When the usage level of the asset is not relevant for the purpose of the public party (that is, it is not of itself a public objective), but it is still paramount that the asset be available for use by the final users, for instance health workers in a hospital, then payment should be based on the availability concept. These schemes are the most common payment regime in social infrastructure, while in some transport sub-sectors this practice is becoming increasingly common (toll free roads, rail, and water transportation). In energy (power generation and transmission infrastructure), the payment scheme is quite similar to the availability concept, with the PPA typically providing a fixed payment as long as the plant is able to supply energy (plus another variable payment depending on the energy effectively supplied).
The government should only pay for the asset as long as it is available (so, never before construction is completed) and to the extent that it is available (but considering “exemptions” such as planned maintenance periods or certain events out of the control of the private partner. This is a matter of risk allocation). For example, only parts of the infrastructure may be available for use or the ability to use others may be limited (for instance, parts may have been downgraded for not meeting certain performance criteria, or they may not be available at all).
As suggested above, availability will generally respond to two general criteria. One related to physical availability for use (that is, the asset can be used effectively) and one related to the condition criteria (while being available for use, the asset may not be deemed available for the purpose of calculating the payment). Each situation can result in different payment adjustments or deductions (see section 4.10.1).
The availability payment mechanism should be unitary (subject to the caveat explained in box 5.12 below), as it is the revenue based on payments by users (no user, no user payment) or on shadow tolls. The unitary concept implies that the payment related to a section or portion of the asset (as described below in section 4.10.2) should be zero for the period of time during which that section or area of the asset is unavailable.
The availability risks are ultimately operational risks, intrinsically manageable and dependent on the performance and management capabilities of the private partner (subject to the availability and performance criteria being reasonable). Also, it is customary to segment the payment into sections of the asset or areas of the facility and then pro-rate the unavailability and payment deductions by time units. Additionally, time will be granted to the private partner to rectify events that cause unavailability (for example, a lane closed because of an accident), which further diminishes the credit risk (see section 10.5).
BOX 5.12: Availability Payments and Immature PPP Markets
The natural structure of a PPP payment mechanism is unitary in nature, as the government is paying for services and if there is no service there should be no payment.
However, certain EMDE countries, particularly those LDCs and any immature PPP markets, are opting for non-standard approaches to availability payments; this has the aim of significantly diminishing the risk perceived (especially by lenders), and in so doing support the bankability of the projects.
This de-risking approach takes two main forms.
Breaking the unitary concept to protect lenders and/or limiting deductions in order to increase the commercial acceptance of the PPP contract may be sensible in country/markets introducing the availability concept for the first time. This can also apply during the time in which the respective market is in its early stages of development. But otherwise, this should be avoided because this payment approach does not create the proper incentive for high standards in performance, and therefore the VfM may suffer.
The contents of this section assume though that all performance requirements are embedded in the payment mechanism, and when breaching any condition criteria this will directly impact the unitary charge.
The performance requirements and ultimately the service levels defined in the performance or condition criteria must be objective and measurable, but also realistic.
The government can set high performance standards without creating excessive risk for the private partner. There are multiple factors that influence the potential payment to be earned, and all of them have to be balanced so as to impose the desirable level of tension and risk on the private party: high standards may be compensated for by generous rectification periods or with modest deductions. High deductions might be associated with low or flexible standards, and so on. All factors should be driven by the value of the availability (or the loss in terms of unavailability) for the public party and the urgency with which breaches or faults affecting availability are remedied. The final definition of these factors has to be carefully analyzed by conducting “dry runs” of the payment mechanism (see box 5.13 about payment mechanism calibration).
Some performance criteria (generally speaking, those that are not regarded as critical for the purpose of the asset, for example, the cleanliness of the waiting room in a hospital rather than in an operating theatre) may be dealt with separately from the unitary payment. This is achieved by allocating a specific payment stream to a specific service, rather than paying for that service through the unitary payments. Deductions relating to that service are only made from that specific payment stream, not from the unitary charge.
Paying for specific services and deducting payments or imposing penalties for services separately from the unitary payment is common in respect of soft services (such as cleaning and catering), as well as in some accommodation-type social infrastructure projects such as prisons, hospitals, court buildings, and schools, where the government is retaining the ultimate public service concomitant to the infrastructure.
The rationale for these compound payment systems is the need to retain flexibility in order to accommodate certain specific risks and the ability to change the scope of soft services during the course of the contract (see section 4.10.10).
The factors that define the availability payment mechanism have to be tailor-made for each specific project. Nevertheless, it is useful to analyze successful precedents for similar projects. It is also very useful to create generic standards in the form of guides or white papers.
The ultimate structure of the payment mechanism, as with volume payments, will most likely be defined during the structuring process. Even if the basic structure of the payment mechanism was defined at an earlier phase, the mechanism should be carefully calibrated before the final definition.
As a general rule, the procuring authority should avoid unnecessary complexity when designing the payment mechanism, as complexity generates increased costs related to reporting and monitoring.
All these features and other elements of the payment mechanism, as well as structuring matters affecting this payment regime, are explained in this section.
4.10.1 Availability or Unavailability Definition. Categories and Faults as Service Level Targets or Condition Criteria
Availability will be determined based on two general sets of criteria: those related to physical availability for use and those related to the condition criteria (while being available for use, the asset may not be deemed available for the purpose of calculating the payment). Each set of criteria may result in different payment adjustments or deductions.
It is challenging to define “deemed unavailability”, as a significant level of judgment is required to define when something should be considered unavailable even though it is being used. The relevant question is to clearly define the performance thresholds that must be met for there to be an acceptable level of service — especially those requirements, which if breached, will represent a danger in terms of safety and security.
Therefore, a paramount factor for the structure is the definition of the level of performance required for each of the requirements or criteria defined,that is, how challenging or demanding the government will be in setting the required service levels or thresholds for the "condition criteria". Examples include: the index of roughness in a road, the luminosity of the lights in a road, the number of lights working properly in each stretch of the road, the number of lights required to be functioning in each of the defined areas of a school, the temperature in specific parts of a hospital, and so on.
To effectively assess unavailability and calculate deductions, a performance monitoring regime must be in place. Considerations as to who performs and who pays for the monitoring, and the monitoring and reporting obligations of the private partner, are covered in section 9. The UK “Standardisation of PFI2 Contracts” guide provides additional information and recommended standards for dealing with monitoring.
In a unitary payment, the payment covers all of the private partner’s costs including the amortization of capital, the cost of the financing, taxes and operations, and the maintenance costs.
However, the payment (and the availability or unavailability concept) should be divided into segments corresponding to parts of the infrastructure when practicable. This is for the purpose of calculating the payment and deductions by sections or areas.
In some infrastructure, a section or part of the asset may be considered self-contained as long as it is providing a complete service with its own economic and social sense. For example, a road project will usually be divided into sections of the road, each section being a stretch of road that allows users to reach one specific destination (for example a stretch of road between two cities, or between one city and the connection with a different highway that is out of the scope of the project).
In many social infrastructure projects, different areas serve different purposes, but all of them together provide a single ultimate service. For example, in a hospital project, rooms or patient places may be available, but if the area dedicated to clinical care is not available, patients will not be provided with the full services expected in that hospital. Even though patients do not have the full range of expected services available to them, it makes no sense for the government to consider that the facility is entirely unavailable, since many of the needs of the patients will still be met. Conversely, the unavailability of the clinical care area will adversely affect the care of some patients, and therefore the private partner should not be entitled to full payment.
Therefore, it is necessary in these projects to identify and define different areas because the relevance of each area for the public service is different. Some areas are regarded as more critical than others, and this should be reflected in the severity of deductions.
Each functional unit of the asset (such as a segment of a road, or a room or department in a hospital) will have its own weighting that determines the size of the payment deduction if that functional unit is unavailable.
In a road project or other transport infrastructure projects whose purpose is to provide users with the ability to move, the weighting criteria is clearly related to that use. Sections of a road that represent a higher demand will affect more users than other sections used by a lesser number of drivers.
In building-based PPPs, this is a more complex exercise, which will ultimately depend on the procuring authority’s own view of what parts or areas of the building are more critical to the service.
The yearly payment will be earned for each section or area by units of time, that is, availability may be measured in units of days or even hours or minutes. In some projects, the payment may be related to other units such as trips (a trip served by a rail system that is not meeting the availability criteria may be deemed as non-existent for the purpose of calculating the payment).
If the availability criteria and the condition criteria (target levels for each of the performance requirements) are not met, the government will be entitled to make deductions from the payments. Deductions should be commensurate with the seriousness of the criteria breached and the seriousness of the breach (that is, the unavailability event).
Other factors that will define the amount of the deduction are as follows:
- The area (or section in a road) that is unavailable, as some areas will have higher weightings than others.
- The length of time for which the area is unavailable.
- Potentially, the specific time of the day and year (month or day of the week), especially in transport projects – see next heading.
- Potentially, the persistence or repetition of the breach.
- Whether the asset (or the area affected) is effectively being used despite the unavailability. This is implicit in the payment mechanism in road projects through a specific category of unavailability which is related to the road section being blocked and to what extent (that is, how many lanes out of the total number of lanes designed and constructed in the respective section).
In road projects, it is not uncommon that the payment is built and calculated from bottom to top, that is, summing up the availability periods of each of the sections and therefore summing up the payment earned per unit of time. The payment related to each unit of time will be adjusted by an unavailability factor which usually ranks between 1 and 0 (including 0.5 or potentially other intermediate grades), with the former being related to full unavailability and the later to full availability.
Depending on the weighting of the area in some PPPs, especially social infrastructure projects, a deduction may be higher than the value of the payment that can be theoretically obtained during the unavailability period, even without considering ratchet factors (a factor that, when applicable, increase the deduction applied to the payment under a particular sub-performance situation or unavailability event – see 4.10.6) and other potential adjustments. Some guides (for example, “Standarisation of PFI2 Contracts” by the UK’s HM Treasury) describe a typical range of the potential value of the unavailability between 150 and 200 percent of the value of the corresponding payment.
This means that payment may be zero before the facility is completely unavailable.
However, it is not customary to require the private partner to make payments to the procuring authority, so the actual deduction will always be limited to 100 percent for the purpose of payment calculations.
The adjustment applied to the payment in an availability period so as to deduct the earned payment is sometimes further adjusted by a time factor. This is quite common in transport projects so as to align the relevance of a lack of availability with the actual time in which the fault is occurring: the impact of unavailability of public transport services is clearly higher at peak hours and on peak days (business days), so the deduction may be higher in those periods and lower in off-peak periods.
Another factor to be defined and which will influence the risk and incentive for optimum performance is the time for rectification.
It is customary that an availability fault or unavailability (that is, the failure to meet the threshold of a condition criteria) will not result in a deduction as long as the partner rectifies the breach within a certain period of time from the moment that the breach was detected (the rectification period).
Technically speaking, the private party is at fault and the asset is unavailable, but the unavailability will not be considered to have occurred for the purpose of the payment calculation if it is rectified within the time frame established for that particular condition criteria. If the fault has not been remedied within the rectification time, unavailability will be considered to have started at the time at which it was detected (and not at the time when the rectification period is finished).
- Some considerations in this respect. Not every breach should have a rectification period. For example, a breach of the specification for serving three meals per day should result in a deduction without a rectification period;
- Rectification periods should not provide a disincentive for proactive or preventive maintenance. Therefore, some critical criteria should not have rectification periods and the extension of the period should be short for other critical criteria. For example, electricity supply is critical in hospitals, so in a hospital PPP the private partner should install and maintain back-up generators. If the external power supply fails and is not immediately rectified by the back-up generators, it should automatically lead to a deduction. The private partner should have maintained the back-up generators so that they immediately took over (a rectification period should not be needed to restore the power supply);
- For requirements and their respective criteria that are not regarded as critical, or when the degree of the breach is not significant, longer rectification periods may be granted;
- Unavailability is considered to start when the unavailability event has been detected or reported. The main means to identify or detect unavailability events is through the monitoring process, that is, through the monitoring systems to be put in place and the inspections to be made of the asset as defined and prescribed in the contract or proposed otherwise by the private partner in its bid. Reports may come from the procuring authority or from an interested party (for example, a teacher in a school PPP); and
- Some events may require an immediate remedy so as to temporarily solve or mitigate the breach of the requirement, regardless of the existence of a rectification period. These are often referred to as “temporary fixes”. For example, if a window in a school PPP breaks, the private partner may have a five day rectification period in which to replace the glass, but may be required to place boards over the window within 24 hours of the breakage as a temporary fix. The requirement for a temporary fix should be clearly set out in the contract. Any failure to perform the urgent remedy will constitute by itself a performance breach, which is usually handled by penalties directly or indirectly through a performance point regime.
Some payment mechanisms apply ratchet factors, increasing the value of deductions to penalize repeated or persistent under-performance, or unavailability that continues unresolved.
When part of the infrastructure (a section or an area) is closed or is not available for use due to maintenance work, this will not constitute an unavailability event as long as the maintenance work is undertaken in accordance with the agreed maintenance plans. Other approaches to permit maintenance work, without affecting payment, are to allocate time weighting factors (TWF) of zero value to some periods of the day (for example, to some periods during nights) or to grant a "bag of maintenance hours" on the condition that these are only used at less disruptive times (which may be particular hours, times of the year, or months).
Other exceptions to unavailability (that is, unavailability periods that will not be deemed unavailable for the purpose of calculating the payment) may include:
- Unavailability due to police orders or other lawful requirements;
- Accidents in road projects (in some project contracts). This is generally not advisable, as accidents can be allowed for by simply providing a sufficient rectification period); and
- In general terms, any excused event to the extent that may affect availability.
Exempted unavailability events should be duly described and regulated in the contract.
Although increasing the curve or slope of the payment stream may be a temptation (for example, back-ending the curve of payments so that the payments are lower in the early years of the contract), but this should be avoided. The price of the service should (at least in real terms) be the same during the whole life of the contract, with the exception of changes in capital works or services.
Indexing the payment to yearly inflation may appear logical on the surface. However, this should be also carefully considered as the cost of the service (including capital costs, that is, repayments to the lenders and investors) is not entirely linked to inflation.
Debt service is not generally related to inflation. Under most of the financial structures commonly applied to PPPs, debt service is a fixed cost because the private partner hedges the interest rate risk, converting the interest payable to a fixed rate for all or a majority of the life of the loan.
Conversely, equity is exposed to inflation risk, and many of the private partner’s operating and maintenance costs will change over time roughly in accordance with the CPI. However, the costs that are linked to inflation typically represent the minority — usually less than 50 percent — of the total pay-outs by the project company.
To better fit with the real cost curve profile of the cash flows, many availability-based projects opt for a compounded indexation formula where only a proportion of the unitary charge is adjusted for inflation. That proportion is based on the expected or actual proportion of the project company’s outflows that will be affected by inflation.
This has the advantage of providing higher payments in nominal terms in the early years of the contract. Therefore, balancing the stream of payments in nominal terms will allow the project company to accelerate the debt repayment compared to a full CPI indexation.
Availability is about paying for the ability to use the asset. Therefore, payments during construction should generally be avoided, regardless of how advanced the construction is.
An exception may be made when there are parts or areas of the infrastructure that can be effectively used and as long as there is an intention to use them, which should be clearly stated in the contract. This could be the case in projects such as the following.
- In some road projects where certain stretches of the road may be effectively used prior to completion of the remainder of the road, it makes social and economic sense to enable that use; and
- In a multi-facility project where different buildings are being contracted and each provides a stand-alone service without heavy dependence on the existence of others (for example, a number of schools on separate sites under the one PPP contract).
Another exception may be projects related to existing infrastructure where there is a need to keep the existing infrastructure in use during the course of the initial works for upgrading. Payment should be granted at least to offset the O&M costs of the existing infrastructure (but not to pre-finance the upgrade works).
The contract should regulate the situation when a project construction is completed before schedule or before the target date prescribed in the contract. In general terms, the procuring authority should not have an obligation to pay for the asset if availability starts before the target date, but should state clearly whether it will accept the asset for use (and therefore start payments) before that date.
Early availability of the infrastructure is usually in the interests of the procuring authority, however the government may, or most likely will not, be ready to make payments if the project is completed early.
In that sense, one possible approach is to establish an early completion floor, before which the procuring authority will not render any payment, and let the best bidder define the completion date. The most efficient bidder in terms of the construction term is likely to be able to offer a lower price, therefore there is a natural incentive to share the benefit of early completion.
If the project company completes the project before the date scheduled (and priced) in its bid, it is customary, in order to capture a part of the benefit of earlier completion, to establish a “cut" in the payment so as to cover O&M costs and some level of incentive or bonus for the equity holders.
In a number of circumstances, there may be additional payments out of the unitary charge or as a complement to it.
Examples include the following.
Inverse volume risk adjustments in transport projects
Availability of transport infrastructure is unrelated to the actual traffic or volume. However, the maintenance and life-cycle costs will vary with the intensity of the use, which will cause a perverse misalignment of interests between the two parties.
There are two approaches for this. The first is to define a shadow payment linked to volume intended to cover the "marginal O&M cost", with the intention to neutralize the marginal increment in costs per vehicle. This is usually based on heavy traffic vehicles, and it may be settled as a payment from vehicle zero or above certain thresholds of demand.
The other common approach is to revisit the unitary charge value if and when traffic consistently exceeds a threshold (for example, for three consecutive years).
The latter has the advantage of being simpler to regulate. However, it is more complex to implement in reality as it will be controversial to calculate and agree on the actual additional cost due to excess traffic. This would, however, be the best method as long as the asset is designed to stand for materially higher levels of traffic than the currently expected demand.
Service payments for some soft services in social infrastructure and benchmarking of costs
Availability based payments are appropriate for most accommodation facilities such as hospitals or schools. The government’s objective in a hospital project is not to receive and treat more sick people. On the contrary, the government would prefer that there were fewer sick people requiring treatment, but it wants to be sure that the facility building is available to accommodate and treat those who require treatment. Therefore, it would not be rational to pay the private partner on the basis of volume (number of users). Instead, payments in accommodation-type hospital projects are usually based on the number of beds available and other availability parameters designed for a certain occupation level specified in the contract.
However, more demand or use of the facility (more people to attend to) clearly implies higher variable costs for the private partner. Therefore, above certain thresholds of occupancy, O&M costs will rise, especially in relation to certain soft services (cleaning, waste management, catering, and so on).
In this context, two main approaches may be taken;
- To provide an adjustment to the unitary payment when occupation is above certain thresholds (to be settled in the contract); and
- To consider a separate ad hoc payment component for certain services, that is, those that are very sensitive to demand.
The CPI or other general indexes may not adequately capture changes in the cost of certain services over time. Some contracts (usually related to facility management, including soft services) provide a degree of protection for the private partner against changes in these costs. The costs are analyzed at specific points in the life of the contract and compared with the current market conditions for such services. If the costs are materially different to those expected at the inception of the contract, the unitary charge is adjusted. Methods to deal with this issue, apart from relying on partial adjustments to CPI in the unitary charge, are market tests and benchmarking of costs.
Subjective adjustments and user satisfaction components of a payment mechanism
Objective criteria should be the basis of the availability system. However, the quality of the service in complex settings such as hospitals or schools cannot easily be wholly reduced to a practical set of availability and performance criteria (Standarisation of PFI Contracts, HM Treasury UK).
Therefore, in some projects, it may be appropriate to implement user satisfaction surveys as part of the overall payment regime, but the financial impact of these surveys should be small, as a recognition of the variability and subjectivity of such methods of performance measurement.
Advisable approaches for user satisfaction measurement may include the following provisions.
- The private partner may be obliged to conduct a survey at its own expense at specified periods (for example, once a year);
- The private partner may be required to carry out a performance audit at its own expense. When the scoring is poor, it will need to develop a remedial plan; and
- Potentially when scoring is poor (and noting the challenge to define what is a poor level or negative result in these surveys), the government may be entitled to apply a direct deduction in the yearly payment (which should be small). Alternatively, it could grant performance points that may have financial implication only to the extent that the aggregated number of performance points (related to other breaches of performance levels or other breaches of contract) reaches certain thresholds.
When there are financial implications in cases of poor performance, the scheme may sometimes also grant a bonus payment which again should be small or modest in comparison with the overall size of the payments within the payment regime.
The general rule is that subjective quality factors such as this should only affect the equity return and only to a modest extent. Otherwise, the private partner may find it necessary to build up a cash contingency to mitigate the risk that payments are reduced due to user survey results, despite good performance by the private partner and its subcontractors.
The contract has to clearly settle when payments will be made and the process to calculate the payment earned (that is, the calculation of the deductions and the process of invoice and actual payment). There is a significant variety of approaches.
The process of monitoring and then calculating deductions is complex. Therefore, some procuring authorities prefer that payments are made no more frequently than quarterly. However, this requires the private partner to finance significant working capital, as many of its expenses will be payable monthly or more frequently. Requiring the private partner to finance significant working capital may not provide VfM, as private financing is more expensive than public financing.
To calculate deductions each month and make payments monthly in arrears is theoretically the most accurate and fair system. However, some procuring authorities opt for a middle path solution, making the monitoring calculations and determining deductions on a quarterly basis but making "advanced payments" within the first two months of the respective quarter. Those advance payments may be done on the basis of a provisional notional deduction (for example paying 90 percent of the availability payment corresponding pro rata to the respective month) or reducing the monthly payment by a proportion of the previous quarter’s deductions.
Timing for payments after receipt of the relevant invoice will generally be subject to the government’s general payments regime. Payments should be made within a reasonable time, and delays should be subject to interest at a rate consistent with the potential delay interest payable by the private partner under the financial agreements.
In all these cases, at the last month of the respective quarter, deductions applicable to the actual performance for the quarter will be calculated. The payment for the final month of the quarter will be adjusted to account for any under- or over-payment made in the previous two quarters.
 When unitary payments were first introduced in PPPs in developed countries, there were concerns that they may not be bankable since, at least in theory, all of the private partner’s revenue is at risk and the private partner may be unable to repay its lenders. However, as the risks are largely in the private partner’s control and there is now an established body of experience with these payment mechanisms, unitary payments no longer give rise to bankability concerns if they are appropriately structured.
 In some payment systems, especially in the field of social Infrastructure and particularity in accommodation or building-based projects, some performance requirements (which are not regarded as critical for this purpose) may not affect the unitary payment directly through measurements of unavailability. Deductions may be better dealt with through performance points and a penalty regime. On other occasions, there may be different thresholds defined for a requirement (therefore different levels of criteria), with the minimum or threshold level directly affecting the unitary charge (in such cases double counting should be avoided).
 See section 20 in “Standarisation of PFI 2 Contracts” (HM Treasury UK, 2012).
 An exception to this would be projects with multi-facilities, that is, two different buildings to accommodate patients and provide different clinical attention would be considered as separate assets for the purpose of payment calculations.
 According to HM Treasury, UK (Standarisation of PFI Contracts), accommodation in a hospital infrastructure PPP will usually be grouped into three areas: the most important area includes accident and emergency facilities and patient spaces including bathrooms, operating theaters, and intensive care; the area of medium importance includes general waiting areas and clinical support areas such as pharmacy, physiotherapy, and chiropody; and the least important areas are office areas and educational facilities.
 However, it should be noted that more recently there has been an increase in the participation of institutional investors in the financing structures, and these investors are frequently interested in assets linked to inflation.
 See chapter 15 of Standarisation of PFI Contracts (“Price variations”) for additional and more specific intelligence on contract tools to deal with this issue.