Canberra’s tram plan looks suss to me.

It was with some excitement that I downloaded the PDF containing the full business case for the new Canberra light rail. 

1. I love public transport.

2. I like to see open and honest planning. The Victorian government has given us no such insight into their East-West Link plan.

But my happiness turned to confusion when I saw the way the calculations had been made. This project supposedly has a benefit-cost ratio of 1.2. But when you look into the $1 billion of benefits, you’ll see that they are a castle made of gossamer. The $823 million of costs look altogether more concrete, and the study estimates a 25 per cent chance they will blow out to be higher still.

capital metro map
Proposed route

The most obvious benefit of the tram is time savings.

These have been calculated over 30 years, even though the operating contract for the metro is for 20 years. I am suspicious here. If you assume traffic conditions deteriorate at a given rate per annum, you could get the model to show very bad traffic in 20 years time. Including those final ten years might be a substantial source of the time savings.

And it goes without saying that trying to calculate seconds saved by tram users in 2049 requires making some heroic assumptions about the way we will live then.

The patronage assumptions also look bold. They assume 15,000 passengers a day in 2021. That’s 7500 in each direction, or 416 passenger boardings each hour for 18 hours of operation. If you assume a tram every 10 minutes that’s 69.4 boardings per tram, and if you assume the average passenger goes 6km along the 12km line, you’re guessing the tram will have on average 35 passengers on board. I’ve been on a lot of ACTION buses and that seems like 35 times as many passengers as the average ACTION bus outside peak hour.

Time savings are not even the biggest source of “benefits” in this highly questionable calculation.

tram benefit graph

The biggest source of benefits is land use.

“Light rail and the anticipated resulting urban renewal will create a more grand entrance to the City befitting its status as the Nation’s Capital”

It is true that the drive into Canberra – down Northbourne Avenue – is grim. I moved to Canberra in 2005, and coming down Northbourne Avenue that fateful day I didn’t realise we were in the heart of the city until it had passed.

But solving that by building a tram is a bit like renovating your kitchen by installing a restaurant in your lounge room. Sure, it provides the impetus, but it’s indirect and probably pretty costly.

If you want to increase density near the centre of Canberra, the big question should be what’s the best way to do this?

If you pose the question that way, using an elaborate system of trams seems less efficient than changing the planning regulations.  The business case more or less admits it:

“Land development decisions undertaken by the ACT Government will influence the realisation of benefits by the Capital Metro project”

But it makes no attempt to say whether those “decisions” could be brought in without the tram.

The next biggest source of benefits is wider economic impacts.

Some people say wider economic benefits are the cheat’s way of boosting your benefit-cost ratio. Infrastructure Australia says that:

  • Wider economic benefits may be negative for some initiatives; and
  • the availability of Australian specific data needed to calculated wider economic benefits is currently sub-optimal.
  • Therefore, Infrastructure Australia will treat wider economic benefits separately to the traditional CBA.

It’s true that the ACT government does two calculations of the benefit cost ratio: with the Wider Economic Impacts and without.


But given the project would not go ahead with a ratio of 1.0, the wider economic benefits are obviously key to the calculation.


Benefit cost analyses are a GIGO model. You put garbage in and you get garbage out. Even more important than the garbage that goes in is what is excluded. here’s a short list.

  • There is no price paid for the public land on which the tram will run. The tracks will be 6.6m wide. The line is 12km long. If 11km is on public land, that is 72,600 square metres.  At $575 per square metre for undeveloped land, that’s $42 million that should arguably be included as an opportunity cost.
  • The tram has priority at all intersections. I see no estimate of how much extra waiting that will mean for pedestrians, cars, cyclists and buses that try to cross its path.
  • Cutting bus services is counted as a benefit, not as a cost.
  • The cost of the funds is not accounted for. Whether raised by debt or taxes, getting $100 for the government to spend does not cost society $100.

The cost of the tram plan is equal to $830 million. To put that in context, it’s $2,168 per ACT resident. And most of them don’t live near the 12k tram line.

The red line is the tram. The areas in the black circles miss out on any benefit
The red line is the tram. The areas in the black circles miss out


The budget for the entire ACT bus system was $123 million in 2012-13. The reason patronage is low and dissatisfaction is high is clear – low frequencies, badly designed routes and poor service hours. You could double patronage a lot more easily, for a fraction of the cost of the tram, by just improving the buses.

How much should we spend to get cycling up to 5 per cent of trips?

Melbourne’s weather is poor. It rains often. The city is huge – 100 km from edge to edge – and vast swathes of it are covered in the kind of densely packed contour lines that make cyclists legs tremble.

In winter, Melbourne’s cycling community shrinks by over a third.


On days like today I suspect the number of cyclists is far smaller.


In short, Melbourne will never be the sort of city where 50 per cent of trips are possible by bike. Cycling (and walking) will never ever do the “heavy lifting” in our transport mix. That role will always be split between public transport and private motorised transport.

At the moment, the mode share split between these three is:

Source: 2011 census

And the trends are these:

Cycling is growing fast, more than doubling in eight years.:

Source: VicRoads

Public transport growth has been its highest in sixty years, with train travel accounting for most of the increase:

Source: PTV

And vehicle kilometres have surged on freeways, while not increasing on arterial roads.

Source: VicRoads

Expenditure on specific infrastructure looks like this:

Nationwide, spending on cycling is $112.8 million. Spending on roads is over 100 times more, at $18 billion.

Source: BITRE

The data is tough to aggregate, but one estimate is that roads get four times the investment of public transport.

All the modes are growing. How do we decide what the data means? And why not let the market decide what modes live or die?

The answer to the second question is that transport is going to be a centrally planned space until we can charge users per kilometre.

Public roads built to accommodate cars push the whole investment process into the world of “second-best.” If subsidising roads is a given, subsidising public transport can be efficient. Subsidising public transport makes policy makers wonder if there are other, cheaper ways to move people around, like bikes.

So if we’re going to be centrally planning our transport mix, we must ask: do we like the current 78/17/5 mix?

I’d argue we should not. I’d argue we should be aiming to grow the share of modes that have fewer negative externalities and greater returns to scale.

I’d hazard a guess that for Melbourne, 10 per cent share evenly split between walking and bike, 30 per cent for public transport, and 60 per cent for cars would be optimal.

Does that mean we should start spending 10 per cent of infrastructure funding on active modes, 30 per cent on public transport and 60 per cent on cars?

Only if we want to move very very slowly.


Infrastructure lasts a long time. That means the stock of existing infrastructure is the single biggest determinant of infrastructure in five years time. Marginal changes in expenditure rates affect outcomes only very gently. If we want to effect change, we need to tip the scales massively in favour of the modes we want to grow, in the short term.

That means that announcements like $650,000 for changes to a cycling bridge in Melbourne’s west should not be cause for widespread congratulation.

In the short term, we could probably usefully spend 60 per cent of the transport infrastructure budget on public transport and 15 per cent on active modes. If we did that for a few years, we would move swiftly towards the outcomes we want, before returning to a “maintenance” split, where expenditure is based on usage.

Spending even $500 million a year on bicycle infrastructure might seem like a lot when the recent budget has been around $30 million. But when you look at what passes for “bicycle infrastructure” and imagine replacing it with global quality bicycle infrastructure, it would be a drop in the ocean.

“Bike boxes” were the sine qua non of Melbourne bicycle infrastructure innovation just a few years ago.

I don’t imagine gold-plated bicycle infrastructure should go everywhere. Far from it. Cycling infrastructure should be optimised in the areas where cycling can thrive, likely to be areas that already see some bicycle traffic. Fixing missing links, creating Copenhagen lanes on major on-road routes, plus widening and lighting off-street bicycle paths would be the top three priorities.

If we want to increase the share of some modes, we need to be bold about throwing money at them, and not be afraid to acknowledge that such a move comes at the expense of other modes.

Faster train journeys – some low-hanging fruit.

The state government is prepared to make big investments to make train travel easier and faster. So they should. They are contemplating a $9 billion tunnel that will make journeys faster and more reliable.

But what if I told you there was a much much cheaper way to improve travel times of the rail system, while making people’s journeys to work more comfortable?

The solution requires thinking outside the box. This is not about putting faster motors on the trains. Not about improving signalling or driver training or anything to do with the train system itself. It’s about cutting walking times to the station.

Let’s look at my local station, Clifton Hill.

Clifton HIl station entrances and exits.
Yellow dots mark the entrances

There are just three entrances, all clustered up the northern end of the station.

The last 100 metres.
Placing the entrance at the northern end means passengers walking from the south have to walk an additional 120 metres to get on the platform. (I reckon 2 minutes is an overestimate by our friends at Google Maps.)

This is not a passenger-centric design, but an operator-centric design.

Where new station entrances should be
Where new station entrances should be (blue dots)

Walking times are some of the most important parts of a train journey. A meta-analysis of the subject cites research that found ten minutes of walking was equivalent to 20 minutes of riding in the vehicle. In other words, the walking section of the journey is an important place to focus improvements.

I’ve made a model for the benefits at Clifton Hill station. Clifton Hill Station got 3,009 boardings per weekday in the 2011-12.

I model it like this: 25 per cent of the users who walk to the station live in line with the platform, so moving entrances to the end is of no benefit, while 75 per cent of users currently walk past the end of the platform to get to the entrance. An amazing isochrone map app I found shows that my model is probably conservative, because of the local geography.

Those inside the yellow circle and not between the red lines can be assumed to benefit from new platform entrances.
Those inside the yellow circle and not between the red lines can be assumed to benefit from new platform entrances.
How far can you get in 12 minutes?
12-minute walking map. Source: Cartoo

There are 90 car parks at Clifton Hill Station, and it is served by buses. There are trams not so far away. I estimate two-thirds of people walk to the station. (2000 users a day.) Let’s say 1500 of them could potentially access the station via the new gates I propose.

But smart public transport users know not all carriages are equally useful. Depending on what station you get off at, and where you’re headed, your train exit may be speeded up most by being at the front, back, or middle of the train.

If you come to Clifton Hill Station from the south, but you want to board the southbound train’s last carriages (and you’re not running late) it provides you no advantage to have an extra gate at the south end of the station.

But if you’re coming from the south and you want to be in the front carriages of that southbound train, you need to walk that distance twice. Once along the street outside the station, and then back again along the platform.

Putting an entrance at each end of the platform, replacing a northern-end entrance
Southerners Middlers Northerners
front carriage riders 150 0 0
mid train riders 75 0 0
last carriage riders 0 0 0

12 per cent of the 2000 walkers will save a conservatively estimated 150 metres, or 90 seconds. 12 per cent of them will save an estimated 75 metres, or 45 seconds. That adds up to 563 minutes on access to the station. That makes 9.4 hours. If the same effect is present when they return home, it’s worth 18.8 hours.

The model does not assign any benefit to all the trains that may now *just* be caught when before they were just missed.

Assuming a value of time of $30 an hour, the value of the additional exits would be $146,000 a year, just measuring weekday trips. For a gap in the fence, a bit of concrete paving and some extra Myki machines, which I estimate to cost perhaps $500,000, it would pay itself off within a few years, yielding a positive rate of return.

You may think a minute here or there is not important, but there is no single change that can cut a train journey’s duration in half. If we want improvements to service we need the operators to accumulate small easy changes like this across the network.

Clifton Hill Station is not even the worst offender. Camberwell station has the entrance to two of its three platforms about 100 metres away from the main road it serves.

Camberwell station served 6,571 passengers each weekday in 2011-12
Camberwell station served 6,571 passengers each weekday in 2011-12.

Why haven’t they thought of these fixes already?

There will be a perceived trade-off with safety. I don’t doubt the ex-post rationalisers are currently saying “but having one entrance allows for surveillance!” But a single entrance also funnels station users past a single choke point. In the same way a narrow alley feels dangerous at night, so can a single station entrance with no alternatives.

In fact, more exits and entrances should mean fewer people spending time at stations, increasing the visibility of anyone loitering with malicious intent and decreasing their opportunities.

I contend the reason this sort of station design is not widespread is institutional. The Public Transport Victoria guidelines for station design seem to support this kind of solution:

“Many aspects of the local context and surrounding urban design will influence the station entry configuration. A thorough study of the station catchment area is required to determine the most appropriate placement of the entry or entries in order to attract patronage by:
a) Encouraging the use of the station by simplifying connections with existing and future urban design;
b) Providing accessibility, convenience, clarity and quality of arrival to and from the station;
c) Providing safe and attractive public spaces that contribute positively to the local identity;”

But in fact, station access runs second to concerns over train boarding patterns. PTV tries to alternate whether station entrances are at one end of the train or the other, to prevent any one carriage getting too full. Their conception of their job focuses on the trains, not the passengers.

But building more station entrances should be a priority. It won’t help just walkers. At the margin, making walking distances to train stations shorter will encourage more people to walk, and free up scarce car parking spots for people who live even further away.

As well as institutional bias from the departments, there is a ribbon-cutting bias in public transport investment. Politicians want grand visions. The Premier can’t imagine himself showing up to the construction of a new station entrance, so he doesn’t push for it to get done. But that doesn’t mean it’s not the best idea.