Menu lock


Sep 24, 2017



Sydney and Melbourne GCCSA population vs with addition of Wollongong and Geelong LGAs (source data ABS)

In this week’s Tw3, The Urbanist comments on:

  • Is Melbourne already bigger than Sydney?
  • Bigger is better – small businesses once employed over half of private sector workers but no longer
  • Public transport is always greener on the other side
  • Doncaster hill’s next major apartment project verges on approval
  • Council wants toll for $5.5b West Gate Tunnel to be set by minister, not operator
  • Sydney park to be named after Green Ban-era community leader Nita McCrae

Is Melbourne already bigger than Sydney?

Sydney has a bigger population than Melbourne, right? Well, maybe. A lot depends on where you draw Sydney’s northern boundary. At the moment the official definition of Greater Sydney stretches all the way to Lake Macquarie, about 120 kilometres north of the CBD.  So if Greater Sydney did not include the Central Coast and Greater Melbourne did include Geelong, the Victorian capital would be Australia’s biggest city by a significant margin.

But conversely, if Greater Melbourne did not include the satellite cities of Melton and Sunbury and the shires beyond, and Greater Sydney did include the City of Wollongong, the NSW capital would be head and shoulders larger than Melbourne i.e. 5.2 million vs 4.3 million.

How long is a piece of string? Defining the geographical extent of cities is fraught, but arguably the best way is to include those outlying locations that have a strong economic and social relationship with the central urbanised area. Labour catchments are a useful measure and they’re mainly what the ABS uses to define its current definitions of Greater Capital City Statistical Areas (see also How big is Melbourne?).

Anyway, the author is a little casual with his measurements. Greater Sydney’s northern boundary isn’t 120 km from the CBD. According to Google Maps it’s 90 km to Point Wollstoncroft on Lake Macquarie. Further, Geelong isn’t “a little closer to Melbourne’s CBD than the central coast hub of Gosford is to downtown Sydney”; Geelong is 65 km and Gosford 50 km from their respective CBDs. While the author emphasises the stretch to the northern boundary, note that Greater Sydney’s southern boundary at Little Garie Point is 38 km from the CBD, much the same as the 39 km Greater Melbourne spreads in the west to Little River.

At present, the ABS estimates the populations of Greater Sydney and Greater Melbourne are 5.0 million and 4.7 million, respectively. If one suspects the ABS might’ve squibbed on including Wollongong and Geelong within the boundaries of Greater Sydney and Greater Melbourne for fear of outraging the locals, adding them reduces Sydney’s margin from 305,000 to 277,000. Not really a lot.

Bigger is better – small businesses once employed over half of private sector workers but no longer

Why have small businesses declined in relative importance? Many factors have probably contributed to the downward trend, including, among others…consumer preferences, with consumers preferring to shop at bigger stores with more variety and lower prices (as the bigger stores benefit from economies of scale)…

The story’s about the massive and rapid fall in small business’s share of jobs (in Qld, from 56% in 2007 to 44% in 2016), but it also illustrates one of the problems with the fashionable idea of the 20-Minute City (see Is the “20-minute city” mostly spin?).

Given travel by car is perceived as low cost, shoppers seek not only the economies of scale offered by malls and big box centres, but also the larger range of products and services they offer compared to smaller local centres. Large retail concentrations enable shoppers to compare prices and to buy multiple goods and services in the one trip. Then there’s air-conditioning, security, clean toilets, ease of parking, and more. They’re not for everyone, but it seems they’re for most everyone (see Do suburban shopping centres deserve all the hatred?).

Public transport is always greener on the other side

In Australia, public transport has to play catch-up constrained by an urban form designed by and for the car. This isn’t an impossible task, by any means, but it suggests that perhaps we’re comparing ourselves to the wrong cities.

What if we made more realistic comparisons? For example Portland, Oregon, is around the same size and has about the same population of Brisbane. It is held up as one of the “best transit cities” in the United States. Yet Brisbane has more public transport trips per capita – around 70 per year – than Portland, which has 58 per year.

The author, Dr Alexa Delbosc, followed up this week asking: “Which is more useful: to compare ourselves to cities with similar land use, or cities with amazing public transport but completely different land use?”.

Cities like London and Paris are useful aspirational models, but their land use pattern is very different to Australian cities. Even in the pre-car era, Australian cities developed at relatively low dwelling densities. Today, Sydney has a population of around 625,000 within the first 5 km radius of the CBD and Melbourne has approximately 430,000; in contrast, London has 1.3 million and Paris has 2.25 million (see Is public transport the only solution to congestion?).

However, as Dr Delbosc points out, our cities look quite good compared to north American cities. I’ve noted before that Sydney does better on active travel than transport poster-child Vancouver (see How does Sydney compare to Vancouver on travel?).

Doncaster hill’s next major apartment project verges on approval

Already holding the consent of Council planners, a positive result at next week’s Council Meeting will see 666 Doncaster Road add to the measured flow of high-density residential projects atop the hill. 

Doncaster Hill already has a growing concentration of high-rise residential buildings, yet it has no rail line or tram line, only buses. Conversely, an inner suburban train station like Dennis has very little high density residential development. The forces that drive development aren’t as simple as “build it and they will come”.

Council wants toll for $5.5b West Gate Tunnel to be set by minister, not operator

The council believes variable pricing could be one way to keep congestion in the CBD down – and wants the government to be able to impose it without having to renegotiate contracts with the $24 billion toll road operator. 

Tolls are set to maximise revenue, not to manage traffic flow. As I’ve noted before, the Government could implement a congestion charging tariff to manage traffic and compensate the operator for foregone revenue (see Is it time to get serious about road pricing?).

Sydney park to be named after Green Ban-era community leader Nita McCrae

The City of Sydney voted on 11 September to name a pocket park in Sydney’s The Rocks after the late community leader Nita McCrae. McCrae was the founder of the Rocks Resident Action Group, which mobilized in opposition to the NSW Government’s plans to redevelop the suburb as a high-rise commercial precinct in the 1970s.

Ms McCrae is doubtless a worthy person, but public bodies should learn the lessons from the international debate over statues and abandon the practice of honouring individuals by naming public spaces after them. As I noted recently, in a world where even Mother Theresa has feet of clay and the failings of the royals are continuously and forensically documented, there’s no one who mightn’t be beyond criticism as community views change over time. It’s especially contentious where the honouree reflects the political preferences of those currently in power (see Should public works be named after politicians?).

Transport - general

Aug 16, 2017


Distribution (%) of trip times for cars and for public transport, Melbourne (source data: VISTA)

It’s evident from the exhibit that the duration of weekday trips in Melbourne by car tend on average to be short while those by public transport tend to be long. For example:

  • 60% of car trips take less than 20 minutes, compared to 3% by public transport
  • 55% of trips by public transport take more than 50 minutes, compared to 12% of car trips.

The average weekday car trip in Melbourne is 20 minutes; the average weekday trip by public transport is 58 minutes. This difference is large even in the inner suburbs, where cars average 19 minutes and public transport averages 48 minutes.

The numbers are from the Victorian Integrated Survey of Travel and Activity (VISTA), a continuous survey that samples around 14,000 residents of Victoria. It’s important to note that the exhibit shows the proportion of trips by time category for each mode. It’s in percentages, so bear in mind the total number of car trips is more than eight times greater than the total number of public transport trips.

For example, while only 5% of car trips take 40-49 minutes compared to 17% of public transport trips, the number of car trips in this category is 484,650 versus 179,258 public transport trips. In terms of absolute numbers, Melburnians favour cars over public transport for all trip durations up to an hour.

That said, the difference shown in the exhibit has obvious and enormous relevance for the practicality of the “20-minute city”, which seeks to set a maximum trip duration by active modes, including public transport. The 20/30-minute city is embraced enthusiastically by politicians at federal and state level but it’s much more about clever politicking than good policy-making (see Is the “20-minute city” mostly spin? and Are the politicians trying to con us on this one?).

Of course, the current pattern of travel is in part a product of the existing transport system. Some argue that public transport isn’t used for short trips for the simple reason the available options aren’t good enough. I think there’s something in that argument, but it can’t be taken very far.

Consider that Melbourne has a relatively dense network of trams, trains and buses in the inner suburbs. Yet only 3% of public transport trips take less than 20 minutes and 2.5% are less than 2 kilometres. In a city with a claimed half a million tram boardings per day on average, a mere 35,000 trips of less than 20 minutes duration per day by all public transport modes isn’t primarily down to poor public transport.

While it’s mostly driving that’s the preferred mode for short trips (there are 5,328,014 car trips taking less than 20 minutes compared to 35,000 by public transport), walking out-competes public transport too. There are over forty times as many walking trips that take up to 20 minutes – and over sixty times as many that cover up to 2 km – as public transport trips. There are also three times as many bicycle trips under 20 minutes – and three times as many under 2 km – as public transport trips.

The key reason public transport is barely used for short duration trips is the time involved in walking to a stop and waiting for a service – even a frequent one – makes up a large percentage of what is a brief trip (the average duration for all trips up to 20 minutes is around 12 minutes). That makes it uncompetitive with the alternatives, especially driving, which requires no wait time, no timetable, usually no walk time at origin or destination, goes direct, travels at higher average speed, and is entirely private. I expect many short trips are also made when congestion is low and hence driving is especially attractive e.g. near home and/or outside peak hours.

Some of those who walk and cycle might choose to take public transport instead if what was on offer were better, but it’s fair to ask if that would be a better outcome and if the cost would be justifiable.

The lessons for policy makers are, firstly, that the greatest potential for public transport is in medium to longer time/distance trips and, secondly, that public transport won’t win a larger share of shorter trips unless car use is actively suppressed in some way. Restricting parking at rail stations is one way to promote use of feeder buses.

At a practical level, there’s a clear implication for the politician’s favourite i.e. the 20-minute city designed around active transport. The travel limit can’t sensibly be set by public transport or, as I’ve argued before, by bicycle; it must be set by the distance the great majority of residents are able to walk.

Cars & traffic

Aug 1, 2017


Population in the inner 100 sq km approx of six cities – Manhattan is the exception; it’s only 59 sq km, giving it a higher average density than central Paris

The Age editorialist reckons the $10 Billion Melbourne Metro rail tunnel now under construction is transformative infrastructure, but it’s not enough to deal with the city’s “congestion crisis” (Melbourne’s congestion crisis: The Metro rail tunnel is just not going to cut it):

A failure by successive Victorian and federal governments to adequately plan and to invest in public transport has created Melbourne’s most pressing and profound problem: congestion.

The writer thinks we should be dealing with congestion by building far more urban and regional public transport:

Multiple train line upgrades and extensions are needed… (The) ultimate solution to congestion is public transport.

The ultimate solution to traffic congestion is public transport, essentially rail? We certainly need to invest more in rail – a lot more – in order to handle projected population growth, but the solution requires more than that. Public transport is part of the solution but it can’t do it by itself:

  • Because building public transport infrastructure does not significantly reduce traffic congestion any more than building motorways does. The space vacated on roads by motorists who shift to the new train is soon taken up by other motorists i.e. induced demand.
  • Because the cost of retrofitting a public transport network that could attract all or even most travellers away from cars would be stratospheric. Paris has 303 stations within circa 5 km of the city centre; Melbourne has just 28 (see Can we build a metro just like the one Paris’s got?).
  • Because our cities are low density compared to the likes of successful transit-oriented cities such as Paris, Manhattan and London and we seem intent on keeping them that way. Paris has 2.25 million residents within the first 5 km; Melbourne has 430,000, most of whom resist higher densities (see How dense are our cities compared to Paris? and Is 16-storeys OK in the inner city?).
  • Because much larger and denser cities like Paris and Manhattan with outstanding rail networks are nevertheless still afflicted with serious traffic congestion. Drastically limiting car use is politically very hard in most places.
  • Because our multi-generational preference for private transport isn’t magically going to go away. Cars are currently much faster on average than public transport for all trips other than those to a few very dense places e.g. the CBD (see Is driving quicker than taking the train?). Autonomous cars promise to make driving even more convenient than at present (see What should we be doing now to prepare for driverless cars?). Electric vehicles powered by renewable energy sources will reduce significantly the environmental problems with cars (see Are electric cars a game-changer?).

The only way to tackle traffic congestion is to ration access to road space in some way. Some Chinese cities use odds and even number plate days without success; the obvious candidate is network congestion pricing. That can reduce the number of vehicles, increase average speed in the peak, and help make space for other road users like buses, trams and two-wheelers (see Is congestion charging just too unfair to bother with?).

The common argument is that congestion pricing can only be implemented if travellers have access to alternative public transport that’s as fast and convenient as driving. That condition would effectively rule road pricing out in Australian cities. However, it’s overly demanding; given driving is under-priced, there’s excessive car travel and hence no warrant for providing the same quantity of transit service. Moreover, motorists can shift the timing of trips to lower cost off-peak periods, or they can chain trips.

But reducing traffic congestion isn’t the only objective and arguably not even the main one; it’s a pity politicians and editorialists focus exclusively on it. The other key purpose is to increase access. New motorways eventually induce congestion at peak periods, but they increase the number of travellers who can get places, albeit slowly in the peak but faster in the off-peak. Public transport doesn’t “solve” traffic congestion either, but it excels at moving large numbers of people who want to go to the same place at the same time, albeit more slowly than driving in uncongested traffic.

If Melbourne really does grow at the projected rate (and that’s by no means certain), I think a number of key actions along the following broad lines are necessary:

  • Increase the supply and coordination of public transport; some new rail lines will be required but the weight of change must come from repurposing existing road space for high frequency buses and trams (see How can public transport work better in cities?).
  • Moderate demand by pricing access to the road network and encouraging vehicles to use motorways rather than high amenity streets. Given the scale of projected population growth, some increase in kilometres of motorway will likely be necessary.
  • Encourage a shift to more space-efficient private vehicles i.e. smaller, slower and kinder to others. The greatest promise is speed-limited electric scooters using dedicated road space. The priority shouldn’t be to eliminate all travel by private modes; it should be to eliminate large, fast, low-occupancy vehicles.
  • Require autonomous passenger vehicles to be shared (i.e. like a driverless taxi), powered by electricity, and charged by trip distance and time-of-day.
  • Reduce barriers to higher residential and employment densities in established suburbs, with greater intensity in locations close to high capacity public transport. Continue to permit well-planned incremental residential expansion at the fringe.

These actions should be supported by higher-level policy initiatives, especially shifting metro electricity generation to clean sources and removing taxation incentives that make housing as much about investment as shelter. The implementation would also vary spatially e.g. higher public transport mode share in denser areas; more private transport in less dense areas.

There are a few other claims in The Age’s editorial that I think warrant more explanation e.g. the idea that Melbourne Metro is “transformative”; and the charge that airport rail “has been forced off the agenda because of surging demand on the Sunbury and Melton lines”. The one I want to comment briefly on though is the claim that:

Evolving technologies will help ease Melbourne’s growing pains – ride sharing and driverless cars will probably reduce gridlock.

If they’re implemented on the current model of private car ownership – i.e. business as usual – I think it’s far more likely driverless cars will increase gridlock, not reduce it. That’s because they’ll significantly reduce the cost of travel. If you send your driverless vehicle up to the automated dispenser to pick up groceries while you sleep in, you probably don’t care too much if it sits in congested traffic for 30 minutes longer than it would off-peak.


Jun 30, 2017


(Image by Joseph Bsharah via So bad So good)


What’s Turnbull’s airport rail adventure all about?

The funding the Turnbull Government says it will provide to Victoria in next month’s budget for a rail line from the CBD to Melbourne Airport is about politics, not good policy

Does Sydney Airport really charge train travellers too much?

The assertion that rail travellers pay a $7.96 access fee to Sydney Airport on top of the Opal fare is wrong and lets the NSW Government off the hook

Is an airport train to the CBD the whole story?

The Victorian and Federal governments have different strategies for Melbourne Airport rail but there’s a lot more to the transport task at Tullamarine than a train to the CBD


Western Sydney: proposed metro/light rail network

Guest writer Dr Garry Glazebrook describes what a comprehensive plan for a metro rail network supported by feeder light rail services in Western Sydney should look like

Is all-night weekend public transport value for money?

Melbourne’s move to all-night public transport on Friday and Saturday nights is certainly appealing but there are many other ways such a huge sum of money could be spent

Melbourne ‘fantasy rail map 2070’

Adam Mattinson’s 2070 fantasy rail map incorporates almost every rail line ever proposed for Melbourne, as well as historic lines decommissioned during the 1950s and 60s

The Melbourne Subway Map (a rail fantasy)

A fantastical vision from mapmaker Adam Mattinson of what a subway underneath Melbourne’s inner suburbs could look like if it were as connected as Tokyo’s famous system

Should Doncaster BRT be a priority?

A private proposal to build and operate a Bus Rapid Transit system in Doncaster looks promising but taxpayers will ultimately pay for it; so it’s vital to make sure it’s a high priority


Is this (suburban) motorway a good idea?

The Andrews Government’s decision to build Melbourne’s North East Link lacks transparency and analysis, but the idea of suburban motorways shouldn’t be dismissed out of hand

Has Albo really found a congestion hollow log?

Anthony Albanese gets it wrong on traffic congestion. Australian cities need politicians who’ll tell us what’s really going on and what really needs to be done


What’s the heritage value of this office block?

Another fashionably handsome modernist building with debatable claims to social and architectural significance is the subject of a heritage battle in Melbourne


Is a cycling “freeway” the way to go?

The schematic proposal for a veloway the Andrews government says it’s going to build as part of the Western Distributor looks good but there’s room for improvement

If it’s not OK to walk it, why is it OK to cycle it?

A simple but shocking image showing how infrastructure purportedly provided for the benefit of cyclists, expects them to ride in situations they feel are dangerous

Should walking and cycling trails be like freeways?

Melbourne’s new Darebin-Yarra Link will be more like a freeway than a shared walking and cycling trail; it’ll be hard – and unsafe – for nearby residents to get access to it

How big are the public health benefits of riding to work?

Those who cycle to work have better health prospects than those who commute by car or transit but it’s unlikely all the public health benefits would scale up if a lot more of us cycled

A Bicycle Boulevard for Melbourne?

There’s an opportunity for the Victorian Government to recognise the key role Yarra Boulevard plays in supporting cycling in Melbourne by giving more of it to riders

A fantasy cycling map for inner Melbourne

Here’s master mapmaker Adam Mattinson’s vision of what a high-quality cycling network could look like in an Australian city. At present, unfortunately, it’s sheer fantasy

What are the prospects for dockless bike share in Australia?

Dockless bike share faces a much bigger challenge in Australia than in countries like China, especially given new entrant oBike must make it work commercially


Have the Premiers finally ‘gotten’ New Urbanism?

A visiting Canadian expert appears to have done what the locals couldn’t: inspired most of the Premiers to improve the health of city dwellers through better urban planning

Is Sydney the new New York?

The Sydney Morning Herald’s comparison of small parts of Sydney with leading world metropolises might grab the attention of readers, but it’s rubbish

Are Australia’s big cities dense enough?

The extensive suburbs in the middle rings of Australia’s east coast capitals explain why the “missing middle” is so much less dense than in comparable Canadian cities

Where’s the toilet?

The location of a humble toilet block in a park might seem a minor issue, but it’s big news in inner suburban Melbourne and highlights some larger issues

Can security bollards be done better?

It’s regrettable it seems necessary to install heavy bollards in Australian cities to protect against vehicle attacks. But it could be a way to improve public spaces

Why has walking to school stumbled so badly?

The proportion of children driven to primary school has raced ahead over the last 30 to 40 years. The standard solutions won’t change that by much, but there’s hope

Could the average trip to work nearly double by 2030?

The average one-way commute could increase by 28 minutes by 2030 according to Melbourne’s Herald Sun. Sounds horrendous but it’s scary tabloid journalism


Do residents of green places live longer?

It’s arguable whether greener residential areas reduce mortality but there are plenty of other good reasons to promote planting more trees, especially along streets


The outlook for infrastructure is improving

Guest writer Dr Garry Glazebrook reckons the federal budget shows the Coalition is at last heading in the right direction on infrastructure funding


Does the Grenfell Tower fire mean we should stop building residential towers?

The Grenfell Tower fire was the result of flawed policy on public housing, not some inherent flaw in the high-rise residential building type

Are towers inherently wrong for public housing?

St Louis’ famously demolished Pruitt-Igoe public housing project has a lesson for the debate over high-rise stemming from the Grenfell Tower fire; it’s not really about the architecture


Suburban sprawl or regional sprawl?

It’s probably a political winner, but it’s not obvious the Victorian Opposition’s new regional sprawl policy is preferable to suburban sprawl or redevelopment of established metro suburbs

See Urbanist Digest, Volume 1, 2017

Cars & traffic

Apr 27, 2017


Possible route options for the North East Link between Metropolitan Ring Road and Eastern Freeway/Eastlink (source: The Age)

Victoria’s Premier, Daniel Andrews, announced yesterday there’ll be $100 million in next month’s budget to start planning the promised North East Link motorway. The Premier emphasised the project will complete the “missing link” in Melbourne’s metropolitan ring road, running from Greensborough in the north-east to the Eastern Freeway/Eastlink in the south-east. It’s expected to take around ten years to complete, create more than 5,000 direct jobs, and cost up to $10 billion. It will be funded by a mixture of government contributions and tolls.

The key benefits, the Premier says, will be “more jobs, and less congestion in the north, east, and south”. Roads Minister Luke Donnellan added: “when we complete the ring road, you’ll be able to drive from Ringwood to Tullamarine without stopping at a single traffic light.” The RACV says it will improve freight movement across the metropolitan area.

Apart from the instinctive appeal of the “missing link” meme, the Government is relying heavily on the recommendation of Infrastructure Victoria, the independent “umpire” established by the Andrews Government in 2015. The organisation identified the North East Link as the next highest priority infrastructure project for Victoria. As I noted here, its preliminary analysis puts the benefit-cost ratio at 1.2 – 1.7, or 1.7 – 2.4 when Wider Economic Benefits (WEBs) are included.

Naturally there’s plenty of concern at the idea of another big motorway concurrent with the Government’s recent commitment to the West Gate Tunnel project. Memories of the East West Link debacle are still fresh too. Opponents believe the project will enable car-dependence; won’t reduce congestion; will increase emissions; will take funding away from public transport; and is bound to have a major impact on the environment. Some argue a public transport option would be a better solution for the region.

I think it’s appalling the Government has committed to a project of this scale without releasing to the public for prior debate a much more detailed analysis than Infrastructure Victoria was able to undertake within the limited timeframe it had available. This will inevitably be a case of developing a business case to support a decision already taken on limited evidence. I know that’s not unusual in our political culture and I know there’s bipartisan political support for the North East Link, but this is huge project and has the potential to be very controversial.


Despite the Premier’s claims, the project won’t do much in the medium term to reduce peak-hour road congestion. That’s due to the well-established phenomenon of induced demand. Yet governments still want to build motorways even while they’re also, as in the case of the Andrews Government, building major rail projects. So, is there a case for the North East Link? Are all motorway projects inherently bad? I want to revisit some points I made last year, when I argued there are some other aspects to consider in assessing the warrant for the North East Link (see Is this motorway obviously a really stupid idea? and Does the impact on trains kill the case for this road?).

While it will eventually congest, the North East Link will nevertheless reduce travel times significantly in non-peak periods i.e. during the day, at night, and on weekends. That’s especially important for freight movements. 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 will also provide a big increase in the number of vehicles that could travel across the region 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.

A large proportion of the capital cost will be paid by motorists via tolls. Tolling 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.

The motorway will reduce the social costs of crashes and improve the amenity of suburbs like Rosanna that currently experience high levels of traffic and truck movements. Notwithstanding the inevitability of congestion on the new motorway, it’s likely it will offer faster peak-period trips than maintaining the status quo.

Public transport isn’t a plausible alternative to the North East Link in this location. It’s an orbital motorway and it’s 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 currently 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).

There are big environmental risks as this is a sensitive 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 many of the risks can be managed. The scale and location of interchanges is likely to be critical.


In looking at a distinctively suburban road project like this one, it’s important to understand the great majority of Melburnians don’t live in the inner city or even the inner suburbs where road projects like the East West Link and West Gate Tunnel are proposed. And the great majority don’t work in the CBD where the standard of public transport is high and cars have limited utility.

Most Melburnians – existing and future – live and work in the low-density middle and outer ring suburbs where 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?).

So it’s vital to take a reasoned view of suburban road projects. There’s definitely a need for better public transport in the suburbs, but it won’t replace significant numbers of private vehicle trips; cars aren’t about to go away or, in the absence of comprehensive road pricing, become uncompetitive relative to other modes.

Apart from the dismaying lack of transparency and analysis, I think the biggest issue will be how the detailed design of the road evolves, particularly in response to environmental concerns. There’s potential for construction costs to increase and thereby compromise the benefit-cost ratio.

Education, justice, health

Apr 11, 2017


Time trends in the percentage of children and young people who use active transport to and/or from school (source: AHKA – Report Card 2014 on physical activity for children)

The exhibit illustrates a now familiar story; the proportion of children travelling to school on foot or by bicycle in Australia fell by 42% between 1971 and 2013. The great majority of primary age children are now driven to school. For example, 74% come by car in Melbourne and 3% by public transport. Of the 22% who travel by active modes, 19.6% walk and 2.6% cycle (see Surely the ’30-minute city’ makes sense for primary school trips?). (1)

What caused this precipitous drop-off in walking and cycling? The exhibit indicates this change isn’t peculiar to Australia or even to “new world” countries. It’s an international phenomenon, so it’s likely there’s a structural explanation. I’ll dismiss dubious arguments like the “kids of today are different”, leaving a range of more plausible explanations. Over the last 30-40 years we’ve seen:

  • Higher car ownership, now close to saturation
  • Increased female workforce participation
  • Higher traffic levels on streets
  • Increased fear of predatory strangers
  • Longer travel distances from home to school e.g. private schools, specialist schools
  • More after-school activities e.g. after-care, “hot-housing
  • Lower tolerance for risk.

Most of these changes are the result of higher living standards. The last one warrants further comment because it’s an under-appreciated phenomenon. Parents’ tolerance for risk has fallen substantially over the last 30- 40 years. This explains why parents often mention “stranger danger” and road safety as the main reason they drive their children to school even though, as active transport advocates often counter, the objective risks are still extraordinarily low. It explains a lot of behaviours; for example, the seemingly excessive fear of cycling on roads (see How dangerous is cycling?).

So, what can be done to increase the proportion of children who walk or cycle to primary school?

Some argue children should be compulsorily zoned to their local school, but the cost would be very high relative to the benefits (see What to do about schools and “rich switch”?). It would restrict parental choice on matters like religious and cultural education and it would limit the ability of schools to specialise or to benefit from economies of scale. Even then, while 35% of Melbourne children who live within one kilometre of school are currently driven , this rises to 78% for those who live between one and two kilometres.

The practical barrier is the car, both in terms of its attractiveness as a mode for the journey to school and as a potential source of danger for children who walk or cycle. It would be necessary to make driving less attractive by taking some road space away from cars and trucks and reallocating it to active modes. Further, vehicles would need to be “tamed” i.e. required to travel at much slower speeds and to give priority to vulnerable travellers (see Is it time for a 40 kmh speed limit in cities?).

New fringe suburbs reflect much of this thinking but I’m not optimistic about retrofitting the required changes to existing suburbs where in any year 98% of the population already live. It wouldn’t be easy; even cities like Paris and Manhattan with very high mode shares for walking and transit are nevertheless congested with traffic.

I expect the sort of compromises that can practically be made with drivers in Australian cities would yield walking and cycling options over time that’re reasonably satisfactory for adults (e.g. segregated on-road cycle paths, Quietways), but I fear they’d still be unacceptable to the risk-averse parents of young children. I think they’d still find the idea of their young ones crossing even moderately busy roads hard to accept. In any event, many parents would strongly resist restrictions that make driving less useful for them.

Hence I’m not confident big reductions in the mode share of driving for primary school trips can be achieved by promoting walking and cycling. But there’s another way.

Really serious gains are only likely to come from on-demand autonomous buses that carry children to and from home door-to-door. They’d address the concerns of parents by providing short waiting times and safe and secure vehicles e.g. access entry passes, video surveillance, perhaps volunteer monitors.

Operational costs would be low and they’d pay for themselves in terms of social costs many times over by reducing the number of cars ferrying children to and from school. Driverless buses wouldn’t help with exercise (they might even reduce active school travel in aggregate) but the key purpose of transport is getting places; there are other ways for parents and schools to protect the health of young children.

The technology isn’t ready today but it’s getting closer(e.g. see Driverless bus trial in Perth notches up 2,000 passengers). It’s likely to make a bigger and earlier impact than the messy business of relying solely on repurposing road space for active transport and trusting parents will let their children use it to walk or cycle to school.


  1. The pattern is different for high school students. While 18% (16.7% walk and 1.3% cycle), 36% travel by public transport. Less than half get driven or drive (see Does Turnbull’s ’30-minute city work for secondary students?).

Airports & aviation

Apr 10, 2017


New and upgraded transport links recommended by Infrastructure Victoria, including a heavy rail line to Melbourne Airport (source: Infrastructure Victoria)

The Herald-Sun reported on Friday that next month’s Federal budget is expected to provide some funding for a rail line to Melbourne Airport (Melbourne airport rail link boost: Turnbull Government to pour cash into long-awaited project):

The (Turnbull) government is understood to be willing to make a multi-year commitment to kickstart it. The funding would be part of a Victorian infrastructure package worth more than $1 billion, which would also finance upgrades to the Bairnsdale, Wodonga and Warrnambool railway lines. The money is expected to flow from a federal-state agreement to pay Victoria more than $1 billion it is owed from the asset recycling fund for the sale of the Port of Melbourne.

The Prime Minister’s motivation has nothing to do with good policy. The intent is to help State Opposition Leader Matthew Guy scratch the gloss on the suite of transport infrastructure projects Victoria’s Andrews Government is implementing in Victoria. These include Melbourne Metro, level crossing removals, Mernda rail extension, signalling upgrades and additional rolling stock.


Airport rail is an ideal political issue for the State Opposition for at least three reasons.

One, it’s a very popular idea. Although the airport has good public transport (SkyBus), it seems Melburnians are embarrassed by being among the handful of large world cities without a rail line to the airport. The Opposition recognised this long ago; it went to the 2014 election promising to build airport rail as part of its version of Melbourne Metro.

Two, it’s a point of difference with the Andrews Government, which relies on the advice of arms-length adviser, Infrastructure Victoria, that airport rail won’t be necessary until around 2030 at the earliest. The Government also contends that since the airport line is planned to join up with Melbourne Metro, it couldn’t in any case start operation until the latter is completed, expected to be in 2026. Premier Daniel Andrews supports the idea of airport rail but says the other projects his Government is pursuing are higher priorities (see Andrews government talks down Melbourne Airport rail as PM pledges cash).

Three, the “cost” to the Commonwealth for this exercise will be tiny in the context of the budget; probably less than a $100 million out of the 2017-18 budget. But it’s more than enough to put political pressure on Daniel Andrews. In any event, it’ll come from the $1.45 Billion “owed” to Victoria under the Asset Recycling Scheme as a reward for selling the Port of Melbourne.

It would be a bonus for Mr Turnbull and Mr Guy if the manoeuvre also leads to conflict between Victorian Premier Daniel Andrews and Federal Opposition Leader Bill Shorten. The Herald Sun reckons there might already be trouble (Opposition leader Bill Shorten says Melbourne Airport rail link is a ‘no brainer’):

The Labor leader has offered to work with Prime Minister Malcolm Turnbull on developing a preferred route for the train to Tullamarine after the Herald Sun revealed funding for the project would be included in the upcoming federal Budget.


There’s little opposition in Victoria to the idea of building a circa $3 – $5 Billion rail line from the CBD to the airport. Most everyone agrees it’s a good thing. The more ardent advocates make exaggerated claims about the costs and benefits, but the key difference is about timing. The Prime Minister’s contrived scheme doesn’t advance serious consideration of this issue, either way; his motivation is to win political advantage.

Timing – whether too early or too late – isn’t the only risk. Another possibility is that the status of Infrastructure Victoria as a source of independent evaluation of proposals could be severely weakened by Mr Turnbull’s political adventure. He’s already shown here he’s prepared to ignore its recommendations when it’s politically useful. Its standing could be weakened further if the Victorian Government caves-in to the political pressure triggered by his intriguing and ends up ignoring the “umpire’s” advice.

I’ve explained before that while it would certainly be a “nice to have”, the case for building airport rail right now isn’t compelling. There are other projects that should have a higher priority in the immediate future. In my view, Infrastructure Victoria has made the right call and the Government should continue to heed its advice (Is it high time Melbourne got a rail line to the airport?).

Nevertheless, it could help to defuse the political dangers if Infrastructure Victoria were to show more clearly how it arrived at its view on airport rail. It’s approach to date had necessarily been “broad brush”, but given the ongoing political machinations it should undertake a more detailed evaluation.

For its part, the Victorian Government needs to be less sanguine. It should show it’s actively investigating and planning for an airport rail line, as I’ve suggested before, so that it’s prepared to move earlier if circumstances change (e.g. see Government warned Melbourne Metro won’t support future airport rail link). It should also show it’s moving quickly to implement the recommendation made by Infrastructure Victoria to improve the existing bus service as an interim measure.

Deliver a high level of onroad priority to bus services linking Melbourne Airport to central Melbourne, including better signalling and managed motorway improvements, over 0-10 years. This will maximise the capacity, efficiency and reliability of these services and defer the need for a more costly investment in a heavy rail line to Melbourne Airport to the 15-30 year period.

It’s important though that further investigations are part of a comprehensive plan to improve access to the airport from all parts of Melbourne and Victoria, not just via the CBD. It should examine all modes, as well as options for reforming airport parking and road-pricing on key access routes (see Should Melbourne Airport rail be put on the front-burner?).

Cars & traffic

Dec 20, 2016


Cumulative benefits and costs including dependent infrastructure upgrades - upper project cost estimate (source: Infrastructure Victoria)
Cumulative benefits and costs including dependent infrastructure upgrades, proposed North East Link motorway – upper project cost estimate (source: Infrastructure Victoria)

The Age gave a lot of prominence on Saturday to a report that the Andrew’s government’s newly announced North East Link motorway would “shift 25,000 rail passengers a day to cars”.

The $10 billion North East Link will create “an estimated reduction in train boardings of 25,000 on an average weekday”, according to Infrastructure Victoria. That is the equivalent of shifting about one in 25 train journeys across Melbourne from rail to road. The government argues the road will cut congestion. But a report has found it will “compete directly with sections of the metropolitan rail network, principally the Hurstbridge, Mernda, Upfield and Craigieburn” railway lines, leading commuters to opt for cars.

This is a motorway so we should expect there’ll be more driving, but 25,000 sounds extraordinary given the North East Link itself will carry 100,000 vehicles a day. The preliminary economic appraisal prepared by Infrastructure Victoria acknowledges the claim, stating:

Analysis indicates that the project would contribute to a small increase in car trips and reduction in public transport trips across the network…This is reflected in an estimated reduction in train boardings of 25,000 on an average weekday.

So, is it as bad as it sounds? At the outset, it’s useful to make clear that few of the train travellers would literally shift to the North East Link. Infra Vic’s report doesn’t go into details, but it’s more likely most of them would start driving on the arterial network because the new motorway would draw traffic away from existing roads and so make them less congested. That’s unexceptional; the same happens with any substantial road improvement that changes the relative “cost” of rail vs road e.g. clearway extension, traffic light priority, level crossing removal.

Some other points to consider:

  • 25,000 boardings is a pretty small number in the context of the 1,073,438 public transport trips made on an average weekday in Melbourne (about 1 in 50 trips) and microscopic compared to the average 12,249,250 total weekday trips by all modes. One reader wrote to The Age yesterday; “I would suggest that the reported ‘reduction’ in train boardings with the North East Link is so small as to be absorbed into the estimation error of any travel forecast”.
  • Those who change from train to car would do so because it’d make them better off; they’d get a faster trip.
  • The motorway is in any event estimated to give public transport users $419 million (NPV) in benefits due to factors like reduced crowding and improved bus reliability.
  • The estimated benefits to users would significantly exceed the negative externalities associated with the motorway (by nearly 15 to 1), including those due to the reduction in train patronage.

There are always various positive and negative changes with any large project, whether road or rail. The key thing is to look at the overall net outcome of the entire investment.

Infrastructure Victoria says that after accounting for negative externalities including pollution, emissions, crashes, and foregone exercise from walking to public transport, the motorway is estimated to give a benefit-cost ratio of 1.4 – 2.4 (increasing to 2.2 – 3.1 when wider economic benefits are included, as they should be). It says while there’d be a small increase in car trips, aggregate time spent travelling on the road network would fall.

The headline “North East Link to shift 25,000 rail passengers to cars” doubtless triggered a reflexive and satisfyingly negative response from many in Fairfax’s target market, but as I wrote last week (Is this motorway obviously a really stupid idea?), it’s critical to understand this is the suburbs.

It’s extraordinarily hard for public transport to provide an attractive alternative to cars in a low density environment where activities are dispersed e.g. public transport’s share of all weekday trips is 9% in the middle ring suburbs and 6% in the outer suburbs (the figures for trains are 6% and 4% respectively). Moreover, the North East Link is for orbital travel – it’s part of the ring road – whereas Melbourne’s legacy rail system is radial, designed around the concentration of activities in the centre.

None of that means the North East Link is necessarily a worthwhile project; in particular, there are still significant environmental risks that need to be evaluated carefully. But as ever it pays to look at the entirety of any project.