The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1030-9616.htm

ARJ
34,6 The public-private partnership
valuation paradox
Stephen Gray
University of Queensland, Brisbane, Australia
546
Jason Hall
Received 26 August 2020 Ross School of Business, University of Michigan, Ann Arbor, Michigan, USA
Revised 21 December 2020
Accepted 2 March 2021 Grant Pollard
Ergon Energy, Townsville, Australia, and
Damien Cannavan
Illinois State University, Normal, Illinois, USA

Abstract
Purpose – In the context of public-private partnerships (PPPs), it has been argued that the standard
valuation framework produces a paradox whereby government appears to be made better off by taking on
more systematic risk. This has led to a range of approaches being applied in practice, none of which are
consistent with the standard valuation approach. The purpose of this paper is to demonstrate that these
approaches are flawed and unnecessary.
Design/methodology/approach – The authors step through the proposed alternative valuation
approaches and demonstrate their inconsistencies and illogical outcomes, using theory, logic and
mathematical proof.
Findings – In this paper, the authors demonstrate that the proposed (alternative) approaches suffer from
internal inconsistencies and produce illogical outcomes in some cases. The authors also show that there is no
problem with the current accepted theory and that the apparent paradox is not the result of a deficiency in the
current theory but is rather caused by its misapplication in practice. In particular, the authors show that the
systematic risk of cash flows is frequently mis-estimated, and the correction of this error solves the apparent
paradox.
Practical implications – Over the past 20 years, PPP activity around the globe amounts to many billions
of dollars. Decisions on major infrastructure funding are of enormous social and economic importance.
Originality/value – To the best of the authors’ knowledge, this study is the first to demonstrate the flaws
and internal inconsistencies with proposed valuation framework alternatives for the purposes of evaluating
PPPs.
Keywords Government, Infrastructure, Public-private partnership, Valuation
Paper type Technical paper

1. Introduction
In recent years, many national and provincial governments have entered into public-private
partnerships (PPPs) with the private sector. A PPP is a contractual arrangement between
the government and the private sector, usually for the delivery of a piece of social
infrastructure or social service. Generally, the term PPP is used when the private sector also
Accounting Research Journal
Vol. 34 No. 6, 2021
finances the project. For example, the contract may be for the design, construction, finance
pp. 546-579
© Emerald Publishing Limited
1030-9616
DOI 10.1108/ARJ-08-2020-0275 JEL classification – H43, H54
and operation of a hospital, road or rail link. It might also be for the provision of school Partnership
classrooms or pathology services over a fixed period of 20 years. These contractual valuation
arrangements between government and the private sector have different names in different
paradox
jurisdictions and each contract has its own unique characteristics, but there are a number of
features that are common in such agreements.
Firstly, the contract usually involves a private sector consortium that includes one or
more banks (involved in financing and structuring), engineering and construction firms 547
(involved in design and construction) and an operations firm (who is responsible for
operations, billing or revenue collection and maintenance).
Secondly, the contract usually requires the government to make a series of payments to
the private sector consortium. It is common for the private sector consortium to own the
asset for a concession period, of say 30 years, at which time the infrastructure is generally
transferred to the government. Consequently, this is sometimes seen as a form of off-balance
sheet financing for the government – even though the government is contractually
committed to make the series of agreed payments over the concession period.
Thirdly, the contract usually involves some sharing of risk between the government and
the private sector. It is common for the private sector consortium to bear some form of
demand risk, for example, uncertainty over the volume of traffic that uses a new toll road or
construction risk, such as the risk that a sub-contractor will fail and will have to be replaced
causing a delay. The government will always bear at least residual delivery risk – if the
consortium fails and is unable to deliver on its contractual commitments, the government
will likely have to step in to ensure the delivery of the particular piece of infrastructure or
social service.
Some examples of the many billions of dollars of public assets that are subject to PPP
arrangements are as follows:

The LaGuardia Airport Central Terminal B in New York, USA. The New York
Government and the Port Authority of New York and New Jersey selected a private
sector consortium to redevelop and operate the terminal. The deal was valued at
$5.1bn and is one of the largest PPPs in the history of the USA. The private sector
consortium has a 35-year concession period.

The Central Java Power Plant project in Batang Regency, Indonesia, valued at US
$4bn, was developed by a private consortium under a Build-Operate-Transfer
scheme with a 25-year concession period.

Royal North Shore Hospital and Community Health Services Redevelopment Project
in Sydney, Australia valued at $1.125bn, involved redeveloping the hospital and
running and operating all “soft services” within the hospital over a 28-year
concession period.

A large number of schools and hospitals in the UK have been structured as PPPs.
Some hospitals have been designed and constructed by private sector consortiums
in return for government subsidies. Some school PPPs have involved land being
sold to developers with school grounds then being leased back to the government.

The central focus of this paper is on the financial evaluation of PPP arrangements. The
process begins with the government determining that the particular piece of infrastructure
or social service should be delivered. This decision is made as a matter of public policy and
is not modelled in this paper [1]. Rather, in this paper, we examine an important aspect of the
analysis of how the project should be procured. Specifically, we consider the computation of
ARJ the discount rate that should be applied when performing financial analyses of different PPP
34,6 and government procurement strategies.
Broadly, there are two procurement methods – government financed procurement and a
PPP arrangement with the private sector. Government financed procurement involves (on
the balance sheet) government financing of the project. This includes traditional
government procurement, where government effectively bears the majority of the risks (and
548 generally all of the systematic risk) of the project. It also includes alternative government
financed procurement options including alliance contracting, design and construction
contracts together with either short or long-term operating agreements where the risks of
the project can be allocated between the government and its counterparties. The selected
government financed procurement option is known as the “public sector comparator” or
PSC.
A PPP requires the government to commit to make a series of future cash flows to the
private sector consortium and generally involves a greater transfer of risk by the
government to the private sector. It is the role of the government to determine which of
the two options, PSC or PPP, provides the best value for money to taxpayers.
It is the risk-sharing aspect of PPPs that is of central importance to proper financial
evaluation. In particular, the task confronting the government is to evaluate the costs and
the risks of the PPP against the PSC. Corporate finance and valuation theory, based on the
capital asset pricing model (CAPM), would suggest that this evaluation can be performed by
first setting out the expected cash flows of each alternative and then discounting them at a
rate that properly reflects their systematic risk. That is, the techniques that have been
adopted as the standard in other areas of valuation practice should apply equally well to the
evaluation of PPPs. Ariel (1998), for example, demonstrates why negative expected cash
flows can still be treated appropriately by standard CAPM valuation approaches.
By contrast, a range of different practices for evaluating PPPs are used by different
national and provincial governments around the world, and none of these are consistent
with corporate finance and valuation theory or with the standard valuation practice that is
used in commercial settings [2].
The reason that is usually given for the ad hoc or “modified” approaches that are adopted
in practice is that there is some sort of paradox when evaluating PPPs against the PSC.
Specifically, it is argued that standard corporate finance and valuation theory must be
rejected (or at least modified) because it implies that the government can be made worse off
by transferring systematic risk to the private sector. As this cannot be true, it is argued, the
standard theory and practice cannot be applied to the evaluation of PPPs.
After a brief review of the literature in Section 2, in Section 3, we set out the standard
valuation theory and practice and show by way of example where the paradox is thought to
arise. Section 4 then sets out the approaches that various jurisdictions have used to “solve”
the paradox and to evaluate PPPs against standard government procurement.
In Section 5, we set out a number of problems with the approaches that have been
proposed, and that are currently being used in practice. We show that whilst the
modification of the accepted theory appears to solve the apparent paradox, it causes a
number of internal inconsistencies and can produce illogical outcomes.
In Section 6, we set out our proposed approach, which is perfectly consistent with
accepted theory and involves discounting expected future cash flows at a rate that properly
reflects their risk. We show that the apparent paradox is not the result of a deficiency in the
current theory but rather is caused by its misapplication in practice. In particular, we show
that the systematic risk of the PSC and PPP cash flows is frequently mis-estimated, and the
correction of this error solves the apparent paradox. Moreover, our proposed approach does
not have the unintended consequences or side effects of the approaches that are currently Partnership
used in practice. valuation
Section 7 contains our conclusions, and the formal mathematical derivations that
underpin our proposed approach are set out in detail in the Online Appendix.
paradox

2. Literature review
A vast literature has developed internationally around PPPs, with a notable increase in
production over the past decade or so. Numerous reviews of the literature have been
549
conducted (see, for example, Cui et al., 2018; Ma et al., 2019; Petersen, 2019; Hodge and Greve,
2017). Due to their multi-faceted nature and broad applicability, PPPs have been analyzed
and documented in a wide variety of ways. Extant studies range across topics such as
operational performance (Saeed et al., 2018), governance (Benítez-Ávila et al., 2019), risk
management (Nisar, 2007), frameworks (Liu et al., 2015a) and project financing (Tang and
Shen, 2013).
The “success” or otherwise of PPPs has been studied by numerous authors. This
involves first designating what is meant by success. Greve and Hodge (2013) discuss
numerous aspects of this issue comprehensively. The use of critical success factors (or
“CSFs”) has emerged as a clear focus for many (Liu et al., 2015b; Zhang, 2005; Ismail, 2013).
One of the key criteria is the concept of “value for money” (Grimsey and Lewis, 2005; Burger
and Hawkesworth, 2011). In some recent work, Iossa and Martimort (2015) and Martimort
and Straub (2016) show that well-designed PPPs may increase efficiency in public service
delivery, though they may be too expensive under high levels of uncertainty and/or when
the projects are highly sophisticated.
One distinct line of research has been that of survey evidence (see, for example, Hu et al.,
2014; Cheung et al., 2009; Cheung et al., 2012). In contrast, another offers theoretical
modelling on mechanism design and contract theory. For example, recent work by Buso
et al. (2020) develops a principal-agent model of complete contracting to resolve the moral
hazard issue. Wang and Liu (2015) also use a principal-agent model and integrate it with a
fairness preference theory to generate optimal incentive mechanisms. Yet another strand of
the literature looks at the use of the real options methodology (see, for example, Alonso-
Conde et al., 2007; Martins et al., 2015). Recent work by Esmaeeli and Heravi (2019), Liu et al.,
2017 and Jin and Liu (2020) continue to develop this area of research.
Risk allocation is one broad category related to our paper. Studies such as those of Jin
(2010). Wang and Liu (2015), Loosemore and Cheung (2015) and Ke et al. (2010) deal with the
topic in various ways.The closest related literature to our paper relates to work relating to
the use of discounted cash flow (DCF) models applying the CAPM (Hellowell and Vecchi,
2018). Whilst not directly related to our paper, studies by Gao et al. (2017), Hellowell and
Vecchi (2018) and Whitfield (2017), which provide evidence that rates of return are high for
private companies participating in PPPs, are of note.
Despite much research, there is no literature on how current approaches to valuation
using the CAPM to discount cash flows in practice suffer from internal inconsistencies and
produce illogical outcomes in some cases. We fill a gap in the literature by showing that the
systematic risk of cash flows is frequently mis-estimated in practice and demonstrate how to
correct the error.

3. Do we need a unique framework for assessing public-private partnerships?

3.1 Focus on discount rates
When comparing a PPP with traditional government procurement, there are two aspects to
consider:
ARJ (1) the cash flows relating to each alternative; and
34,6 (2) the discount rate that should be applied to each set of cash flows.

In relation to the cash flows, a number of papers report that cost overruns and time delays
are more common for government projects. For example, Malone (2005, p. 422) concludes
that “[t]here is a recognition that large public sector infrastructure projects have historically
550 been delivered with large time and cost overruns.” In this paper, we set aside issues relating
to cost management efficiencies and cash flows more generally and focuses on the discount
rate that should be applied to a series of expected cash flows.
Our focus is on the evaluation of different cash flow streams from the perspective of the
government, and specifically on the discount rate that should be applied to each potential
series of cash flows. Our main point is that government should apply a discount rate that
properly reflects the systematic risk of the cash flows being evaluated. Our objective is to
show how government should compute the present value of the series of cash flows to be
provided by the government. We note that this is independent of the tax treatment of those
cash flows in the hands of the recipient.

3.2 Standard valuation practice

Standard DCF valuation practice is to take a set of expected cash flows and to discount them
back to present value using a discount rate that properly reflects the risk of those cash flows.
Where there are two or more mutually exclusive alternatives, the standard approach is
applied to each alternative and the alternatives can be ranked in terms of their present
values. The different alternatives will almost certainly involve different expected cash flows
and if the risk of the cash flows differs between alternatives different discount rates would
be used, commensurate with their risk.
In this section, we set out the reasons that have been used to justify the need for a unique
approach for evaluating PPPs. In essence, the proposed reasons are based on the notion that
standard DCF valuation practice works well for all other projects but leads to implausible
outcomes, unintended consequences and a paradox when applied to PPPs.

3.3 Discounted cash flow valuation

Traditionally, finance practitioners have used DCF analysis, grounded in corporate finance
and valuation theory, to value project proposals. This methodology requires the estimation
of the project’s expected cash flow stream which is then discounted to present value using a
risk-adjusted discount rate, computed as:

X
T
E ½CFt 
Value ¼
t¼1
ð1 þ rp Þt

where:
E[CFt] = expected cash flow of the project in year t; and
rp = is the project’s required return on capital.
The project’s required return on capital is usually derived from the CAPM in which the
required return is the sum of the risk-free rate of interest and compensation for bearing
systematic risk:
 rp ¼ rf þ b a  MRP Partnership
where: valuation
rf = is the risk-free rate of interest; paradox
b a = is the asset beta, which is a measure of the systematic risk of the project [3]; and
MRP = is the market risk premium, which is the expected return on the market portfolio
of all risky assets relative to the risk-free rate.
551
Under standard DCF valuation, the expected cash flows of a project are discounted at a rate
that reflects the risk of those cash flows. The sum of the present values of the cash flows is
called the net present value (NPV). The NPV is an estimate of the value of the project as a
lump sum in today’s dollars that is equivalent (in value) to the future cash flow stream that
the project is expected to generate.

3.4 Negative net present value projects

Whilst private sector commercial projects tend to have positive NPVs (that is, they create
value for owners) PPP projects most commonly have negative NPVs. That is, the present
value of the cash flows is negative, and the project can only proceed if it is subsidised by a
government who decides that it should proceed for reasons other than the stand-alone
financial viability of the project. That is, in these cases, the government must either:

undertake the project itself and bear all the associated risks and costs; or

invite a private sector party to undertake the project via a PPP with the government
providing some compensation to offset the losses that would otherwise be incurred
by the private investor.

From the government’s perspective, the most beneficial proposal is the one which results in the
lowest negative NPV, and thus requires the lowest present value commitment by the government [4].

3.5 Anomalous outcomes for negative net present value projects?

The primary argument for the need for a special valuation approach for evaluating PPPs is
that the standard DCF approach leads to a paradox when applied to negative NPV projects.
As PPPs tend to have negative NPVs, the argument is that a special evaluation approach is
required. The reasoning behind this argument can be best explained via a simple example.
Suppose that the project is to build a new rail link with a construction cost of $100m.
Over the 5-year concession period, the allowed passenger charge is insufficient to cover
operating expenses, with a loss of $40m per year anticipated [5]. That is, the project’s
expected cash flows, from the perspective of the government, are as follows:

Year 0 1 2 3 4 5

Cash flow 100 40 40 40 40 40

Assuming that the risk-free rate is 4%, the MRP is 6% and the asset beta is 0.5 (inferred
from an analysis of listed firms), the project’s required return on capital is 7% (derived from
the CAPM). By applying this discount rate to the above cash flows, the NPV of the project
is –$264m. That is, to construct the rail link and operate it for five years, the government
would expect to pay the series of cash flows set out above, and this series of cash flows has
an equivalent lump sum present value of –$264m.
ARJ Now suppose that something can be done to alter the project risks so that the expected
34,6 cash flows remain unchanged, but the risk to the government is increased. That is, the
operating losses are still expected to be $40m per year, but the range of possible outcomes is
greatly increased – previously the operating losses might be a little above or a little below
$40m, but now they might be greatly above or below $40m. Now the government is exposed
to more risk, but all other things are equal. The new riskier project is clearly inferior to the
552 otherwise identical, but much less risky, the project as the expected cash flows are the same
but the government is now bearing more risk.
However, if we assume that the systematic risk of the project increases such that the
asset beta is now equal to 1.0 (that is, the relevant measure of risk doubles), the riskier
project with the same expected cash flows would have an NPV of –$252m, applying a
discount rate of 10%, commensurate with the increase in risk.
If the government were to make its investment decision purely on the basis of which
project had the lowest negative NPV, the riskier project would be favoured. However, this
outcome appears to be inconsistent with the intuitive view that the government should
prefer the project that is less risky, but otherwise identical in all respects. The reason for this
inconsistency is that a higher discount rate results in a lower NPV for the future negative
cash flows. This outcome is seen as a paradox that results from a deficiency in the accepted
theory, at least insofar as it applies to the evaluation of PPPs.
To avoid this paradox and apparently incongruous results, different governments have
adopted different methodologies to rank PPP proposals and to compare them with PSC
alternatives – as set out in the following section. In the subsequent discussion, we
demonstrate that the apparent paradox presented above does not present an issue to be
solved and that standard valuation theory does in fact generate sensible project evaluations
for negative NPV projects.

4. Approaches used in practice

In this section, we set out the approaches that various governments have used to “solve” the
apparent paradox and to evaluate PPPs against standard government procurement. For
example, the UK Government has determined that all cash flows, whether under a PPP or
government procurement and regardless of the systematic risk of the cash flows, are to be
discounted at the same constant social time preference discount rate. By contrast, the
approach adopted in Australia and Canada (British Columbia) has been to retain as many
features of the accepted corporate finance and valuation theory as possible but to make
adjustments or modifications to it to avoid the apparent paradox discussed in Section 2.
It is generally recognised that these modifications to the standard framework are ad hoc
and not supported by theory – they are designed only to avoid the apparent paradox. For
example, Partnerships Victoria (2003, p. 23), from the Treasury of the Australian state of
Victoria, argues that its modified approach will preserve the correct ranking amongst
alternatives even though the actual final numbers derived “will have no direct meaning.”

4.1 US approach
The USA does not currently have a unified approach or framework for evaluating PPPs.
Historically, PPP deals have been less common in the USA than in other jurisdictions,
although there are signs of change [6]. Nevertheless, the USA has been lagging and “many
states lack the technical capacity to consider such deals and fully protect the public interest”
(Istrate and Puentes, 2011) [7].
 In short, the assessment of PPPs is only now becoming an important issue in the USA, Partnership
and the approaches and frameworks that are used in other jurisdictions appear to be of valuation
interest to government agencies and policymakers.
paradox
4.2 Single discount rate approach
The UK Treasury recommends that all PPP-type projects should be evaluated using a single
discount rate. Under this approach, all expected cash flows are discounted at the same rate 553
regardless of whether they are made under a PPP or PSC arrangement. Consequently, no
assessment of the risk of a particular series of cash flows is required – the same discount
rate will be applied regardless.
The UK approach is set out in the UK Treasury’s “Green Book”. The single discount rate
applied to all projects is based on an estimate of the social time preference rate [8].
This approach technically avoids the paradox of government being made better off by
taking on more risk. Under the single discount rate approach, the present value of a
particular set of cash flows is completely independent of the risk of those cash flows.
Consequently, an increase or decrease in systematic risk will have no impact at all on the
estimated present value of a set of cash flows.
This approach is inconsistent with the notion that the present value of a future cash flow
depends on its systematic risk. Moreover, it leads to its own set of illogical and unintended
consequences. Suppose, for example, that the proposed PPP had expected cash flows that,
year-by-year, were slightly lower than the expected cash flows under the PSC, but that the
PPP cash flows involved dramatically higher risk to the government. The single discount
rate approach would rank the PPP ahead of the PSC, even though this would clearly be a
detrimental outcome for taxpayers.

4.3 Private sector discount rate corporate finance approach

In some jurisdictions, the private sector cost of capital is used as the appropriate discount
rate for all procurement decisions. Under this approach, the private sector cost of capital is
estimated as it would be in a standard corporate finance setting. This involves the
estimation of the systematic risk (that is, beta) for the particular type of project and the use
of the CAPM to estimate the discount rate. This same discount rate, reflective of the usual
overall systematic risk of that kind of project, is applied to all cash flow streams under both
PSC and PPP alternatives. For example, the British Columbia Government has used a
market-based discount rate [9], being an estimate of the private sector weighted average cost
of capital for a project of a similar type, for the evaluation of PPP proposals and the PSC.
This approach is also inconsistent with the standard valuation framework that is based
on the CAPM, which requires that all cash flows be discounted at a rate that properly
reflects their systematic risk. A single discount rate applied to all alternatives does not
reflect the amount of systematic risk that may have been transferred to the private sector
under a PPP arrangement.
In relation to risk transfer, we note that there are no constraints on the risk-sharing
arrangements that might be agreed upon by the government and the private sector. The
parties are free to structure arrangements in any way. We note that some of the risks that
might be shared are systematic in nature (e.g. traffic volumes are likely to be correlated with
economic conditions) or non-systematic (e.g. natural disaster risk). Non-systematic risk is
typically dealt with in the cash flows (e.g. via insurance costs), whereas the transfer of
systematic risk requires an adjustment to the discount rate. The latter case is the focus of the
analysis below.
ARJ We show below that computation of the appropriate discount rate requires a proper
34,6 assessment of the systematic risk that is retained by the government. Which components of
systematic risk might be transferred or who they might be transferred to are secondary
considerations. For example, one does not need to know that the risk of construction
overruns has been transferred to a construction partner and that traffic volume risk has
been transferred to an operations partner – one simply needs to know the quantum of
554 systematic risk that remains with the government. Then that depends on the correlation
between the cash flows that the government is required to make [10] and the state of the
economy/market. The systematic risk retained by the government could be computed via
scenario modelling or via a bottom-up allocation approach as described in the Australian
PPP Guidelines. Our analysis below takes this as given and shows how the resulting
systematic risk flows through to a proper computation of discount rates.

4.4 Modified capital asset pricing model approach

The approach that is closest to being consistent with the standard CAPM valuation
framework is the approach that is currently adopted in Australia [11]. This approach
focusses on capturing differential allocation of systematic risk as the driver of different
discount rates between project options. Under this approach, there is an attempt to recognise
the sharing of systematic risk between the parties. Under different arrangements, different
parties will bear different amounts of systematic risk. As systematic risk varies, so too
should the discount rate that is applied. The systematic risk transfer approach follows a
step-by-step process to quantifying the amount of systematic risk involved in a project and
the way it is divided amongst the parties.
The vast majority of PPP arrangements involve social infrastructure projects that
require some form of subsidy from the government to make them viable for the private
sector. The modified CAPM approach deems such cases as “net cost projects”, defined to be
projects where the sum of the cash flows (from the perspective of government) is negative.
For a net cost project:

The PSC cash flows are discounted at the risk-free rate; and

The PPP cash flows are discounted at a rate that reflects the amount of systematic
risk transferred to the private sector.

The modified CAPM approach applies different discount rates to net cost projects (where the
sum of the cash flows is negative from the perspective of the government) and a net revenue
project (where the sum of the cash flows is positive from the perspective of the government).
This assumes that in relation to net revenue projects structured as PPPs government will
bear no systematic risk.
A net revenue project operated solely by the government is simply a standard project.
Cash flows, in this case, should be discounted at the usual CAPM-based project rate. The
discount rates to be applied under the modified CAPM approach are summarised in Table 1
below, in which assumptions are incorporated as to the level of the risk-free rate (5%) and
the project’s systematic risk premium (3%).
To illustrate the application of the modified CAPM approach, we consider an example
based on the following data:

The risk-free rate is 5% p.a. This would be estimated as the yield on long-term
government bonds;

The total systematic risk of the project is 3%. This is estimated within a CAPM
framework where the beta is estimated with reference to a set of exchange-listed
 firms (for example, beta estimate of 0.5) and the MRP is estimated with reference to Partnership
historical stock and government bond returns or from equity prices and earnings valuation
forecasts (for example, MRP of 6%). This step is performed in the same way as for
any standard CAPM estimate of a required return;
paradox

Under the proposed PPP, two-thirds of the systematic risk is transferred to the
private sector, making the PPP discount rate 7% (that is, 5% plus two-thirds of 3%)
under the modified CAPM approach; and 555

The cash flows under the PPP and PSC alternatives are as set out in Table 2 below.
Under the PSC there is a 3-year construction phase followed by a 30-year operations
phase. Under the PPP, the government makes annual concession payments over the
30-year operations phase of the project.

Under the modified CAPM, the PSC cash flows are discounted at the risk-free rate (5%)
and the PPP cash flows are discounted at a rate that reflects the systematic risk transferred
to the private sector (7%). The present value calculations set out in Table 3 below indicate
that the PPP alternative would be preferred in this case.

5. Problems with current approaches

5.1 There is no advantage from government cost of capital
It is sometimes argued that government procurement has a natural advantage as the
government’s cost of funds, as measured by the government bond rate, will be lower than
the private sector cost of capital. Such arguments have been rejected in favour of the view
that the appropriate discount rate to be applied to a series of future cash flows depends upon
the systematic or factor risk of those future cash flows, and that the identity of the owner of

PPP
Govt. bears Govt. transfers two-thirds Govt. bears
PSC all risk of systematic risk no risk Table 1.
Net cost project (sum of Risk-free Risk-free Risk-free rate plus two-thirds of Project The discount rate
cash flows is negative) rate (5%) rate (5%) systematic risk premium (7%) rate (8%) approach under the
Net revenue project (sum Project rate n/a n/a Risk-free modified CAPM
of cash flows is positive) (8%) rate (5%) approach

Year 0 1 2 3 4 5 ... 33
Table 2.
PSC cash flow 100 100 100 10 10 ... 10 Net cash flos flows
PPP cash flow 35 35 ... 35 government

Year 0 1 2 3 4 5 ... 33

PSC cash flow 100 100 100 10 10 ... 10 Table 3.
NPV @ 5% 405 NPV calculations
PPP cash flow 35 35 ... 35 under the modified
NPV @ 7% 355 CAPM approach
ARJ those cash flows is irrelevant [12]. Consequently, the approach of discounting all cash flows
34,6 at the government cost of borrowing is inconsistent with the standard notion that there is a
relationship between the risk of a set of cash flows and the required return in relation to
them.

5.2 Attempts to align discount rates to systematic risk

556 Consistent with the existing literature, we reject approaches that apply the same discount
rate to cash flow streams that have demonstrably different systematic risks. Consequently,
our focus in the remainder of this section is on the modified CAPM approach, which seeks to
incorporate the effect of systematic risk transfer and to apply a discount rate that reflects the
systematic risk of the relevant cash flows.
In the remainder of this section, we set out a number of shortcomings of the modified
CAPM approach. We show that:

a guaranteed cash flow from the government is effectively a government bond and
should be valued accordingly, but the modified CAPM approach does not do so;

perfectly offsetting cash flows provide no net benefit to the government, but the
modified CAPM approach can lead to different conclusions; and

in some circumstances, the modified CAPM approach can lead to the conclusion that
the government is made worse off by an unambiguous improvement in the cash
flows.

The alternative approach that we propose in the following section does not suffer from any
of these problems, is based on the proper application of standard valuation techniques and
provides valuation estimates that are directly interpretable as the present value of a
particular stream of cash flows to the government.

5.3 Guaranteed cash flows from government should be valued as a government bond
Consider a PPP under which government agrees to make a fixed payment of exactly $40m
per year for five years to the private sector partner, ignoring any abatement mechanism [13].
Under this arrangement, the government is bearing no risk because the payments from the
government are fixed. It is the private sector partner that bears all of the demand and
inflation risk. Consequently, the modified CAPM approach would require that these cash
flows be discounted at the project rate, 7% in our earlier example, giving a present value of
$164m.
However, this series of cash flows is effectively a government bond – it is a series of fixed
payments to be made by the government at specified times. There are five “coupon”
payments of $40m each. Because this is a government bond, it must be valued as a
government bond. The CAPM requires that the cash flows must be discounted at the
relevant government bond rate, 4% in our earlier example, giving a present value of $178m.
That is, the government could raise $178m by issuing bonds that were backed by the
promise to make a payment of $40m per year for five years, and would record this amount
as a liability in its accounts – that being the present value of its commitment to making these
future cash flows. However, if exactly the same set of cash flows is to be made in a PPP
context, the modified CAPM approach requires them to be discounted at 7%, in which case
the present value is only $164m. This clearly understates the value of the government’s
commitment to make future cash flows, which can be an important consideration to the
extent that this figure flows through to government accounts.
 Moreover, the modified CAPM approach can also lead to the government being Partnership
arbitraged by the private sector to the extent the private sector can repackage the cash flows valuation
for a profit. Suppose a government did follow the modified CAPM approach and estimated
the present value of this negative $40m annuity to be –$164m. If, in return, the project
paradox
offered services or infrastructure (that is, social benefits or economic externalities) that the
government valued at $165m and which cost the private sector $165m to supply, this would
appear to be a good arrangement from the perspective of government [14]. However, to
properly consider whether or not this is a favourable deal for the government, consider what 557
would happen if the proposal proceeds. The private partner could immediately securitize the
guaranteed payment series (of $40m per year) from the government and sell it off as
(effectively) a government bond. The market would pay $178m for these fixed payments
from the government, valuing the stream of fixed payments as a government bond as set out
above and the private sector partner would supply the infrastructure or services for $165m
and would pocket the difference.
An example of such arbitrage is the Angel Train case in the late 1990s. Nomura and its
partners purchased the right to receive a series of essentially government guaranteed
payments; the seller discounting the scheduled payments at a rate that reflected the risk of
operating a train network. Nomura then securitised the payments and sold them as an
effectively government-guaranteed series of payments, booking a £440m profit on a £700m
outlay (Smith, 1999).

5.4 Perfectly offsetting cash flow streams provide no net benefit

Consider two local councils that sit adjacent to one another and are otherwise identical in all
respects. Each spends $40 per year on the maintenance of parks and gardens. Now suppose
that each council contracts with the other to provide the maintenance service for $40 per
year over five years. That is, Council A will maintain the parks and gardens of Council B
and vice versa. Clearly, neither council is made better or worse off by this arrangement –
they both continue to pay $40 per year and they continue to have their parks and gardens
maintained.
The arrangement by which Council A contracts with an external provider (Council B) to
provide a service at a prescribed standard for a fixed period of time is in nature a PPP. As
Council A is bearing no risk (the payment of $40 per year is fixed and does not vary with
demand, inflation or any other economic variable) the modified CAPM approach requires
that the cash flows must be discounted at the project rate, 7% in our earlier example. In this
case, the present value of the cash flows, from the perspective of Council A, is –$164.
Next, the arrangement by which Council A agrees to receive $40 per year in exchange for
performing maintenance services for Council B is a net revenue project from the perspective
of Council A. The modified CAPM approach requires these cash flows to be discounted at
the risk-free rate, 4% in our earlier example, giving a present value of þ$178.
That is, the modified CAPM approach suggests that Council A is made $14 better off by
this arrangement. Moreover, Council B will perform the same exercise from its perspective
and will conclude that it is also $14 better off. Then this approach will also conclude that
both Councils can increase the amount of value they create by charging each other a higher
amount each year. In reality, of course, such an arrangement creates no value at all beyond
having the respective parks and gardens maintained at the cost of $40 per year.
In such a transparent case as this example, it is likely that common sense would prevent
the modified CAPM approach from being applied mechanically and that a more sensible
conclusion would be reached. However, the example does illustrate that there is an
inconsistency in the way that cash inflows and cash outflows are evaluated. In a large state
ARJ government, for example, there may be a diverse mix of net cost and net revenue projects
34,6 across a range of portfolios. The modified CAPM approach would indicate a positive net
benefit even though, from a whole of government perspective, there are cash inflows and
outflows that cancel each other out.

5.5 Inconsistent evaluation of net revenue and net cost projects

558 The modified CAPM approach requires that, for the PSC, cash flows for net cost projects are
discounted at the risk-free rate, whereas cash flows for net revenue projects are discounted
at the project rate. This arbitrary distinction can lead to projects with similar cash flows and
similar risks having significantly different NPVs.
Consider the following project cash flows from the perspective of the government under
the PSC. In this case, the project loses money and requires net cash outflows early in its life,
but there is some fee for service and demand is expected to grow over the life of the project to
the extent that net cash flows in later years are expected to be positive.

Year 1 2 3 4 5

Cash flow 75 40 5 30 87

The sum of these cash flows is –$3m, meaning this project is deemed to be a net cost project.
As such, the modified CAPM approach requires that these cash flows be discounted at the
risk-free rate (4%) to give a present value of –$16.4m.
Now suppose that the expected cash flows remain the same except that the last cash flow
at Time 5 is increased to þ$93m from þ$87m. This unambiguously improves the project
and should make it more attractive in any financial evaluation.
The sum of the expected cash flows is now þ$3m, meaning this project is deemed to be a net

Year 1 2 3 4 5

Cash flow 75 40 5 30 93

revenue project. As such, the modified CAPM approach requires that the cash flows be
discounted at the project rate (7%), giving a present value of –$19.9m.
That is, under the modified CAPM approach an unambiguous improvement in the
expected cash flows appears to have made the government worse off. This stems from the
fact that there is a discontinuity in the discount rate that is applied to net cost versus net
revenue projects – the discount rate jumps immediately from 4% to 7% at a specific point.
The cash flow series in the two cases set out above are all but identical and must have
immaterially different systematic risk profiles – yet they are discounted at materially
different rates under the modified CAPM approach.

5.6 Conclusion
The modified CAPM approach to assessing PPPs can lead to counter-intuitive results. In the
following section, we propose a new approach that:

provides the correct ranking in terms of value for money to the government;

provides outputs that are economically meaningful in that they measure the true
present value of government’s liability to make a future series of cash outflows;

evaluates the merits of the PPP proposal from the perspective of the government by Partnership
determining the present value to the government of cash flows paid by the valuation
government;
paradox

produces robust and reliable results that are consistent with the CAPM framework
that is the basis of the analysis; and

is simple, easy to understand and easy to implement.
559
6. Using standard discounted cash flow valuation to evaluate public-private
partnerships
6.1 Overview – sensible and meaningful valuation outcomes
Our proposal is to apply standard DCF valuation techniques to determine the
appropriate discount rate. The alternative approaches used in different jurisdictions
were originally motivated by a belief that the standard valuation approach did not
work in the PPP setting in that it produced perverse outcomes under which bearing
higher risk seemed to make the government better off [15]. In our view, the problem
does not lie in an inherent flaw in standard valuation techniques, but in the application
of the techniques by analysts in the PPP setting. That is, the problem is not with the
technique itself, but with its misapplication. We illustrate how the standard valuation
techniques that are applied in all other settings can be properly implemented in the PPP
setting.

6.2 Perspective of government

The starting point for our approach is that the analysis should be performed from the
perspective of the government. The government has decided, for community benefit
and social policy reasons, that a particular project should be undertaken. The goal is to
maximise the value for money in relation to the delivery of the project. For ease of
exposition, suppose that there are two ways of achieving an identical project outcome –
a government procurement (PSC) option and a PPP option. As both options produce the
same result, the least cost alternative should be preferred. Thus, we have two
alternative series of cash flows to be made by the government. Each needs to be
discounted back to present value using a discount rate that properly reflects the
systematic risk.
The cash flows are to be made by the government and the analysis is being
performed to estimate the present value of the cash flows to the government, so the
relevant discount rate is one that reflects the systematic risk of those cash flows to the
government. If the cash flows to be made by the government involve little or no
systematic risk to government, that should be reflected in the discount rate applied to
them. Conversely, if the cash flows to be made by the government involve substantial
systematic risk to the government, that should also be reflected in the discount rate
applied to them.
By contrast, the focus of the modified CAPM approach is on the perspective of the
private sector partner. Indeed the discount rate that is applied to PPP cash flows depends on
the amount of systematic risk that is transferred to the private sector, rather than on the
systematic risk that is retained by the government. We demonstrate below that this is the
starting point for error and inconsistency in the proposed application of valuation principles
to the PPP setting.
ARJ 6.3 Negative betas? [16]
34,6 The next point to note is that systematic risk or beta, from the perspective of government,
can be (and often is) negative for the sorts of cases that are likely to be examined. Beta is
usually estimated by examining a number of exchange-listed comparable firms. For each of
the comparable firms, beta is estimated using some form of regression of stock returns on
broad market returns. Listed companies tend to have positive betas because their stock
560 returns are positively correlated with broad market returns – individual stock prices tend to
go up (on average) when the market is up and down when the market is down.
The issue of whether regression analysis of historical returns generates reliable beta
estimates for estimating expected returns is a separate but related issue. A limitation of the
use of comparable listed firms for estimating the systematic risk of a PPP is that these listed
firms, by construction, will be different to PPPs. If the project under consideration for a PPP
were a positive NPV project, more than likely a private sector firm would be prepared to pay
the government for the opportunity to undertake this project. In contrast, the issue
addressed here only arises because projects under consideration are predominantly negative
NPV projects. Hence, an implicit assumption of using comparable listed firms for estimating
project beta is that the magnitude of the systematic risks of cash flows will be the same for
the listed firms and the PPP, but that the sign of expected cash flows is different and/or the
PPP could be structured in a manner whereby the private sector participants bear positive or
negative systematic risk [17].
The source of the positive correlation between stock returns and market returns lies in
the fact that a company’s cash flows tend to be better than expected when the economy is
expanding and the market is up and worse than expected when the economy is contracting
and the market is down. That is, cov(CFt,rm,t) > 0.
The corporate finance literature has long recognised the concept of a cash flow
beta (b) [18] – the systematic risk of a particular cash flow – defined as:
 
cov CFt ; rm;t
b¼ :
varðrm Þ

The relationship between the cash flow beta and the standard returns beta ( b ) is:

b ¼ b  PV ½CFt :

Consequently:
 
cov CFt ; rm;t
b ¼ :
varðrm ÞPV ½CFt 

Note here that the sign of the standard returns beta depends on the sign of the expected cash
flow, PV[CFt]. In particular, consider a set of comparable listed firms. It is likely that the
cash flows from these firms are higher than expected when the market is up and vice versa
so that cov(CFt,rm,t) > 0. It is also likely that future cash flows to the firm are expected to be
positive (if the firm is to remain solvent), so that PV[CFt] > 0. Since var(rm) > 0 by definition,
this all implies that b > 0 so the firm has a positive beta and (under the CAPM) a required
return on equity that exceeds the risk-free rate [19].
However, now consider the government provision of a social infrastructure or service
project. In this case, the government is making a series of cash outflows, so that PV[CFt] < 0.
If it remains the case that this project generates cash flows that are better than expected Partnership
when the market is up and worse than expected when the market is down, as is the case with valuation
the listed comparables, we have cov(CFt,rm,t) > 0. The implication then is that b < 0, in
which case the appropriate discount rate is less than the risk-free rate.
paradox
To see this by way of a simple example, consider the cash flow from a listed comparable
that is expected to be $10. Suppose that this cash flow will be $1 better than expected if the
market is up ($11) and $1 worse than expected if the market is down ($9) and that there is a
50/50 chance of an up or down market. In this case, note that the return is þ10% when the 561
market is up and 10% when the market is down, so there is a positive relationship
between returns for the company and returns on the market and a positive returns beta:

11  10 9  10
¼ þ10%; ¼ 10%:
10 10

Now consider the government provision of a comparable project that is not economically
viable in its own right. In this case, the project does not generate surplus cash flows, but
rather requires a subsidy from the government. Consider the case where the cash flow to be
made by the government is expected to be –$10. Again suppose that, as for the listed
comparable, this cash flow will be $1 better than expected if the market is up (–$9) and $1
worse than expected if the market is down (–$11). In this case, note that the return is 10%
when the market is up and þ10% when the market is down so there is a negative
relationship between returns for the company and returns on the market and a negative
returns beta:

ð9Þ  ð10Þ ð11Þ  ð10Þ

¼ 10%; ¼ þ10%:
ð10Þ ð10Þ

6.4 Example of application of our proposed approach

We continue the example from the previous section in which the risk-free rate is 5%, the
MRP is estimated at 6%, beta is estimated at 0.5 and the proposed PPP would involve the
government transferring two-thirds of the systematic risk to the private sector. Recall that
the expected cash outflows under the PSC are $100 per year for 3 years followed by $10 per
year for 30 years.
In this case, the government bears all of the systematic risks of the project under the PSC
because there is no other party to share any of that risk. If those PSC cash flows were
completely free of risk, the appropriate discount rate would be the risk-free rate (5%) and the
present value would be –$405. However, the government is bearing detrimental systematic
risk under the PSC, in which case the net “cost” of this project to the government must be
more than –$405.
Suppose it is the case that cov(CFt,rm,t) > 0 for the listed comparables and the PSC. In
this case, the sign on the beta estimate is reversed and becomes 0.5 and the premium for
systematic risk is 3%, as set out above. Consequently, the appropriate discount rate is 2%
(5%–3%). The present value of the PSC cash flows at a discount rate of 2% is –$499. That is,
the present value of the PSC cash flows, if they were risk-free, is –$405. However, they are
not risk-free and the value of that risk (from the perspective of the government) is –$94. This
is summarised in Table 4 below.
Under the PPP, a different series of cash flows is required from the perspective of the
government. Recall that these expected cash outflows are $35 per year for 30 years,
ARJ beginning in the fourth year. If those cash flows were completely free of risk, the appropriate
34,6 discount rate would be the risk-free rate, which produces a present value of –$465.
However, under the PPP government bears one third of the detrimental systematic risk
and, consequently, an adjustment must be made for one third of the systematic risk
premium. Again, as the cash flows are negative, in this case, but cov(CFt,rm,t) > 0, the
adjustment for systematic risk must also be negative. The resulting discount rate is 4%
562 (5%–1%) and the present value of the PPP cash flows is –$538. That is, the present value of
the PPP cash flows, if they were risk-free, is –$465. However, they are not risk-free and the
value of that risk (from the perspective of the government) is –$73. This is summarised in
Table 5 below.
Table 6 below sets out a summary of the calculations under our proposed approach.
Firstly, note that the government is bearing $94 of systematic risk under the PSC but
only $73 of systematic risk under the PPP [20]. The cost of systematic risk is lower under the
PPP because some of it has been transferred to the private sector. Even though under
the PPP two-thirds of the systematic risk is transferred to the private sector, the reduction in
the cost of systematic risk is relatively small ($94 to $73). This is because the present value
of the payment liability to the government is considerably higher under the PPP for a
considerably longer period of time. This is set out in Figure 1 below which shows the
present value of the payment liability to the government under each alternative. In each
case, the present value of the payment liability to the government is initially set at the NPV
from Table 4 above. Each year, the value of the payment liability is inflated by the relevant
rate (5% for PSC and 7% for PPP) and reduced by the amount of any payment from the
government as set out in Table 3 above. This is essentially an amortisation schedule where
the balance grows with “interest” and is reduced by “payments”. Although the premium for
systematic risk is reduced from 3% to 1%, it is applied to a substantially higher payment
liability over the life of the project.

Year 0 1 2 3 4 5 ... 33
Table 4. PSC cash flow 100 100 100 10 10 ... 10
Evaluation of PSC NPV @ 5% 405
cash flows under NPV @ 2% 499
proposed approach Dollar value of systematic risk 94

Year 0 1 2 3 4 5 ... 33
Table 5. PPP cash flow 35 35 ... 35
Evaluation of PPP NPV @ 5% 465
cash flows under NPV @ 4% 538
proposed approach Dollar value of systematic risk 73

Present value of cash flows at risk-free rate Dollar value of systematic risk Net position
Table 6.
Summary of PSC −405 −94 −499
proposed estimation PPP −465 −73 −538
method Difference −60 +21 −39
 $700.00 Partnership
valuation
$600.00 paradox
PPP
PV of payment liability to government

$500.00
PSC

$400.00
563

$300.00

$200.00

$100.00
Figure 1.
Present value of
$0.00 payment liability to
0 5 10 15 20 25 30
government
Year

Next, we note that the present value of the cash flows at the risk-free rate is substantially
higher (that is, more negative) under the PPP. These figures represent the present value of
the two cash flow streams putting aside any consideration of risk – only the time value of
money is taken into account, with the effects of systematic risk being separately accounted
for in the second column of Table 4 above. Table 4 shows that the risk-free present value of
the cash flows is $60 higher under the PPP.
In summary, by adopting the PPP rather than PSC, the government reduces the cost of
bearing systematic risk by $21. However, for this benefit government must pay the private
sector a present value of $60 in terms of additional cash flows. This leaves the government
worse off by $39. That is, the additional magnitude of the cash flows (in present value terms)
under the PPP more than offsets the value of systematic risk transferred.
By contrast, the modified CAPM approach suggests that the government is made $50
better off under the PPP (see the calculations in Section 2 above). The difference in the
present value of the cash flow streams is still $60, as set out in the first column of Table 4
above. However, the value of systematic risk transferred is estimated to be $110, which more
than offsets the $60 difference in the risk-free present value of cash flows.

6.5 Beneficial risks and detrimental risks

Thus far, our analysis has assumed that cash flows will be better than expected when the
market is up and worse than expected when the market is down so that cov(CFt,rm,t) > 0 for
the set of listed comparables and the project being evaluated. A PPP arrangement involves a
contract between the government and the private sector under which government agrees to
make a series of payments to induce the private sector into providing a piece of social
infrastructure or service that would not be economically viable but for the government
subsidy. These contracts can take a number of different forms and the form of the contract
can affect the relationship between cash flows and market returns so that cov(CFt,rm,t) is no
longer positive, therefore, no longer exposes the project to detrimental market risk. This, in
ARJ turn, will affect the sign of the returns beta that is used in the CAPM to estimate an
34,6 appropriate discount rate.
Systematic risk is generally considered to be “detrimental” because it increases the
uncertainty of payment/receipt of the cash flows, and therefore will require a premium in the
form of higher expected returns. An asset with a positive beta (or positive systematic risk)
tends to generate higher than average returns when the economy is strong and the market is
564 up and lower than average returns when the economy is weak and the market is down. This
is viewed unfavourably by investors – this market price of this asset tends to fall when the
price of other assets falls and the investor is in greatest need for some positive returns, and
its price rises when the investor already has plenty of return from other assets. This is a
detrimental systematic risk and investors will require higher average returns to attract them
to such an asset.
However, systematic risk is not always positive and therefore detrimental. An
investment that pays off more when returns in the broader market are declining but less
when returns in the broader market are increasing is a valuable hedge asset (that is, a
beneficial risk) for which the market will pay a positive price. Such an asset has a negative
beta and requires a negative risk premium – that is, the market will require a return below
the risk-free rate for such an asset that “insures” against market movements.
In a PPP context, systematic risks borne by the government can be beneficial or
detrimental depending on how the deal is structured. Alternatively, it is possible to eliminate
the systematic risk to be borne by the government either by transferring it to the private
sector or by writing a contract that involves the government making a fixed series of
payments that do not vary with general economic conditions [21]. In any case, the key
question that must be answered is, from the government’s perspective, what is the risk of
the cash flows that the government is required to make?
Consider the following three payment structures under a PPP (ignoring any abatement
mechanism):
(1) Case 1: The government pays the private sector a fixed amount each year;
(2) Case 2: The government payment is based on demand for a service and this
demand is positively related to general economic conditions; and
(3) Case 3: The government effectively immunises private sector losses.

In Case 1, the government agrees to make a fixed payment (for example, $40m) each year to
subsidise the project irrespective of the impact of external factors on the actual cash flows of
the project. This case is akin to a series of availability payments whereby the payments are
known in advance and guaranteed so long as the project proponent meets a set of minimum
“availability” standards. These fixed payments involve no risk at all to the government and
should therefore be discounted at a risk-free rate.
In Case 2, the government may agree to pay a subsidy of $2 per car that uses a toll road
or per passenger who uses a new rail link. Suppose that the volume of traffic is positively
correlated with economic growth. In this case, the volume of traffic (and therefore the
government payment) will be higher when the economy is expanding and lower during a
recession. Suppose this results in the following payment structure:

–$44m in an economic expansion when the broad market is up; and

–$36m in an economic recession when the broad market is down.

This payment structure is equivalent to a guaranteed fixed payment of $40m per year, plus
a hedge contract of:

–$4m in an economic expansion when the broad market is up; and Partnership

þ$4m in an economic recession when the broad market is down. valuation
paradox
In the case of a recession when the government’s tax receipts are expected to fall and welfare
payments are expected to increase, the government will be paying less to the private
operator (that is, the hedge contract has a positive pay-off of $4m in a recession). In other
words, when the government’s other assets are falling (or its liabilities are rising), the value 565
of this hedge asset is increasing. This payoff structure is a beneficial risk for which the
government would be willing to pay.
That is, this series of uncertain cash flows is better, from the perspective of the
government, than the fixed payment of $40m every year. Under the fixed payment option,
the government is bound to pay $40m whether its tax receipts are high or low. Under the
alternative, the government is required to make a lower payment when tax receipts are
down and a higher payment when tax receipts are up. This is a hedge asset for the
government, which reduces the variability of its budget bottom line. Consequently, it should
be preferred to the fixed payment option.
In Case 3, the government agrees to effectively immunise the private sector operator for
losses incurred over the period. These losses, and hence the government payment, will be
higher in a recession as traffic volume and toll revenue is lower. Suppose this results in the
following payment structure:

–$36m in an economic expansion when the broad market is up; and

–$44m in an economic recession when the broad market is down.

This payment structure is equivalent to a guaranteed fixed payment of $40m per year, plus
a risky cash flow of:

þ$4m in an economic expansion when the broad market is up; and

–$4m in an economic recession when the broad market is down.

In the case of a recession when the government’s tax receipts are expected to fall and welfare
payments are expected to rise, the government will be paying more to the private operator
(that is, the government must pay an additional $4m in a recession). When the government’s
other assets are falling (or its liabilities are rising), the government will incur an additional
cost to subsidise the private sector operator for their additional losses. This payoff structure
is a detrimental risk and the government should be prepared to pay to have it removed.

6.6 Measuring systematic risk: Conclusions

When estimating systematic risk, there are two key considerations relating to the sign of the
risk, that is, whether the relevant beta should be positive or negative:

Whether the cash flow will be higher than expected when the market is up and
lower than expected when the market is down or vice versa; and

Whether the expected cash flow is positive or negative.

Both of these considerations are vital to the proper estimation of systematic risk and the
proper implementation of standard valuation techniques. Our approach summarises these
considerations in Table 7 below.
Of most interest in the PPP setting will be the negative cash flows (from the perspective
of government) that are required for social infrastructure projects (or any project that is not
ARJ economically viable in its own right and requires a contribution from the government). This
34,6 is the right-hand column of the table above.
If the cash flows to be made by the government are likely to be better (lower) than
expected during economic expansions and worse (higher) than expected during recessions,
cov(CFt,rm,t) > 0 and the systematic risk is detrimental as the government will have to make
a larger cash payment in circumstances when its revenues are already under pressure. If the
566 expected cash flows are negative, the sign applied to the beta estimate must be reversed and
made negative. That is, the cash flows from the perspective of government are:

similar to those of exchange-listed comparable firms in that cov(CFt,rm,t) > 0, but

different from those of listed comparables in that PV(CFt) < 0.

This requires the sign applied to beta to be reversed and made negative.
If the cash flows to be made by the government are likely to be worse (higher) than
expected during economic expansions, cov(CFt,rm,t) < 0 and the systematic risk is beneficial
as the government will have to make a smaller cash payment in circumstances when its
revenues are under pressure. In this case, the cash flows from the perspective of the
government are:

different from those of exchange-listed comparable firms in that cov(CFt,rm,t) < 0
and

different from those of listed comparables in that PV(CFt) < 0.

This requires no sign change to be applied to the estimate of beta – the two differences
effectively cancel each other out.
In summary, the initial beta estimate and resulting systematic risk premium for the
project will be computed with reference to a set of listed comparable firms. The projects that
make up these firms tend to lie in the top left cell of Table 7 above. The expected cash flows
for these projects tend to be positive (consistent with the firm continuing to exist) and the
cash flows tend to be better than expected when the market is up and worse than expected
when the market is down (consistent with the beta estimates for these firms being positive).
If for a set of PPP or PSC cash flows, one or other of these features is reversed, the sign of the
beta estimate (and systematic risk premium) must also be reversed. If both of these features
are reversed no sign change is required (of course there are actually two sign changes, but
these cancel each other out).
We also note that our findings are consistent with the rich literature on discounting
negative cash flows in other settings [22]. Following earlier authors such as Miles and Choi
(1979) and Hull (1986), Ariel (1998), for example, demonstrates that, provided correct
distinctions are made between the cash flow beta and the standard returns beta, the
standard CAPM-related approach to valuation remains valid for negative expected cash
flows.

Sign of expected cash flow

Positive Negative

Relationship between cash flow and market return

Table 7. Detrimental systematic + –
Sign applied to beta risk: cov(CFt,rm,t) > 0
under our proposed Beneficial systematic – +
approach risk: cov(CFt,rm,t) < 0
7. Conclusions Partnership
Every year governments around the globe enter into PPP arrangements with a combined valuation
value of many billions of dollars. To date, there has been no academic literature on the
financial evaluation of PPPs relative to traditional government procurement methods. In
paradox
this paper, we examine a range of approaches that have been adopted by different national
and provincial governments for the financial evaluation of PPPs. We show that the
approaches that are currently adopted are inconsistent with the standard CAPM/DCF
approach to valuation that is used in other settings. The current approaches produce
567
illogical outcomes in some settings and are likely to mis-rank alternatives.
We show that there is no problem with the standard CAPM-based valuation approach and
that the alleged paradox is not the result of a deficiency in the current theory, but rather is
caused by its misapplication in practice. In particular, we show that the systematic risk of cash
flows is frequently mis-estimated, and the correction of this error solves the apparent paradox.
The key contribution of this paper is in the proper understanding of how systematic risk
(beta) should be estimated. Beta estimates are based on an analysis of comparable firms
listed on a stock exchange. These listed comparables have two key properties:
(1) expected cash flows are positive (if they were not, the firm would not be
economically viable and would not be listed); and
(2) on average, cash flows are higher than expected when the market is up and lower
than expected when the market is down. This implies that stock returns are
higher than average during economic expansions when the stock market is up and
lower than expected during recessions when the stock market is down (which is
why beta estimates for these firms are positive).

We show that (other things equal) if one of these properties is reversed, the sign on the beta
estimate must be reversed and made negative. If both of these properties is reversed, there is
a cancelling effect and no change is required to the sign of beta.
With properly signed estimates of beta, the standard CAPM-based valuation framework
can be applied to the PPP setting and produces correct rankings and economically
meaningful output. Specifically, the output of this process is a direct estimate of the present
value of the liability from the perspective of the government.

Notes
1. We note here also the literature that suggests that how information is reported can be material in
the consideration of public infrastructure investment – see, for example, Chatterjee et al. (2017).
2. Variations in fundamental reporting requirements across government entities, even within a
single country, have been documented in other contexts – see, for example, Wines and
Scarborough (2015).
3. In statistical terms, asset beta is the covariance of expected returns of a particular project with
returns on the market portfolio of all risky assets, divided by the variance of returns on the
market portfolio, that is, b a = COV (ra, rm)/s m2.
4. Throughout this section, we assume that an identical project is delivered by the PSC and all PPP
bids so they can be ranked on present value cost alone. That is, we deal with one complexity and
one issue at a time.
5. Of course, PPPs tend to involve construction periods of 2–3 years followed by a concession
period of 20 years or more. We have kept this example as simple as possible to illustrate the
point.
ARJ 6. Between 2005 and 2014, for example, 48 infrastructure PPP transactions worth $61bn reached the
formal announcement phase, with 40 successfully closed (Deye, 2015).
34,6
7. The US Office of Management and Budget (1992) has published a set of guidelines for how
government agencies should compare the costs and benefits of proposed programs. These
guidelines propose that cash flows from proposed alternatives should be compared by
discounting to present value using a government bond rate.
568 8. HM Treasury in The Green Book: Appraisal and Evaluation in Central Government, Annex 6:
Discount Rate, page 97.
9. British Columbia Ministry of Transportation: Project Report: Achieving Value for Money William
R. Bennett Bridge Project, September 2005, p. 8. “For the comparison, both proposals and the PSC
used a market-based discount rate of eight per cent, which is an estimate of the private sector
weighted average cost of capital for a project of this type.”
10. Or likely to make – for example, if the PPP is for an essential service, it is likely that government
will step in as a provider of last resort in the event of the collapse of the private sector partner,
even if there is no contractual obligation for it to do so.
11. In December 2008, Infrastructure Australia released the National Public Private Partnership
Guidelines: Discount Rate Methodology Guidance. In August 2013, a revised version was
published, which did not feature any substantial changes. The Australian Guidelines set out an
approach for evaluating PPPs that is designed to overcome the perceived limitations of
traditional valuation approaches in this setting. They are based on the framework that had been
previously proposed by the state governments of Victoria and New South Wales.
12. For example, Brealey et al. (1997, pp. 23–24) note that “The UK Government applies the same
discount rate of 6% across the vast majority of public-sector projects. Yet it is widely accepted
that the discount rate should vary with a project’s exposure to factor risk” and that “when assets
are exchanged between the public and private sectors, the spurious apparent value may be
created by the use of an inappropriate discount rate.” They conclude that “the cost of capital is
the same in the public and private sectors.” Along the same lines, Klein (997, p. 30) concludes that
“the apparent cheapness of sovereign funds reflects the fact that the taxpayers, who effectively
provide credit insurance to the sovereign, are not remunerated for the contingent liability they
assume. If they were to be remunerated properly, then the advantage of sovereign finance
would – almost by definition – disappear”.
13. That is, to make the relevant point here we consider the case where the government will make a
completely fixed set of cash flows. In practice, abatement mechanisms may be common and
result in the cash flow stream from the government varying. At this juncture, we seek to make a
simple point relating to the valuation of a fixed set of cash flows from the government.
14. For example, the $165m might be the PSC estimate for providing the same service or
infrastructure or it might simply be the government’s estimate of the value of the project broadly
defined to include social benefits, etc. The point here is that government is or believes that it is,
receiving more than $164m in benefits from the project.
15. There are also some other motivations for approaches used in other jurisdictions. For example,
the UK use of a social time preference discount rate is also consistent with a view that this is the
appropriate rate to be applied to all social infrastructure regardless of the risk associated with
cash flows.
16. The material in this section draws on the formal mathematical derivations that are set out in
detail in the Online Appendix.
17. An alternative technique is to directly estimate the project’s beta from scenario analysis on the
project itself, specifically asking, “What would be the expected cash flows in states where the
market return was above or below expectations?” However, for the purposes of this paper, we
 focus only on the application of beta estimates in the valuation of PPPs, not the technique in Partnership
which those estimates are made.
valuation
18. See, for example, Brealey et al. (2006), 8th ed., p. 227, and the derivations on the accompanying paradox
website at www.mhhe.com/mba8e
19. We note that throughout this analysis we consider the extent to which systematic risk is driven
by the correlation between cash flows and market returns (as a proxy for aggregate wealth).
Correlation between stock and market returns can also be driven by changes in the market price 569
of risk. See Campbell and Mei (1993) and Davis (2005).
20. More formally, “the cost to government of bearing systematic risk” has a present value of $94
under the PSC and $73 under the PPP.
21. This assumes that the contract remains materially the same over the contract term and the
private sector operator fulfils its obligations under the contract.
22. See, for example, Beedles (1978), Miles and Choi (1979), Lewellen (1977), Ariel (1998), Berry and
Dyson (1980, 1983) and Hull (1986).
23. See, for example, Brealey et al. (2006), 8th ed., p. 227, and the derivations on the accompanying
website at www.mhhe.com/mba8e

References
Alonso-Conde, A.B., Brown, C. and Rojo-Suarez, J. (2007), “Public private partnerships: incentives, risk
transfer and real options”, Review of Financial Economics, Vol. 16 No. 4, pp. 335-349.
Ariel, R. (1998), “Risk adjusted discount rates and the present value of risky costs”, The Financial
Review, Vol. 33 No. 1, pp. 17-30.
Beedles, W.L. (1978), “Evaluating negative benefits”, The Journal of Financial and Quantitative
Analysis, Vol. 13 No. 1, pp. 173-176.
Benítez-Ávila, C., Hartmann, A. and Dewulf, G. (2019), “Contractual and relational governance as
positioned-practices in ongoing public–private partnership projects”, Project Management
Journal, Vol. 50 No. 6, pp. 716-733.
Berry, R.H. and Dyson, R.G. (1980), “On the negative risk premium for risk adjusted discount rates”,
Journal of Business Finance and Accounting, Vol. 7 No. 3, pp. 427-436.
Berry, R.H. and Dyson, R.G. (1983), “On the negative risk premium for risk adjusted discount rates: a
reply”, Journal of Business Finance and Accounting, Vol. 10 No. 1, pp. 157-159.
Brealey, R.A., Cooper, I.A. and Habib, M.A. (1997), “Investment appraisal in the public sector”, Oxford
Review of Economic Policy, Vol. 13 No. 4, pp. 12-28.
Brealey, R., Myers, S. and Allen, F. (2006), Principles of Corporate Finance, 8th ed., McGraw-Hill,
New York, NY.
Burger, P. and Hawkesworth, I. (2011), “How to attain value for money: comparing PPP and traditional
infrastructure public procurement”, OECD Journal on Budgeting, Vol. 11 No. 1, pp. 91-146.
Buso, M., Moretto, M. and Zormpas, D. (2020), “Excess returns in public-private partnerships: do
governments pay too much? (No. 0246)”, Dipartimento di Scienze Economiche, working paper.
Campbell, J. and Mei, L. (1993), “Where do betas come from?”, Review of Financial Studies, Vol. 6 No. 3,
pp. 567-592.
Chatterjee, B., Mir, M.Z., Eddie, I.A. and Wise, V. (2017), “Infrastructure reporting by New Zealand local
authorities – perceptions and expectations”, Accounting Research Journal, Vol. 30 No. 1,
pp. 36-57.
Cheung, E., Chan, A.P. and Kajewski, S. (2009), “Reasons for implementing public private partnership
projects”, Journal of Property Investment and Finance, Vol. 27 No. 1.
ARJ Cheung, E., Chan, A.P. and Kajewski, S. (2012), “Factors contributing to successful public private
partnership projects”, Journal of Facilities Management, Vol. 10 No. 1.
34,6
Cui, C., Liu, Y., Hope, A. and Wang, J. (2018), “Review of studies on the public–private partnerships
(PPP) for infrastructure projects”, International Journal of Project Management, Vol. 36 No. 5,
pp. 773-794.
Davis, K. (2005), “PPPs and infrastructure investment”, The Australian Economic Review, Vol. 38 No. 4,
570 pp. 439-444.
Deye, A. (2015), “US infrastructure public-private partnerships: ready for takeoff?”, Kennedy School
Review, Vol. 15, pp. 6-12.
Esmaeeli, A.N. and Heravi, G. (2019), “Real options approach versus conventional approaches to
valuing highway projects under uncertainty”, Journal of Construction Engineering and
Management, Vol. 145 No. 11, p. 04019068.
Gao, H., Xiang, Y. and Xie, S. (2017), “Private capital rate of return on investment in PPP projects based
on the perspective of government regulation”, in Wu, Y., Zheng, S., Luo, J., Wang, W., Mo, Z. and
Shan, L. (Eds), Proceedings of the 20th International Symposium on Advancement of
Construction Management and Real Estate, Springer, Singapore.
Greve, C. and Hodge, G. (Eds) (2013), Rethinking Public-Private Partnerships: Strategies for Turbulent
Times, Routledge.
Grimsey, D. and Lewis, M.K. (2005), “Are public private partnerships value for money? Evaluating
alternative approaches and comparing academic and practitioner views”, Accounting Forum,
Vol. 29 No. 4, pp. 345-378.
Hellowell, M. and Vecchi, V. (2018), “Assessing the cost of capital for PPP contracts”, in Public-Private
Partnerships in Health, Palgrave Macmillan, Cham, pp. 85-109.
Hodge, G.A. and Greve, C. (2017), “On public–private partnership performance: a contemporary
review”, Public Works Management and Policy, Vol. 22 No. 1, pp. 55-78.
Hu, Z., Chen, S. and Zhang, X. (2014), “Value for money and its influential factors: an empirical
study of PPP projects in Japan”, Built Environment Project and Asset Management, Vol. 4
No. 2.
Hull, J.C. (1986), “A note on the risk adjusted discount rate method”, Journal of Business Finance and
Accounting, Vol. 13 No. 3, pp. 445-450.
Iossa, E. and Martimort, D. (2015), “The simple microeconomics of public private partnerships”, Journal
of Public Economic Theory, Vol. 17 No. 1, pp. 4-48.
Ismail, S. (2013), “Critical success factors of public private partnership (PPP) implementation in
Malaysia”, Asia-Pacific Journal of Business Administration, Vol. 5 No. 1.
Istrate, E. and Puentes, R. (2011), “Moving forward on public private partnerships: US and international
experience with PPP units”, Brookings-Rockefeller Project on State and Metropolitan Innovation,
December.
Jin, X.H. (2010), “Determinants of efficient risk allocation in privately financed public infrastructure
projects in Australia”, Journal of Construction Engineering and Management, Vol. 136 No. 2,
pp. 138-150.
Jin, H. and Liu, C. (2020), “Optimal design of government guarantee and revenue cap agreements in
public–private partnership contracts”, International Journal of Innovation, Management and
Technology, Vol. 11 No. 1.
Ke, Y., Wang, S., Chan, A.P. and Lam, P.T. (2010), “Preferred risk allocation in China’s public–private
partnership (PPP) projects”, International Journal of Project Management, Vol. 28 No. 5,
pp. 482-492.
Klein, M. (1997), “The risk premium for evaluating public projects”, Oxford Review of Economic Policy,
Vol. 13 No. 4, pp. 29-42.
Lewellen, W.G. (1977), “Some observations on risk-adjusted discount rates”, The Journal of Finance, Partnership
Vol. 32 No. 4, pp. 1331-1337.
valuation
Liu, J., Gao, R. and Cheah, C.Y.J. (2017), “Pricing mechanism of early termination of PPP projects based
on real option theory”, Journal of Management in Engineering, Vol. 33 No. 6, p. 04017035.
paradox
Liu, J., Love, P.E., Carey, B., Smith, J. and Regan, M. (2015a), “Ex-ante evaluation of public-private
partnerships: macroeconomic analysis”, Journal of Infrastructure Systems, Vol. 21 No. 2,
p. 04014038.
Liu, J., Love, P.E., Smith, J., Regan, M. and Davis, P.R. (2015b), “Life cycle critical success factors for
571
public-private partnership infrastructure projects”, Journal of Management in Engineering,
Vol. 31 No. 5, p. 04014073.
Loosemore, M. and Cheung, E. (2015), “Implementing systems thinking to manage risk in public
private partnership projects”, International Journal of Project Management, Vol. 33 No. 6,
pp. 1325-1334.
Ma, L., Li, J., Jin, R. and Ke, Y. (2019), “A holistic review of public-private partnership literature
published between 2008 and 2018”, Advances in Civil Engineering, Vol. 2019.
Malone, N. (2005), “The evolution of private financing of government infrastructure in Australia – 2005
and beyond”, The Australian Economic Review, Vol. 38 No. 4, pp. 420-430.
Martimort, D. and Straub, S. (2016), “How to design infrastructure contracts in a warming world: a
critical appraisal of public-private partnerships”, International Economic Review, Vol. 57 No. 1,
pp. 61-88.
Martins, J., Marques, R. and Cruz, C. (2015), “Real options in infrastructure: revisiting the literature”,
Journal of Infrastructure Systems, Vol. 21 No. 1.
Miles, J. and Choi, D. (1979), “Comment: evaluating negative benefits”, The Journal of Financial and
Quantitative Analysis, Vol. 14 No. 5, pp. 1095-1099.
Nisar, T.M. (2007), “Risk management in public–private partnership contracts”, Public Organization
Review, Vol. 7 No. 1, pp. 1-19.
Partnerships Victoria (2003), “Technical note: use of discount rates in the partnerships Victoria
process”, Victorian Department of Treasury and Finance, Melbourne.
Petersen, O. (2019), “Evaluating the costs, quality, and value for money of infrastructure public-private
partnerships: a systematic literature review”, Annals of Public and Cooperative Economics,
Vol. 90 No. 2, pp. 227-244.
Saeed, A.M., Duffield, C. and Hui, F.K.P. (2018), “An enhanced framework for assessing the operational
performance of public-private partnership school projects”, Built Environment Project and Asset
Management, Vol. 8 No. 2, pp. 194-214.
Smith, J. (1999), “So what is principal finance?”, Structured Finance International, April, pp. 1-5.
Tang, L. and Shen, G.Q. (2013), “Finance-related critical success factors for the briefing of PPP projects
in construction”, In ICCREM 2013: Construction and Operation in the Context of Sustainability,
pp. 427-437.
Wang, Y. and Liu, J. (2015), “Evaluation of the excess revenue sharing ratio in PPP projects using
principal–agent models”, International Journal of Project Management, Vol. 33 No. 6,
pp. 1317-1324.
Whitfield, D. (2017), “PPP profiteering and offshoring: new evidence”, ESSU Research Report No.10.
Wines, G. and Scarborough, H. (2015), “Australian government budget balance numbers: the
hybrid nature of public sector accrual accounting”, Accounting Research Journal, Vol. 28
No. 2, pp. 120-142.
Zhang, X. (2005), “Critical success factors for public–private partnerships in infrastructure
development”, Journal of Construction Engineering and Management, Vol. 131 No. 1,
pp. 3-14.
ARJ Further reading
34,6 Booth, L.D. (1983), “On the negative risk premium for risk adjusted discount rates: a
comment and extension”, Journal of Business Finance and Accounting, Vol. 10 No. 1,
pp. 147-155.
British Columbia Ministry of Transportation (2005), “Project report: achieving value for money William
R. Bennett bridge project”, Infrastructure BC, Vancouver.
572 Federal Highway Administration Office of Innovative Program Delivery (2012), Value for Money
Assessment for Public-Private Partnerships: A Primer, Washington, DC, Federal Highway
Administration.
HM Treasury (2014), “The green book: appraisal and evaluation in central government”, available at:
www.hm-treasury.gov.uk/d/green_book_complete.pdf
Infrastructure Australia (2013), “National public private partnership guidelines – volume 5: discount
rate methodology guidance”, Australian Government, Canberra.
Mostaan, K. and Ashuri, B. (2017), “Challenges and enablers for private sector involvement in delivery
of highway public–private partnerships in the United States”, Journal of Management in
Engineering, Vol. 33 No. 3, p. 04016047.
New South Wales Government (2007), “Technical paper: determination of appropriate discount rates
for the evaluation of private financing proposals”, NSW Treasury, Sydney.
US Department of Transportation (2009), “Public-private partnerships for highway infrastructure:
capitalising on international experience”, American Trade Initiatives Report, available at: http://
international.fhwa.dot.gov/pubs/pl09010/pl09010.pdf
US Office of Management and Budget (1992), “Guidelines and discount rates for benefit-cost analysis of
federal programs, transmittal memo no. 64”, available at: www.whitehouse.gov/omb/rewrite/
circulars/a094/a094.html

Appendix. Derivation of beta estimation framework

Context
In this appendix, we set out the mathematical derivation of our proposed beta estimation framework.
The goal of the appendix is to establish that our proposed framework is mathematically rigorous and
consistent with finance and valuation theory and that it produces economically sensible and
meaningful results. We begin by considering a one-period example in which there is a single cash
flow to be made one period from now. We also consider all projects to be 100% equity financed so
that the asset beta is the relevant measure of systematic risk and there is no need to consider re-
levering of equity betas and so on. Indeed, for the remainder of this appendix, we refer to the asset
beta as simply “beta”.

Definition of beta
In a CAPM context, beta is formally defined as:

covðrp ; rm Þ
b ¼
varðrm Þ

where cov(rp,rm) is the covariance between the return on the project (rp) and the return on the market
(rm, usually proxied by the return on a broad market index such as the All Ordinaries Index) and var
(rm) is the variance of the returns on the market portfolio.
Definition of returns Partnership
For the one-period examples in this appendix, we define the present time to be Time 0 and the end of valuation
the period (when the cash flow is to occur) as Time 1.
In general, the return on an asset over the period can be written as:
paradox
P1
r¼ 1
P0
573
where P0 is the price of the asset today and P1 is the price of the asset at the end of the period. For
example, if the price of an asset increases from 100 to 110 over the period, the return is 10%.
In the case at hand, we will be considering the payment of a single cash flow at the end of the
period. This payment will be made by the government to the private partner and will be denoted by
CF1. The present value of this cash flow (at Time 0, the beginning of the period) is denoted PV[CF1] –
the present value of the cash flow that will be made at Time 1, the end of the period.
The return in relation to this cash flow is therefore:

CF1
r¼ 1
PV ½CF1 

Suppose the present value of the cash flow is 100 and the actual cash flow made at the end of the
period is 110. This would represent a 10% return over the period.

Definition of the present value of a cash flow

We have not yet discussed how to compute the present value of a cash flow. Standard finance and
valuation practice is to estimate the present value of a cash flow by discounting the expected cash
flow using an expected return:

E ½CF1 
PV ½CF1  ¼ 1
1 þ E ½rp 

Here, we recognise that CF1 is risky in the sense that the amount of the payment that will occur at
Time 1 is unknown at Time 0. That is, the payment to be made at Time 1 might be contingent on the
number of cars that use a rail link or the number of patients that use a hospital over a period. So
the amount of the payment is uncertain. If there is a 50/50 chance of that payment being 100 or 90, the
expected cash flow is 100. It is this expected cash flow that is discounted back to present value.

Definition of beta in terms of cash flows

We substitute in expressions for the return and present value of cash flows to write the definition of
beta in terms of cash flows as follows:

covðrp ; rm Þ
b ¼
varðrm Þ
! !
CF1
cov    1 ; rm
E ½CF1 = 1 þ E ½rp 
¼
varðrm Þ
ARJ Since E[rp] and 1 are both constants, we have:
34,6    
1 þ E ½rp  CF1
b ¼ cov ; rm
varðrm Þ E ½CF1 

That is, beta (or systematic risk) for a particular cash flow depends on the covariance
  between the
574 return on the market (rm) and the actual cash flow relative to its expected value ECF
½CF1  .
1

Aside: Reconciliation with cash flow betas

The cash flow beta is defined as:

covðCF1 ; rm Þ
b¼
varðrm Þ

The relationship between the cash flow beta and the standard returns beta is:

b ¼ b  PV ½CF1 

Consequently:

covðCF1 ; rm Þ
b ¼
varðrm ÞPV ½CF1 

and using the definition of PV[CF1] from above yields:

covðCF1 ; rm Þð1 þ rp Þ
b ¼
varðrm ÞE ½CF1 

which can be rearranged as:

   
1 þ E ½rp  CF1
b ¼ cov ; rm :
varðrm Þ E ½CF1 

That is, our derivation above is consistent with the standard definition of a cash flow beta and the
well-established relationship between cash flow and returns betas. The cash flow beta and
the relationships set out above are standard results in finance that are central to valuation using the
certainty equivalent approach [23].

Further derivations in relation to beta

We have noted above that:
   
1 þ E ½rp  CF1
b ¼ cov ; rm
varðrm Þ E ½CF1 

It is most common to use the CAPM to estimate the expected (or required) return on a project:
 E ½rp  ¼ rf þ b  MRP Partnership
valuation
where rf is the risk-free rate of interest and MRP is the market risk premium. paradox
However, this means that beta appears on both sides of the equation above – on the left hand
side and on the right hand side as a component of E[rp]. Substituting in for E[rp] and rearranging the
expression yields:
  575
cov CF1
;r
1 þ rf
E ½CF1  m
b ¼  
½CF1  ; rm
varðrm Þ  MRP  cov ECF 1

This is an unwieldy expression, but it does not have to be evaluated at all when implementing
our proposed framework. We only report the derivation in this appendix to demonstrate that
beta can be written in terms of the key covariance term and a number of other basic quantities
that can be estimated in standard ways. That is, there are well-established techniques for
estimating the risk-free rate (r f ), MRP and market volatility (var(r m ). All of these terms are
estimated in precisely the same way whether it is a PPP or any other project being analysed.
This leaves only the key covariance term to be estimated, which is the focus of this appendix.
An understanding of how this covariance term works, and how it drives beta, is fundamental to
an understanding of our proposed approach and the problems with the approach set out in the
Guidelines.
We also note that the derivation above is consistent with Hull (1986), who expresses a similar
result in terms of returns as:
  
varðrm Þ cov E ½CF1  ; rm 1 þ rf
MRP CF1

r ¼ rf þ  
1  var
MRP
ðrm Þ
cov CF1
E ½CF1 
; rm

Definition of variance
The statistical definition of variance, illustrated in terms of the market return, is:
 2
varðrm Þ ¼ E rm  E ½rm 

Suppose, for example, that the uncertainty around the market return is such that there is a 50/50
chance of the return being þ30% or 10% over the period.
In this case, the expected return on the market is:

E ½rm  ¼ 0:5  ðþ30%Þ þ 0:5  ð10%Þ ¼ 10%

The variance of the return on the market in this case is:

2 2
varðrm Þ ¼ 0:5  ð30%  10%Þ þ 0:5  ð10%  10%Þ ¼ 0:04
ARJ Definition of covariance
34,6 The statistical definition of covariance, illustrated in terms relevant to the case at hand, is:
 
CF1 CF1 CF1 
cov ; rm ¼ E E rm  E ½rm 
E ½CF1  E ½CF1  E ½CF1 

576 Since:

CF1 E ½CF1 
E ¼ ¼1
E ½CF1  E ½CF1 

we have:
 
CF1 CF1 
cov ; rm ¼ E  1 rm  E ½rm  :
E ½CF1  E ½CF1 

Extending our earlier example, suppose that a cash flow will be 110 if the market is up and the
economy is doing well and 90 if the market is down and the economy is doing poorly. In this case, the
expected cash flow is:

E ½CF1  ¼ 0:5  110 þ 0:5  90 ¼ 100

and the covariance will be:

 
CF1
cov ; rm ¼ 0:5ð1:1  1Þð0:3  0:1Þ þ 0:5ð0:9  1Þð0:1  0:1Þ ¼ 0:02:
E ½CF1 

Calculation of beta
In extending this example further, suppose that the risk-free rate is 4% and the MRP is 6%. Note that
this is consistent with the expected return on the market portfolio being 10%, as in our example
above.
We can compute the beta for this case as:
  
cov 1 þ rf
CF1
;r
E ½CF1  m 0:02ð1 þ 0:04Þ
b ¼  ¼ ¼ 0:54:
varðrm Þ  MRP  cov E ½CF1  ; rm
CF1 0:04  0:06  0:02

Using the CAPM to estimate the required return yields:

E ½rp  ¼ 4% þ 0:54  6% ¼ 7:2%:

Negative cash flow, better than expected when the market is down
In this case, we extend our example from above but consider a negative cash flow. Suppose that the
government has agreed to pay a subsidy to the private partner and that the cash flow (from the
perspective of the government) will be 44 if the market is up and the economy is doing well or 36
if the market is down and the economy is doing poorly. For example, the government may be paying Partnership
an amount for each service provided by the private sector partner, and demand for the service might valuation
be positively related to the state of the economy. We continue to assume that there is a 50/50 chance
paradox
of the market return being þ30% or 10% over the period.
In this case, we have:

E ½CF1  ¼ 0:5  ð44Þ þ 0:5  ð36Þ ¼ 40 577

and the outcome (from the perspective of the government) is worse than expected if the market is up
and better than expected if the market is down.
Note that this is the reverse of the standard case for a profitable project with positive cash flows,
where cash flow outcomes tend to be better than expected if the market is up and worse than
expected if the market is down. Intuition might lead one to conclude that it would be appropriate in
these circumstances to use the opposite beta that would be applied to a comparable, but profitable,
project. However, this is not correct. The sign of the key covariance term and, consequently beta, is
altered by the fact that the expected cash flow is negative. In particular, we have:
 
CF1
cov ; rm ¼ 0:5ð1:1  1Þð0:3  0:1Þ þ 0:5ð0:9  1Þð0:1  0:1Þ ¼ 0:02:
E ½CF1 

We can compute the beta for this case as:

  
cov 1 þ rfCF1
;r
E ½CF1  m 0:02ð1 þ 0:04Þ
b ¼  ¼ ¼ 0:54:
varðrm Þ  MRP  cov E ½CF1  ; rm
CF1 0:04  0:06  ð0:02Þ

Using the CAPM to estimate the required return yields:

E ½rp  ¼ 4% þ 0:54  6% ¼ 7:2%:

Negative cash flow, guaranteed

Next consider the case where the government is to make a guaranteed payment to the private sector
partner, an absolutely fixed payment. In this case, the covariance between the cash flow and the
return on the market is zero – there is no relationship between these variables because the cash flow
is fixed. Substituting this into our equation for beta produces a beta estimate of zero, which implies
that these cash flows should be discounted at a risk-free rate. This accords with basic intuition. A
series of cash flows that are guaranteed by the government should be discounted at the relevant risk-
free rate. In other words, a government bond should be valued as a government bond.

Negative cash flow, better than expected when the market is up

In this final case, we extend our example from above but consider a negative cash flow that is better
than expected when the market is up (i.e. the reverse of the previous example of a risky payment to be
made by the government). For example, the government may have agreed to pay a subsidy to the
private sector partner so that the cash flow (from the perspective of the government) will be 36 if
the market is up and the economy is doing well or 44 if the market is down and the economy is
ARJ doing poorly. That is, the government is effectively subsidising the losses of the private partner and
34,6 those losses will be lower if the market is up and the economy is doing well. We continue to assume
that there is a 50/50 chance of the market return being þ30% or 10% over the period.
In this case, we have:

E ½CF1  ¼ 0:5  ð36Þ þ 0:5  ð44Þ ¼ 40

578
and the outcome (from the perspective of the government) is better than expected if the market is up
and worse than expected if the market is down.
This is similar to the standard case for a profitable project with positive cash flows – cash flow
outcomes tend to be better than expected if the market is up and worse than expected if the market is
down. Intuition might lead one to conclude that it would be appropriate in these circumstances to use
the same beta that would be applied to a comparable, but profitable, project. However, this is not
correct. The sign of the key covariance term and, consequently beta, is altered by the fact that the
expected cash flow is negative. In particular, we have:
 
CF1
cov ; rm ¼ 0:5ð0:9  1Þð0:3  0:1Þ þ 0:5ð1:1  1Þð0:1  0:1Þ ¼ 0:02:
E ½CF1 

We can compute the beta for this case as:

  
cov CF1
1 þ rf
;r
E ½CF1  m 0:02ð1 þ 0:04Þ
b ¼  ¼ ¼ 0:50:
varðrm Þ  MRP  cov E ½CF1  ; rm
CF1 0:04  0:06  ð0:02Þ

Using the CAPM to estimate the required return yields:

E ½rp  ¼ 4% þ ð0:50Þ  6% ¼ 1:0%:

Note that the beta estimate, in this case, is not the exact opposite of the previous case involving risky
cash flows. In the previous case, the cash flow was 10% worse than expected when the market was
up and 10% better than expected when the market was down. In this case, the reverse is true. In the
previous case, the key covariance term
 
CF1
cov ; rm
E ½CF1 

was þ0.02 and in this case, it is 0.02. That is, this covariance term has the same magnitude but a
different sign. However, beta is not a perfect linear function of the key covariance term. In particular,
the denominator of the expression for beta contains a second-order term that involves the product if
the covariance term and the MRP. However, this results in a minor variation so that the beta is 0.54
for the former case and 0.50 in the latter.
In this regard, we note that beta cannot be estimated with any great precision (certainly not with
respect to the second decimal place) and that the Guidelines propose that beta estimates be obtained
from the relevant broad risk band. The minor variation in the magnitude of beta in the example
above would certainly not result in a movement from one risk band to another. The obviously more
important effect is in obtaining the correct sign on the beta estimate. Consequently, the focus of our Partnership
proposed approach is on obtaining the correct sign for the beta estimate and takes the magnitude of valuation
beta from the relevant risk band (consistent with the approach set out in the Guidelines in this
respect).
paradox

Summary
This appendix shows that beta depends upon the covariance between a cash flow relative to its 579
expected value and market returns. It also shows that this covariance (and, consequently beta)
depends crucially on:

whether the expected cash flow is positive or negative; and

whether the cash flow is more likely to be better than expected when the market is up
and worse than expected when the market is down or vice versa.
Consequently, when determining the sign for beta (positive or negative) these two considerations
must be addressed. Our proposed approach sets out a simple 2  2 grid for this purpose, as illustrated
in Table 7 above.

Corresponding author
Stephen Gray can be contacted at: [email protected]

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]