Does NSW transport need a big new idea or just clearer purpose?

NSW Minister for Transport and Infrastructure, Andrew Constance, announced yesterday he was "calling on the world's brightest tech minds to find the next big idea that would shake up transport in NSW".

It is good to seek ideas from elsewhere - I've done it myself - but technology is not a substitute for purpose. If you don't know what port you're sailing to, no wind is favourable.

Building new infrastructure, introducing smartcards or changing the fare structure do not of themselves constitute a purpose. They may be beneficial in some way, but what precisely are they directed towards achieving? Australian public transport planning has long failed to articulate a purpose or set of guiding principles for achieving it. 

This shortcoming is puzzling. There is an established "science" of public transport which is taught in tertiary institutions across Australia, often assisted by funding from transport departments.  Many overseas countries practice this science assiduously (the Swiss do it best). But strangely, in this country, the science rarely seems to percolate to the surface of public policy. More often than not, a public transport "solution" is proposed in the context of individual urban developments. The bigger picture of providing better mobility for city and suburbs as a whole seems to fall below the radar. 

One particular mode - say a Metro or Light Rail - are frequently advocated as "solutions". Often the proponents are urban planners or architects who practice the art of professional imperialism - extending the reach of their particular discipline to subjects about which they have no real expertise.

Economists are the great masters of professional imperialism, but a kit bag of arcane econometric models is a poor substitute for in depth operational knowledge of transport systems. This has not stopped the Independent Pricing and Regulatory Tribunal (IPART), an agency blissfully unencumbered by knowledge of the complex science of public transport, from being given the task of redesigning NSW's public transport fare structure. 

So what are the main features of the science of "good" public transport, which seems so curiously hidden and poorly understood? This blog post can't do it justice, but here are a few pointers.

Public Transport objective

To be effective, public transport must compete with the private car, not just for travel to work or travel to the central business district, but for all journeys - shopping, social and recreational travel. Public transport can only compete with the car if it enables the user to:

• get from wherever they are 
• to wherever they need to go
• at a time that suits them.

In other words, a city's public transport network must be multi-destinational. As only 15% of all journeys in Sydney, for example, are trips to and from work, a public transport network which focuses primarily on moving commuters to work will fail to compete with the private car. It won't accomplish its purpose.

How to achieve this objective

1. Good network design

If people are to get from wherever they are to wherever they need to go, it is not practical or efficient for the network to connect all origin-destination pairs with single, direct trips. The best public transport networks are high frequency grids, where users can make transfers at connecting nodes with short waiting times - like the London Underground or the Tokyo Metro.  But it doesn't have to be a metro system. It may be a combination of modes, including suburban trains, buses, light rail and ferries.

The key is quality network design. The mode selected for individual corridors is based on technical, geographic and cost considerations, which are contextual and pragmatic. So no one mode can be said to be better than any other. 

In areas of lower demand, where high frequency services can't be justified, the network should be designed as a pulse timetable, so waiting times are short at nodes even if service intervals are 30 or 60 minutes.

Well designed networks are also highly legible - stopping patterns are consistent and timetables are clockfaced. It is easy for passengers to figure out how to use the system.     

2. Network design guides infrastructure

The Swiss Federal Railways design timetables 20 years in advance. This allows them to prioritise infrastructure projects needed to achieve improvements in the timetable. This is the most efficient way of planning and building infrastructure because it ensures that what is built - and the technology used - is only what is necessary.

Technology has a role to play, but it should never lead. Technology should be the servant of network design. 

3. Fare structures that encourage public transport travel

The fare structure needed for successful public transport systems is the opposite of what mainstream economists think we should have. Economists like "pay as you go" fares, because then users have the tendency to ration their use of taxpayer subsidised services. Many users also like the idea of only paying for what they use.

But if our objective is for public transport to compete effectively with the car, the fare structure should provide an incentive to people to use public transport for as many trips as possible, including shopping, social and recreational trips. This is best achieved with highly discounted periodical fares (weekly, monthly, quarterly and annual travel passes), so every additional trip taken appears to the user to be free. 

Public transport fares: do we need more than economics to make them efficient?

There are some recommendations in this week's draft report on public transport fares from the NSW Independent Pricing and Regulatory Tribunal (IPART) which make a lot of good sense.

But then there are the weird bits.

Let's be positive and start with the good recommendations.

The first is the proposal that journeys made with two or more modes should be priced as a single journey based on distance travelled. Most observers were scratching their heads when the Opal card was introduced with penalties for passengers who need to transfer between modes to complete a journey. Good on IPART for recommending a fix to this.

Changes to policy on week-end fares and seniors discounts are also fair and logical. 

By replacing the Sunday cap of $2.50 with a week-end daily cap of $7.20, the perverse incentive to delay a week-end ferry ride until Sunday will be removed. And let's face it, $2.50 for travel anywhere on a Sunday is overly generous and distorts demand with very detrimental effects on service, especially for ferries.

The changes to eligibility for the Gold Opal card, which mean Seniors who are not pensioners will need to pay fares at the regular concession rate, is another dose of common sense.

Recommendations I place in the category of Weird City mostly stem from IPART's understanding of the word "efficiency". The name of IPART's report "More efficient, more integrated Opal fares" is significant. Early on at university, economics students are taught a new meaning of the word "efficiency". "Economics", students are instructed, "is about the efficient allocation of scarce resources". The instruction must be effective, because when graduates find roles in cloistered academia, consulting firms and Government agencies (like IPART), they seem compelled to look for anything that looks remotely like a scarce resource which is in need of efficient allocation, preferably by a price mechanism. 

When the only tool you have is a hammer, everything looks like a nail.

This is why IPART is so fond of "pay as you go" fare structures, because if a user pays for each individual trip, they will feel inclined to "ration" their use of public transport. It explains their recommendation to lift weekly Opal card weekly caps from $60 to $75 in 2018 and to reduce the number of "free" trips for frequent users. And there is no recommendation to re-introduce a periodical or travelpass style of ticket.

Economics is not of course the only set of ideas relevant to efficiency in business or the business of government. There is another quite different tradition of operations research and systems science, with a lineage going back to the American philosophers CS Peirce and CI Lewis, the statisticians Walter Shewart and Ed Deming and notable advocates like Stafford Beer, Gregory Bateson, C West Churchman, Russell Ackoff and Fred Emery. Their understanding of efficiency is a practical one which sits more comfortably with the average business owner. Efficiency in business (or public transport) is about making process improvements to the system to provide better products and services to customers, at lower cost. 

This tradition is mostly concerned with the way a business (or system) operates as a whole, and less about its component parts, because the whole is not the sum of the parts. 


For a public transport operator, efficiency improvement can be achieved by many practical activities, including minimising the time the fleet is lying idle and maximising the time in their rosters when crews are driving boats (trains/buses). These should also be concerns of government, because efficient use of assets and labour by operators leads ultimately to lower costs to the taxpayer.   This is what real world efficiency is all about. 

An economist's idea of efficiency is more abstract. You need look no further than this week's IPART report to see how abstract and remote from the real world this can be. Here is a quote from page 36:

"In theory, a certain number of journeys on a public transport service will maximise the welfare (or net benefits to individuals and the wider community) generated by the service. In economics, this is known as the socially optimal level of consumption. Fares set to achieve this level of consumption are known as the ‘socially optimal fares’. Socially optimal fares encourage both efficient use of public transport and efficient delivery of public transport – our two ‘efficiency’ criteria.

At the socially optimal number of journeys, the cost of providing the service to the last passenger is equal to the benefit of the service to that passenger and to the wider community. This last passenger is known as the ‘marginal’ passenger, and the costs and benefits associated with serving the marginal passenger are known as the ‘marginal costs’ and ‘marginal benefits’. 

At the socially optimal number of journeys, the costs to society of any additional journeys would outweigh the benefits to society associated with those additional journeys"

Further on page 37: 

"We developed a mathematical optimisation model that we used to estimate the socially optimal fares for single journeys on each mode...

It aims to identify the fares that will balance the following two effects: 
1. Setting fares above the socially optimal level would lead to excessive use of private cars and underutilisation of existing and planned public transport capacity, leading to higher external costs associated with road congestion, emissions and road accidents. 
2. Setting fares below the socially optimal level would lead to excessive crowding on public transport, underutilisation of existing and planned road capacity, and excessive public transport operating losses which must be funded from taxation." 

In short, IPART believes there is an optimal number of public transport trips in Sydney - not too many and not too few - and by pulling the right leavers on fares, this optimal level of trips will be reached.  After all, according to IPART, we really don't want people travelling too much by public transport or else there could be "underutilisation of existing and planned road capacity". 

What the heck?

This is the first time I have heard anyone say that the existing or planned capacity of Sydney roads may be underutilised if too many people use public transport. When did that become a problem? The thrust of the NSW Transport Masterplan (which doesn't rate a single mention in IPART's 106 page report) is to promote a mode shift away from private cars to public transport so Sydney becomes a more vibrant, liveable and better connected city. 

The report doesn't state explicitly what IPART thinks the socially optimal number of public transport trips should be in Sydney, but it appears to be not too different from what it is now - about 120 per resident per year. By comparison with most European cities, this is woefully low.  It's about 400 in Zurich and 500 in Vienna. 

IPART needs to forget the idea that public transport can be made more efficient by rationing trips via the price mechanism. What actually needs to happen is the opposite of this. Sydney needs to have a fare structure which discourages rationing of public transport travel. This is achieved by offering heavily discounted periodical fares (weekly, monthly, quarterly and annual travel passes). If you already have a periodical ticket, you are more likely to use public transport for non work commuting travel - like going shopping or meeting friends. 

Travel for purposes other than commuting to work make up more than 80% of all trips by Sydney residents! Public transport must be made a more attractive option for these types of trips if a substantial mode shift is to be achieved. Unfortunately, IPART disparagingly refers to them as "discretionary" travel (as if we would all be better off staying behind closed doors at home when not at work).

What really matters is increasing patronage and lowering taxpayer subsidies. Most European cities achieve far higher levels of real world efficiency in their public transport systems because their utilisation of assets is higher.  This can be done by drawing on the fine practical traditions of operations research and systems science, not the abstract, "mind experiment" concepts of efficiency so loved by economists. 

Fares are not everything. If average public transport trips per resident were to lift to a modest 250 trips per year, then capacity would have to increase and networks designed better to provide for multi-destinational travel, not just commuting to work in the CBD. 

But fare structures are part of the overall public transport jigsaw puzzle that we need to get right in Sydney. Unfortunately IPART is leading us down the wrong path.
   

Why are ferry work commuter numbers declining?

Every six months, a week long census is conducted by Harbour City Ferries to count passengers boarding and disembarking by individual wharf and route. The data has a high level of granularity as it includes details such as time of boarding and disembarkation.

Thanks to the good work of the Bureau of Transport Statistics, the information is publicly available on its website .

Caution should be exercised in using the data - the census covers just seven days and there may be special circumstances affecting the count in a particular week. For example, services might have been cancelled due to fog or industrial action or a wharf may have been closed for upgrade. The May 2013 count was particularly affected by fog cancellations.

Notwithstanding these cautionary notes, the data does enable the ferry enthusiast to delve a little more deeply into customer demand trends. While total ferry patronage has grown since 2010, one trend that is clear is a shift in ferry patronage away from work commuters to so called off peak periods, especially week-ends.

A proxy measure of work commuters is the count of passengers disembarking at Circular Quay, Darling Harbour, McMahons Point and Milsons Point between 7.00 am and 9.00 am on week-days. This has dropped from a high of 6,600 in November 2010 to just 5,249 in the May 2015 survey. Disembarkations at these times and at these locations include very few leisure travellers. The vast majority are people travelling by ferry to work.

Derived from BTS data set, Ferry Load Census Data May 2010 - May 2015

The decline in commuter numbers has happened despite an increase in numbers of daily AM peak services of around 12 per cent since May 2010.

So what is behind the drop?

There are probably many contributing reasons, but some stand out as leading candidates:

• Competition from the Manly Fast Ferry. Frequency of the Manly Fast Ferry has increased to six per hour in the AM peak and fares are now very competitive with the Opal Card. Disembarkations at Circular Quay between 7:00 am and 9:00 am from the iconic, regular Manly Ferry services has declined from a high of 1,765 in November 2010 to 1,230 in May 2015.

• Changes in fare structures. If you are commuting to work by ferry, you may need to catch a bus to your local wharf, or take a train from Circular Quay to reach your office. Changes in fares structures have made these intermodal transfers less attractive. My Multi ticket rules (the periodical travel pass ticket product) changed in September 2013 by moving Manly Ferry passengers from the cheapest MyZone 1 category to the most expensive MyZone 3. Inner Harbour passengers were moved to MyZone 2. Monthly, Quarterly and Annual My Multi tickets ceased to be available from August 2014.

More importantly, Opal card users, now the majority of passengers, are required to pay an extra fare if they need to transfer to a different mode as part of their journey.      

Let's hope the current IPART review into public transport fares will start a process to remove disincentives for intermodal transfers. Without such changes, the role of ferries in Sydney commuter travel will continue to decline.

The myth that ferries don't compete on cost with other public transport modes

There is a common assumption that urban transit ferries cost more to operate per passenger journey - and attract greater taxpayer subsidies - than other modes of public transport. 

Often this is true, but it is not always the case. 

A previous post on this blog debunks the myth that Sydney Ferries attract much higher subsidies than buses or trains. 

But the misunderstanding persists and often leads to  
rhetorical questions like "why should battlers in the western suburbs subsidise rich ferry riders from the north shore or eastern suburbs?" I have even heard it raised at IPART hearings. 

So it's time to restate the facts.

The NSW Auditor-General's Financial Audit Report - Volume Seven 2014 - provides a helpful table showing the operating costs and revenue per passenger of three modes of public transport in NSW (trains, buses and ferries) in the 12 months ending June 2014. A graphical version is shown below:

Source: NSW Auditor-General's Financial Audit Report Volume 7 2014, p. 17

Net cost per passenger journey, after deducting farebox revenue, is $10.61 for rail, $6.24 for ferries and $5.88 for buses. This means the subsidy per passenger for rail is more than $4 higher than ferries. The difference between buses and ferries is almost negligible. 

Now there will be arguments about what the Audit Office includes as a cost. And the average rail journey is longer than either ferry or bus journeys (this raises another question - why aren't train fares commensurately higher for longer trips?). 

All of this does muddy the water, but the difference in operating costs are really not that great.  

When other issues come into play, such as the impact on CBD congestion caused by conga lines of buses entering and leaving the city, the argument for ferries becomes more compelling. Ferries certainly have a competitive advantage over buses in near city suburbs with good waterway connections. 

Putting the horse before the cart in ferry infrastructure

Not since the opening of the Harbour Bridge in 1932 has Sydney seen more major projects planned to support its ferry network. The recent announcement of a major upgrade to Circular Quay is one of many major ferry infrastructure investments, including a new hub to be built at Barangaroo South and the acquisition of six new inner harbour ferries.

The thought of shiny new terminals with glamourous retail areas affording sparkling views of Sydney Harbour can bring a rush of excitement. But let’s not forget the main purpose of a ferry terminal is to provide for the safe and efficient berthing of vessels and the safe and efficient exchange of passengers. And before we can say with confidence that the new terminals will perform these functions adequately, a more fundamental question needs to be answered: what should the ferry network of the future look like? Only then can the operational specifications of the new terminals and vessels be settled.

Why does the network matter?

One obvious example is the level of connectivity that passengers should expect. If Sydney was to adopt world’s best practice in network design, all lines hubbing at Circular Quay would operate to a pulse timetable, as practised by the Swiss and in other parts of continental Europe.  All ferry services would arrive at Circular Quay shortly before the hour and half hour and depart shortly after the hour and half hour. This would enable lines intersecting at Circular Quay (or at Barangaroo for that matter) to connect conveniently with just a short transfer wait time.

It would also make Sydney Cove safer to navigate, as arriving vessels would no longer interfere with those that are departing.

Opting for quality in network design has consequences.  A well connected network must be punctual or planned connections will fail. This requires more advanced gangway technology at wharves and on vessels, so services are not delayed by slow passenger loading.  And if timed connections become the norm, pontoon ramps and the passenger concourse at Circular Quay and Barangaroo must allow free movement of passengers transferring between vessels.

One of the biggest questions is how to make better use of Barangaroo in the network. The current King Street Darling Harbour terminal is a poor relation to Circular Quay, but this is now changing with developments on the western side of the CBD. By the time King Street wharf is replaced by a four berth terminal at Barangaroo in late 2016, ferry passengers coming from the west of the city will find this to be easily the most convenient point of entry to Sydney’s CBD. This is thanks in part to the Wynyard Walk project, which will provide easy pedestrian access to Wynyard Station and George Street.

The best option would be for all ferry lines west of the harbour bridge to hub at Barangaroo, including a new line to the White Bay Power Station in anticipation of the Bays Precinct development. There is no reason why the Darling Harbour, White Bay and Parramatta lines should not all pulse at Barangaroo, adding significantly to the utility and liveliness of this terminal.

The building of a new transport interchange does not of itself improve the connectivity of a network. Only timetables make connections, but they need well designed infrastructure to support them. It is time for Sydney to stop putting the cart before the horse. Let’s design our networks first before building the supporting infrastructure.

Pyrmont Bay Terminal Review

The new ferry wharf at Pyrmont Bay opened on Tuesday 8 September. It is one of the many new terminals built by Roads and Maritime services over the last five years, starting with Milsons Point (completed December 2010).

At the outset, it must be questioned whether Pyrmont Bay has a long term future as a ferry terminal. As the crow flies, it sits 200 metres from King Street wharf, a distance perfectly suited to a pedestrian/cycle bridge. That would be a much more convenient way for people to move between Pyrmont Bay and King Street or Barangaroo South.

A ferry terminal on Darling Island, situated on almost a direct line from Barangaroo South to the Rozelle Power Station in the Bays Precinct, would make a lot more sense than Pyrmont Bay, especially if the current cruise terminal at Balmain is developed into a major residential precinct. A new line from Barangaroo with stops at Darling Island, Balmain South, Glebe Island and Rozelle Power Station is a far better option.

But aside from whether or not it was necessary to build a new terminal at Pyrmont Bay, how does the wharf perform?

Well, it is better than the old stepped wharf, but not much better.

A previous post pointed out some serious design problems, which have unfortunately not been addressed. The small pontoon and single ramp are not adequate for crowds likely during major events, especially the Vivid Festival. Disembarking passengers will be met by a wall of people trying to board. That's not safe and will cause delays. 

The other curious feature is that although the pontoon is perpendicular to the shoreline, there is only one berthing face. This is a wharf used by other ferry operators and party boats, so two berth faces would seem an obvious requirement.

And when did Opal readers stop being located close to the pontoon? At both Pyrmont Bay and the recently opened Balmain East wharf, the only Opal card readers are on the landside, a sizeable walk from the pontoon. Woe betide the poor passenger who only remembers to tap on just as the gangway is planked. The sprint back to the Opal card reader may cause them to miss the ferry.  

It is time for a re-think about the Sydney Ferry network and some fastidious planning to make sure wharf infrastructure investments better match operational and customer requirements. 

Why the inconvenience of carrying whitegoods on a train is not very important

As restful as it may be to leave our prejudices quietly undisturbed, there are times when habits of thought need to be shaken up by a dose of empirical evidence. The great American philosopher Charles Sanders Peirce called it the irritation of doubt.

And where better to start this unsettling journey than a visit to the NSW Bureau of Transport Statistics website?

We can be lulled by the belief that the "main game" in Sydney's transport is moving people into and out of the central business district. That's probably why Sydney's transport network is a radial pattern, funnelling passengers on trains, ferries and conga lines of buses into the small area bounded by Circular Quay to the north and Central Station to the south.  

There can also be a lingering assumption that most trips are journeys to and from work.

None of these beliefs are true.

According to the BTS Household Travel Survey, Sydney's residents make an average of 16.7 million trips each week-day. Even if the Central Business District is broadly defined as the Sydney LGA (covering the area from Alexandria (south) to Circular Quay (north) and from Glebe (west) to Moore Park (east)), then this is the destination of just 10% of those trips. 

Some may pretend that travel that is not work related is unimportant, or "discretionary" (perhaps an "indulgence"), but we are human beings and as humans we meet friends, shop, watch movies and go on picnics. The data shows that commuting to work is not the main purpose of our travel. There are more shopping trips each day than journeys to work!    

   
The BTS Journey to Work data provides a more detailed picture of commuter travel.  This shows that only 12 per cent of jobs in Sydney are located in the central business district.

If travelling to work in the CBD is such a small component of all our travel needs, why is Sydney's public transport network so focussed on the CBD?       

This dose of reality goes some way to explain why the private motor car is still king and public transport makes up only 12% of the trips we make each week-day. It is not taking us where we need to go. This is a more fundamental cause of low mode share in this city than, say, the inconvenience of carrying whitegoods on a train. 

We need only look to continental Europe to see what needs to be done to achieve the mode shift necessary to make Sydney a more liveable, less congested city - designing public transport so residents can get from wherever they are to wherever they need to go, at a time that suits them. This, in combination with a little demand management of private car use, can lead to some habit changing about when to ride a bus or catch a train.  

There must also be a rethink about fare structures to encourage greater use of public transport, not ration it, which is what the "pay as you go" Opal card does. And let's not penalise transfers between modes nor differentiate prices based on mode. 

None of this is about extravagant infrastructure projects. The city's mobility problems will not be solved by a new cable car to Barangaroo or a big new rail project. We just need a change in thinking, as uncomfortable as that may be.