Major suburban employment clusters in Melbourne by size and mode of journey to work (source: Charting Transport)

The Andrews government emphasises how its promised $50 billion Melbourne rail loop will link 15 suburban employment centres and thereby catalyse the development of a polycentric urban form, bringing jobs closer to where workers live. The government says the line will:

form a ring through Melbourne’s suburbs, connecting the Monash, La Trobe, Sunshine and Werribee National Employment and Innovation Clusters (NEIC) with key precincts such as Box Hill, Burwood, Broadmeadows and the Airport.

There’s no doubt the project will have an impact on commuting patterns, but will it be a significant impact? Most importantly, will the effect be commensurate with the monumental scale of spending required to build the line? Will it justify prioritising spending on the loop over other transport priorities, or over capital works in other key budget areas like education and health?

I think there’re several reasons why the loop is unlikely to have a big impact on suburban job centres or on commuting patterns.

Few major centres, few jobs

First, only a tiny proportion of Melbourne’s jobs are in suburban centres. A recent analysis by Chris Loader from Charting Transport identified the number and size of employment clusters in the suburbs of Australia’s capital cities, using minimum thresholds for the number of jobs and employment density.

Although he used “fairly low” thresholds, Mr Loader found there are only 9 suburban employment clusters in Melbourne that meet his criteria. These major centres don’t contain many of the city’s jobs; collectively, they account for a mere 5% of all employment in the metropolitan area. Compare that with the 28% of metro jobs located within 4 km of the city centre

Most suburban jobs in Melbourne are dispersed; 93% of those located more than 4 km from the city centre are in diffuse locations e.g. stand-alone addresses, business parks, or small centres.

Don’t assume there’s a bias against centres in the methodology; Mr Loader found Sydney has 24 major centres. He found they collectively account for 19% of all metro jobs i.e. nearly four times as much as in Melbourne. While the two cities are close in population, there are important structural differences e.g. in topography, industry structure, and the capacity (‘elasticity’) of the CBD/inner city to grow.

Three major centres on loop

Second, just three of the major suburban job clusters identified by Mr Loader are on the route of the promised loop i.e. Clayton, Box Hill and Heidelberg. They’re the biggest major centres, but collectively they still only account for around 4% of metro jobs.

Centres mostly minnows

Third, most of the other 12 centres on the loop are relative minnows in terms of job numbers. My own analysis shows that while the cluster comprising the airport and surrounding industrial area accounts for 1.5% of metropolitan jobs, all the other centres each have fewer than 1% of jobs (see Where are the suburban jobs?).

For example, Doncaster has 0.1% of metropolitan jobs and Sunshine, Broadmeadows, Burwood and La Trobe each have 0.2%. Although I use very low thresholds for total jobs and employment density (much lower than those used by Mr Loader), three of the centres on the loop – Cheltenham, Fawkner and Werribee – don’t qualify.

Density of centres is modest

Fourth, the 15 centres on the suburban loop are all a far cry from the CBD, which accounts for most commuting by train in Melbourne. The expanded CBD, including Docklands and Southbank, is four times denser than Box Hill and seven times denser than Clayton. The Hoddle Grid is nine times denser than Box Hill and twelve times denser than Clayton.

Centres have big footprint

Fifth, a number of the 15 centres are geographically expansive, making them less suitable for high-capacity transit systems. The draft planning framework for the Clayton NEIC covers an area of circa 16 sq. km. Even the Sunshine NEIC, although tiny in terms of jobs, extends over approx. 15 sq. km. Similarly, La Trobe NEIC is about 13 sq. km. In contrast, the CBD rail loop encloses an area of 2 sq. km. and services it with 5 stations.

Most centres already on rail

Sixth, 13 of the 15 centres on the promised suburban loop are already on radial rail lines. Yet these 13 haven’t grown into massive job hubs. In fact, the largest suburban cluster, Clayton, has developed despite very poor rail access. This suggests that even radial lines connected directly to the massive job concentration in the city centre aren’t a silver bullet for growth or density in suburban centres; there are other factors that matter more. It also implies an orbital rail line that mostly connects minnows with each other is unlikely of itself to be a catalyst for significant jobs growth.

Major centres not increasing share

Seventh, the argument that the suburban centres will grow dramatically in the future isn’t convincing. Melbourne got much larger over the last 50 years and for much of that period jobs grew faster in the suburbs than in the inner city. Yet the proportion of suburban jobs in centres declined for most of the period. That’s in line with the experience in American cities too.

More recently, the suburbs have lost out to the inner city notwithstanding spectacular population growth across Melbourne as a whole. The proportion of jobs within 4 km of the centre increased from 26% in 2006 to 28% in 2016. The proportion of jobs in the suburbs correspondingly declined and, importantly, the share of metro jobs in major suburban centres remained static at 5% over the period.

The growth of the inner city reflects the continuing transition to the knowledge economy and the importance of factors like the demand for face-to-face contact in business. While Melbourne’s suburban centres remain important places for local population-serving businesses and space-intensive industries, they aren’t proving attractive for firms that value the level of density that justifies mass transit. They have better options.


Melbourne isn’t a polycentric city; rather, it’s mostly a dispersed city. It has a dominant inner city with 28% of jobs, but the great bulk of employment – 72% – is in the suburbs. And contrary to the received wisdom, 93% of those suburban jobs are scattered; only a small fraction of suburban jobs are in major suburban centres (just 7%; or 5% of metro jobs).

Costly high-capacity systems like underground heavy rail make obvious sense in the CBD with its huge concentration of activities at very high densities. Commuting within Melbourne’s geographically extensive suburbs, however, requires a transport system that deals more effectively with low levels of concentration than mass transit. That’s been the domain of unrestricted car use, but with increasing congestion there’s a pressing need for other solutions, like road pricing and more flexible transit systems (e.g. see Isn’t there a much, much better way to do cross-city public transport?).

There might be other justifications for spending $50 billion on a suburban loop rail line (although I think that’s questionable – see Has Daniel Andrews gone loopy on rail?), but it’s doubtful if growing suburban centres is one of them.