Here are the main critiques Engel, Fischer, and Galetovic set forth in their Regulation article, drawn from their National Bureau of Economic Research working paper “When and How to Use Public-Private Partnerships in Infrastructure: Lessons from the International Experience”:
- P3s “are routinely renegotiated” after the winner has been selected, suggesting a kind of crony capitalism.
- The only real benefit of P3s is lower maintenance cost, not greater value for taxpayers via transferring risks to investors.
- The cost to the government is the same whether the project uses availability payments (explained below) or tolls (assuming the government would have chosen to use tolls in the first place).
- Revenue risk is too great for the market to bear, so if a project is financed based on revenue risk borne by the company, it will not be financeable without government revenue guarantees (which Engel, Fischer, and Galetovic say are common).
Many of those points do apply to Latin America and to some countries in Europe, such as Germany and the United Kingdom. But they are not problems in either Australia or the United States, where highway P3s are generally toll-financed, with major risks transferred from the state (i.e., taxpayers) to the P3 investors.
Outsourcing risk/ It is important to understand that two different models are used worldwide for large-scale highway P3 projects. In Canada, Germany, much of Latin America, and the UK, these projects are generally financed based on the government agreeing to provide annual “availability payments” (APs) over the life of the long-term P3 agreement (typically 30–35 years). Bond buyers finance the deals based on the government’s commitment to making those payments every year. Some of the projects do include tolls, but the tolls are paid directly to the government, which can then use the revenue to cover some of the cost of making the annual APs.
By contrast, most U.S. and Australian toll P3 projects are financed by toll revenues paid directly to the P3 companies by their customers. This means the risk of insufficient traffic and revenue is borne by the company, not the state (as in AP-financed P3s). Other risks transferred include construction cost overruns and late completion. At the outset of the project, the value of these risk transfers, as well as hidden state costs such as insurance (and the P3 company paying taxes to the extent it is profitable), are compared with traditional government procurement in a detailed Value-for-Money analysis to determine if the project is better done traditionally or as a P3. Hence, there is far more value in a revenue-risk P3 than just assured maintenance. Also, in Australia and the United States there is no history of “renegotiation” of the deal after the winning bidder has been selected. Indeed, P3 companies can go bankrupt if they badly miscalculate.
As for the cost to the state being the same for a P3 and conventional procurement, this might be the case for the AP type of P3. But it is clearly not the case when the private sector provider makes the public case for tolling as the only feasible way to get the project done in the near term and willingly takes on the revenue risk. (This has been called “the outsourcing of political risk.”) Such projects might have waited several decades for the state to eventually amass the resources to build them — or might not be done at all.
The most challenging claim in Engel, Fischer, and Galetovic’s article is that revenue risk is a major obstacle to using the P3 mechanism unless the state provides revenue guarantees. The authors created an alternative model under which competition for the project is based on the lowest net present value of revenue each bidder would require over the life of the project. The agreement would thus have a duration that is variable: the company would go on collecting tolls until it has achieved that total, whether that takes 20 years or 95. This changes the model from “revenue risk” to eventual guaranteed profit. That would certainly subject this form of P3 to denigration as a form of “crony capitalism” if it were tried in this country.
In their 2017 article titled “Sharing the Big Risk: Assessment Framework for Revenue Risk Sharing Mechanisms in Transportation Public-Private Partnerships” (Journal of Construction Engineering Management), Stanford’s Michael Bennon and colleagues analyzed various ways in which revenue risk could be transferred in long-term P3s. They included having government take all the risk (via the AP model), Engel’s variable-length model, minimum-revenue guarantees by the state, and some form of revenue sharing. Bennon et al. evaluated each based on its effect on the borrowing capacity of the project. Their conclusion was that the least-bad approach would be a kind of revenue-sharing structure that included a small minimum-revenue guarantee and offered high potential upside for the company. Their model explicitly finds that the variable-length concession “does little to increase project leverage and hence reduce the cost of financing.”
Focusing on consumers/ Even more important, in my view, is that by providing essentially a guaranteed rate of return to the investors, the AP concession takes away the incentive to earn a higher return by doing a superb job of attracting toll-paying customers and delivering high value to them. In a toll road project done as an AP concession, the company has no incentive to design and operate the road so as to maximize convenient opportunities for would-be customers to enter it or seek to win their ongoing loyalty. The firm’s only two tasks are to build the project competently on time and then maintain it to the state transportation department’s standards. That is why this form of concession company can be caricatured as “a construction company plus a contract maintenance firm.”
