Written evidence submitted by Dr Andrew
Edkins, Graham Ive and Alex Murray, University of London |
Evidence submission representing the views of Dr
Andrew Edkins (Senior Lecturer), Graham Ive (Senior Lecturer)
and Alex Murray (Research Assistant), all of whom are employed
within the Bartlett School of Construction and Project Management,
University College London (UCL). Graham and Andrew are experienced
academics in the fields of construction economics and project
management respectively. They have worked with Alex over the past
two years in developing comparative benchmarks of the operational
performance of PFI and non-PFI social infrastructure (schools
procurement methods offer different and inconsistent benefits
to the purchaser. The purchaser should consider what aspect they
place the greatest importance on achieving.
level to which risk can be transferred from the public to the
private sector depends on three main aspects:
1. extent to which specifications and standards
of the asset or service can be defined;
2. ability to put in place objective measurement
systems accurately to determine if contractual performance has
been achieved; and
3. private capital expenditure should be realised
before payment to maintain incentives for delivery.
date, based on limited public information, relatively few PFI
projects have failed. Without access to commercially confidential
information to ascertain project specific performance, this suggests
PFI is performing, in terms of delivery, certainly no worse than
publicly funded projects.
with non-PFI operated social infrastructure suggest PFI facilities
are performing comparably in terms of cost, with some higher levels
of performance and evidence of investment in the long term maintenance
of physical infrastructure.
1. What are the strengths and weaknesses of
different public procurement methods?
1.1 A full answer to this would exceed the word
limit. This submission will therefore only consider procurement
method involved in the provision of substantial assets, such as
buildings, civil engineering works or complex IT systems. A summary
of generally accepted strengths and weaknesses for a broad category
of procurement methods is presented:
Separate procurement for design, build, operation and maintenance contracts (D+B+O)
|||Choose separate tailored parties for|
|||Can take longer to deliver with opportunities for disagreement
||Opportunities for difference between bid prices and out-turn costs and delivery times
Single contract for design and build, separated from operation (D&B+O)
|||Reduced contracting so lower transaction costs
||||Reduced input from client in design process
|||Incentivises for communications between D & B to improve "buildability"
||||"Buildability" may be achieved at cost of operational performance
|Professional: outsourced procurement process as a form of cost plus procurement
||||Operational specifications guide design and build
||||Client retains significant final cost risk
|||Able to deliver novel projects quickly
|Fully integrated: Design, build, finance and operation wrapped up together into one contract eg PFI (DBFO)
||||Fully incentivised whole life factors between D, B & O
||||Inherently inflexible beyond contract close
|||Risk transfer, client does not pay if agreed service/availability is not delivered
||||Risk transfer can be costly
|||Access to private capital for public infrastructure
||||Private capital is more expensive than public
1.2 Construction clients suffer both hold-up and quality measurement
problems. No procurement method is best at mitigating both problems.
By making the constructor (and project company which it part owns)
bear the consequences of inferior quality (lower service payments,
higher maintenance and life cycle costs), PFI can claim to be
superior at solving the problem that some construction contracts
give the contractor a perverse incentive to "shade"
1.3 However PFI also makes the client particularly vulnerable,
potentially, to opportunistic pricing of post-contract changes
introduced by the client (the form the hold-up problem usually
takes in construction projects). It therefore will tend to be
a good choice for the public client in cases where there is little
risk of requiring changes, and where the quality giving attributes
of an asset are capable of objective measurement and can be linked
to contract payments.
1.4 PFI will also tend to be a good choice where the constructor
has strong reasons not to take advantage of client vulnerability
post contract signature, because they are trying to build a reputation
and thus be selected (out of many competitors) to bid for future
PFI projects. This is particularly relevant where the public client
has an active public programme and where the number of competitors
is fairly high.
2. If PFI debt had been on-balance sheet rather than off-balance
sheet would PFI projects have been used as much? How should PFI
deals be accounted for?
2.1 The authors of this submission are not qualified accountants
and do not reside in an accountancy orientated academic department.
We therefore consider ourselves not qualified to answer these
two questions. However, we are of the view that the fundamental
objective of the PFI is to achieve better value for money for
the delivery of services.
3. How far can risk really be transferred from the public
to the private sector?
3.1 A fundamental principle of PFI is the allocation of risk
to the party best placed to manage it. The types of risks and
duration over which risks are considered are generally service
orientated, wide-ranging and long term. This has the fundamental
effect of reframing the nature of the challenge to one that includes
whole life cost considerations and stable provision.
3.2 Three aspects are relevant in considering limits to the
kind of risks that can be transferred. The first is the extent
to which specifications of required services or standards can
be clearly defined in testable, measurable ways. It is this problem
that bedevilled the early and most complex IT projects (NIRS2)
and Ministry Of Defence equipment projects (FSTA). The second
is the ability to put in place objective performance measurement
systems. These are often associated with human assessment where
there is no objective test. Soft FM services (such as reception
services) present such a challenge in determining standards. It
is to be noted that this is a tractable problem as the highly
successful prison PFI projects have proved, where performance
standards are largely driven by measurements and metrics relating
to human performance. It is therefore an exercise of development
and refinement of appropriate metrics and measurement systems
that frame the questions of how far risks can be transferred.
The third aspect is that PFI projects realise capital expenditure
before public payments begin, so lenders are at risk (with no
government guarantees of debt, except in instances such as the
London Underground PPPs). In effect, the public sector relies
on the lenders to ensure that project sponsors perform.
3.3 The extent to which risk has and is being transferred
in present projects can be deduced from two streams of evidence.
The first is where either too much or the wrong type of risk has
been transferred, and the consequences are such that the project
fundamentally fails. The extreme would require the public sector
to take back the risks and responsibility for the project. Examples
include the Croydon Tramlink and the two major London underground
PPPs. We are not aware of a definitive national list of these
failed projects being currently available. Compiling such an official
list would be a valuable exercise so that further analysis of
any common factors can be undertaken.
3.4 The second stream of evidence to inform the limits of
risk transfer is the inverse of the first. Here the evidence base
is the population of projects that are operational. The definitive
source for this is the HM Treasury PFI signed projects list.
This only includes active projects (operational or in-construction).
Considering just the operational projects, 569 projects are recorded
as being in receipt of unitary charge payments (UCPs) in 2008-09
with a total value of £6.49 billion.
The respective figure for 2009-10 is 619 projects being in receipt
of UCPs with a total value of £7.37 billion. The net gain
in the number of projects and increase in UCPs value has factored
in the removal of instances of projects that have been taken back
to public sector responsibility, such as Croydon tramlink and
the two London Underground upgrade projects. The logical deduction
from this is that for the 619 projects in receipt of UCPs, risk
of the type and size considered "normal" (construction,
maintenance and life cycle cost risks) has been transferred and
is being managed by the private sector. The number of projects
and sum of UCPs, is therefore evidence of real risk being transferred.
3.5 The presumption for these operational PFI projects is
that UCPs are only being made after due assessment that the service
provided meets contractual obligations for delivery of acceptable
standards of service. Where such service standards or availability
are not, as specified, being provided, it has to be assumed that
penalty points are being incurred and payment deductions made.
Sufficient data to provide more detailed commentary on the performance
of individual projects is not publicly available for commercial
confidentiality reasons. We believe such data would be extremely
useful in providing a true evidence base for the consideration
of the extent to which risk has been transferred in such projects.
The presence of very few instances of project failure (not all
of which were due to issues of risk), coupled with the requirement
in such projects for transfer of design, construction and operational
risk, reinforces the evidence that significant risk has been transferred.
4. Are there particular kinds of risk which are particularly
appropriate for transfer through PFI deals, or particular projects
which are suited for PFI?
4.1 Others such as the National Audit Office have examined
and reported on the issue of design and construction risk performance
 The body of evidence
is that PFI has tended to handle this risk relatively well. There
is less objective evidence of the risks that lie within the operational
services provided by such PFI contracts. For the majority of asset-based
projects, these can be broadly considered as the risks associated
with soft (people centric) and hard (asset centric) facility management.
4.2 The authors, as research active staff within the Bartlett
School of Construction and Project Management at UCL, have developed
an ability to benchmark the operational cost and performance of
aspects of soft and hard services within core social infrastructures,
specifically hospitals and schools. This research has been driven
by our curiosity to look at the impact of the procurement method
on the operational performance of facilities and wider infrastructure.
4.3 We have developed comparisons derived from independently
collected data sources in a bid to provide objective evidence
to what has largely been a debate driven by perception. This is
an area that is poorly developed and served. This is a view held
by others, including Infrastructure UK which notes that data on
the cost and performance of infrastructure is "uncoordinated".
We believe that our endeavours have demonstrated that it is possible
to generate objective assessment of the cost and performance of
operational services and the assets that underpin them. It is
our recommendation that to help solve this issue of poor data
on public infrastructure, a national asset register be established
to record at the individual asset level, the key information on
the asset's creation, expected lifespan, and its operational performance
and associated costs. Such an undertaking is recognised as significant.
The Canadian Statistical Office collate such data setting a precedent
for what can be achieved.
4.4 Taking the example of what our research revealed in the
important area of hospital cleaning, the findings from our analyses
comparing new hospitals conclude that PFI facilities witness higher
levels of cleanliness and quality of patient environment, with
no associated statistically significant higher cost of cleaning
services. In fact,
the dispersion of cleaning costs seen within PFI facilities is
significantly lower than in non-PFI new hospitals. We consider
this provides procuring authorities with greater certainty on
the on going cost of operation. These findings are based on publicly
assessed and available data.
4.5 A forthcoming paper of ours,
looking into the operational services expenditure between schools
renewed via PFI and non-PFI, reveals PFI costs to be higher in
six of the nine years into operation. However, these differences
are not statistically significant and PFI schools cost less overall
in three of the nine years examined. What can also be observed
also is PFI facilities are allocating funds to invest in building
maintenance with considerable increases in hard FM expenditure
following the fifth year of operation. The corresponding expenditure
in public funded renewed schools remains roughly flat throughout.
The issue of responsible stewardship of infrastructure is as relevant
today as it was when PFI was developed in the early nineties.
Emerging pressures to cut these budgets in PFI facilities is a
case of falling back into old habits. This is an area at the heart
of the recent James Review.
4.6 The main risk that cannot be transferred is the reputational
risk that is inherent with any public service. The general public
will not, and indeed should not, seek to hold accountable any
party other than the public sector for any major failing in a
public service that is then made the subject of a PFI/PPP contract.
However, the alternative is also true, with the public sector
enjoying the reputational benefits from improved services from
4.7 Those projects where clearly understood fixed assets are
at the centre of the service delivery are well suited for the
transfer of risk for the successful completion of the design,
build, financing and operation of the asset. Examples of these
include schools, hospitals, residential and office accommodation
buildings that make up much of the present PFI/PPP market. Additionally,
there are civil engineering dominated projects that form parts
of the road, rail, water, and energy sectors.
5. What state guarantees are explicit or implicit in PFI
5.1 We are not aware of any general area where explicit guarantees
are made in PFI projects, but would defer to those legal practitioners
in the area of PFI contract law for more detailed comment.
5.2 As we understand the question, the foremost implicit guarantee
is of persistence of presence. Without such surety, the private
sector would place higher prices on the risks of both "walk-away"
and or "hold-up". It is the implicit trust in the State,
its government and its agencies that they will endure and be obligated
by their pre-existing contractual obligations that provides the
comfort needed by the private sector and its financiers. The prevalence
of availability payments means that the private sector is guaranteed
the equivalent of off-take agreements. Whether the service made
available is actually needed (so for example a school building
that is open, but for which there are no pupils to require teaching)
is a concern for the wider society and tax-payer.
6. In what circumstances are PFI deals suitable for delivery
6.1 There must always be a strong service element in any PFI
as otherwise it is just a form of lease or affermage. The circumstances
where the service is best suited for PFI are those where there
is a combination of clarity of the output to be specified and
established and agreed protocols and metrics for assessing the
performance of the service. This set of criteria is most easily
accomplished in projects dominated by the technical performance
of a relatively stable assetsuch as road or simple building.
These represent clearly understood entities where not only is
the "what is required" clearly defined, but so
is the "way it will be measured". This leads
to a set of limits where if there is too much novelty, complexity
or opportunity for principled disagreement, PFI is not suitable.
6.2 Where straight forward objectively measurable service
outputs can be specified, risk of delivery can be priced and wrapped
into an integrated contract. Where complex subjective service
outcomes are present, risk of delivery will be impossible to price
accurately, so will likely include margins of safety which render
them not value for money. The recent decision to incentivise the
Doncaster prison service provider for the reduction of reoffending
rates can be considered as objectively measurable, except it is
not a service output (directly within control of the provider),
rather an outcome (a result of interaction between service outputs
and service users, such as prisoners). In this case, the service
provider is not taking the risk of deductions from increased re-offending
Ive, G and Chang, C-Y (2007) The principle of inconsistent trinity in the selection of procurement systems. Construction Management and Economics, Vol
25, pp 677-690. Back
Updated in March 2011 and available at: www.hm-treasury.gov.uk/ppp_pfi_stats Back
Un-indexed values used. Back
Performance of PFI Construction, National Audit Office, 2009. Back
PFI: Construction Performance, NAO, 2003. Back
Point 3.36, National Infrastructure Plan, Infrastructure UK. Back
Age of Public Infrastructure: A Provincial Perspective, by Mychèle
Gagnon, Valérie Gaudreault and Donald Overton, Investment
and Capital Stock Division, Statistics Canada, 2009. Back
G Ive, A Murray, A Edkins and K Rintala, Cost and performance
comparison of PFI and non-PFI Healthcare Infrastructure in England,
presented at the 3rd Annual Conference of the Health and Care
Infrastructure Research and Innovation Centre, Edinburgh, September,
Finding of initial analyses to be presented at the 2011 RICS COBRA
conference. Paper to be submitted early June. Back
Review of Education Capital, Sebastian James, April 2011. Back