Technical ReportPDF Available
Contract Design in
Public-Private Partnerships
Report prepared for the World Bank
BY
Elisabetta Iossa
Giancarlo Spagnolo
Mercedes Vellez
FINAL VERSION
September, 2007
1
TABLE OF CONTENTS
1. INTRODUCTION ......................................................................................... 3
2. THE ROLE OF CONTRACT DESIGN ......................................................... 9
2.1 Contract misspecifications, standardized contracts and PPP Advisory units ............................. 9
2.2 Transparency, Confidentiality and Governance in PPP contracting ..........................................12
2.2.1 Transparency and Governance ..................................................................................................12
2.2.2 Transparency in Public Procurement .........................................................................................13
2.2.3 Transparency in Public-Private Partnerships .............................................................................15
3. RISK ALLOCATION, INCENTIVES, AND TYPES OF PPPS ....................17
3.1 The main characteristics of PPPs ...................................................................................................17
3.2 Categories of risks............................................................................................................................18
3.3 Efficient Risk allocation ..................................................................................................................20
3.4 Limits to risk transfer......................................................................................................................25
3.5 Types of PPPs and risk allocation ..................................................................................................26
3.6 Joint Ventures ..................................................................................................................................31
3.7 Service unbundling ..........................................................................................................................31
4. PAYMENT MECHANISM ...........................................................................33
4.1 Output specification, payment mechanism, and risk allocation ..................................................33
4.2 The provision of incentives through the payment mechanisms ...................................................34
4.2.1 Cost-plus payment......................................................................................................................35
4.2.2 Fixed-price payment ..................................................................................................................36
4.2.3 Incentive payment......................................................................................................................37
4.3 Payment mechanism in PPPs..........................................................................................................40
4.3.1 User charges...............................................................................................................................41
4.3.2 Usage payments .........................................................................................................................44
4.3.3 Availability payments ................................................................................................................45
4.3.4 Performance payments...............................................................................................................46
4.3.5 Liquidated damages and performance bonds .............................................................................47
4.4 Contractual incompleteness, non contractible quality and reputational forces .........................49
4.5 Price variations ................................................................................................................................51
4.5.1 Inflation indexation....................................................................................................................52
4.5.2 Price reviews, market testing, and benchmarking......................................................................53
4.5.3 Tariff regulation and price adjustment clauses...........................................................................54
4.5.4 Price adjustment and strategic behavior (gaming) .....................................................................55
5. FLEXIBILITY AND RENEGOTIATION.......................................................57
2
5.1 General Trade offs: Flexibility, Predictability and Renegotiation ..............................................57
5.2 More on the costs of abusing flexibility and renegotiation...........................................................60
5.2.1 Incentive Distortions..................................................................................................................62
5.2.2 Bidding distortion ......................................................................................................................65
5.2.3 Enforcement problems, rent-shifting and corruption .................................................................66
5.2.4 Political interests........................................................................................................................67
5.3 How contract design can help to maintain flexibility while limiting abuses ...............................69
6. CONTRACT DURATION............................................................................73
6.1 Contract duration and investment .................................................................................................73
6.2 Contract duration and flexibility....................................................................................................75
6.3 Contract duration and competition................................................................................................76
6.4 Contract duration, renewals, and performance incentives ..........................................................78
7. OTHER CONTRACTUAL ISSUES.............................................................79
7.1 Refinancing.......................................................................................................................................79
7.2 Dispute resolution ............................................................................................................................80
7.3 Step-in rights ....................................................................................................................................81
7.4 Early Termination ...........................................................................................................................83
References...............................................................................................................................................90
ANNEX ...................................................................................................................................................96
Table 1: Risk allocation and incentives in BOT contracts .................................................................97
Table 2: Risk allocation and incentives in DBFO contracts...............................................................98
Table 3: Risk allocation and incentives in Management contracts....................................................99
Table 4: Risk allocation and incentives in Lease contracts ..............................................................100
3
1. INTRODUCTION
Public-Private Partnerships (PPPs) are long-term contractual arrangements between
the public and private sectors in which the private sector has responsibility for significant
aspects of the building and operation of an infrastructure for the delivery of public services.
For the purpose of this paper, we shall use the term PPP to refer to any contractual
arrangement between a public-sector party and a private-sector party for the provision of
public services with the following four main characteristics: (i) the bundling of project phases
into a single contract, (ii) an output specification approach, (iii) a high level of risk transfer to
the private-sector party, and (iv) a long-term contract duration (see Section 3 for a detailed
analysis of these four points; see Hodge and Greve (2007) for a survey on the definitions of
PPPs used in the public management literature).
PPPs have been implemented broadly around the world. In the 1980s, the United
Kingdom pioneered the development of a particular form of PPPs, creating the Private
Finance Initiative (PFI) in 1992 to further promote PPP agreements. As of December 2006,
794 PFI projects had been signed involving around £ 55 billion of capital value (CBI, 2007;
HM Treasury, 2006). Other European countries have also invested in PPPs, especially Ireland,
Portugal, Greece, the Netherlands, and Spain (PWC, 2005; EIB, 2004) and large PPP projects
have been implemented in the US. In a review of PPP activity, PricewaterhouseCoopers
(PWC, 2005) reports that 206 PPP contracts were signed worldwide in 2004-5 involving USD
52 billion in investments.
PPP agreements in developing countries have grown steadily since the 1990s.
According to the World Bank’s Private Participation in Infrastructure1 database, 2750
infrastructure projects involving private and public investment for capital value of USD 786
billion have been implemented in 1990-2003 (in 2002 constant dollars). Around 1000 projects
and 47% of the investment took place in Latin American and the Caribbean (LAC) countries,
where Chile and Mexico were pioneers in the use of PPPs (IMF, 2004).
1 Private Participation in Infrastructure (PPI) is an alternative way to name PPPs in the development-
financing sector (see Yescombe (2007) p. 4)
4
Since the 1980s, and particularly after the United Kingdom developed the PFI
program, PPPs have been introduced in many sectors, mainly water and sanitation, transports,
energy, and telecommunications. In recent years, countries with a large experience in this
kind of private participation in public services provision, such as the United Kingdom,
Australia, Canada, Ireland and the Netherlands have introduced PPPs in areas such as
education, health, water and waste management (Hammami et al., 2006).
In LAC countries, PPPs were implemented mainly through concession contracts: in
the 1990s, 3 out 4 PPP agreements were of this type (Fay and Morrison, 2005). Concessions
targeted water services and transports, including ports, airports, roads, and railways
(Hammami et al., 2006). The private partner was given the right to operate a service for a long
period of time, while the public-sector party retained asset ownership and regulatory power.
Since the private partners operating in these sectors were often foreign companies,
privatizations became a politically sensitive issue and governments opted for concession
contracts (Guasch et al., 2003). However, privatizations were the norm in energy and
telecommunication sectors (Hammami et al., 2006). For example, Argentina, Bolivia, Brazil,
Chile, Colombia, and Peru privatized the great majority of their energy distribution and
generation facilities. Transferring public asset ownership to the private partner required deep
legal reforms (and even constitutional reforms) to lift institutional constraints on state-owned
enterprises divestiture and on land expropriation.
In assessing PPP performance, a number of circumstances are typically associated
with poor PPP outcomes: contract renegotiations, delays in construction and service
provision, failures to meet quality standards, service disruptions because of disputes between
the contracting parties, and early project terminations. 2
Renegotiations are particularly severe in LAC countries; they have led to substantial
changes in the original contract terms to the benefit of the private partner and at the expense
of final users, the public sector, and taxpayers. A typical contract revision delivers
adjustments in several key variables, including tariffs, cost components passed through tariffs,
2 These circumstances reflect the historical experience in LAC. However, there are other sources of
evidence that can be used to assess PPP outcomes; in this regard, Hodge (2004) distinguishes between the policy
rhetoric, the legal contract, and the historical experience as alternative sources.
5
annual fees due to the public partner, obligation and timing of investments, asset base used to
impute rates of return, and length of the concession period. In this regard, Guasch (2004)
reports that renegotiations of concession contracts allowed for a tariff increase (62 % of the
cases), an increase in the number of cost components passed through tariffs (59 %), a
reduction in annual fees due to the public sector (31 %), and delays and reduction in
investments (69 %). In a quite small number of cases, tariffs were reduced (19 %), annual fees
raised (17 %), and asset base modified undermining private returns (22 %).
World-wide experience with PPP suggests there is no a ‘one-size-fits-all’ principle
that might simplify the design of a PPP contract for a given objective and sector. However,
the empirical evidence suggests that some factors heavily influence the likelihood of
performance failure in a PPP agreement (though these factors are not specific to PPPs). The
first three factors listed below could be considered as ‘external’ to the contract; and the fourth
factor as ‘internal’:
i. the characteristics of the targeted sector and the market structure
ii. the degree of macroeconomic instability
iii. the country’s regulatory and institutional framework
iv. the contract design and management, in particular the payment mechanism and the
risk allocation built-in the contractual terms
The characteristics of the sector targeted by a PPP contract and the prevailing market
structure help in explaining PPP performance, as it is suggested by differences across sectors
in the incidence of contract renegotiation for LAC countries. The highest renegotiation
incidence corresponds to concessions in transports and water services, where 55 % and 74 %
of the contracts ended up under revision (Guasch, 2004). Some authors suggest the recurrent
revisions implied that reforms in these sectors were far from successful (Estache et al., 2003).
A lower incidence is observed in privatizations in energy and telecommunications, where the
private sector participation and market competition were relatively higher.
The degree of macroeconomic instability also matters in accounting for PPP
outcomes. In LAC, the incidence of contract renegotiation also differs across countries with
different degree of macroeconomic instability, suggesting that aggregate risks influence some
6
aspects of a PPP agreement. In an uncertain macroeconomic environment, contract design
failures are more likely since it is difficult for the contracting parties to envisage future
contingencies and write the contract terms accordingly (thus aggravating problems arising
from contract incompleteness). Ex ante, a high aggregate risk level discourages long-term
contractual relationships and weakens incentives to undertake investments in infrastructure
projects that typically have long maturity. Moreover, macroeconomic instability translates
into the occurrence of unexpected shocks of an economic and financial nature (e.g. aggregate
demand contractions, increases in interest rates, exchange rate devaluations, etc.) that may
undermine the PPP project sustainability and lead to the insolvency of the private partner.
To the extent that macroeconomic crises could dramatically change the market
conditions surrounding a concession, it is more likely that the partners will end up calling for
a contract revision. In fact, empirical evidence shows a correlation between macroeconomics
instability and renegotiation incidence: the peaks of contract revision occurred when negative
shocks hit LAC countries and triggered severe macroeconomic crises. For instance,
generalized renegotiations were observed soon after the Argentine hyperinflation in 1990, the
Mexican crisis in 1995, the Brazilian devaluation in 1999, the Colombian recession in 2000,
and the Argentine crisis in 2001 (Guasch et al., 2003).
An important source of risk for PPP agreements in LAC countries was currency risk.
In general, the private partner collected local-currency denominated revenues but financed
investments using foreign currency-denominated debt. Hence, the private-sector party was
heavily exposed to real exchange rate fluctuations. The currency crises increased the real
value of the private partners’ liabilities, causing large financial losses and threatening the
project’s sustainability. Undergoing financial stress, private partners had strong incentives to
call for a contract revision. In some cases, contract clauses sought to insure the private partner
against aggregate risks, but the public sector reneged on these clauses when a severe
macroeconomic crisis occurred. For instance, the Buenos Aires water concession intended to
protect the private partner against currency risk by indexing local-currency denominated
tariffs to the US dollar. However, after a devaluation of the local currency, Congress passed
an economic emergence law that nullified these guarantees (Lobina and Hall, 2003).
Regarding the regulatory and institutional framework, the quality of contract
enforceability and governance are critical factors affecting PPP agreements. In LAC
7
countries, weak governance and the government’s lack of commitment not to renegotiate also
accounted for the recurrent contract revisions. In many of these countries, the regulatory
agencies were rarely given training and instruments to carry on their mandate with
competence and even lacked political support from the government.
Moreover, in some cases, the government had political control over them, raising
concerns on autonomy and accountability issues (Estache, 2006). There were instances in
which the private partner considered its main counterparts to be ministers and secretaries
rather than the regulatory agency. For example, in the Buenos Aires water concession, the
Secretary of Environment and Sustainable Development bypassed the regulators arguing
regulation of water provision was too complex to be managed by the regulatory agency.
In addition, some private partners insisted on commercial confidentiality issues, or the
governments raised concerns about the public interest, and thus contract documents became
secret files, undermining the transparency of PPP contract design (Lobina and Hall, 2003). In
this regard, it has been argued that, to the extent that a regulatory agency does not depend on
the fiscal budget to finance their activities, it enjoys a higher autonomy. In fact, in some LAC
countries, regulatory agencies are financed by levying duties on the gross revenues of the
regulated industries (Fay and Morrison, 2005).
However, when the funds financing the regulatory agency are closely linked to the
tariffs collected by the concessionaire, the regulator’s interests appear to be aligned with the
private partner’s, and thus the autonomy and accountability issues remain unsettled (Lobina
and Hall, 2003). Overall, the LAC experience suggests that competent, autonomous, and
accountable regulatory bodies can make a crucial contribution to the development and success
of PPP arrangements.
Aspects of the contract design, such as the risk allocation or the payment mechanism,
significantly affect the PPP outcomes. In LAC countries, contract renegotiations also resulted
from misallocation of project-idiosyncratic risks between the public and private partner,
suggesting the distribution of risks between the contracting parties matter. Project-related
risks, such as construction risk, cost overruns risk, and demand risk, are allocated through the
contract design. The contract may either transfer a wrong amount of risk to the private-sector
party or transfer the wrong type of risks. In LAC, most cases of renegotiation or contract
termination were due to contract design failing to manage risks (Guasch, 2004). But also in
the EU risk assessment and allocation seem to be problematic issues, leading to contract
8
revisions and unanticipated financial burdens for the public sector (Renda and Schrefler,
2006).
PPP failures also occur because of contracts misspecifications and poorly designed
contracts. For example, there have been instances where contractual deductions for poor
performance were insufficient to incentivize the private partner (e.g. the London Underground
case) or where the contract omitted to specify what would happen to the facility once the
contract expired.
Furthermore, the sheer complexity of PPP contracts makes contract design a key issue
for the success of a PPP project. Even countries with a long tradition on PPP contracts review
their standardized contracts on a regular basis by letting information from past projects feed
into the design of better practices (see for example the UK).
In this paper, we shall focus on the main issues in contract design for PPP agreements.
Particular emphasis will be given to aspects of the contract that may have an impact on the
likelihood of success of a PPP.
In section 2, we shall discuss the importance of standardized contracts and advisory
units to reduce the risk of contract misspecification. We shall also discuss governance and
transparency issues associated with the public procurement process and the PPPs.
In section 3, we shall discuss the efficient risk allocation and describe the incentive
structure and risk transfer between the contracting parties for the main forms of PPPs (tables
describing the risk allocation in different PPP contracts are presented in the annex).
In section 4, we shall review alternative payment mechanisms suggested by the
economic literature on contracts, followed by an analysis of payment schemes and price
adjustment provisions used in practice.
In section 5, we shall discuss the trade-offs between flexibility, investment protection
and predictability, and contract renegotiation.
In section 6, we shall discuss the implication of contract duration on investment
decisions and PPP performance.
Finally, in section 7 we shall discuss other important issues concerning PPP contract
design like refinancing, dispute resolution procedures, step-in rights, and early contract
termination.
9
2. THE ROLE OF CONTRACT DESIGN
2.1 Contract misspecifications, standardized contracts and PPP Advisory
units
PPP contracts are based on an output specification approach: the public-sector party
defines the basic standards of service whilst the private-sector party chooses how to meet and
possibly improve upon these basic standards. This approach incentivizes innovative solutions,
allowing for private sector's skills and knowledge to feed into public service provision, but
comes at the cost of greater risk of contract misspecifications for the public sector.
In particular, problems may arise because the output characteristics specified in the
contract, which form the basis of the contractual obligations, may be ill or not clearly
described. Problems may also arise to the extent that the output specifications are inconsistent
with the infrastructure needs that the PPP intends to satisfy, and that should be identified by
conducting a careful assessment previous to the contract drafting. Mistakes at the contract
drafting stage can then be very costly for the public-sector party because of the long-term
nature of most PPP contracts.
A number of factors can lead to contract misspecifications. For example, they can be
due to mistakes resulting from an incompetent public-sector party (or whoever acts on its
behalf) in charge of contract drafting. Lack of appropriate incentives for the public-sector
party and thus inadequate effort in information acquisition and processing can also result in
contract misspecifications. Corruption and favouritism explain contract misspecifications that
lead to the private-sector party receiving very favourable contract terms.
Making the public-sector party accountable for its actions so as to provide adequate
incentives is not an easy task. Labour market regulations often constrain the use of incentive
mechanisms for the public sector and lack of financial stakes make it difficult to provide
incentives in the first place. Also, PPP contracts are generally long-term contracts, so when
mistakes are discovered, the public sector employee may have moved job already.
In the early stages of PPPs, mistakes at the contract drafting stage have often arisen
simply because of lack of experience of public administrations on the writing of PPP
10
contracts. In Europe, a challenge for the public sector has been to rapidly build up the
capacity and knowledge to devise and implement PPPs, and to manage the PPP contractual
relationships over the long-run. The public sector's progress on this front has not kept pace
with that of private sector partners. In the UK, recognition of this problem by the National
Audit Office has led to a series of programs aimed at training public officials and to an
extensive use of external consultants.
However, the use of external consultants can be a means to acquire competence and
knowledge for the benefit of the public-sector party, but it also has limits. First, the use of
consultants does not address the problem of corruption. A priori, there is no reason to believe
that external consultants maybe less prone to corruption than public sector bureaucrats,
although it could be argued that large consultancy firms may have more to loose from being
involved in a corruption scandal than single individuals.
Second, large consultancy firms that have the specific knowledge to deal with
contracts as complex as PPP contracts find themselves hired by the public sector in one
contract and by the private sector in another. Working for both sides, although at different
points in time, is likely to be conducive to favoritism.
To the extent that experience with PPP in a specific sector accumulates, i.e. the PPP
market matures, the public authorities could standardize parts of the contracts for that specific
sector as a means to reduce the likelihood of contract and output misspecification. From
ongoing projects in that sector it is possible to learn what the generic risks are, and then use
this information to design a standard contract (or some parts of it) that allocates these risks
properly. Thus, the use of standardized contracts also reduces transaction costs resulting from
agreeing and drafting the contract, provided that both parties are willing to accept the
standardized terms. 3
Finally, legislation for PPP contracts providing for the compulsory use of standardized
contracts could reduce the incidence of corruption by making it impossible to offer bribes in
exchange of favourable contract terms. Standardized contracts are widely used for example in
the UK (see HM Treasury, 2007).
3 In PPP markets which are not mature, the use of guidelines on commercial principles could be a
(probably imperfect) substitute for standardized contracts.
11
However, the compulsory use of standardized contracts also brings the cost of
increased rigidity and, in particular, of valuable local or specific information not feeding into
the standardized contract. This problem could be overcome by adopting an ‘intermediate’
approach towards contract standardization. The process of standardization could be governed
by more general guidelines on how to write PPP contracts, more general and less rigid that a
standardized contract. Alternatively, the public-sector party could be given the option to
introduce motivated changes into the standardized contract. The benefit of this increased
flexibility would then have to be weighed against the cost of a higher risk of corruption and
favouritism.
For these reasons, it might be advisable to also develop public sector ‘special units’
that advice local administrations on the drafting of contracts (this entity could also advice
local administrations in their evaluation of PPP projects, tender documents and bids, and in
negotiations with private partners). The PPP Unit Partnerships UK in the UK and the Unità
Tecnica della Finanza di Progetto in Italy serve these functions. These entities could lead the
‘learning and standardizing’ process by helping to internalize the informational externalities
that exist across the different contracts, and could be made accountable for their actions.
As long as the entities are independent (e.g. the UK PPP unit is not), they could also
be in charge of accepting or rejecting proposals for modifications to the standardized contracts
made by local authorities, which would allow the use of flexible standardized contracts whilst
maintaining control on possible abuses or mistakes. Advice, oversight and approval of
modifications of the contracts during renegotiation could also be delegated to these special
entities. Compared to a local public-sector party, the reputational concerns of the entity
should be stronger, thus possibly limiting its incentives to renegotiate contract terms with the
private-sector party. 4
As suggested by Monteiro (2005), these units specialized in PPP contract design and
management should be separated from industry regulators (if any) in order to avoid a conflict
of interest. A yet separate unit would then have to monitor the activity of the PPP unit, by
surveying public-private relationships, collecting, analyzing, and then disseminating
information on PPPs.
4 See Bennett and Iossa (2006) for a study on how delegation of contract management and renegotiation
to a PPP unit affects the investment incentives of the private-sector party.
12
In fact, the very development of PPPs enticed many competent public sector
employees to join the private sector. While this has probably raised overall productivity, it
made the objective of creating value for money for the public sector through PPPs even more
challenging. A skill retention policy, e.g. increase remuneration of project manager or module
training qualification linked to an existing training qualification, may help reducing this
problem.
2.2 Transparency, Confidentiality and Governance in PPP contracting
Disclosure requirements of contractual terms are a crucial tool in the governance of all
economic transactions. When public money is involved, governance problems tend to
increase because of the many layers of delegation that often protect public agents and the lack
of market-based governance mechanisms. Disclosure requirements therefore tend to be more
stringent for transactions involving public partners in the hope that higher transparency can
increase accountability by facilitating taxpayer control. In the particular case of PPP
procurement, accountability is made even more difficult by the complexity and specificity of
each procurement.
In the following, we briefly review the crucial link between disclosure rules,
transparency and accountability/governance in general, and then in the specific of public
procurement and PPP.
2.2.1 Transparency and Governance
Since the ‘Cadbury report’ identified the lack of disclosure in CEOs and directors
compensation contracts and the side transactions as crucial determinants of corporate
governance failure, most corporate governance codes and regulations incorporate stringent
disclosure requirements on the details of crucial contracts (see e.g. OECD, 2004).
Increased disclosure of contractual terms and performance information is widely
thought to improve shareholders’ and more generally financial markets’ ability to properly
evaluate firms’ management and prospects, and to act consequently.
13
One main benefit of this, for good corporations, is that - by reducing information
asymmetries with potential investors - increased transparency tends to reduce their cost of
finance. Another important benefit, on the governance side, is that increased transparency
may help markets to provide the right incentives to firm management.5
Among the costs associated with increased disclosure by a firm are the operational
cost of correctly disclosing more information and, more importantly, the competitive cost of
having competitors that know more about the firm’s situation (so called ‘commercially
sensitive’ information).6
The widespread governance failures around the world in the beginning of the last
decade (Enron, WorldCom, etc.) led to a further reinforcement of disclosure requirements,
both in voluntary codes and in binding regulations.7 Disclosure requirements are also a crucial
tool in banking regulation. It has recently been shown through cross-country comparisons that
the regulatory approaches that act by increasing disclosure requirements to the market, rather
than by having a powerful regulator dictating behaviour to financial institutions, are those that
work best (Barth et al., 2004; Beck et al., 2006).
2.2.2 Transparency in Public Procurement
For Public Procurement in general, where governance, and in particular corruption
problems are still widespread even in OECD countries, in large part because market forces
cannot discipline politically-protected public buyers that misbehave, stringent disclosure
requirements are also seen as a potentially powerful remedy (see e.g. Rose-Ackerman, 1999;
Kaufmann, 2005).
5 Leuz and Wysocki (2007) provide a review of what is currently known about the effects of disclosure
requirements on corporations.
6 Hermalin and Weisbach (2007), however, show that increased transparency may also imply costs in
terms of higher employees’ compensation.
7 For a good example of non-binding codes see CALPERS (2007). For compulsory regulation, in the
US the Sarbanes Oxley Act of 2002 substantially increased compulsory disclosure requirements for corporations,
extending them to a detailed reporting of off-balance sheet financing and special purpose entities; and in 2006
the Security and Exchange Commission further increased the level of details of compensation contracts that
listed firms must report (see e.g. SEC, 2006; Sutherland et al., 2006).
14
It is interesting that Coppier and Piga (2007) find an inverse cross-country relationship
between the level of transparency in public procurement and the perceived level of corruption
in a country, as measured by Transparency International. According to the authors, this
relationship is the result of governments’ attempts to fight corruption by increasing disclosure
and transparency in countries where this is a major problem. This relationship testifies for the
widespread belief that transparency is a force that limits corruption and improves governance,
but also that transparency is a relatively weak force..
The direct costs of disclosing information on contract terms and performance
evaluation appear to be rather small in general (see e.g. Leuz, 2007), and even more so for
repeated procurements and large infrastructure projects like PPP. Disclosure costs in terms of
potential competitive harm for the private-sector party (and for potential private partners in
the phases that precede the signature of the contract) should be rather small when disclosure
regards contractual and output-related performance measures, and much larger when
disclosure refers to investment choices and other input-related variables that may convey
delicate information about production processes and strategic choices.
When the buyer is a public entity, another potential cost of disclosure adds to the
competitive harm for the private partners: information can be deemed sensitive to the public
interest, for example on national defence grounds or on the ground of weakening the future
bargaining position of the public sector in future procurements.
One cost of disclosure rules that is not often discussed, but that has been identified
early in relation to public procurement rules, is that public knowledge on the price and
quality conditions offered by the winning private partners may harm the competitive process
by facilitating anti-competitive agreements among competitors. Bid ringging tries to illegally
keep procurement prices at a level higher than the competitive one, trying to prevent
competitors from cutting prices by threatening to punish them with ‘price wars’. And, as the
Nobel-prize-winner George Stigler wrote regarding public procurement auctions: The
system of sealed bids, publicly opened with full identification of each bidder's price and
specification, is the ideal instrument for the detection of price cutting.” (1964, p. 48).
15
2.2.3 Transparency in Public-Private Partnerships
In comparison with the public procurement of standardized products or services, PPPs
are likely to be even more problematic from a governance point of view, as PPP procurements
tend to be infrequent for the particular public buyer, much larger, more complex, and often
specific to particular assets. These features make benchmarking and other standard forms of
outside control more complex. At the same time, stakes are higher than in standard
procurement, so bad governance can be much more costly.
Early reports on PPP best practices recognized well the particular governance
problems of PPP procurement, and therefore suggested (IPPR, 2001) or prescribed (NHS,
2003) great levels of transparency and a widespread and proactive disclosure of contractual
terms.
Ex post analyses such as Gosling (2004), however, have revealed that even in a
country like the UK, with a good general level of accountability and a lively public debate,
non-binding ‘best practice’ recommendations to disclose information were seldom followed
by public administrations, even when directly asked for the information. It is clear, therefore,
that in countries with weaker general accountability and public debates, non-binding
disclosure requirements are likely to have little or no impact.
We mentioned that the standard ground for labeling a piece of information as
‘confidential’ and not disclosing it is that the piece of information, if disclosed, could damage
a private party by undermining its situation relative to competitors. This suggests there is a
trade off between accountability and willingness of private parties to disclose delicate
information on how certain needs could be faced in a PPP. Such information is crucial, for
example, in the early stage of assessing the suitability of PPP for an infrastructure project..
The trade off discussed above should be relevant for information about the production
processes and strategic choices of the private-sector party. The disclosure cost in terms of
potential competitive harm for the private-sector party should therefore be small or absent
when disclosure is limited to contractual terms (payment schemes, quality standards,
deductions, prices, etc.) and other output-related measures (revenues).
16
However, since PPP contracts are based on output specifications and the assessment
and selection phases are in the past, from a PPP governance perspective the most important
contractual information that needs to be disclosed once the contract is signed is exactly that
about output variables.8 Moreover, information should be disclosed within a pre-established,
short, and binding time limit from its emergence. The time limit could be established by law,
by the national auditor, and/or by the contract, but in any case it must be short and strictly
binding. Notice that such information should imply little or no competitive harm for the
private-sector party.
The case for confidentiality on contractual and output features that convey no
information on inputs after the contract has been signed is, therefore, extremely weak.
The other cost of disclosure mentioned earlier, the direct cost of making information
public, appears simply negligible if compared to the very large size typical of PPP
procurements. It is not a case, therefore, that most reports and best practice documents
suggest maximal disclosure of PPP documentation, limiting confidentiality to a very strict set
of information (HM Treasury, 2007).
The competitive cost of disclosing information on the awarding price highlighted by
Stigler, however, is likely to be relevant for PPP, as PPPs regard often such large projects that
potential competitors can be counted on the fingers of one hand, making distortions of
competition highly likely. This could be a reason not to disclose exact price information in the
contract.
We have been writing until now about full disclosure requirements, unrestricted in
terms of recipients. Disclosure restricted to bodies like Agencies and Accounting offices that
would maintain the information relative to competitors under secret does not cause the costs
discussed, and therefore such a disclosure must be complete (HM Treasury, 2007).
8 Also, a case could be made for disclosing the criteria used in the PPP suitability assessment (e.g. the
public sector comparator) and the business outline (i.e. the formal presentation of the project for approval by the
public sector).
17
3. RISK ALLOCATION, INCENTIVES, AND
TYPES OF PPPS
3.1 The main characteristics of PPPs
A unanimous definition of PPP does not exist. As it was mentioned in the
Introduction, we use the term PPP to refer to any contractual arrangement between a public-
sector party and a private-sector party for the provision of public services with the following
four main characteristics: (i) the bundling of project phases into a single contract, (ii) an
output specification approach, (iii) a high level of risk transfer to the private-sector party, and
(iv) a long-term contract duration.9
The bundling of project phases into a single contract is the main characteristics of PPP
contracts. If we consider the different stages of a project as comprising the design (D), the
building (B), the finance (F) and the operation and management (O), we have that PPPs differ
in terms of which of these four stages are delegated to the private sector. However, the term
PPP is generally used to indicate a substantial involvement of the private sector in at least the
building (or renovation) and operation of the infrastructure for the public-service provision.
The bundling of project phases encourages the private-sector party (typically a consortium of
firms) to think about the implications of its actions on different stages of the project (from the
building to the operation) and thus favours a whole-life costing approach (see Bennett and
Iossa (2006), and Martimort and Pouyet (2007) for an in depth discussion).
PPP contracts are based on output specifications in the sense that the public-sector
party defines only basic standards of service, leaving the private-sector party with the choice
as to how to meet and possibly improve upon these basic standards. The idea of the output
specification is to provide incentives for innovative approaches, allowing for private sector’s
skills and knowledge to feed into public service provision. By leaving the private partner’s
9 See Iossa and Martimort (2007) for a discussion of the incentive effects of the main characteristics of
PPPs discussed above.
18
discretion over how to meet the output requirements, scope is created for innovation and
flexibility at the project design stage and throughout the contractual period.
PPP contracts are characterized by a relevant level of risk transfer to the private-sector
party, although the specific risk allocation varies with the form of PPP used for the project, as
different is the scope of activities delegated to the private sector. For each type of PPP
contract, risk is allocated to the private-sector party through contractual incentives and
penalties incorporated within the payment mechanism, and through the activities for which
the private-sector party is responsible.10
PPP contracts are generally long-term contracts with duration increasing with the level
of financial involvement of the private sector in the provision of investments. Upon contract
expiry, the public-sector party regains possession of the assets and can re-tender aspects of the
service provision to other providers or take provision in-house.
In this section we shall discuss the main risks involved in a typical infrastructure
project and the criteria for efficient risk allocation. We shall then describe different types of
PPP contracts and how each type of contract allocates those risks between the public and
private partners.
3.2 Categories of risks
In every infrastructure project there are risks associated with the nature of the project.
For example, a project to construct a highway to diminish travel times is subject to the risk of
construction delay due to unexpectedly bad geological conditions, inefficient construction
practices, additional environmental studies to be done before the works begin, unexpected
problems raised by their results, or simply construction permits not released on time. Once the
construction phase ends, the project faces the risk of not raising sufficient revenues to cover
operation costs and recoup invested funds because of traffic volume being lower than
expected.
We list below the main types of risk at general level. For a further discussion, see Li
10 See Grimsey and Lewis (2002) for a discussion of how risk allocation in PPPs differs from that in
traditional procurement.
19
Bing et al. (2005) and Loosemore (2007).
Statutory/Planning risk
It includes the risk that planning permission for the construction of the infrastructure
project may be refused, the risk that unacceptable conditions may be applied to any planning
permission granted, and the risk that the planning process may take longer than anticipated
and cost more than expected.
Misspecification of output requirements risk
It refers to the possibility that the output characteristics specified in the contract and
which form the basis of the contractual obligations are ill or not clearly described.
Design risk
It includes the risk of failing to complete the design process in time and within the
budgeted costs, the risk of failing to deliver a solution that works satisfactorily and meets the
requirements set by the public-sector party. This risk also includes the possibility of changes
in technical standards during the design phase.
Construction risk and time schedule risk
It includes the risk that factors such as changes in labour and materials costs,
inadequate cost management, inefficient construction practices, adverse site and weather
conditions, protester action, delays in obtaining approvals and permits, and the failure of
private partners to perform may lead to construction delay and cost overruns.
Operation risk
It includes the risk factors that may lead to an increase in operation costs and failure to
meet availability and performance standards. These factors include shortage of skilled labor,
labor disputes, late delivery of equipment, poor maintenance schedule, inadequate cost
management, etc.
Demand risk
It is the risk of making lower-than-expected revenues if the actual demand for service
falls short of the demand initially forecasted.
20
Risk of changes in public needs
It is the risk that the output specifications set up in the initial contract become
inadequate because of changes in society’s preferences. The relative importance of this risk
increases with contract length, as for a longer contract the chance of changes in public needs
is greater.
Legislative/Regulatory risk
It is the risk of changes in the legislative and regulatory framework, e.g. unexpected
modifications in tax legislation, tariff-setting rules, and contractual obligations regarding
investment and quality standards.
Financial risk
Operating and capital losses may result from interest and exchange rate fluctuations,
capital controls restricting convertibility and transferability of profits, etc.
Residual value risk
It is the risk of holding a facility (e.g. land, buildings, water plant) whose value at the
end of the contract is lower than that anticipated at the start.
3.3 Efficient Risk allocation
In a contract, the risk allocation between the contractual parties should accomplish two
sets of goals: to provide incentives for the parties to undertake efficient actions when these
actions cannot be directly contracted upon (because they are not immediately observable) and
to provide insurance to a risk averse party against the risks of the project. In a PPP contract, in
particular, the goals of risk allocation should be:
(i) to provide incentives to reduce the long-term cost of a project;
(ii) to provide incentives to complete the project in time and within budget;
(iii) to provide incentives to improve the quality of services and revenues yield; and
(iv) to insure the risk averse public and private partners against risk. Risk insurance for
the public partner helps to improve its profile of expenditure on the project by
21
converting variable operation and capital costs into predictable unitary payments.
Risk insurance for the private partner helps to reduce the cost of capital.
In order to accomplish the above goals in the most effective way, two principles
should guide the allocation of risk between the public and private partners:11
(P1) given partners with similar risk-aversion, the risk should be allocated to the
party that is responsible or has relatively more control over the risk factor, and
(P2) given partners with similar responsibility or control over the risk factor, the
risk should be allocated to the party that is more able to bear it, i.e. the less risk-averse party.
The intuition and the implications of the above principles can be understood by way of
an example. Consider a PPP where the private-sector party is in charge of constructing a
facility for the provision of a public service. Suppose that there are two construction practices
that can be followed, practice G (good) and practice B (bad), and that the type of construction
practice adopted by the private-sector party is not observable by the public-sector party.
Suppose that if practice G is chosen, then with probability 10% problems arise during
construction that lead to an increase in the cost of the project by $10.000. If instead practice B
is chosen, then the corresponding probability is 50%. That is, we are in a situation where the
efficiency of the construction practice affects the likelihood of cost overruns. Suppose also
that practice G has a non-contractible cost for the private-sector party of $7.000 whilst
practice B has a non-contractible cost of $6.000. Non-contractibility implies that this cost is
borne by the private-sector party who then has incentives to take it into account when
choosing between practice G and B.
Finally, consider the possibility that the private-sector party and the public-sector party
are risk averse. A risk averse party faces a cost for bearing risk (we call it ‘risk premium’)
which generally increases with: (i) the degree of risk aversion of the party, (ii) the likelihood
11 See Martimort and Sand-Zantiman (2006) for a discussion of how risk allocation can also help to
address situations where the public-sector party is more informed than the private sector party about the quality
of the assets.
22
that adverse events will occur, and (iii) the amount of risk borne, in the example given by the
expected cost overruns. Denote by rpr and rpu the parameters, with values between 0 and 1,
that capture the degree of risk aversion of the private-sector party and of the public-sector
party, respectively. Denote by x the fraction of risk borne by the private-sector party as
provided by the contract; 1-x is the fraction of risk borne by the public-sector party.
For the sake of simplicity, we can use the following expressions as a simplified
measure of the cost of bearing risk for the private-sector party in our example:
rpr · x · 10% · $10.000 when practice G is used,
rpr · x · 50% · $10.000 when practice B is used.
Whilst the cost of bearing risk for the public-sector party is:
rpu · (1-x) · 10% · $10.000 for practice G,
rpu · (1-x) · 50% · $10.000 for practice B.
Consider now the expected total cost of any given practice. This is given by the sum
of the expected cost overruns, the non-contractible cost for the private-sector party, and the
risk premium for each of the party that bears the risk.
For practice G the expected total cost is given by:
CG = [10% · $10.000] + $7.000 + [rpr · x · 10% · $10.000 + rpu · (1-x) · 10% · $10.000]
= $8.000 + $1.000 · [rpu + x · (rpr - rpu)].
Whilst the expected total cost of practice B is:
CB = [50% · $10.000] + $6.000 + [rpr · x · 50% · $10.000 + rpu · (1-x) · 50% · $10.000]
= $11.000 + $5.000 · [rpu + x · (rpr - rpu)].
It then follows that practice G always minimizes the cost of the project since CG is
always lower than CB. Therefore, the efficient practice is practice G.
23
The issue is then whether the private-sector party has incentives to select practice B,
given that its choice of practice cannot be contracted upon. As we show below, a problem
may arise because the private-sector party only takes into account its own cost and not the
total cost of the project, and because of the presence of some non-contractible costs.
In particular, consider the rational choice for the private-sector party. If the private-
sector party chooses practice G, it will face the cost:
[x · 10% · $10.000] + $7.000 + rpr · x · 10% · $10.000 = $7.000 + x · $1.000 · (1 + rpr).
Whilst if it chooses practice B, it will face the cost:
[x · 50% · $10.000] + $6.000 + rpr · x · 50% · $10.000 = $6.000 + x · $5.000 · (1 + rpr).
By comparing the above two formulas, it is immediate that unless the private-sector
party bears sufficient risk (i.e. x is sufficiently high), it will choose practice B instead of the
efficient practice G. In fact, consider the case where the public-sector party bears all the
construction risk and therefore pays for all cost overruns, i.e. x=0. In this case, the private-
sector party will not have incentives to choose practice G since practice B will cost it less
($6.000 instead of $7.000).
If instead the private-sector party bears a sufficient amount of risk, it will choose the
efficient practice G. In particular, for any x at least equal to 25%, the private-sector party will
have incentives to choose practice G independently of its level of risk aversion rpr. This is
because for any x at least equal to 25%, the cost of practice G for the private-sector party is
never greater than the cost of practice B:
$7.000 + 25% · $1.000 · (1 + rpr) $6.000 + 25% · $5.000 · (1 + rpr).
The above example shows the rationale behind principle (P1), which leads to the
following criterion: when the private-sector has relatively more control over a risk factor, then
transferring the risk to the private-sector party (i.e. setting x sufficiently high) helps to provide
incentives for efficient actions (i.e. choice of practice G rather than B).
24
Note that for x between 0 and 25% the inequality above is satisfied only for
sufficiently high rpr. In other words, the more risk averse is the private party, the smaller is
the risk it must bear to have correct incentives to perform.
Note also that the non-contractibility of the action undertaken by the private-sector
party to control the risk factor (in our example given by the type of construction practice) is
critical for principle (P1) to be relevant. Should the action undertaken by the private-sector
party be perfectly observable by the public-sector party, it would suffice for efficiency that the
public-sector party specified in the contract which particular action it wishes the private sector
to undertake (in our example, to choose construction practice G).
The above example also explains principle (P2). Consider again the total cost of
practice G, as given above by:
CG = $8.000 + $1.000 · [rpu + x · (rpr - rpu)].
We see immediately that transferring risk to the private partner increases the total cost
of the project whenever the private-sector party is more risk averse than the public-sector
party. Formally, whenever rpr>rpu, the greater the risk transfer x, the greater the total cost CG.
More importantly, the total cost of practice G is minimized by letting the party with the lower
degree of risk aversion bear most of the risk. That is, by choosing x that minimizes the term
[rpu + x · (rpr - rpu)].
The aim to minimize the total cost of the project, isolating risk averse parties from
risk, explains principle (P2) and yields the following criteria that abstract from the effects of
risk allocation on incentives (i.e., they hold assuming that project G is selected). First, when
the public-sector party is more risk averse than the private-sector party (rpu>rpr), then the total
cost of the project is minimized by letting the private-sector party bear all the risk; this can be
achieved by setting x=1. Second, when instead the public-sector party is less risk averse than
the private-sector party (rpu<rpr), then (in the absence of incentive problems) the total cost of
the project is minimized by letting the public-sector party bear all the risk; this can be
achieved by setting x=0.
25
When we put together both the issue of incentives and that of risk premiums
minimization, it is then clear that risk is optimally allocated if the following holds:
i. When the public-sector party is more risk averse than the private-sector party
(rpu>rpr), then risk transfer to the private-sector party helps both to ensure
incentives over non-contractible actions and to minimize the total cost of the
project. The optimal risk allocation then calls for the private-sector party to
bear all the risk: x=1.
ii. When the public-sector party is less risk averse than the private-sector party
(rpu<rpr), then risk transfer to the private-sector party generates a trade-off: it
helps to ensure incentives but it may lead to an excessive risk premium.
Typically, however, the incentives consideration prevails and the efficient risk
allocation has the private-sector party bearing a substantial amount of risk, the
more the less risk averse it is.
3.4 Limits to risk transfer
The public-sector party should keep in mind that certain residual risks cannot be
transferred, e.g. the risk of political discontent if the public service provision deteriorates as
the private-sector party underperforms (objectively or according to the public’s perception).
Further, risk allocation in practice may differ from what it has been planned and
originally envisaged because the public-sector party is the provider of last resort in all PPP
projects. If, for instance, a PPP school were found to be massively behind schedule and over
budget during the construction stage, the public sector would have to take back full control of
the project. The need to ensure service continuation is often one of the reasons behind the
decision of the public-sector party to renege on contracts provision either by bailing out a
private-sector party in difficulty or by not levying penalties. These practices tend however to
have serious negative effect in the long run, and should be avoided as far as possible, as
discussed at greater length in Section 5.
26
3.5 Types of PPPs and risk allocation
Following our definition of PPPs we list below different types of PPPs ranging from
the least sophisticated modes to the most complex forms of PPPs according to the activities
and risks assigned to the private-sector partner.
Build-Operate-Transfer (BOT) contract
In a BOT contract, the private-sector party takes responsibility for building (B),
operating and managing (O) assets. The bundling of these activities is meant to provide
incentives for the private-sector party to take into account the cost of operating the asset not
only in the operation phase but also early in the construction phase. In the BOT contract,
investment in capital assets is undertaken by the private-sector party but it is financed by the
public-sector party, which retains the financial risk. Upon contract expiry, the ownership of
the assets is transferred (T) to the public-sector party sector under the terms of the original
agreement, unless a contract extension or renewal is eventually granted.12
Design-Build-Finance-Operate (DBFO) contract: concessions and operation PFI
In a DBFO contract, the private-sector party (typically a consortium of firms) is in
charge of all the stages of a project for the provision of a public service. This PPP contract
involves the design (D), building (B), finance (F) and operation (O) of the project. The
bundling of project phases encourages the consortium to think about the implications that a
design and/or construction decision will have on the operating effectiveness and the costs of
managing and maintaining the facility during its operational life.
The consortium is normally formed by a joint venture between a range of
organizations including private partners, facilities managers, banks investors and suppliers,
which are willing to commit equity and/or resources to the project. The consortium generally
12 Variations of the BOT contract can be implemented in practice: Design-Build (DB), Design-Build-
Operate (DBO), Build-Own-Operate-Transfer (BOOT), Build-Lease-Operate-Transfer (BLOT), Build-Transfer-
Operate (BTO), etc. For instance, in a DB contract, the main benefits arise from making the private-sector party
responsible for designing (D) and building (B) a facility, e.g. a road, while allocating financial, operation, and
demand risk to the public-sector party. In a DBO contract, additional benefits arise from the fact that the private-
sector party operates (O) the facility, and then it has incentives to design and build taking future operation costs
into account.
27
sets up a single purpose entity known as Special Purpose Vehicle (SPV) to manage the
different stages of the project and to allocate risks among the parties of the SPV.
DBFO contracts can be distinguished into two different types: (i) the financially free
standing projects, such as concession contracts, where the private-sector party’s main source
of income is constituted by user fees, and (ii) the operation Private Finance Initiative (PFI)
projects, where the private-sector party mainly sells the service to the public-sector party. We
shall refer to the first type of projects as ‘concessions’, and to the second type of projects as
‘operation PFI’.
(i) Concessions
In a concession, the private-sector party has the responsibility for constructing and
financing a new asset, or modernizing and expanding an existing facility. The concessionaire
is given the right to operate the facility for a specified period, and the public-sector party
regains asset ownership when the contract expires. A typical concession is a long-term
contract with duration ranging from 25 to 30 years but it can go up beyond 60 years.
Concession contracts generally involve the construction or extension of the facility. If
the concessionaire charges final users, it collects no revenues until the building phase is
completed and service provision commences. To the extent that the public-sector party makes
no payment to the private-sector party, the construction risk is fully transferred to the private-
sector party in order to provide incentives to complete the building phase on time and to
reduce costs. However, in some cases, the public sector pays a fixed amount during the
construction. Liquidated damages can be imposed to the private-sector party if the
construction delays as a consequence of its own actions.
In a concession, the private-sector party typically finances capital expenditure, and
thus it bears the financial risk. Typically, the SPV borrows funds in capital markets to finance
investments, pledging as collateral the revenue stream that results from charging users during
the operation phase (IMF, 2004). The providers of finance look to the cash flow of the project
as the source of funds for repayments. In this regard, financial security against the SPV is
hardly sought because the SPV has minimal assets, and because the financing is without
recourse to the sponsor companies. Indeed, the objective behind large PPP projects is to
28
achieve a financial structure with as little recourse to the sponsors as possible, whilst at the
same time providing sufficient credit support so that the lenders are satisfied with the credit
risk.
Under certain financing structures, the financial risk may be an important issue for the
private-sector party. In particular, exposure to interest and exchange rate risk results from
using short-term, foreign-currency denominated debt to finance long-term, domestic currency
revenue-generating assets. This exposure may be large in economies with weak currencies
and high macroeconomic and financial instability. By transferring financial risk to the private
partner, incentives are given to reduce exposure, e.g. discouraging purchases of imported
goods produced by concessionaire-related firms, and borrowing in foreign currency (Lobina
and Hall, 2003).
During the operation phase, the SPV receives income based on the usage of the facility
assuming that the service provided meets a range of key performance indicators. There are
normally abatement clauses in the concession contract, which can penalize the SPV for
providing the services below the agreed standards. There are also penalty points, which if
accumulated to a certain level, can lead to termination of the contract for poor performance.
In the concession, the demand risk is typically transferred to the private-sector party.
The concessionaire charges consumers, e.g. toll roads, and so it bears the demand risk and
suffers from demand falling short of forecasted levels. The public-sector party may however
set a minimum revenues guarantee to reduce the risk borne by the private-sector party (see
section 4 on payment mechanisms for further details).
Changes in the legislative and regulatory framework that have effects on operation
costs and profits are likely to occur during the concession. When these changes are of a
general nature and affect the whole industry, e.g. modifications in tax legislation, the rising
costs can be either transfer to the private party or shared with the public-sector party. For
instance, indexation provisions may allow the concessionaire to pass on the rising costs to
consumers through price increases. On the other hand, when changes in law and regulation
have a specific nature and affect only the concession project, it is often the public-sector party
who bears the risk of rising costs (HM Treasury, 2007).
29
Concession contracts raise important issues for the public-sector party regarding
public accounting practices and the ownership of residual assets. The private finance aspect of
concessions can allow the public sector to finance the construction of infrastructure ‘off the
balance sheet’. The accounting treatment of payments by the public-sector party to the private
sector can make the fiscal budget look healthier than it actually is, thereby undervaluing the
cost of PPP financed infrastructure. This can bias decisions in favor of PPPs as opposed to
more traditional procurement arrangements, and also make PPPs a means to unduly transfer
costs from current to future generations (see IPPR (2001) for a discussion of how off balance
sheet considerations mattered in the early PPP projects in the UK).
The public-sector party generally owns all assets when the concession expires, both
the existing facility and the new infrastructure built by the private-sector party. Hence, the
public partner bears the residual value risk. However, for facilities which have an alternative
use if retained by the private sector (such as leisure centres and accommodation), the
ownership of the asset can be retained by the private-sector party. The advantage in this case
is that, to the extent that innovative facility design positively affect the residual value of the
facility at the end of the contract, innovation in PPPs is more likely to occur when the private-
sector party also owns the facility at the end of the contract (see Bennett and Iossa, 2006).
In fact, when the asset reverts back to the public sector once the concession expires,
the private-sector party’s incentives to maintain the asset towards the end of the contract are
attenuated, and the more so the closer the contract is to expiry. This may create a problem
because there are aspects of maintenance that cannot be perfectly verified. In this regard, PPP
contracts often include provisions establishing a final compensation to be paid to the private-
sector party at the end of the contract that is made conditional on the state of the asset verified
by a third-party auditor. However, these provisions cannot solve the problem if it is caused by
unverifiable maintenance investment whose negative effects come with a long delay and so
the auditor cannot detect any wrongdoing. Notice that if the maintenance needed to preserve
the asset value were fully verifiable, then specifying clauses making the private-sector party
responsible for them would suffice.
(ii) Operation PFI
30
In an operation PFI contract, the private-sector party sells a service directly to the
public-sector party rather than charging user fees. The public sector commits to make future
payments to the private-sector party for the service that the latter will provide using the
infrastructure. Providers of PPP schools and prisons receive their funding in this manner.
These PPPs work in almost all respects as concession contracts, the only difference being that
demand risk (e.g. the risk that the number of inmates in a prison goes down) is generally
allocated to the public sector.
Other two types of contracts are usually considered as a form of PPP: Management
and Lease contracts (see for example IMF (2004) and Renda and Schrefler (2006)). Therefore,
we briefly describe them below although the bundling of the project phases occurs to a
limited or no extent, leaving them outside our definition of PPP13.
Management contract
A management contract is an agreement where the private-sector party takes
responsibility for operating and managing state-owned assets. The private-sector party is
remunerated by the public partner, who pays fixed fees or incentive payments contingent on
meeting performance targets. Operation risk is allocated to the private partner in order to
provide incentives to undertake cost-reducing efforts, especially when fixed fees are in place.
On the contrary, neither construction risk nor financial risks are transferred to the private
partner since the public-sector party remains responsible for investing in the facility and
financing capital expenditure. The public-sector party also retains the right to charge the final
users of the service, thus bearing the demand risk if it actually chooses to charge the service
provision.
Lease contract
A lease contract is an agreement where the private-sector party takes responsibility for
operating and managing state-owned assets, but, contrary to a management contract, it
purchases from the public-sector party the right to charge the final users of the service, e.g. by
imposing tariffs. Hence, both operation risk and demand risk are transferred to the private-
sector party in order to provide incentives to reduce costs and improve service quality.
13 Tables describing the risk allocation and incentives given to the private sector in Management and
Lease contracts are included in the annex for a comparison purpose.
31
Similarly to the management contract, the lease makes the public-sector party responsible for
investing and financing.
3.6 Joint Ventures
A Joint Venture (JV) agreement is an alternative approach to a traditional PPP contract
where the public-sector party and the private-sector party create a joint company for the
provision of public services. The JV company is typically a separate legal identity that carries
out the common enterprise of the public and private partners. Both partners own the JV shares
and place representatives in a board of directors.
In a JV each party brings its own expertise on a particular area of the project and risks
are shared among the contracting parties. For example, the public-sector party may have more
expertise in dealing with planning issues, while the private-sector party may have a technical
advantage in designing issues. Being a shareholder in the JV, the public-sector party has
greater input as regards to the management and outcome of the project compared to other PPP
agreements.
The public-sector party in a JV shares not only risks but also rewards, i.e. profits.
Since it holds a controlling interest in the company, it directly enjoys the benefits of the
project for the JV. However, conflicts of interest and corporate governance problems may
arise as long as the public sector is not only a JV shareholder but also the regulator of the
sector in which the company develops. It may then happen that the public-sector party faces a
trade-off between increasing the project’s profits and safeguarding the public needs. For
example, a representative of the public-sector party serving on the board of directors may find
inconsistencies between her fiduciary duties to the JV and her responsibilities as public
servant.
3.7 Service unbundling
We have discussed the main attractiveness of a PPP when the contract bundles the
different activities involved in the project. In some cases, however, the contract un-bundles
32
services related to the management of the facility undertaken by the private-sector party. Two
types of services may be distinguished: ‘soft’ facility-management services (e.g. cleaning,
catering, security), and ‘hard’ facility-management services (e.g. routine and/or life-cycle
maintenance of buildings and equipment). Soft services neither require significant capital
outlays to be provided nor affect the value of the project’s major capital assets, whilst hard
services do involve capital outlays and affect the value of capital assets.
In choosing between bundling and unbundling of soft and hard services, there is a
trade off to consider. Bundling soft and hard facility-management services in the contract has
the advantage that, being responsible for providing both soft and hard services, the private-
sector party cannot argue availability failures are not its fault but an otherwise independent
soft service provider’s.
On the other hand, there are benefits for the public-sector party in unbundling soft and
hard services, and thus in dealing with separate soft-service providers. These benefits arise,
for example, because soft-services provision generally requires less capital investment (if any)
than hard-services provision. As discussed in more depth in section 6 this makes it desirable
to choose shorter-term contracts for soft services so as to benefit from the competitive
pressure that more frequent tenders guarantee. Hard services instead, being more capital
intensive, tend to require longer-term contracts for the protection of investment. The cost of
this is the reduction in the competitive pressure.
In addition, separate tendering for soft services favors the participation of small firms
to the tendering process and thus helps the competitive pressure; small firms tend not to
participate in large and complex tenders for PPP projects.
A decision to unbundle services is to be made considering not only the above trade off,
but also other sector- and country-specific factors. For instance, in the health sector, there are
no uniform experiences across countries regarding service unbundling: in Portugal, a PPP for
the construction and operation of a hospital typically includes the provision of clinical service,
while in the UK hospital projects do not include these services.
33
4. PAYMENT MECHANISM
A key element in every PPP contract is the payment mechanism. The structure of the
payment mechanism is the principal means for the public-sector party to allocate risks and
give incentives to the private-sector party. In order to provide appropriate incentives and
encourage good performance, payments are generally contingent on the project’s results, not
on the inputs and processes needed to deliver the service. The project’s results are to be
defined as output specifications in the contract.
4.1 Output specification, payment mechanism, and risk allocation
An output specification defines the performance standards to be achieved by the
delivery of the service. By setting the output specification, the public-sector party delegates to
the private-sector party the decision of how to organize the service provision. For the purpose
of control and monitoring the service provision, the performance standards should be clearly
defined, measurable in quantitative and qualitative terms, and verifiable by third parties like
courts or arbitrators. For example, in the water sector, an output specification could be the
frequency of water testing and threshold values of water purity, preferably a third party
implementing the tests; in the transport sector, an output specification may include a certain
level of lighting, signage, and road surface maintenance.
The output specification is important not only for control and monitoring but also to
design an incentive-oriented payment mechanism. Well-defined performance standards allow
to reward the private partner for meeting them and to penalize it for failing to do it, i.e. to
establish a system of carrots and sticks. A results-based payment mechanism has the
advantages of targeting directly the project’s outcomes the public-sector party wishes to
achieve while giving freedom to the private-sector party to select the process systems and
equipment. In other words, this payment mechanism turns the private partner’s creativity and
know-how instrumental to the pursuit of the public sector’s aims.
In specifying a payment mechanism, it is desirable to meet some basic criteria. There
should be simple and transparent rules linking payments to the measurable project
34
deliverables determined in the output specification. These rules should be carefully design to
provide strong and appropriate incentives for the private-sector party to perform. As discussed
below, the risk allocation resulting from them is a key element in this regard. The rules should
specify clearly and in detail the payments to make when performance criteria are met, and
charges to deduct when they are not. Since the payment mechanism heavily affects the
financial structure of the project, the rules should ensure the private-sector party is able to
finance the project given the risks allocated to it, and that the public-sector party can afford
the pledged payments. Otherwise, financing the project may become difficult.
To the extent that there is consistency between output specification, payment
mechanism, and risk allocation, the contract design increases the likelihood that the project
delivers value for money. Decisions to be taken in formulating a payment mechanism will be
informed by the output specification and the project’s risk assessment; similarly, the payment
mechanism may also lead to further refinement of the output specification and risk
assessment. As an introductory example of the interactions between these three elements,
consider a standard case in the transport sector. The output specification may target lighting,
signage, and road maintenance that are closely related to the quality of the service. A payment
mechanism could be based on tolls paid by users and set by the private-sector party. Thus, the
demand risk is fully transferred to the private partner who has strong incentives to improve
service quality and availability in order to raise revenues. An alternative mechanism could be
based on tolls paid by users but set by the public-sector party, who makes unitary payments to
the private-sector party. Thus, the demand risk is shared by both parties. Incentives for the
private-sector party to perform are given if the unitary payments are based on usage,
availability, and service performance.
4.2 The provision of incentives through the payment mechanisms
The economic literature on explicit procurement contracting and regulation
emphasizes the role of the payment mechanism in the allocation of risks and in the provision
of incentives to the private-sector party for cost reduction and good performance. The
literature identifies three main categories of payment methods: cost-plus, fixed-price, and
incentive payments (Laffont and Tirole, 1993).
35
4.2.1 Cost-plus payment
Under this payment mechanism, the public-sector party agrees to reimburse
documented construction and operation costs associated with the infrastructure project plus a
fixed and possibly a variable fee (thus giving rise to the name cost-plus). Formally, the
payment P would be a linear function of the cost level P =F+(1+m)C, where C is cost, F0 is
a fixed fee and m0 is the mark up the private-sector party can charge on each unit of
documented cost. Since the private-sector party is fully insured against any cost increase that
can happen in the construction and operation phases, a cost-plus payment gives no incentives
to the private-sector party to exert any extra effort to reduce cost, as lower cost implies lower
profits for the private-sector party (Albano et al., 2006a).
A cost-plus payment may therefore not be an appropriate mechanism when the
project’s total costs heavily depend on the private partner’s actions, precisely because
incentives to save on costs are weak. The mechanism has also a negative effect on the
tendering process to select a private partner when there is a high degree of uncertainty about
the bidders’ efficiency. As long as the cost-plus payment reimburses all costs, both efficient
and inefficient suppliers have an incentive to submit the same offer. But, because of the
uncertainty, the public-sector party cannot distinguish between them. Hence, it is likely that
the tendering process ends up selecting a private partner who is not the most efficient
provider, which may increase total cost.
On the other hand, the cost-plus payment has an advantage in terms of flexibility to
cope with uncertainty regarding circumstances that may arise during the project. Unforeseen
changes in the environment and modifications in the output specification may call both parties
for a contract revision. In the renegotiation, the private-sector party knows that any verifiable
cost increase resulting from the new contract terms will be reimbursed by the public-sector
party. Thus, the cost-plus mechanism narrows the scope for disputes and renegotiation costs
(Bajari and Tadelis, 2001, 2006). Also, since the private partner does not bear the risk of cost
overruns, to the extent that the firm is risk averse, payments to the private-sector party will
not need to account for a risk premium and as such will be lower than when risk is
transferred.
36
Finally, since the private-sector party is fully insured against any cost increase that can
happen in the construction and operation phases, a cost-plus payment gives good incentives to
the private-sector party to comply with the public sector needs and quality requirements.
However, these incentives may not suffice so it is always preferable for the cost-plus payment
to be adjusted for quality. To ensure incentives to deliver the quality standards contracted
upon, a scheme of deductions must be in place. The payment scheme will comprise a
component -d(Qs-Q) where (Qs-Q) is the difference between the agreed quality standard and
the provided quality, and d is a parameter that determines the size of deductions, that should
increase in the importance and costlyness of the quality dimension(s).
In the context of regulation of privatized utilities, the pricing rule in a rate of return
regime resembles a cost-plus payment in procurement contracting. In this case, the tariff level
set by the regulator is determined ensuring that the expected revenues at that price cover the
expected operation costs plus a return on the capital invested by the private partner. Since the
price received by the private-sector party covers its costs, the incentive structure of cost-plus
also applies.
4.2.2 Fixed-price payment
In this case, the public-sector party agrees to pay the private-sector party a fixed
amount for the service provision that must achieve certain quality standards. It is then critical
that a scheme of deductions is in place to guarantee that the quality standard is respected, so
that payment scheme will be P = F - d(Qs-Q), with the previously clarified notation. Since
payment do not change with cost, the private partner bears all the costs associated with the
project and fully appropriates the benefits of cost-savings activities. Hence, a fixed-price
payment gives the private-sector party strong incentives to undertake cost-reducing efforts.
Fixed-price payments perform well when potential private partners are large
companies with a diversified portfolio of activities (hence able to bear and manage substantial
risk), and the quality dimensions important for the public sector are easy to monitor, so that
deductions for quality below the standard are effective in guaranteeing that the quality
standard will be respected. Fixed-price contracts with well specified output have the
additional advantage to make competitive tendering processes extremely effective in terms of
both selecting the most efficient private partner and reducing the cost of service for the public
37
sector. In other words, well designed fixed-price contracts and competitive tendering go along
very well. On the contrary, competitive tendering may even be harmful for the public sector
when cost-plus contracts are adopted (Bajari and Tadelis, 2006).
A fixed-price payment may not be an appropriate mechanism when important quality
dimensions are hard to monitor and sustain through payment deductions, so that cost-saving
actions taken by the private-sector party are likely to cut down substantially the quality of
service. Similarly, the fixed-price payment may discourage quality-enhancing efforts that
raise the private-sector party’s costs if these efforts or quality aspects are hard to observe and
contract upon. On the other hand, if the quality dimension is easy to monitor also for third
parties, an additional payment rewarding performance above standards could be added to the
fixed-price contract, eliminating the drawback. Another advantage of fixed-price payment is
that minimizes transaction costs as no cost information is required for its implementation.
In the context of regulation, the pricing rule in a price cap regime resembles a fixed-
price payment in procurement contracting. In this case, the tariff level is chosen by the private
partner subject to a maximum cap imposed by the regulator. The private-sector party is
allowed to appropriate all the benefits resulting from cost-saving actions. Thus, price cap
regulation attempts to overcome the inefficiency in rate of return regulation that arises from
the weak incentive to undertake cost-reducing efforts when the price review is frequent and
backward-looking.14 A pure version of price cap regulation has no price review, and thus the
regulatory lag is of an infinite length. On the contrary, a pure rate of return regulation would
allow the regulator to examine cost and profitability of the private-sector party on a
continuous basis (Armstrong et al., 1998).15
4.2.3 Incentive payment
An incentive payment lies between the extreme cases of cost-plus and fixed-price. In
general, for a given quality standard the incentive payment is the sum of two components: a
fixed-amount plus a variable payment that partially compensates for the costs incurred. For
14 Price cap regulation has been implemented world-wide because of its theoretical advantages;
however, LAC countries did not account for its full range of implications (Guasch, 2004).
15 In practice, periodic price reviews are allowed both in price cap and rate of return regulations, but
typically the regulatory lag under price cap is larger than in rate of return regulation (see Price variations below).
38
example, the payment P can be a linear function of the cost level P = F + bC. When the
public-sector party bears all costs and the private-sector party is fully reimbursed, b=1 and the
incentive payment becomes a cost-plus. In contrast, the when the public-sector party bears no
cost and the private-sector party is not reimbursed, b=0 and the incentive payment becomes a
fixed-price.
By partially transferring the cost overruns risk to the private-sector party, the incentive
payment encourages it to undertake cost-reducing efforts. Since these efforts reduce both cost
and payment, but the payment decreases less than the cost, the private partner’s profits
increase. Therefore, the parameter b representing the fraction of costs born by the public-
sector party turns out to be a key element to induce the private partner to save on costs. In
particular, the lower the value of b, the larger the private-sector party’s responsibility for cost
overruns; thus, the more the private-sector party benefits from cost-reducing efforts, the
higher the power of the incentive scheme.
The public sector’s choice of the cost-sharing parameter b depends mainly on three
factors. The first factor is the ability of the private partner to bear the cost overruns risk
resulting from his degree of risk aversion. To the extent that the private-sector party is risk
averse, it will require higher overall compensation for bearing the risk of suffering unexpected
cost increases. The second factor relates to the predictability of shocks affecting the project’s
costs, which determines overall risk. When these shocks are largely unpredictable, the
private-sector party will be less willing to accept an incentive payment with a low cost-
sharing parameter b that transfers it the bulk of the cost overruns risk. It will then require
higher payments for bearing risk. The final factor is the degree in which the private partner’s
cost-reducing activities impact on the actual cost structure. The larger the expected effect of
cost-reducing activities on the project’s costs, the lower will be the payment required by the
private-sector party.
Implementing an incentive payment mechanism typically involves transaction costs.
For instance, it is costly to collect the information needed to compute the payment, to provide
accounting measures of the costs incurred, and to measure the quality of service. Thus, it may
happen that transaction costs are so large that outweigh the expected benefits of setting an
incentive payment mechanism. Under these circumstances, the public sector itself would
prefer to save on transaction costs and adopt an alternative payment scheme. For instance, it
39
may consider a payment mechanism easier to manage and less costly in terms of information,
such as a fixed-price payment.
Of course, also in the case of incentive contracts the private-sector party has incentives
to reduce the provision of costly quality, so that a quality standard should be established and
‘defended’ by an appropriate set of deductions. Taking into account deductions for quality
supplied below the contracted standard the payment scheme would be something like P = F +
bC - d(Qs-Q), where all symbols have the meaning explained earlier.
An incentive payment mechanism can also target quality-enhancing activities rather
than cost-reducing efforts for a given quality standard. Analogously to cost incentive
payments, quality incentive schemes may be used to encourage high quality in service
provision and good performance on the part of the private partner. In general, a quality-
oriented payment specifies a fixed amount for a low minimum performance level guaranteed
by harsh sanctions, like very high (total) deductions or even private-sector party replacement
in case of violation, and in addition different bonuses corresponding to higher quality levels.
Formally, the contract would look like this: P = 0 if Q<Qm;
P = F + bC +β(Q-Qm) if Q>Qm, where Qm is the minimum performance level and β
is the bonus increasing with the quality actually provided.
One benefit of using positive rewards for higher quality, in contrast to deductions (or
liquidated damages, were present) for lower quality, is that the latter appear highly ‘punitive’
for the private-sector party who sees its payment reduced, although as agreed upon in the
contract - while the former does not.16 Deductions for low quality are often not exercised by
public sector buyers. Some official argue this is also because of the fear to ‘spoil the
relationship’ with the supplier. Recasting the incentives in positive terms may not have this
drawback: the bonus may be seen as ‘something more’ the private-sector party gets only if
particularly high quality is effectively delivered, while if it is not nothing is ‘taken away’ from
the supplier.
16 Kahneman and Tversky (1973) first showed how strong is the asymmetry with which the same pay
for performance is typically evaluated when it is presented as a loss and as a gain from some reference points;
see also Camerer (2003) for many examples of this asymmetry.
40
4.3 Payment mechanism in PPPs
PPP arrangements aim at maximizing incentives for the private-sector party to take
into account whole-life costing and to undertake innovative approaches leading to cost-
reduction and quality improvement. For this purpose, the payment scheme under PPPs must
be output oriented, in the spirit of fixed price arrangements: the payment must depend on
output (and thus on the service being provided) rather than on input (the cost of the service)
and payments should only occur once the service is provided (and not before the
infrastructure is built).
Also, risk allocation plays a critical role as it is via an appropriate risk transfer that
incentives are provided. In general, risk allocation must follow two main principles: provision
of incentives and risk insurance. In particular:
(i) risk should be transferred to the party whose actions can have an impact on risk
(and hence on the likelihood that adverse events occur).
(ii) risk should be transferred to the party that is most able to bear it (i.e. that is
relatively less risk averse or better able to insure the risk)
Point (i) ensures that incentives are provided through appropriate risk transfer; point
(ii) ensures that risk averse parties are adequately insured against risk. Note that, whenever a
risk averse private-sector party is asked to bear risk for incentive purposes, it will demand a
higher compensation as a risk premium. Since in many instances, the public-sector party is
more risk neutral than the private-sector party, this suggests that risk should only be
transferred if risk transfer has a positive impact on the incentives of the private-sector party.
Inappropriately transferred risk will result in no effect on incentives and in a higher risk
premium.
The above two principles must lead the allocation of the main risks that affect PPPs,
which are cost overruns, demand risk, legislative/regulatory risk, and financial risk as we
discussed in the previous section. Payment mechanisms are means to allocate these risks
between the public and private parties. To ensure the private-sector party does not bear the
risks it cannot control, indexation and/or adjustment rules may be included in the payment
41
agreement (see Price variations below). For instance, the payment structure may allow an
automatic adjustment of tariffs for exchange rate variations.
In the practice of PPP there is a broad array of payment mechanisms that are used to
ensure:
(i) that the private-sector party receives adequate compensation for its investment
and operational cost,
(ii) that incentives are provided for the private-sector party to perform,
(iii) that risk is appropriately transferred (which also has impact on incentives)
In some payment schemes, the public-sector party makes unitary payments to the
private-sector party following different criteria, such as service usage, availability, and
performance. Some other mechanisms entitle the private-sector party to receive revenues from
the service users, e.g. user charges. There are also many schemes where payments to the
private-sector party comprise both user charges and public sector contributions. In what
follows we shall discuss each of these payment forms in isolation.
4.3.1 User charges
In a payment mechanism based on user charges, the private-sector party gets revenues
directly through charges on private end users of the infrastructure facility. By this payment
method, the public-sector party fully transfers the demand risk to the private-sector party. To
the extent that users pay the corresponding fees, there is no cross-subsidy from taxpayers to
users that might compromise public finances. Bearing the demand risk, the private-sector
party has direct incentives to improve performance in order to encourage service usage and
thus increase revenues. Moreover, it is possible to manage demand by setting a pricing rule
that charges users according to usage, e.g. in the transport sector, tolls can vary by vehicle
type and time of day.
If the private-sector party is able to control demand risk, it is efficient to fully transfer
this risk to the private-sector party by implementing a user charge-based payment mechanism.
With risk transfer, the uncertainty surrounding revenue streams and forecasts on service
demand influences the cost of capital facing the project, with more uncertainty leading to a
higher cost of capital. When user charges are the only source of revenue for the private-sector
42
party, higher cost of capital calls for a higher user charge level to reimburse investments.
Also, contract length may have to be increased. As explained above, risk transfer can help
incentives but it comes at a cost. Therefore, demand risk plays a crucial role in determining
the bankability of the project by affecting the project’s cost of capital.
The link between demand risk and tariff level affects the public sector’s choice
regarding the payment mechanism to implement. It may happen that an excessive demand risk
wipes out the project’s benefits because the resulting cost-covering user charges are so large
that users prefer to seek for alternative services, thus making the project unbankable. Under
these circumstances, the public-sector party could modify the user charge-based payment
mechanism in such a way that demand risk is not fully but partially transferred to the private-
sector party. A possible modification is a payment scheme where the public sector sets tariffs
and pays a revenue subvention to complement the fees collected by the private-sector party
from the service users. By this mechanism, the public sector can set the service tariff at a level
that maximizes social benefits of the project. In addition, the private partner and its providers
of capital reduce their exposure to demand and collection risks. Hence, the cost of capital
should be lower, and also the amount of support needed from the public sector.
Criteria to set tariffs
When the private partner’s revenues in a PPP project are based on user chargers, tariff
setting criteria become a relevant issue. Tariff levels should follow three main criteria:
(i) allocative efficiency
(ii) bankability of the project
(iii) distribution considerations
(iv) other factors
(i) Allocative efficiency calls for a pricing rule which sets tariffs according to
marginal costs. From the public sector perspective, the tariff setting should take into account
the social costs and benefits associated with the project. Service provision involves not only
the private production costs facing the private-sector party but also any cost imposed on other
activities, e.g. externalities like pollution. Hence, a cost-covering tariff based on both private
43
and social costs and benefits may be higher or lower than the tariff level that maximizes the
net sum of private benefit and cost.
(ii) When user charges are the only source of revenue for the private-sector party, the
level of user charges should be also consistent with recouping initial investments and sunk
costs.17 The tariff level should be such that revenues cover operation costs and investments in
the service provision, allowing a commercial rate of return. The effect on revenues of
changing user charges depends on the price-elasticity of demand for the service.18 In turn, the
volume of demand affects the operation costs of providing the service.
(iii) To the extent that the public sector is concerned with the project’s social benefits,
e.g. improvement of health from installing proper sewerage systems, pricing service below
marginal costs might be the preferred pricing structure. But to ensure bankability and
incentives to invest in service expansion and quality improvements for the private partner, the
public sector must subsidize the project where it is at a loss.19 Distributional issues arise also
within actual users as long as they differ in terms of income and service usage. Seeking equity
of treatment regarding user charges, differentiated tariffs can be set according to types of
services, categories of consumers, etc. It is a common practice to establish systems of cross-
subsidies where some groups pay tariffs below costs and others pay tariffs above costs to
compensate. But from the perspective of efficiency, these systems have negative
17 Projects with increasing returns to scale display average cost below marginal cost, so pricing
according to the latter leads to economic losses.
18 For example, in the water sector, industrial and residential demands are inelastic, thus revenues
increase along with tariffs (Howe, 2003).
19 International experience in the developing world clearly demonstrates that pro-poor arguments fail to
justify subsidies to infrastructure services because the poorest typically have no access to them (Kerf, 1998).
However, empirical evidence suggests it is difficult to reconcile the profit motive with delivering services to the
poor without public subsidies or specific contractual agreements. For instance, in the water sector, serving poor
households is not profitable since these cannot afford to pay for the connection or to consume enough water to
cover the costs of service provision. Private partners have developed different approaches to this problem,
ranging from defining contracts in a way to ring-fence profitable users to introducing cross-subsidization of
tariffs. Interestingly, arguments supporting water privatization included a criticism of the public sector for
subsidizing excessively and failing to set cost-recovery tariff levels, but nowadays subsidies through public
finance are seen as a key to sustain the presence of private partners (Lobina and Hall, 2003).
44
consequences to the extent that consumption patterns are distorted for users enjoying
subsidies, and for users contributing to finance them (Kerf, 1998).
(iv) Tariff setting should cope also with other factors such as the collection risk, i.e.
the risk that users of the service try to avoid paying the user charge. Since the users’ incentive
not to pay increases along with the tariff level, a too high user charge may lead to the private-
sector party taking excessive collection risk.
A number of subvention schemes may be offered to the private-sector party when the
expected revenues are not sufficient to ensure project bankability.
First, capital expenditures contributions (capex) may be offered by the public-sector
party taking the form of: (i) capital grants; (ii) loans; (iii) equity. In order to avoid public
sector exposure to construction risk, the construction phase should be funded by the private-
sector party, and capex contributions should be injected only after project completion. Capital
grants minimize conflicts of interest, but the public-sector party should take into account that
a capital grant is a sunk cost which is not refunded in the event of contract termination.
Moreover, as long as senior debt will be repaid first in case of private-sector party default, the
use of capex contributions may not be appropriate.
Second, revenue support to improve the private partner overall cash flow may be
structured as depending on the number of users, thus being equivalent to user charges in terms
of incentives and risk transfer. Alternatively, it could be set in such a way that payments
decrease over time, thus giving incentives for the private-sector party to encourage usage; or
that payments decrease as usage level increases, thus avoiding windfall earnings under
unexpected increase in demand. Revenue support can also take the form of revenue
guarantees payable only in years where revenues from user charges fall short of a specified
level.
Third, debt guarantees could be provided by the public-sector party, thus having no
effects on the fiscal budget as long as it is not called upon.
4.3.2 Usage payments
A payment mechanism based on usage can be seen as a variant of user charges, but
where it is the public-sector party that pays the private-sector party instead of service users. In
45
this scheme, the public-sector party sets tariffs to be charged on users according to its
objectives. After receiving the associated revenues, the public-sector party makes unitary
payments to the private partner depending on the actual usage level.
In most cases, there are bands for usage levels determining the payments, and thus
setting limits to the demand risk transferred to the private partner. Usage payments imply less
risk for the private-sector party as final users do not pay for the service, so in principle
demand levels are not affected by income shock. As it was mentioned above, a lower
exposure to demand risk leads to a lower cost of capital for the project. Using bands at low
usage levels bounds the risk to the private-sector party that service demand is lower than
expected. In practice, lower bands provide a certain minimum usage payment to cover debt
service, but not to ensure a positive return on equity. On the other hand, using bands at high
usage levels caps the number of users for which the public-sector party should make
payments, thus bounding the public-sector party’s financial liability.
A usage payment has advantages in terms of incentive design since the private
partner’s actions regarding service availability and quality affect the usage level on which the
payment depends. Besides, the scheme is easy to implement because it is quite consistent with
the traditional practices of financing infrastructure projects out of public funds. However, the
scheme involves financial risks for the public-sector party because payments are uncertain ex
ante, and this can cause difficulties in budget planning. Usage payment also involves
distributional issues: service users may pay a relatively low fee while taxpayers end up
subsidizing them.
4.3.3 Availability payments
In a payment mechanism based on availability, the public-sector party rewards the
private-sector party for making the service available regardless of the actual service usage.
The payment scheme typically involves also deductions if the private partner fails to comply
with availability targets. In any case, objective measures defining service availability should
be established in the contract, e.g. lanes ready-to-use in roads and capacity to undertake water
treatment works. In addition, deductions should depend on whether the private-sector party
can or cannot control the events causing unavailability, the private partner’s effort to provide
alternative services, the spread and recurrence of unavailability episodes, and the rectification
46
period needed. The weightings of deductions are important: if deductions are too low, it may
be convenient for the private-sector party to under-perform; if they are too high, risks increase
and the contract may require a higher pricing (HM Treasury, 2007).
In practice, availability payments are used in projects related to infrastructure
construction and management. There is no demand risk for the private-sector party, so the
cost of capital is likely to be lower. Moreover, since the construction phase should be
completed before services can be made available, the invested funds represent a fixed cost at
the outset. In terms of incentives, an availability payment encourages efforts to ensure
potential service provision. It could be coupled with user charges or usage payments to the
extent that it is efficient to allocate some demand risk to the private-sector party to promote
quality- enhancing efforts.
4.3.4 Performance payments
A payment mechanism based on performance rewards the private-sector party for
meeting certain standards of the service provision. In practice, performance payments
complement other payment method such as usage or availability. The scheme sets charges for
performance failures which are deducted from unitary payments. Alternatively, as discussed
earlier, the contract could establish ‘bonuses’ to be awarded if and only if certain target
performance levels are reached.
Several issues must be taken into account when the public-sector party uses
performance payments. In order to stimulate quality innovations, the contract should specified
objective measures defining service performance, e.g. quality of water in terms of pH levels,
on-time delivery of the service, and adequate road signs. In particular, the standards of
performance must be monitored at a low cost.
The payment structure must define the consequences of a failure to meet the required
quality level of the service. The simplest approach is to categorize various types of
performance shortcomings and use a grid of monetary deductions. An alternative, two-stage
approach is to assign penalty points to the private-sector party any time a performance failure
occurs, eventually attaching more points to a serious and recurrent failure, and to set a rule
that translates points into monetary deductions. Generally, deductions are made when a
47
certain number of points have been assigned to the private-sector party within a defined time
period. To provide incentives to perform, there should be an adequate calibration between the
seriousness and frequency of a failure, the number of penalty points assigned where
applicable, and the financial impact of deductions on the private partner (HM Treasury,
2007).
When the private partner’s performance is continuously poor and early failures are not
rectified within a certain period, the public sector may include a ‘ratchet mechanism’ in
payment deductions. For example, a simple ratchet mechanism increases the number of
penalty points incurred for any given failure that is repeatedly observed over a time period.
The ratchet mechanism informs other parties involved in the contract, e.g. lenders of capital,
the private-sector party fails to perform. Thus, it encourages the private partner to take early
remedy actions to avoid reputation losses.
Other guarantees are available to incentive the private-sector party to comply with the
contract terms, such as liquidated damages and performance bonds. Moreover, non-financial
mechanisms can be implemented, ranging from formal warnings to contract termination as we
will discuss in section 7.
4.3.5 Liquidated damages and performance bonds
When the private-sector party fails to deliver the service on time or to meet the
performance standards, the public-sector party may use contractual protections such as
liquidated damages and performance bonds. Liquidated damages are rules that set in advance
an amount to be paid by the private-sector party to compensate for the estimated economic
losses incurred by the beneficiary in case of certain breaches of contract. Liquidated damages
are often calculated as a percentage of the contract price that depends on the project
complexity, e.g. it ranges from 10 to 15 percent in gas pipelines projects, and from 35 to 40
percent in coal-fired power generation (Kerf, 1998).
Facing liquidated damages, the private partner may increase the price required for the
project to ensure itself ex ante against the risk of late service delivery. The private partner will
increase its tendering price anticipating any contingency that makes liquidated damages
payable, e.g. the private-sector party can raise the estimated construction cost or require a
48
longer construction period. Therefore, using liquidated damages is likely to increase the
unitary payments or user charges the private-sector party receives, and also to lengthen the
time schedule of the project.
The HM Treasury (2007) suggests that since liquidated damages can compensate for
the beneficiary’s economic losses but at the same time they have a negative effect on pricing,
the use of liquidated damages in addition to deductions or rewards for quality is cost-effective
only as long as the expected losses suffered by the beneficiary are greater than the payment
increase required by the private-sector party. However, this does not take into account the
gain that liquidated damages may create in terms providing incentives to invest and perform
when it is not possible to introduce sufficiently strong deductions or bonuses linked to quality.
By increasing incentives, liquidated damages can be suitable also when the expected losses
suffered by the beneficiary are lower than the payment increase required by the private-sector
party.
Performance or ‘surety’ bonds are used as a guarantee of construction completion in
case the private-sector party goes bankrupt and thus the project remains unfinished (Engel et
al., 2006).20 The bond is issued by a bank, an insurance company or a specialized ‘surety
bonds’ company in favor of the public-sector party and guarantees a payment (typically
around 10 percent of the value of the construction contract) or even the private partner
replacement and project completion (in the case of surety bond companies) in case of non
performance by the original private partner. When the public-sector party requires a
performance bond to the private-sector party, the risk of losing the bond encourages the
private-sector party to comply (Gausch, 2004; Engel et al. 2006). In addition, it prevents the
private-sector party from ‘walking away’ from the project if disputes arise.21 On the other
20 In the US in 1990-1997, more than 80,000 private partners went bankrupt leaving unfinished project
with liabilities of USD 21 billion (Engel et al., 2006).
21 For example, in the Tucuman water project (Argentina), a conflict arose between the private partner
and the authorities on matters of water quality and pricing. The performance bond posted by the private-sector
party did have an impact in solving the conflict (Kerf, 1998). However, in the water sector, it is debatable
whether performance bonds are an effective device to make private partners assume risk for poor performance
(Lobina and Hall, 2003). When a public authority is relatively weak, e.g. a local authority dealing with a
transnational firm, it may be reluctant to call in a performance bond. The public sector may fear retaliation from
49
hand, as with liquidated damages, a private partner asked to post a performance bond is likely
to react by passing through the cost and time schedule the risk of loosing the bond in the
future (HM Treasury, 2007).
Contractual protection payments are not recommended if the public sector does not
suffer any significant loss as a consequence of the private partner’s bad performance, e.g.
because payment deductions for poor quality have been sufficiently strong. As it has been
argued, both liquidated damages and performance bonds are likely to increase the price and
lengthen the project schedule, so the public sector should consider their effects on the value
for money (HM Treasury, 2007).
4.4 Contractual incompleteness, non contractible quality and reputational
forces
All was written earlier in this section assumed that all output specifications are simple
to monitor, clearly defined in the contract, and easy to verify by a third party (court or
arbitrator) in case of disagreement/litigation. Under this condition, the acquisition of service
at the root of the PPP can be seen as an acquisition of standardized supply, which is
reasonable to pursue with a detailed contract specification and competitive tendering (Albano
et al., 2006a; Bajari and Tadelis, 2006).
In any kind of service, however, there are qualitative aspects or tasks that are too
costly or simply impossible to specify in a contract so that they can be monitored and verified
by third parties. For example, it is very hard to specify a contractually usable standard for how
‘good’ is food provided in a canteen or how ‘smart’ are the doctors and nurses providing
hospital services. Tasks with poorly measurable or verifiable quality dimension, if bundled in
the same contract with other easily measurable tasks, may lead an opportunistic provider to
underperform on these non-measurable tasks to cut cost and maximize profits. If only the
main contract is there regulating the service, this may take place without any consequence for
the supplier (in terms of deduction) because the fall in quality cannot be verified by who is in
charge of enforcing the contract.
the private-sector party who could withhold payments of lease or concession fees, or threaten to discontinue the
service provision.
50
These types of tasks -sometimes very important- cannot be regulated by any of the
contractual forms discussed earlier. Other forms of incentives should be adopted. First, in
some cases it is advisable to ‘unbundle’ the PPP by removing tasks of this type from the main
PPP contract, and have them regulated separately (Kerr, 1975; Holmstrom and Milgrom,
1991).
Second, customer satisfaction surveys run by a third independent party could be run
regularly, and deductions for poor quality should depend on the level of customer satisfaction
achieved (Albano et al. 2006a).
Third, contract duration on these tasks should be very short, and renewal should be
used as an incentive, i.e. conditioning contract renewal to the achievement of the minimum
level of customer satisfaction required (Calzolari and Spagnolo, 2006). As we will discuss in
the section on contract duration, under certain circumstances there is a case for separating soft
and hard services and setting out contracts for them that have different duration (unless it is
customized to the financing or results in cost saving by economy of scope). Unbundling and
different contract duration can in fact help to ensure that contract duration follows incentives
and efficiency considerations for each type of service.
More generally, to ensure an overall high quality level of service the public sector
agency in charge of the procurement contract can use past performance information (PPI) as
selection criteria in new tenders (Kelman, 1990). The public sector could set up a system of
private sector performance evaluation on a national basis, as done in the US for standard
procurement, requiring public sector agents in charge of each PPP operate a recurrent
evaluation of private-sector party performance, centered on results from independent
customer satisfaction surveys. The database would then be made available to other public
sector agents that are selecting private partners for related projects, and these should be
required to substantially penalize poorly performing private partners in the new tendering
process (US DISA, 2003).
Past performance information may both be verifiable and non-verifiable or ‘soft’ (like
‘satisfaction indicators’). Particularly for the second type of PPI these schemes rely on the
public sector agencies being accountable, as it may be difficult to verify whether the
information they provided are entirely truthful. To reduce dependence on public sector
51
officials accountability, collection of ‘soft’ information may be delegated to independent third
parties, firms specialized in customer satisfaction surveys, and may involved a large number
of anonymous interviews to users.
This system would act as a high quality ‘reputation’ or ‘brand’ acts and sustains high
quality in standard markets (Calzolari and Spagnolo, 2006). For firms also active on the
private market, the incentive power of customer satisfaction surveys and past performance
evaluations could be further increased by making their results easily accessible to the public
(Dellarocas et al., 2006). The well performing firms will then be able to increase their
deserved reputation for high quality, while the badly performing firms will be publicly
unveiled and avoided by future public and private customers. This certification role’ could
further strengthen incentives for the provision of overall high quality.
Of course, if poorly observable tasks or quality dimension are dominant in a planned
acquisition, then the acquisition is not suitable for an output oriented PPP, and will be much
better handled through an input oriented cost-plus contract awarded through negotiations
(Bajari and Tadelis, 2006).
4.5 Price variations
To the extent that the service price is fixed during long periods of time, there is the
risk that unforeseen changes in costs of major inputs, service requirements, or the regulatory
environment render that price insufficient to cover operation costs and financial obligations.
This risk affects both tariffs paid by final users and unitary payments paid by the public-sector
party as long as they remain unchanged over time, and it is apparent in long-term contracts
such as a 30-year concession. Thus, it is convenient for both parties to introduce provisions in
the contract design to adjust the price in certain specified circumstances.
By reducing the risk exposure of the private-sector party, the provisions are likely to
increase the initial bid price paid by the private-sector party in a concession (or lower the
initial price paid to the private-sector party in a procurement agreement) in the tendering
process.
52
Besides, the provisions help in reducing the cost-covering service charge, either the
tariff or the unitary payments, that is needed to ensure bankability. This is so because, when
projections on demand and/or costs are highly uncertain, the private-sector party may seek to
negotiate a relatively high service charge, thus requesting a sort of premium to compensate it
for the possibility that the price fixed in the contract will become insufficient in the future.
Under these circumstances, allowing the service charge to adjust according to the (observable)
costs outside the private partner’s control, the price adjustment clauses help in increasing the
initial bid price paid by the private-sector party in a concession.22
The price-change provisions should be carefully formulated taking risk allocation and
transaction cost issues into account. As we have seen before, there is no gain from
transferring risk to the private-sector party that it cannot control. Thus, the clauses should
provide the private partner with a hedge against unforeseeable cost overruns outside its
control, while maintaining the incentives to undertake cost-reducing efforts and to seek
efficiency.23
Further, the price adjustment rule and procedures should try to economize on the costs
of collecting and processing information, and to minimize the scope for future disputes on
price changes.
4.5.1 Inflation indexation
Inflation indexation is a typical provision to adjust tariffs and unitary payments in a
continuous basis. Since price and cost indexes used are publicly available, this provision
economizes on transaction costs. To adjust service charges by indexation, a choice should be
made concerning the price or cost index to apply, the proportion of tariff subject to
adjustment, and whether the indexation rule itself will be revised periodically. Regarding the
price or cost index to be applied, there is an important trade off to consider. If the index used
22 In addition, as long as most of the cost variation experienced by the private-sector party is of an
industry-wide nature, e.g. changes in prices of major inputs used in the sector, the price adjustment provisions
may prevent the difference between the service charge and the market price of similar services from varying
excessively over time (HM Treasury, 2007).
23 In terms of the formula for incentive payment mechanism, the service charge could be determined by
P = F+bCc+Cn, where Cc are costs controlled by the private-sector party, and Cn are non-controlled costs. More
generally, the provisions should not distort operating, investment, and financial decisions of the private-sector
party.
53
is not specific to the sector, e.g. the retail price index (RPI), it is likely that its variations do
not mimic the changes in the private partner’s non-controlled costs. This may lead to price
adjustments failing to track the relevant cost changes, thus distorting incentives. On the other
hand, if the index is too industry-specific, it is likely that its variations could be influenced by
the tariff level of the regulated service, and so manipulated by the private-sector party itself
(Armstrong et al., 1994). The proportion of tariff subject to indexation also matters: to
provide a proper hedge against observable, non-controlled cost overruns without distorting
incentives, the proportion of tariff subject to indexation should match the proportion of
variable costs in total costs (HM Treasury, 2007). Since fixed costs are known in advance,
involving no risk, the proportion of the tariff that covers them should not be indexed.
4.5.2 Price reviews, market testing, and benchmarking
Provisions could also allow for a periodic price review taking place once during a
certain specified period of time, say every three or five years.24 These reviews are useful to
adequate tariffs or unitary payments to long-run changes in the private partner’s uncontrolled
costs, e.g. technological progress modifying the cost structure, introduction of new inputs
whose prices are not tracked closely by the available indexes, etc.
The ‘value testing’ provision, for instance, establishes how to adjust prices
periodically according to the evolution of the costs of service provision. In ‘value testing’
procedures, information on costs is collected directly, so implementing the procedures
involves higher transaction costs compared to a simple, mechanic inflation indexation. But
there is an important advantage: an adjustment of service charges based on accurate, specific
information on costs closely tracks the private partner’s uncontrolled costs, and thus provides
incentives to control costs and properly select suppliers.
In practice, the main procedures conducted to test value are ‘market testing’ and
‘benchmarking’. The ‘market testing’ aims at ascertaining the market value of the main inputs
involved in providing the service through a re-tendering among potential suppliers of these
inputs. The information collected is used in the price review for tariffs or unitary payments. In
the ‘benchmarking’, information on market prices of inputs is gathered to compare the private
partner’s costs and adjust prices.
24 In LAC countries, the duration of regulatory lag is usually 5 years (Guasch, 2004).
54
‘Market testing’ and ‘benchmarking’ are useful procedures for soft services when
there are competitive markets providing comparable data. In contrast, to the extent that
markets are concentrated, the current supplier is unlikely to be undercut by other competitors.
Furthermore, these procedures are vulnerable to collusion between the private-sector party
and the supplier, where the supplier could choose not to participate (or make a good offer) in
the tendering process in exchange for a monetary reward from the private-sector party. Thus,
the price review may simply update tariffs or unitary payments according to the current
suppliers’ prices (HM Treasury, 2007).25 When value testing procedures were implemented in
the UK, they proved to be a lengthy process taking around 2 years to be completed, and some
difficulties arose in finding suitable benchmark data to compare with.
In implementing these procedures, a choice should be made on when price reviews
and testing will take place. In fact, a trade off exists regarding the length of the regulatory lag.
If the first review is planned to occur in an early phase of the project, a potential operator
could bid aggressively, offering a low tariff and expecting the review to increase it soon after
the contract is awarded (this kind of incentive distortion arising from expected contract
revisions will be discussed in detail in section 5). On the other hand, if there is a lengthy
period before the first review, the private-sector party is largely exposed to the risk of
misalignments between the initial fixed price and the operation costs, and thus it may require
a higher service charge level (HM Treasury, 2007).
4.5.3 Tariff regulation and price adjustment clauses
In practice, price adjustment clauses are an important element in the tariff regulation
for services where the private partner’s revenues result from charging final users. If the price
review is frequent and the price adjustment is backward-looking, i.e. the regulatory lag is
small and past changes in costs are considered to compute a new tariff level, the private-sector
party has little incentive to undertake cost-reducing efforts. This is so because, if the private-
sector party anticipates that any cost reduction in the present will lead to a tariff reduction in
the future, it may prefer not to exert cost-savings efforts. The literature refers to this incentive
25 To address the problem of concentrated markets, a ‘benchmarking’ procedure could be conducted
using cost information of comparable firms in other regions or countries, or cost estimates resulting from a
simulated model that formalizes the behaviour of a hypothetical efficient firm. In practice, a model of an
efficient firm is used to regulate the electricity sector in Chile (Kerf, 1998; Di Tella and Dyck, 2002) and Peru
(Guasch, 2004).
55
distortion phenomenon as the ‘ratchet-effect’ (Milgrom and Roberts, 1992). The situation is
the opposite when the regulatory lag is large and the price adjustment is forward-looking, i.e.
expected future changes in costs are considered in tariff determination. In this case, the
private-sector party do have incentives to undertake cost-savings efforts because it can reap
the benefits from lower than expected costs until the next price review (Laffont and Tirole,
1993; Armstrong et al., 1994).26
In the review, the tariff cap should be adjusted by inflation and efficiency gains in a
forward-looking way, i.e. using expected values for inflation and productivity growth, so that
ratchet effects are avoided. By introducing a forward-looking price indexation, the regulator
attempts to compensate the private-sector party for expected costs overruns outside its
control. And by subtracting expected efficiency gains, a transfer is made to consumers
through a lower relative price of the service. In addition, excessive profit-making by the
private-sector party is deterred; to be precise, the private partner appropriates additional
benefits to the extent that cost-savings efforts increase efficiency above the expected level
already discounted in the current tariff.27
4.5.4 Price adjustment and strategic behavior (gaming)
To the extent that price reviews use costs information provided by the firm itself, the
firm has incentives to misreport its costs and to manipulate the information provided in such a
way that the subsequent price adjustment favors it. In the literature, this is known as ‘strategic
behavior’ or ‘gaming’. In practice, there are a number of gaming actions. For instance, firms
having private information about costs can either report them truly or make ‘creative
accounting’ shifting cost across periods and categories, e.g. cost padding. ‘Creative
accounting’ allows for information manipulation, but it may involve costs for the firm since it
has to manage at least two parallel accounting books (Laffont and Tirole, 1993)
26 In practice, however, ‘expected’ costs tend to be computed using projections based on historical costs,
so the line dividing forward- and backward-looking tariff determination blurs.
27 From this observation, an implication is derived for choosing the optimal regulatory lag in a price cap
regulation. A trade-off arises when a lengthy lag is considered: on the one hand, the private-sector party has
incentives to undertake cost-reducing efforts because it might exceed the expected efficiency gains; on the other
hand, there is a higher probability of allocative inefficiency arising from the excessive profits that will be
observed if the private-sector party actually exceeds the expected efficiency gains (Armstrong et al., 1994).
56
As we have seen, in ‘value testing’ procedures, the public-sector party makes an effort
to collect information on costs directly from markets trading the main inputs. But under
certain circumstances, even these procedures are vulnerable to gaming. For instance, when the
market is highly concentrated and a simulated model is used as a benchmark, it is likely that
the information needed for model calibration is heavily influenced by the costs of the very
firm. Thus, the firm could game the public-sector party providing distorted information that
leads to price adjustments favorable to it.
This problem is well known in privatized industries under by price cap regulation,
where the adjustment of the cap requires information to estimate prospective efficiency gains.
Benchmarking in regulated sectors is vulnerable to gaming because these sectors are typically
concentrated markets. Thus, since there is a small number of firms reporting information, the
probability of each firm’s reports influencing the future price caps is high.
Di Tella and Dyck (2002) document a case of gaming in the Chilean electricity sector.
The authors observe a U-shaped pattern for the reported costs-to-revenues ratio in the 4-year
periods between cap reviews. Reductions in the ratio reached 1.2 % a year, and were reverted
in the years prior to regulatory reviews. The authors show that the strongest reported cost
reductions occur early in a regulatory period. But, on the basis of stock market information
complementing the costs and revenues reported to the regulator, they argue that firms may
have shifted the incurred costs towards the end of the regulatory period by engaging in
‘creative accounting’. Firms had incentives to do so as long as they expected the high cost
level reported at the end of the period to justify a higher price cap in the next review. In other
words, firm had a long time to benefit from cost reductions, and revealed cost later in an
attempt to influence future price caps, gaming the system.
57
5. FLEXIBILITY AND RENEGOTIATION
Several recent reports on PPP contracting highlight the need for enhanced contractual
flexibility, in particular aimed at taking into account possible changes in user needs that – in
the presence of rigid contracts - have sometimes triggered very costly contract renegotiation
processes.28 Enhanced flexibility, in particular directed to accommodate changes in user
needs, is important for the long-term projects typical of PPP, and may be achievable through
well designed change-management contractual clauses necessary to limit potential abuses.
However, enhanced flexibility will inevitably come at the cost of lower predictability and
higher risk for the investing private-sector party, and of reduced effectiveness of the
competitive selection process.
Below we briefly discuss the trade-offs between flexibility, investment protection and
predictability, and contract renegotiation. We then discuss the risks that large post-award
contractual changes and renegotiation in general pose on the effectiveness of the procurement,
and the consequent need to limit and structure such processes as much as possible. We
conclude discussing how contract design can help in minimizing the costs arising from
contractual flexibility. This section will not focus on how contract duration can be limited in
order to enhance flexibility (on this, see Section 6), nor on anticipated changes in payments
such as tariff indexation (on this, see Section 4).
5.1 General Trade offs: Flexibility, Predictability and Renegotiation
PPP procurements often develop along a long time horizon, 25-30 years or more. In
such long period many things can change, so that there is a need for flexibility and adaptation
of the contractual relationship far greater than in a more standard type of procurement.
Some of the possible changes can be anticipated, in which case they may be specified
in and regulated by the initial contract (e.g. changes in capacity). Other possible changes,
however, may be hard to specify in the original contract, or may be totally unexpected (e.g.
new incoming technologies that change substantially users’ needs).
28 See e.g. HM Treasury (2006, ch.5).
58
Contracts are legally binding instruments that prescribe a course of action the
contracting parties agreed upon, and that make it costly for each party to unilaterally change
that course (i.e. contracts make it costly to violate the promises they contain). Contracts,
therefore, have the precise objective to reduce the parties’ ability to choose in the future, to
‘reduce their flexibility’, in order to increase the predictability of the future actions. This
ensures that each party can better foresee what the other party will do, and so can act relying
on that forecast rather than under a much higher uncertainty.
In particular, when a party must undertake non-contractible investments that are
specific to the business relationship, a contract protects it by limiting the other party’s ability
to ‘hold up’ the investing party asking for new terms of trade after the first party has
committed its investment and is ‘locked in’ with the contractual relationship.
Moreover, the benefits of competition between potential sellers/suppliers at the
selection stage can only accrue to a buyer when the object of the competition – be it a good or
a service - is well defined and cannot be easily modified after the competitive process is over.
In procurement, in particular, a contractual definition of the procured good or service that is as
clear and complete as possible, along with a strong and credible commitment not to change
that definition after the contract is awarded unless extreme unexpected events occur, are
essential requirements for a competitive bidding process to be effective in selecting the best
supplier and offer at the efficient terms of trade (i.e., the so called ‘sanctity of the bid’ is
crucial to the whole procurement).
Besides the benefits of generating certainty, fostering investments, and allowing for
effective competition, the rigidity generated by contracts also brings about the costs resulting
from reducing the parties’ ability to adapt to novel circumstances that could not be envisaged
at the contract drafting stage.
The costs of reduced flexibility are larger the less the contract itself is adaptable to
changes in the environment, that is, the fewer possible contingencies have been anticipated,
described and regulated by the initial contract; and the more the environment and the parties’
objectives may change in an unanticipated way along the contract life.
An important trade off between the benefits from contractual protection and the
predictability and costs of contractual rigidity is therefore present in any long-term transaction
and is inevitable.
59
The larger the investment required from the parties, the larger the benefits from
contract rigidity relative to the cost of lost flexibility. As will be discussed in section 6 on
contract duration, the longer the duration of a contract and the large the scope for
unforecastable changes in the environment and the objectives of the parties, the larger the cost
of lost flexibility relative to the benefits from enhanced predictability.
In PPPs, in particular, the large investment at the core of the project is the source of
gains from contractual completeness and rigidity, while the long horizon for the service
provision is at the root of the need for enhanced flexibility.
This trade off can be softened by increasing the initial investment in forecasting future
contingencies (e.g. possible changes in knowledge or technology) and in describing and
regulating them in the contract. That is, the flexibility/predictability trade off can be
attenuated by incurring in the cost of increasing built-in flexibility/adaptability of the contract.
An example of this are the built-in adjustment mechanisms for tariffs and other payments, like
the indexation clauses linking payments to price or cost indexes.
There are limits, however, to what a costly investment in a more complex, complete,
and adaptable contract design can do to enhance flexibility, because the kind of built-in
flexibility obtained by enriching a contract can only adapt to changes in the environment that
are verifiable by a third party like a court.
In particular, when the need for flexibility derives from changes in the contracting
parties objectives/preferences, that are hard and sometimes impossible to verify by a court,
then there is little more that a contract can do than allocating all authority to one of the parties
(i.e. the right to decide what to do) and prescribing limits to what can be required to the other
party and to cost-based compensation.
Of course, when substantial changes occur that make the original plan technologically
or economically inappropriate, any contractual agreement can be modified by the mutual
consent of the parties. Contract renegotiation is a (set of) change(s) in the original contract
terms that is agreed upon by both parties and formalized by legally binding changes in the
contract terms, and is an element that can always increase flexibility if the parties agree it is
needed.29
29 Renegotiation may also include additional (complementary) contracting which completes and
integrates the original contract.
60
However, contract renegotiation typically occurs in a very different situation than the
original contract drafting and awarding did, in particular in a bilateral ‘lock-in’ rather than in
a multilateral competitive situation. Because of this, contract renegotiation, as well as any
analogous form of contractual change (modification or simply enrichment) taking place after
the contract is signed, are subject to the risk of abuse either from the private partner (thanks to
its now strong bargaining position) or from both contracting parties when one of them is a
public entity managing third party money (see the section below).
Moreover, potential private partners’ expectation or even hope to abuse renegotiation
tends to distort the functioning of the competitive selection process, thereby implying high
costs in terms of wrong selection of supplier, projects, and terms of trade. Therefore, as we
will argue, contracts need to be carefully drafted to prevent renegotiation and analogous
contractual changes as far as feasible. In addition, when flexibility in terms of contractual
changes is absolutely necessary, contracts need to be carefully drafted to prevent abuses
through such contract changes, and more importantly, to prevent the expectation or hope of
such abuses from potential private partners.
5.2 More on the costs of abusing flexibility and renegotiation
Contract theory predicts that as long as the contract is complete, i.e. it specifies all
possible contingencies and includes a full description of the renegotiation process, the
contract will never actually need to be renegotiated (it is renegotiation-proof). This result does
not depend on the ability of both parties to commit not to renegotiate the contract in the
future, but in reality it is hard to imagine a complete contract, particularly for complex and
long-lasting business relations. Renegotiation, therefore, cannot be excluded. When it is not
abused, renegotiation is efficiency-enhancing an typically arises because the contract is
incomplete to some extent, as not all possible future contingencies can be forecasted and
regulated by the contract at reasonable cost.
The literature offers a number of reasons why contracts are incomplete and they
cannot specify all possible contingencies. First, some circumstances that may affect the
contract terms are not predictable ex ante, so unforeseen events cannot be incorporated in the
initial contract. Second, there may be some non-contractible contingencies, either because
they depend on non-observable variables, such as effort exerted by the private-sector party, or
because they are non-verifiable by third parties and thus cannot be enforced in court. Third, it
61
may be too costly to write down clauses accounting for all possible events affecting the
contractual relation, so the parties must decide which contingencies to include and which not,
in order to save transaction costs. Fourth, economic agents may display bounded rationality,
being unable to identify and order all the contingencies that are truly relevant for the
contractual relation. In this regard, agents may learn along the contract period and rely on
renegotiation or additional side-contracting to correct past decisions that turned out to be
wrong or to complete aspects that were left unregulated.
In a context of incomplete contracts, and looking at it ex post, renegotiation is a
‘Pareto-improving’ mechanism to redress inefficiencies caused by incompleteness or
mistakes. To be specific, renegotiation provides the parties with the opportunity to adequate
the original contract terms when unforeseen events occur, agents learn more on project
design, new information becomes available, etc.30 Thus, the higher the degree of contract
incompleteness, the more likely renegotiations or additional side (complementary) contracting
will occur.
Contract renegotiations are also affected by the interactions between project
complexity, contract incompleteness, and features of the payment mechanisms. In this regard,
Bajari and Tadelis (2001, 2006) argue that, in an optimal contract design, there is a trade off
between the costs of ex post renegotiation of contracts with different payment mechanisms,
and the ex ante incentives provided by these mechanisms. According to the authors, contract
design costs are increasing both in the desired degree of completeness and in the complexity
of the project to be implemented.
Consider two possible payment mechanisms for any given contract: a fixed-price and
a cost-plus payment scheme. It has been discussed in section 4 that the fixed-price mechanism
is a high-powered incentive scheme, but it makes renegotiations more costly: for the public-
sector party, locked-in by the well specified contractual obligations of fixed-price contracts;
30 Despite the fact that revisions due to unforeseen events allow for improving the contract terms, the
negotiations may fail to accomplish this aim, and end up terminating the contractual relation. For instance, in a
tender to manage a motorway in Hungary, a key variable in the award criteria was the tariff level required by the
private partner. Thus, the DBFO contract was expected to allocate the traffic risk to the private-sector party.
However, the award criteria induced excessively optimistic forecasts on traffic volume, and the private partner’s
revenues happened to be half of what had been estimated. This led to litigation on tolls, suspension of investment
and loan disbursements, and a default on the private-sector party debt. Both the concession and debt obligations
were taken over by the public sector (European Commission, 2004).
62
and in general, because asymmetric information and haggling over prices cause efficiency
losses. In contrast, the cost-plus mechanism is a low-powered incentive scheme, pushing less
towards cost-efficiency but facilitating contract revisions because the public-sector party
reimburses any cost increase resulting from changes in the contract terms. Hence, when
choosing the contract design, the public-sector party faces a trade-off between providing
incentives to reduce costs (in fixed-price) and facilitating efficient contract revisions and
information sharing (in cost-plus).
In this context, the authors characterize the optimal contract design. A fixed-price
payment is optimal for a project of low complexity that has a design with low degree of
incompleteness, and where renegotiation is less likely. In other words, when the project is not
too complex and/or uncertain, it is better not to save on design costs and write a more
complete contract; as a consequence, renegotiation is less likely, allowing to take advantage
of the fixed-price mechanism in terms of incentives.
On the other hand, a cost-plus payment is optimal for a project of high complexity and
subject to high uncertainty that has a design with high degree of incompleteness, and where
renegotiation is more likely. In other words, when the project is highly complex, it may be
better to save on design costs and write a less complete contract; as a consequence,
renegotiation is more likely, but the cost-plus mechanism ensures it won’t be too costly.
Despite the benefits of contract renegotiation in terms of improving the original
contract terms ex post to cope with unforeseen events or to introduce learning, there are many
undesirable outcomes that typically arise when revisions take place or are expected in a
context of imperfect enforcement and opportunistic behavior by the contracting parties,
including rent-shifting activities, politically-motivated investments and corruption. All these
negative effects, even if only anticipated, tend then to distort bidding behavior during the
private partner selection phase and thereby to generate an inefficient selection of private
partners and terms of trade (overly aggressive bids).
5.2.1 Incentive Distortions
Whilst in the previous section we have highlighted the ‘Pareto efficient’ side of
renegotiation, it is important to note that renegotiations can have negative consequences on ex
ante efficiency because, when the private-sector party anticipates future contract revisions, it
63
faces distorted incentives and misbehaves in the tendering process. This line of reasoning is
akin to the soft budget constraint problem (Kornai, 1986; Rajan, 1992; Dewatripont and
Maskin, 1995), which may also lead to inefficient project being financed.
The literature on soft budget constraint (SBC) helps in identifying negative effects of
recurrent, and thus expected, contract renegotiations on the private partner’s incentives and
risk exposure. The original formulation of the SBC problem by Kornai (1979, 1980, 1986)
was developed to explain the survival of persistent money-losing state-owned enterprises in
socialist and transition economies. The interpretation of SBC as a dynamic commitment
problem, due to Dewatripont and Maskin (1995), asserts that entrepreneurial incentives are
distorted by the managers’ belief that the public sector will bail out firms in the future if they
fail or underperform. This belief is supported by an objective fact: it is often in the interest of
the public sector to rescue firms to avoid the social costs of investment project termination.
Ex ante, the public sector should commit not to bail out firms in order to get entrepreneurial
incentives right and to discourage entrepreneurs from undertaking bad investment projects.
However, such a commitment is not credible because ex post the public sector will do better
by bailing out all bad projects already undertaken. Thus soft budget constraint may lead to
more bad projects being financed.
The SBC problem provides an insight to account for the pattern of PPP renegotiations
in LAC countries. According to the empirical evidence reported below, concession revisions
changed essential contract terms and tended to benefit the private partners by reducing or
rescheduling investment requirements, increasing tariffs, granting subsidies or tax
exemptions, lengthening the contract duration, etc. Ex ante, private partners expecting future
renegotiations to favor them face weak incentives to perform, to reduce costs, to improve
quality of service, and to innovate.
In addition, they enjoy a de facto low exposure to risks leading to financial
disequilibrium. For these reasons, ex ante a benevolent public sector wishes to commit itself
not to change the original contract terms in the future, at least not in a way that simply
benefits the concessionaire and do not address Pareto-improving issues resulting from
contract incompleteness. But the public sector’s threat of no-renegotiation may be empty to
the extent that, ex post, the public sector prefers to rescue the firm and avoid contract
64
disruptions. Under these circumstances, the private partners’ expectative on future contract
revisions is strong, and thus incentive distortions and risk misallocation arise at the outset.
Furthermore, renegotiation brings about a cost in terms of commitment loss: if abused
in the past, the public sector’s reputation may be ruined, and this can reduce the incentive
power of future contracts and distort competition in future tenders.
There are a number of reasons why ex post the public sector may prefer to support a
concessionaire calling for a contract revision. For instance, if an unfavorable event occurs and
the private-sector party undergoes financial stress, it could threaten to abandon the concession
unless the public sector accepts to change the original contract terms.31 The concessionaire’s
threat is effective when it has an informational advantage in the operation of infrastructure
and the provision of service that renders it difficult to substitute, or when its sunk costs and
specific investment are not large, so its losses are not heavy if it leaves the concession.
On the other hand, the public sector’s bargaining position is weak when arranging a
re-tendering to select a new private operator involves time and large transaction costs, even
more when the market structure is not competitive and few operators are ready to take over
the concession (Kerf, 1998). Thus, a public sector refusing to renegotiate takes the risk of
suffering disruptions in the service provision when the private-sector party abandons the
31 There are cases in which the public-sector party refused to renegotiate the contract terms precisely
because the bidding behavior of the private partner seemed to be strategically aggressive. For instance, in a
tender to provide water services in Buenos Aires, a key variable was the lump-sum fee due to the provincial
public sector. The winner offered to pay a fee 18 times larger than the strongest competitor’s, but soon after
awarding the concession it sought to renegotiate the contract terms. Cross accusations of non-compliance led the
public sector to refuse revising the original contract terms. Then, the private partner abandoned the concession
and the public authorities reassumed responsibility for providing water services (Guasch, 2004). Another case is
the electricity distribution privatization in Peru. The public authority sold 30 % of the state-owned enterprises’
assets to the private sector, planning to divestiture the remaining assets in subsequent phases. The winning bid
exceeded the other competing offers by such a large amount that some analysts estimated that current tariffs
would be insufficient for the private-sector party to recover its initial investment. After winning the privatization,
the private partner requested both a tariff increase and a valuation of her assets based on the large amount she
had paid in the privatization. Since regulators refused to change the established valuation criteria, the firm
refused to purchase an additional 30 % of shares and sued the public sector for breach of contract. Subsequently,
the public sector regained control of the distribution companies (Guasch, 2004).
65
concession. Therefore, the public-sector party is likely to accept the private partner’s demands
regarding changes in the contract.32
5.2.2 Bidding distortion
So far, the discussion has focused on the negative effect of anticipated renegotiations
on the private partner’s incentives to perform and risk exposure, assuming that it has already
awarded the concession. Thus, expected contract revisions create a moral hazard problem. But
it is clear that renegotiations also distort the private sector’s bidding behavior in the initial
tendering process, i.e. they create an adverse selection problem. To the extent that firms
bidding in the tendering anticipate that a contract revision could be called for later, they do
not deem the original contract terms proposed by the public-sector party as binding
commitments. Hence, bidders seek to win the concession at any cost, regardless of what the
current contract terms are. To maximize the probability of being awarded, bidders have
incentives to bid aggressively, offering low tariffs or high transfers to public-sector party,
ambitious investment plans, etc. Once the concession is granted, the private-sector party calls
for a renegotiation to change its formal contractual commitments.33
Interestingly, a high degree of competition in the tendering process may worsen the
negative effect of anticipated renegotiations on bidding behavior. If a private operator faces
strong competition in the tendering, it will bid very aggressively to undercut the competitors
and win. Next, if the operator wins offering very generous terms for the public sector, it is
more likely to seek for a renegotiation. Guasch (2004) reports empirical evidence on
concessions in LAC countries supporting this point. Most of the contracts were awarded
32 Outright political affiliation between the public sector and the concessionaire also explains the
willingness of the former to favor the latter in a contract revision. In fact, contract revisions were observed in
sectors like highways where the private-sector party did not have an advantage in terms of know-how, and thus
could have been replaced at no significant costs for the public sector, i.e. there was no information-related hold
up problem (Engel et al., 2006).
33 The case of the Lima airport illustrates an aggressive bidding under expectations of a future
renegotiation. In the tendering process, a key variable was the percentage of gross revenues to be transferred
from the private-sector party to the public-sector party. The winner consortium submitted a very attractive bid
whose financial viability was questionable: it offered to transfer 47 % of revenues and to undertake an ambitious
investment plan. But after winning the concession, the consortium demanded a renegotiation. Not surprisingly,
the revision adjusted downwards both the revenue transfer and investment obligations during the concession
period (Guasch, 2004).
66
through competitive bidding (78 %), and a few through direct adjudication and bilateral
negotiation (22 %). Among the concessions awarded by competitive bidding, 46 % were
revised at least once. In contrast, only 8 % of the concessions awarded by non-competitive
procedures underwent renegotiations.34 Hence, competition in a tendering process might have
induced aggressive, financially-unsustainable bids by the participants, and thus made it
convenient for the winner to revise contracts afterwards. On the other hand, bilateral
negotiation in non-competitive procedures might have granted favorable terms to the private
partner, ruling out financially-unsustainable proposals but allowing for corruption and rent-
seeking, and thus lessened the incentives to call for a concession renegotiation.35
It is noteworthy that competition and market structure have different effects on the
contract renegotiation issue. On the one hand, higher competition in the renegotiation stage
increases the bargaining power of the public sector, making it more able to resist unacceptable
proposals backed by the private partner’s threat of contract termination. Hence, since contract
revisions that simply benefit the private-sector party are less likely to happen, competition
reduces the incentive distortion created by renegotiations (Segal, 1998).
On the other hand, it was argued lack of competition might rule out aggressive bids
that are financially doubtful, thus decreasing the probability of a future renegotiation that aims
precisely at benefiting the concessionaire (Guasch, 2004). In this sense, someone would think
that competition may increase the incentive distortion. But it is likely that it is not competition
that generates more renegotiation but the lack of guarantees and weak enforceability of
contracts, which leads to expected renegotiation and distortive behavior in the bidding stage.
5.2.3 Enforcement problems, rent-shifting and corruption
From the incomplete contracts perspective, renegotiation is beneficial for both parties
because it allows them to correct inefficient outcomes that would arise because of contract
incompleteness, i.e. they can make mutually advantageous deals on issues not envisaged
34 The figures exclude telecommunications concessions.
35 The case of water concession in Cochabamba (Bolivia) illustrates why bilateral negotiation involving
political affiliation reduces the probability of renegotiation. The Aguas del Tunari consortium was the only
bidder in the tendering. Since the local partner in the consortium was owned by one of the most influential men
in Bolivia, Aguas del Tunari awarded the concession despite omissions and irregularities in the tendering
process. The contract entitled the private partner to a 15 % return for 40 years guaranteed by the public sector, so
there was little incentive for the concessionaire to change the contract terms (Lobina and Hall, 2003).
67
before. But the previous discussion on SBC has emphasized that anticipated renegotiations
distort incentives ex ante, and that the bargaining power of both parties affect the outcomes
resulting from a contract revision. In particular, it has suggested the threat of concession
termination by the private partner could back demands for new contract terms that only
benefits it and have nothing to do with correcting inefficient outcomes. Indeed, fearing
disruptions in the service provision, the public sector may end up satisfying the private
partner’s demands.
The risk that the private partner abuses renegotiation increases with the existence of
enforcement problems. If the public-sector party finds itself unable to force the private-sector
party to meet the original contract terms, it can hardly oppose a change in these terms
required by the private-sector party, no matter how unacceptable the change could be from the
public sector ’s perspective (provided that this new contract terms are enforceable). Thus, the
renegotiation may entail a rent-shifting benefiting the private partner, who calls for the
revision aiming to exploit the public sector’s inability to enforce the original contract (Guasch
et al., 2006). Moreover, the lower is the enforcement, the higher is the probability of shifting
rents, and thus the stronger is the incentive for the private-sector party to call for a contract
revision.
Beyond enforcement problems, the public sector may lack transparency and
accountability, thus providing incentives for corruption (Lobina and Hall, 2003). It was
argued the quality of enforcement is negatively related with the level of corruption, which
also affects the probability of renegotiation (Guasch, 2004). Corruption may be related with
illegal payments made by the private-sector party. Bribes aim to align the public sector’s
interest with the private partner’s rather than to the public interest. As long as the private-
sector party believes the public sector can be bribed, it will bribes in order to increase its
chances of appropriating rents in a contract renegotiation. Hence, lack of public accountability
mechanisms alters the bargaining relationship in renegotiations, while corruption
opportunities distort the parties’ behavior.
5.2.4 Political interests
Letting aside the mutual gains resulting from completing the original contract terms,
the arguments of SBC and imperfect enforcement emphasize it is the private partner that
68
benefits from the contract revision, thus suggesting it has strong incentives to call for the
revision.36
However, in the study of Guasch, Laffont, and Straub (2006), almost two-thirds of the
concession renegotiations were initiated by the public-sector party, and most of them took
place in periods surrounding elections. Consistently with these facts, a number of political
motives have been proposed in the literature to explain the interests of the public-sector party
itself in revising PPP contracts. Naturally, renegotiations involving political interests are
likely to go far beyond contract incompleteness issues.
For instance, a government wishing to increase its chances to be re-elected may seek
for financing investment and expenditure in public works that create jobs and boost economic
activity. Facing fiscal or political constraints to expand spending, the government may
attempt to get private partners to do it (Guasch, 2004). In a regular budgetary process, issuing
public debt to finance additional spending requires approval from the opposition parties,
which may block the public sector’s spending plan if it reduces their own electoral chances.
Thus, instead of negotiating with the political opposition, the public sector may prefer to call
for PPP contract revisions and increase the investment requirements to be financed by private
partners (Engel et al., 2006). Since these renegotiations are not included in the regular
budgetary process, the government circumvents the opposition’s scrutiny and reaps the
political benefits resulting from higher present spending, e.g. a higher probability of being re-
elected. Private partners, on the other hand, may obtain better contract terms, higher user
charges, revenue subsidies, an extension in the contract duration, etc.37
36 In the SBC argument, the public sector’s gain results from avoiding the costs of a concession
termination.
37 Engel et al. (2006) describe two types of concession renegotiation often used in Chile to anticipate
spending: the public sector changed the contract terms to obtain additional works from the concessionaire, or
both parties exchanged an ‘insurance premium’, i.e. the private-sector party made an upfront cash payment and
was entitled to a potential concession extension. Twelve out of the sixteen highway projects awarded by 1998
were revised in the following five years. The revisions required the provision of additional infrastructure,
increasing 15% the value of the original investment plans. Renegotiations set the so-called ‘complementary
contracts’ which were agreed bilaterally and without public review. Since Chile had undergone a recession in
1998-2002, traffic flows grew less than expected, and tax revenue were insufficient to undertake public works.
Concessionaires were granted a public sector guarantee for the toll revenue that would have collected if traffic
had grown at certain annual rates during the contract life. Should actual revenues increase below such rates, the
69
In the short-run, both the public and private partners benefit from politically-oriented
renegotiations, but the resulting bias towards over-spending in the present may imply an
intertemporal resource reallocation that negatively affects social welfare. In this context,
ensuring the public sector is accountable to the Parliament or independent oversight agencies
for the PPP agreements it negotiates would help to reduce opportunistic behavior leading to
politically-oriented contract revisions. Beyond the accountability issue, the public sector
agency running a concession program should be given appropriate incentives. For instance, if
the agency is part of the ministry of public works, whose purpose is naturally to build new
projects, it will probably attempt to expand the public works program (Engel et al., 2006).
5.3 How contract design can help to maintain flexibility while limiting
abuses
As the discussion above emphasizes, while there is a clear need of maintaining
flexibility in time to deliver what is really needed at that time, there is also a clear need to
structure this flexibility so as to limit as far as possible the potentially very high costs it may
imply, particularly if abused. As a general principle, therefore, post-award contractual
changes should be avoided as far as possible; they should be rare and exceptional events,
particularly when changes are requested by the private-sector party.
One general and somewhat obvious principle of contract design that may help to
soften this trade off is trying to build in the original contract clauses for all anticipated
potential changes. This sums up to try ‘making the contract as complete as possible’, taking
into account the cost of writing a complex contract that details many possible contingencies
from which only few will effectively realize. A contract that regulates ex ante anticipated
potential changes produces a sort of built-in flexibility that reduces the need for contractual
changes and is not subject to most of the downsides of renegotiation discussed earlier.
concession would be lengthen by up to almost 10 years and the public-sector party would pay the remaining
difference. Thus, the renegotiation proposed an asymmetric variable-term contract to replace the original fixed-
term contract. Concessionaires, on the other hand, had to pay an ‘insurance premium’, around 8 % of the
guaranteed revenue, in the form of additional investments. Thus, the public sector engineered an intertemporal
transfer, receiving additional current infrastructure in exchange for a guaranteed income that future
administrations would pay.
70
If the anticipated potential changes can be properly specified in terms of output, then
the price of these changes may already be established at the initial competitive selection stage.
If they cannot, then clauses dealing with contractual changes should determine cost-based
compensation for the changes (as in cost-plus contracts).
Although investments in contract design that regulate potential changes are welcome,
a complete contract is not something one can aim at in a 30-year long procurement
relationship. The complexity and long-term horizon typical of PPPs are bound to make any
such contract incomplete and subject to requests for changes linked to unanticipated events.
Even the contractual provisions for anticipated changes may easily become obsolete over
time, and their adaptation may become necessary in the light of unexpected major
technological changes. In this context, post-award contract changes, renegotiation, and
contract completion can be efficient means to address issues arising from contract
incompleteness, and so they should not be ruled out. The challenge for contract design is then
to identify and support efficiency-improving contract revisions.
To the extent that a revision may lead to ex ante undesirable outcomes, such as rent
shifting and politically-motivated investments, there is a case for the design of contracts that
explicitly foresee future renegotiations and pre-emptively establish principles and procedures
to rule the revisions if the parties call for them. Let us call them ‘bad renegotiation’-proof
contracts. The literature provides useful insights on how contract design can ensure that future
renegotiations will contribute to achieving the PPP objectives. Contract design can do
something about renegotiation (influencing its occurrence and outcomes) not only because it
can directly affect the contract characteristics determining the degree of incompleteness and
the likelihood of revisions (as documented by Guasch et al., 2006), but also because it can
require compulsory and structured renegotiation processes that limit the scope for abuse.
In this regard, the initial PPP contract should already address as clearly as possible:
(i) the circumstances that justify tariff and output adjustments,
(ii) when and how to implement benchmarking and market testing to test the value
for money of the proposed changes,
(iii) the circumstances under which the contracting parties are entitled to call for a
more general contract renegotiation,
71
(iv) specific principles and procedures to rule the revision.
A basic and somewhat obvious criterion often advanced to deal with contract changes
required by the private-sector party is linked to the size of the shock leading to the proposed
changes. Small unexpected shocks are unlikely to constitute a real threat to the financial
stability of the firm. Thus, small shocks do not justify the costs and risks linked to a contract
revision. Very large unanticipated (hence exceptional) shocks, instead, may threaten the
financial stability of the firm in such a way that - unless a contractual change is introduced -
the private-sector party may not be in a position to continue fulfilling its service obligations.
As mentioned earlier, under these conditions renegotiation often occurs in reality,
particularly if the cost of service disruptions is considered large. However, one also has to be
sure that the shock is really unanticipated, exceptional, and independent of the private
partner’s efforts before going for a costly bail out of a failing private-sector party through
renegotiation, rather than going for its replacement. The risk, again, is disrupting the whole
procurement process by favoring private partners that are unable to anticipate shocks or to act
so as to minimize their impact, i.e. less able private partners, and by selecting their overly
aggressive offers – which are probably cheap because they do not anticipate well the possible
shocks – rather than more appropriate and expensive ones.
Another basic principle is that, given their negative effects on governance and
efficiency, renegotiations should be extremely open and transparent procedures. To improve
on transparency, the contract may envisage calling a third party, e.g. an arbitrator, an
independent commission, or a group of experts, to evaluate the case and seek to conciliate the
needs of both parties without too much harm for the taxpayer. To limit discretion and
disagreement the contract may also provide a limit as to the amount that can be renegotiated
without calling for a new tendering process.
Problems such as rent shifting and politically-motivated anticipated spending, i.e.
problems that arise from weak enforcement, political interests, and failures in the regulatory
environment, are probably beyond the range of influence of the contract design. However, the
contract design should manage efficiently the risks resulting from these problems, e.g.
72
regulatory risk resulting from weak institutions, cost overruns and financial risks resulting
from macroeconomic shocks as we discussed in section 3.
It is noteworthy that some of the regulatory and institutional factors are not outside the
range of influence of the institutions ruling- and agencies implementing PPP agreements.
Thus, certain actions can be taken to address some of the problems distorting renegotiations
outcomes, i.e. leading to ‘bad renegotiations’. For instance, improving skills of regulators,
ensuring the agency negotiating PPP contracts do not depend on the ministry of public works,
etc.
73
6. CONTRACT DURATION
An important element in contract design is the length of the contract. The literature
emphasizes the implications of contract duration for
(i) investment decisions;
(ii) contract flexibility;
(iii) competition and performance incentives;
(iv) transaction costs.
6.1 Contract duration and investment
Contract duration can affect the investment decisions of the private-sector party both
when investments are contractible (or verifiable) and therefore describable in the initial
contract, and when they are not observable and therefore non-contractible.
Consider first the case of verifiable investment. With projects where the public-sector
party can pay a contribution to the private-sector party, contract duration should play a limited
role on investment levels since investment levels can be specified in the contract and adequate
payments by the public-sector party can ensure the project bankability whatever the contract
length. Any changes in the duration of the contract can be accompanied by changes in the
payment by the public-sector party to the private-sector party that leaves the level of
bankability unchanged.
With financially free standing projects the private-sector party does not receive
contributions or payments from the public-sector party, and its sole source of revenues is the
cash flow that the project can generate through user charges. In these cases, the duration of
the contract must ensure bankability and thus be consistent with expected cash flows and the
time needed to recoup invested funds. As long as supplied quantity and correspondent
revenue increase with contract duration, the larger the investment, the longer the optimal
contract duration. If the contract is shortened, then user charges are modified so as to ensure
greater revenues to the private-sector party, or the investment obligation is reduced.
74
However, contract duration could be determined endogenously. Engel et al. (2000,
2001) develop an award criterion based on the Least Present Value of Revenues (LPVR) that
bidders must submit when tendering. The contract lasts until the winner private partner
receives the LPVR it had submitted. In LAC countries, this criterion was used in highway
concessions, one in Chile and three in Peru (Guasch, 2004). A similar scheme has been used
also for the Dartford Tunnel in the UK (Klein, 1998). The advantage of this scheme is that it
reduces the demand risk for the private-sector party as a fall in demand and hence in revenues
immediately translates into a lengthening of the contract. The disadvantage is that by being
more protected against demand risk, the private-sector party has less incentive to make
investment that can help to control demand risk.
Consider now the case of unverifiable investment, where the investment is not itself
directly contracted/contractible upon and therefore cannot be described and protected in a
specific clause in the contract. If the investment is at least partly specific to the public sector’s
needs or to the project (we refer to this as ‘specific investment’) it will have limited value for
the private-sector party if used for alternative purposes outside the contractual relationship. In
this case, if the investment helps the private-sector party to increase its profits (say by
reducing the costs of the project) and if the private-sector party can appropriate (at least some)
of these increase in profits (say because the original contract is fixed price), then a long-term
contract provides more incentives to undertake unverifiable investments.
This is because specific investments imply larger losses for the incumbent in case of
terminating the relationship with the public sector. In particular, when both parties negotiate a
renewal of a short-term contract, the public sector can hold up the private-sector party by the
threat of terminating the relationship, asking for price reductions or changes in the original
contract terms. To the extent that the private-sector party fears holdups that reduce investment
returns in the future, it has then less incentive ex ante to undertake specific investments. Since
a long-term contract fixes the contract terms, it protects the incumbent against holdup. In this
respect, the higher the investment specificity, the longer the optimal contract duration (see
Ellman, 2006 for an in depth discussion). 38
38 Ellman (2006) warns that self-investments may raise the private partner’s payoffs but in an inefficient
way, e.g. resources are wasted in trying to hide low quality aspects of the services.
75
6.2 Contract duration and flexibility
In sectors with a rapid pace of technological change and changing user needs, the
public-sector party may prefer to have flexibility to modify the contract terms adapting the
service provision to the incoming innovations (HM Treasury, 2007).39 In order to adapt the
service provision as required by the public sector, the private-sector party would undertake
investments referred to as adaptation investment.40 The contract must then provide for the
right of the public-sector party to demand changes in service provision and for the right of the
private-sector party to be adequately rewarded for his adaptation investment.
In practice, requiring changes in service provision so as to adapt the service to new
user needs can be very costly for the public sector as it needs to bargain with the private-
sector party for the implementation and remuneration of these changes. Contract duration is
then important as it affects the bargaining position of the public-sector party and thus the cost
the public sector will have to pay for adaptation investment. This in turn affects the net gain
for the public sector from adapting the contract to new user needs and thus the flexibility of
the contract.
In particular, a long-term contract implies less flexibility because the public sector has
to wait longer if it wants to switch provider. The availability of alternative providers improves
the bargaining position of the public sector when requiring service adaptation from the current
provider. In addition, with a long-term contract the public-sector party’s threat of no renewal
during the contract is less powerful because a long-term contract implies the realization of the
threat would occur far in the future, and thus the present discounted value of the cost of
failing to renew the contract is negligible for the incumbent. Therefore, for the public sector a
long-term contract also increases the cost renegotiating efficient adaptations within the
contract period.
On the contrary, when the contract is short-term, the public-sector party‘s threat to
contract with a competitor when the current contract expires can be more powerful and induce
39 Hospitals and health services are typical examples of highly innovative sectors where PPPs have been
implemented and where user needs evolve rapidly over time.
40 Adaptation investment is a term coined by Ellman (2006).