Infrastructure is a sturdy perennial on the G8 and G20 agendas. Invariably there is a plea for more public-private partnerships (PPPs). With the prospect of budget austerity as far ahead as the eye can see, the case for getting the private sector to pay for infrastructure seems compelling.
If these international discussions could help clarify just how the private and public sector should work together, the benefit would be substantial.
Good policy needs a clear analytical starting point, and this had been largely absent in the complex world of infrastructure. Almost universally, PPPs have been the default response when vital infrastructure needs have run up against tight government funding constraints. Indonesia, for example, plans its infrastructure program on the fanciful basis that the private sector will fund more than half of the projects, with this figure being a simple residual determined by legislated budget constraints.
Funding is a curious starting point for the PPP debate.
Governments everywhere are criticised for operational deficiencies, lack of dynamism, inadequate cost management, technical conservatism and worse. It would be easy for the private sector to out-perform the public sector in many aspects of infrastructure. Yet leaving aside the basket-cases, any half-respectable government can borrow more cheaply than its private sector can. The private sector can help in many areas of infrastructure projects, but why privatise the funding component?
Two answers are routinely offered. First, governments have a well-deserved reputation for excessive borrowing. Thus the constraints are self-imposed deficit limits, or imposed externally by credit-rating agencies. The inability to fund infrastructure from the budget is collateral damage from this generalised reputation for profligacy. [fold]
Second, private funding may provide a valuable filter or hurdle to prevent 'white elephant' infrastructure from being built. If the project has to earn a commercial return and if the private sector bears all the risk inherent in the project, PPPs provide discipline on a sector not famous for this quality.
In practice, however, this discipline has rarely worked well.
Few infrastructure projects lend themselves to simple revenue collection. A good part of the benefits are often external, unable to be collected as revenue (a toll road delivers congestion relief to those who don't use it, as well as to those who pay). In many cases there are irresistible social pressures to supply the service at less than cost. In just as many cases, the service is supplied in a monopolistic way (no point in having two competing water pipes), so pricing can't be left to the free market.
Moreover, the universal experience is that the private sector is particularly skilled at shifting residual risk to the public sector. When, after long experience, risk transfer was more firmly tied down (the Sydney cross-city and Lane Cove tunnels, for example) the experience led the private sector to withdraw from this PPP model.
Reflecting this experience, a different model seems to be evolving in Australia: the government builds the infrastructure and then sells it to the private sector to operate. We might hope that this model will evolve further because it is unsatisfactory in its present form. The prospect of sale may impose useful discipline on project selection and implementation, but the experience so far is that the funding cost for a risk-free project (eg. the Sydney desalination plant, which is already operating, has a take-or-pay contract) is around twice that of a simple bond issuance.
The funding issues raise a wider problem: the inadequate analytical tools used by financial markets to assess government borrowing. Credit-rating agencies have demonstrated their inability to provide useful guidance for sovereign debt. They were tardy in identifying the problems in the European periphery and just as slow to see the enhanced borrowing potential in emerging Asia. The professional debate on government debt is weaker still. The flawed analysis of Reinhart and Rogoff not only went unchallenged, but was used as a rule of thumb to identify borrowing problems; its misleading legacy remains apparent.
PPPs illustrate the inadequacy of current debt evaluation approaches.
Suppose a government improves its debt position by selling a hospital to the private sector, contracting ongoing routine payments to the new owner for service provision. Has the government's fundamental financial position improved? Or, on the other hand, if the government borrows to build and operate a viable infrastructure project, is its financial position weakened by the additional debt, when this is offset by an additional earning asset? We still don't have a framework to assess this degree of funding complexity, only simplistic rules of thumb.
There is little doubt that a closer working relationship between governments and the private sector in the provision of infrastructure has the potential for huge efficiency gains. But these will come from providing services more competitively (such as in telecommunications, where technological advance allowed what had been a government monopoly to be fully privatised in a competitive environment), or where privatisation facilitates full price recovery (toll roads), or when the sharper accountability of the private sector allows operational benefits and better project appraisal.
It's less likely that substantial benefits will come simply from shifting ownership to dodge the current set of mis-specified constraints on government borrowing, either self-imposed or imposed by financial markets.
Australians pioneered some of the initial PPP models and our investment banks have taken them to the rest of the world. There will be opportunities, as we improve the current imperfect models, to use this experience to put substantive content into the international discussion.
Image by Flickr user 1950sUnlimited.