Murdoch station on the Perth to Mandurah rail line opened in 2007

For around $2 billion in today’s dollars, the WA government constructed a 71 km rail line from Perth CBD to Mandurah. About 30 km of the route is located in the median of the Kwinana freeway.

In contrast, this week the Napthine government released a feasibility study estimating the cost of connecting suburban Doncaster to the CBD by rail – about 20 km of track – would be around $8-$12 billion (fn 1).

The two rail lines have some similarities. Much of the proposed Doncaster line would be located in the median of the existing Eastern Freeway. It was designed at the time to accommodate a future Doncaster train service.

The Doncaster costing isn’t unusual. The 13 km Melbourne Metro is likely to require funding in the region of $8 billion and even 19 km of light rail to serve Sydney’s eastern suburbs is costed at $1.6 billion.

I’ve looked before at possible reasons for the high cost of infrastructure, including public transport (e.g. see herehere and here).

The new Doncaster feasibility study is especially interesting because it adds to that discussion. Unusually, it explicitly seeks to explain and justify why it’s expected to cost much more than the Mandurah line.

It notes that “cost comparisons have been drawn by some”, undoubtedly a reference to the report prepared last year by a team led by Curtin University’s Professor Peter Newman.

His report said a line from Doncaster to the CBD along a similar alignment to that now proposed could be constructed for circa $1.2 billion. It was prepared for a group of eastern suburbs municipalities lobbying for the line.

The new study recognises there are some superficial similarities between Mandurah and Doncaster. Both utilise a freeway median for some of their alignment and both rely on park-and-ride and feeder bus services over walk-up patronage.

But it cautions against comparisons:

It is the opinion of the study team however that these comparisons seek to over-simplify the very different nature of these distinct rail lines, drawing comparisons where they are alike, but failing to recognise some very important differences between the two

The report offers seven explanations for the difference in the cost of constructing the two lines.

First, Doncaster is a “brownfields” construction site. Construction would be difficult because of the narrowness of the median and the need to keep the existing freeway operational.

Mandurah was essentially a “greenfields” site. Although it was partly constructed on an existing busway, this was for only a relatively short section of the entire route.

The southern section was constructed within a very wide median or, in some sections, before the second freeway carriageway was built. Access to the majority of the railway site was therefore relatively easy for construction.

The report argues it would be reasonable to assume a ten-fold increase in costs attributable to working in a “brownfields” versus a “greenfields” environment.

Second, Doncaster requires a number of underground stations, whereas most of those required for Mandurah were at-grade. Where underground stations were built, they were at considerably shallower depths than Doncaster would require.

Underground stations are generally in the order of five times more expensive than at-grade stations to construct.

Third, there’s been considerable inflation in construction costs since Mandurah opened in 2007.

Construction inflation could easily add 33% to any construction prices over the intervening period.

Fourth, Doncaster would require 1500v DC electrification in order to be compatible with the rest of Melbourne’s rail network. The Mandurah line uses a 25kV AC system.

A DC system would typically cost twice as much as a comparable AC system.

Fifth, Doncaster requires a number of large bridges. With the exception of the Narrows Bridge, most of those constructed for Mandurah were relatively minor.

Sixth, the ground conditions faced by a Doncaster line are relatively complex, involving basalts, siltstones, and soft alluvial deposits. The Mandurah line was largely constructed on permeable sand deposits.

Seventh, a significant proportion of the Doncaster alignment requires tunnelling because it’s located within an urban area. Much of the Mandurah route was constructed through greenfield “pre-urban” areas and consequently only 1% of the alignment (less than 1 km) involved tunnelling.

Tunnelling operations are largely unique and vary in cost depending upon a number of factors. Typically however, tunnelling costs in the order of ten times as much as construction above ground.

The report doesn’t mention them, but there are other factors that might also help explain the low cost of Mandurah compared to practically every other urban rail proposal (e.g. see herehere and here).

In addition to differences bearing directly on the cost of construction, the report also signals that a Doncaster line would be unlikely to generate the same rate of increase in patronage as the Mandurah line.

The Doncaster area is an established, middle ring region experiencing relatively slow growth (less than 1% p.a), mostly via redevelopment. Mandurah on the other hand serves an outer suburban growth corridor where population is increasing at 5% p.a.

No one likes the high cost of urban rail infrastructure, but insisting the costs of Mandurah are applicable to other projects in today’s environment is just blatant politicking.

So is arguing that governments deliberately and consciously inflate project costs in order to argue they can’t be built. If anything, the history of cost blowouts on major projects suggests the opposite (although I think it’s primarily optimism bias rather than deceptiveness).


(fn 1): For a mostly at-grade alignment (freeway median) from Doncaster Hill to Clifton Hill and a below-grade connection (tunnel) for the “decoupled” Sth Morang line from Clifton Hill to Flagstaff. Alignments and costs are as cited in the feasibility study.