Change in average journey-to-work travel distance by region (km). Data from BITRE

The Age says jobs in Melbourne are losing pace with sprawl – it cites a new study by BITRE which predicts “an increase in the average commuting distance” by 2026 and a rise in journeys to work involving a road distance of more than 30 kilometres.

If a rigorous, hard-nosed body like the Bureau of Transport, Infrastructure and Regional Economics is saying things are going to get worse in the future, it’s worth sitting up and taking notice, right? It’s true BITRE does say that, but it’s also true the media tends to err toward a sensational rather than a sober interpretation of any given facts. In this instance the story is a bit of a beat-up.

For a start, it’s hardly news that commuting distances could “increase” over a period of 15 years given the spectacular growth in population projected for Melbourne. What matters is the size of any increase – if it’s only a 1% increase over the entire period, that’s an infinitesimal 0.06% p.a. However if it’s (say) 15%, i.e. 1% p.a., that’s worth taking note of. However The Age is silent on this score.

BITRE doesn’t say anything about the size of the predicted increase either. There’s a good reason for that. BITRE’s study isn’t an authoritative prediction of future commute distances as implied by The Age’s story. It doesn’t make forecasts based on the latest data, using innovative modelling techniques and complex algorithms as one might expect. In fact the report isn’t even about the future! – it’s actually about historical population, employment and commuting patterns in Melbourne up to 2006.

The Age relies on what is in effect an ill-advised throwaway line by BITRE. The report states (p 333) that if the Victorian Government’s spatial projections of population and employment through to 2026 are realised, the likely commuting implications include….”an increase in journeys to work involving a road distance of more than 30 kilometres and an increase in the average commuting distance”. There’s no analysis or supporting information behind this assertion, so too much shouldn’t be made of it. The prominence given to it by The Age suggests BITRE should’ve thought a bit harder before including it in a report about the past and the present.

However what BITRE actually has analysed in-depth is the historical change in travel distances – and here the picture is if anything somewhat mixed. The report looks first at what’s happened over 2001-06 (see exhibit). That isn’t necessarily a guide to what will happen in 2026, but it shows how current patterns are trending. The picture it reveals isn’t one of rampant increases in commute distances but rather one of relative stability.

BITRE found the average commute in Melbourne increased from 14.7 to 14.8 km, or by just 100 metres over five years. That’s a 0.7% increase, or a miniscule 0.1% p.a. Surprisingly, the average commute increased proportionally less in the outer suburbs than in the inner city – in fact as the exhibit shows, the average commute shortened in absolute terms in the Outer South, Outer East and the Outer West.

This is the real news! It’s important because commute distances have historically increased significantly, while commute times have remained relatively stable. So reliable evidence that commute distances have stabilised, even for five years, is noteworthy.

Some caution is needed though. While not as reliable as Census data, ABS surveys suggest the average commute distance in Melbourne may have increased at a faster rate since 2006. If so it’s not clear why – perhaps it reflects the greater use of trains for the journey to work over the last five years? It’s a pity BITRE chose to undertake this extensive study so close to the release of the 2011 Census results, as the latter would’ve provided more up-to-date information.

Overall though, there’s no objective basis for The Age’s story. I’m even more mystified by the paper’s claim in its opening para that the number of people driving more than 30 kilometres to work in Melbourne will increase in coming years.

I can’t find anything in the BITRE report to support that contention. Apart from having very little to say about the future, the report doesn’t disaggregate the historical data on commuting distance by mode. My best guess is The Age has interpreted BITRE’s use of the term “road distance” to mean driving, whereas BITRE probably (somewhat clumsily) meant non-Euclidean distance.

There’s still the related issue of whether commute distances are simply too long. I’ve looked before at commute times, but not specifically at distance – I’ll do that another time.

Animation of Tokyo from farmland to megalopolis - click and wait a second

It might seem intuitively obvious that any comparisons between cars and public transport should be on a “per kilometre” basis. After all, as Steven Smith at Market Urbanism points out, “people take trips of varying length, and longer trips are more expensive than shorter trips, so the desire to standardize and compare makes us want to simply divide the trips by their length and call it even”.

However here are four observers of the US transport and planning scene who all say the concept of comparing transport modes on a per mile basis (or more correctly, per passenger mile basis) is deeply flawed. Steven Smith says both supporters and opponents of light rail use per passenger mile costs and subsidies to justify their positions, but the problem is that the purpose of transit is not to travel long distances:

These are not pleasure travellers trying to get as far from home as possible, but rather commuters trying to get to wherever their jobs and schools are located. But the distance to this “somewhere” is not a variable to be held constant – it actually varies with population and job density, which is highly correlated with mode of transit. Places with train lines generally have and allow for denser development and thus less distance between your house and your workplace or school – the difference in average commute distance between urban and exurban areas could be as much as an order of magnitude.

Michael Lewin reckons per passenger mile comparisons are flawed because they assume that trips involve the same mileage on any mode. However in the real world our choice, he says, is not between a city with a 20 mile bus commute and one with a 20 mile car commute:

Rather, our choice is: do we want to make cities more compact, thus increasing the number of short commutes (some of which will typically involve transit, for the reasons stated above) or do we want to create a relatively spread-out city with lots of long commutes (most of which will usually be by car)?

In the compact city, fewer passenger-miles will be travelled, which means that all the negative externalities of travel (e.g. pollution, collisions, public costs) will be lower. And because people will be somewhat more likely to use transit and carpool, both cars and transit vehicles will be more fuel-efficient, because cars and buses are more fuel-efficient when they have more passengers. By contrast, in the car-oriented, spread-out city, both car and transit commutes will typically be longer, and both cars and buses will have fewer passengers.

Alon Levy at Pedestrian Observations is annoyed that subsidies for roads look much lower when they’re divided by the appropriate number of trillions and expressed in terms of passenger miles of travel. He says:

Passenger miles don’t vote. They’re not a unit of deservedness of subsidy. They’re one unit of transportation consumption. They’re like tons of staple as a unit of food production, or calories as a unit of consumption. We don’t subsidize food based on cents per calorie.

Even as a unit of consumption, there are flaws in passenger miles as a concept, when it comes to intermodal comparisons. The reason: at equal de facto mobility, transit riders travel shorter distances than drivers….. Transit is slower than driving on uncongested roads, but has higher capacity than any road. In addition, transit is at its best at high frequency, which requires high intensity of uses, whereas cars are the opposite. The result is that transit cities are denser than car cities – in other words, need less passenger miles.

Matt Yglesias at Think Progress reckons the concept of passenger miles is borderline incoherent and senselessly biased in favour of auto-oriented road projects:

The use of passenger miles as a unit of measures embeds the assumption that the goal of a regional intra-urban transportation system is to have people travelling as far as possible. Now you could imagine a city in which individuals, firms, structures, natural resources, etc. are just strewn about at random. If that was the case, then you probably would want to organize transportation to maximize distance travelled. People would have arbitrary transportation needs, might need to get very far away, etc. But when you’re talking about a real growing city, a focus on passenger miles just implies a focus on spreading your urban area out as widely as possible.

Do these guys all make sense? I think so. Passenger miles (kilometres) are not a sensible way to compare modes. The fact that a mode endogenously fosters shorter trips should be recognised otherwise it will be put at a disadvantage. The appropriate measure is passenger numbers or person journeys. Of course these views all pertain to the US where public transport trips tend to be short and car trips long.

There’s a sting in the tail, though. In Melbourne, it’s the reverse – car trips are shorter. For example, according to the VISTA travel survey, the median distance for non-work trips is 8.5 km by car, compared to 14.9 km by public transport. These non-work trips account for around 70% of all travel in Melbourne (Vista says median work trip distances for the two modes are the same). This is also another reminder of the need for care when comparing between cities and across countries.

Where residents of Brimbank travel to (%)

The stereotype of people travelling long distances in Australian cities is wrong but persistent. The reality is that most trips are relatively close to home.

For example, the accompanying chart and map show that 52% of all weekday trips (all purposes, all modes) by residents of the middle suburban municipality of Brimbank in Melbourne’s west are made within the municipality. Further, 79% of trips are made either within Brimbank or to contiguous municipalities.

In fact more than 90% of trips by Brimbank residents can be accounted for if just one more destination – the city centre – is added to the list above.

This pattern also holds for the other parts of Melbourne.

More than 80% of trips by residents of Monash, Cardinia and Casey, for example, are likewise made within their home municipality or to neighbouring ones.

While another 8% of trips by Monash residents are to the city centre, the corresponding figure is less for far-flung municipalities – just 4% for Casey and less than 1% for Cardinia (see more charts below).

It thus makes sense to think of a city like Melbourne as a number of regions rather than as a very big, singular entity. In terms of what people physically do within the urban area, there are multiple ‘Melbournes’.

Each little ‘Melbourne’ or region is centred on a home LGA and has a limited ‘home range’. The median weekday trip distance (all purposes, all modes) for residents of Cardinia is 3.7 km. Monash is 4.9 km and Brimbank is 6.3 km.

For the great bulk of residents, metropolitan Melbourne is more of a construct – an idea – than something that has a real presence in their day-to-day lives. With the exception of the city centre, few people venture much beyond their own region.

This limited ‘home range’ is a product of many forces. In the case of the inner city municipalities – Melbourne, Yarra and Port Phillip – the predominance of professional jobs and social attractions in the city centre is an obvious and powerful factor. The median weekday trip distance for residents of Yarra is 2.8 km.

For the great bulk of the population who live in the suburbs, the factors explaining this limited geography would include the suburbanisation of jobs (more than 80% of jobs in metropolitan Melbourne are outside the CBD), slower travel speeds on roads due to increasing traffic and the desire to live close to family and friends.

It’s natural therefore to emphasise the importance of regions and devise typologies like The Age uses in its real estate pages e.g. inner east, outer east. But that assumes a community of interest at the regional level. And it assumes it can be defined within fixed boundaries.

There are 31 municipalities in metropolitan Melbourne and they each display a localised pattern like the ones mentioned above. There are, in other words, 31 unique and overlapping ‘regions’. Even contiguous municipalities have very different aggregate ‘home ranges’.

Consider, for example, the adjoining outer suburban municipalities of Cardinia and Casey. Just over 86% of weekday trips by Cardinia residents are made within the municipality itself; to neighbouring Casey; or to locations beyond the metropolitan area. Casey however is quite different. 86% of trips by its residents are to Gt Dandenong, Monash, Melbourne, Kingston, Cardinia and, of course, within Casey itself.

31 regions is of course too many to be useful in most government programs and the idea that they are not exclusive – i.e. they overlap – would make any administrator tear his or her hair out.  Nevertheless, understanding travel patterns should give those who rely on fixed regional definitions some pause for thought about the limitations of their frameworks. There is scope to use travel data to improve the functionality of regional definitions.

The bigger issue is that localised travel has major implications for how these trips might be made. Short trips make public transport less competitive, partly because they tend to be dispersed and partly because waiting time becomes a larger proportion of total journey time. They promote travel by car because of its ‘on demand’ convenience, but also favour walking, cycling and travel by scooters and motor bikes.

Below: what other municipalities do residents of Monash, Yarra, Cardinia and Casey municipalities travel to (percent)? Note: travel data is from VISTA.

The issue I have with satellites is they’re O.K. if they have plenty of local jobs or if workers commute by public transport to the nucleus or host city, but they’re a very bad idea if neither of these conditions apply.

Melton was made a satellite city in 1974. According to historian David Moloney, satellite cities were a response to “urban quality of life issues: large cities and unrelieved urban sprawl were seen as too congested, uncongenial and economically inefficient”. They were, he says, a product of the rise of the town planning profession in the 1960s.

The Shire of Melton is in two parts. The main part with a population of around 40,000 is Caroline Springs – it is contiguous with the metropolitan area. Melton township is a further 9 km to the west and separated from Caroline Springs by green wedge.

Melton and Caroline Springs. The area shaded brown is the Investigation Area for possible future urban development. It would make sense to "fill-in" the gap. (click to enlarge)

Melton township has a population of around 40,000 (see here) but it does not have a commensurate level of economic activity. Approximately 60% of the workforce is employed outside the township (see here). Some work outside the Melbourne Statistical Division but most commute across the green wedge to ‘mainland’ Melbourne. As only 8% of Melton Shire’s workers commute to the city centre (City of Melbourne), the vast bulk of trips made by township residents, whether for work or otherwise, are made by car.

The median journey-to-work distance of Shire workers is 33 km – considerably longer than the 20 km combined median for six other outer suburban LGAs (Wyndham, Frankston, Mornington Peninsula, Hume, Casey and Cardinia).

This pattern is true for all trips indicating that it is a general characteristic of the Shire. For example, the median weekday distance travelled by Shire residents (all modes, all purposes) is 42 km, whereas that for residents of the other six outer suburban LGAs is 27 km.

Further, only 44% of weekday trips originating in Melton Shire (all purposes, all modes) also terminate within the LGA – the comparable figures for Cardinia, Wyndham and Casey are 68%, 63% and 58% respectively. This is unusual because outer suburban LGAs are usually relatively self-contained.

I don’t have data specifically for the residents of Melton township, but I expect their travel distances would be even longer than those for the Shire as a whole. So why do the the citizens of Melton township travel such long distances? There could be a number of factors at play here, each of which is related at least in part to the township’s satellite geography.

First, the township has a serious deficiency in jobs. When the decentralisation programs of the 70s were designed, manufacturing was thought to be the ideal industry for regional centres and satellites towns because it was less dependent on proximity to the city centre than other industries. Manufacturing employment has of course fallen precipitously since the 70s. In later years back office functions like call centres were also prospective candidates but there were many prospective locations for these jobs, both within Australia and overseas.

Even so, it is difficult to explain why a town of 40,000 people has such a poor workforce/jobs ratio. It might be that Melton township is too close to the metropolitan area to offer all the advantages of a regional location (e.g. low labour turnover) yet too far to offer all the advantages of agglomeration (e.g. access to a large labour market). At the very least this hypothesis warrants further examination.

The second explanation for such long travel distances is that the residents of Melton township have to travel across the green wedge just to be on the same footing as ‘mainland’ residents. That is a minimum extra distance of around 9 km from the eastern edge of Melton to the western edge of Caroline Springs.

Third, Melton township’s isolation from the main fabric of the city and its high standard freeway access mean that traffic congestion is lower and speeds are higher. Residents drive further because they can. In suburbs that grow incrementally at the edge of the main metropolitan area, subsequent waves of growth add to traffic on roads and discourage residents from journeying long distances.

A view of the Green Wedge separating Melton from 'mainland Melbourne'. (click to enlarge)

In Melton township’s case it might even be difficult for local service businesses to get a foothold if residents can easily drive to established suppliers in ‘mainland’ Melbourne. Those suppliers can in turn offer the benefits of economies of scale that make it hard for any start-up in Melton to compete.

I’ve little doubt that in due course the township will achieve a much more balanced ratio of local jobs to workers. But a lot of time has already passed and there might still be a long wait. A positive sign is that the Investigation Area for possible future urban development shown in Melbourne @ 5 Million envisages the possibility that Melton might be connected via development with Melbourne ‘proper’.

The lesson to draw from the Melton experience is that the natural advantage of proximity to the established metropolitan areas should never be discarded lightly. The underlying drivers of Melton’s successes and failures warrant detailed study – there are bound to be some powerful lessons for how growth should be handled.