TRANSACTION COSTS

Another controversial area of P3s is the higher transaction costs-relative to conventionally delivered projects- incurred by both the public sector and the private sector bidders to execute the transaction and to manage the project and monitor outcomes through to the end of the contract term. In principle, higher transaction costs in P3s result from the same factors that drive the efficiency gains:

◆  greater due diligence in risk assessment and allocation effort, which is reflected in the provisions of the partnership agreement; and

◆  the private financing that needs to be put in place, including the additional due diligence undertaken by the equity investors and lenders, each of which has its own set of legal, commercial, and technical advisors for the project.

Even public sector client departments tend to exercise more due diligence in P3 projects, perhaps because they are making long-term commitments that are difficult to change. However, critics argue that these transaction costs are too high and tend to erode the benefits of P3 procurement methods.33

Transaction cost data on P3 projects are seldom available, particularly data on costs incurred by the bidders. However, a study by Dudkin and Valila based on 55 U.K. P3 projects from six different sectors (hospitals, schools, roads, prisons, government buildings, and information technology) found that the public sector's bidding and contract negotiation costs were on average 3.5 per cent of the capital value of projects. The winning bidders' costs averaged 3.8 per cent of the capital value of the projects while the failed bidders' costs averaged about 5 per cent.34 These costs were limited to the procurement period and did not capture the expected costs of project management and monitoring during the contract term. Nor was there any attempt to compare these figures with transaction costs for conventional procurements. It is worth noting that the transaction costs for winning and losing bidders are not necessarily additive, as implied in Dudkin and Valila. This is because a steady-state scenario suggests that the winning bidders would factor in more than the transaction costs of their current bid: They would also include the transaction costs arising from the likely number of losing bids required to secure the winning bid, subject to the competitive pressures of each bid contest.

By comparison, the second wave of Canadian P3s appears to have a similar level of transaction costs, at least for the AFP projects undertaken in Ontario. However, Canadian data also indicate that the incremental portion of transaction costs, attributable to undertaking P3 transactions rather than conventional procurements, is relatively modest. It is also important to note that transaction costs under P3 procurements are not simply higher than those under conventional procurements: These costs, notably those related to due diligence, tend to be incurred much earlier than in a conventional procurement process characterized by a succession of procurement exercises.

According to the VfM studies issued by Infrastructure Ontario, the public sector transaction costs for 28 AFP projects (see Table 3) for which data are available vary from 1.5 per cent of the AFP budget in the case of the Niagara Health System project to 6.7 per cent of the AFP budget in the case of the Sudbury Regional Hospital (unweighted averages). On average, these transaction costs are 3.5 per cent of the AFP budget (unweighted average) and, unlike costs in the U.K. study, include the advisory fees and project management costs for the entire contract term.35 The Infrastructure Ontario VfM studies also provide an estimate of what the public sector transaction costs would have been for the same AFP projects if these had been procured conventionally. The incremental transaction costs attributable to the AFP approach, which are calculated as the difference between the transaction costs incurred under the AFP approach and the (lower) transaction costs that would have been incurred under a conventional project delivery, are 1.8 per cent of the AFP budget (unweighted average). In other words, the incremental transaction costs attributable to the AFP procurement approach are relatively modest, at least when compared with the typical cost overruns in conventionally procured projects. Moreover, these incremental transaction costs have been declining over time because Infrastructure Ontario relies on standardized documentation and other savings from multiple transactions in similar asset classes.36

The incremental transaction costs for the British Columbia P3s in our evidence base appear to be of a similar order of magnitude. In the two projects for which the incremental transaction data were published-the Abbotsford Hospital project and the Kicking Horse Canyon Phase 2 - the incremental transaction costs were 1.9 per cent and 3.5 per cent of the respective public sector P3 budgets.37

The transaction costs associated with the second wave of Canadian P3s appear quite reasonable relative to the U.K. levels, particularly since the United Kingdom has historically imposed a high degree of document standardization across its private finance initiative (PFI) project agreements (which would help keep down costs).38 However, the United Kingdom does not have a specialized central agency to manage PFI procurements. (Partnerships UK does not manage PFI procurements.) Every line department has its own PFI unit and manages its own procurements.

In contrast, the second wave of Canadian P3s has benefited from the establishment of specialized provincial P3 agencies, including those within central agencies of provincial governments. This may have resulted in some standardization of P3 agreements, particularly within each province and within asset classes with many transactions, such as hospitals. More importantly, each of these agencies has introduced a standardized procurement process that is replicated for each transaction. Although any comparison with other jurisdictions such as the United Kingdom remains highly speculative, it is quite likely that the procurement processes for the second wave of Canadian P3 transactions have reduced transaction costs relative to those incurred by comparable first-wave Canadian P3s. Most of the first-wave procurements were managed as one-off processes, in some cases by local authorities for which the P3 deal represented the first and last procurement of its type.

___________________________________________________________________________________________________________

33  For example, Vining and Boardman argue that "the ten case studies [reviewed in their article] indicate that the potential benefits of P3s are often outweighed by high contracting costs." See Vining and Boardman, "Public-Private Partnerships," p. 9.

34  Dudkin and Valila, "Transaction Costs." Note that the 5 per cent represents the costs not of one but all the failed bidders on an average project.

35  If seven of the projects that have an operating period are excluded, the transaction costs are 3.8 per cent of the AFP budget. The ratio of incremental transaction costs to the AFP budget rises slightly to 1.9 per cent when the same seven projects are excluded.

36  By grouping the AFP projects based on the year in which they reached financial close, we find that the 16 projects that reached financial close in 2007 had incremental transaction costs that amounted to 2 per cent of the AFP budget. The average incremental transaction costs dropped to 1.7 per cent for the seven projects that reached financial close in 2008, and to 1.5 per cent for the four projects that closed in 2009. (VfM results are not yet available for the Windsor Regional Hospital and the Toronto South Detention Centre projects, which also closed in 2009.) All averages are unweighted.

37  The VfM reports issued by Partnerships BC usually include transaction cost data for the P3 project under review. However, they do not usually include an estimate of the transaction costs under the public sector comparator. The latter are required in order to identify the incremental transaction costs for each project.

38  See HM Treasury, Standardisation of PFI Contracts.