Cars & traffic

May 20, 2012

Should cars be subsidised?

It’s a truth universally acknowledged that private cars are massively subsidised. Indeed, cars are an object lesson in the consequences of under-pricing. Because we don’t tak

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

Costs per passenger-km for cars and public transport in Sydney, 2005/6 (from Glazebrook, 2009)

It’s a truth universally acknowledged that private cars are massively subsidised. Indeed, cars are an object lesson in the consequences of under-pricing.

Because we don’t take account of the social costs of driving when we get behind the wheel, we drive too often, too far and in vehicles that are too large. Our cities sprawl, our roads are choked by congestion and public transport languishes for want of passengers.

All that’s pretty familiar stuff, yet despite the central role under-priced travel plays in shaping our cities, it’s astonishing how difficult it is to find reliable and objective data on the level of subsidy for cars.

The Bureau of Infrastructure, Transport and Regional Economics (BITRE) published figures last year showing expenditure on roads by all levels of government in 2008-09 was $15.8 billion. Revenue from charges and taxes levied on motorists (excluding GST) was $15.6 billion in the same year.

However this doesn’t take account of social costs. The topic appears to be so under-studied that the Victorian Competition and Efficiency Commission (VCEC) identified only one set of numbers in its 2006 report on traffic congestion. Those weren’t prepared by a technical expert but came from a submission made to VCEC by an advocacy group, the Public Transport Users Association (PTUA).

The PTUA says motorists cost the Australian community $42 billion annually in cash outlays and negative externalities. However they only pay $27 billion in taxes and charges, therefore requiring a net subsidy of $15 billion.

The PTUA counts costs like road construction, the value of land used by roads, traffic accidents, tax concessions for cars, pollution and emissions. On the payments side, the calculation counts fuel excise, GST, registration, insurance and tolls.

Unfortunately the PTUA provides little explanation for the assumptions used in its submission. VCEC makes it clear it has misgivings about the methodology adopted. I’m in any event a little cautious about the calculations of activist organisations of all colours.

Another source of data is a more recent study by one of Australia’s leading transport researchers, Dr Garry Glazebrook, from Sydney’s University of Technology, published in 2009. Although he doesn’t account for payments made by motorists, his work is especially valuable because he estimates the private and social cost of travel by both car and public transport.

His calculation of the private costs of car travel (in Sydney) comprises the cost of purchase plus running costs, including fuel, registration, insurance, servicing, tolls and paid parking. The external or social costs include traffic congestion, accidents, emissions, pollution, noise and unpaid parking.

He treats fares as a private cost of train and bus travel. He estimates the external costs from emissions and traffic congestion imposed by these modes (he says buses account for 10-15% of traffic congestion in Sydney), but the big social costs of public transport come from financial subsidies paid by the State Government to keep fares down.

As the exhibit shows, Dr Glazebrook finds the private costs of car travel are much higher than those of public transport, indicating travellers are prepared to pay a lot for the benefits of private transport.

More interestingly, he finds the social cost of all three modes – i.e. the sum of subsidies and negative externalities – is essentially the same at around $0.38 per passenger kilometre.

So cars are subsidised as much as public transport. There’s an important difference though, which explains why one subsidy gets more attention from governments than the other.

The subsidy for public transport is mostly a financial cost – it is funded direct from the current budget. Hence governments worry about it because they have to find the funds in the current term.

The subsidy for cars, however, is mostly an economic cost. The cost of externalities like congestion, pollution and emissions are diffuse and are mostly borne by the community at large. Even where these costs impact the budget – e.g. higher health costs from pollution – they mostly manifest years down the track and hence are of little interest to the current government.

Because they’re not paid out in cash at the time of the journey (like fares are), it’s hard for both politicians and the public to see the real costs cars impose on the community. Governments are accordingly less responsive to the need to take action.

In the past it could be argued there were practical difficulties to making motorists pay their real costs. However recent technological advances, especially in computing and communications, mean the failure to take action is now due more to political contraints than technical ones.

Dr Glazebrook ignores the taxes and charges drivers currently pay. That’s arguable in the case of the fuel excise because it’s the largest payment made by motorists and directly influences their perceived cost of travel. However it only adds around $0.04 per (vehicle) kilometre to the cost of driving – that’s modest in the context of a social cost of $0.38 per passenger kilometre.

So yes, we should reduce the level of subsidy for car travel. The politics of change however would be truly horrendous. For example, fully recovering that $0.38 per passenger kilometre social cost would theoretically require the price of petrol to rise by roughly $3.45 per litre. It wouldn’t need to be that big in practice because motorists would inevitably find ways to adapt – for example by shifting to vastly more fuel-efficient cars – but it would nevertheless be very painful.

A large increase in the cost of driving would inevitably drive up demand for public transport (in fact it’s an essential condition – whether the increase is in cash or time – to achieve significant increases in public transport’s share of travel). Finding ways of funding large-scale public transport improvements would be an extremely important goal for policy.

(Visited 5 times, 1 visits today)


Leave a comment

Your email address will not be published. Required fields are marked *

19 thoughts on “Should cars be subsidised?

  1. Burke John

    Ok Alan I stand chastised. I feel lucky not to be banned from figures are usually loose I admit it and I have pre-positions as well 🙁 In this instance though out of respect I’ll half my estimates and round them off to 100 billion in annual subsidies for motor cars. Nothing much to be concerned about for “award winning” treasurers and people with bigger economic fish to fry.
    Where do those awards come from anyway…the same place that real estate agents, banks and phone companies get theirs?

  2. Burke John

    Thank you sixd..I’ve never been assessed as “right” before. I doubt anyone has tallied the true costs and subsidies or anything remotely close. The right questions have to be asked first and data collected such as below.
    What is the real estate value of every car garage attached to every bungalow in Sydney, bricks and mortar + land? How much of this is counted as a subsidy to motoring and how much is counted against the increased cost of public transport as a result of the added distances? A figure substantial enough to make Gina Reinhart sneeze no doubt.

  3. Alan Davies

    Burke John and Flip

    Need to take care though that you aren’t double counting the costs.

  4. sixd

    Good on Alan for tying to highlight the quantum of the subsidy. Main point, though is that @BourkeJohn is right..most people DONT know about the subsidy and those that might have an idea would surely underestimate it. The real goal (which is where Alan has started) is to reveal a complete cost, i.e. one that incorporates @BourkeJohn’s fantastic list above of both time and $. Can anyone shed light on whether they have been tallied/are available??

  5. Flip

    First point the statistics used in this article are so old to be almost irrelevant. Second point, they refer only to Sydney. Third point, when referring to “cars” – is that the manufacture of cars or roads, bridges etc? Up shot – this article is to ambiguous. It gets an F.

  6. Burke John

    It is indeed an easy sport thinking of uncounted car costs for Australia. Here is my second new one for today. How about the war in Afganistan? Here is one connection. Al Queida is largely oil-funded.
    Anyhow I’ve cut and paste some more below just to help anyone wishing to make a start.
    “Twenty years ago, in one of his punchy little books called Energy and Equity, Ivan Illich pointed out that if one factors in the time spent parking, servicing, washing, and doing paperwork for our urban commuter car, its average speed over the 20,000km per year that most of them do, drops well below the average speed attained in actual driving. In addition to this Illich pointed out that if we consider the time spent earning the money to pay for the car and its various parking, servicing and paperwork demands, the average speed declines again. If we now factor in the time taken to generate the infrastructure requirements of the car, such as road and street construction and maintenance services, police, EPA- recognised environmental services, hospital, medical, legal, political, roadside repair, tow truck, ambulance and insurance services, almost all of which are currently debited to our social and bureaucratic resources, the average speed of the commuter car comes down to something our shoes would be ashamed of and the average commuter cyclist would have no trouble exceeding. Coupled with an extensive and fully used metrorail network, the potential average speed of bike/rail would take some beating. To underscore the point, factor in the currently unrecognised time spent on environmental, personal and social traumas, and efficiency in relation to the private car as a means of urban commuting becomes a complete non sequitir.”
    Prof. Frank Fisher, Director of the Graduate School of Environmental Science at the University of Monash, Australia, 1997

  7. Burke John

    Suburbanite is on the right track. Alan states that is a universal truth that it is widely acknowledged that cars are massively subsidized. I would suggest that a very few hold that perspective and when they do usually identify a mere tiny proportion of the real subsidy.
    So yes Suburbanite keep going and you should be able to identify at least 50 billion in money wasted on cars per every year in this country.
    For example suburban sprawl is a product of car culture. Suburban sprawl increases the costs of public transport exponentially. Therefore a large % of public transport costs, quite possibly the majority should in fairness be marked on the ledger as a subsidy to cars.
    You can think of a few billion wasted every day for several weeks without much too much effort. Eventually you may be like me surprised that after subsidizing cars there is any money left over for anything! Quibbling over Newstart increases seems absurd.
    A bumper sticker I read last week is an appropriate source for this discussion.
    It read “At least we are winning the war on education”.

  8. suburbanite

    Other costs associated with cars are noise and disposal. Are these included in any of the calculations? There has got to be some costs associated with using 2 tonnes of metal and plastic just to move one motorist. I wonder how this resource use compares to buses, trains and bicycles?

  9. Krammer56

    While I certainly acknowledge the societal costs of transport, valuation will always be fraught due to overlapping costs and benefits.

    For example, costing all road space against cars misses their use by peds, cyclists, buses, garbage trucks, services and landscaping. In fact, many of Melbourne’s widest streets were laid out well before the car (I particularly love some of the old photos of places like Dandenong, which show massively wide streets, with lots of dust and a few horse drawn carriges).

    Are pedestrian deaths at level crossing a train health impact or a pedestrian health impact? Are the high numbers of pedestrians killed near pubs with high blood alcohol levels really the car’s fault? What share of the health costs of a sedentary life style are due to car use and what due to western lifestyles (office jobs, overeating, watching TV, etc) generally?

    Adding bus priority measures can reduce delays for bus pasengers but can also increase delays for motorists.

    Best guesses are probably all anybody can do, and I doubt there will ever be agreement as the guesses will be coloured by the point of view.

  10. lindsayb

    I would love to see cycling added to the above graph, just as a point of comparison.

  11. Stephen

    Sydney’s deliberate sabotage of public transport in favour of private tollways has been supported by both parties for a generation. Over which period Perth (even Los Angeles) are cities that have gone about it a different way, to their great benefit.

  12. hk

    For the few who consider the health impact of urban systems in any discussion on cross subsidy and sustainability the following recent reference provides thought provoking leads.
    Hoehner, C. et al (2012). Commuting Distance, Cardiorespiratory Fitness, and Metabolic Risk retrieved from 21 May, 2012 from

    hk: It cost $31.50 to read the full article! Here’s a link to an ungated version (possibly slightly different to the published version). AD

  13. Wiz Aus

    Surely why it may be true currently that the rate of subsidisation per passenger km travelled in a private vehicle vs a PT is similar, we could add NEW passengers to existing PT at a lower cost than we could add new private vehicle drivers? For instance, let’s say the number of people needing to travel from Werribee to somewhere within reasonable distance of a train station close to the city during peak hours is expected to double in the next 20 years. The immediate financial cost of providing extra freeway space for those drivers is surely more than the cost of running a few extra train services – as it is, the freeway is already frequently at capacity, whereas the train line is not even close.

    On a slightly related note – has anyone considered the viability of a train service that ran from, say, spotswood, then went over or under the river, and then through port/south melbourne (probably mostly underground) to the city? Probably completely crazy but when you see how congested the Westgate bridge gets and with the talk about developing Fishermen’s bend etc. etc., seems to have some sort of potential…

  14. Daniel Bowen

    AD: “Those numbers are different to those in the PTUA’s submission”

    That’s correct – the numbers here have been steadily updated as new data has become available, following the submission to VCEC going in several years ago.

    Older versions available here.

  15. SBH

    I reckon my bicycle gets subsidised at the rate of f#ck all so next time some peanut says ‘do you pay registration’ (yes two lots actually) I’ll know what to say

  16. hk

    The word subsidy needs meaningful definition when it is bandied about and applied to the transportation sector. Some of us who are fortunate enough to live totally fulfilling lives without needing to use private cars, have no objection to our taxes contributing to our road based transport system for PT, freight, local delivery and pick-ups such as postal services, emergency, trade, service and similar vehicles. The main question is how much of our tax take should go into many of the inherent and avoidable inefficiencies in the road transportation system.
    There are many examples of where the high level of cross subsidy for private vehicle provision can be eliminated. Only two of the many follow:
    1) For example why should there be subsidizing of so much on street car parking outside residences? Car owners could pay the full cost for garaging their cars off the surface roads; a widely practiced solution to parking in urban communities in Europe.
    2) Also when the road system exceeds capacity, the system often runs at about 70% capacity with the consequential increased operating cost to all road users. The capacity demand situation is easily solved by ramp or entry metering. Ramp metering is already in operation by VicRoads to optimize freeway use in several locations. The delay of entry into congested road sections could be extended elsewhere to also optimize network operation by intelligent transport technology applications.

    To really stir the pot on full cost implications of private car use, consideration needs to be given to the sedentary behaviour of car occupants and their contribution to the burden of disease, which in turn increases life long health costs for the individual and the public.
    When the realization dawns that the current level of private car use in urban systems is not economically sustainable both in terms of infrastructure investment and adverse health impact, the level of cross-subsidy to private cars may be quantified and reviewed along the lines suggested by AD.

  17. Daniel Bowen

    Hi Alan, in case you missed it, the PTUA’s VCEC submission links to this page: – which towards the bottom details where the numbers came from.

    Daniel: Those numbers are different to those in the PTUA’s submission (see p90 of VCEC report). Are they a separate calculation or have they been updated? AD

  18. drsmithy

    All the calculations above seem to make (at least) two implicit assumptions:
    * that cars deliver no positives (social, economic, or otherwise).
    * that eliminating cars would deliver no negatives.

    Both assumptions seem ridiculous.

    drsmithy: They’re essentially the same point. And it’s a point that’s only really relevant if it’s thought necessary to subsidise car travel in order to enjoy its economic and social benefits. I can’t see that it’s necessary – in fact I think the existing subsidy might actually reduce the benefits. AD

  19. Jim Wright

    Any analysis of the private car versus public transport question should take into account the need to travel. A large proportion of the greater city travel is composed of driving to work, forced upon drivers because of the centralisation of business activity. Trains are effective as point-to-point modes of travel but cannot service all needs. Buses can stop to pick up passengers, but this increases the journey time. Tax incentives to encourage decentralisation (taking advantage of modern communications technologies), thus allowing more people to work locally, could substantially reduce the needs for subsidies.

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details