As many of you will know, Auckland Transport (AT) recently released the draft Regional Public Transport Plan (RPTP). One of the more interesting parts of the plan relates to fare systems, which proposes to establish several geographic zones that more or less radiate out from the city centre, as well as some changes to ticket products.
You can read the draft RPTP here and more about the zones in particular in this post.
In this post I want to present some of my thoughts on fares. I should say from the outset that I fully understand that I may not be right, nor that everyone will agree, but I want to get these views out there so that we can collectively come to a more informed position than where we started.
First let’s consider the following comment from Lennart:
Why have zones at all? The HOP card should make it easy to calculate fares on the number of km’s traveled. If you include a “tag on fare” as a minimum fare and a maximum fare for travel within the city boundaries you should be done. Can’t be that difficult to let the HOP card calculate the distance traveled between the stop you tagged on and the stop you tagged off…
Is Lennart correct when he suggests that “it’s not that difficult” to implement a distance based fare system with minimum and maximum fares? Yes he is, and such systems already exist (and seem to work well) in cities such as Singapore and Amsterdam. For those who have not been exposed to them, distance based fare systems work as follows:
- When you tag-on you immediately incur a “flag fall” fare of say $2; and
- When you tag-off HOP multiplies the distance traveled by a cost per km, say $0.20.
So, if you have travel 5km then your fare would be $2 + (5km*$0.20) = $3. The key advantage of distance-based fare systems is that you remove the need for zones altogether and all the boundary problems that they entail. Boundary issues will always end up disadvantaging someone. Not having zones also remove the whiff of political favouritism, and also removes the incentive for communities to lobby to be in one zone or another.
One of the more interesting developments in distance-based fares is that you can define distance in terms of the shortest alternative route by road between the stops where you tag on/off stops as if you were to make your trip by car. This means that the fare you pay is quite closely related to the main transport mode with which we want to compete, i.e. cars. This effect is useful, I think, in situations where Auckland’s geography and/or road network mean that there is a much more direct route than what is taken by public transport.
- Myth #1 – If your bus takes a squiggly route then you will get charged more. Not true, because the distance you travel is calculated based on the stop where you tag on and off, not the route the vehicle travels between those two stops.
- Myth #2 – If you travel long distances you will get pinged. Not true, because the “flag fall” means that short trips pay a higher cost per kilometre. The higher the flag fall component the more you favour long distance trips, and vice versa.
- Myth #3 – It does not support an integrated network. Not true, because the flag fall component can be waived when you get off one service and onto another within a specified period of time, e.g. 1 hour. So transfers remain free, and if you do transfer then the distance could be calculated based on your ultimate end stop – not your intermediary stops, so that you don’t get pinged when the network takes you out of the way in order to connect.
- A distance-based fare system based on stored-value HOP cards.
- Fares are calculated in two-parts: a flag-fall component plus a variable component. The variable component simply multiplies the distance you travel by a rate per kilometre.
- “Bulk buy” or “frequent traveler” discounts/caps can then be applied once people reach certain levels of expenditure or travel, e.g.:
- Daily/monthly limits, after which a discount or cap is automatically applied to subsequent trips. This replicates what our periodical tickets are doing now, except that you don’t have to decide in advance whether you need one. I think this is a huge advantage; and/or
- Regular trips (e.g. 10 per week or 40 per month) have a “frequent traveler” cap that automatically kicks in when, for example, people make the same trip more than 4 times per week or 20 times per month.
This fare system simultaneously removes the need for both zones and periodical tickets. Passengers can travel knowing that the system will automatically calculate the best fare product for them given their travel patterns.
I know many of my fellow bloggers will raise other objections to distance-based fares, so I will now move onto another issue with fares and let them do the hard work of balancing my ideas 
.
Another issue I wish to raise relates to statements along the lines of “we should try to get everyone onto monthly passes“.
Such statements imply we should price monthly passes extremely competitively, so to encourage people to buy them. The problem I have with this idea stems from differences in the types of people that tend to buy different ticket types. First, monthly passes tend to appeal to commuters. Commuters, in turn, tend to 1) travel at peak times and 2) have a job. Or to use economic jargon, they tend to impose high marginal costs while also having high willingness to pay.
On the other hand, passengers who travel by stored-value (or cash) are a more more diverse group and difficult to pin down but I suspect that a greater proportion of them are from lower income households. Suggesting we should discount monthly passes so as to “encourage” people to switch from stored value seems to presume that people who use stored value really want (and could benefit from) a monthly pass.
Personally, I’m not convinced such – and here’s a teary-eyed example of why:
Violet worked her whole life and never had time to learn how to drive, let alone the money to buy a car. As a result Violet always relied on public transport, even when she was a solo mother raising three young children and working three jobs, 7 days per week (as an aside, at that time women were paid less than half of what a man would for the same job). When she retired, Violent bought a unit in Glen Eden. Despite being wholly reliant on public transport, Violet was never in a position where she “needed” a monthly pass because she simply did not travel that much. So Violet would always pay by stored value.
The main point of this (true) story is this: Violet (my grandmother) had no need for a monthly pass no matter what the “discount”. In the long run the only impact that heavily discounted monthly passes would have for people like my grandmother is that it would tend to push up the cost of other tickets, namely cash fares and stored value travel. So I would resist the suggestion that monthly tickets (if we even need them) should be discounted to the point where they inflate the cost of other tickets, especially stored value – because to do so is likely to be regressive.
Finally, I’d like to make a small point of supporting more targeted concessions. For example, I don’t think that SuperGold card holders should travel for free and certainly not during the afternoon peak. Instead, I would support the idea of a more comprehensive “community discount” for anyone who is on a benefit or comes from a low income household. And a discount for off-peak travel, which I note is identified in the RPTP, is something I would support some analysis of.
One interesting thing to note is that if we were to have a system where more people used stored value (as proposed earlier in this post), then it becomes much more effective to deliver these kinds of targeted discounts. It’s quite hard, for example, to combine targeted discounts and/or concessions with monthly passes, unless you complicate the system by having multiple passes (e.g. a standard and an off-peak monthly pass).
That’s the end of my rant really – which ultimately just comes down to three things:
- Distance-based systems have some advantages over zone-based systems;
- Discounts for monthly passes can be perniciously regressive if they result in higher prices for cash or stored-value tickets ; and
- More targeted concessions and discounts – including a new “community discount” for people who live in low-income households, as well as an off-peak discount.
Discuss and disagree; I’d appreciate your feedback because all of this will ultimately inform my submission on the draft RPTP. And I should finish by saying that I think the draft RPTP is a great advance; credit to AT for seizing the initiative on some many fronts.
Even though I prefer distance over zone based systems, I fully acknowledge that the proposed zone structure are a vast improvement on the current fare system and do in most places make logical sense.
I think you raise a good point about monthly fares not being appropriate for everyone, particularly from a socio-economic point of view. One of the saddest observations I’ve made about people catching PT is that the poorest areas also tend to be the areas that have the highest proportion of passengers paying with cash, rather than using a card (whether that card be stored value, multi-journey or whatever). The very people who most need the discounts are missing out, presumably because they lead busy lives and can’t get around to sorting out a card or because they really struggled to justify shelling out $10 for a card in the first place.
I do also agree that having a zonal system rather than a distance system will lead to winners and losers in terms of shorter trips being more expensive than longer trips – just because of a zone boundary. I think that’s part of the reason why I have a preference for more, smaller, zones compared with what AT’s proposing. At least with more zones which are smaller, the unfairness can be minimised while still getting the simplicity benefits of a zone-based system.
Yes I’ve observed that too – I wonder whether primary/secondary school students should be given a HOP card?
Consider the following options:
1. A system based around many small zones with daily/weekly/monthly ticket types and stored value; or
2. A system based on a distance based fare with automatic discounts for bulk buy/regular travel.
In my mind #2 is considerably simpler than #1, but you go for #1. Such is the colour of life!
I agree on that. Many small zones has all the complexities of stage or distance systems, with the added complexity of border effects all over the show, and without and practical benefit from in-zone travel.
“So transfers remain free, but the distance you travel to get there does not (although if you do transfer then the distance could be calculated based on your ultimate end stop – not your intermediary stops, so that you don’t get pinged when the network takes you out of the way in order to connect).”
That is my main concern with distance based pricing, but charging only one flag fall per journey and taking the distance from initial origin to final destination would absolve that. My question is do any systems actually do that, and how do they differentiate between one long linked journey and a series of short separate journeys? I guess a fairly generous transfer time allowance would do it, but then people could also game the system for short return errands (although whether many people would bother is a different story).
If you don’t account for that then your ‘wiggly line’ myth does actually exists in a way, as a ‘ou of the way transfer point’ problem. That would result in journeys via transfer costing more that one seat trips between the same places. That maintains some transfer penalty, and could make people demand more ‘everywhere to everywhere’ services.
Anyway, I’m under firm instructions not to get stuck in here, so I’ll put my own thoughts together for a separate post.
Yes I’ve since found out that the technology is sophisticated enough to account for the impacts of intermediary connections on travel distance. It requires some complex business rules, but nothing that some good software programmers can’t code up. While such complexity makes it hard for you to calculate it in your head, once you’ve made a trip once (or have jumped onto a journey planner) then you will know what it costs.
To be honest despite my passionate support for distance based systems I’m fairly open to the zone structure Auckland seems to be heading towards (with a few tweaks on the North Shore). What I will be interested in (and something we can analyse after it’s implemented) is how much the distortions in fares introduced by the zone boundaries suppress demand for some trips. It may be that people simply ignore them and travel without thinking about the zones, so maybe I’m brewing up a storm in a tea-cup.
But I’d be particularly interested in how focus groups of PT users would respond to the two fare systems, because what we’re discussing here is what other people want, and that’s always dangerous when you’re a weird outlier such as myself
.
when travelling frequently by PT, I always chose to use stored value because of its flexibility, monthlies are good for people who commute, but if you are to live a PT lifestyle and use the bus train or ferry as the first choice for any trip, then stored value is the obvious choice – isn’t that lifestyle what we should be advocating?
second point, while a tag on/off distance based system is good for card users, zonal fares with all their issues probably work better for cash fare payers, simpler for the driver to process (although a possible technology solution would be having the ticket machine linked to a route map on a touch screen where you could enter a cash fare user’s destination which then calculates the fare and issues a ticket)
I guess my major concern (I’m starting to sound like a cracked record, I know) is the CBD fixation of the proposed fare system, it should be more orientated to travel between the metropolitan centres including the CBD
Nothing (except simplicity) says you have to have a linear relationship between distance and price (eg. X cents per km). Instead you could have Xc per km for the first 2km, then Yc per km for the next 2km, then Zc per km for the next 2km… or whatever. A ‘flag fall’ will always penalise short journeys more than long journeys, but by having a graduated pricing plan for distance travelled this could also be avoided. Something similar could also apply for very long journeys.
Yes that is correct.
I think your point about commuter discounts is particularly valid. We already have two such distortions in the car system: Subsidised parking and Early Bird parking. If any parkers should receive discounts it shouldn’t be the peak driving commuters who, as you say of peak PT commuters both cause the highest marginal costs on society [over-building of road network to serve the peak] and have the highest ability to pay market rates. And as for employer subsidised parking hopefully this will diminish with the FBT just announced, although I see it already looks like it is getting watered down.
You present a good case for your tidy-minded system, transfer issues aside [Nick's comment above], but I also agree that an improved zone map to the one released is not too difficult to conceive and it’s clear that Auckland is geographically extremely well suited to a zone system- certainly more so than London for example- with its natural boundary features.
Agreed – even though I personally prefer distance based systems, I freely admit that Auckland is uniquely well-suited to a zone based system.
But I do want to raise awareness of it because it may be something we want to consider in the future, especially if we start pushing more long distance services north/south.
I also sense that once the numbers are run on what the fares would actually have to be to introduce the zone system in a fare-neutral way, some of the results will be pretty scary and AT will possibly want to look at a bunch of alternatives a bit closer.
Yes I got that feeling too looking at the zones – large zones are all very well in principle, but what they mean is that the jumps in fares crossing zone boundaries are more dramatic, which in turn means that the relative importance of zone boundaries will increase. Zones are a bummer if you travel one bus stop too far!
Distance based charging though – less of a problem
.
Whats so natural about Londons zones? Always seemed arbitrary to me
But remember that PT travellers during the peak create the greatest positive externalities through their decongestion benefits.This is probably particularly true for CBD-bound radial commuters.
That’s only if you believe in decongestion benefits
. But yes, peak commuters do generate more external benefits.
I think it probably works on two levels. In the short term, focussing on getting peak radial commuters off congested motorways and onto PT creates strong decongestion benefits (or the inverse “growth with the same level of traffic” benefit). But in the long therm, a focus on getting local and crosstown all-day travel onto PT would have the greatest actual decongestion benefit, as people start to do away with the second and third car per household and generally just drive less overall.
One thing with monthly or weekly passes is it can add to speed of loading on the bus. A great example of this is on the 881 at Constellation. A large number of students have weekly passes or hop cards. They can board almost double file – monthly passes shown to driver, hop cards tagging on. This way 30 people can board in 90 seconds ( I have timed the avergae baord time here and it is about 3 to 5 seconds per person). I expect this woud be twice as fast as the NEX at Britomart which seems painfully slow in comparison. I think I recall reading elasticity of bus time is about -0.5 meaning slow a trip down 1% and 0.5% less people use it. On a 50 minute trip with 60 people boarding adding 4 seconds to boarding time will add 4 minutes to the trip (8% longer) . Therfore you would expect 4% less people ( 2 people per bus) to use the service.
I think you will find that the days of “showing” the monthly pass to the driver are gone – all passes will need to be swiped in the future. But I agree with the importance of boarding times, that’s why I’d suggest Auckland really needs to consider:
1. No change given onboard – the norm in most North American systems
2. Automatic coin counters – mean that cash paying customers board almost as fast as electronic card payments
3. Double door entry/exit – allows for two people to swipe on/off simultaneously.
I’m agnostic about zones or distance charging, as long as Waiheke is included in the same zone as the suburb that is the same distance from the CBD. Currently Half Moon Bay to CBD by ferry costs $13.80 return while from Waiheke (only a third of the distance further) it costs $35. We can’t wait to be integrated!
*can’t*, of course. Where is the edit button, admin?
Just edited it for you
If we can’t be integrated, can we get rid of the Supergold card integration? It doesn’t make sense to have transport mode and ticket integration for one age group only.
And also, why give Waiheke Buses an $300,000 AT subsidy but keep it out of integration? Its fares for distance travelled are already the highest in the region. A 2-stage fare, which can be as little as 1km of travel is $3.10.
I agree.
We could bring ferries into the zone system, and treat the Waiheke buses as their own zone. One way to do it would be to assume that ferries pass through all the zones that would be applicable with the shortest equivalent trip by road. So on the harbour ferries it’s pretty straightforward to work out, probably two zones for the Shore ones and three for the west and east Auckland runs, a four for maraetai. in the case of Waiheke you might start in the local zone, then assume the most direct route passes to Maraetai, then however many zones from there to downtown. That would probably make Waiheke the equivalent of a train from Pukekohe or a bus from Helensville.
Yes, that is acceptable. Pukekohe is the same distance to the CBD as Waiheke – and the routes by ferry and train are straightforward. So they should have the same far structure.
Don’t know how you worked that one out. Pukekohe – Britomart = 41km, Matiatia – Britomart = 21km. Both as the crow files
Charging on a per km basis will work well if all passengers pay on a HOP card. The system would lend it self to a range of shoulder peak and off peak charging rates, or multiple fare caps based on time of first boarding.
This gives the ability to manipulate demand without totally confusing the passengers
When Charging per km It become quite difficult to sell individual cash fares, particularly on buses. Or should we get rid of cash fares altogether? But that becomes unfair to those on low incomes, tourists and other one off users.
On monthly fares just because users have a regular income doesn’t mean they can afford higher PT fares. Peak period commuters are the bread and butter of PT and it is a market that PT can actually compete with the cars. Yes we build PT for peak demand and it costs a lot of money, Without peak demand we don’t need rail or busways so we end up with a bus system that is used as transport of last resort.
Yes cash fares do become more difficult with a distance based system.
For rail you won’t be able to pay onboard, so immediately you’re either a) using a machine or b) dealing with a person. So they just need to specify their destination station and pay accordingly. For bus a time-limited single-trip cash fare would probably work best.
My comments are general: On average, people with monthly passes will have a job and therefore be less sensitive to price. And that’s supported by evidence showing that off-peak price elasticities are typically much higher than peak period elasticities. That’s because we’re pinging our nanas more in the off-peak.
I don’t think you can properly claim that peak hour commuters are the ones that cost the most to serve at the margin – and as an argument for not givng those customers discounts, I think it is short-sighted. PT systems, especially fledgling ones like Auckland’s need volume to keep costs under better control, and bulk discounts of one kind or another (commuter passes, discounts for pre-paying and so on) are probably the best way to build volumes to where we need them to be. In any cases, surely the pre-pay aspects of users buying monthly passes and so on yield a cashflow benefit that it’s worth granting a discount for?
I think I can say that peak hour commuters cost more – half the bus fleet, three quarters of the rail rolling, and the last $2 billion in CAPEX are for peak commuters. Yes PT benefits from economies of scale, but these are not as strong as you think. Even on rail our service costs have grown only slightly slower than revenue. That will decouple a little more with EMUs and HOP (reducing labour input into services) but still not mean that peak commuters are the saviour of the system.
Note that I’m not arguing against bulk buy discounts; I’m arguing against pre-determined monthly passes. Think about these options:
1. People on stored value that caps out at the monthly (or annual) rate; or
2. People have to predetermine whether they want a monthly pass in a particular month.
There’s little cashflow different between the two: In fact in #1 is probably better because the loading of stored value would be a smoother profile than lumpy monthly pass revenue. And I think users would definitely go with #1 – and why not? It gives you the flexibility of stored value with forgoing the benefits of a monthly pass, in the months where you need it.
I can live with your option 1, so even if we come at it from different perspectives, we’re on the same page. I do think it’s important that regular users get discounts, regardless of when they travel. After all, if that increases the ‘baseline’ volume, it presumably moves the point at which an extra passenger starts to become expensive… I do really think that Auckland’s adoption of PT is potentially quite fragile and needs to encouraged for the collective good. Call me a (market) socialist, if you like, but there it is…
The major problem with as the crow flies distance charging is: how do you handle the case where the connecting leg results in a reduction in the distance travelled? If you refund that money, you allow people to make return journeys within the transfer window for the price of the flag fall.
Not crow flies distance charging; “shortest path” distance charging – big difference.
The problem you identify simply requires business rules that detect return trip behaviour so that it’s “not really worth your while” e.g:
1. Business rules that pick up when someone is making a return trip to the same and/or nearby stop.
2. Business rules that pick up when someone is making a return trip via the same route but in opposing direction.
Basically you just need to code up rules that pick up when someone is making a trip that geometrically looks like a return trip.
I disagree that the problem is coding up the business rules.
Anyway, even if “shortest path” business rules can be devised – they present certain problems for movements which cannot be done driving, such as travelling along train lines – they represent confusion for the average user.
There will still be cases where an acute angle multilegged trip results in a negative “shortest path” contribution from the second or later sector.
All systems have limitations. Zonal isn’t too bad and neither is charging by route distance – circuitous routes should be removed from the system anyway.
Negative contributions for subsequent legs can be managed.
Let’s consider an example to make sure we understand what each other is saying:
- If A->B is 10km and B->C is 5km then someone travels the equivalent of 15km.
- But A->C by shortest path is only 7km. That means they get charged 7km, which is less than the 10km for the first trip.
So in that case all you need are business rules that provide credit your balance when you tag off the second trip. Easy peasy lemon squeasy.
P.s. Agree with your suggestion that all systems have their limitations and that circuitous routes should be removed from the system.
Once you do that, return trips within the transfer window are likely to be able to attract a negative charge. Where I live (not AKL), I can walk about 600m to pick up an inbound bus. On the outbound I sometimes use a bus going a completely different way and walk 800-1000m from that bus from the exact opposite direction. So if I did so within the transfer window, that would be considered a connection with a negative charge. At least a zone system will charge the full price for one direction.
This may have been answered elsewhere, but can someone with more brains than I please explain what is good or bad about time based ticketing i.e. one ticket allows you two hours travel anywhere on the network – ANYWHERE. So if you complete your shopping journey within 2 hours you’re sweet. If you tack on another chore that takes you over two hours, then you purchase another ticket.
And I’ve no idea why my post at 3.52pm is placed above earlier posts.
Time based ticketing is good when you have a system where all vehicles travel at approximately the same speed and there is not so much variety in trip length. E.g. Edinburgh just has a bus system, they all putter along about the same speed, and you can get from one side of the city in about an hour.
In Auckland it’s more complicated because two hours on the rail gets you all the way from Pukekohe to Britomart, which is a $10+ trip right now. But on the other hand everyone knows their destination station. So distance-based cash fares are more appropriate to the rail/busway network.
On the other hand time-based tickets are more suited to the bus network – where there’s lots of stops (for which people don’t know the name).
Thanks, appreciated.
Hopefully at some point to assist those writing up submissions to the RPTP someone could kindly summarise all the points made on this page about the fare zoning proposals…
Stu, I have a few questions about your grandmother Violet. When she was working three different jobs, seven days a week, how was she getting to these jobs? Walking or public transport? Did she need to travel to more than one job on the same day? My experience with monthly passes is that they’re usually priced around the equivalent of a weekday’s commuters travel habits: so the same price as a return journey, 20 days a month. If those are the only trips you take, the advantage is purely convenience – but any extra trips you take on top of that become effectively “free”. It’s quite possible that a monthly card could have been very advantageous for Violet: she was travelling 7 days a week instead of 5, she might have had to travel between jobs on the same day, and trips she had to make on behalf of her children wouldn’t cost her any extra. Monthly passes really incentivise you to use PT for an many extra journeys as possible to get value out of the investment.
A zone-based system can support both pay-per-trip and monthly passes as payment methods, so users can select the option that best suits them. A distance-based charging system categorically rules out the possibility of weekly or monthly passes.
To me, zone-based systems offer simplicity and predictability. With a distance-based system, really the only way to know in advance how much you’ll pay is to run the exact route before hand and see how much you get charged when you step off.
I think that perhaps the point Stu was trying to make is that while a monthly pass would have certainly been best for Violet, she just didn’t have a big chunk of money spare to pay for it all at once.
I would question the use of on-the-ground distance charging. It’s a piece of cake to write the rules to calculate straight-line distance based on coordinates of tag-on/tag-off transactions, giving an instant bonus over private vehicle traffic by calculating using a route that is impossible by any form of transport, and if there’s a window to allow for transfers it’s simple for the system to tot up the various pairs and determine the final cost for any given journey. It is, in fact, easier to do it that way than to determine the shortest road-bound distance between two points, and means that people who are forced by geography to travel a longer distance (going around the edge of the harbour, for example) are not disadvantaged compared to people who have the option of taking a shorter journey. I could see some edge cases that might be odd, such as a person who takes the Half Moon Bay ferry to town being charged the same as the person who takes a bus, but that’s only a problem if we are trying to treat use of public transport as a journey rather than an enabler of journeys.
Stu I will give my favourite example for you to get your thoughts on as I don’t think you have answered it when I have raised it before. Due to our unique geography there there a few places where we would get some interesting anomalies, in particular would be the case of Henderson and Sunnyvale train stations. From a PT perspective, Sunnyvale is closer to town being a stop earlier than Henderson yet road it is further away. If you charge based on the closest road distance like you suggest, this leaves you with the odd situation where you get charged more for getting off a service earlier.
On a separate note, why couldn’t you have both zonal and distance based fare options available and let people choose which one they would prefer. If it could be done as a option if you register your card then we can have the best of both worlds as people would just select what was best for them. Each option would have to be priced in a way that ended up roughly fare neutral. Doing it this way you might get someone that mainly does a short distance trip that crosses a zonal boundry setting up their card to use distance based fares but then someone else who mainly does trips close to home may opt for a zonal based fare system on their card.
A bit off point, but has anyone worked out what the price would have to be based on the zones as set out in the paper. With the zones being so big, and AT after a 50% recovery, short trips are going to be very expensive?!?!?!
Does anyone have a link for where this has be looked into?
not quite OT, something Octopus in Hong Kong does, is a collection of discount processor machines scattered around the city in places where the authorities would like to encourage more walking (instead of mechanised transport of some sort) users swipe their cards as they walk past for a $1HKD discount off their next trip.
Large zones, big ‘overlaps’, and low fares.
I love Auckland. But I want Melbourne’s fare system.