Transport - general

Jan 21, 2015

Why do commuters drive to work instead of taking transit?

The great majority of commuters in Australia’s major cities work in locations where public transport is so uncompetitive relative to cars they're prepared to pay a big premium to drive.

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

How much the Australasian Railways Association reckons car-owning commuters could potentially save annually by leaving the car in the garage and taking public transport to work instead (source: ARA)

The Australasian Railways Association (ARA) released its annual report yesterday claiming that commuters could save big money if they use public transport instead of driving (media reports here and here).

The report, The cost of commuting, says the average potential saving for capital city commuters could be $1,725 if they left their vehicle at home and travelled to work on the train, bus, tram or ferry instead; or $9,974 if they went without a vehicle entirely and relied solely on public transport. (1)

These are substantial savings; it seems commuters must be insane if they drive to work. Yet averaged across Australia’s capital cities, 75% of workers insist on getting to work by private vehicle.

Even in Sydney, where public transport has by far the highest mode share for the journey to work, 67% of workers commute by car.

Part of the explanation is that the Australian Railways Association (ARA) has been imaginative with the facts; it’s seriously exaggerated the cost of driving to work. Sadly, I don’t find that surprising; after all it’s a lobby group and in today’s public discourse it seems pretty much anything goes.

One way it’s been especially creative is to overstate the costs of driving; it’s assigned all the standing costs of owning a car – things like depreciation and registration – to commuting. (2)

That greatly helps the ARA’s agenda because fixed costs are usually much higher than operating costs. But it’s misleading because commuting is only a minority trip purpose.

In Sydney, for example, commuting only accounts on average for 26% of all kilometres travelled on a weekday. The ARA should’ve assigned three quarters of standing costs to non-work purposes, like social and recreational travel. (3)

That makes a big difference in terms of the relative cost of the two modes for commuting but even so, the most important explanation for the popularity of driving to work is straightforward: it offers substantial and tangible benefits relative to what the available public transport offering provides.

For most commuters driving is a far more attractive option and worth paying the considerable extra cost. Compared to public transport it’s usually quicker, it’s available on demand, it’s point-to-point, it’s more reliable, it’s more private, it’s more flexible, and so on. That’s not rocket science.

Commuters know they could save a lot of money if they were to forgo owning a second or first car and take the train or bus instead. It’s just that the (private) benefits cars provide compared to the available public transport options are worth the additional expense.

We already know that commuters act rationally when public transport is competitive with cars; the one place where that applies in Australian cities on any sort of scale is the CBD.

Upwards of 70% of all CBD workers in Sydney and Melbourne already commute by public transport, especially by train. They make that choice for two main reasons.

The key one is that traffic congestion around the dense centre in combination with high parking costs makes driving to work much less competitive against trains, buses, trams and ferries.

The other reason is that the CBD is the one location in Australia’s metropolitan areas where the level of service provided by public transport is exceptionally good.

Public transport systems in Australian cities are predominantly radial with nearly all routes converging on the CBD. Moreover, trains have their own right-of-way and so provide a faster trip to the city centre during peak hours than cars.

The minority who continue to commute by car to the CBD in places like Sydney (18%) and Melbourne (26%) are either strongly committed car users; prejudiced against public transport; start and end work outside the peak; or have their car use subsidised by their employer and/or the tax payer.

There’s still scope in places such as the centre of Adelaide – where cars have 54% mode share for the journey to work – to increase public transport patronage at the expense of cars. But in cities like Sydney and Melbourne the marginal CBD worker who drives will be much harder to persuade.

The problem for commuters is that even though the public transport rich CBDs of Australia’s major cities constitute by far the largest single concentration of jobs, the great majority of metropolitan jobs aren’t in the CBD.

For example, only 15% of all jobs in Melbourne are within 1.5 km of the town hall. In fact 71% of jobs are more than 5 km from the centre and of these, only 20% are in significant suburban centres; the other 80% are relatively dispersed.

Jobs located outside the city centre are considerably harder to access by public transport from the rest of the metropolitan area than those in the CBD. Parking costs are also generally lower, so driving is more competitive too. (4)

At present, the great majority of commuters in Australia’s major cities work in locations where they regard public transport as so uncompetitive relative to cars that they’re prepared to pay a big premium to drive. (5)

There’s a suite of measures available to policy makers to improve transport in cities, including improved public transport, higher densities and better management of road use. However as I’ve discussed before, the solutions aren’t easy e.g. see How easy would it be to shift commuters out of their cars?


  1. How the ARA arrived at many of its specific numbers isn’t clear from the report. Note also there’s no mention of cycling (the ARA is a rail industry association after all) but commuters could save even more by cycling instead of driving.
  2. The ARA’s been “creative” in other ways too. For example, in it’s discussion of Melbourne, it cites two cases: one is a 24 km one-way commute and the other is a 63 km one-way commute. However the median one-way commute in Melbourne is around 16 km; plugging it in to the ARA’s formula gives an annual saving from leaving a large car (Commodore) at home of $629. I’ve previously made comments about the methodology in reference to last year’s report; see How much extra does it cost to commute by car?
  3. Commuting only accounts for 15% of travel in Sydney when the purpose is measured by the number of trips. However commutes are by far the longest trip purpose in Sydney, so they make up a larger share of travel when measured by kilometres.
  4. In Melbourne, there’s a very large increase in car use (and fall-off in public transport mode share) as soon as you get beyond walking distance of the city loop.
  5. Not all locations are the same; public transport generally captures a higher share of commuters in centres and the closer jobs are to the CBD.


Leave a comment

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

25 thoughts on “Why do commuters drive to work instead of taking transit?

  1. Oz (Horst) Kayak

    The main reason why so many people drive to commute to work is because they can afford to pay the related costs of car ownership etc. People who reason that investing in sustainable travel and therefore do not consider cost the main issue then face the problem of the time-budget to be allocated to travel; a travel time requirement allocation which is magnified on many irritating occasions by low levels of time based PT service delivery.

  2. Tom the first and best


    If you buy a car only for commuting and then sometimes use it for non commuting trips, whether trips that would not have otherwise been car trips or trips that would have been in another car, all the costs of owning that are still because it was bought for commuting.

  3. Alan Davies

    Gavin Moodie #21:

    Yes, there’s room for improvement in the peak hour public transport services to the CBDs of Australia’s major cities, especially in terms of capacity and reliability.

    But relative to the option of driving to work in the CBD, they’re certainly plenty good enough to offer a highly competitive alternative. Let’s not forget that driving to the CBD isn’t perfect either; it also has capacity constraints (congestion) and reliability problems (as Steve777 points out at #22). The shortcomings of public transport are not a key explanation for why a minority continue to drive to work in the CBD, at least in Sydney and Melbourne.

  4. Steve777

    One drawback of travelling by car in Sydney that is not often mentioned is the unpredictability of journey times. Sydney’s road system operates on the verge of chaos. Any disruption – normally a traffic accident but sometimes roadworks, special events or bad weather (not uncommon in Sydney) quickly causes gridlock. Because Sydney’s road system is full of choke-points, it is normally near impossible to avoid such congestion when it occurs.

    What this means is that if arrival at your destination is time-critical (e.g. an important work commitment, a wedding, a specialist appointment) , you need to allow an extra half hour if your journey crosses more than a couple of suburbs.

    In the meantime, as Scott @ 20 says, the reliability of public transport, especially trains, has greatly improved in Sydney, so equivalent problems on public transport are much less common than they used to be. The time advantage of private cars is not as great as it seems unless it doesn’t matter if you’re half an hour late about 10% of the time.

  5. Gavin Moodie

    Of course Australian public transport’s service to the cbd is ‘exceptionally good’ only in comparison to its exceptionally poor service outside most of the cbd.

    There is no need for a public transport timetable in Berlin, London or Toronto because the services are so frequent that there’s no need to time one’s trip to suit public transport. Neither have I ever been passed at those cities by a service because it is too full to take on more passengers, unlike my frequent experience in Brisbane and Melbourne.

  6. Scott

    Another factor is the cost of public transport e.g. in Sydney it costs me nearly $15 a day for a return trip that is only 19km (includes a train and bus). It’s incredibly expensive for a “public” mode of transport.

    Although I had to say the train in particular has HUGELY improved in terms of reliability compared to the last time I lived in Sydney (pre 2003).

    The author is also correct in terms of the radial nature of the train network – it’s still nearly always easier to drive if you’re wanting to do a e.g. south to west trip (rather than catching the train in to the city and out again).

  7. Alan Davies

    Dudley Horscroft #17:

    If you buy a car for commuting and use it 100% for commuting then fine. But if you end up using it 50% for commuting and 50% for shopping, then only half the standing costs should be apportioned to the cost of commuting. Your original intention is irrelevant; what matters is how you use it.

    Cars used for commuting in Australian capital cities are a mixture of first, second and even third cars. Very few would be used entirely for commuting (although I met a guy once who had a $1,000 rattler just to get to and from the station).

    I disagree with you on the quality of the ARA’s methodology. Apart from the issue of standing costs, it’s not at all clear how they arrived at the dollar savings for each city. Moreover, I suspect they were highly selective and picked a handful of “typical” commuter routes (mostly long ones) to arrive at their cost comparisons (see my footnote 2).

  8. Dudley Horscroft

    Re The Pav’s 16, the quote was:

    “We have no history of funding urban rail and I think it’s important that we stick to our knitting,” the Federal Opposition Leader declared. “And the Commonwealth’s knitting when it comes to funding infrastructure is roads.”

    One of the worst statements made by Mr Abbott. The Commonwealth’s ‘knitting’ was National Roads and National Rail, all other roads and rail are for the States to play with, and just as tolls are charged on rail, so should they be on roads. If the Commonwealth is willing to throw money away on WasteConnex and the unlamented East-West Link, there should be no opposition to funding the Doncaster Light Rail or the direct Parramatta – Lidcombe – Dulwich Hill – Airport railway.

  9. Dudley Horscroft

    Re your 15. Apportioning costs is wrong. It is a matter of economics. Why do you buy a car? If it is to drive to work and back, then 100% of the standing costs should be allocated to the cost of commuting by car. It you buy a car for the pleasure of driving into the country at weekends, or because you just like the sheer joy of driving, then you allocate 0% to the cost of commuting by car. In each case the costs should be allocated to the reason for buying the car. The car is ‘free’ with regard to the other purposes – you have already bought it.

    That said, it appears the ARA methodology is good.

    Incidentally, the figure in the Sydney diagram of 65,570 persons travelling from the SE suburbs indicates why it is necessary to go for light rail (CSELR)in that area, not served by rail. No doubt a similar figure would be applicable for areas either side of the Parramatta Road.

    I would suggest that a lot of the demand for “WasteConnex” is because the trains from western Sydney are too full and too infrequent – which could be remedied with more trains, and possibly the branch from Ashfield direct to Barangaroo – costly, yes, but far cheaper than WasteConnex.

  10. The Pav

    A week or so ago the Mayor of Fremantle had an article printed that raised an aspect of the Public Transport/ Car debate that I had really quantified or considered.

    He made the point that each car needs three parking spaces ( home/work/shop)

    For Perth to meet the expected traffic growth in the coming decades this will result in urban sprawl increasing by 100 sq km just to meet the parking need.

    Forget any other factor the economic cost alone is prohibitive which makes the PMs comment that Railways aren’t in the Liberal DNA and that roads are the way to go even more stupid.

  11. Alan Davies

    Saugoof #11:

    Thanks, that’s clear now. Agree that a car used for commuting is likely to rack up proportionately more kilometres for this purpose than one that isn’t. I’m not aware of any data that shows what the difference is though. But even if it were 50%, the point still holds that ARA should’ve apportioned the total standing costs between commuting and other travel purposes.

    Prompts the question whether or not cars used for commuting are also substantially different from the overall fleet in some other way. For example, do work-related tax breaks significantly shape the standing and operating costs of the commuting fleet? Or are purchasers more concious of running costs when they purchase a vehicle for commuting? Are there differences between first vs second cars used for commuting in terms of fixed and operating costs?

    Update: Just stumbled on ABS stats showing 27% of all travel by passenger vehicles in Australia in 2011/12 was for commuting. A further 19.9% was for business travel. Some of that business travel will almost certainly be commuting.

  12. Dylan Nicholson

    Saugoof, it’s not that hard to believe…apparently the average person drives 15000 km a year, or ~300 km a week. So that would mean on average they commute ~75 km a week, and drive an extra 225km on non-commuting trips. Sounds a lot, I agree but that’s the available stats. And while obviously quite a lot of people commute a lot more than 75km a week by car (that’s only 7.5 km per trip for a full-time job), quite a lot of us don’t commute at all by car (because we use other transport, work from home, or don’t have full-time jobs that require commuting as such).
    Also that “non-commuting” km presumably includes driving by those whose job it is to drive (taxi-drivers, truck drivers etc. etc.), which is probably quite a high percentage of the total.

  13. michael r james

    #12 Jacob HSR

    Description from blurb of Donald Shoup’s book referenced in #7.

    [One of APA’s most popular and influential books is finally in PAPE, with a new preface from the author on how thinking about parking has changed since this book was first published. In this no-holds-barred treatise, Shoup argues that free parking has contributed to auto dependence, rapid urban sprawl, extravagant energy use, and a host of other problems. Planners mandate free parking to alleviate congestion but end up distorting transportation choices, debasing urban design, damaging the economy, and degrading the environment. Ubiquitous free parking helps explain why our cities sprawl on a scale fit more for cars than for people, and why American motor vehicles now consume one-eighth of the world’s total oil production.]

  14. Jacob HSR

    Parking is too cheap.

    When I used to work in the CBD, all of my colleagues got to the office by train or bicycle.

    But when the office moved to Port Melbourne, near a tram stop, I could not believe how many of my colleagues started driving to work!

    I kept my faith in PT and always used the train + tram to get to the new office.

  15. Saugoof

    Alan Davis #9
    The problem I have with how you interpret ARA’s numbers is that you claim that because commuting takes only 26% of a car’s kilometres travelled, only a third of the registration and depreciation costs should have been used. I’ve mentioned before why I have a problem with that 26% number, but either way, it’s not applicable here. When you take this number, you take all cars into account, regardless of whether the car is used by people commuting to work, or whether the car is owned by a pensioner, or it’s a “weekend getaway” car, or belongs to a stay-at-home mum, etc. But the ARA’s claim is about commuters saving money. For drive-to-work commuters, the percentage that commuting takes up of the total distance the car travels will be far, far higher than 26%.

  16. michael r james

    #9 AD:
    [assigned all the standing costs to commuting rather than just the portion of travel related to commuting; what’s harsh about exposing that? ]

    Goes to my question (#7):
    [how many households would reduce their car-ownership by one, if the commute (by at least one household member) was done by PT? Almost certainly this would not be trivial]

    It would be entirely fair to attribute 100% of the capital costs (& insurance & maintenance) of ownership to commuting for the fraction that only own that extra car because of commuting.
    In fact David Owen (of Green Metropolis) explained the situation he got into when living in an “idyllic” rural environment in upstate NY: both adults needed a car for work (even though he was mostly a work-at-home author) plus–after some painful experiences–a third car to cover those times when one car was in for servicing/repairs! When he moved to Manhattan: zero cars. He summarizes the effect of living in semi-rural or exurban environments with witticisms such as “(when moving into the country/exurbs) … you are really moving into a car” and the mortgage-seeking advice (very pertinent to younger Australians): “drive until you qualify” (then pay forever in multiple car costs, repairs, insurance, petrol and most of all quality-of-life).

    The point is that no clinical analysis of statistics, especially dubious (and they are almost always dubious) monetary ones, will ever capture the most important factors related to urban planning or such lifestyle choices. I know this is counter to what economic-rationalists think, that absolutely everything of any significance can be captured by a dollar figure. They are wrong.

    And #4 Roger Clifton, seems you have made some awful lifechoices if the only relief is to spend hours in your car each day. You should find a situation where ideally you can walk to work (or use rail) so you don’t have the urge to beat your wife and kids so much (joke). Maybe those non-driving New Yorkers like David Owen spend their discretionary budgets not spent on cars instead on shrinks? Seriously, keep in mind what Kierkegaard (not accidentally a Copenhagen resident) claimed:

    [Above all, do not lose your desire to walk; every day I walk myself into a state of well-being and walk away from every illness. I have walked myself into my best thoughts, and I know of no thought so burdensome that one cannot walk away from it.]

  17. Alan Davies

    michael r james #7:

    Good to see AD applying scepticism to a report’s claim. I just wish he would do it to all reports that he references.

    That’s because the ARA provided enough information in their report to make that call; in the case of the CESLR, the NSW government hasn’t provided enough info to evaluate the business plan (although the headline numbers look plausible). It probably comes down to a lobby group wanting publicity and hence feeling they have to provide a modicum of info to be taken seriously, whereas the NSW government is more concerned about avoiding the sort of publicity that might involve scrutiny.

    Saugoof #8:

    Still at a loss to understand why you’ve concluded I’ve interpreted ARA’s numbers too harshly. The ARA report assigned all the standing costs to commuting rather than just the portion of travel related to commuting; what’s harsh about exposing that? Have a look at footnote No. 2; if I wanted to be harsh there’s lots more where that example came from!

    The NSW Bureau of Statistics data is from their annual household survey. The Bureau says:

    The 2012/13 estimates featured here are based on three years of pooled data collected from July 2010 to June 2013 from a combined gross sample of 14,636 GMA households of which 9,859 (67%) responded.

    Perhaps you’re in the 49% of commutes (can’t break it down by mode) that the Bureau says are less than 10 km one-way (or even the 28% whose one-way commute is less than 5 km).

  18. Saugoof

    Alan Davis #8
    My dispute is that while accusing the ARA to interpreting the numbers too favourably, you’re going the other way and interpret the numbers too harshly.

    As for those 26%, yes that’s from the bureau of transport statistics, but I still find that very hard to believe. I’m not accusing anyone of fudging the numbers, but I just have a hard time believing that these are accurate. I know personal experience counts for nothing and my circle of friends and work colleagues is not really representative, etc. But nevertheless, I can’t think of anyone I know whose work commutes wouldn’t take up around 80% of their total car kilometres travelled. I mean, just think of it, an average commute into work is about 30 minutes, so that’s an hour a day, 5 hours per week. If this was only a quarter of your weekly commutes, you’d either spend each weekend driving 15 hours or you have some very busy evenings.

  19. michael r james

    Good to see AD applying scepticism to a report’s claim. I just wish he would do it to all reports that he references.

    An interesting factor to know is how many households would reduce their car-ownership by one, if the commute (by at least one household member) was done by PT? Almost certainly this would not be trivial and would partly counter AD’s correction to the stats. However it is also true that many drivers don’t truly consider those issues unless really under strong financial pressure; ie. they tend to totally discount the cost of (multiple) car ownership.

    No one should need reminding but Australia has one of the world’s highest, if not actual highest, car ownership per capita, an obvious consequence of our awful urban sprawl combined with Olympic-medal-winning levels of neglect of public transport. Incidentally radial PT networks needn’t be inefficient if the urban fabric is designed with that in mind; ie. linear development; beads-on-a-string model etc. This is even true for roads which is why many cities tend to develop in quasi-linear mode by default.

    While using kilometres travelled is clearly more relevant than trip number, it is also significant that commuting has worse impact because 1. it is more single-occupancy than any other; 2. driving at the most congested times, ie. more stop-start and more polluting etc. 3. grotesquely inefficient use of resources for car parking in CBDs (or anywhere really, see:
    The High Cost of Free Parking
    Shoup, D. 2011
    And incidentally this is one of the big problems of the imagined Nirvana of self-driving cars.

  20. Alan Davies

    Saugoof #2:

    The data on commuting’s share of travel in Sydney is from the NSW Bureau of Transport Statistics‘ latest household survey.

    I’m not sure what your getting at when you dispute my interpretation of how ARA interpreted the numbers. The ARA is describing the current costs of driving vs public transport, so the current commute number (“26%”) is the relevant one. You seem to be addressing a different question?

  21. Norman Hanscombe

    As long as people are using vehicles paid for by entities other than themselves, is it surprising if they don’t use public transport? When I had a free car and parking space at my job, but chose to walk to work in all weather other than events such as driving hail, colleagues used to laugh about it.
    Financial penalties are almost the only thing which effectively changes most people’s habits.

  22. Roger Clifton

    One thing that driving one’s car, or at least, sitting in it, gives is privacy. For many of us the commute is a treasured patch of time alone, protected from the intrusions of family at one end and co-workers at the other. We can explore ideas, rehearse what we should have said, dream of what we could be doing next week, next month and next year.

    It is also time when our personas shift out to a distance. Many of us cease to be a human and become just the black car with wog windows growling threateningly at the lights. Others of us are alone on a stage with an admiring audience, puzzling others in the traffic with our gestures and muffled rants.

    Of course, music can soothe the unsettled soul. But the cassette player also allowed us to talk to ourselves thoughtfully and repeatedly – I used to speed-read my journals etc, dictating the interesting or intricate bits into a cassette for repeated playing. Then I would mull over the interesting bits as the traffic allowed.

    The DVD player can also teach – I spent a year of commuting learning French for a trip to France. Less successful were my attempts at learning to sing, though I must confess it was much more fun to arrive at work out of breath and smiling. No doubt some people have secret devices to plug into the cigarette lighter too, but of course I know nothing of that.

  23. Jill Baird

    I am a big fan of public transport, but when I had a load of stuff to take to work, kids to drop off and pick up on the way, shopping to do after work etc, guess what won.

  24. Saugoof

    I don’t think I can agree with your interpretation of those numbers vs. how ARA interpreted them. Even if, for example, only 26% of a car’s kilometres in Sydney is for work trips (something I find a bit hard to believe, but let’s agree that this is the case), increased and better public transport that comes with higher numbers of passengers means that a lot of the remaining kilometres end up getting done on public transport too.

    I’ve had the benefit of living in countries with very good public transport where it is not just the norm to go to work on public transport (at least in larger cities), but even business trips to anywhere in the country were normally done by train too.

    I now live in the Melbourne CBD. It’s true that the majority of people who work here commute by train, but I’m still astonished that anyone would choose to drive into the CBD during business hours. Horrible traffic, it’s slow, expensive (though sadly a lot of companies pay for employee car parks) and really strikes me as the worst choice of transport that you could pick.

  25. Dylan Nicholson

    On the other hand, I don’t see much mention of what to me are the much *greater* benefits of using PT over cars for commuting than the immediate dollar benefits: more exercise (unless you happen to live and work opposite train stations/bus/tram stops), more productive time usage (you can’t read while driving), less pollution, less wear and tear on the roads (that we all pay for ultimately) etc. etc. A dollar value could easily be put on all of these too, if it was really necessary to prove a point.
    But aside from the time-usage issue, all of these are advantages that are purely long-term and, except exercise, apply to society as a whole, with minimal immediate benefit to individuals. Most people understandably discount those benefits heavily compared to immediate personal benefits.

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