5.3 Comparing public procurement and private ownership of facilities and services

Hart (2003) discusses the relative merits of the conventional provision of services, whereby the government contracts with one private party to build a facility and then contracts with another private party to operate it (“unbundling”), and PPP, whereby the government contracts with a single private party to build and operate the facility (“bundling”). In either case, the private party responsible for building the facility has the option to spend more on construction. “Productive” investment raises the benefits of the services produced by the facility and also lowers the (operating) cost of providing them. In contrast, “unproductive” investment lowers both the benefits and operating costs of providing these services.

Under conventional provision, the building party simply minimizes the construction cost, and so does not invest in either productive or unproductive investment. In contrast, under a PPP arrangement, the private party internalizes operating costs (but not benefits), which leads it to over-invest in unproductive investment (since this allows it to lower its own operating costs, and it is unconcerned by the associated lowering of benefits in the process) and to under-invest in productive investment (since it is not motivated by the associated raising of benefits).

Hart concludes that unbundling is good, if the quality of the facility can be well specified whereas the quality of the service cannot be, as this avoids the overinvestment in unproductive investment that would result from a PPP arrangement (since there is little the government could do to penalize poor service provision). He also concludes that bundling is good, if the quality of service can be well specified in the initial contract and the quality of building cannot be (since a PPP partner can be incentivized to improve outcomes on the basis of quality of service).

Contractual incompleteness makes it impossible to contract on some things, but this still leaves freedom regarding how to contract over the issues that are contractible. For example, the choice of delivery date determines the amount of timing flexibility the “builder” has. Giving the firm more timing flexibility is giving it a potentially valuable real option, and this should be reflected in a relatively low tender price. This means that allowing a contractor more time to build a road allows it greater flexibility to avoid carrying out construction during periods of labour shortages or when other inputs are expensive. This benefit to the public needs to be measured against the cost of delayed introduction of the services the investment is to provide.

However, there are other issues to consider when deciding how much flexibility to allow in a contract, which are related to Hart's distinction between productive and unproductive investment. The private party responsible for the “building” stage has real options to engage in productive investment and real options to engage in unproductive investment. The terms of the contract will determine how much flexibility the party has, which will determine the value of these two options. Increasing the value of the party's real option to engage in productive investment is socially beneficial, as is reducing the value of the party's real option to engage in unproductive investment. For example, the builder might have the option to adopt emerging technology that lowers operating costs. If use of this technology is not something that can be contracted on, then giving greater flexibility regarding delivery dates will give the builder the option to wait and adopt the technology if it becomes economically viable (perhaps the technology becomes cheaper or the operating costs avoided become greater due to labour shortages, etc). This may be socially beneficial, or not, depending on the effect that adopting the new technology has on the benefits generated by the contracted services. In general, the choice of how much flexibility to include in a contract depends on how the firm will use that flexibility.

Hart's story, in its most basic terms, deals with situations where it is impossible to write a contract that will force a private party to operate in a way chosen by government. However, it is possible to predict how the private provider will behave, and that prediction helps government to decide whether or not to adopt a PPP arrangement and, if so, how to design the underlying contract. Circumstances may change during the course of a contract. For example, initially market conditions may be such that we would predict that the firm prefers bundling; perhaps the technology to undertake “unproductive” investment is currently so expensive that government believes that a PPP arrangement is suitable. If the price of this technology falls over the life of the construction phase of the contract, then the firm might actually undertake the unproductive investment, so that an initially desirable PPP may end up becoming an undesirable one. If this possibility seems likely, then contractual elements such as delivery dates might be tightened to reduce the flexibility available to the private party.

In an extension of Hart (2003), Bennett and Iossa (2006) consider a setting where a public entity delegates to private firms the construction and management of a facility that will be used to supply a public service. They model a situation in which specialisation requires separate firms undertake construction and management, but where construction and management firms can form a consortium to undertake both aspects of the project, and where there is the potential for investment in welfare-enhancing innovations that could not be anticipated at the time that the initial contract is signed. Within this structure they consider the impact of allocating the residual ownership rights to the government (traditional procurement) or the private sector (under a type of public-private partnership that they term “private finance initiative” (PFI)).

Hence they consider ownership by:

• The construction company.
• The management company.
• The government with separate procurement from a construction company and a management company.
• The government with procurement from a consortium.
• A private consortium.

As with other applications of the literature on incomplete contracts, these ownership structures matter, because value is created by research into innovative approaches to carrying out the tasks of constructing and delivering services with the facility. The discovery and implementation of any innovations are not contractible ex ante, but they are verifiable ex post. This means that ownership rights are associated with the residual decision rights to determine whether investment in research on innovations will take place, and whether the innovation will be implemented. Under private ownership, and provided that basic requirements in the contract are not violated, the private firm makes that decision. Under traditional procurement with public ownership, renegotiation between the public manager and the private firm must take place before the innovation is implemented.[26]

If there is a positive externality associated with innovation at the construction stage, then it will always be optimal for construction and management to be undertaken by a consortium, whether the residual ownership rights rest with the government (procurement) or the private sector (PFI). This is because separate contracting for construction and management will always create the potential for underinvestment in value-enhancing innovations because the gains will be shared with government (under procurement) and since ownership by either the firm constructing the facility or managing the facility will result in only one of the potential innovations in construction or management being implemented.

Bennett and Iossa (2006) also consider a model in which there are three distinct phases: construction of the facility, management and delivery of services using the facility, and the residual value of the facility at the end of the contract period. This allows them to explore the impact of both positive and negative externalities between construction and service delivery, as well as the impact of residual value, on the project.

In their model, PFIs should be used for the construction and operation of public service facilities where:

• The potential externalities from innovation are more strongly positive.
• The effect of innovations on social benefits that cannot be internalised in the contract is small.
• The effect of innovation on residual value is large.
• It is possible either to leave the facility in private hands in the long term, or assess its value as part of any renegotiation of the terms on which the facility would be transferred back to the public sector.
• The higher is the residual value for the private sector created by innovations in the contract.

The model generating these results differs from that of Hart (2003) in that it assumes that investments that result in innovations in building design or service delivery are verifiable. If they were not verifiable this would increase the range over which private ownership was optimal, because it would reduce the potential for firms operating under public procurement contracts to renegotiate contracts to obtain a share of the benefits from any innovation that they developed. Similarly the assumptions of the model that the contracting process is “one time” rather than a repeated game, and that information is symmetric, are unlikely to have strengthened the results in favour of private ownership and provision. The development of reputation in repeated rounds of contracting for similar PFI projects may increase the ability of public entities to identify private-sector owners whose construction and management of facilities will generate welfare benefits, and asymmetric information is more likely to increase the challenges of obtaining welfare-enhancing outcomes where private-sector firms working under procurement contracts have to negotiate with government to obtain a share of the value that their potential innovations will create.