Could we pay travellers not to use over-crowded trains?

The John West technique: maximising carrying capacity on peak hour trains, Japan

If you think crowding of trains in Australia’s capital cities is bad, have a look at this extraordinary video of how they cram passengers onto trains in Japan! John West could learn a thing or two! Peak crowding is uncomfortable for passengers and increases operating costs – more capacity is needed to handle the peak, but much of it is unused in the off-peak period. That extra capacity might take many forms, such as more carriages, more trains, more staff, etc.

There could potentially be big savings if some of this peak demand were shifted to earlier or later periods. This applies to trains, buses and roads and indeed to many activities that experience peaking e.g. cinemas, concerts. Apart from the disincentive of being treated like a sardine, the standard approach is to charge a higher price in peak periods relative to the off-peak. However political constraints mean public transport operators in Australia tend to conceive of differential pricing as an off-peak concession rather than as an active way of managing peak demand.

Here’s another way of approaching this problem. The Economist reports Singapore is planning a pilot scheme offering public transport passengers a greater chance of winning a prize if they choose to go off-peak. All travellers are entered into a pool with a chance to win cash in weekly lotteries, but those who travel off-peak will effectively get three times as many ‘tickets’. The principle is that small rewards will pay for themselves in lower capital and operating costs.

The Economist quotes Stanford University academic, Balaji Prabhakar, who says lotteries rely on the behavioural-economics insight that the average person is risk-seeking when stakes are small:

Offer individuals 20p to leave the house an hour earlier, and most will say no. But a 1-in-50 chance of winning £10 may seem more enticing. The risk-seeking effect is amplified in small networks: regularly hearing about other winners leads individuals to overestimate their own chances of success.

The idea of carrots rather than sticks is not new. For example, long-standing readers might recall this proposal to reward drivers who don’t speed with a cash reward. Fines from speeders are paid into a pot and redistributed randomly as prizes to motorists who are ‘caught’ by speed cameras driving within the designated limit. The Capital Bikeshare scheme in Washington DC offers prizes to riders who travel against the dominant flow, thus reducing the cost of rebalancing the (geographical) distribution of bikes. This study of the effectiveness of a lottery in reducing car travel found it had a positive effect, although it disappeared when the lottery was stopped (note very small sample size). Read the rest of this entry »


Is commuting killing us?

Assembly kit for Manhattan street grid

Long commutes cause obesity, neck pain, loneliness, divorce, stress, and insomnia. Your commute is in fact killing you, according to this story published in Slate last week. And it’s bad for others too – in his Melbourne address last month, Robert Putnam argued that a ten minute increase in commute time reduces social capital by 10%. Richard Florida says it’s time to put commuting right beside smoking and obesity on the list of priorities for improving the health and well-being of Americans.

I’m always bemused by these sorts of claims. Apart from the fact that the majority of commutes are relatively short, they neglect the salient fact that people spend time commuting because it’s worth it – that’s how they earn their living. And in general, the further they go, the better the job and/or the better the house. Commuting is a bit like having children – it costs a squillion, but for most people it’s worth it!

The reality is that most people prefer to commute some distance. This study of US commuters by Redmond and Mokhtarian found that 42% of their sample are happy with their current commute i.e. their actual travel time and their ‘ideal’ commute time coincide. People seem to like some space between work and home. They found that 7% actually say their commute isn’t long enough!

Nevertheless, the study also found that just over half feel their commute time is too long compared to their ‘ideal’ commute time. That finding, however, doesn’t really say much. The trouble is people don’t make unconstrained judgements like this in real life. If asked, rational people will of course say they would like less of the boring things in life and more of the interesting and exciting things. If they’re not forced explicitly to consider the cost, people will naturally acquiesce when they’re posed questions of this sort. It’s a difficult concept to measure, so a much better guide to commuting time preferences is what people actually choose to do in the face of real-world constraints.

It turns out workers don’t tend to spend inordinate amounts of time commuting. This analysis of US Census data shows that 45% of one-way commutes in US metropolitan areas take less than 20 minutes and only 8% take more than 60 minutes. This US survey found that 81% of commuters spend less than half an hour getting to work. In Melbourne, more than half of all trips to work (54%) take less than 30 minutes. Only 12% of commutes take longer than an hour and only 3% more than 90 minutes.

Having said that, whether or not an hour a day spent commuting to and from work is ‘inordinate’, depends on what it yields. The question can’t be addressed sensibly without considering the benefits as well as the costs. We spend time on a host of activities like sleeping, cooking and taking the kids to sport because we feel they are necessary to derive the associated benefits. Likewise, commuting provides something that’s extremely valuable – income. That’s a basic, a necessity. But work also provides a host of associated benefits like status and social interaction. The bottom line is we commute because it’s worth it – we’ll minimise commute time subject to other constraints but we don’t expect it to cost nothing. Read the rest of this entry »


Should fare zones be rationalised?

I like the principle underlying Melbourne’s two zone public transport fares system – if you travel further, you pay more. Travel within roughly 11 km of Flinders St Station is a Zone 1 fare and travel within the area beyond that is a Zone 2 fare. Those who travel between zones pay a Zone 1-2 fare. Sensibly, there’s a three station ‘grace’ overlap between zones.

The logic of the tariff is premised on the idea that trips are to the CBD. If you live in Epping (say), you are 19 km from the CBD and in Zone 2, whereas someone living near Bell station in Preston is only 9 km out and is in Zone 1. The standard Two Hour Zone 1 fare is $3.80 and the Zone 2 fare is $2.90. If you cross zones, the fare for the longer distances expected with Zone 1-2 trips is $6.00.

That all sounds good, but inevitably there are anomalies. For example, a traveller boarding a train on the Epping line at Reservoir station can travel through 15 stations to Flinders St for a Zone 1 fare. However anyone boarding at the next station out, Ruthven, must pay a Zone 1-2 fare to travel to Flinders St.

Residents who live near the overlap stations and make short trips face a very high marginal cost. For example, someone boarding at Bell station on the Epping line can only travel 3 stations away from the CBD before incurring a Zone 1-2 fare.  Similarly the four station inbound trip from Ruthven station to Bell is a Zone 1-2 fare. The same pattern happens on other lines.

Another anomaly is the cost of Zone 1-2 fares compared to Zone 1 and Zone 2 fares. With so much of the cost of operating a rail service tied up in capital items, it’s hard to see why a Zone 1-2 fare costs 60% more than a Zone 1 fare and over 100% more than a Zone 2 fare. That might possibly make sense for buses but seems a real stretch for trains.

These sorts of anomalies could be seen as unfortunate but inevitable compromises – after all, most travellers are bound for the city centre. However as discussed here, the great bulk of jobs are now in the suburbs, not the CBD. Moreover, most Melburnians live relatively close to their key travel destinations, usually in the same or a contiguous municipality.

We should expect that public transport travellers will increasingly be seeking to make intra-suburban trips via a grid of well coordinated train and bus services. Indeed, many advocate such an approach. A two-part tariff based on distance from the CBD makes less sense in that situation. While the CBD will remain the main market for trains for a considerable time yet, they nevertheless are likely to have a growing role in intra-suburban travel. Policy-makers should be thinking about how to make them work harder in this role. Read the rest of this entry »


Why do we dislike buses?

Houston, we have a problem...

Melburnian’s seem to love trains and dislike buses. Melbourne Airport and Doncaster are both served by high-frequency bus services with a wide span of operating hours, yet large numbers of people want to spend billions replacing them with trains.

The list of criticisms of buses – relative to trains – is long. Right at the top is slowness. Buses operate in traffic, follow circuitous routes, stop frequently and idle while passengers dig out spare change to pay the driver. They’re uncomfortable too. Shelters are perfunctory, the ride is jerky and difficult if standing, seats are jammed too close together and too many drivers don’t seem to actually like passengers. And buses are unreliable. They’re invariably late and if you miss one the wait for the next one will seem interminable. If it’s night or a weekend there’s a good chance that was the last bus for the day.

Then there are systemic criticisms. Buses aren’t ‘legible’ — prospective passengers can’t see where the route goes. Sometimes routes vary over the course of a day. Buses are also impermanent. Developers are less inclined to risk investing along a route if it can be changed overnight or even removed. And there’re hard-nosed criticisms, too. Buses don’t carry many passengers. Operating costs are high because each vehicles requires a driver. Per capita GHG emissions and energy consumption are no better than cars. Engines are noisy and polluting.

As things stand, buses look pretty bad compared to trains, even given the unreliability and crowding that characterises peak hour train services in many Australian cities. Buses have a serious image problem, not just here but in many western cities.

But it’s an unfair comparison*. The key reason buses are perceived so poorly relative to rail is they are mostly assigned to marginal routes with low patronage. Operators follow indirect routes and stop frequently to maximise revenue; they reduce frequency and hours of operation to minimise costs. They can do that because their customers aren’t usually sophisticated CBD workers but travellers who are mostly “captive” to public transport. In other words buses mostly operate in a different market to trains. But it’s given them a bad reputation.

Most of the criticisms can be fully addressed, or at least softened, when the comparison is like-for-like i.e. when buses operate the sorts of long-haul commuter services that urban rail is customarily used for in Australia. Bus Rapid Transit (BRT) systems typically have high patronage and operate in their own right of way (just like trains) or in dedicated on-road bus lanes. Stops are major interchanges spaced kilometres apart and tickets are bought before entry.

BRT vehicles can be made with the internal look and ‘feel’ of (light) rail and the jerky ride can be minimised with electric engines. Articulated buses carry large numbers of passengers and can provide better leg room. According to Corinne Mulley, Chair in Public Transport at Sydney University, Brisbane’s South East Busway carries 15,000 passengers per hour and in Bogota buses carry 45,000 per hour.  She says “US evidence points to infrastructure costs for dedicated buses being approximately one third of light rail costs”. As a point of reference, Eddington forecast that a Doncaster rail line would carry up to 12,250 one-way trips per day. Read the rest of this entry »


How do world transit systems compare?

This is an eye-opener. First thing you need to do is go here where artist and urban planner Neil Freeman has prepared maps of many more cities at the same scale  (none from Australia unfortunately). What these maps show is the enormous variation in the scale and density of different systems. No doubt the artist has had to make assumptions about what to include and exclude*, but this is nevertheless very important information to bear in mind when comparing the performance of Australian rail networks with systems elsewhere.

I think Melbourne’s rail system is a leviathan compared to some of these e.g. the networks in Toronto, Vancouver, Amsterdam, Budapest and Brussels. Note the scale (5 km) beside the map of Toronto above. As a point of reference, it’s about 5 km as the crow flies from Flinders St station to Merri station on the Epping line, Hawthorne station on the Lilydale line and Newmarket station on the Broadmeadows line. Five kilometres is the boundary of the inner city. Read the rest of this entry »


– Are cars the key challenge?

Share of journeys to work by car, 2006 (click). Map from Charting Transport

The accompanying map from Charting Transport shows the proportion of journeys to work in Melbourne’s inner city undertaken by car in 2006 — the rest were by public transport, walking and cycling (note that green denotes a low share for cars and red a high share*).

There is a clearly defined area — the ‘golden mile’, bounded by Spring, Flinders, Spencer and La Trobe — where cars capture less than 30% of all journeys. Around Swanston Street the share of journeys taken by car is as low as 24% (click to enlarge map).

But the really striking thing to my mind is the steep increase in car’s mode share as soon as you move beyond walking distance of the rail loop. You only have to go as far as the northern side of Victoria Parade and that share jumps to 50– 60%. Go toward the Hoddle Street end and it’s more than 70%. Once you get into the suburbs, most employment concentrations have a mode split that is 80% – 90% car.

One interpretation of the data is that sustainable modes are doing pretty well in the golden mile. Another is that they’re not doing well enough – maybe no more than 10% of journeys should be considered the ‘natural’ share of the car in this sort of key business area – and hence there’s an opportunity for public transport to increase its share.

I incline to the latter view, but I think there’s another interpretation of this data that’s arguably more significant. Consider what it takes, in terms of density and transit supply, to get the share of journeys to work by car down below 30%.

The ‘golden mile’ is just 2 km2 out of a built up area of around 2,500 km2 (the entire MSD is over 7,500 km2). That little area is the largest single concentration of jobs, by far, in the metropolitan area. It accounts for 10% of all metropolitan employment and is denser by an order of magnitude than any other activity centre in Melbourne, as exemplified by the number of very tall buildings it contains. And despite its diminutive size, it’s the focus of the entire metropolitan train and tram systems. It’s the ‘hub’ where all the ‘spokes’ meet – a giant transport interchange – and is accordingly, again by an order of magnitude, the most accessible place in Melbourne by public transport.

Yet for all the astonishing advantages of job density and public transport quality, a quarter to a third of all those who work in this tiny area still drive to work. Further, literally within one or two blocks of the rail loop, the car’s share of journeys to work rises to 50% and more, notwithstanding that these near-CBD areas themselves have a job density and level of public transport service that is far higher than the rest of the metropolitan area. And this is the journey to work – the trip most likely to be taken by public transport! Read the rest of this entry »


Why is transit’s share low at Latrobe Uni?

Mode shares (%), 2006 Census

Workers who commute to Melbourne University at Parkville are much more inclined to use public transport than their colleagues who work at suburban Monash or Latrobe universities. The chart shows that at the 2006 Census, 41% of Melbourne University workers reported they drove to work compared to 83% at Monash and 84% at Latrobe universities. Many more staff at Melbourne also walked and cycled – 24% compared to 6-7% at the other two institutions.

Melbourne University’s lower car use is explained by a few key factors. The main one is that it is located on the edge of the CBD where car use is limited by high levels of traffic congestion and expensive all-day parking charges. For many staff, driving would take too long, generate too much angst and be too expensive. If the value of driving is marginal, the decision to choose an alternative will be tipped by the high quality of public transport service available to Parkville workers. Although it’s not served directly by rail (none of these universities are), Melbourne University has easy access by multiple tram lines to the CBD and thence to the many radial train and tram lines linking to the larger metropolitan area. For many Melbourne University workers public transport would be a no-brainer.

Melbourne University’s high level of walking can largely be attributed to the relatively high residential densities in the nearby CBD and inner city environs. If transport is expensive in outlays and time, it makes sense for workers to live close to the university. In this case, living close to the university also means living close to the many activities and opportunities offered by the inner city.

The suburban setting of Monash and Latrobe provides a very different environment. Although these universities are not without their challenges, they generally experience less traffic congestion and enjoy cheaper parking than Melbourne University. Low suburban residential densities and large open space and industrial uses mean fewer staff can live within walking distance. The level of public transport service is actually pretty reasonable by prevailing standards (for example, see here) but obviously not as good as Melbourne University, which benefits greatly from its proximity to the CBD. Read the rest of this entry »