The system by which roads are provided in Australia is in a mess. Roads are funded from revenues collected from fuel excises in the most part by the Commonwealth which builds some roads. Most roads however are provided by local government (which gets funds from the Commonwealth, from local taxes and from parking fines) and by the states which again gets funds from the Commonwealth but which also collects vehicle registration fees, licence fees, speed fines and taxes on car insurance premiums.
Road construction and road maintenance decisions are typically made by engineers and bureaucrats on the basis of cash budgets determined by a political bargaining process between the states and the Commonwealth. A cash budget is provided and roads are designed on the basis of this. The amounts of money involved are huge.
Moreover while motorists pay for the costs of roads the payments are not designed to provide accurate signals that help determine the appropriate use of particular roads.
There is a real case for reform that assigns a substantial role to economics. I addressed some of these issues here but in this post I emphasise institutional issues.
The Australian Transport Council (ATC, 2009) have been concerned with the implementing the first-phase of the COAG road reform agenda namely providing what is hope to be an effective and efficient transport marketplace for Australia. This involves delivering improved road, rail and public transport networks to service the anticipated doubling of the Australian freight task over the next 20 years and greatly increased passenger movements. The main issues are strategic network planning, infrastructure funding, investment decision-making and pricing for infrastructure services. The idea is to impose market and commercial opportunities and disciplines particularly with respect to supplying road services.
My view is that the supply-side issues associated with such reforms are very complex. Indeed my view is that few of the really intractable supply issues have been resolved.
The demand side issues provide practical if not theoretical difficulties. Eventually it should be possible to selectively price travel on the basis of congestion, axle loads and type of road – this is mass-distance-location charging Therefore suppose the marginal social costs of using a road can be identified and billed to road users as an ‘attributable cost’. The capital costs of roads not affected by the extent of usage (‘common costs’) can be collected by levying a fixed annual registration charge on road users. There are difficult issues here such as recouping specific costs of investing in road durability but these can be ignored here. Suppose that a two-part tariff can be designed which recovers the capital and usage-related costs of each road. Such charges will determine patterns of usage on roads and net revenues that can be attributed to roads.
Of course while such revenues can be identified ex post they can only be forecast ex ante. Such forecasts will need to inform supply decisions both in regard to the upgrading or maintenance of existing roads and the construction of new roads.
It is not the suggestion here to suggest that difficult issues of user pricing do not arise but simply to suggest that they can be clearly articulated using well-known theory. The practical issues are obvious in relation to assessing road damage costs. ATC (2009, p. 23) state:
‘Pavements across Australia vary widely in strength and quality with many sections which are fragile, particularly during wet periods. Road pavement and bridge performance varies with construction specification, age, traffic and maintenance history, source of construction materials and climate, making engineering analysis and overall conclusions challenging. Road use/cost relationships are complex and not well understood. Empirical research to improve knowledge is necessarily long-term given road life and the number of axle passes needed for modelling impact’.
This claim seems incontrovertible although such issues can in principle be addressed. What seems more problematic is to come up with institutional designs that encourage road supply agencies to optimise road construction on the basis of such empirical knowledge and to provide a regulatory environment that encourages agencies to levy user charges which reflect such marginal costs.
To simply suggest that road reform can follow the lead established by rail, electricity and telecommunications reforms is not helpful. This does not seem obvious given the difficulties of assessing and monitoring highly-decentralised road use demands and in assigning charges. Road commercialisation seems much more complex than commercialisation in these other industries.
Road supplies. What is sought is a road supply system which matches revenues received from roads with the costs of constructing and maintaining roads.
A starting point is for a national agency, in coordination with other jurisdictions, to devise a strategy for land transport. This would include short-run information on committed rail and road transport investments over the next 5 years and a medium to long-term plan that provides forecasts of various traffic volumes over a longer horizon. This would inform infrastructure road supply decisions in various jurisdictions, improve coordination across jurisdictions and facilitate forward planning.
Then regional, state and perhaps national road supply agencies would then be set up to deliver efficiency by assigning them commercial incentives and accountabilities. Incentives to respond to consumer demands would be created by linking revenues accruing to the supply agencies with road use charges levied on the specific roads they manage. The intention then is for roads to be constructed when capital and maintenance costs are covered. Then road agencies would not be constrained by current budgets but by the profitability of investments in roads. Agencies could then issue equity of borrow in capital markets to fund road provision on the basis of expected revenues and costs. Road agencies would be subject to capital market disciplines.
Such agencies would be ‘regulated’ to limit the abuse of local monopoly power and, subject to this, would develop road projects only if and only if ‘revenues’ from user charges and imputed community benefits exceeded costs. The test would have to include alternative options such as public transport and would extend to regional groups supplying local roads. The regulation would permit prices to be set to permit a ‘reasonable return on investment’.
Several comments seem appropriate.
First, the taskforce report does not contain economic analysis. This can be a useful adjunct towards considering what is an inherently difficult issue. As a minor issue it ignores the substantial literature on when optimally designed roads do self-fund on the basis of user charges. Contributions on this highlight the role of road damage costs – see Mohring et al (1962), Small et al. (1989). This literature discusses the self-funding issue when there are economies of scale and hence the need for ongoing road subsidies. Small (1999) extends these results to allow for imperfections in input markets such as land markets. This is important since, once transactions cost obstacles to user charging are resolved, it is important to understand the circumstances where such user charges will self-fund. Clearly if roads deliver community service obligations and other non-commercial objectives they will not self-fund. But they may also not self-fund when certain returns-to-scale characteristics which are purely technical features of the roads arise. This is a ‘loose end’ that is conceptually easy to tie down. The main issue is to determine the returns to scale properties of different types of road constructions.
Second, and more substantively, the issue of ensuring economy-wide efficiency when decentralised road supply agencies make independent assessments of road investment needs is not addressed. The independent agencies will have monopoly power and that will be regulated to yield a ambiguously defined ‘reasonable’ rate of return. Even ignoring community service obligations (CSOs) this is a non-straightforward task. Regulators need to evaluate how pricing rules influence social surplus levels which include consumer benefits as well as supplier benefits. Interventions and supplementary public funding will also be necessary to meet CSOs that are not going to be self-financing. There is in fact no such thing as a ‘competitive national transport market’ and it is doubtful there ever will be. Competitive markets might exist for road construction but they do not exist for road provision.
In part CSOs and capital costs of road construction that are not durability investments based on the needs of specific vehicle types would be funded from fixed charges that are currently collected as registration fees. These would not necessarily disappear under a user-charges based system. Mechanisms for transferring a component of such fixed charges to specific road projects need to be devised.
Generally it is not at all clear how commercialisation arrangements with respect to road provision can be made operational. Major roads, such as interstate highways, would need to be centrally determined by a national highway agency which would account for the effects of expanding or upgrading certain parts of the network on traffic flows in other parts of the network. If such roads are provided by decentralised jurisdictions but on a basis that is consistent with a national road transport strategy that amounts to the same thing as central provision. This is a difficult investment planning task in itself since, apart from complicated network effects, forecasts of traffic would need to be made on the basis of new and improved roads that are subject to charging. This would necessary draw on centrally provided forecasts of freight and passenger traffic demands as well as forecast impacts of new urban and lower tier road developments but, in addition, assumptions would need to be made about demand responses to charging. Possible privatisations of the management of such efforts would be limited by the high upfront capital costs involved and the high uncertainty of demands for the final road demands. Plausibly the lower capital costs enjoyed by government would favour public operation. Private operations would almost certainly result in contracts which guaranteed minimum rates of return thereby introducing moral hazard into road design behaviour.
Given the plans for major national roads, lower level roads could be planned recursively by local agencies given central and decentralised information about local transport demands together with information about forecast provisions of major roads by the national authority. The benefits of decentralisation here rest in the use of local knowledge. Provision local plans need to be fed back into the planning mechanisms of the major roads provider. Lower level road networks in urban areas would need to be able to user price to limit bottleneck ‘last kilometre’ features.
This type of recursiveness which assists the design of a coherent road network is most applicable to links between national highways and urban settings. It is much less applicable to road planning issues in large cities which planning of major roads needs to occur in the context of broader planning issues which link optimal road hierarchy issues with land use planning. Major arterials and ring roads in a metropolis need to be designed with the inclusion of smaller feeder roads in mind. There is not much recursiveness to simplify planning here.
Generally the mechanisms by which comprehensive plans for high level roads can be linked to local planning are not secondary or simple issues and need to be set out.
Finally, the issues of shifting toward a system based on user charges and productive efficiency is not only an issue of implementing charges. These concerns are important but there is also the issue of how charges should be directed toward the existing network of Australian roads which may not be optimally designed with respect to features such as durability. Here it might not be efficient to pursue cost-recovery on the basis of user charges if, for example, roads have sub-optimal durability.
These comments should not be taken to imply a rejection of the COAG reform process but rather suggest directions for refining it or assisting its development. The importance of some type of reform itself is undeniable. This was highlighted by announcements on 27 April, 2009 when state and Federal governments announced a $37b package mainly devoted to building roads in the eastern Australian states (AFR, 2009). Although linked to longer term productivity issues the main stated motivation for such a road building program was to stimulate employment over the next 5 years. Apart from this there seemed no explicit motivation for a package which costs around 3 per cent of Australian GDP. The Business Council of Australia’s President Greg Gailey is quoted in the AFR report as stating:
‘Our infrastructure problems are not only driven by poor planning and unclear objectives, but also by inappropriate infrastructure pricing, the absence of effective national infrastructure markets and poor regulation.’
Almost any market-based planning scheme would outperform current approaches to road policy-making.
Australian Transport Council, COAG Road Reform Plan Phase 1 Report, Australian Transport Council Report to The Council of Australian Governments, March 2009.
A. Hepworth, L. Dodson & M. Dunckley, ‘States Agree $32bn boost for transport’, Australian Financial Review (AFR), 27th April 2009, p1, p6.
H. Mohring & M. Harwitz, Highway Benefits: An Analytical Framework, Evanston, Illinois, Northwestern University Press, 1962, 70-87.
D. Newbery, ‘Pricing and Congestion: Economic Principles Relevant to Pricing Roads’, Oxford Review of Economic Policy, 6, 2, 1990, 22-38.
K.A. Small, ‘Economies of Scale and Self-Financing Rules with Non-Competitive Markets’, Journal of Public Economics, 74, 1999, 431-450.
K.A. Small, C. Winston & C.A. Evans, Road Work: A New Highway Pricing and Investment Policy, Brookings Institution, Washington DC, 1989.