I’ve been in Sydney most of this week attending the Australia’s Future Tax System – A Post-Henry Tax Review. It had some excellent speakers and was for me – not a taxation specialist – very informative. I particularly liked John Freebairn’s overview and a superb paper by Ben Smith which clarified my views on the resource super tax. My own contribution – joint with David Prentice – is on the road transport sector and is reproduced in draft format over the fold. Comments are very welcome as the paper is now being revised. All the papers from this meeting will be published in a forthcoming book. The opening address by Ken Henry was characteristically forthright but in my view terribly misrepresented in blogs, the press and by commentators such as Warwick McKibben and others. Henry was suggesting that economists get unreasonably cantankerous about second-order issues and, on climate change policy, I agree. I’ve got some more substantial comments on this issue that I will defer. I have much respect for Ken Henry and on how he operates the Commonwealth Treasury. He is a straightforward person with high intellectual honesty who has vast experience at dealing with government. He is also amazingly knowledgeable on tax issues – hardly surprising given his background.
After listening to Ben Smith I’ve changed my mind on the resource super tax. It is a tax grab but, if implemented as stated, will not harm exploration effort unless there is the assumption that imposing it raises future further sovereign risk issues. There is a slight interaction issue, detected by Professor Jack Mintz, that means a non-neutrality will arise if the resource tax is imposed with a company tax but this is easily sorted out by making up-front cost refunds accruing to the miners non-taxable. Otherwise the debate between governments and the miners is just a ‘cake-eating task’ – who gets what share of the cake? The industry is lying about the distortionary effects on exploration of the tax to protect the stake in their firms owned by shareholders. This might be justified – shareholder values will be diminished – but it is still a blatant distortion since exploration activity will not be inhibited.
The tax has neutral effects on exploration and the scale of the industry because its not really a tax but a 40% government shareholding in ventures with the government contributing 40% to costs as well as getting 40% of dividends. There are efficiency gains in cutting royalties by eliminating them but refunding to the states from the Commonwealth’s income share these royalties. I’ll refrain from correcting some misleading comments in earlier posts – that suggested non-neutrality with respect to exploration effort – and make a general mea culpa here. When I get the time I’ll insert links between these correcting comments and the earlier posts.
Reforming Taxes and Charges on Australian Road Transport (Draft: Comments welcome).
Abstract: This paper questions how taxes and charges on road transport in Australia should be set to best achieve economic efficiency. It is concerned both with avoiding taxes and charges that impose significant efficiency costs and with setting alternatives that promote economy-wide efficiencies. It also considers whether such ideally set taxes and charges should be hypothecated to cover road sector costs of capital – some argue that revenue signals can be used by planners to generate supply-side efficiencies. The case for shifting taxes and charges to a user pays basis has strengthened over recent years. However arguments for relating supply decisions to tax and charge revenues involve issues of meeting community service obligations and of the devising appropriate relationships between decentralised and centralised decision making. Road network design issues need to be centralised although local governments, with revenue hypothecation, as well as the community at large, will benefit through their ability to efficiently resolve ‘last mile’ problems. The case for user charging is strong and independent of doubts raised concerning the case for supply-side hypothecations.
1. Introduction. Australia’s road transport sector underpins our national productivity. Australia is a geographically large country with a low average population density which is, however, highly urbanised particularly in eastern seaboard capital cities. Trucking and rail transport are an important means of transporting raw materials to population centres and ports while road transport is important for transporting people between what can be distant locations.
Significant congestion issues arise in the capital cities where supply-side solutions to growing transport problems face rising costs. Outside of the major cities there are significant issues of best designing roads where traffic densities are often low relative to those of other developed countries. Constructing European-style high-durability roads is not feasible in Australia given lower traffic volumes and the scale of freight tasks so the cost effective management of road damage costs from heavy vehicle use of roads is a significant issue. For the most part congestion issues in urban centres and heavy vehicle issues elsewhere are separate concerns. There are, however, interactions because of the pressure on trucking operators to increase loads to improve operational economies and to make precise origin-to-destination freight deliveries which avoid reloading. Such deliveries involve ‘first’ or ‘last mile’ issues whereby local governments must decide whether to grant heavy vehicle access to low durability local road infrastructure. Without either pricing of such access, or inefficient lump-sum compensations, local governments will incur costs without proportionate benefits, and respond by failing to build optimal road infrastructure and inefficiently restrict access to existing roads. Similarly State and regional governments, whose roads carry high levels of freight, are also subject to disproportionate cost burdens unless damage is priced and revenues hypothecated, and a similarly suboptimal use of roads will result.
Like the earlier Taxation Review Committee (Asprey, 1975) this paper accounts for revenue-gathering efficiencies. However, the Australia’s Future Tax System Review (the Review) focus, on promoting economy-wide transport sector efficiencies is also pursued. The present paper provides a non-technical synthesis of the report by Clarke and Prentice (2009) (CP) to the Review and incorporates views from the final Review report (AFTSR, 2010).
This paper is structured as follows. Section 2 provides background. Section 3 discusses fuel excises and other vehicle and driver-related charges. Section 4 discusses congestion, pollution as well as accident and insurance externalities. Section 5 deals with road damage charges and supply-side hypothecation. Section 6 synthesises conclusions. The approach employed throughout uses partial equilibrium analysis that aggregates inefficiencies across markets with each market viewed as operating independently. The alternative general equilibrium approach is discussed in CP. General equilibrium qualifications to the predominantly partial equilibrium conclusions presented are noted.
2. Background. Travel by private vehicle incurs private costs borne by those travelling (time costs, fuel, vehicle capital, maintenance and depreciation costs, traffic accident and insurance costs, car registration, licence fees, fines and sometimes road use and parking charges) and unpriced social costs born by the larger community (congestion, unpriced parking costs, pollution, noise, road damage and some traffic accident costs). These latter costs are externalities.
In a ‘first-best’ economic world social costs would be internalised by levying a specific charge on the aspect of travel which creates the costs. This charge would often depend on the identity of the driver, the type and mass of the vehicle driven as well as on both when and where travel occurred.
If road use and parking are underpriced, there will be socially excessive demands for roads and parking. If attempts are made to satisfy these under-priced demands, too much will be invested in capital items such as new road supplies, parking spots and road maintenance. Moreover, road expansions may have limited efficacy in reducing social costs because of latent demands for travel – more commuters will elect to make road journeys when congestion falls with road supply increases. Indeed, paradoxes can arise for micro-level road network planning if congestion externalities are unpriced. Problems of excess supply and counterintuitive planning outcomes never arise if roads are correctly priced (Arnott and Small, 1994).
Finally, the failure of private road travellers to internalise the social costs of private vehicle use increases the relative price of transport alternatives, such as bus or train services compared to private vehicle use. If public services are priced at marginal cost this reduces demands for mass transport below what is socially optimal. These transport alternatives often provide lower negative externalities per commuter or per unit of freight than does private vehicle use.
With substantial fixed capital costs of providing train or bus transport services there is an economies-of-scale rationale for public subsidies directed towards their provision. There are also arguments for subsidies as a ‘second best’ means of promoting efficiency if there are constraints on the ability to fully price private road use.
Many external costs of road travel are not priced in Australia. Instead, a mix of fixed and variable charges – a two-part tariff – is applied to use of private vehicles. There are fixed charges of driving private vehicles including registration and vehicle licensing costs. There are variable charges mainly based on fuel consumption that is linked to distance travelled. In part charges reflect revenue-raising objectives. Many transport-related activities (such as fuel purchases) are in inelastic demand which provides a conventional efficiency-based rationale for concentrating excises on these items to raise public revenues.
Charges include goods and services tax, tariffs on imported vehicles, taxes on insurance, stamp duty on trading motor vehicles, licence and registration fees. Partly these charges recover costs of publicly provided services. For example car registration charges partly cover car registration services plus compulsory third-party personal insurance. Licence tests encourage minimum standards of driver education while the institution of licensing itself limits driving by unsafe drivers. Sales taxes, import tariffs and stamp duty charges can be used to promote use of particular types of cars (domestically-produced, non-luxury, low fuel use, alternate fuel use and those inflicting low road damage costs). These charges, being fixed, have their main impact on once-and-for-all vehicle purchasing decisions but with limited impact on ongoing social costs once purchase has occurred.
Vehicle-specific fixed charges capture some externalities. As mentioned, road damage costs are related to the mass of vehicles (specifically their ‘axle load’), to distance travelled and travel location. These damages can be more accurately addressed by vehicle specific charges related directly to axle loads and the extent and location of road use.
Excises on fuels are an important transport charge that provides significant revenue to governments around the world. Costs to motorists vary with the extent of vehicle use. These taxes are popular with governments because fuels provide a broad tax base with relatively price-inelastic demands. There is also often an implicit motivation to tax certain fuels to encourage substitution away from exhaustible oil resources if it is supposed current prices do not reflect future oil scarcities. This is public insurance that reduces the adjustment costs fuel users would experience in the future were they exposed to possible rapid fuel price increases. Given the substantial uncertainties about fuel supplies the appropriate policy may be to encourage agents to make anticipatory adjustments now.
Fuel excises can also be motivated as an ex post attempt to approximately capture social costs generated in the absence of comprehensive user charges. This rationale for fuel excises can be based on transactions cost obstacles to user charging. However this approximation is rough. Many external costs depend on the type of vehicle driven and where/when it is driven rather than on its fuel consumption. Using fuel excises to address external costs saves transactions costs and helps promote fuel efficiency but will only imperfectly internalise many road use externalities.
For example, congestion costs are not well captured by fuel taxes since travel in uncongested areas, where no externality arises, is taxed at close to the same rate as travel in congested areas. Similarly local pollution and noise damage associated with travel is setting-dependent. Traffic accident costs are linked to distance travelled, and hence to fuel usage, but also depends on where travel occurs and on driver characteristics.
Fuel taxes can be used to internalise global public bad characteristics of fuel consumption – carbon-based fuels release CO2 that contributes to climate change. Such emissions are not restricted to the transport sector suggesting the need to rely on broader policies, such as the proposed Carbon Pollution Reduction Scheme – a ‘cap-and-trade’ emissions trading scheme. This scheme has recently been put on hold until at least 2012 so fuel excises can be viewed as capturing such otherwise unpriced external costs at least until that date.
3. Excises on Road Transport Fuels and Other Government Vehicle-Related Charges.
The current system of taxation of road transport is best described as a set of two-part tariffs. The fixed component is composed of the charges associated with purchasing a vehicle, such as stamp duties and annual operating charges such as registration. The variable component is fuel excise. The fixed components vary by state because registration and stamp duties vary by state. In addition, imported and luxury cars attract additional taxes. The variable component varies by the type of fuel used, type of use (business or personal) and the extent of use. There are two issues with these taxes. First, are these taxes an efficient way to raise general revenue? Secondly, does this system of taxation provide incentives to improve the efficiency of the transport sector by, for example, internalising externalities? In this section the first of these questions is examined. The second question is dealt with in subsequent sections.
Fixed Charges. Here the taxes and charges that do not vary with the use of the vehicle are examined. The most important of these are tariffs on imported vehicles, the luxury car tax, annual registration charges, stamp duties when vehicles are sold and levies on vehicle insurance.
Tariffs and the luxury car tax are imposed on certain types of cars. The tariff imposed on new passenger motor vehicles, off road vehicles, second hand cars, light commercial vehicles and vehicle parts was reduced to 5 per cent on January 1, 2010. Used vehicles, imported in volume, are subject to an additional charge of $12,000. These measures apply to a substantial proportion of new vehicles. According to Glass’s Information Services (2008) in 2007 about 75 per cent of passenger vehicles, 88 per cent of light trucks and 60 per cent of heavy trucks were imported. However the breadth of protection is limited because a substantial share of these vehicles are subject to lower tariffs under free trade agreements with the US and Thailand and as a consequence of the Australian Competitiveness and Investment Scheme (ACIS). The Productivity Commission estimated the economy-wide efficiency losses from higher tariffs and the ACIS scheme at $500 million per year (Productivity Commission, 2008). Other estimates, such as those in Dixon (2009), are smaller, though CP concluded that the true losses were unlikely to be as low as Dixon’s. The tariff effectively raises the price of most vehicles (both imported and import competitors), reduces purchases of new cars and distorts the pattern of vehicle purchases. These distortions create some welfare losses – though they will not be huge at a 5 per cent tariff level – CP concluded that the tariff on cars and the charge on imports of used cars should be abolished.
The luxury car tax is imposed on cars with a pre-tax price above a threshold of about $57,000. CP estimate the tax was applied to about 10 per cent of cars purchased in 2007. There are no estimates of the elasticity of demand for luxury cars so CP estimate the welfare losses using a range of plausible elasticity estimates. Because the market for luxury cars is small, the total welfare losses are not large, though the loss per dollar collected is higher if the elasticity of demand is higher. The luxury car tax also includes a higher threshold for energy efficient cars. CP concluded that the tax should be abolished. It distorts choices among luxury items and seemingly has no arguments favouring it as a redistributive tool compared to income taxation. It is disappointing that the Rudd government ruled out its abolition in its initial response to the Review which also recommended its abolition. The higher threshold on energy efficient cars is unlikely to be effective as evidence from Prentice and Yin (2004) suggests energy efficiency is of little concern to purchasers of luxury and sports car.
A second set of charges are imposed on the ownership of vehicles – annual registration fees and taxes on car insurance. For 2006-2007 CP calculated from the Budget papers for each state that annual motor vehicle registration fees and taxes raised about $3.8 billion across all states while surcharges and levies on compulsory third party insurance raised another 0.2$ billion. Registration fees and other taxes are similar across states but there are differences. Transaction costs would be saved by making them uniform across states. Widespread car ownership suggests these taxes do not discourage car ownership per se but as they do tend to increase with car size they may discourage consumers from purchasing larger cars. The popularity of SUVs before the recent increase in fuel prices suggests these effects are unlikely to be large.
A third set of fixed charges are imposed on trading cars. These stamp duties are imposed by the states and are about 3 per cent of market values (New South Wales Treasury, 2007). For 2006-2007, CP calculate, from the Budget papers for each state, they raised around $2b for the states. By discouraging trading, these duties inhibit consumers from trading toward their preferred vehicles. Thus if the fuel prices rise consumers will be more reluctant to swap a SUV for a less environmentally damaging hybrid car.
Taxes such as annual registration fees that differentiate between different passenger cars have low efficiency costs. Indeed, these taxes have an administrative cost-recovery component. Although there is no explicit hypothecation, such charges do also contribute toward the capital costs of roads. Taxes on trading vehicles restrict the types of cars held by individual consumers by making it more costly to trade to a preferred type when circumstances change, imposing welfare losses. They are less likely to significantly affect the total number of cars held. There are no efficiency arguments for these taxes so CP argue that they be abolished. The revenue could be replaced by increasing registration fees which incur low welfare losses.
Another issue of concern is whether the statutory formula used to estimate the fringe benefits tax on vehicles encourages excessive driving. The fringe benefits tax is used to tax employee income received in kind. This is difficult to estimate so a formula is applied to approximate the proportion of vehicle driving associated with private use. This formula assumes that the proportion of all driving for private use declines with the miles driven as certain threshold mileages are exceeded. Hence the tax paid declines as certain mileage thresholds are exceeded. This provides an incentive for unnecessary driving to reach higher mileage thresholds to reduce the tax paid. Kraal et al (2008) provide evidence supporting this view by noting of clustering of mileages just above thresholds at which the tax is lowered.
Fuel Excises. Fuel excises are the main variable charge for transportation. They are a major tax in Australia raising more than $13b in 2007-2008. (Budget Papers No. 1, 2009-2010). The tax paid varies directly with the consumption of certain fuels but not others. The role of fuel excise is discussed here in three stages. First, traditional arguments for including a fuel excise as part of a set of taxes for raising revenue are advanced. Second, arguments for removing support for specific excises are examined. Finally, proposals for reforming fuel excises to accommodate technological change are discussed.
An important feature of the demand for petrol is its extreme price inelasticity. Breunig and Gisz (2009) provide a recent estimate of own price elasticities for Australia of -0.2. This is very low – by comparison the demand elasticity for an addictive good, cigarettes, is around -0.4 (Scollo et al, 2003) which is twice that for petrol. Australian petrol demand is also relatively inelastic compared to demand in other nations (Litman, 2008).
The traditional Ramsey analysis for designing efficient commodity taxes places greater taxes on the products for which demand is relatively inelastic. Then quantity adjustments to taxes will be proportionally similar across products, minimising distortions to demands for all products. Given the low fuel price elasticities the Ramsey analysis supports a higher than average fuel excise. CP adapted, for Australia, a model of the optimal fuel excise for the UK and USA developed by Parry and Small (2005). The fuel excise, in their model, is used both for raising general revenue (along with an income tax) and as a charge for the negative externalities from fuel consumption, such as pollution, congestion and traffic accidents. CP’s simulations calculated that an excise of about $0.50 was required for reducing the negative externalities and that a tax of $1.50 could be justified for raising general revenue. This estimate is higher than Parry and Small’s estimates for the US, probably due to the relatively inelastic demand in Australia for petrol.
The Ramsey approach has been questioned as an approach to tax design because it presupposes targeting of a fixed amount to be collected and excludes as a policy variable income taxes. Kaplow (2008) argues that, if labour and leisure are weakly separable from the demand for goods such as fuels, and income taxes are available as a policy instrument, that there is no case for differential commodity excises.
Furthermore, introducing commodity taxes, in addition to labour taxes, can create byproduct efficiency losses beyond the markets for each commodity. This argument against differentiated excises observes that income taxes create inefficiency, because labour is effectively taxed but leisure is not. Thus income taxes encourage leisure and discourage labour supply. If commodities are complementary with labour or substitutes for leisure, then having specific commodity taxes increases the efficiency losses in labour market from labour taxes. Commodity taxes on top of income taxes encourage increased consumption of untaxed leisure. Goulder and Williams (2003) show, if fuel consumption has average substitutability with leisure, that ignoring the effect on commodity taxes on labour markets understates the inefficiencies from a fuel excise. Their numerical analysis suggests these understatement effects are large. This is because the labour market is large so that even small distortions are quantitatively important. Then it follows that excises on fuels should be set low to avoid efficiency losses. Broad-based commodity taxes are a more efficient way of raising revenue than differentiated excises.
The assumption that the taxed commodity is average in terms of its substitutability with labour is an important proviso. If the commodity is only weakly substitutable or strongly complementary with leisure the argument against specific excises fails. For example, if leisure is a complement with gasoline an increased gasoline excise reduces demand for untaxed leisure and reduces distortions in labour markets. If there is neither substitutability nor complementarity then the traditional partial equilibrium formulation works. Whether fuel is a substitute or complement to leisure is an empirical question requiring estimation of the cross-price elasticity of demand for leisure with respect to fuel prices. West and Williams (2007) provide evidence that gasoline is a complement to leisure. Furthermore, they compute optimal excises that are much larger than those suggested in earlier work. The intuition is that driving is a relatively time-intensive activity which is likely to be strongly complementary with leisure. More accurately the demand for leisure driving may be more price-elastic with respect to fuel prices than the demand for work commutes is, so a higher tax leads to less leisure-based travel. People drive less for their holidays when fuel prices are higher.
Based on the evidence of West and Williams (2007), CP use the estimate of the elasticity of demand by Breunig and Gisz (2009) to estimate the annual efficiency losses from the excises on petrol and diesel and conclude they are small. They also emphasise other advantages of the fuel excise. Costs of collecting taxes are also welfare losses as are resources spent in avoiding or minimising taxes. Income taxes require considerable resources for collection and much is spent avoiding them. By contrast, the simple nature of fuel as a product and the highly concentrated and vertically integrated petroleum industry means that tax collection costs are low and opportunities to avoid the tax are low.
Given that there remains a case for imposing an excise on fuel, there also remain important questions on the design of the tax. It is currently imposed on petrol and diesel with exemptions for business use. The exemptions for business use are appropriate as they avoid distorting the use of inputs as well as choices of consumption goods as pointed out by Diamond and Mirrlees (1971). Currently exemptions are granted only for use by vehicles above a certain size. CP suggest that, if avoidance issues can be dealt with, exemptions could be granted for business use by all vehicles – not only large ones.
Currently most fuels that are consumed for road transport are considered for excise. ABARE statistics, quoted in CP, for 2005-06 report that 92 per cent of the energy consumed in road transportation came from either unleaded auto gasoline or diesel. 7.5 percent came from liquid petroleum gas and the remainder from natural gas and bio-fuels. Recently hybrid cars, using both gasoline and electricity, have been commercially launched. Furthermore technological change, particularly in bio-fuels, and substantial charges placed on carbon-based fuels, may increase the importance of alternative fuels. If fuel excise is being used to raise general revenue, then there is no reason for setting differential excises across the fuels. There may be practical difficulties in collecting excises with greater fuel coverage. Provided alternative fuels and electricity for hybrid cars, are distributed through retailers such as service stations, collection is straightforward. However, for hybrid vehicles that are recharged at homes it may be more difficult to collect such excises.
Recommendations of Australia’s Future Tax System Review. Transport would be taxed differently if recommendations of the AFTRS (2010) are implemented. AFTRS recommend that both the fuel excise and most of the fixed charges be replaced with road user charges. This is consistent with their general approach which is that general revenue is to be mainly raised from personal and business income taxes, consumption taxes and resource rent taxes including land taxes. Externality taxes, such as road user charges, are seen by AFTRS as correcting market failures not as providing major revenue sources. Though, as discussed earlier, there is a case for abolishing some fixed charges, the limited empirical evidence suggests fuel excises could be retained given the large revenues gathered with negligible deadweight losses and the low transaction costs of collection. Finally, fixed road construction charges of road need to be recouped from a source other than user charges. Registration charges provide a convenient base for recouping such costs. If fuel excises are retained, AFTSR advocate taxing comparably competing fuels that inflict the same external costs as petrol, a view that is endorsed here. They also suggest that if fuel excises are retained they be indexed to the CPI. Consistent with this CP pointed out that failing to index fuel excises had effectively reduced the revenue-gathering ability of this excise.
4. Urban Externalities. The efficiency costs associated with congestion, parking, air pollution, noise externalities, traffic accidents and several other externalities are discussed. Core policy debates in Australia focus on congestion so this is emphasised. Road damage costs are addressed in Section 5.
Congestion. Congestion costs impact on all Australian capital cities but are most pronounced in Sydney and Melbourne. They are expected to increase across all Australian capital cities from $9.4b in 2005 to $20.4b in 2020 unless addressed by policy (BTRE, 2005).
The case for addressing congestion problems using user charges has sharpened in recent years. There have been advances in understanding how gains from congestion pricing increase significantly with the inclusion of gains from reduced traffic bottlenecks, from improved travel reliability and from the recognition of heterogeneous travellers who select between high- and low-value journeys (Arnott et al., 1993). Attention has been devoted on the political economy of introducing road pricing and on the need to win community support for such moves by offering of pricing trials (as in the case of Stockholm), through offering enhanced modal choices including choices of travelling on congested or uncongested roads and through advances in techniques for dealing with occasional users of priced city zones (Oehry, 2010). In addition, there is now much more practical experience of congestion pricing (in cities such as London, Oslo, Singapore, Stockholm and in the Netherlands as a whole) that those contemplating congestion pricing can draw on. Road pricing transaction costs have fallen as gantry- and GPS-based technologies have matured. Finally, supply-side congestion policies have escalated in cost as land values and construction costs in large cities have risen.
There is almost no Australian experience with congestion pricing and only rudimentary modelling of congestion. For example, the widely-discussed BTRE (2007) congestion cost estimates rely on a model using simple aggregative ‘stock-flow’ traffic dynamics. Such specifications understate gains from congestion management by ignoring gains from resolving traffic bottlenecks. Moreover, with the exception of the Sydney Harbour Bridge tolls, current road pricing is geared to cost-recovery without a peak load element and hence does not assist in smoothing peak travel demands. Actual road pricing – e.g. on CityLink and Eastlink in Melbourne – is geared to ensuring cost-plus recovery by private operators not to addressing congestion. Reforms need to be devised which renegotiate these contracts so short-term marginal costs not uniform charges are levied (Clarke and Hawkins, 2006). Future tolling contacts should require efficient pricing with, perhaps, compensatory transfers to pre-existing operators.
A wide range of congestion pricing options is now available. These range from comprehensive electronic pricing of whole countries to partial schemes which price roads in only key parts of a city or indeed only on particular lanes on a single road. Popular partial schemes – for example, cordon pricing and/or density-dependent tolls on major ring roads and arterials – impose spill-over costs from traffic diversions onto unpriced roads. Comprehensive pricing, using telematic or gantry-based technology, avoids such spill-overs. The actual choice of technology is often not critical – there are various mature technologies available and the choice should be based on what meets congestion control needs at minimum cost (Oehry, 2010).
Policies also need to be devised which address congestion arising from non-recurring events, such as accidents and bad weather. These can create severe congestion and reduced travel time reliability. Policies for addressing such events are not considered here though, with the congestion pricing, such events will have reduced impact anyway.
A ‘first-best’ partial equilibrium efficiency-based case for comprehensive road pricing obtains provided costs of pricing do not exceed the inefficiency costs avoided. Two complications limit the applicability of this prescription. First, significant implementation costs may mean that comprehensive pricing is impractical so only pricing of a subset of roads – including perhaps a cordon around a city’s CBD – is feasible. This leads to a policy problem where design is conditioned by second-best constraints. Second, general equilibrium impacts of pricing may trigger by-product distortions in such areas as labour markets. The ‘double dividend’ question examines whether such by-product costs can be offset by using congestion tax revenues to cut other distortionary taxes.
If congestion pricing is restricted to a city cordon or to major arterial or ring roads, the gains from congestion pricing can only be realised with auxiliary policies to address second-best constraints. Even with such ‘side’ policies, gains are reduced. For example, with a pricing cordon, congestion that develops on a cordon boundary needs to be addressed through parking restrictions and other policies. Pricing only major roads can lead to ‘rat-running’ along alternative unpriced routes that create urban disamenities. Tolls must be set lower with partial compared to comprehensive pricing so fewer inefficiency costs are addressed (Clarke, 2008).
An important issue in Australian cities is that congestion arises on roads very distant from the city CBD. This is difficult to address since if comprehensive pricing is impractical severe ‘second-best’ issues arise with partial pricing. It is difficult to provide public transport infrastructure that encourages sought modal shifts since traffic densities are often low. Clarke and Hawkins (2006) argue for liberalising provision of bus and mini bus services as a palliative but there remain difficult issues of transport planning in the periphery of Australia’s large, low-average density cities.
In terms of partial approaches to congestion pricing much attention has focused on learning from the London ‘area pricing scheme’. Set-up and operational costs of this scheme were higher than expected but this scheme is now seen as a success (Leape, 2007). The London scheme might be seen as an intermediate policy in a longer-term move to full-scale electronic pricing of all roads in the UK. However the success of the London scheme suggests that such an extension may be uneconomic since congestion pricing of London accounts for 80 per cent of the benefits from a national scheme (DOT, 2004). In fact, on interurban roads, Newbery (2005) estimates that national distance-based pricing would be uneconomic and that a better broad scheme is to rely on existing fuel taxes which adequately capture congestion costs. This is an important insight for Australia.
Finally, the time profile of phasing in reform programs needs to be addressed. Three stylised classes of policy sequencing are plausible. First, adopting piecemeal reforms (cordons, pricing of major roads) that are followed by a switch in the future to comprehensive electronic pricing that includes broad assessment of various externalities. Second, maintaining piecemeal reforms for a period into the future or even indefinitely. Third, there is the possibility of avoiding piecemeal reforms and simply switching directly to lower cost comprehensive electronic pricing schemes in the future when comprehensive approaches become more technologically viable and commercially feasible.
The cost of a program of piecemeal reform is high as evidenced by the cost of the London scheme and the effectiveness of selective pricing schemes is lower than with comprehensive pricing. Thus there are arguments for delaying moves to act until comprehensive pricing can be introduced. The difficulty with this argument is that offering politicians the chance to procrastinate in dealing with severe emerging congestion problems is something those seeking better demand management of Australian urban roads are reluctant to endorse. Australian governments have been slow to recognise the emerging costs of traffic congestion so that advocating delays in charging might have the perverse effects of further weakening already feeble political spines.
It is not sensible policy to avoid introducing any road demand management reforms in the lead-up to comprehensive pricing. At least some roads and access to CBD areas should be priced. Piecemeal reforms provide relief from congestion and information about likely traffic responses to pricing. Case studies of current commercial pricing schemes can add to this information base.
Most importantly attempts to introduce congestion pricing in Australia must be premised on gaining community support first. An active debate on congestion pricing that elucidates the main ideas and corrects misperceptions needs to begin. Essential to this debate is imparting the knowledge that only congested road travel will be priced, that marginal not large adjustments in road travel behaviour are sought and that potentially because efficiency gains are yielded by such pricing that everyone can be made better-off. Most importantly the idea of levying congestion tolls to limit social costs not to raise revenues must be emphasised.
Finally, in advocating a case for congestion pricing, the general equilibrium (GE) arguments of Goulder and Williams (2003), discussed in Section 3, need consideration. Imposing a congestion tax will reduce deadweight losses within markets for road transport but can impose them elsewhere on ‘non-environmental’ welfare. Particular concern has been addressed to effects on labour markets. Secondary impacts of congestion pricing on tax-distorted fuel markets are likely to be small in Australia given low fuel price elasticities.
Clearly the direction of any tax relief provided on the basis of congestion tax revenues will influence judgements about overall efficiency gains. A simple approach to selecting a tax base on which to focus cuts would be to select that base where current marginal excess burdens are highest. These will lie outside the transport sector and involve taxes on such things as labour. Parry and Bento (2001) have shown that channelling revenues from a congestion tax into reducing labour taxes provides net benefits. Such switches however are likely to be politically unpopular if congestion taxes are seen as local taxes and the support of local residents must be achieved to make the switch to pricing feasible. It is plausible to suppose such support will be greater if revenues are used to improve local public transport. This is a also a politically attractive since those ‘tolled-off’ roads will see improved mass transport provision as compensation. If there are external benefits from funding extra public transport then the optimal congestion tax is higher than it otherwise would be.
An alternative option is to hypothecate congestion charge revenues to contribute towards the costs of road supply as is discussed in Section 5. Attention must be paid to how the proceeds of congestion charging are used as well as to how congestion pricing schemes are designed to work.
Parking. Many vehicle journeys involve the need to park a vehicle. There are two separate tax and charge policy concerns: (i) To set parking charges efficiently so that parking markets clear and parking spot search times are optimised; (ii) to use parking policies to provide a ‘second best’ surrogate for congestion pricing. The first issue has not been addressed in Australia while the latter has in our most congested cities, namely Sydney and Melbourne.
Parking charges should be set to reflect peak load demands so that, on average, search costs for parking spots are close to zero. Traffic engineers recommend that keeping 15 per cent of parking places vacant to insure easy ingress and egress out of spots will ensure that parking markets clear. Setting time-varying charges using appropriate metering technology to achieve this vacancy rate avoids wasteful cruising (Shoup, 2005). Cruising to search for a parking spot is itself an independent cause of traffic congestion. Shoup (2005) estimates that 30 per cent of traffic in 11 large US cities is caused by cruising for parking with the average motorist taking 8 minutes to find a spot. Motorists have high willingness to pay to avoid such costs – in Sydney they will pay up to 3.5 times their wage to avoid such searches (Henscher & King, 2001). Subsidised on-street parking creates particular congestion externalities in Australian cities through its effects in encouraging searches by motorists to search for cheap on-street parking spots.
This is again a tax/charge reform issue since it suggests public on-street parking should be priced in parity with off-street parking and that taxes should be levied on the latter so city-wide search costs for parking are close to zero.
Separately from such arguments tax-augmented parking charges can act as a second-best surrogate for congestion charges by governments that are reluctant to price congestion. This is not first-best pricing since such charges only influence traffic terminating costs at a destination. To the extent they reduce such traffic there are increased incentives for increased ‘through’ and ‘drop-off’ traffic which will experience less congestion. In addition, such charges capture only congestion costs in the neighbourhood of a destination. Different vehicles will have taken different journeys to reach this destination and hence will have inflicted different congestion costs en route.
Ideally, congestion should be priced efficiently and then parking priced to clear parking markets. Introducing both forms of efficient pricing will reduce the charges that would need to be applied using only one approach. But Shoup (2005, p. 215-217) provides compelling evidence that, in some cities at least, correctly determined parking charges alone can provide many benefits that jointly efficient congestion-cum-parking policies yield.
Attempts to adjust parking charges as a second-best CBD congestion policy are popular in Australia. The Victorian Government, in 2006, introduced an $820 per parking spot levy on long-stay car parking spaces in Melbourne’s CBD to ease congestion and to reduce greenhouse gas emissions. In Sydney, a Parking Space Levy (PSL) has operated since 1992 to discourage car use. The PSL is hypothecated to fund infrastructure projects that make it easier to access public transport. These include ‘kiss and ride’ facilities that allow temporary parking for those dropping off or picking up travellers, park-and-ride facilities, bus shelters, taxi stands, transport mode interchanges, passenger information and security services.
In Australian cities levels of parking are also regulated by controlling on- and off-street parking. The trend has been to heavily restrict on-street parking but to keep it priced cheaply. The resulting excess demands are rationed primarily by restrictions on the time a spot can be occupied, the type of vehicle that can be parked and by fines on over-stayers and illegal parking. Private off-street parking has encouraged as a high-priced alternative to on-street parking. A policy objective is to discourage longer-term parking which helps retailers optimise sales but which has perverse effect of increasing congestion.
At the same time as inner city parking costs are being increased in Melbourne’s CBD additional free parking spots are being provided in outer Melbourne on major train corridors into Melbourne City. This encourages ‘park-and-ride’ commutes as a ‘second-best’ attempt to deal with congestion without pricing it directly. The same policy is being pursued in Sydney under the umbrella of ‘transport interchange policy’.
These measures reduce congestion by encouraging a modal shift to train use but at considerable cost. Land near railway stations has a high opportunity cost and multi-level ‘structured’ parking stations involve substantial construction costs and will only be effective when land values are high. The cost of providing a parking spot at a train station in Melbourne has been estimated by Green advocates at up to $17,000. If anything the cost of providing a parking place near a Sydney railway station would exceed that in Melbourne given higher land values there. This is an expensive way of switching passengers from car to rail to reduce road congestion (Houston and Perkins, 2008).
Parking spots are costly to provide though many ‘park-and-ride’ and other schemes leave parking unpriced. Since private benefits are being delivered they should be recouped via metered charges. Otherwise land resources may be inappropriately converted into parking spaces and cost ineffective policies employed to address congestion.
In Australia free parking provided to an employee is subject to fringe benefits tax of 46.5 per cent so that the ‘free parking’ distortions that operate in countries such as the US do not operate here.
It is finally worth emphasising that GPS and other technologies can further reduce parking search costs. In San Francisco vacancies at 6,000 of 24,000 metered on street parking spots and 11,500 off-street car parks are displayed on street signs or on maps on screens of mobile phones (Ensha, 2008). Drivers can pay for parking using their phone and can top up the parking meter remotely using their phones without returning to their car. This trial seeks to find an alternative to congestion pricing of roads although a modification, whereby congestion is also priced, is of clear interest.
Such technologies could be used in Melbourne and Sydney as an accompaniment to congestion pricing or as a substitute for it. Such schemes would reduce parking search costs and congestion. They also operate as a ‘sweetener’ in seeking community approval for using telematic devices in vehicles to effect congestion pricing. These devices could become not only a way to provide congestion pricing but also as an aid to drivers who seek to park without incurring search costs.
Traffic Accidents and Insurance Reforms. Traffic accidents generate private costs through deaths, injuries and property damage. In addition, when a vehicle joins a traffic stream, accident probabilities of other motorists increase creating an externality. As the possibility of collisions increases, traffic will slow as densities and driver care increase, thus reducing rates of serious traffic accidents and deaths. Empirical evidence on the net accident effects of increased density needs to be understood. Opposing cost effects (more accidents, less serious accidents) are sometimes assumed to net out to zero so that accident externalities can be ignored. As studies confirm that net accident externality costs are 52-74 per cent of private costs of accidents (Small and Verhoef, 2007, p. 102) this is poor practise. In fact a key empirical regularity is the strong, net positive dependence of not only accident frequencies but also aggregate accident costs on traffic densities (e.g. Litman, 2008).
Apart from density-related causes there are other accident externalities. A pedestrian struck by a vehicle is an external cost. In addition, those driving under the influence of alcohol have much higher external accident risks providing a justification for high penalties on drink-driving (Levitt and Porter, 2001). Finally, motorists who drive ‘light trucks’ (vans, pickups or four-wheel drive utilities) raise the external costs of accidents while improving the safety of their vehicle occupants (White, 2004). These issues nuance the way accident externalities must be addressed using corrective taxes and fines.
If negative traffic accident externalities are unpriced individuals will undertake a socially excessive number of journeys and generate a socially excessive number of accidents.
There are no reliable Australian estimates of accident externality costs. Estimates made are simply assumed to be some arbitrary proportion of total accident costs – often 10-50 per cent (Martin, 2005). These costs are subject to further controversy because of difficulties in valuing a human life which turns out to be a crucial issue in assessing accident costs. This lack of evidence inhibits informed Australian policy discussion since inferences must be drawn using international data that, with different traffic densities and road conditions, may not reflect local conditions. This evidence failure is serious because there are reasons for believing accident externality costs are very large. The discussion to follow is largely contingent on use of US evidence.
An approach to pricing traffic accident externalities is to set a fixed charge (e.g. via car registration) reflecting person-specific expected accident costs and to add to this a variable charge depending on how far a motorist travels and their personal characteristics. Accident frequency is linked to driver age and gender (young hormonal males have high accident rates), to alcohol and drug consumption, use of mobile phones and other distracting devices, and to influences such as the weather.
The bulk of vehicle crashes are multi-vehicle – in the US over 70 per cent (Edlin and Karaca-Mandic, 2006, p. 933) (hereafter EM). EM argue that, when two cars crash, if the average damage to a single vehicle is D the social damage is 1.7D. Motorists pay only the average not the marginal cost, creating an externality. They pay too little for driving and drive too far thereby causing ‘too many’ accidents.
EM estimate US external costs using panel data on state average insurance premiums and loss costs which capture effects of reduced accident severity in denser traffic. External costs are substantial in traffic-dense states and negligible in low density states. In traffic-dense California, an increase in density due to a single extra driver raises state-wide insurance costs by $1725-$3239 depending on car model. This added $744 to the insurance premiums an average driver paid in 1996. A Pigovian tax to internalize this externality would raise $44 billion annually – more than all US state taxes combined! A national corrective tax would raise $113 billion (Edlin & Karaca-Mandic, 2007 correct a numerical error in EM). These cost estimates understate true costs since damages are ignored where parties are uninsured. If uninsured costs behave as insured costs then EM estimate accident externalities could be 3.5 times those estimated or $10,000 per motorist annually in California.
While Pigovian tax bases (per mile, per driver or per litre of fuel) might deal with this externality EM’s first-best solution is to tax car insurance premiums and to force motorists to be fully insured. Implementing this in California would require a 200-400 per cent levy on insurance premiums a solution EM recognizes is probably infeasible. In addition, without an effective mechanism ensuring that motorists fully insure, such a premium encourages drivers not to insure. A second-best compromise is to leave overall insurance costs the same but to increase marginal costs by linking premiums to distances travelled. Drivers then have incentives to cut insurance costs by driving less which is the sought outcome. The charges can be driver-specific with high per km premiums on high accident risk drivers who live in high traffic density areas. Edlin (2003) estimates US gains from such policy reforms at $12.7b per year.
A difficulty is that the chance of having an accident depends on when travel occurs. If driving at certain times means that few cars are encountered, accident probabilities involving other vehicles are low. Ideally one should congestion price and add to this toll a charge reflecting the contemporaneous accident externality. Such a charge would be higher in peak periods than a pure congestion toll. This policy refinement might become feasible with developments in telematic technology.
Parry et al. (2007) criticise EM because the externality evidence is based on insurance data which mainly covers property damage. Damage costs are only 14 per cent of what Parry et al. (2007) see as the total social costs of accidents. EM only provide inconclusive evidence on whether fatal traffic accidents – which account for 33 per cent of the Parry et al. estimated social costs – increase with density. Parry et al. estimate marginal external costs inclusive of pedestrian and cyclist injuries at 1.2-4.8 cents US per kilometre for the year 2000 which is still very high. It is 13-44 per cent of average US accident costs. These figures provide a rough clue on the likely scale of Australian costs – the Parry per km figures are between 20-80% of the congestion costs in Australian capital cities.
Alternatively the 200-400 per cent premium figure provided by EM provides a method of estimating accident externality costs in Australia ignoring the issue of uninsured motorists and time delay costs of accidents. Accepting the Connelly and Supangan (2006) estimate of the costs of Australian accident costs as 2.3 per cent of GDP this suggests a tax on insurance costs of 4.6-6.9 per cent of GDP, an enormous figure. This is an approximate figure which ignores reductions in driving as a consequence of the tax. There may also be less high-density driving in Australia and hence fewer multi-vehicle accidents.
The proposal for distance-related insurance charges is part of a package of reforms private insurance companies might seek, or be induced to seek. Such charges are more effective than increased excises on fuels since the primary behavioural response involves less driving not a fuel economy switch. Litman (2008) estimates the desired charge as around 6 cents US per mile for the average US motorist or 7.5 cents Australian per kilometre. This figure is strikingly large and of the same order of magnitude as unit congestion costs in Australia’s major cities. Traffic accident externalities it would seem cannot be ignored as a source of unpriced social costs. The actual charges levied would depend upon the motorist and the vehicle relative risk factor.
There are various other pricing options such as using self-declarations backed by penalties for false declarations, using estimated mileage as a rating factor, paying a surcharge at the petrol pump to fund basic vehicle insurance, prepaying for a certain mileage cover, per minute premiums and using GPS-based pricing. There are implementation costs of such schemes although all satisfy basic cost-benefit tests even if estimated transaction costs of inspection per vehicle are $10-$150 annually (Litman, 2008). Private insurance firms have adopted specific schemes already and can be expected to increasingly offer such products as technologies for monitoring vehicle mileage improve.
EM argue this solution must be regulated since individual insurers may not adopt per km policies since gains accrue to other insurers while monitoring costs are internal. However, as Litman (2008) shows, some firms are already developing such policies. There is limited endorsement of them by firms in Australia by who now advertise such policies.
The possibility of linking third-party insurance with distances travelled by imposing a second-best compulsory excise levy on petrol has the advantage of simplicity and helps bring the uninsured into the payment system. A disadvantage is that it fails to reflect driver heterogeneity. In addition fuel-efficient vehicles pay lower charges even if they impose the same accident costs. Fuel surcharges are most effective in targeting reduced petrol consumption while pay-as-you-drive insurance better targets distance travelled. Parry (2005) shows that distance-based charges outperform excises on fuels once congestion, local accident and accident reduction benefits are accounted for.
Per kilometre charges could be imposed in states such as Victoria by modifying the Transport Accident Charge (TAC) levied as a compulsory component of the annual vehicle registration charge and which covers third party personal damages. The registration charge could be based on a per km charge that is driver and vehicle specific. The motorist could select scope of cover and insurance company to cover additional non-third party damages. Consumer objections to such schemes could be eliminated by offering motorists the chance to stick with a high fixed charge or utilise per km charging. For example the standard registration fee for a private motor vehicle in Victoria is $593.80 including a TAC of $378. If the per distance accident charge to cover third party personal damages was 6 cents per km for the average motorist driving a typical motor vehicle this fee would cover 10,080 km annually. Higher risk individuals would receive this cover by paying a higher fee as would motorists seeking to drive further. If motorists reacted to the higher marginal cost of travel by cutting back distances travelled by 10 per cent they would drive 1008 kilometres less and traffic accident costs, using estimates of Litman (2008), would fall by 14-18 per cent. These are huge savings and they would be accompanied by reduced traffic congestion and local pollution.
Since this insurance covers only third party person accident costs the appropriate per km charge might be less than 6 cents. Other forms of insurance such as comprehensive policies covering personal accident cover and property damage could be taken out by a public or private insurer.
The AFTSR (2010) report did not take up the suggestion by CP for moving towards distance-related insurance policies. It did however argue for a more careful monitoring of personal characteristics in devising premiums and did mention distance-related insurance charges being adopted in the private sector. This is a useful reform but does not achieve effective user charging for traffic accidents because accident probabilities depend not only on driver identity but also on the extent of exposure to accidents as measured by distance driven.
Greenhouse gas emissions. Greenhouse emission costs can be addressed via petrol excises although, at carbon charges of around $20/t, they are negligible at around 5 cents per lire. With low fuel price elasticities such charges will have close to zero impact. Since greenhouse gas emissions arise in many sectors of the economy an economy-wide policy, such as an emissions trading scheme (ETS), to address them makes sense. Consideration of such a scheme has been postponed in Australia until 2012 so that current fuel excises can be interpreted as capturing such costs.
Other substance pollutions and noise/vibration externalities. Gasoline-using vehicles emit carbon monoxide, nitrogen oxides and volatile compounds. These compounds have adverse health consequences for humans – the damages are linked to location, distance travelled, fuel consumed and emission performance. Recent epidemiological studies suggest greater sensitivity of human health to such pollutants than previously believed. Litman (2003) argues such pollutants cause as many deaths as motor vehicle accidents although, because these deaths are concentrated among older people, years of life lost are lower.
Total estimated Australian pollution heath costs due to car travel are 36 per cent of total congestion costs in 2000 BTRE (2007, p.109). If this same percentage held in 2005 then, taking the per kilometre congestion cost as 6.8 cents, the cost of air pollution externalities in Australia in 2005 would be 2.4 cents per kilometre then. This is not a negligible cost. Policies for addressing it include taxes on fuels and incentives/regulations to improve emission performance.
The costs of travel-related to noise and vibration have long been discussed by road engineers. Noise has detrimental effects on human health that are proportional to the square of distance from the source. Hence concern in regulating noise is for land uses close to highways. Even here however damages are difficult to assess since homes and businesses can be insulated to protect against noise or retrofitted with such insulation. The costs of noise have been estimated by looking at its effects on property values. Delucchi and Hsu (1998), for example, estimate costs of 0-0.4 US cents per mile for passenger vehicles.
While these costs are low compared to other externalities they are high in specific areas. Given locational specificity they are best addressed by regulation not taxes and charges. Regulation is necessary since although quiet engines are not difficult to design, car makers have low incentives to do so given that individual vehicles generate only a small fraction of noise costs.
Miscellaneous external costs. Sedentary lifestyles are claimed to lead to a broader range of health costs although many such costs (e.g. obesity) are privately borne and hence are not externalities. Such costs are not analysed further.
Others draw attention to externalities associated with urban sprawl which is claimed to arise because road travel is underpriced. This is not a clear-cut unpriced social cost if the inefficiency here lies in failure to price road use since this suggests that it is road use not sprawl that should be priced. Moreover, as Bruegmann (2005) points out, preferences for living in large homes on a city’s periphery can reflect informed market choices. There can be distortion-based preferences for such housing if tax concessions are provided towards capital gains on housing or if the external and infrastructure costs of new development are underpriced but these are features separate from transport issues and need to be addressed separately.
Finally, links have been drawn between agglomeration economies and urban traffic congestion. Arnott et al (2005) point out that there are agglomeration benefits in business and personal life that accrue from living in concentrated cities. Such benefits provide a rationale for clustering of activity. These benefits are difficult to quantify but suggest an offset to the social costs of congestion. The upshot of this view is that congestion charges might be set below marginal congestion costs to encourage interactions. This however ignores the fact that while agglomeration externalities result in land use being insufficiently concentrated so too do congestion externalities. . Without data on agglomeration economies and a clear theory of how such economies interact with congestion conclusions are not clear. Potentially such data could change the way congestion modelling is interpreted.
5. Road Damage and Road Supply Hypothecation issues. Public spending on roads is a major public expenditure item which is split between various levels of Australian government. While revenues collected by road use taxes and charges more than balance road costs (PC, 2007), the charges levied reflect average not specific user costs and hence there are inefficiencies. A major variable cost associated with production of road services is that of repairing road wear caused by heavy vehicles. User charges reflecting actual costs of road use provide better usage decisions than do increased registration charges which recoup costs on average. Indeed there is an efficiency dividend from user charging of road damage costs that can be returned to truckers as a lower average charge. This section examines user charges and proposals to utilise them in a commercialised road supply sector.
Australia has the largest and heaviest road-legal vehicles in the world. These trucks are used for transporting livestock, fuel, mineral ores and general freight. They deliver cost-effective transport which links remote areas to population centres and ports.
Large trucks impose almost all road damage costs because the damage done to a road is proportional to the Nth power of a vehicle’s axle weight where N > 3 (Bridle and Porter, 2002). The precise value of N depends on truck design and road durability. It can be large in Australia where roads have low durability reflecting their low construction durability investment. Generally roads should be designed with a forecast maintenance schedule in mind. A road can be very durable with high investment in initial thickness but this is expensive and, for low utilisation routes, unwarranted. Similarly there can be underinvestment in roads used by heavy vehicles resulting in excessively large maintenance costs. There are important tradeoffs between durability and maintenance costs that should be efficiently resolved.
Australian road supply decisions are often not related to demands but are often instead engineering-based and politically-driven. Proposed reforms relate road supply decisions to forecast demands based on user charges – particularly charges related to the road damages. Then charge revenues themselves – or forecasts of such – could be used to signal desired road capacity and durability. Indeed it is known that, under certain conditions, correctly levied user charges will generate revenues that recover the capital and maintenance costs of roads. They also signal when roads are correctly priced and when roads should be expanded or upgraded.
A further reform proposal is that such decisions should be undertaken on a commercial basis so roads should be managed to optimise their value as net revenue-yielding assets. However commercialisation of roads is not as simple as commercialising public utilities such as power generation. Attention must be directed at community service obligation roles of roads outside urban areas, indivisibilities in supplying roads of minimum capacity and to the network character of roads.
Economics typically opposes hypothecation arguments that favour assigning taxes and charges derived from road transport to road capital and maintenance expenditures. There are however supply-side efficiency reasons for doing this. Most importantly road capacities and durabilities need to be related to the benefits roads can be anticipated to deliver. There should be incentives to deliver roads of appropriate capacity and durability given projected road demands. One signal for this is the aggregated user charges roads can be expected to generate compared to their provision costs. This shifts the focus of road planning away from optimisation subject to a budget towards the optimisation of the net benefits from roads.
Road pricing and investment. In competitive markets producers adjust outputs so short-run marginal costs equal market price. Total revenues cover total costs and provide a rent to cover costs of fixed factors. If rents exceed costs of providing the fixed factors then entrepreneurs have incentives to acquire more fixed inputs. If enough entrepreneurs do this prices will fall until rents just cover fixed input costs. The market is then in equilibrium and no firm has incentives to expand.
A optimal road network can be described analogously. Newbery (1989), drawing on seminal work of Mohring and Harwitz (1962), considered congestion and road damage costs in a road investment and pricing setting where designers can invest in durability to reduce the damages heavy vehicles cause on roads. With constant returns to scale in road construction for roads of given durability and with strictly constant returns to road use – so heavy vehicles distribute themselves uniformly over a road’s width – Newbery shows that optimal road user charges – congestion plus road damage – will recover all maintenance and interest on capital costs irrespective of economies of scale in road construction. There is exact self-financing.
The result provides a necessary condition for optimal capacity and pricing so that, whether it is met or not, provides a check on road supply efficiency. It is not a sufficient condition since there are other combinations of tolls and road capacities that cost-recover but only one is optimal. The result also shows that taxes are not necessary to support a road system since it can self-fund if appropriately designed and priced. It also encourages public acceptance of road pricing since it is consistent with ‘user pays’ and the cost recovery process is transparent and observable. Finally charge revenues provide a measure of when a road should be expanded. Small and Verhoef (2007, p. 166-170) discuss the result in detail and include various complicating issues – discreteness in capacity, bottleneck congestion, network issues and user heterogeneity.
Note the constant returns requirement in use. Even with constant returns in construction heavy vehicles spend a higher fraction of their time in slow moving lanes. Then optimal road user charges will recover more than total road costs. Newbery shows that if heavy vehicles cause all road damage and only use the slow lane, the charges will recover all overheads and interest on capital plus exactly twice the total costs of maintenance.
These self-financing results also break down if roads of minimum scale (‘indivisibilities’) must be provided to meet community service obligations (CSOs) or if roads are sub-optimally priced. Then interest centres on the case where no toll is charged so self-financing cannot occur since revenues are zero. Optimal road capacity without pricing must be higher than with pricing so having efficient pricing in place saves capital resources.
With higher investment in road durability roads last longer and need costly resurfacing less frequently. On high-volume interstate highways in the US Small and Winston (1988) show that net savings from reduced resurfacings save 40 per cent of resurfacing costs. Indeed, with optimal durability the trucking industry claim that on high-volume roads trucks would not damage pavements were they designed optimally, is confirmed. Trucks should still pay most damage costs since the need for enhanced durability is attributable to them, though payments are for capital not maintenance and paid through fixed charges. Indeed, there is a strengthened role for fixed license and registration charges with such optimal durabilities. If optimal durability is low – as on parts of the Australian network – variable user charges will still have an important role in cost recovery.
Road supply reforms. In Australia policies governing charging for heavy vehicle use of roads have been part of a continuing reform process since 1991 when the National Road Transport Commission (NTC) was set up by the Commonwealth to implement a reform program for heavy vehicles. Following a request from the Council of Australian Governments (COAG) the Productivity Commission (2007) (PC) presented a report which dealt with road and rail freight pricing in Australia. This has led to recent rounds of proposed pricing reforms. Although set up partly to address competitive neutrality concerns arising because rail services alone pay on a commercial basis for their infrastructure, the PC determined that neutrality concerns were insignificant. Indeed road and rail were seen more as complements than substitutes which meant that, to a substantial degree, their efficient provision could be analysed separately. There were however inefficiencies in road pricing that needed to be resolved.
The PC argued that heavy vehicles were paying the road damage and access costs that could reasonably be attributed to them so cost-recovery was not an issue. Moreover, with recent NTC reforms, cross-subsidies between vehicles were subsequently eliminated – the levies imposed are derived on a pay-as-you-go with costs recouped via registration charges and a fuel excise that reflected their weight on the basis of historical costs imposed. The common levies across Australia are imposed only on heavy vehicles – for light vehicles charges vary by jurisdiction. These heavy vehicles charges are averaged over different roads and hence do not reflect actual costs. What are needed are mass-distance-location charges.
One specific efficiency-promoting reform has been the proposal to admit heavy vehicles to less durable roads provided they pay an ‘incremental price’ reflecting additional damage to the road caused by incremental mass above current maximum weight limits. This reform employs the telematic GPS technology currently used by trucking companies to monitor vehicle movements for fleet management. This makes it possible to charge on the basis of time, distance and location of travel although currently not on the basis of loaded weight. In fact telematic technologies are now available which will enable road managers to price road use on loaded ‘axle-load’ weight. These technologies can now capture all externalities generated by both light and heavy transport usage of roads of varying durability and which are subject to differential substance and congestion-based externality costs. With an adequate data base determining how road damages can be related to types of heavy vehicle use, pricing of vehicle-induced road damages on a mass-distance-location basis is feasible now.
What remains to be determined is how such revenues should be disbursed to road construction authorities in different jurisdictions to facilitate efficient road supply responses in terms of capacity and durability characteristics.
COAG and Telematics. Communication devices in vehicles can change and improve the way we travel. In-vehicle (telematic) units (IVU) units can be installed to meet various regulatory and commercial needs. They can provide information about journeys to external agencies such as regulators or to vehicle-fleet managers and receive information from such agencies in a form accessible to drivers. For regulation IVU devices can be used for congestion pricing, for mass-distance-location pricing of road damage costs, for charging for vehicle pollution emissions, for tracking delivery of dangerous goods and for monitoring driver-specific information on rest breaks and distances travelled to ensure road safety standards. There are also commercial applications for tracking/tracing vehicles, for pay-as-you drive insurance, as anti-theft devices, for vehicle-to-vehicle communications and navigation services and for providing emergency help in the event of an accident. They could be used to provide information about and to manage parking in a congested city.
IVU are being used in Australia under the Intelligent Access Program to manage access of vehicles onto the Higher Mass Limits (HML) roads in NSW and Queensland and there has been an application to heavy mobile cranes in Victoria. The IAP is a voluntary program providing heavy vehicles with access, or improved access, to roads in return for monitoring of compliance with specific access conditions using vehicle telematic solutions. Experience with regulatory applications is developing rapidly in Australia. The devices are necessary if the ‘incremental pricing’ reform discussed by COAG is to be implemented. This scheme permits heavy vehicles to drive on less durable roads – from which they are currently banned – by paying an appropriate fee. Incremental pricing schemes are to some extent being displaced by the IAP.
A crucial issue in Australia is that many trucks shift heavy but variable freight loads. Many journeys by heavy vehicles in Australia involve taking mineral products on one-way journeys to ports so that, almost 50 per cent of the time, such large vehicles are travelling without load. Possibilities for triangulating freight tasks are low. In Australia a vehicle of given size can therefore carry widely differing tare loads. This is a distinctive feature of Australian compared to European conditions and means that inferences about the mass of a load cannot be readily inferred from the size of a vehicle. Self declarations are subject to deliberate and inadvertent error but can be monitored using existing weighing stations and inspection procedures. Alternatively tare weight can be assessed directly using telematics. Across a variety of truck types total costs ranged from $3,200-$13,500 per vehicle with installation costs varying between $200-$2500 (TCA, 2009, p. 43-44).
COAG is committed to a Road Reform Plan designed to improve the effectiveness of Australian road planning. The intent is to link revenues from road use with better decision-making procedures by road providers.
Ideally road providers need to design roads which recover all costs together with a competitive return on capital by charging road users the capital costs of the roads used as well as maintenance, congestion and operational costs incurred. If providers know that this will happen they can draw on local and international capital markets to supply desired road funding for efficient road designs. Public budgets then do not limit the achievement of design optimality.
This proposal is analogous to commercial ‘road fund’ proposals (Gwilliam and Shaizi, 1999). Essential national or regional road agencies are set up as regulated monopolistic public utilities. These institutions derive their revenues from road user charges and would be responsible for maintenance and capital investments in their segment of the market whether it was national highways or regional roads. Regulation would be essential both to restrict the exercise of monopoly power and to encourage the development of projects with high social value that might require cross subsidies or direct funding by government. The experience of other countries can be drawn on in coming up with specific road agency designs. An instance of such arrangements is Transit New Zealand (subsequently the NZ Transport Agency) (NZTA, 2009) which manages New Zealand’s roads.
The issue of selling such technological fixes to politicians and ultimately, via politicians, to road users has not been discussed. Ignoring cross subsidies the attractiveness of all such schemes is that they potentially reduce costs of road usage by delivering efficiency gains. Incremental pricing schemes make this transparent since they offer trucking operators additional freedom of choice to select road transport options at a price. Local government will be particularly attracted to a fully priced IAP since they can then cost-effectively resolve ‘last mile’ problems implicit in providing point-to-point freight deliveries.
The AFTSR (2010) report did not go as far as others have in seeking to tie supply decisions to profitability once optimal user charges are in place. The suggested start of using strategic planning buttressed by ‘hard-headed’ cost-benefit analysis is however useful. There are however complex issues involved in applying cost-benefit analysis to assets that are part of a network, which deliver CSOs and which are subject to indivisibilities in supply. According to AFTSR supply decisions should optimise road asset values but these would be made on the basis of time savings not user charges and current capital as well as road maintenance costs. Leaving out user charges means that supply decisions are distorted since drivers are not making their decisions to travel on the basis of the actual costs they impose.
6. Conclusions. The key tax reform issue in the Australian road transport sector is whether governments should move away from collecting revenues via fixed charges (such as vehicle registration) and variable charges (such as taxes based primarily on fuels) toward a system based on user charges that account directly for the specific costs that road users impose.
A case for excises on fuel can be based purely on their revenue gaining potential independently of their role in muting externalities. Given that price elasticities of demand for transport fuels are low and the likelihood that leisure and fuels are complementary items in consumer utilities, the case for excise tax differentiation is strong.
The revenue case for excises on fuel is strengthened by their low collection costs and by the limited opportunities agents have to waste resources evading them. Independently too, even disregarding revenue-gathering objectives, there is a case for a fuel excise which exceeds the current excise simply as a proxy for non-existent user charges. The fixed capital costs of using roads can also be collected using fuel use as a surrogate for road use. A specific case for a portion of the fuel excise can be derived from its role in capturing greenhouse gas emission costs.
The other major externalities (traffic accident, road damages and congestion) are only imperfectly linked to fuel usage.
Traffic accident externalities increase under congested conditions so, if fuel excises reduce driving, they will prove helpful in reducing traffic accident externalities. However these externalities can be better addressed using direct distance-based insurance charges since fuel use can be reduced by investing in fuel efficient cars rather than driving less. Moreover accident probabilities reflect characteristics of the driver as well as their location. These factors can be incorporated into personalised two-part insurance tariffs that reflect distance travelled and personal characteristics of the driver. Taxes can be appended by government to such tariffs or to the fixed component of registration charges on order to capture residual external costs. There is a role for government to promote distance-based insurance given that external benefits flow from imposing them.
Another significant source of external cost is vehicle-caused road damage. Fuel use reflects some road damage costs since heavy trucks that impose most damage do consume much fuel. Road damage costs however depend on the character of roads driven on, distances driven and vehicle mass. The technology for measuring the mass- distance-location of travel is now viable so road damage costs can be accurately monitored at relatively low cost.
As a prelude to such an innovation, ‘incremental pricing’ will confirm the efficiency gains that accrue to trucking as well as to public budgets from a more direct approach to charging for road damage costs. Pricing here relaxes usage constraints on road users. An efficiency dividend from better pricing of freight can be returned to heavy vehicle operators. The core difficulty of assessing the loaded mass of a vehicle can be addressed either by using advances in telematics or by using driver self-declarations of mass and monitoring such declarations using pre-existing inspection procedures.
There are strong arguments for addressing the levying specific congestion charges in large cities. This again is also only imperfectly captured by excises on fuels because increases in fuel use that occur under congested conditions do not match congestion costs. The most plausible role for congestion pricing is in the east coast Australian state capital cities, Melbourne and Sydney. There is a strong case for moving toward a comprehensive congestion pricing schema in these cities.
User charges also provide signals that help ensure efficient road supply decisions. Hypothecation arguments are usually unpopular among economists but there are efficiency reasons for considering them here. Road supply decisions should reflect forecast road demands and these are reflected in forecast road user charge revenues. These provide useful signals that guide road construction, expansion and maintenance. Constructing a road provides a community with a long-lived community asset which should be efficiently used, managed and, if necessary, expanded or upgraded through time. Such assets should have capacities and durabilities which provide competitive returns on capital employed, after accounting for maintenance costs, that are comparable to other long-lived public and private sector assets.
The difficulty here is not so much to devise an effective system of charging but to come up with appropriate institutions to forecast and respond to such signals. There is a role for central direction in providing forecasts and for constructing major state roads. The regulatory issues that arise in governance of decentralised service providers to prevent inefficient strategic behaviours and to meet community service obligations are more difficult. One suggestion is for a central authority to declare the revenue reimbursements paid to local governments from allowing various types of heavy vehicle use of their roads but to allow local institutions to make a judgement on the admissibility of such use based on a cost-benefit analysis of incremental revenues versus incremental infrastructure costs. This efficiently resolves the ‘first’ and ‘last mile’ problem.
This paper has also drawn several conclusions about the way fuel excises operate independently of their revenue-gathering characteristics. There are positive features of the current system. As currently implemented fuel excises do not impact on business inputs if large vehicles are used. Moreover, the excise system has become simpler and now includes a broader range of fuels. If hybrid cars become popular, electricity used for transportation may also be considered for inclusion as a fuel. Significantly there is no implication that excises on fuels invariably mean excises on liquid carbon-based fuels alone. A criticism of current excises is that lower rates of excises are proposed for alternative fuels, such as those based on ethanol and LPG, when the environmental grounds for giving ethanol fuels, in particular, lower excise rates are not strong. This will be even truer if an emission trading scheme operates.
This report has also provided views on the efficiency effects of other taxes and charges in the economy. While the scale of inefficiencies associated with low tariffs is the subject of dispute there is no convincing case for retaining them. Similarly while the luxury car tax causes only numerically small excess burdens there is no sound distributional or efficiency-based reason for retaining it. On the various state taxes further study is required. Making charges uniform across states would save transaction costs. It would also be simpler to remove transaction taxes and replace them with larger registration fees. Tax expenditures from application of the current statutory formula for fringe benefits tax create a moral hazard problem resulting in excess driving. However, application of the statutory formula reduces collection and administration costs of this tax. Further study can determine if cost savings are greater than losses from induced excess driving.
Transport is a crucial input into all aspects of the Australian economy. Australia is a large country with a dispersed population distant from many key international centres. Transport in Australia involves large public costs and transport activities are quite heavily taxed. Transport yields a significant range of external costs that impact on all. The more efficient are transport taxes and investments in transport infrastructure the better will be Australia’s productivity and quality of life. Australia will improve its economic advantages if it moves to a more efficient transport sector.
R. Arnott & K. Small, ‘The Economics of Traffic Congestion’, American Scientist 82, 1994, 446–455.
R. Arnott, A. De Palma & R. Lindsay, ‘A Structural Model of Peak-Load Congestion: A Traffic Bottleneck with Elastic Demand’, American Economic Review, March 1993, 161-179.
K. Asprey, J. Lloyd, R. Parsons, & K. Wood, (‘Asprey Report’) Taxation Review Committee, Full Report, Australian Government Publishing Service, Canberra. 1975.
Australia’s Future Tax System Review (AFTSR), Report to the Treasurer, Part 1 and Part 2, 3 volumes, Canberra, 2010.
R. Bridle & J. Porter (eds.) The Motorway Achievement: Frontiers of Knowledge and Practice. Thomas Telford. 2002.
R.V. Breunig & C. Gisz, ‘An Exploration of Australian Petrol Demand: Unobservable Habits, Irreversibility and Some Updated Estimates’, Economic Record, 85, 2009, 73-91.
R. Bruegmann, Sprawl: A Compact History, University of Chicago Press, Chicago 2005.
Bureau of Transport and Regional Economics (BTRE), Estimating Urban Traffic and Congestion Cost Trends for Australian Cities, Working Paper 71, BTRE, Canberra, ACT. 2007.
Bureau of Transport and Regional Economics (BTRE), Health Impacts of Transport Emissions in Australia: Economic Costs, Working Paper 63, BTRE, Canberra, ACT. 2005.
H. Clarke, ‘Targeting Urban Congestion: Equity and Second-Best Issues’, The Australian Economic Review, 41, 2, 2008, 177-186.
H. Clarke & A. Hawkins, ‘Economic Framework for Melbourne Traffic Planning’, Agenda, 13, 2006, 63-90.
H. Clarke & D. Prentice, (CP) ‘A Conceptual Framework for the Reform of Taxes Related to Roads and Transport’, Paper prepared for Australia’s Future Tax System Review, The Treasury, Canberra 2009. http://taxreview.treasury.gov.au/content/html/commissioned_work/downloads/Clarke_and_Prentice.pdf.
L.B. Connelly & R. Supangan, ‘The Economic Costs of Road Traffic Crashes: Australia, States and Territories’, Accident Analysis and Prevention, 38, 6, 2006, 1087-93.
Department of Transport, Feasibility Study of Road Pricing in the UK, London DFT, 2004.
M.A. Delucchi & S. Hu, ‘The External Damage Cost of Noise Emitted from Motor Vehicles’, Journal of Transportation and Statistics, 1, 3, 1998, 1-24.
P. Diamond & J. Mirrlees, ‘Optimal Taxation and Public Production, I: Production Efficiency’, American Economic Review, 61, 1971, 8-27.
P.B. Dixon, ‘Comments on the Productivity Commission’s Modelling of the Economy-Wide Effects of Future Automotive Assistance’, Economic Papers, 28, 1, 2009, 11-18.
A.S. Edlin, ‘Per Mile Premiums for Auto Insurance’ in R. Arnott, B. Greenwald, R. Kanbur & B. Nalebuff (eds), Economics for an Imperfect World: Essays in Honour of Joseph E. Stiglitz, MIT Press, MA, 2003.
A.S. Edlin & P. Karaca-Mandic, ‘The Accident Externality from Driving’, Journal of Political Economy, 114, 5, 2006, 931-955.
A.S. Edlin & P. Karaca-Mandic, ‘Erratum: “The Accident Externality from Driving”’, Journal of Political Economy, 115, 4, 2007, 704-705.
A. Ensha, ‘Can Parking Policy Ease Congestion’, The New York Times, July 1, 2008.
Glass’s Information Services, 2008 Black & White Data Book, Glass’s Information Services, 2008.
L.H. Goulder & R.C. Williams III, ‘The Substantial Bias from Ignoring General Equilibrium Effects in Estimating Excess Burden, and a Practical Solution’, Journal of Political Economy, 111, 4, 2003, 898-927.
D. Henscher & J. King, ‘Parking Demand and Responsiveness to Supply, Pricing and Location in the Sydney Central Business District’, Transportation Research Part A, 35, 2001, 177-196.
C. Houston & M. Perkins, ‘Latest Problem on Train Networks – Car Parks’, The Age, May 27, 2008.
L. Kaplow, The Theory of Taxation and Public Economics, Princeton University Press, Princeton, 2008.
D. Kraal, P.W. Senarath Yapa & D. Harvey, ‘The Impact of Australia’s Fringe Benefits Tax for Cars on Petrol Consumption and Greenhouse Emissions’, Australian Tax Forum, 23, 2008, 192 – 223.
J. Leape, ‘The London Congestion Charge’, Journal of Economic Perspectives, 20, 4, 2007, 157-176.
S.D. Levitt & J. Porter, ‘How Dangerous Are drinking Drivers?’, Journal of Political Economy, 109, 2001, 1198-1237.
T. Litman, ‘Distance-Based Vehicle Insurance as a TDM Strategy’, Victoria Transport Institute, Victoria, BC, Canada, 18 November 2008.
T. Litman, ‘Integrating Public Health Objectives in Transportation Decision Making’, American Journal of Health Promotion, 2003. Preprint at http://www.vtpi.org/AJHP.pdf.
T. Litman, 2008. ‘Transportation Elasticities’, Victoria Transport Policy Institute. Available at http://www.vtpi.org/elasticities.pdf as at March 21, 2009.
L. Martin, ‘External Accident Costs of Motor Vehicles Revisited’, Staff Paper, Department of Transport and Regional Services, 28th Australasian Transport Research Forum, September 2005.
H. Mohring & M. Harwitz, Highway Benefits: An Analytical Framework, Evanston, Illinois, Northwestern University Press, 1962, 70-87.
D. Newbery, ‘Cost Recovery from Optimally Designed Roads’, Economica, 56, 1989, 165-185.
D. Newbery, ‘Road User Charging and Finance’, Taxation Seminar, London School of Economics, December 2005.
New South Wales Treasury, Interstate Comparison of Taxes 2007-08, Office of Financial Management Research & Information Paper, TRP07-2, November 2007.
New Zealand Transport Agency, Statement of Intent 2008-2011, New Zealand Transport Agency, March 2009, http://www.nzta.govt.nz/publications/docs/statement-of-intent.pdf, (Accessed April 2009).
B. Oehry, ‘Critical Success Factors for Implementing Road Charging Systems’, Prepared for the ITF/OECD Round Table of 4-5 February, 2010 on Implementing Congestion Pricing’, Paris 2010.
I. Parry, ‘Comparing Alternative Policies to Reduce Traffic Accidents’, Resources for the Future, 2005.
I. Parry & A. Bento, ‘Revenue Recycling and the Welfare Effects of Road Pricing’, Scandinavian Journal of Economics, 99, 2001, 261-279.
I. Parry, M. Wells & W. Harrington, ‘Automobile Externalities and Policies’, Journal of Economic Literature, 45, 2, 2007, 373-399.
I. Parry & K. Small, ‘Does Britain or the United States Have the Right Gasoline Tax?’, American Economic Review, 95, 4, 2005, 1276 – 1289.
D. Prentice & X. Yin, ‘Constructing a Quality-Adjusted Price Index for a Heterogeneous Oligopoly’, The Manchester School, 72, 2004, 423-442.
Productivity Commission, Modelling Economy-wide Effects of Future Automotive Assistance. Productivity Research Report, May 2008.
Productivity Commission, Road and Rail Freight Infrastructure Pricing, Final Report, Melbourne, April 2007.
M.S. Scollo, M. Younie & M. Wake, J. Freeman and F. Iscasiano. 2003. ‘Impact of Tobacco Tax Reforms on Tobacco Prices and Tobacco Use in Australia’, Tobacco Control, 12 (Suppl II) :ii59-ii66.
D. Shoup, The High Cost of Free Parking, Planners Press, Washington 2005.
K.A. Small & C. Winston, ‘Optimal Highway Durability’, American Economic Review, 78, 3, 1988, 560-569.
K. Small & E. Verhoef, The Economics of Urban Transportation, Routledge, London & New York, 2007.
Transport Certification Australia (TCA), On Board Mass Monitoring Test Report (Draft V2), Draft – March 2009, Transport Certification Australia.
S. West & R.C. Williams, ‘Optimal Taxation and Cross-Price Effects on Labor Supply: Estimates of the Optimal Gas Tax’, Journal of Public Economics, 91, 2007, 593-617.
M. C. White, ‘The ‘Arms Race’ on American Roads: The Effect of Sport Utility Vehicles and Pickup Trucks on Traffic Safety’, Journal of Law and Economics, 47, 2004, 333-355. (2498)