Adelaide, SA 5001, Australia
Abstract
In an earlier article in this journal (Grimsey, D., & Lewis, M. K. (2002b). Accounting for Public Pri-
vate Partnerships. Accounting Forum, 26(3), 245–270), we examined the intricacies of the accounting
issues raised by Public Private Partnerships (PPPs). It was argued that the critical accounting question
from the public sector’s viewpoint is not one of whether the arrangement is on or off balance sheet,
but whether it represents good value for money. However, determining value for money for a PPP
is an area in which, despite strong criticisms by a number of academic writers of the methods used
by practitioners to evaluate value for money, surprisingly little engagement has taken place between
the practitioners and the academics on the issues involved. This paper attempts to provide such an
engagement. At the same time, because many of the academic critiques focus on the situation in one
country (particularly the UK or Australia), we try to put matters into a broader, comparative context
by considering approaches to value for money tests in a number of countries. Our examination is
thus comparative in the sense of considering value for money tests in different countries, while also
comparing the views of academics and practitioners.
© 2005 Published by Elsevier Ltd.
Keywords: Public Private Partnerships, PPP; Private Finance Initiative, PFI; Value for money; Public Sector
Comparator, PSC; Risk; Uncertainty; Discount rate
∗ Corresponding author. Tel.: +61 8 8302 0536; fax: +61 8 8302 0992.
E-mail addresses: darrin.grimsey@au.pwcglobal.com (D. Grimsey),
mervyn.lewis@unisa.edu.au (M.K. Lewis).
1 Tel.: +61 3 8603 3655; fax: +61 3 8603 4046.
0155-9982/$ – see front matter © 2005 Published by Elsevier Ltd.

doi:10.1016/j.accfor.2005.01.001
346 D. Grimsey, M.K. Lewis / Accounting Forum 29 (2005) 345–378
1. The PPP route
Public Private Partnerships (PPPs) are a refinement of the private financing initiatives for
infrastructure that started in the early 1990s and describe the provision of public assets and
services through the participation of the government, the private sector and the consumers.
There is no single definition of a PPP. Depending on the country concerned, the term can
cover a variety of transactions where the private sector is given the right to operate, for an
extended period, a service traditionally the responsibility of the public sector alone, ranging
from relatively short term management contracts (with little or no capital expenditure),
through concession contracts (which may encompass the design and build of substantial
capital assets along with the provision of a range of services and the financing of the entire
construction and operation), to joint ventures where there is a sharing of ownership between
the public and private sectors. Generally speaking, PPPs fill a space between traditionally
procured government projects and full privatisation.
Although many commentators consider PPPs to be a new version of privatisation (Minow,
2003), in our view PPPs are not privatisation because with privatisation the government
no longer has a direct role in ongoing operations, whereas with a PPP the government
retains ultimate responsibility.2 Nor do PPPs involve simply the one-off engagement of a
private contractor to provide goods or services under a normal commercial arrangement.
Instead, the emphasis is on long-term contracts and strict performance regimes, such as
build-operate-transfer (BOT) or design-build-finance-operate (DBFO) projects to design,
construct, finance, manage and operate infrastructure under a concession, with revenues
(either from government or users) according to services supplied. The private sector partner
is paid for the delivery of the services to specified levels and must provide all the managerial,
financial and technical resources needed to achieve the required standards. Importantly, the
private sector must also bear the risks of achieving the service specification.
There are various reasons as to why governments might undertake PPPs, although
paramount is the objective of achieving improved value for money (VFM), or improved
services for the same amount of money, as the public sector would spend to deliver a simi-
lar project. There is a long history of publicly procured contracts being delayed and turning
out to be more expensive than budgeted. Transferring these risks to the private sector under
a PPP structure and having it bear the cost of design and construction over-runs is one way
in which a PPP can potentially add value for money in a public project.
However, construction risk is not the only aspect of public procurement that needs to be
addressed. There are also risks attached to site use, building standards, operations, revenue,
financial conditions, service performance, obsolescence and residual asset value, amongst
others, to be taken into account when evaluating whether the PPP route to public procurement
constitutes good value for money. In fact, based on experience with Private Finance Initiative
(PFI)3 projects in the UK, there is an acceptance amongst public service project managers
2 Of course, this depends on how privatisation is defined. Often in the United States privatisation is defined as
any shift in the locus of the production of services from public to private, whereas in the UK the term is reserved
for the explicit transfer of public assets to private ownership. For an extended discussion on this point, see Starr
(1988) and Hood (1995), both reproduced in Grimsey and Lewis (2005).
3 Private Finance Initiative is the UK programme encompassing PPP arrangements whereby a consortium of
private sector partners come together to provide an asset-based public sector under contract to a public body. These,
D. Grimsey, M.K. Lewis / Accounting Forum 29 (2005) 345–378 347
that there are six main determinants of value for money (Arthur Andersen, 2000), namely:
risk transfer; the long-term nature of contracts (including whole-of-life cycle costing);
the use of an output specification; competition; performance measurement and incentives;
private sector management skills. Of these, competition and risk are seen to be the most
important.
It follows that what would seem to be required to achieve value for money, defined as
‘the optimum combination of whole life cost and quality (or fitness for purpose) to meet
the user’s requirement’,4 is that
• projects be awarded in a competitive environment;
• economic appraisal techniques, including proper appreciation of risk, be rigorously
applied, and that risk is allocated between the public and private sectors so that the
expected value for money is maximized;
• comparisons between publicly and privately financed options be fair, realistic and com-
prehensive.
These considerations are normally examined on a case-by-case basis (although one study
does focus on the overall rates of return across a large number of PFI projects),5 on the
grounds that risk allocation depends on each project’s risk profile, while the competitiveness
of the market for bids will vary from project to project and from one time to another.
However, while competition and risk allocation are the important pre-conditions, with value
for money enhanced by transferring an appropriate degree of risk to the private entity,
they do not guarantee value for money. Rather, the possibility of achieving extra value for
money by implementing a PPP can be estimated (under the approach in the UK and some
other countries) with a two-fold analysis conducted prior to the PPP implementation. It
comprises, first, the calculation of the benchmark cost of providing the specified service
under traditional procurement and, second, a comparison of this benchmark cost with the
cost of providing the specified service under a PPP scheme. This benchmark is known as
the public sector comparator (PSC).
2. Approaches to value for money
This section examines some different approaches to value for money testing applied in
different countries. To our knowledge, there is no comprehensive study of the PPP market
at a global level. In order to give some perspective on the topic, Table 1 provides details of
PPP activity in 29 countries for which information is available, drawing on practitioner and
other sources.
Broadly speaking, four main alternative approaches can be discerned. Ordered from the
most to the least complex, these are: first, a full cost-benefit analysis of the most likely public
along with the introduction of private sector ownership into state-owned enterprises and the sale of government
services into wider markets, constitute the current Partnership UK agenda.
4 Office of Government Commerce (2002, p. 6).
5 The study was undertaken by Pricewaterhouse Coopers (2002), commissioned by the Office of Government
Commerce, and is discussed later in this paper.

Table 1
Public Private Partnerships in various countries
Country Experience with PPPs
Argentina Extensive use has been made of the concession model for water, ports and railways. A
concession contract was awarded in 1993 to Aquas Argentinas (led by Lyonnaise des Eaux
and Dumez) to operate the greater Buenos Aires water supply and sanitation system. As
compared with the state-owned authority, labour efficiency has increased, water quality
improved, unaccounted-for-losses and illegal connections reduced and water rates cut by 17
percent. The six main terminals at the port of Buenos were awarded to concessionaries. Like
some other Latin American countries (Brazil, Bolivia, Chile) the concession model was
used for railways, with five lines totalling over 22,000 km let to concessionaries. Value for
money was sought by competitive bidding to supply rail services on government-owned
assets. The Argentinian system was the first to require freight concessionaries to share
tracks with passenger services, and also the first to use negative bids to subsidize suburban
passenger services.
Austria PPPs in procurement in health and transport sectors, including the EIB-financed Grazer
Frasht Terminal (PPP) A&B. PSC not used, but a PPP unit has been established to promote
PPP projects.
Australia PPP projects are developed at the federal and state level, with projects underway in all
states, plus Northern Territory and Commonwealth. Victoria is the dominant market,
followed by NSW. A full cost-benefit analysis is undertaken prior to a decision to commit to
a major infrastructure project. A PPP’s value for money (VFM) is assessed against a PSC
and a public interest test. Differences exist between the states in terms of risk adjustment
and discount rate treatment.
Belgium PPP market is not developed. Airport, social housing and road projects are in procurement.
The principal public service that has been structured with a real PPP approach to date is the
Brussels North Waste Water Treatment. The project has been procured as a BOOT with no
preparation or consideration for VFM measurement or PSC.
Bulgaria The 15 year Sofia Water and Wastewater Concession Project, which reached financial close
in October 2000, is the major municipal infrastructure concession in Bulgaria and one of the
first water concessions to be financed on a limited recourse basis in Eastern Europe via a
special purpose vehicle.
Canada PPPs include roads, bridges, airports, seaports and harbours, energy, hospital facilities,
waste water facilities, social housing and schools. British Columbia has recently established
Partnerships BC (along lines of Partnerships UK and Ontario’s Superbuild) and is
developing a PSC approach.
Croatia The government’s policy is favourable to the use of BOT schemes for transport (Istrian toll
road), energy (Lukovo Sugarje power project) and water (wastewater treatment plant for
Zagreb). New legislation is designed to facilitate concessions.
Czech Republic Joint ventures have taken place between public institutions and private entities in the energy
sector, telecommunications and water and waste water treatment, mainly as a result of
privatisation. Toll roads have been rejected with two BOT projects not realized. A task force
was created in 2000 to develop PPPs, in order to complete the road network.
Finland The Helsinki–Lahti motorway, conceived in 1995 and begun in 1997 is the first and largest
PPP in Finland, involving equity from the UK, Sweden and local entities. A pilot PPP
project was to build a sixth form college specialising in IT, but market remains undeveloped.
France France has a long-established tradition of public–private cooperation (especially in sectors
such as water) using the concession structure. PPPs are not permitted in social infrastructure
area. The tunnel Prado-Carrenage in Marseille was toll-financed. Three major road projects
have been launched under PPPs since 2000 (Millau Viaduct, A19 and A28) and
cross-border projects such as the Perpignon–Figueras high speed link and the Lyon–Turin
high speed link have involved recourse to PPPs. Because private sector involvement has
been through concessions, those awarded have been evaluated on bids and a separate value
for money assessment has not been included.
Table 1 (Continued )
Germany Germany has no formal PPP programme, but four projects are EIB funded PPPs (Elbe Tunnel,
Engelbergbasistunnel PPP, Warnowquerung Rostock and Weser Tunnel), some of which did involve
risk transfer to the private sector under a concession framework. A BOT law has been passed. A
PSC has not been used in Germany, but it is a legal requirement that adequate “economic feasibility
studies” are prepared to support public investment and private companies may be required to
demonstrate if and how far public duties can be executed by private parties (to equivalent standard
and for comparable/lower) costs.
Greece PPP projects include Athens International Airport, New Athens International Airport, Essi
Motorway, Rion-Antirion Bridge and Olympic Stadium. The government launched a PPP
programme in 2000, as well as setting up a central PPP Unit.
Hungary An early project was the nationwide digital cellular network. Some transport projects have been
developed by PPPs (e.g., M5 BOT project), but others have not been realized or transferred to
National Highway Agency (e.g., M1). The Szechenyi Plan sought to expand PPPs, with projects in
schools and water treatment, and others in health, prisons and sports facilities in procurement.
Hong Kong Hong Kong has been a leader with PPPs involving BOT contracts for transport beginning with the
first Cross Harbour Tunnel in 1972. Since then three further tunnels (Cross Harbour Tunnel,
Eastern Harbour Crossing and Western Harbour Crossing) and a tolled road with bored tunnel have
been undertaken using the BOT method. Other PPP projects include two container terminals,
exhibition centre, refuse collection, transfer and disposal schemes and water treatment. The
government has taken considerable care to ensure that BOT projects are economically viable,
commissioning their own independent economic evaluations of proposed project, with traditional
procurement preferred if the project is not ‘bankable’ (e.g., Tsing Ma Bridge). More recently, PSCs
have been used in project evaluation, and have gained acceptance as a mandatory component of the
procurement process.
India Extensive use of BOO projects for power projects (e.g., Dabhol Power Project). Other sectors with
projects include oil and gas, construction infrastructure, hospitals and water sanitation. India has
not used a PSC tool systematically, but financial advisors to government bodies procurring a
concession are usually required to develop cost estimates and model the bid. If the bids come in at
levels substantially different from estimates, much analytical work is called for to justify action on
the bid, in particular when bids require the government to pay more than estimated.
Ireland The M4 PPP DBFO Toll Motorway project (from Kinnegad to Kilcock) is the first Irish road PPP
and is part of a group of 11 proposals (the Roads PPP Programme). A light rail system was initiated
and toll bridges, government offices, schools and colleges, and prisons have been designed, built,
financed and operated by the private sector. There is a strong commitment to a formal PPP
programme. A clear legislative framework is in place, a dedicated PPP unit has been set up and
central committees facilitate PPPs. With the M4 PPP project, the Irish National Roads Authority
worked with the EIB, international banks and the project consortium to identify an acceptable
allocation of risks and a satisfactory deal, but a PSC was not developed.
Italy The Merloni Bill in 1994 and 1998 set the framework for using private sector contractors and later a
special PPP taskforce, UFP, was created and its powers reinforced in 2001. There have been
projects in the water and power sectors in particular which involve the private sector, on a
concession-style basis. Other projects have been in roads, light railway and health services.
However, there are administrative complexities associated with the civil code.
Japan PPPs in Japan date back to the ‘third sector’ approach introduced in the mid-1980s, bringing
together the public (the ‘first sector’) and the private sector (the ‘second sector’) to form
project-based companies, engaged in urban developments, leisure/resort developments, transport,
telecommunications and other regional activities. Landlord ports are operated with berths leased to
private ocean carriers to manage. With projects under government guidelines, the public sector
must construct a PSC and demonstrate that the PPP option provides better VFM, disclosing the
extent of VFM, either in the form of percent or absolute amount. There is no specific requirement
under the guidelines to verify VFM when bids are submitted and/or contract is finalized. This is
because there is a separate procurement rule where bids become ineligible if the price exceeds the
government-planned price (i.e., budgeted price), which is normally set based on the PSC.
Netherlands Kennis-centrum PPPs was set up in 1999 and a major pilot project HSL Zuid PPP (the high
speed rail) under the Trans-European Network (TEN) was started. Projects underway since
then include road, railway harbours and water projects, i.e., Zuiderzeelijn/Randstad Circle
Line (magnetic levitation technique), 2nd Maasvlakte (enlargement harbour Rotterdam) and
the Delfland Waste Water Treatment Project (involving EIB participation). The Delfland
project involved the preparation of a PSC, and the final project NPV indicated a saving of
around 15 percent vis-à-vis the PSC, and in excess of the target of 10 percent for the project.
New Zealand Little PPP activity to date. Although a number of policy papers have been prepared, projects
have been confined to prison management contract services and consessions or franchises
for the operation of water and wastewater infrasructure facilities.
Philippines The Philippines BOT law is regarded as best practice among developing Asia Pacific
countries, and includes a range of contractual structures (e.g., BOO, BTO, BLT, etc.) in
addition to BOT. It is supported by the BOT Centre to encourage investors, co-ordinate and
monitor projects and assist with bidding. Over 80 projects are in various stages of
development, implementation or operation. Two 25 year concessions, developed with World
Bank assistance, for vertically integrated water supply and sanitation for east and west
Manila, are also regarded as ‘best practice’ models for the concession approach. Value for
money was evidenced by the winning bidders providing concessions at an average water
supply price 57 and 27 percent below pre-concession tariffs, while meeting compliance
standards and improving and expanding the system.
Poland The A4 Katowice–Krakow is the first toll highway in Poland. The government is anxious to
facilitate PPPs and two bridges, a port and water treatment facilities identified as PPP
projects. The legal, accounting and taxation system hinder the implementation of PPPs, and
legislative changes are underway.
Portugal A large number of transport PPP projects have been developed, including five motorways
under the SCUT programme and four others, all with EIB involvement, to develop the
Portugese transport infrastructure. Other projects include water and sewerage, subways,
local transportation (e.g., light railway) and museums.
Romania Concession-based financing techniques are favoured. In 2000, the French utility company
Vivendi was awarded a 25 year concession to provide water and pipeline rehabilitation
services to Bucharest, in the form of a new treatment system and modernising the existing
water system. Commercialisation of road maintenance, health services and sporting
facilities under procurement, with PPPs being promoted for infrastructure development as
part of Romania’s preparations for EU accession.
South Africa A number of projects developed including the defence sector, hospitals and an EIB funded
transport PPP project. PPP regulations require formal National Treasury authorisation of any
national and provincial PPP in three steps. Step one authorisation is based on a completed
option analysis/feasibility study. There are three evaluation criteria—affordability (budget
impact), value for money and transfer of appropriate technical, operational and financial
risk. A key required element is therefore a PSC, as a mandatory tool.
Slovenia An EBRD assisted project is investigating private investment in the maintenance of the
national road network. Development of a private finance concession-based highway
maintenance scheme is a planned pilot for PPPs.
Spain Fourteen EIB funded PPP projects in transport, and one in water (Aguas de Sevilla, PPP).
However, the legal framework is not supportive, there is no law to cover concessions and
few other projects are underway.
United Kingdom The British Government launched its PPP development policy in 1992 under the ‘Private
Finance Initiative’. Since then, the technique has been applied systematically to virtually
every area of significant government capital spending in the UK. Partnership UK was
established in 2000 to promote PPP/PFI concepts. It also works on local authority projects.
Calculation of a PSC is required for value for money evaluation.
United States PPPs operate for policy-level and project-level partnerships. Examples of policy-level
partnerships are the Federal highway assistance program, urban mass transit assistance
program and initiatives under the Intermodel Surface Transportation Act (ISTEA).
Project-level partnerships focus on specific projects or situations. Examples are the Dulles
Toll Road Extension in Virginia under a BOO concession, and State Routes 91 and 125 in
California under BTO schemes. A number of cities have formed partnerships to operate
and maintain wastewater treatment systems. Examples are: the Milwaukee Metropolitan
Sewerage District, City of Indianapolis, Wastewater Treatment Plants and Collection
System, Buffalo Water Treatment System including customer billing and collection
services. Partnerships have also operated for technology, health, prisons, welfare and
urban regeneration. However, the market is fragmented and very diverse, with initiatives
at the federal, state, local and municipal level. There are no established procedures for risk
transfer, risk valuation, and establishing value for money.
Sources: Grimsey and Lewis (2004c), PricewaterhouseCoopers (2004), European Investment Bank (2004).
and private sector alternatives; second, a PSC–PPP comparison before bids are invited; third,
a UK-style PSC–PPP VFM test after bids; fourth, reliance on a competitive bidding process
to determine VFM once PPP ‘road-testing’ has been established. Since the third of these,
the UK-style PPP–PSC comparison is the one that has been subject to academic critique, it
is this with which we commence.
Decisions about options to follow in traditional public procurement are usually based on
cost-benefit analysis in which there is assessed the full range of economic costs, risks and
benefits, taking account of their timing, with these discounted to obtain an NPV. Other, less
quantifiable, impacts are also brought into consideration in the calculation. Normally, the
initial cost-benefit analysis does not look at the different ways of procuring a given project
but assumes the most likely commercial approach. In most cases, this means provision by
the public sector. Once a procurement approach is decided in detail, the public sector then
sets in motion a competition between bidders to ensure that value for money is achieved.
In effect, the evaluation of competing bids, where price and non-price factors are assessed,
equates to a value for money test.
In a situation where a different procurement method, such as a PPP, is under consideration,
there is a need to establish that these alternative commercial arrangements deliver good
value for money. While a competitive market between bidders will ensure that, for any given
commercial deal, the best value options are selected, the choice of the particular commercial
arrangement must be tested in some way to ensure that it is capable of delivering value for
money.
In the UK, this need arose when PPPs were made feasible after 1989 by the relaxation
of the ‘Ryrie Rules’ (NEDC, 1981) and the subsequent launching of the Private Finance
Initiative in 1992. Engaging the private sector in funding major infrastructure projects
implied higher explicit financing costs because the return required by lending and investing
institutions exceeded the cost of issuing public sector debt. Experience on early projects had
suggested, however, that this cost of capital could be more than balanced by the efficiencies
delivered by the private sector in both cost and risk management.
It was, therefore, ruled that the public sector could enter into PFI transactions and, through
them, borrow from the private sector, provided that in doing so the overall NPV of cost was
lower than if the project were conventionally procured and financed by government funds.
Value for money in this context can be thought of as the best price for a given quantity and
standard of output, measured in terms of relative financial benefit. It is assessed through a
comparative analysis of the costs of different solutions that generate the same outputs in the
form of a comparison between bidders’ cash flows and those of the public sector alternative.
Such a test is undertaken prior to final approval of the deal, by preparing a hypothetical
set of costs for the public procurement of a project delivering the same output, including
a full evaluation of the project risks borne by the public sector. This hypothetical costing
is compared with the actual cash flows to be paid to the private sector provider, plus the
value of any residual costs and risks not transferred to the PFI and therefore retained by the
public sector. This comparison constitutes the PSC, which largely considers the financial
differences between two procurement options for the same project.
Calculation of the PSC is necessarily theoretical since any attempt at a ‘live testing’ of
procurement options, through running separate (or multi-tracked) competitions under two
different procurement methods, would unlikely be well received by bidders. Thus, the PSC
is based on estimates of full costs, revenues and risks, set out in cash flow terms, discounted
at a public sector rate to an NPV. It is compared with the discounted value of payments
under the PSC along with the adjustment for risks and costs retained. Often this comparison
is of the two single-point NPVs, but the use of ranges is more robust as explained below.
Construction and application of a PSC is an integral component of all PPPs in the UK
and Australia, and many jurisdictions active in the PPP market have adopted this concept.
However, this approach is by no means universal, and many countries where PPPs are
used either do not have established procedures or do not use a PSC-based approach when
assessing value for money.
For example, in the United States, PPPs have covered transport, private prisons, detention
centres, technology, water treatment, welfare provision, health and medical services, and
a range of community activities from schooling to urban regeneration and environmental
policy. Most of these operate at the state level and the market is a fragmented one with
few conventions (Rosenau, 2000). In some cases, value for money is sought through the
tendering process. Most of the contracts by state governments in the United States for
private prisons require that private firms offer the service at 5–10 percent below what it
would have cost the state, and the available evidence would seem to suggest that there is a
cost advantage (slight) of private prisons over state facilities (Schneider, 1999).
Another country that has taken a different route to that in the UK is France, where there is
a long history of PPPs, dating back to the 17th century, in the form of concession contracts
which are now an established way of involving private sector resources and management
of municipal services, such as water supply, sanitation, urban heating, waste management,
urban transportation and sporting facilities. Two PPP operators, Lyonnaise des Eaux and
Vivendi, controlled 62 percent of water distribution, 36 percent of sewerage disposal, 75
percent of urban central heating, 60 percent of refuse treatment, 55 percent of cable operation
and 36 percent of refuse collection in 1995, all under PPPs (Ribault, 2001). In most countries,
these figures would be remarkable. They are not in France because PPPs have been around
for a long time, and there is a track record of performance. Indeed, it is France that can claim
to be the pioneer of PPPs and the inventor, in the form of the concession approach, of the
one PPP model that can be said to have stood the test of time. Because concessions are an
established contractual form with a long history, those awarded have not included a VFM
assessment involving a PSC. The concession model has been emulated in many countries
in Latin America, Eastern Europe and the Francophone countries of Africa.6
Turning now to some of the countries that do use explicit value for money tests, in
Australia all projects in Victoria, NSW and Queensland are using a fully developed UK-
type PSC approach, and there is general consensus that this is a sound basis upon which to
proceed, despite some differences in risk adjustment and discount rate treatment. In Japan,
under government guidelines, the public sector must construct a PSC and demonstrate that
the PPP option provides better VFM, before adopting a PPP project, and is required to
disclose the extent of VFM, either in the form of percent or absolute amount. South Africa
also uses a PSC calculation as a mandatory tool.
In some other locations, the situation may be similar but not so formalized. For example,
a recent PPP in Hong Kong utilized a PSC approach drawing on the UK/Australian toolkit.
India has not used a PSC tool systematically, but financial advisors to government bodies
procuring a concession are usually required to develop cost estimates and model the bid.
PPPs (or P3s) in Canada have involved the federal government, provinces, cities and towns,
and there are no common VFM conventions, However, the province of British Columbia is
developing a framework for VFM assessment involving the use of PSCs to aid in procure-
ment evaluation. The same is true of Ireland, which in other respects has generally followed
the UK model.
Nevertheless, the PSC method of focusing on the financial impact of the procurement
method is not the only approach being followed. An alternative way of testing VFM is to
undertake a full economic analysis of a feasible public sector option and a real PPP bid. This
course is taken in some countries, for example, Germany. There it is a legal requirement that
adequate economic feasibility studies be prepared to support public investment, and private
companies may be required to demonstrate if and how far public duties can be executed by
private parties to an equivalent standard and for comparable or lower costs. The problem
with this approach is that it entails a greater amount of work, a much greater degree of
subjectivity in arriving at the assumptions needed to evaluate economic costs and benefits,
and hence more ambiguity as to which option will deliver VFM. The financial PSC is easier
to compile and, for all that it is hypothetical, perhaps more transparent.
There is another approach currently being employed to evaluate VFM for PPPs. This
is to check the VFM before bids are invited, using a hypothetical PSC and a ‘shadow’
PPP. For example, in Japan an early indication of VFM is a pre-requisite for approval of a
PPP to proceed, and once VFM has been shown theoretically possible, then the procure-
ment can go ahead. The original assumption of VFM may be subsequently rechecked with
a PSC. Such an approach is employed in The Netherlands. The second PSC test would
appear to be worthwhile because initial estimates of bidders’ prices will often diverge
widely from outturn, particularly where there are novel commercial arrangements and risks
transferred.
6 The worldwide PPP market is examined in Grimsey and Lewis (2004c, 2005).
3. Calculation of the PSC
From the preceding survey, it is apparent that, although it is certainly not the only way of
evaluating value for money, some jurisdictions active in the PPP market have adopted the
concept of a PSC in some form to provide the core test as to whether a PPP achieves a lower
overall NPV of cost (or better value for money) than it would if the project were traditionally
procured and publicly financed. Fashioning the PSC performs the following roles:
• It promotes full costing at an early stage in project development.
• It provides a key management tool during the procurement process by focusing attention
on the output specification, risk allocation and comprehensive costing.
• It provides a means for testing value for money.
• It provides a consistent benchmark and evaluation tool.
• It encourages competition by generating confidence in the market that financial rigour
and probity principles are being applied.
In UK and Australian practice, the PSC is constructed and refined during the feasibility
and business case stages of a project prior to release of the tender documents, and thus before
bids are received. There are two reasons for this. First, it is essential that the PSC is a ‘pure’
public sector option, not influenced or potentially influenced by ideas coming forward from
PPP bidders. Second, its wider value to the procurement process will be enhanced if it is
prepared early as it will help with detailed project definition and in anticipating what a
private sector bid will need to deliver in order to improve VFM compared to the PSC.
The PSC will develop from the preferred option selected after analysis of its costs and
benefits. It will usually involve a higher degree of refinement of the numbers than the option
analysis. Elements of the option analysis, however, may be left unrefined; an example would
be the assessment of benefits, which are less likely to be a differentiator between the PSC
and bids.
As the procurement progresses, the PSC should be kept up to date so that it remains
a valid comparison in terms of the scope of project and the required service performance
standards. Comparison with the bids is usually made at more than one point: initially when
bids are received, in order to have an initial test of potential for VFM; then prior to selection
of preferred bidder, using revised bid figures; finally, prior to closing the deal, as a pre-
requisite for approval to do the deal.
As a modelling of the cost of procuring a project through traditional public procurement,
the PSC should take into account all the assets, services, staff, consumables and other
elements required to deliver the project to the same standards and level of certainty required
of the PPP partner. It should also allow for the risks that the public sector encounters in
procurement as well as a realistic assessment of how the public sector might deliver future
efficiencies.
In order to be a valid comparative model of traditional procurement, the PSC calculation
must use the same assumptions as the PPP in respect of the following elements:
• Timing—The PSC assumes the same commencement date and project term as the PPP.
It ignores the fact that, to opt for the public option after pursuing a PPP bid, the timing
of the project would have to be deferred.
• Funding—The PSC assumes the capital funds are available for the up-front investment
required to deliver the same output specification as the PPP.
• Procurement costs—Only the costs associated with implementing the reference project
should be included in the PSC. The costs of running the overall PPP process need to be
considered but should not be included in the PSC. These costs and any associated risks
should be added to the NPV of the PPP bids. As a general rule these costs should only
be taken into account at the level at which they would occur in a mature market, i.e., any
development costs are sunk and in the past.
• Output specification and performance standards—The reference project and PSC must
be developed to achieve the same standards as under the PPP irrespective of whether past
experience indicates that these standards have not been maintained by the public sector.
There are some differences from country to country in these elements. In particular,
procurement costs and delays caused by switching to a new procurement route could rea-
sonably be included. In Australia the differential costs of procurement (that are not already
sunk costs) may be brought into consideration in comparing the two options, although in
reality many of the procurement costs will have been sunk by the time of the comparator
and the costs included are more likely to refer to the re-let process. Where an estimate of
VFM is carried out before the bid process, as in The Netherlands, procurement costs may
also be included.
Timing and procurement costs may all be included if a full cost-benefit analysis of the
two procurement options is to be carried out as in Germany, where a full economic appraisal
of the private sector’s ability to improve on the public sector provision is required. In these
cases differential benefits can be evaluated, rather than assuming a constant output between
public provision and PPP.
The absence of capital funds for a PSC project is a factor that drives many countries
away even from producing a hypothetical VFM test. If procurement by the public sector
is unaffordable in any way, there is seen to be no need for this test. In other words, the
assumption of available public capital is too unrealistic to make a PSC credible—a point of
some contention in the UK.
Finally, although developed under the equivalent output and performance specifica-
tions, the PSC must be based on the most likely public procurement method. This
is assumed to incorporate some efficiency in procurement. For example, a design
and build contract might reasonably be assumed and efficiencies might be achieved
through contract writing, good management practice or competition from external ser-
vice providers. Nevertheless, such efficiencies have to be feasible and untried methods
avoided.
In both the UK and Australia, PSCs are generally categorized into four core elements:
raw PSC (base costs), transferable risk, retained risk and competitive neutrality.
3.1. Raw PSC (base costs)
The raw PSC should provide a base costing including capital and operating costs,
and represent a full and fair estimate of all of the costs of delivering publicly the same
volume and level of performance, service and residual asset value that is required from
the private sector under the PPP alternative. Those aspects of the contract that could not
feasibly be provided in-house by the public sector should be priced under the commer-
cial terms that would be normally expected to apply to their procurement. The costs will
include:
• preliminary set-up and planning work, approvals and permits;

• costs of achieving any efficiencies or innovation;
• design and capital procurement, including replacement;
• opportunity cost of using existing assets;
• management and facility overheads;
• operating costs, including maintenance, consumables, contracted services, staff, utilities,
etc.;
• any decommissioning costs at the end of the project.
Wider socio-economic costs may be excluded, on the grounds that they are not likely to
differ greatly between the public and private provision. These should be raw costs without
any contingencies included—these will be covered in the evaluation of risk.
3.2. Transferable risk
The optimal allocation of risk is the key objective of all PPPs and the value of transferable
risk needs to be included in the PSC. As in any complex procurement, a detailed risk
register is required, analysing risks in terms of likely impact and probability of occurrence,
drawing on specialist technical input and analysis wherever necessary. Once this register
is completed, the allocation of risk between the public sector and the bidder needs to be
ascertained. For the PSC, it is advisable to include the full risk adjustment as this helps
to give a more accurate picture of the cost of the project. For the PPP comparison, it will
be necessary to deduct from these risks the risk transferred to the private sector, with only
retained risk added to the PPP cost. The transferred risk is often a key determinant of VFM
in PPPs, and one that may need to be updated as negotiations proceed, to allow for variations
in risk allocation.
Determining the value of risks using historic and current data as benchmarks is highly
dependent on the availability of such data and countries vary as to the extent to which
relevant data have been collated and made available. This factor may explain some
differences between the UK and Australia on transferred risks, despite a very similar
methodology being employed. Practitioners in Australia report that the average value of
transferred risks on PPP projects across Australia appears to be around 8 percent. The
UK seems to have applied greater premiums to the transferred risks in its PFI projects.
Transferred risk as a percentage for total PSC in the UK appears to be within the range
of 10–15 percent and averages around 12 percent. This discrepancy may be due to the
UK’s experiences with cost overruns (as evidenced by the recent UK studies into opti-
mism bias) and could therefore reflect a better informed view of what values should be
attributable to transferred risk. Equally, it could indicate that Australia has experienced
less in the way of cost blowouts under traditional procurement than is the case in the
UK.
3.3. Retained risk
Any risk that is not transferred under the PPP contract is a retained risk. The value for
retained risk needs to be included in the PSC using the same methodology as applied
for transferred risks. Any asset residual value should be credited to the project at the
end particularly if residual value risk is transferred to the private sector (as is the case
in some projects in Australia, e.g., Victoria’s County Court). If, however, this would
be the same under the PPP, which is typically the case if the residual value risk is
retained, the figures could be omitted from the VFM comparison and consequently the
PSC.
3.4. Competitive neutrality
Competitive neutrality adjustments remove any net competitive advantages that accrue
to a government business by virtue of its public ownership. In some countries a tax adjust-
ment is made to reflect the fact that some of the cost of the PPP represents tax revenue,
which returns to the government. Of course, to the extent that the PSC includes pur-
chases from the private sector, that will also result in a tax flow back to government
too. Rather than credit both PPP and PSC with the full tax adjustment, which would in
any case be extremely hard to estimate, the approach taken (where there is one) has been
to identify the additional tax flows that arise because of the PPP structure. In Australia
where these issues are covered by the Competition Principles Agreement (see English &
Guthrie, 2003), these factors have been identified as land, local government, payroll and
capital transaction taxes. In the UK, the company tax paid by the PPP corporate vehicle
has been identified as an add-back for the PPP (or cost to the PSC). Competitive dis-
advantages also arise and need to be factored into the PSC adjustment for competitive
neutrality. These include heightened security and reporting requirements not faced by private
enterprise.
3.5. The VFM comparison
Once the NPVs of both PSC and PPP have been prepared and adjusted to a comparable
basis, then a simple comparison of the two can be carried out. Fig. 1 illustrates a possible
value for money comparison between a PSC and a PPP bid. Assuming all things equal
(i.e., quality and risk allocation), value for money is demonstrated when the total present
value cost of private sector supply is less than the net present value of the base cost of the
service, adjusted for the cost of risks to be retained by the government, cost adjustments for
transferable risk, and competitive neutrality effects.
4. Some practical considerations
This section examines some practical matters associated with the PPP–PSC value for
money tests, before going on to consider some criticisms of the VFM methodology.
Fig. 1. PSC and value for money.
4.1. Is the PSC too subjective?
The PSC provides a benchmark against which to consider whether the general level of
bids represents value for money to the public sector and whether a change in the commercial
terms generally improves VFM. The level of detail in the analysis should be sufficient to
enable a meaningful comparison between the PSC and bids and to understand the differ-
ences between them. Numbers for the PSC must be realistic. However, arriving at realistic
estimates can be a costly and complex task, not dissimilar to the preparation of business
plans.
While realism is sought, the PSC calculation remains a hypothetical one, and the amounts
are not ever put through any real market test. It is, therefore, arguable that the PSC is too
subjective and can be manipulated to show whatever its users require it to show. However,
if the PSC forms the basis of the project budget, then it will not be in the public sector
interest to understate the amounts involved. Nevertheless, it remains necessary for the PSC
to be seen to be based on rigorous and objective estimates of costs and benefits in the public
sector procurement process.
In order to estimate the costs of public sector procurement in any detail there will be a
need for a clear statement of the project requirements so that the scope of work is known
(the output specification). Preparation of a PSC based on such a specification ensures that
there is a set of common reference or baseline financial data to underpin the projections and
provide the reference point for assumptions about costs and revenues. In countries where a
PPP model is prepared as part of the planning process, the same applies.
4.2. Is the PSC needed?
Given that the PSC is certainly not costless, it might be argued that the checking of a
procurement option is needed only where that method is new and untested. Drawing on the
approach followed in many countries, procurement could be assumed as the base case for
projects where PPP is established as the method of delivery, and there have been sufficient
tests of VFM, along with estimates of the cost structure, to make it indisputable that this is the
preferred commercial arrangement. This is effectively what happens in France and other
countries where the French concession approach is a well-established tool for involving
private operators in the provision of infrastructure services, especially in the area of water
and sanitation, and there is a track record of cost effective performance standards reached.
Provided there is a competition, VFM should be ensured by competitive tension between
bidders. The same is true of BOT projects for roads, tunnels, bridges, etc., especially in
developing countries where the concessionaires are responsible for marshalling the capital
resources for the project, along with the specialized design, construction and operating
skills. Again, once the procurement route is well established, competition in the bidding
process is relied upon to ensure VFM.
However, there are some offsetting factors to be considered, most notably if public
accountability is an important factor. Then, the PSC forms a vital component of the decision-
making audit trail. Also, on purely financial grounds, the PSC can provide a negotiating tool
to the public sector in ensuring the best possible deal is obtained from the private sector.
Often the knowledge that a PSC has been prepared and that bids will be measured against
it can bring real competitive tension to a procurement process. Similar arguments apply
to the preparation of a PPP model. A PSC or PPP model should not replace competition
between bidders wherever it can be achieved. But selective disclosure of the PSC (at least
at a summary level) can be used in the later stages of a procurement to ensure the best
possible terms—especially once a preferred bidder has been selected, or in highly technical
procurements where there are few acceptable potential suppliers.
4.3. Is the PSC too simplistic?
In many countries that have not yet moved to a PSC-type methodology for bid evaluation,
cost-benefit analysis underpins the decision to proceed with a project as there is a clear need
to demonstrate that an assessment has been made of the benefits relative to the costs of a
project before proceeding. This type of analysis enables different options to be considered
in full, rather than the more limited cost comparison used in the PSC. Germany uses this
approach, although the government is considering the desirability of adopting a PSC tool
as well.
A cost-benefit analysis can be a powerful tool in establishing the priorities of projects to

take forward, but it comes at a cost in that it can be more complicated and less accurate than
an approach that focuses primarily on financial costs, because it requires a greater number
of assumptions, for example, about the accrual and valuation of benefits. Effectively, the
PSC approach is a halfway house between a full cost-benefit study of all feasible options
on the one hand, and simply allowing a bidding process to ensure VFM on the other. Some
of the broader issues that might be revealed in a cost-benefit analysis could, nevertheless,
be picked up in the assessment of qualitative factors.
4.4. Unquantifiable elements
The VFM comparison is primarily a financial one between the NPVs of the cash flows
of two options. Normally, PPP guidance material also directs that the decision as to VFM
should also incorporate qualitative factors. Qualitative factors by definition are not costed
in the PSC as they are not accurately quantifiable. These factors may include the following:
• Material costs (including risk) that are not capable of being quantified.
• The identity, credit standing and proven reputation of the bidder.
• Differences in the service provision between bids that cannot be quantified.
• Any wider benefits or costs that may flow, for example social benefits.
• The accuracy and the comprehensiveness of the information used and the assumptions
made in the PSC.
Nevertheless, it must be said that while most governments talk about qualitative factors
and broader value for money considerations, in reality the quantitative comparison usually
takes precedence in the evaluation.
4.5. The PSC is more risky
The PSC usually remains a statement of initial estimates, while PPP bids contain numbers
to which bidders must commit contractually. Despite the rhetoric, the exercise of assessing
the PPP relative to the PSC calculation is not strictly one that compares like with like.
The PSC calculation is a theoretical one of what the project might cost using the public
procurement route, with many unknowns attached to the costings, particularly with respect
to the evaluation and treatment of risks and the likelihood of cost overruns resulting. By
comparison, the PPP figure is actually a firm bid, put on the table by a consortium willing,
able and ready to do the job and with a desire to start delivering the infrastructure services
as quickly as possible.
There may, therefore, be a greater degree of risk priced into the PPP, which is not matched
by the hypothetical risk adjustments in the PSC. The PSC is to that extent subject to a greater
amount of ‘unprovided’ risk, which is not accounted for and understates its NPV. This can
be allowed for some extent by risk analysis and the use of benchmarks as to historic levels of
risk in publicly procured projects, although not all countries have built up a comprehensive
data set.
4.6. The PSC is incomplete
Most PPP contracts in the UK and Australia are for at least 20–30 years (and some much
longer), and consequently they all have some time to run. Value for money is projected at the
point of the PSC comparison, and actual outturn costs may differ from those projected by the
PSC. Where these fall well below the level of projections, then VFM will not be maintained.
Over a long period the impact of this factor on the public sector could be significant. This
problem can be addressed through provisions in the contract to monitor on-going value for
money and re-price where market prices have fallen below those predicted at contract close.
Contract provisions to achieve this include: benchmarking of aspects of the costs through
periodic review of performance against peers and re-pricing to adjust; re-competition of
underlying subcontracts and pass though of price reductions; the ability to refinance and
share cost reduction gains arising from the refinancing; and open book accounting with
full disclosure of actual outturn compared to projections by the supplier, and a right of
(selective) audit for the public sector customer—with a sharing of profit levels above an
agreed ceiling.7
4.7. The comparative advantage rule
While there is no simple pass or fail with the PSC–PPP examination, there is the expec-
tation that the private sector bid must deliver better value for money if it is to proceed.
The essence of this VFM test would seem to be what Officer (2003) calls the principle of
comparative advantage in service provision. On this basis, responsibilities would be allo-
cated between the public sector and private sector according to which sector can add the
greatest value to the community; government should provide services when it is better than
the market at doing so, but leave it to the market if that is not the case.
This pragmatic principle is unacceptable to those on both sides of the political spectrum.
To economic liberals, the presumption that PPPs must demonstrate value for money over
the PSC in order to proceed is the wrong-way round. They would point to the Adam
Smith principle, that government should do only what cannot be done in the market. To do
otherwise and go beyond this point, they argue, is ultimately corrosive to the market, and
inhibiting to the development of free enterprise (Seldon, 1990).
To many others, the comparison is essentially flawed. Construction of the PSC is a hypo-
thetical exercise, and the potential for disconnection between this hypothetical comparator
and what the public sector can and would actually do is a cause of criticism and objection to
PSCs by both PPP bidders and public sector alike, especially when there is no likely funding
for public sector procurement. It is considered that in these circumstances there may be a
‘subconscious psychological bias’ to private sector involvement, reinforced by the private
sector analysts employed to give advice who often are ‘not neutral referees but interested
7 This is a difficult area, brought to a head in the UK by the refinancing in the Fazakerley PFI prison project,
the ‘windfall gains’ from which were the subject of a special National Audit Office report (NAO, 2000). From the
viewpoint of the financiers offering better terms a track record of operations had been established and perceived
and actual risk factors removed. In a real sense, it can be argued that the private contractor benefited from its own
successful risk-taking.
players’ (Heald, 2003, p. 361). This leads us to the criticisms of the value for money tests
made in the academic literature.
5. Criticisms of the PSC approach
Prominent amongst those recently questioning the validity of the value for money
methodology are Broadbent, Gill, and Laughlin (2003), Broadbent and Laughlin (2003),
Heald (2003), and Shaoul (2005).8 On the specific issue of the application of PFI to the UK’s
National Health Service (NHS), earlier studies include those of Mayston (1999), Pollock,
Dunnigan, Gaffney, Macfarlane, and Majeed (1997) and Pollock (2000). Some of these
papers, along with other studies concerned with VFM and accountability issues in PPPs,
are surveyed in Demirag, Dubnick, and Khadaroo (2004). It is not possible to do justice here
to all of the points in these articles. Nor can we give an extended account of their arguments.
Instead, the major concerns raised are summarized in terms of general issues related to the
PSC and VFM test and those specifically related to the applications for the NHS.
The general concerns would seem to be as follows:
1. The value for money evaluation usually comes down to a choice between two very large
net present values, with the difference between them often small and reliant on the risk
transfer calculations included in the PSC. Because the PSC is entirely hypothetical,
its value can be altered by the assumptions made, especially about risk transfer to the
private sector, which is usually the crucial element in establishing expected VFM of a
PPP transaction over the PSC.
2. The discount rate methodology is faulty. It does not provide a measure of social time
preference. Because of the discounting inherent in calculating NPVs, even small changes
in the discount rate applied will vary the outcome as to which scheme is the best VFM.
3. Irrespective of how much risk is transferred to the private sector, the main risks (obso-
lescence, changing needs and service performance outcomes) are still held by the public
sector and costs fall upon the general public. Further, the real issue is uncertainty not
risk, and the significance of this distinction renders risk calculations problematical.
4. With contracts lasting sometimes up to 60 years, financial evaluations related to cost
estimates, discount rates and risk allocation are incomplete bases to draw conclusions
about the viability of proceeding with the PPP option, and more emphasis needs to be
given to non-financial elements in a longer term evaluation.
In terms of the special case of the application PFI to the NHS, there is a more lengthy
list and the concerns can be summarized as follows:
1. NHS decision-makers have not been able to choose freely between publicly and privately
provided infrastructure, and have been corralled into more expensive PFI options because
of cuts to direct public funding budgets. PFI is perceived as the ‘only game in town’.
8 There are also recent studies considering particular aspects such as taxation (e.g., Lehman & Tregoning, 2004),
accounting (e.g., Rutherford, 2003) and accountability (e.g., Watson, 2003) that space limitations prevent us from
examining here. Our earlier views on these three topics are in Grimsey and Lewis (2002b).
2. There can be no presumption that private entities possess better design capabilities
than the accumulated experience of the NHS itself, and no reason to assume that the
chosen private contractor will be the most appropriate for facilities management or for
providing finance (the ‘bundling’ issue). Speed has often been the major feature of the
design process.
3. The high degree of secrecy and commercial confidentiality surrounding many PFI deals
has hindered the process of public accountability and accounting for the contracts. Staff
has rarely been consulted.
4. Involvement of major construction companies in the PFI creates a bias toward new
building rather than maintenance and refurbishment of existing NHS capital facilities.
5. Tendering, negotiation, transactions (contracting), and monitoring costs are high relative
to the value of many NHS schemes. Consultancy fees charged to the NHS add to the
cost of PFI arrangements. Financiers and consultants may be the chief beneficiaries.
6. Payments under PFI schemes involve significant fixed payments to cover interest charges
and capital repayment. These debt repayment charges introduced into the finances of
health authorities limit the later flexibility of NHS bodies, by ‘mortgaging the future’ in
return for immediate gains (the ‘affordability’ issue).
7. Insolvency of the special purpose vehicle created by private sector sponsors to deliver
PFI services might result in lengthy litigation with no guarantee of financial redress,
leaving the public authority to ‘pick up the pieces’ and the patients to bear the cost of
service unavailability.
The general points raised by the academics are discussed in the next section. With respect
to the evaluation of VFM in the context of the NHS and the way PFI has been applied to the
health sector in the UK, a number of observations can be made. First, many of the issues
seem to be specific to the UK. Although, in general, we are wary about drawing too much
from individual cases, without supplementing these with the overall weight of evidence,
there are significant differences between the case studies of the Berwick hospital PPP in
Victoria, Australia and the West Middlesex hospital PPP in England, in terms of the PPP
model employed and the value for money test (Grimsey & Lewis, 2004c).9
Second, PPPs separate service provision from asset ownership and offer different financ-
ing options to public service managers, much like leasing facilities or hire purchase enable
payments to be spread over time as asset services are consumed. The economic liberal will
see these new instruments and options as welfare improving, in the sense that additional
choices are now available, even if not all public sector managers make sensible decisions
about their use.
Third, in terms of design, the issue is not just experience and knowledge of NHS
operations and practices but also importantly one of incentives. PPPs are no magic cure-
all for public services provision but often simply a way of providing decision-makers
with the appropriate economic incentives. In one Australian hospital PPP of which we
9 The former PPP project was undertaken using the ‘financier-led’ approach that has gained some popularity in
Australia whereas the UK PPP utilized the conventional ‘contractor-led’ model. Also, the projected cost saving of
9 percent against the net present risk-adjusted cost of the PSC in the Berwick project was more clear-cut than in
the West Middlesex project.
are aware, the private party responded to the whole-of-life cycle contract, and the abil-
ity to integrate design, construction, operations and maintenance, by designing dedicated
access space in the roof area and allowing access corridors, enabling maintenance of
the plant and machinery without having to close down ward space. As Peter Drucker
(1984) argues, innovation is often not about grand architectural design but about the
cumulative impact of a large number of small, perhaps unexciting, but cost-effective
changes.10
Fourth, to some degree concerns about the application of PFI to the NHS may be specific
to that sector. Allen (2001) argues that, in terms of VFM, there are considerable differences
across PFI types with road and prison projects achieving ‘reasonable efficiency gains’,
while those for schools and hospitals showing ‘minimal gains’ (p. 32). This difference is
attributed to two factors. The first is that for road and prison projects, there is no partition
of core and ancillary services, enabling the private contractor to make design and build
innovations since they will also operate and maintain services. In PFI health projects, core
and ancillary remain segmented, perhaps reducing some of the potential for innovation.
The other reason is that for roads and prisons there is a single, central government agency
handling the contracting, whereas with health and education private sponsors must deal with
a number of bodies such as NHS trusts, local education authorities and school governing
councils.
Finally, it has to be appreciated that health continues to be a highly charged political
area in the UK. Many people remain attached to the ideal of universal ‘free’ hospital and
medical care, and look askance at the cost of private medicine in the United States. Mayston
(1999) sees the PFI as a continuation of ‘contracting out’ in which the ownership and
management of major capital assets such as hospitals will be in private hands. He sees
parallels between PFI and earlier rounds of asset sales in the public sector under privatisation,
and regards the PFI as a further step along the path that one writer described as the ‘hollowing
out of the state’ (Rhodes, 1994). Pollock (2004) regards PPPs as part of a political plot
to privatize the NHS, notwithstanding evidence that public cash expenditure on health
has, in fact, been rising by about 10 percent per annum and is forecast to expand to 8
percent of GDP by 2007. Nevertheless, despite this expenditure, public funding is still
struggling to keep up with an ageing population and advances in medical technology, and
in such an environment, if PFI can deliver genuine cost savings or efficiency gains then
the experiment seems worthwhile. Yet PFI is no panacea and cannot overcome the basic
issue of ‘affordability’, highlighted by one of us some years ago, which arises because
the ‘Trusts’ expectations exceed the realities of their budgets’ (Grimsey & Graham, 1997,
p. 220).
6. Engaging the critics
This section addresses the general concerns raised in the recent academic literature. It is
worth pointing out at the outset that the four issues highlighted have featured in the practi-
10 The incentive structure and the bidder’s approach to innovations in design were seen by HM Prison Service to
have reduced cost by approximately 30 percent in the case of the Bridgend PPP prison project (NAO, 1997).
tioner and policy-based material, although in some cases, as we will show, the points have
been taken in a somewhat different direction and given a different emphasis. Nevertheless,
it does underscore the need for improved interchange between the two circuits because it
would seem that each has proceeded on its own separate tracks.
6.1. PPP–PSC comparison
Consider, first, the criticism that the difference between the PPP and PSC will in many
cases be small, and will depend on assumptions made, in particular about risk. Indeed, in
our experience this position is almost invariably the case, and a number of techniques have
been adopted for dealing with the problem, especially from those approaching the topic
from an engineering background. For example, consider Smith (1999):
‘. . . when the two alternatives are reasonably close together . . . minor changes in the
underlying assumption will cause the model to give completely different results. It
is therefore usual to construct dynamic cash flow models in which changes to the
key variables can be interactively tested, often using a computer-based spreadsheet.
Such models then allow the various parameters to be progressively varied in order
to simulate project performance over a wide range of conditions and to isolate the
variables to which the model is particularly sensitive’ (p. 181).
Thus, a key issue with the PPP–PSC assessment is whether a single number or a range of
values should be examined. There are clearly many ambiguities about the projected figures
for the PSC as is recognized by practitioners and academics alike. For this reason, it may
be sensible to compare ranges of outcomes using sensitivity analysis for the most important
assumptions and assessed risks, rather than rely on point estimates. Such a comparison
avoids the possibility that spurious accuracy finds its way into the comparisons and imma-
terial differences between very large total cash flows are used to justify proceeding with a
transaction.
However, this alternative modelling approach must satisfy its own value for money
requirements. The PSC approach is itself not a ‘first best’ approach but a cost-effective
compromise between a full cost-benefit analysis of all project options (as in Germany) and
simply selecting the ‘best’ private bid (as in France) which at the same time ensures that
all projects are treated in a like for like way and are subjected to a broadly similar and
systematic test for VFM.
In general terms, the requirements of the project and the quality and completeness
of the data are likely to determine the approach to modelling risks. Full statistical
analysis of outcomes using Monte Carlo simulation provides the most detailed compar-
isons, but is only worth undertaking where the quality of data is good and the range
of outcomes is sufficiently wide that different scenarios could change the choice of
procurement route. For these reasons, simple probability valuation methods are often
sufficient for all practical purposes. These are point estimates based on the subjective
probability and impact of each risk. The sum of these individual estimates provides
a single expected outcome for the transferred risk. Nevertheless, in the right circum-
stances, Monte Carlo simulations (run on packages such as @RISK or Crystal Ball) can
give more realistic analysis and representation of risk in terms of a range of possible
outcomes.11
Based on these techniques, the VFM comparison between the PSC and PPP bids can
then consider a range of potential outcomes for each method of procurement. The ranges
will probably differ, because there is likely to be a narrower range of outcomes under a PPP
reflecting the degree of risk transfer to the private sector. This spread of values highlights
the volatility of the project and introduces risk exposure into the PPP evaluation. In certain
circumstances it may be possible to justify taking forward a PPP project that may have a
similar expected outcome to the PSC but delivers to government a much lower exposure to
risk.
6.2. The discount rate
Large changes in practice have occurred over the last year, and to a considerable extent the
academic criticisms of the discount rate methodology have been left behind by developments
in policy circles and practitioner thinking, which were underway when many of the articles
were written. For this reason, it may be useful to provide some perspective on the revisions
to policy analysis, which found reflection in a number of draft documents circulated for
discussion amongst market participants.
The PSC is assessed over the life of the project in NPV terms, and the rate that is used
to discount cash flows to a present value has a big impact (which is obviously true also for
cost-benefit analysis). The discount rate can be seen as having two elements:
• The basic ‘Social Time Preference Rate’ (STPR). This represents the rate that society is
willing to pay for receiving something now rather than in the future. Calculations (e.g.,
HM Treasury, 2003a) suggest that in most developed countries this is around 3.5–4.0
percent in real terms (i.e., before allowing for price inflation).
• Some allowance for other factors, principally to ensure that the public sector does not
assess the benefit of projects without taking account of the risk to which it exposes
taxpayers in the process (for example, the potential to incur additional costs if things go
wrong).
Some public authorities in fact use their long-term borrowing rate as a proxy for the
discount rate on the grounds that they use additional borrowing to fund incremental expen-
diture. In authorities with an AAA credit rating this rate will tend to be close to the STPR
and below a risk-adjusted discount rate.
Until recently the UK public sector authorities assessed projects with a 6.0 percent real
discount rate on the grounds that the discount rate used should reflect the fact that resource
expenditure by government has an opportunity cost, with 6 percent real as the low-risk
cost of private capital, funded by a mix of debt and equity. This practice was followed in
other countries, such as Australia, notably in Victoria (Partnerships Victoria, 2001). It is
11 Government guidance material on PPPs (e.g., Partnerships Victoria, 2001) and project evaluation (e.g., HM
Treasury, 2003a) provide a good coverage of risk analysis. For a practical illustration of the various techniques
applied to a particular project, see Grimsey and Lewis (2002a).
the 6 percent real norm that is criticized in the academic studies cited earlier, which instead
favoured something lower such as the risk free rate on government borrowing or the STPR.
In fact, from April 2003, a 3.5 percent STPR rate was adopted in the UK, and accom-
panying the change new guidance material issued about how public authorities should take
separate account of the other factors previously reflected in the discount rates, particularly
risk. This requirement recognises that there will inevitably be variation between estimates
and actual costs over and above the impact of optimism bias (or project risk). The Green
Book (HM Treasury, 2003a) calls these variations ‘the cost of variability in outcomes’.
Interestingly, the Partnerships Victoria discount rate methodology also changed in 2003,
but in a different direction to the UK approach. In the new Partnerships Victoria (2003)
guidance material on discount rates, this is recommended to be a rate indicative of the
project risk, on the grounds that the cost of capital or discount rate is specific to each
project and is a function of the risks for the particular project in question. The guidance
material is based on the application of the Capital Asset Pricing Model (CAPM) to PPP
project evaluation.12 A key benefit of the CAPM approach is that it recognises that the
cost of capital/discount rate is specific to each project and is a function of the risks for
the specific project. In a perfect market, this would lead to the conclusion that, as long as
there is sufficient competition to drive every component of the deal to maximum efficiency,
the appropriate discount rate would be the rate of return implicit in the winning bid, and
therefore one would not need to develop a specific discount rate for analysis.
These two approaches to setting the discount rate may appear to be opposed to one
another, but the principles underlying them are essentially the same. The discounted cash
flow analysis involves discounting the expected cash flows associated with a project to
produce a risk adjusted present value figure that takes into account all financially measurable
benefits, costs and risks for the project. One method of adjusting for risk is to add a risk
margin to an appropriate risk-free discount rate. An alternative treatment is to value risk in
the cash flows so that a risk free discount rate is applied to cash flow forecasts that have been
adjusted from their risky form to ‘certainty-equivalents’, in the terminology of Brealey and
Myers (2003). In effect, in terms of the conventional discounting formula, in one approach
risk is allowed for in the denominator, while in the other method a risk adjustment is made
in the numerator. In spite of the formal equivalence of the two ways of allowing for risk, it is
easy to understand how the process of putting theory into practice could result in differences
arising. If the risk-adjusted interest rate method is used, then this must reflect systematic
risk rather than idiosyncratic project risk, since the latter can be pooled and diversified in
such a way that no one individual bears any significant risk. The reward for bearing risk
depends only on the systematic risk of an investment because systematic risks are highly
correlated and pooling and diversification has little effect other than to redistribute a given
risk across the community. But the CAPM from which this result derives is likely to provide
only a broad range of appropriate discount rates, and practitioners may have to fall back
on an initiative approach to estimating discount rates, based on precedents and judgment,
rather than market information.
12 It also follows the methodology contained in the document, Determination of Appropriate Discount Rates
for the Evaluation of Private Financing Proposals 7 June 2002, prepared by the Australian Commonwealth
Government which at the time of writing is still in draft form.
If the alternative of an STPR approach is adopted then it is necessary for risks in the
cash flow forecasts to be adjusted using the certainty equivalence method (or ‘the cost of
variability in outcomes’ as defined in the UK Green Book). This involves estimating, for
example, what the government would be willing to pay in order to know with certainty
that its outlay under the procurement will be the expected cash flow payment. Because it
is difficult to price systemic risk as a cash flow, available market data may be even less
useful than for discount rate assessments and there is little in the way of guidance available
from existing theory as to how this calculation is to be made. Practitioners are left to make
their own assessment of the value of risk for a specific set of expected cash flows (e.g.,
using predictions of insurance premiums required to avoid risk or CAPM-based formula to
estimate separate adjustment factors for each key cash flow element, based on individual
beta estimates for each cash flow factor).
In conclusion, while the idea of using a risk-free discount rate as advocated in the
academic criticisms has an intuitive appeal, the rest of the exercise of adjusting the cash
flows appropriately seems particularly difficult to carry through.13 Heald (2003) in fact
foreshadowed that the implications for PFI appraisal are likely to be more complicated
than a simple reduction in the discount rate from 6 to 3.5 percent, and this assessment is
something on which we can agree.
6.3. Uncertainty
Risk transfer is a key element in the PPP–PSC comparison. Shaoul (2005) argues that
the measurement and methodology of risk transfer is rendered problematical because all
possible outcomes cannot be predicted and weighted, and the complete array of results
covering all eventualities compiled, when the issue is uncertainty not risk. This emphasis
on the importance of uncertainty as well as risk in VFM evaluation is certainly something
with which we can agree (Grimsey & Lewis, 1999, 2002a, 2004a, 2004c). The difficulty is
how to deal with uncertainty in the context of VFM evaluation in general, and PPP–PSC in
particular.
Part of the problem is that uncertainty is poorly understood. For example, uncertainty is
mentioned in the UK Green Book in a number of places, yet in the one practical illustration
provided of allowing for uncertainty in an analysis of costs (Box 4.5), reference is made
to the ‘probability distributions specified for each variable’ indicating that it is ‘risk’ rather
than ‘true’ uncertainty that is being discussed. In the case of risk, the probabilities of the
various future outcomes are known (either exactly mathematically, or from past experience
of similar situations). In the case of uncertainty, the probabilities of the various future
outcomes are merely ‘wild guesses’ in the description of Frank Knight (1921), because the
‘instance’ in question is so unique. September 11 is an obvious example, and a statistical
basis for assigning a probability distribution to its occurrence did not exist in any real sense.
At a practical level, governments traditionally do not budget for systematic risks or
uncertainty, and consequently PSC calculations only contain project specific risks that are
identified and quantified with no adjustment for systematic risk or uncertainty. Systematic
13 These issues are discussed further in Grimsey and Lewis (2004a).

risks have now been taken care of in the discount rate or ‘the cost of variability in outcomes’,
but not uncertainty. Public sectors as a whole might be able to ignore uncertainty across
their whole portfolio but bidders for a PPP project cannot. Yet the fact is that PPPs move
the project out of the government’s portfolio and in a sense provide some insurance to the
government against uncertainty. Of course, this is not true for the private sector and this fact
would seem to have implications for the setting of discount rates and the value for money
test.
Consider, for example, the study by PricewaterhouseCoopers (2002) that provides a
different way of assessing value for money than the case-by-case PPP–PSC comparisons.
Utilizing the capital asset pricing model, the study analyzed the projected rates of return on a
sample of 64 PFI projects to ascertain whether the returns that the private sector expected to
earn for managing and bearing risk were excessive or in line with what might be anticipated
from a competitive market amongst bidders. Across all projects in the sample, the internal
rate of return is found to average 7.7 percent per annum, whereas the average weighted
average cost of capital is estimated to be 5.3 percent per annum. Thus, the ‘spread’, the
amount by which the average project internal rate of return is higher than the cost of capital,
is 2.4 percent per annum, described as an estimate of the total excess projected return on
PFI projects above the cost of finance. However, of this figure 1.7 percent is thought to be
accounted for by two factors: unrecovered bid costs on other projects (about 1 percent);
and the higher cost of underlying rates for private sector borrowing compared with public
sector borrowing, primarily caused by the cost of swaps compared with gilts (about 0.7
percent). After taking account of these two factors, the excess projected return to project
investors is estimated as being about 0.7 percent. These excess returns are attributed to
‘structural issues’ that in the past have limited competition in the PFI market. For instance,
the length of PFI procurements and the level of bid costs incurred by the private sector
are thought still to create barriers to entry to the market. Our objection is that the study
focused on risk and ignored that bidders may seek to incorporate a margin for uncertainty
into their calculations. If this is so, some part of the ‘excess’ returns may be attributable to
a failure by the researchers to take account of the private sector’s responses to the presence
of uncertainty.
Uncertainty also has implications for the discount rate. Broadbent et al. (2003) noted
the argument of Grout (1997), later developed further in Grout (2003), to the effect that the
VFM test is biased against the public sector. His argument runs as follows. When public
sector provision is being valued a discount rate is applied to a cost cash flow. This cash flow
represents the cost of building the facility if it is done in the public sector. In contrast, for
valuing the private sector provision a discount rate is applied to a stream that constitutes an
outlay for the public sector but is a revenue item to the private entity and is being valued
from the revenue side. With a PPP, this revenue stream is not the equivalent cost of building
the facility. It is the cash flow associated with the flow of benefits valued at the price in the
contract. There is no reason to suppose that the risk characteristics are equivalent for these
two cash flows. Indeed, Grout argues that there is every reason to suppose that they are
not, because in general costs are less risky than revenues (particularly when the revenues
depend on services of a suitable quality being provided). Under what appear to be plausible
conditions, he contends that a higher discount rate should be used for the PPP than for the
public sector equivalent. Failure to do so will suggest that private provision is less efficient
than public since the present value of private provision will be overestimated relative to
public procurement.
Grout’s proposition hints at what are fundamental differences in terms of risk between
the cash flows built up for the PSC and those bid as service payments as a PPP. Those who
adhere to the Knightian distinction between risk and uncertainty would contend that ‘true’
uncertainty may also need to be taken into account. The PPP cash flows include premiums
for project or idiosyncratic risks, systematic or market based risks, and may also incorporate
an allowance for uncertainty. If so, the fact that uncertainty exists and will be priced into
PPP cash flows reinforces Grout’s argument that a higher discount rate may be warranted
for PPP discounting than for the PSC.
6.4. Long-term VFM evaluation
Broadbent et al. (2003) stress the need for a long-term evaluation of PPPs, arguing that
undue emphasis is given to the pre-decision stage when considering VFM. To this end,
they seek to develop a framework for long-term post-project evaluation, focusing on risk
allocation (whether risk assessment and allocation is as predicted), facilities management
(whether the facilities achieve the standards intended), and the non-financial aspects of the
project. However, with contracts lasting up to 60 years, it is recognized that any meaningful
evaluation will take years to complete, although some tentative conclusions as to VFM
might be able to be reached 5–7 years into the contract.
Valuable though such an evaluation process may be, from the viewpoint of policy imple-
mentation the more pressing need is ensuring that the relationship is managed in such a
way that the outcomes actually are realized. That is, the public sector agency’s immedi-
ate concern is with governing the contract, rather than evaluating it down the track, and
this aim is reflected in the practitioner literature (e.g., Partnerships Victoria, 2001). The
emphasis in this literature is different from Broadbent et al. in two ways. First, instead of
‘post-project’, the focus is upon the ‘realization’ phase or ‘post-financial close’ phase of
the project. Bringing a PPP to financial close involves a major investment by all parties
to the contract, but it cannot be the end of the matter. Indeed, it is the conversion of the
contract into delivery of the outputs that is essential to meeting the overall project objec-
tives. Second, once the contract has been signed, it then has to be managed by the public
sector agency, which is responsible for monitoring and implementing the contract. Rather
than a post-project evaluation system, the over-riding priority is to establish a contractual
governance system in order to oversee project delivery, deal with contract variations and
handle re-pricing adjustments to maintain VFM, monitor the service outputs and detect any
problems at an early stage.
Elsewhere (Grimsey & Lewis, 2004b) we have outlined a contractual management and
reporting regime to aid this function, indicating the degree to which project objectives are
being achieved and providing an early warning of potential problems. Special emphasis has
been given to establishing, first, key reporting requirements for the meeting of performance
quality standards and, second, a performance monitoring system for assessing the ‘health’
of the contracting enterprise and the risk of default—an important factor when government
is unable to walk away from providing the core services, which is a point of concern to a
number of the authors referred to earlier.
In this framework, contract management can be seen as processes undertaken to main-

tain the integrity of the contract, and ensure that the roles and responsibilities contractually
demarcated are fully understood and are carried out to the contracted standard. The con-
tractual performance specifications and the associated incentive payment mechanism are
the primary legal means by which the public sector can enforce the project objectives and
risk transfer provisions. The specification sets out the scope and quality of services that
need to be provided. In turn, the payment mechanism sets out precisely how the SPV is
paid for providing the specified services. Performance regimes and payment mechanisms
are central to the success of the project in that they provide commercial incentives for the
private sector to manage risks properly and deliver value for money.
7. Do PPPs deliver value for money?
The previous section has shown that there exists a considerable overlap between issues
raised in the practitioner and the academic literature referred to earlier, revolving around
risk transfer, discount rates, uncertainty and the need for long-term assessment of PPP
performance. Nevertheless, our perception is that a considerable gulf still remains between
market participants and the accounting and public policy academics on the central issue
posed in this paper: Are PPPs value for money? By way of background, we would note
on this point that there is another literature on PPPs, namely, that in the engineering-based
journals.14 In general, however, many of the articles have a practitioner, often case study,
focus to them and in this respect are probably closer to practitioner views. Another point
that needs to be made is that, despite the attention given to PPPs in the literature, they are
in reality the smaller part of the public infrastructure market. PFI accounts for between
only 10–14 percent of public sector investment in the UK and PPPs for about 10 percent
in Victoria, the leading market in Australia.15 Most public infrastructure projects are still
traditionally procured, and PPPs can hardly be said to be taking over the market for public
procurement.16
In seeking to develop an understanding of what merits PPPs bring to the procurement
market, this section first considers the available evidence on PPPs. It then endeavours to
provide a conceptual underpinning for PPP performance and thus explains some of the
reasons why practitioners and others in the policy arena look to PPPs as a procurement
option.
PPPs have appeal (especially to those in charge of allocating public sector resources)
because they offer one way of resolving the large cost overruns and delays in traditional
public procurement methods for infrastructure—phenomena known as ‘optimism bias’.
Two studies of optimum bias in 2002 confirmed the results of earlier research by Pickrell
14 Here we refer the reader to journals such as Engineering, Construction and Architectural Management, Inter-
national Journal of Project Management, Journal of Project Finance, Infrastructure, Transportation Quarterly and
so on. Articles on PPPs from all of these journals are contained in Grimsey and Lewis (2005).
15 The UK data are from HM Treasury (2003b).
16 Thus, the title of Watson’s (2003) article ‘The rise and rise of public private partnerships’ somewhat overstates
the position.
(1990) and Fouracre, Allport, and Thomson (1990). In the first study, Flyvbjerg, Holm, and
Buhl (2002) examined 258 large transport infrastructure projects covering 20 countries,
the overwhelming majority of which were developed using conventional approaches to
public procurement. Costs were found to be underestimated in 90 percent of the projects, in
most cases by substantial amounts. In the other major study, the UK Treasury commissioned
MacDonald (2002) to review 50 large public procurement projects in the UK over the last 20
years, 11 of which were undertaken under PPP/PFI. On average, the PPP/PFI projects came
in under-time (compared to 17 percent over-time for the others), and capital expenditure
resulted in a 1 percent cost overrun on average (relative to an average cost overrun of 47
percent for traditional procurement projects).
Further evidence for the UK comes from HM Treasury (2003b) in its own review of 61
PFI projects, where it was found that 89 percent of the PFI projects were delivered on time or
early and that all PFI projects were within budget. To these results we can add the findings
of the UK National Audit Office (2003) of PFI construction performance. In contrast to
traditionally procured projects, the PFI projects were largely being delivered on time (76
percent of PFI versus 30 percent for conventional procurement) and on budget (78 percent
versus 27 percent). Significantly, in no case did the public sector bear the cost of construction
overruns, a major change from previous non-PFI experience where the financial costs of
projects that ran into difficulties were absorbed into government budgets.
These figures relate to construction cost and time overruns. Value for money tests based on
the PPP–PSC comparisons, however flawed they may be in some eyes, enable us to put some
numbers on the expected overall gains from PPPs. In the UK, a review, commissioned by
the Treasury Taskforce (now Partnerships UK) and published jointly by the London School
of Economics and Arthur Andersen in January 2000, analysed 29 public sector projects that
used the PFI and it was calculated that on average the predicted savings from using the PFI,
compared with conventional procurement, was 17 percent (Arthur Andersen, 2000). In its
own separate analysis, the National Audit Office has produced value for money reports on 15
projects, 7 of which were evaluated for value for money against a public sector comparator
for traditional public procurement. Overall, the total cost savings of these projects was 20
percent (NAO, 2001).
Studies for some particular sectors in the UK report broadly consistent results. Parker
and Hartley (2003) record claims that PPP contracts for UK defence services have resulted
in cost savings between 5 and 40 percent compared with conventional public procurement,
although the authors are concerned as to whether these apparent cost savings will be realized
over the projects’ whole-of-life due to the inherent uncertainties of long-term contracting
(which reinforces our comments in the previous section as to the importance of contract
management in realizing VFM in a PPP). In the case of roads, an earlier study (NAO,
1998) of the first four DBFO road contracts found that the four contracts appear likely
to generate net quantifiable financial savings of around 100 million pounds (13 percent).
Hodgson (1995), reviewing the switch from conventional procurement methods to DBFO
arrangements for roads, notes that for traditional procurement involving the award of design
and construct contracts ‘the public sector’s record in the design and construction of capital
schemes is poor. Time and cost overruns are common. Part of the reason lies in the attitudes
and culture of the public sector. In the construction sector this often results in conservative
or over-engineered designs’ (p. 68).
In Australia, a recent review of the Partnerships Victoria policy found evidence of net
benefits of PPPs, and that for the eight PPP projects examined by the review the weighted
average savings was 9 percent across all projects (Fitzgerald, 2004). Directly comparable
evidence for other countries is sparse, in part because French-style concessions are widely
used and a PSC is not employed as a systematic test of VFM. However, in The Netherlands,
the Wastewater Treatment Delfland PPP, closed in 2003, achieved an expected saving of
around 15 percent compared with the PSC.
Why? What aspects of a PPP might bring about this marked difference in performance?
Here, we are led back to Officer’s comparative advantage approach to public and private
sector participation in the delivery of public services. He argued that, on a number of scores,
the government would appear to have a ‘comparative disadvantage’ in undertaking many
business activities. Principal amongst these is the difficulty it has in creating incentives for
managers to act in commercially oriented ways. By contrast, market competition can be
seen as a form of coordination with intrinsic advantages over bureaucratic organizational
forms and the discipline and incentives embodied in market contracting arrangements are
valuable in injecting greater efficiency into infrastructure delivery, in three ways. First, it is a
means of resolving agency problems, which arise due to the divergence between ownership
and control of assets, and are easier to manage in the private sector through managerial
incentives and market disciplines. Second, the transfer of risk to the private sector provides
an incentive for private entities to maximize efficiency. Third, resources are more efficiently
allocated in cases where clear markets for property rights can be established.
The latter two are especially important in a partnership arrangement. An effective transfer
of risk from the public sector to the private sector is needed, since it is the acceptance of risk
that gives the private entity the incentive to price and produce efficiently. However, for this
transfer to happen, a clear property right needs to be created (Alchian, 1965). Within the
ambit of a PPP, this suggests an allocation of responsibilities that might have the government
retaining those areas where it is difficult to establish a clear contractual specification, leaving
the private sector to undertake those activities for which a clear and unambiguous contract
or property right can be formed. One of the merits of going through the PSC exercise is in
focusing attention on these issues in order to determine whether these preconditions exist
and whether a PPP approach could, if used, add value for money through risk transfer and
the incentives that follow.
Incentives to performance under PPPs come from integrating within a private sector party
all (or most of) the functions of design, building, financing, operating, and maintenance of
the facilities or assets in question, often in the form of a special purpose vehicle created
for the specific project, under a contractual arrangement based on services provided rather
than asset provision Grout (1997). Writing the contract in terms of the flow of services from
the infrastructure facility rather than the process of construction can change the incentive
system markedly from the ‘claims culture’ and continuous redesign that often marks con-
ventional public procurement (Heald, 2003).17 Having the same entity responsible for both
construction and supplying the services, but remunerated only for the successful provision
of services of a suitable quality, encourages the private party to build the correct facility,
17 Heald describes the ‘claims culture’ as the practice of winning the contract by bidding low and then securing
additional payments.
get the process of delivery right and contain costs while not sacrificing quality. The public
sector procurer, by shifting risks and responsibilities on to the private contractor under a
PPP, while gearing payment only to successful delivery of the services, sends out a strong
signal that time and cost overruns and service quality lapses will not be accepted and that
payments for services will be at risk.
However, if we are to understand the full set of incentives operating upon PPPs we need
to take this argument further. In particular, the financial side of a PPP is important too.
Indeed, we contend that finance should be seen as the ‘glue’ that binds the risk allocation
in a PPP together. Financiers have incentives to make sure that services are supplied on
time and to the requisite standard when the revenue stream that is generated represents
the main source for repaying debt. There is also the participation of private risk capi-
tal, which brings with it with the expectation that those with equity at risk will insist on
there being a harder-nosed approach to project evaluation, risk management and project
implementation—effectively, at all stages of the project. Project-financing techniques have
to be ‘engineered’ to take account of the risks involved, sources of finance, accounting and
tax regulations, etc. Project design requires expert analysis of all of the attendant risks and
then a structuring of the contractual arrangements prior to competitive tendering that allo-
cates risk burdens appropriately. Project implementation then requires expert management
of all of the component parts of the delivery of the services, such as construction, commis-
sioning, operation and maintenance. This last stage is essential to the successful treatment
of all of the attendant risks and, from the senior financier’s point of view, secures control of
the sole source of security against the cash flows. In effect, having the privately provided
finance at risk, acts as a catalyst to inject risk management techniques into the project in a
way that is not possible under government financing.18
Focusing upon risks and their allocation is central to understanding what is special
about a PPP relative to conventional procurement. Transferring the responsibility for design
of both the facility and the delivery system on to the private contractor encourages it to
choose designs that will work, and explore innovations that can improve quality and reduce
maintenance and operating costs. Giving the responsibility for construction and project
management to the private entity creates the incentive for it to keep the project on track and
to prevent construction delays and cost overruns. Involving the private sector in financing
the project means that the financiers will look to the security and timeliness of the revenue
stream, and put in place controls over the operators that will minimize the risk of project
failure. Requiring the private body to operate and maintain the facility, as well as design and
construct it, reduces any incentive to skimp on the quality of materials used, while encour-
aging decisions that maintain services to the desired level and keep costs at a minimum. It is
this integration of upfront design and financial engineering to downstream management of
the construction costs and revenue flow that gives the PPP its distinctive incentive compat-
ibility characteristics—characteristics that are difficult to replicate within the public sector.
18 Here, we note Heald’s (2003) argument that ‘finance providers may impose stricter ex ante controls on the
construction stage, thus ensuring that the project is fully costed and that the contractor does not anticipate that
claims for additional work will meet a soft budget constraint’ (p. 366). However, we would add that finance
providers may also impose strict controls ex post (following construction) over delivery to ensure that senior debt
is repaid.
8. Concluding remarks
This paper has had two aims. At one level it has endeavoured to provide an overview,
based on over twenty countries, of value for money assessment in Public Private Partnerships
focusing in particular on the role played by the Public Sector Comparator. PPPs are well-
established practice in the UK, Australia, The Netherlands, South Africa, Canada and Japan,
and in all these countries PSCs are regarded as valuable for assessing value for money
in procurement and evaluating and quantifying risk. However, the use of a PSC is not
universal and elsewhere, where PPPs are employed, our survey reveals that other methods
are employed. Overall, we find that there are four alternative approaches to evaluating value
for money, although a number of countries that have used cost-benefit analysis or other
techniques are now investigating the use of PSCs for ascertaining whether PPPs constitute
good value for money.
At a second level, the paper has sought to engage some of the academics voicing concerns
about PPPs, examining some of the similarities and differences between their views and
those of practitioners active in the market. Amongst some academics, the value for money
tests involving PPP–PSC comparisons are criticized for the seeming arbitrariness of the
assumptions underlying the PSC, especially with respect to risk transfer and the discount
rate (where small changes can have a big impact on project choice), as well as for down-
playing uncertainty and over-emphasizing financial factors relative to issues of long-term
service delivery. Some of these matters (e.g., the assumptions underlining the PSC and
risk transfer) have been considered by practitioners, who use statistical techniques such as
sensitivity analysis and Monte Carlo methods to examine the robustness of the estimates.
Other matters raised (e.g., the treatment of risk and the discount rate methodology) have,
to a considerable degree, been overtaken by developments in policy analysis in both the
UK and Australia, although aspects of the application of the new methodology remained
unresolved. Uncertainty and the need to manage long-term service delivery are issues that
remain on the policy agenda, although practitioner views on these matters have taken some
different directions to the academic literature.
Rather than go over some of the same ground, we would like to conclude with some
general points. First, our survey shows not only that there are alternatives to the PSC
approach but that there are many complexities and ambiguities involved in it which suggest
that the calculation of the PSC needs to be seen merely as one factor, albeit significant, in
procurement decisions. Its development constitutes a valuable discipline upon public sector
procurement in assisting decision-makers to understand the project, the risks involved and
how to deal with them contractually. In this respect, the risk analysis required for the PSC
can be seen as part of the broader process of risk identification, allocation and management
within the project. In many cases, the difference between the PSC and the private sector
proposal will be relatively narrow and the procurer has to make professional judgements as
to the value for money to be derived from contracting with the private sector and the risks
which that route involves, while not ignoring that there are also large risks in the public
procurement route, as indicated by the ‘optimism bias’ documented earlier.
Second, the value for money test frequently comes down to a simple, single point com-
parison between two procurement options. In our view, the problem is that value for money
is more often than not poorly understood and often equated with lowest cost. Such an
approach fundamentally understates value in the context of achieving a project’s objec-

tives and protecting the government from adverse outcomes. Value is a complex trade off
between cost, risk and performance, and in this framework it is important fully to understand
the government’s exposure to risk, defined as volatility of outcomes. Probability analysis
overcomes the limitations of the simple point estimate value for money approach by spec-
ifying a probability distribution for each risk, and then considering the effects of the risks
in combination. Nevertheless, it is important to remember that uncertainty is not the same
as risk, and that consideration needs to be given to the unexpected as well as to measurable
risks. Also, value for money as measured by financial comparisons should be augmented
by a consideration of the policy and strategy context and the wider socio-economic costs
and benefits, and we have no quarrel with the academic critics on that general point. This
is one advantage of the approach where there is a full cost-benefit economic analysis of the
feasible public sector option and the PPP route to procurement, but this alternative does
come at a considerable cost in terms of a commitment of time and resources that the more
practically oriented PSC approach to some degree obviates.
Third, PPPs are not, and probably never will be, the dominant method of infrastructure
acquisition. They are too complex, and costly, for many small projects. In some cases, they
may be beyond the capacity of the public sector agency to implement and manage. For
other projects the tight specification of the outputs required may be difficult to detail for
an extended period. In other cases, a sound business rationale may not exist. Nonetheless,
for those projects that are suitable, they are a way of introducing very different incentives
into the procurement process. These incentives, we argue, are distinctive to PPPs relative
to other procurement approaches and come about as a result of the fusion of the upfront
engineering of the design and the finance with the downstream management of construction
and service delivery. Private risk finance using a mix of equity and debt is central to this
process, and provides the ‘glue’ that holds together the transaction and the risk allocation
amongst the various parties. Moreover, there is evidence that the private sector does respond
to these signals and gets it right more often than not. About 75 percent of major infrastructure
projects in the UK were late and over budget before PPPs came into play. Under PPP/PFI
arrangements, 75 percent of projects are on time and to budget.
References
Alchian, A. A. (1965). Some economics of property rights. Il Politico, 30, 816.

Allen, G. (2001). The private finance initiative (PFI). House of Commons Research Paper 01/117. House of
Commons Library, London, December.
Arthur Andersen. (2000). Value for money drivers in the private finance initiative. London: Arthur Andersen and
Enterprise LSE.
Brealey, R., & Myers, S. (2003). Principles of corporate finance (7th ed.). Boston: Irwin McGraw-Hill.
Broadbent, J., Gill, J., & Laughlin, R. (2003). Evaluating the private finance initiative in the national health service.
Accounting, Auditing and Accountability Journal, 16(3), 422–445.
Broadbent, J., & Laughlin, R. (2003). Public private partnerships: An introduction. Accounting, Auditing and
Accountability Journal, 16(3), 332–511.
Demirag, I., Dubnick, M., & Khadaroo, M. I. (2004). A framework for examining accountability and value for
money in the UK’s private finance initiative. Journal of Corporate Citizenship, 15, 63–76.
Drucker, P. (1984). Innovation and entrepreneurship. New York: Harper Collins.
English, L. M., & Guthrie, J. (2003). Driving privately financed projects in Australia: What makes them tick?
Accounting, Auditing and Accoutability Journal, 16(3), 493–511.
European Investment Bank. (2004). The EIB and PPPs. EIB Information, 116, 22–24.
Fitzgerald, P. (2004). Review of partnerships victoria provided infrastructure. Melbourne: Review of Partnerships
Victoria.
Flyvbjerg, B., Holm, M. S., & Buhl, S. (2002). Underestimating costs in public works projects: Error or lie?
Journal of the American Planning Association, 68(3), 279–295.
Fouracre, P. R., Allport, R. J., Thomson, J. M. (1990). The performance and impact of rail mass transit in developing
countries. TRRL Research Report 278. Crowthorne, UK: Transport and Road Research Laboratory.
Grimsey, D., & Graham, R. (1997). PFI in the NHS. Engineering, Construction and Architectural Management,
4(3), 215–231.
Grimsey, D., & Lewis, M. K. (1999). Risk analysis in public sector/private sector infrastructure arrangements. In
The third international Stockholm seminar on risk behaviour and risk management.
Grimsey, D., & Lewis, M. K. (2002a). Evaluating the risks of public private partnerships for infrastructure projects.
International Journal of Project Management, 20, 107–118.
Grimsey, D., & Lewis, M. K. (2002b). Accounting for public private partnerships. Accounting Forum, 26(3),
245–270.
Grimsey, D., & Lewis, M. K. (2004a). Discount debates: Rates, risk, uncertainty and value for money in PPPs.
Public Infrastructure Bulletin, March, 3, 4–7.
Grimsey, D., & Lewis, M. K. (2004b). The governance of contractual relationships in public private partnerships.
Journal of Corporate Citizenship, 15, 91–109.
Grimsey, D., & Lewis, M. K. (2004c). Public private partnerships: The worldwide revolution in infrastructure
provision and project finance. Cheltenham: Edward Elgar.
Grimsey, D., & Lewis, M. K. (2005). The economics of public private partnerships. Cheltenham: Edward Elgar,
in press.
Grout, P. (1997). The economics of the private finance initiative. Oxford Review of Economic Policy, 13(4), 53–
66.
Grout, P. A. (2003). Public and private sector discount rates in public private partnerships. Economic Journal,
113(486), C62–C68.
Heald, D. (2003). Value for money tests and accounting treatment in PFI schemes. Accounting, Auditing and
Accountability Journal, 16(3), 342–371.
HM Treasury. (2003a). The Green Book—Appraisal and evaluation in central government. London: TSO.
HM Treasury. (2003b). PFI: Meeting the investment challenge. Norwich: HMSO.
Hodgson, G. J. (1995). Design and build—effects of contractor design on highway schemes. Proceedings of Civil
Engineers, 108(May), 64–76.
Hood, C. (1995). The new public management in the 1980s: Variations on a theme. Accounting, Organizations
and Society, 20(2–3), 93–109.
Knight, F. (1921). Risk, uncertainty and profit. Boston: Houghton Mifflin.
Lehman, G., & Tregoning, I. (2004). Public private partnerships: Taxation and Civil Society. Journal of Corporate
Citizenship, 15, 77–89.
Mayston, D. J. (1999). The private finance initiative in the national health service: An unhealthy development in
new public management? Financial Accountability and Management, 15(3–4), 249–274.
Minow, M. (2003). Public and private partnerships: Accounting for the new religion. Harvard Law Review, 116,
1229–1237.
MacDonald, M. (2002). Review of large public procurement in the UK. London: HM Treasury.
National Audit Office. (1997). The PFI contracts for Bridgend and Fazakerley prisons, HC 253 Session 1997–98,
31 October 1997.
National Audit Office. (1998). The private finance initiative: The first four, design, build, finance and operate
roads contracts, HC 476, Parliamentary Session 1997–98. London: HMSO.
National Audit Office. (2000). The refinancing of the Fazakerley PFI, HC 584 Session, 1999–2000. London:
HMSO.
National Audit Office. (2001). Managing the relationship to secure a successful partnership in PFI projects, HC
375 Session 2001/2002. London: HMSO.
National Audit Office. (2003). PFI: Construction performance, report by the comptroller and auditor general.
London: HMSO.
NEDC (1981). Report of the National Economic Development Council Working Group on Nationalised Industries
Investment, London: NEDC.
Office of Government Commerce (2002). Green public private partnerships. Norwich: OGC.
Officer, R. R. (2003). The respective roles of government and the private sector and private/public partnerships.
Public Private Partnerships Forum (The Accounting Foundation, The University of Sydney, 8 December).
Parker, D., & Hartley, K. (2003). Transaction costs, relational contracting and public private partnerships: A case
study of UK defence. Journal of Purchasing and Supply Management, 9(3), 97–108.
Partnerships Victoria. (2001). Guidance material. Melbourne: Department of Treasury and Finance.
Partnerships Victoria. (2003). Use of discount rates in the Partnerships Victoria process—Technical note. Mel-
bourne: Department of Treasury and Finance.
Pickrell, D. H. (1990). Urban rail transit projects: Forecast versus actual ridership and cost. Washington, DC:
US Department of Transportation.
Pollock, A. M., Dunnigan, M., Gaffney, D., Macfarlane, A., & Azeem Majeed, F. (1997). What happens when
the private sector plans hospital services for the NHS: Three case studies under the private finance initiative.
British Medical Journal, 314, 1266–1271.
Pollock, A. M. (2000). PFI is bad for your health. Public Finance, 30–31.
Pollock, A. M. (2004). NHS plc: The privatisation of our health care. London: Verso.
PricewaterhouseCoopers. (2002). Study into rates of return bid on PFI projects. London: OGC.
PricewaterhouseCoopers. (2004). Developing public private partnerships in new Europe. London: Pricewater-
houseCoopers.
Rhodes, R. (1994). The hollowing out of the state: The changing nature of the public sector in Britain. Political
Quarterly, 65(2), 138–158.
Ribault, A. (2001, Spring). Lessons from the French experience in public and private partnership. Irish Banking
Review, 49–60.
Rosenau, P. V. (2000). Public private policy partnerships. Cambridge, Mass: The MIT Press.
Rutherford, B. A. (2003). The social construction of financial statement elements under private finance initiative
schemes. Accounting, Auditing and Accountability Journal, 10(3), 372–396.
Schneider, A. L. (1999). Public private partnerships in the U.S. Prison System. Behavioral Scientist, 43(1), 192–208.
Seldon, A. (1990). Capitalism. Oxford: Basil Blackwell.
Shaoul, J. (2005). A critical financial analysis of the Private Finance Initiative: Selecting a financing method or
allocating economic wealth? Critical Perspectives in Accounting, 16(4), 441–471.
Smith, A. J. (1999). Privatized infrastructure: The role of government. London: Thomas Telford.
Starr, P. (1988). The meaning of privatization. Yale Law and Policy Review, 6, 6–41.
Watson, D. (2003). The rise and rise of public private partnerships: Challenges for public accountability. Australian
Accounting Review, 13(3), 2–14.

Are Public Private Partnerships Value For Money PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Are Public Private Partnerships Value For Money PDF

Uploaded by

Copyright:

Available Formats

Accounting Forum 29 (2005) 345–378

Are Public Private Partnerships value for money?

0155-9982/$ – see front matter © 2005 Published by Elsevier Ltd.

1. The PPP route

2. Approaches to value for money

Commerce, and is discussed later in this paper.

3. Calculation of the PSC

3.1. Raw PSC (base costs)

• preliminary set-up and planning work, approvals and permits;

3.2. Transferable risk

3.3. Retained risk

3.4. Competitive neutrality

3.5. The VFM comparison

4. Some practical considerations

Fig. 1. PSC and value for money.

4.1. Is the PSC too subjective?

4.2. Is the PSC needed?

4.3. Is the PSC too simplistic?

A cost-benefit analysis can be a powerful tool in establishing the priorities of projects to

4.4. Unquantiﬁable elements

4.5. The PSC is more risky

4.6. The PSC is incomplete

4.7. The comparative advantage rule

5. Criticisms of the PSC approach

6. Engaging the critics

6.1. PPP–PSC comparison

6.2. The discount rate

13 These issues are discussed further in Grimsey and Lewis (2004a).

6.4. Long-term VFM evaluation

In this framework, contract management can be seen as processes undertaken to main-

7. Do PPPs deliver value for money?

approach fundamentally understates value in the context of achieving a project’s objec-

Alchian, A. A. (1965). Some economics of property rights. Il Politico, 30, 816.

You might also like