The Andrews government’s proposed orbital rail route for Melbourne

The Andrews government announced yesterday that, if re-elected, it will start construction in 2022 of a 90-kilometre underground “loop” rail line running through Melbourne’s outer suburbs. The government claims it will cost up to $50 billion to build, carry 400,000 passenger a day, take 200,000 vehicles off the roads, and be completed circa 2051.

This is a blatantly political promise; the government would be in its third term before it was required to start putting up really serious money. Nevertheless, what’s not to like about “the biggest transformation of public transport in Australian history”? I’ve been banging on for ages about how the great majority of jobs and residents in Australia’s capitals are located in the suburbs, so it should be an appealing project, right (e.g. see The jobs are already in the suburbs?)?

Well, let’s see. There’s not much info to go on so far but I think there are a few things that should be said straightaway.

As expected, there’s no business case, no benefit-cost analysis, no account of how the patronage projection was derived, and no explanation of how the $50 billion was calculated. As is customary, the idea hasn’t been endorsed by the government’s independent adviser, Infrastructure Victoria, either.

Even if we (kindly) assume $50 billion is a reasonable estimate, it’s a huge sum of money to devote to a single project. It’s more than the estimated cost of building a high-speed rail line from Melbourne to Sydney. It would necessarily suck funding from other public transport projects and from other areas of the budget, especially health and education.

And even if we (credulously) accept the government’s claim that demand would be 400,000 one-way trips per day by 2051, that’s a tiny pay-off for the scale of investment required. Massive investments in public transport generally have a small effect on mode shift because they mostly replace existing public transport services (e.g. see Will simply building more public transport seriously suppress car use? and Should cycling get a huge increase in funding?).

There are around 13.5 million trips every day in Melbourne at present; if the population doubles by circa 2050 as projected, the new line will account for around 1.5% of metropolitan trips on an average week day. Most of those travellers would otherwise have travelled on other public transport services e.g. Melbourne’s existing high frequency suburban orbital SmarBus services, Route 901Route 902Route 903. Sure, they aren’t as comfortable as a train and not as fast, but they only cost around $20 million per orbital route to set up.

The key issue here is that the suburbs aren’t like the city centre, which generates most existing public transport use in Australia’s capitals. Driving is a much more competitive option in the suburbs because congestion is lower, parking is easier, and most trips are short. The government could make public transport more competitive for these kinds of trips if it made driving less attractive by (say) implementing road pricing as recommended by Infra Vic, but it’s already ruled it out.

Travellers will choose to drive while ever it out-competes public transport. For this reason, I think the claim that the new line will take 200,000 vehicles off the road should be taken with a grain of salt. But even if it’s taken at face value, Daniel Andrew’s proposition is to spend $50 billion to remove less than 1% of metro driving trips by 2051. Not only is the pay-off modest relative to the outlay, there’s a good chance that by then most of those trips would otherwise be in electric vehicles powered largely by green energy. And in the absence of road pricing, the road space vacated by any drivers shifting to the new rail line will of course be taken by other vehicles.

Betting the entire basket on a single mass transit line makes little sense in the suburbs where both travellers and destinations are dispersed at low densities. I don’t have data for Melbourne, but in Sydney 76% of those who travel by train live within one km of a station. It would make much more sense to have multiple orbital routes spaced to be within (say) one kilometre of all residences. The only way that’d be financially feasible would be to use on-road services like trams and buses. These modes could be improved by providing routes every two km (say) from the city centre, scaling up frequency and vehicle size in line with demand, and providing both dedicated road space and automated priority at intersections.

The government stresses how the line connects 15 suburban centres. I can confirm the proposed alignment includes three of the largest suburban employment concentrations in Melbourne i.e. Clayton/Monash, Tullamarine and Box Hill. All up, the line connects half of all jobs in suburban centres. But most suburban jobs aren’t in centres; only one fifth is in the 31 largest centres, with the great majority relatively dispersed.

Nor has the case hasn’t been made that there are existing or potential flows between the 15 centres that are so compelling they justify upgrading any sort of transport connection, much less a mass transit rail connection that costs $50 billion. Why, for example, is underground rail necessary between Doncaster and adjacent Heidelberg? Or between three minnows like Reservoir, Fawkner, and Broadmeadows?

Another key issue is that that most of the large suburban “centres” are low density compared to the CBD and cover an extended area; for example, the largest suburban employment concentration, Clayton/Monash, covers an area of 10 sq. km. While it makes sense in the small and ultra-dense CBD, a single high capacity rail line isn’t the appropriate way to approach public transport in Melbourne’s mostly sprawling suburban centres.

There are many other projects that should have higher priority for funding than this one. In terms of rail, they include more rolling stock, Metro 2, airport rail, electrifications and extensions, line duplications, signalling upgrades, level crossing removals, and more. There’s an even more pressing need to expand bus and tram services and build related road works to increase priority.


Suburban public transport is crying out for improvement, but the suburbs are not the CBD or the inner city. They’re low density and even the largest job centres are tiny in terms of job numbers and sprawling in terms of geography in comparison with the city centre. Driving within and across the suburbs is a lot more competitive with public transport than driving to the CBD. While ever governments refuse to countenance policies (like road pricing) that make driving less attractive, the demand for public transport for cross-suburban trips will continue, at best, to be modest.

A glamorous solution like a single high-capacity rail line would consume a generation’s worth of public infrastructure spending for essentially no gain. And it’s not necessary. Irrespective of cost, a much better solution would be to provide a multiplicity of parallel orbital bus or tram services that put all suburban residents within reasonable walking distance of high frequency public transport (e.g. see How can public transport work better in cities?).

Once the monumental cost is considered, the government’s proposal is lunacy. This is not a “visionary” project as the gullible would have us believe; it’s cynical, self-serving and massively wasteful.