Menu lock

Cars & traffic

Dec 14, 2016


Government video on North East Link
Government video on North East Link

It’s a truism that networks like to get denser; so there’ll always be pressure to fill in “black holes” in radial train networks and “missing links” in motorway networks. No surprise then that Victoria’s Andrews government yesterday promised to join up the southern and northern ends of Melbourne’s suburban metropolitan ring road. According to Roads Minister Luke Donnellan:

Joining the ring road is a no brainer to take thousands of cars and trucks off local streets and congested freeways – but governments have put it in the too-hard basket for decades.

The project is expected to take up to 10 years to build, cost up to $10 Billion, and be funded by a combination of tolls and government contributions. The government has allocated $35 million to develop the business case and select a route prior to the 2018 election.

Critics were quick to pounce. The Opposition says the announcement is light on detail; the Public Transport Users’ Association (PTUA) reckons any reduction in congestion will be short-lived and the sensible course is to build more rail; and any number of social media commenters warned of the environmental risks in Melbourne’s sensitive north-east.

The promise is indeed short on detail and in my view is premature. But the government emphasises the project is recommended by the state’s arms-length “umpire”, Infrastructure Victoria, for construction within 5-10 years. Infra Vic’s preliminary analysis puts the benefit-cost ratio at 1.4 – 2.1 (2.2 – 3.1 when wider economic benefits are included). That’s very strong; the ratio for the Melbourne Metro rail tunnel is 1.2. And this project is much superior to that other “missing link”, the now abandoned East-West Link, with its sub-unity BC ratio (e.g. see What can we learn from the East-West Link debacle?).

The project illustrates the failure of politicians to embrace road pricing. If the government implemented a metropolitan-wide congestion charging regime over the next ten years to ration access to road space, the demand for the road could probably be delayed for decades. However, that’s politically unrealistic; probably the only plausible implementation over the next 10-20 years is a cordon system around the city centre like London, Singapore and Stockholm have put in place (and perhaps soon Beijing).

The PTUA is of course almost certainly right to say the new motorway will eventually become congested in peak periods. But that doesn’t necessarily mean it would be a waste or shouldn’t be built. Absent a comprehensive pricing scheme, there are some other aspects to consider in assessing the warrant for this “missing link”:

  • It would reduce travel times significantly in non-peak periods i.e. during the day, at night, and on weekends. More trips are made off-peak than in the two peaks combined. The Eastern Freeway opened nearly 40 years ago and still provides uncongested travel in non-peak periods even after connection to Eastlink.
  • It would provide a big increase in the number of vehicles that could travel along the alignment in peak periods, albeit (eventually) at congested speeds. Expanding capacity is an important consideration given Melbourne’s population is projected to double by around 2050.
  • It will be tolled; that will moderate growth in peak period use compared to an unpriced road. Tolling is aimed at maximising revenue so it isn’t as effective at managing congestion as pricing, but the option is open to the government to design (or negotiate) a tariff that focusses on demand management.
  • Even with congestion in peak periods, it would reduce the social costs of crashes and improve the amenity of local neighbourhoods, relative to leaving cars and trucks to continue to congest local streets. It’s probable peak period trips on the new road will be faster than maintaining the status quo.

Suggesting better public transport is an alternative is a glib response; it simply isn’t a realistic option in this location. The “missing link” is in the suburbs, between Ringwood and Greensborough, where car use dominates; 74% of weekday trips in Melbourne’s middle ring suburbs and 81% in the outer suburbs are made by car. Travel is mostly from dispersed origins to dispersed destinations and accordingly favours private travel by car (and, if it were safer, by bicycle).

This isn’t the city centre where the combination of high parking costs, congestion and the legacy radial public transport system means cars aren’t competitive; out here in the ‘burbs travel by car is much, much faster and more convenient than public transport could ever hope to be. The average car trip in Melbourne’s outer suburbs takes 29 minutes; that’s much quicker than the average 48 minute public transport trip in even the relatively well-endowed inner ring suburbs (e.g. see Will simply building more public transport seriously suppress car use?).

I suspect a lot of the opposition to the motorway I’m seeing reflects a knee-jerk antipathy to any road; on this view of the world there should be no new roads whatsoever. Adherents seem unaware that expenditure on public transport has increased markedly over the last two decades e.g. the Victorian government is spending $11 Billion on Melbourne Metro, $8 Billion on level crossing removals, $1.8 Billion on outer suburban public transport improvements, and billions more on upgrading tracks, signalling and rolling stock (see also Do governments spend too much on roads?).

Most tellingly, it ignores the fact most Melburnians don’t live in the inner city or even the inner suburbs; and most don’t work in the CBD where public transport is high quality and cars have limited utility. In the real world where most residents live and work in the middle and outer ring suburbs – especially a world on the verge of autonomous vehicles – it’s vital to take a more reasoned and selective view of road projects. There’s definitely a need for better public transport in the suburbs, but cars aren’t about to go away or become uncompetitive.

There are big potential environmental issues with this project as this is an environmentally valuable area, but it doesn’t follow that it must necessarily be a disaster. It’s non-negotiable that most of the route will be in tunnel so the risks can be managed. Still, this is an aspect that must be closely watched. It suggests the government’s cost estimate – and hence the BC ratio – shouldn’t be seen as hard and fast.

I think it’s too early to say definitively if the North East Link is, or isn’t, a worthwhile project because the necessary research hasn’t been done. It’s disappointing the government has unnecessarily committed the state to such a massive project off the back of what is essentially preliminary work by Infrastructure Victoria. It should’ve commissioned an in-depth evaluation and consulted with the public before making the decision; it smacks of opportunism.

Transport - general

Jun 11, 2015


Percent trips by distance category and mode on an average weekday in 2012/13, Sydney Statistical Division (source: BTS)

The first exhibit shows the proportion of trips taken by the main transport modes in Sydney for different distance bands (kilometres). It’s based on data from the NSW Bureau of Transport Statistics’ (BTS) household travel survey 2012/13. The second exhibit presents the absolute number of trips. (1)

It can be seen that:

  • Walking is the preferred mode for distances up to one km. It’s used for 71% of these short trips. It’s also very important for distances between 1 to 2 km, accounting for 24% of those trips.
  • Driving is by far the favoured mode for all distances above one km. It’s the mode of choice for at least 80% of all trips longer than 2 km.
  • Train and bus are very much minority modes. Overall, they comprise 6% and 5% respectively of all trips, compared to 69% for driving and 18% for walking.
  • Train does significantly better for longer trips than shorter ones. It’s share of trips longer than 20 km is 16%, compared to 4% for trips of 2 – 5 km. Indeed, it’s the only mode where the absolute number of travellers increases consistently with trip length.

The exhibits complement the data on the duration of trips in Sydney I discussed recently in Is driving quicker than taking the train?

The two measures – kilometres and minutes – show the same overall pattern; there are big differences between modes. The sorts of trips made by train are much longer, both in terms of distance and duration, than those made by car. (3)

Of course the exhibits describe the existing travel behaviour of Sydneysiders. The pattern is the result of a number of factors, especially historical land use and transport infrastructure development decisions.

It also reflects policies like the absence of a price on use of road space, limited coordination of modes, and restrictions on residential development. For better or worse, this pattern is the starting point for policy-makers.

The exhibits indicate the big challenge for public transport is to capture mode share from cars. To do that it will have to be much more competitive for short trips; more than half (58%) of all trips in Sydney are less than 5 km and 76% are less than 10 km. Only 10% are longer than 20 km, where trains currently do best. (2)

Public transport patronage will increase if improvements like higher frequencies, better coordination within and between modes, strategic investments in infrastructure, and a multi-modal fare structure are implemented.

But generating a significant increase in public transport mode share – something that seriously changes the relative heights of the bars in the first exhibit – will be far more challenging.

The other big challenge for policy-makers is to do something about cars. They’re by far the dominant mode in the metropolitan area and, notwithstanding increasing congestion, their share of trips is falling slowly (by 0.17% p.a. averaged over the last 10 years).

Public transport will not see a really serious increase in mode share at the metropolitan level unless system improvements are accompanied by measures that make cars relatively less competitive.

As I noted in the earlier discussion of trip duration in Sydney, policy-makers need to think seriously about how to ration demand for road space in congested locations (e.g. by congestion pricing); how to make cars cleaner, quieter and more fuel-efficient; how to make them safer for other street users; and how to make them more respectful of urban amenity.


  1. The mode share estimates in the exhibits allow for cycling, taxi, ferry, etc, but I haven’t shown them because their shares are too small i.e. 2% combined. These are unlinked trips and in-vehicle distance only. ‘Walk-only’ trips are those where the whole trip is made by walking and no change of mode is involved
  2. Public transport and car trips differ in other ways too; I’m limiting my attention to trip distance/time in this discussion.
  3. If walking is omitted, 49% of trips by motorised modes are less than 5 km and 71% are less than 10 km.
Number of trips ('000) by distance category and mode on an average weekday in 2012/13, Sydney Statistical Division (source: BTS)

Cars & traffic

Nov 12, 2013


Changes in the level of weekday light and heavy vehicle road travel in Melbourne with and without the East-West Link (data from GHD - EWL Traffic Impact Assessment)

Investing in big urban transport infrastructure projects, especially motorway and rail projects, is the common response to the growth pressures faced by our major cities.

Sydney’s population is expected to grow by 1.3 million over 2011-31 i.e. from 4.3 million to 5.6 million. It will require an additional 545,000 dwellings and 625,000 jobs.

Melbourne’s population grew by 600,000 over the last decade. It’s projected to increase from 4.25 million at present to 6.5 million by 2050.

But big transport projects are extremely expensive and Governments at best only budget for one every five years or so. It’s also not clear that all big projects provide a significant improvement outside their immediate area.

Consider for example Stage 1 of the East-West Link motorway proposed for Melbourne’s inner northern suburbs. It’s expected to cost $6-8 billion to construct and may monopolise new transport infrastructure spending for some years.

The first exhibit compares key indicators of forecast metropolitan-wide light and heavy vehicle road travel in 2031. It uses data from a report, East-West Link Traffic Impact Assessment, prepared by consultants GHD for the Linking Melbourne Authority.

It shows that building the East-West Link would increase the total number of trips on a weekday in metropolitan Melbourne by 6,000 compared to the no-build scenario. It would also increase total metropolitan travel by 800,000 km per day and reduce aggregate travel time by 4,300 hours per day.

These aren’t trifling changes; for example, if time is valued at (say) $50 an hour, the economic value of the forecast saving in travelling time would be nearly $80 million p.a. by 2031 (1).

But when looked at in the context of total metropolitan travel, the impact of the East-West Link would be decidedly modest. The number of trips in Melbourne would increase by just 0.04% (i.e. by four one hundredths of one percent); kilometres of travel by 0.4%; and aggregate travel time would reduce by 0.11%.

Whether or not the East-West Link is built will barely have any effect on the large changes in metropolitan-scale travel forecast over the period from 2011 to 2031.

There would be much bigger changes at the regional scale but they’re relatively modest too. The second exhibit (see below) shows the East-West Link would increase traffic flows and average speeds at key locations within the orbit of the proposed motorway by only 2-4% relative to the no-build case.

Speeds on the East-West Link itself would be much faster in the off-peak (e.g. average 20 minute saving from Chandler Hwy to CityLink, with reduced variability) but the motorway would hardly be the “game changer” for the city as a whole the Premier claims it will be.

This isn’t just a roads issue; there’s a similar story with some proposed rail lines. Consider the Rowville rail line promised for Melbourne’s outer south-eastern suburbs in the new metropolitan strategy, Plan Melbourne.

It’s likely to cost at least $4 billion to construct but by 2046 would only increase public transport’s mode share in the metropolitan area by 0.1%. Moreover, 57% of patronage would be siphoned away from other rail lines and it would reduce the number of car trips on a typical weekday in 2046 by a trifling 15,000.

Modelling for the Doncaster rail line, which is also promised in Plan Melbourne, indicates it would have virtually no impact on public transport’s mode share at either the local or metropolitan scales.

As discussed here (Does this freeway make any sense?; Is investment in transport a game changer?), adding a new link to a mature and well-developed network will generally only yield a small improvement.

Forecasts like those cited above suggest it’s worth asking if governments will be able to deal satisfactorily with the scale of growth anticipated for our cities just by building a major new piece of road or rail infrastructure once every five years or so. It’ll be an even more pertinent question if they continue to select projects for primarily political reasons rather than good policy reasons.

Governments need to find ways to increase the infrastructure build rate without threatening other important areas like education. They also need to ensure they build the best projects.

Most importantly though, they need to think much harder about how to make the huge stock of existing metropolitan-wide infrastructure work better e.g. by implementing road pricing. Just building major infrastructure projects isn’t likely to be enough.


  1. It’s impossible to know without seeing the detail in the business case, but in light of the Government’s advice that the BCR for East-West Link is 1.4, I’d have expected bigger time savings than suggested by 4,300 hours in 2031.  Perhaps the savings in other areas like vehicle operating costs, accident savings and “wider economic benefits” are large enough to get the BCR into positive territory?
Changes in the trips and average speeds at key locations within the East-West Link region (data from GHD - EWL Traffic Impact Assessment)



Jun 14, 2013


An Adventure of links for urbanists

Comparative subway construction costs, ABC Radio listeners have problems with cyclists, rock and roll economics, bicycle parking robotics in Tokyo, the complications of urban agriculture, musicians don't get science, stop thinking in terms of mpg...

How petrol consumption varies with average population density

Assorted links from around the traps:

Recent collections of links here:

Spirit of Process. Just in – a mysterious, unsolicited 4th entry in Barangaroo hotel and casino design competition. The architects say they’ve moved beyond orthdox ideas like responding to a sense of place: their design captures the Spirit of Process

Cars & traffic

Nov 13, 2012


Is this the world's Most Swimmable City? (source: Mail Online)

There’s a consensus among urban policy analysts that road pricing should be introduced into our large cities as soon as possible, particularly at peak times. Yet some groups, like the Greens and the Public Transport Users Association, aren’t convinced it’s all good news.

Road pricing involves charging motorists for the use of road space. It’s usually thought of as a means of addressing traffic congestion, but in principle it can be applied at all times.

Although it depends on how it’s implemented (e.g. cordon or per km charge), the key areas where road pricing promises benefits are well established. It could potentially:

  • Reduce the pollution, emissions, fuel use and disamenity associated with congested driving conditions
  • Lower the economic cost of congestion – principally delays – which BITRE estimates will reach $20 billion p.a. nationally by 2020
  • Give priority when roads are congested to high-value trips over those of marginal value
  • Moderate popular demands to “build our way out of congestion”
  • Lower the total level of off-peak car use by making drivers more conscious of the marginal cost of driving e.g. shifting some costs from standing charges to per km charges
  • Generate surplus revenue that can be applied to other purposes e.g. public transport improvements
  • Increase the demand for public transport by raising the relative cost of driving
  • Enhance horizontal equity by requiring those who drive more to pay more.

But like everything in public policy, road pricing also has its downsides. One that looms large is it’s a hard sell politically.

The key criticism in policy terms is it’s vertically inequitable – those with fewer resources would pay a higher proportion of their income in charges.

That’s true, but it’s true for public transport and other basic services like electricity, gas and water too. In fact in these cases the marginal charge increases with consumption.

It’s not good policy to encourage excess consumption of a scarce resource in the name of a single objective. A better approach would be to compensate at-risk populations adversely affected by road pricing.

However there’s another objection that hasn’t gotten much attention. It’s most clearly articulated by Dr Paul Mees in his influential book, Transport for suburbia, and in this submission he made to Infrastructure Australia.

Dr Mees argues that congestion charging makes driving more attractive because it offers higher speeds. The inevitable consequence is that on average drivers will use the extra speed to make longer trips, using more fuel and generating more pollution and emissions in the process.

Rather than seeking to manage congestion, the alternative is to see it as a way of deterring driving and encouraging higher public transport use. In his book (p47), Dr Mees cites the experience of Vancouver, which “reduced journey times by promoting congestion.”

In its 1993 regional plan, Vancouver positively embraced congestion as “part of a package designed to promote self-containment and mode shift away from the car.” The pay-off was that despite rapid population growth over the same period:

Vancouver was the only Canadian urban region where the average time taken for the journey to work….declined, from 70 minutes in 1992 to 67 in 2005. By contrast times in Montreal jumped from 62 to 76 minutes.

I think it’s a plausible argument. Motorists would indeed be likely to drive further on average if speeds increased. That’s likely to happen whether the speed increase results from congestion pricing, road works, or some other change.

I’m not sure that it’s such a big problem, though. All the pricing proposals I’ve seen are aimed at increasing speeds enough to clear out logjams and get traffic moving at a modest speed that’s still well below the speed limit.

Moreover any increase in speeds needs to be interpreted in context. The extra travel might be offset in whole or in part by those marginal drivers discouraged by pricing.

And longer trips might also be offset by a fall in low value trips in the off-peak if the pricing regime extends beyond peak hours. Further, if levied on a per kilometre basis it could encourage shorter average trip lengths.

In any event, the benefits of longer trips have to be taken into account along with the costs. Having the choice to drive further could mean, for example, that a worker has the choice of a better job (more efficient labour and job matching).

The key benefit of road pricing relative to “promoting congestion” is it sorts the traffic according to value. Someone with an urgent business or personal meeting to get to, or a load of goods to deliver, will welcome the greater certainty provided by less congested conditions.

I don’t put a lot of store by the Vancouver numbers. Dr Mees elaborates on them further in this paper, but he doesn’t show a causal relationship. There might be other factors that explain the observed reduction in travel times.

I am in any event wary of relying on the experience of just one or two other cities to draw general conclusions, let alone ones that are applicable elsewhere. There’s immense variability between cities.

For example, this writer claims that the introduction of road pricing in Stockholm in 2006 had a similar outcome to that attributed to congestion in Vancouver. Commute times dropped and public transport use increased.

Yet even if it’s accepted that road pricing is as costly in terms of induced travel as Dr Mees implies, it’s not a “policy stopper”. Of course road pricing has negatives as well as positives – all policy initiatives do.

What matters is how those costs and benefits compare. On balance, I think it makes more sense to charge for road space than tolerate congestion, although it will depend ultimately on what sort of implementation is politically feasible.

It’s true congestion pricing won’t produce a wholesale reduction in car use – that’s not what it’s intended to do. It’s very likely that in some form or other cars will be with us for a long time yet so we need to find ways to manage them better. Road pricing should be one of those ways (more on road pricing here, here, here and here).

Cars & traffic

Oct 26, 2010


What Americans mean by a 'McMansion'. Click for slideshow.

I’m not aware of anyone who disagrees seriously with the contention that car travel is underpriced. The consequence of this inefficiency is we drive more than we otherwise would and more than is socially optimal.

The idea of road pricing is that drivers should pay the real costs they impose on others through traffic congestion, pollution, noise and carbon emissions.

There’s also another force at play here which exacerbates the problem of excessive driving. There are some costs that drivers actually do pay – standing costs like depreciation, insurance and registration – that are “disconnected” from the perceived cost of travel.

A person deciding whether or not to drive somewhere will tend to take account of the cost of their time plus petrol, but they usually don’t perceive the standing costs. This under-estimation promotes more driving.

There have been various experiments with road pricing, such as the well known Singapore and London central city cordons (giving rise to amusing interpretations such as this one by Boris Johnson). However this is a technologically outdated approach – transponders and/or GIS technology mean it is now feasible to charge motorists in relation both to distance and traffic conditions i.e by location and time.

A driver who paid a price for a litre of petrol that included both external and standing costs would have a strong incentive to drive less. A gauge on the dash showing the total cost ticking over with every kilometre would provide an even more powerful nudge to think long and hard about the wisdom of driving.

Road pricing can be thought of in simple terms as a two-part per kilometre tariff that recovers both external costs and those standing costs that can be disaggregated. One part is a charge reflecting the general cost of using the roads. The other is a variable price reflecting specific costs like congestion in peak periods.

There are potentially some important benefits for the wider community from road pricing: Continue reading “Why is road pricing a good idea?”

Cars & traffic

Oct 12, 2010


High St Reservoir - where road meets rail!

I’ve concluded before that the most plausible scenario in the forseeable future is that cars will continue to be used for the majority of trips in Australian cities. Increasingly, these cars will tend to be powered by clean energy sources and will be slower and more civilised than today’s vehicles.

I expect growth in public transport and cycling will be much faster but the absolute number of cars will very probably still increase. It is therefore inevitable that there will be continuing pressure for new freeways.

So is there any sort of case for freeways or should all new infrastructure funding be reserved exclusively for public transport, as proposed by the Independent Inquiry into Sydney’s long-term transport needs?

The key criticisms of freeways, most of which are pretty familiar by now, are that they:

  • generate more car travel and higher speeds, which in turn produces more emissions and pollution and consumes more oil
  • promote a sprawled, car-dependent urban form – the higher speeds provided by freeways mean people tend to live further away from activities
  • undermine the viability of public transport where they compete directly
  • impact on neighbouring uses – the amenity of adjoining land uses is diminished by noise and pollution
  • crowd out investment in transit – governments prioritise funding to roads and investment in public transport is neglected
  • sever social linkages and networks when they’re superimposed on existing communities
  • cannot deliver very large numbers of people to concentrated locations, like CBDs, without becoming congested relatively quickly

In fairness, it should be acknowledged that efforts have been made to ameliorate some of these issues. Much of the investment in freeways over the last twenty years has been by the private sector. Governments have built sound barriers along new and existing freeways and the almost mandatory use of tunnels in built up areas means severance is no longer the issue it once was. Continue reading “Is there a case for freeways?”

Cars & traffic

Oct 10, 2010


World's tallest buildings in 1884
47,000 tickets for the Big Day Out at Flemington race course sold out within minutes of going on sale at 9am on Friday. Tickets for the Sydney and Gold Coast events also sold out in record time. The organisers have arranged an additional event in Sydney (tickets on sale October 11).

The issue of how to deal with excess demand is a common one in cities, whether it’s concerts, road congestion or the (until recently unfamiliar) problem of public transport congestion.

As I went through the stressful process of buying tickets for the Big Day Out for my son (successfully as it turns out), I got wondering if ‘first-in, first served’ is really the ideal way to go about selling tickets for such a high-demand event. The organisers appear to think it isn’t.

They made a decision on Friday, apparently on the fly, to hold back 3,000 tickets for the Melbourne event. These will now be made available via a randomly drawn online ballot conducted once a week for ten weeks. Prices will be the same as they were for the original sale.

The organisers quite clearly could charge more than $164 per ticket (including booking fee), so why don’t they? It could be that they’re simply not very good at estimating people’s willingness to pay for a particular set of bands, or perhaps it’s such a black art that they’re not prepared to take too many risks. Or it might be that they’re in it for the long haul and don’t want to alienate concert-goers from returning in future years – they might want to protect the “brand”. Continue reading “How to charge for the Big Day Out?”