Possible alignment (Blue) for Doncaster rail line along Eastern Freeway median. Source: Doncaster Rail Study

Here’s some great news. A consulting team led by Curtin Uni’s Peter Newman has apparently found a way to reduce significantly the stratospheric cost of new urban rail lines in Australia’s capital cities.

Professor Newman reckons a 12 km rail line running down the central median of Melbourne’s Eastern Freeway from inner city Clifton Hill to suburban Doncaster could be built for a mere $840 million. The costing includes five stations and 2-3 kilometres of tunnel.

And for another $300 million, the team reckons a further 3 km of tunnel could be constructed between Clifton Hill and near-city Parkville. And it could fund itself! A tax on the increase in property values arising directly from the rail line would raise enough revenue to fund construction.

Compare it to the likely $5-$7 billion cost of the proposed Melbourne Metro, and Doncaster rail is a bargain. Compare it to the $560 million the Brumby Government spent to extend the Epping Line just 3 km to South Morang and it’s the bargain of the century.

Around half of that South Morang money went on indirect works, but even so the comparison is extraordinary. Professor Newman and his team appear to have found the gunzel’s equivalent of Valyria.

Professor Newman is a credible authority. He’s on the board of Infrastructure Australia and, according to The Age, was the architect of WA’s fabulously cheap 70 km Mandurah rail line (circa $2 billion in today’s dollars). He’s also done this sort of work before – in 2004 he was part of the team who prepared a feasibility study for a rail line to suburban Rowville in Melbourne’s south-east.

I’ve argued before that a rail line to Doncaster can’t be justified for a whole host of reasons (here, here and here). At this price however, I’d have to seriously consider eating my words. The Age’s leader writer also thinks it’s a great idea (no surprise there, though).

The absence of a rail line is a longstanding issue in the Doncaster region. The Baillieu Government promised during the 2010 State election to undertake a feasibility study and it’s now underway and well advanced. The separate Newman study, which was commissioned by the City of Manningham and five other municipalities, is no doubt designed to put pressure on the Government to commit to the line.

Unfortunately I haven’t read the Newman report because it’s still secret. It was splashed across the front page of The Age earlier in the week, but the six municipalities won’t even meet to consider it until Tuesday.

I must say I think it’s appalling that extravagant claims with important public policy implications are made publicly without the supporting information being made available for examination. However since I’m assured by the City of Manningham that the facts reported by The Age, including the costs, are perfectly consistent with what’s in the report, I’ll press on.

The trouble with those reported costs is they’re too good to be true. The most credible estimate we have for the Doncaster rail line was prepared by the Eddington Task Force in 2008. It put the price at $1.8-$2.1 billion. Allow for optimism bias and cost increases and it’s likely to cost $3 billion and counting just to get the line from Doncaster to Clifton Hill.

There’s no spare capacity from Clifton Hill to the city, so an additional 3 km tunnel would have to be constructed to Parkville where it would meet up with the proposed Melbourne Metro. To suggest that would cost a mere $300 million in the current cost environment is……well, to be polite, too good to be true.

As another point of comparison, the Federal Government’s High Speed Rail feasibility study estimated it would cost $1 billion just to build underground station facilities at Southern Cross.

The reason the cost estimates are optimistic is they’re reportedly referenced from WA’s Mandurah rail line, which commenced operation in December 2007. While they’re presumably inflated to today’s dollars, this method is just plain disingenuous because the Mandurah line is a classic outlier.

It’s not clear why Mandurah cost so little compared to more recent projects. It might be because it was built on sand. Or because much of it was constructed through relatively undeveloped country. Or because it was built with a freeway. Or maybe some of the contractors took a bath.

Like most, I suspect a key reason is it had the good fortune to be tendered before the full impact of the resources boom fed through to construction costs (construction commenced in early 2004). However costs have sky-rocketed since Mandurah was built.

Why that is so is an extremely important question and one I’ve discussed on these pages a number of times before (e.g. herehere and here). But the reasons appear to be structural so they can’t just be wished away. Mandurah is simply not even remotely representative of current costs.

There is in any event enormous variability in the cost of projects both at the national and international level. Basing costs on just one project in another State tendered quite some years ago simply isn’t a valid approach, especially when that one project is cherry-picked.

However that isn’t likely to worry the six municipalities who funded the study. As it’s a political stunt, they just want to pressure the Government.

I’m not sure they support taxing the increase in property values though. I like it, although I doubt there’ll be as much value to capture as Professor Newman and his team assume. But it’s too politically difficult to be anything but a liability for most politicians.

There are other problems with a rail line to Doncaster which I’ve discussed before (here, here and here). For example, it would replace one form of public transport (SmartBus) with another (train). I’ll look at these other issues again in the near future when the Newman report is (hopefully) released publicly.

However it’s worth noting some pertinent findings from the recently released Rowville rail stage 1 feasibility study undertaken by the Victorian Government. The proposed Rowville and Doncaster lines have a number of similarities e.g. in length, stations, proximity to existing lines.

Whereas the Newman study claims a Doncaster line would carry 100,000 passengers per day, modelling indicates the Rowville line will only carry 68,000 per day by 2046. That suggests the patronage claim is ambitious, to say the least.

The really interesting bit though is the Rowville line would increase the share of all trips carried by public transport in the metropolitan area in 2046 from 12.6% to 12.7%. Moreover, 57% of the rail patronage would be siphoned away from other rail lines. And the Rowville line would reduce the number of car trips on a typical weekday in 2046 by a relatively tiny 15,000.

There are other transport projects in Melbourne that are far more compelling than a rail line to Doncaster. They include elimination of the city’s 170 odd level crossings and better bus services in the outer suburbs. What is needed is a sensible debate, with reliable evidence. Based on what’s been reported (i.e. leaked), it doesn’t seem like this study will give us that.

Update 27 July 2012: see public statements issued by Arup and the Doncaster Rail Study Team in Comments below, #s 15 and 16.

Update 31 July 2012: see comment (#34) by Peter Newman.