35 years of Shinjuku skyscrapers in 10 seconds (From July 1969 To July 2004)

Jeff Kennett reckons Melbourne needs to puts its rail lines underground. The former Premier of Victoria says it would be the “most important infrastructure needed for Victoria’s future.”

Addressing the Australian Property Institute’s Pan Pacific conference on Monday, he said:

It would cost hundreds of millions of dollars, but I can assure you when you look back in 50 years, or 100 years, whatever you pay today would seem cheap. We can hardly accommodate the traffic on the surface of our community in an efficient way and it is only going to get worse.

Peter Newman, Distinguished Professor of Sustainability at Curtin University, endorses Mr Kennett’s call for a subway system. He says “it’s the kind of far-sighted thinking we desperately need in Melbourne’s politics.”

Of course it would be delightful to have an underground rail system like London’s in any of our capital cities. It could possibly make sense in Melbourne because the city has a very serious problem with level crossings.

Sydney got rid of virtually all its level crossings years ago, however Melbourne still has 170. They contribute to localised traffic congestion and severely limit the number of train services that can be operated in peak periods.

Putting the rail system underground would also provide land that could be developed or used for civic purposes. No doubt that’s of great interest to the members of the Australian Property Institute.

But Mr Kennet’s estimate of “hundreds of millions of dollars” is sheer fantasy – it’s way off the mark.

The average cost of eliminating Melbourne’s 170 level crossings is around $100 million each. The cost of constructing the proposed 13 km Rowville rail line would be in the order of $2 billion and it would be mostly at-grade (surface) and in-structure (elevated).

A rule of thumb in the US is elevated lines cost four times surface lines. Underground lines cost eight times surface lines. Costs of subways are much higher in the US and Britain than elsewhere but are more relevant to recent Australian costs.

Britain’s Crossrail cost $1 billion per kilometre and the Jubilee Line Extension $450 million per kilometre. In New York, the East Side Access, Second Avenue Subway and the 7 Extension cost $4 billion, $1.7 billion and $1.3 billion per kilometre respectively.

Assuming an average cost of $300 million per kilometre, it would probably cost in the region of $25 billion to underground all existing rail lines within a modest 10 km radius of Flinders Street station. That’s a conservative estimate though, given the cost would be increased by the number of rail stations and roads that would need to be reconstructed.

That’s a huge outlay. It’s considerably larger, for example, than the Commonwealth Government’s recession-busting national Building the Education Revolution (BER) program, which cost $17 billion. It’s the kind of money that would have an enormous impact on other budget priorities

Some of that cost could be recouped from selling development rights but I expect the revenue would fall well short of the cost. The outlays for each stage would be incurred upfront but the revenue from sales would trickle in over an extended time frame and should be discounted accordingly.

Moreover most of the land currently used for inner urban rail lines is narrow and close to existing land uses, so its development potential would be modest e.g. three story apartments rather than high rise units. There would also be pressure to reserve a lot of it for parks and community uses.

The key issue though is the operational benefits of undergrounding – i.e. the improvement in the level of service of the rail system – would be relatively modest compared to both the cost and to other potential improvements the funds might be applied to. It wouldn’t even deliver on elimination of level crossings, as only a small proportion are within 10 km of the city centre.

Mr Kennett is no doubt impressed by subway systems he’s seen in Europe and elsewhere. However as the outgoing President of the Public Transport Users Association, Daniel Bowen, says, the benefits from those metro systems don’t derive from the fact that they’re underground.

We shouldn’t be fooled into thinking that for our trains to run efficiently we need them to run underground. You don’t get better rail services just by running them underground. The key is to run them frequently and to run them all day, every day and ensure they’re reliable.

What’s much more important than undergrounding is to increase the connectivity of the network and expand the capacity of existing lines to support more frequent services. That requires measures like improved signalling and elimination of level crossings.

There’s also a need for some new rail services on high capacity routes (like the proposed Melbourne Metro rail tunnel). However expansion of the public transport network doesn’t have to come primarily via expensive underground rail lines.

Australian cities already have an enormous reservoir of untapped capacity that can be exploited to expand the network at relatively low cost. That capacity is the huge existing network of roads and freeways, currently used mostly by relatively low-occupancy cars.

Road space can be reallocated to bus rapid transit and, where warranted by the level of demand, to light rail. The key requirement is that private passenger vehicles give up both road space and road priority to ensure transit isn’t unduly delayed. Drivers, of course, can shift from cars to transit.

I get why Jeff Kennett would put this sort of proposal to a meeting of the Property Institute of Australia. What I don’t understand is why Professor Newman would endorse it.