Do We Really Want Fares to be “Fair”?

For blog item on fares Fares are a tough issue for transit agencies and the politicians they answer to.  Environmentalists and people concerned with social inclusion want fares to be low, or even eliminated entirely.  The financial bottom line of most agencies requires a certain level of fare revenue.  Many agencies worry about their “farebox return” (sometimes opaquely called the “operating ratio”), that is, the percentage of operating cost paid for by fares.  If it’s over 100%, you have a profitable business, but in the developed world it’s usually much less, which is why we have government subsidies.

In this post, I want to focus on fare structures rather than fare levels. You can separate the two issues by thinking in terms of an average fare per passenger that a system needs to collect in order to balance its books.  It may collect this by charging some riders more and other less, but it needs to come out at that point on average.  That’s a fare level.  You can argue that it should be higher or lower overall, but that’s a separate question from fare structure.  Fare structure is about how what you charge to different kinds of people, for different kinds of trips, in order to hit the average fare overall.

For example, there seems to be a fairly broad consensus in most developed countries that there should be discounts for senior citizens, disabled persons, and (in some cases) students.  These discounts need to be offset by charging a slightly above-average fare to everyone else.

But these discounts are the easiest question of fare structure.  The hard questions include:

  • Should people pay more to travel further?  If so, how do we keep track of this?  Some systems keep track of very short intervals of distance, increasing the fare slightly with each unit.  Others use zone systems and count the number of zone boundaries crossed.  There’s huge variety of policies within each of these.
  • Should people pay extra if their trip requires a connection?  There’s a strong argument to be made that connections are to the benefit of the transit agency and the city, and should be encouraged rather than penalized.  But there are also logistical and security issues — especially with paper transfer slips — that have led many systems away from this practice, not to mention the jurisdictional issues that come up around connections from one agency’s service to another.
  • Should there be “frequent rider discounts,” such as a reduced price for tickets bought in bulk or a discounted “pass” for unlimited rides in a day, or a week, or a month, or a year?

All of these issues are matters of active debate in the industry, and there are still huge differences from one agency to another.

After dealing with these issues for years as a consultant, I’m beginning to suspect that we might do well to set aside the idea of “fairness” or “equity” in fares.  When we’re talking about fares, everyone wants (a) to pay less and (b) to perceive that the system is fair or equitable. But a simple thought experiment should tell us that we don’t really want fairness at all:

Suppose that in our perfect transit system, a fare policy really did charge everyone for exactly
their fair share of the cost of providing the service that they use.  If you really wanted to do that, you’d do exactly what you’d do if you and a few other people were sharing a ride in a private vehicle.  You’d want everyone to pay the cost of running the trip divided by the number of people riding.

Clearly, your fare would depend on the distance you travel, but it would depend much more dramatically on the number of other passengers on your bus, train, or ferry.  If you were the only passenger on a 40-seat bus, your fare would be 40 times higher than if every seat was filled.  Your fare would be calculated in tiny increments extending from one bus stop or train station to the next.  As the bus or train filled up, the increments would get cheaper.  As it emptied out, their cost would go up.  The sum of all these increments is what you’d pay for the trip.

As fare technology continues to develop, it is not hard to imagine a future very-smart card system that could actually calculate this “perfectly equitable fare” in real time.   It would know not just where you started and ended your trip, but also how many other smart cards were on the same vehicle with you.  At the end of your trip it would debit your card for the cost of each increment of the trip, divided by the number of people who used that increment.

This system, and only this system, could be called “equitable” in the sense that everyone, at every moment, would be paying exactly their fair share of the costs of running their service.

Obviously, no transit agency does anything like this, and I have yet to see it seriously proposed even for the new era of smart-cards.  I don’t propose it myself, but I do think it’s a valid representation of fairness on an analogy to the way we share costs “fairly” among ourselves when sharing some other good, whether it’s a dinner at a restaurant or the fuel costs of a trip we take together.

As a practical policy it would have some bizarre consequences, but perhaps these seem bizarre only because we’re not used to them.  People would stop cursing the transit agency for offering insufficient rush-hour capacity.  Instead, they’d actively try to pack themselves onto already-crowded buses and trains.  At the other extreme, low-ridership services would become prohibitively expensive to ride, so those services would vanish fast.  For better or worse, the policy would certainly yield a very high-ridership transit system, because high-ridership service would be the only service most people could afford to use.

But it would all be perfectly fair, wouldn’t it?

I think this thought experiment tells us several important things:

  • Equity begets complexity.  The more precisely equitable a fare system tries to be, the more complicated it becomes.   For example, cities that charge for very small units of distance have much more complicated fares than cities that charge a flat fare of have just a few fare zones.  Our “perfectly equitable fare” achieve the extreme of complexity: It is so intricate that nobody can know, when they set out, what their fare will
    turn out to be.
  • Complexity for the passenger is different from complexity for the agency,
    largely thanks to smart cards.  Until the advent of smart cards, agencies and their customers had a shared interest in fare simplicity, because a system that was too complex for customers to understand was usually also too complex to administer and enforce.  Smart cards now make it possible to administer and enforce extremely complex fare systems.  This does not necessarily mean that we should welcome these systems, because ….
  • Fare complexity will still be a major problem for people who need to know their fare in advance.  Some customers have enough financial security that they don’t need to worry about small increments of cost; for such a customer, the hassle of figuring out complicated fares is a much bigger ridership deterrent than the amounts of money involved.  Smart cards will liberate this passenger from that hassle, regardless of the fare system used.  However, some passengers will always need the ability to predict what they will spend. Others may not need this ability but will insist that they have a right to it.  Fare complexity will continue to weigh primarily on these customers, and if the agency chooses to continue to value simplicity in its fare structure, it is these customers who will benefit.

Equity in the sense of “fair share of cost” is clearly related to how many people are using a given service, yet this enormous factor is almost never expressed in conventional fare structures. It can be crudely approximated, for example by charging more for
services that carry fewer passengers by design, such as Dial-a-Rides and other small-bus services, but this is almost never done in current fare practice.  As noted above, with each
such factor, we increase the equity of a fare structure but also increase its
complexity.

Finally, any fare system that is simpler than the one described above is an extremely crude approximation of equity. Any reasonably simple fare system will be unfair to someone by any commonsense definition of fairness.  Thus, any transit agency that is unprepared to implement “perfectly equitable” fares, as described above, must accept being accused of inequity as an inevitable consequence of whatever decision it makes.  This does not mean that the decision is unimportant or can be taken lightly, but only that we must be realistic about the potential for achieving “equity” in any measurable sense.

UPDATE:  See the comments ection for some great further discussion on this!

24 Responses to Do We Really Want Fares to be “Fair”?

  1. Mike May 13, 2009 at 3:37 am #

    The London Oyster system seems to do an admirable job at combining complicated fare structures ( underground zones + private rail services ) with the simple ( flat fee bus service independent of distance ) for casual travel. After a certain amount of travel then costs ceiling out at day rates. Regular commuter journeys can also be accommodated with discounted monthly or annual fares managed with the same card ( or you could always opt for a second card ).
    The flexibility this yields makes London much more connected and public transport is more attractive than taking a car. I can have day outings in London that are inconceivable in Sydney because of the fare penalties, usually of the form train from X to Y, walk to Z, another train or bus back to X.
    In Sydney I look at the cost vs opportunity options for me to travel from the Inner West into the CBD for an evening’s entertainment. A one-way bus fare is roughly the same as a return train fare. At the start of the evening, the bus is generally going to be a slower option, so it’s better to take a train. However at the end of an evening performance, the train frequency has dropped off to the point where there is such a long wait that I may as well catch a bus and throw away the return leg of the train fare.
    My interim solution is to drive part-way and park somewhere near Newtown or Erskineville stations, which are on different branch lines. Late in the evening I can mitigate an extra 20-30 minute wait for the right branch train with the knowledge that it’s only 5 mins walk between these two stations if I take an earlier train for another line.
    If I were in London I wouldn’t even be thinking about this.
    On another note, the costs and fairness you talk about above are worked out within the simple context of the provision of a transport service. For publicly funded schemes, this means that subsidising public transport is not weighed against savings in building road infrastructure, improved health due to less traffic pollution or other harder-to-define benefits of living in a more connected city. It’s so much harder to get friends together in a city that emphasises car use over public means.

  2. Mike May 13, 2009 at 3:38 am #

    The London Oyster system seems to do an admirable job at combining complicated fare structures ( underground zones + private rail services ) with the simple ( flat fee bus service independent of distance ) for casual travel. After a certain amount of travel then costs ceiling out at day rates. Regular commuter journeys can also be accommodated with discounted monthly or annual fares managed with the same card ( or you could always opt for a second card ).
    The flexibility this yields makes London much more connected and public transport is more attractive than taking a car. I can have day outings in London that are inconceivable in Sydney because of the fare penalties, usually of the form train from X to Y, walk to Z, another train or bus back to X.
    In Sydney I look at the cost vs opportunity options for me to travel from the Inner West into the CBD for an evening’s entertainment. A one-way bus fare is roughly the same as a return train fare. At the start of the evening, the bus is generally going to be a slower option, so it’s better to take a train. However at the end of an evening performance, the train frequency has dropped off to the point where there is such a long wait that I may as well catch a bus and throw away the return leg of the train fare.
    My interim solution is to drive part-way and park somewhere near Newtown or Erskineville stations, which are on different branch lines. Late in the evening I can mitigate an extra 20-30 minute wait for the right branch train with the knowledge that it’s only 5 mins walk between these two stations if I take an earlier train for another line.
    If I were in London I wouldn’t even be thinking about this.
    On another note, the costs and fairness you talk about above are worked out within the simple context of the provision of a transport service. For publicly funded schemes, this means that subsidising public transport is not weighed against savings in building road infrastructure, improved health due to less traffic pollution or other harder-to-define benefits of living in a more connected city. It’s so much harder to get friends together in a city that emphasises car use over public means.

  3. Peter Parker May 13, 2009 at 6:30 am #

    Jarrett – thought-provoking post. A couple of points:
    1. A parallel is made between a trip on public transport and a restaurant meal. There is however a major difference that I think affects how it should be charged.
    If you have a meal served to you at the restaurant but I take it from you, then I am nourished while you starved. You are denied of food while I gain at your expense.
    If you are travelling on a moderately loaded train and I board then, unlike the food incident, you suffer no loss of amenity from me also having the benefit of that same train trip.
    If the train is crowded and you are standing then comfort for you is reduced as others board as more bodies are pressing against yours. While comfort is compromised, all those on the train still get the benefit of the train trip.
    It is only when the train is so crowded and passengers are left behind at stations that we have an analogy similar to the food example; ie those who get the food (board the train) get the benefit while those who are denied food (cannot board) suffer.
    At other times public transport is a different type of good, more analogous to knowledge; ie just because you benefit does not mean I miss out.
    2. The cost of providing the service.
    When patronage is light this is very high per passenger as you point out.
    As patronage builds up then it drops per passenger.
    But only to a point.
    If you have enough buses/trains to run more services just by taking on a few more drivers and depot staff then there is a cost to more passengers but it’s fairly low.
    If you need to buy more rolling stock (that might only be fully utilised during a few hours of the day) then the cost increases further.
    Where complex infrastructure is needed to handle more trains then the cost goes up even more as per: //melbourneontransit.blogspot.com/2008/05/service-costs-and-spreading-peak-what.html
    This is the problem we currently face in Melbourne – our network is stretched and patronage is booming.
    Fares do not cover operating costs (let alone capital improvements) but if they did then it would seem rational to seek to spread the peak.
    Fare pricing to relate to cost of service would charge peak travel higher than off-peak travel so that peak travellers make a contribution to the much higher marginal cost of their trips.
    If you believe that people respond to economic incentives (not answers in themselves, if shoulder-peak service levels aren’t up to scratch) this would increase efficiency as the peak would be broadened and flattened. A more steady and less peaky patronage pattern would allow more people to be carried with a given amount of infrastructure, rolling stock and drivers.
    While a packed train may have a lower cost per passenger than a quieter train, if the train is run during peak hour then due to its much higher marginal cost of running then cost per passenger might not necessarily be lower.
    Under an economically-rational fare system, a passenger who switches from a packed peak train to a quieter shoulder peak train should be rewarded through a lower fare rather than a higher fare as your method proposes.
    Relatively dearer off-peak fares would encourage even more peaky travel patterns, which from a driver, rolling stock and infrastructure point of view would raise running costs and decrease efficiency.
    3. Pricing quieter (ie off-peak) services more than peak services. There are equity arguments that I won’t go into here. However if we take this to extremes and let off-peak services fall by the wayside then we’re actually also harming peak travellers as well.
    As well as contributing towards an inefficient, peaky commuter-only service, you’re undermining patronage, even of people who only rarely use (say) evening trains.
    Because if you have to go home early due to an emergency (or have to stay late) and there’s no service then that makes public transport less attractive and will harm its modal share.
    As an example, here in Melbourne many of our local bus routes have been extended to run on Sundays and until 9pm. Even though you see evening buses with not many people on them, the patronage of upgraded routes has boomed, at all times of the day, not the times that the extra services were added. Routes that were not upgraded have grown a little in patronage, but by far less than routes that got the extra service hours.
    Even though you might not use certain services very often, the opportunity of being able to do so is very important and affects a decision to use public transport and contributes to patronage at all hours.
    So in summary, I don’t agree with the restaurant meal analogy, I’m not sure if you factored in the much higher cost of running peak services and the extent of the perverse incentives make basing fares on loading quite unfair and detrimental in practice.

  4. Jarrett at HumanTransit.org May 13, 2009 at 7:58 am #

    Peter.
    Note first of all that this is a thought-experiment rather than a proposal.
    But I don’t see how the sharing of a ride is different from sharing the cost of a meal. Nobody’s taking anything from anyone. We’re simply dividing the cost of the service over the number of people benefitting from it. Like all analogies there are differences in the two cases. In the meal analogy, of course, the meal is presumably bigger if more people are eating, but not necessarily so, while in the transit case the service is probably the same regardless of how many people are riding, but again not necessarily so.
    You are absolutely right about the higher cost of running peak-only services, but I didn’t mean to imply that operating costs would be treated as constant. Indeed, in the idealised world of this example, the fare calculation system is obviously so robust that it could calculate the differential cost of peak-only services and charge a higher cost accordingly. (At this point, it’s worth noting that we’re talking about a fare calculation that’s basically a form of computer modeling, and there’s no longer much technological barrier to doing this in real time.)
    Your third point raises the most interesting issue. Yes, the availability of a transit trip sometimes contributes to ridership that doesn’t happen on that trip. The best documented example is that the last trip of the night always has low ridership, but if you eliminate it the previous trip’s ridership generally falls. People don’t like being on the last trip of the night, so there has to be a nearly empty last bus for the rest of the schedule to attract riders. Similar arguments are made about the importance of midday and evening service in attracting peak riders, because it provides the “guaranteed ride home” if your commute plans change unexpectedly during the day. I am constantly making this argument against people who want to judge individual trips of a route based on ridership.
    But while these effects are hard to quantify now I wouldn’t put it past the next generation of modelling to be able to calculate all these side benefits of a service and combine them into the price. For example, in this case, part of the cost of all midday and evening service would be paid by people riding on the peak, based on the observed degree to which those services really do determine the utility of the service they’re using.
    Note also that the thought-experiment doesn’t posit that this kind of fare would be the only funding for a transit system. You could have this fare system and still choose to subsidize service in other ways to achieve the other benefits of transit that can’t be measured by ridership.
    But again, my goal is not to make a proposal but rather to make a point about fairness. My point is that while it’s impossible to objectively describe what a completely fair or equitable fare system would look like, the closest thing to one that we can describe looks utterly different from what we actually do.

  5. Aaron Antrim May 13, 2009 at 12:44 pm #

    How should transit agencies evaluate the cost of fare complexity (increased barrier to riding, especially in cases where the agency hasn’t implemented a smart card) system?
    My thought is an agency without smart fareboxes should just pick a few fares and keep the fare table simple. But it would be useful to have numbers to show them why this is the case.
    Right now I have a small rural client (~75,000 boardings a year, county pop. ~45,000) with a complicated fare structure. They have about 38 different possible fare amounts, including some that differ by $0.05. To represent the zone-based fare table in the Google Transit Feed Specification, it takes about 260 fare rules.
    This adds a few additional hours of a consultant’s time for Google Transit, and requires a surprising amount of extra paper to print on the timetable. More significant, however, are increased dwell times for passengers to pay fares and the rider confusion.
    But that’s just me, an IT consultant, spinning my wheels. Is there a study or quantitative method to prove this point?

  6. jarbury May 13, 2009 at 9:59 pm #

    I’m a big fan of “unlimited rides” fare structures. You have a series of zones and sell daily, weekly or monthly passes that offer unlimited rides within 1, 2, 3 or more zones.
    This means that people are encouraged to use public transport for trips other than commuting, and it’s effectively already paid for and free. It also provides for easy transfering between bus & train or between bus & bus.

  7. Jarrett at HumanTransit.org May 13, 2009 at 10:45 pm #

    Jarbury. Yes, I like zone systems too, and completely flat fare when you can get away with that. It’s interesting that the number of zones an agency thinks it needs doesn’t have anything to do with the size of the area. Portland has three zones, Melbourne has two, and as you infer, Sydney and Auckland effectively have hundreds.
    I also like the approach used in Portland and Melbourne, among others, whereby a single cash fare is treated as a two-hour pass.
    Simple, saves crucial operating time.

  8. Jarrett at HumanTransit.org May 13, 2009 at 10:53 pm #

    Aaron:
    A small agency could be asking its drivers, and if necessary, doing some ride-along observations to see how much fare transactions are contributing to running time. The TCRP report on Fare policy may have something on this, and I’d be confident that if it isn’t there, it doesn’t exist.
    Quantifying ridership loss due to fare complexity is almost impossible, as is quantifying ridership loss about most causes. You’re asking about the behavior of people who aren’t on the buses, and you’re asking a question with many possible answers, so that means a stated-preference telephone survey with a very large sample. In short, the research requirements would be so onerous, expensive, and error-prone that it would easy to dismiss the data if they contradict common sense. So it’s probably best to make a common-sense argument, with the addition of testimonials from drivers about how the fare system affects their work on the front line.
    How much data do you need to show you that more complicated fares are more work for both customer and driver? How much data do you need to show that if every trip a customer might make involves a different fare, you’re training your riders to use transit only for regular trips that they make all the time, and not to consider it for all their other local travel needs?

  9. arcady May 13, 2009 at 11:20 pm #

    You can try going the other way and quantifying benefits from fare simplification. One potential source of data is Rhode Island’s RIPTA, which fairly recently went from a zone system to a statewide flat fare with transfers. I think there might be quantifiable benefits to treating a single fare as a 2 hour pass as well: it enables more connections and makes things somewhat easier for passengers too.

  10. jarbury May 14, 2009 at 12:02 am #

    I’m certainly a big fan of time-based fares rather than trip-based one. Makes life much easier for transfers.
    And yes, Auckland’s fare system is hopeless. I did design them a new one though:
    http://transportblog.co.nz/2008/11/19/ticketing/

  11. fpteditors May 14, 2009 at 6:05 am #

    Keep it simple — fare-free. End sprawl, oil wars, global warming. Move to town, educate all children, give the suburbs to the organic farmers.

  12. Mike May 14, 2009 at 5:15 pm #

    “How much data do you need to show that if every trip a customer might make involves a different fare, you’re training your riders to use transit only for regular trips that they make all the time, and not to consider it for all their other local travel needs?”
    Actually the need to have approx the right change for all the conceivable fares I have in Sydney is one such deterrent. Having to plan the change in my pocket for a few casual trips is more trouble than it’s worth. Back in the car we go…

  13. Roger May 26, 2009 at 2:38 pm #

    I would disagree that equity requires a complex fare structure. It really depends on how you weigh purpose. A fare can be thought of as gaining access to the transit system. It is up to the rider to figure if the trip is worth the fare. The more complicated the fare structure, the more questions the rider needs to answer before determining if the trip is worth the fare. A system that charges based on occupancy and distance traveled would leave the fare unknown to the passengers and still would not divide all costs equally – would riders be charged more for going uphill instead of down? Most people already do not ask all of the right questions and think they fully pay the cost of owning and operating an automobile through registration fees and fuel tax. The greatest challenge for transit is that revenue and cost numbers are straight forward and easy to understand while automobile costs are not.
    Noting your comment on Portland’s three-zone system: I grew up in the Portland area back when they actually had a five-zone system. With little money at the time but a lot of youthful energy, I would regularly get off the bus at the last stop before crossing a zone to save the difference. It would not have cost TriMet any more to take me further – actually less if you figure I stopped the bus. Today you can still travel from Forest Grove to Beaverton (a distance of 16 miles) in one zone. Going from Beaverton to Portland (less than 9 miles), however, would be through three zones.

  14. Jarrett at HumanTransit.org May 26, 2009 at 8:58 pm #

    Roger
    The point of my thought experiment is to imagine a fare system based on “accurate division of the total cost of service” in every increment. Such a system could easily be enriched through more subtle costing of each increment, e.g. the higher energy costs of going uphill, if you wanted to be perfectly micro-equitable, and I agree this would be in the spirit of the exercise.
    I do think that such a system is the only one that’s truly equitable, because it matches the way we would equitably divide a cost among friends who were sharing a similar kind of good — e.g. the cost of a shared ride or the cost of a meal.
    But my point is to suggest that total equity is not a viable objective for real-world fare systems, especially if we want simplicity and understand that to include “ability to figure your fare in advance.”
    (Note though that motorists do not expect to figure their fuel costs in advance, except in very general terms. They know exactly what it costs to fill the tank but may have only a vague sense of how much of their tank was consumed on each trip they made.)
    Re zone systems, yes, they always have a distorting affect around zone boundaries, where a sudden cost differential applies between adjacent stops. I’ve always thought that the easy solution for this is to make the first zone boundary free, so that zone fares only kick in when you’ve gone a significant distance. But zone systems and their geography are a matter for another post.

  15. Rob June 1, 2009 at 1:08 pm #

    Hi Jarrett,
    I realize I’m a little late to the game on the fair fares post, but I just found your blog from a link over at the Seattle Transit Blog.
    Anyhow, what are your thoughts on free ride areas like in Seattle and Portland? I’m doing some research on Seattle’s RFA but I haven’t found any data regarding numbers of trips within the RFA and how many of those are made by passholders, or delays from confusion and increased dwell time from the two-step (wait for alighting passengers, then board) process for outbound Metro trips from downtown.
    Any thoughts (or data) would be greatly appreciated.
    Kind regards,
    – Rob

  16. Jason June 3, 2009 at 4:23 am #

    Hi
    a few scraps of idea that this post has provoked
    2-hour tickets are great (I live in Melbourne) but the Melbourne system introduces perverse incentives as the period starts from the next whole hour. I bet the validating machines fall silent in the 5 minutes before the hour and then rev twice as hard in the five minutes after!
    Canberra’s fare system was recently changed to a one-fare per ride system, removing the mess of zones that used to exist. this was met with broad approval. Zonal or distance based fares make sense to me as a planner and economist, but as a patron, I hate them!
    I agree with the idea that fairness begets complexity, I think the restaurant analogy used above is helpful… in that i immediately thought it it was wrong and had to think through the ways in which sharing the cost of one exhaustible good among equals is different from riding on a train…
    i. imagine a restaurant meal where some attendees are your children, some are your invalid parents, some people have arrived exactly at the time when the restaurant is busy and forced them to buy out the real estate next door. A fair system is one that recognises and reflects the dynamic effects of patronage on costs and the divergent effect of fares on the lives of customers.
    I was fascinated to learn that the last train is always under-filled, no matter when it runs. I’ve always suspected we should run some empty services to help keep the full ones full…
    This fact is fuel to the fire of my favourite PT theory, which is that the system has to be run holistically, as an integrated transport alternative. Anecdotally, people will buy a car to support one rare but high value kind of trip, eg. going surfing on the weekend, or getting across town to their girlfriend’s house at odd hours. If you run loss-making night-time or rural PT you can actually prevent people from makign the one decision that is most likely to inhibit their PT use, buying a car.
    i.e. empty night-time buses are the PT company’s promise to the consumer that PT will always be there for you, so the customer never invests in an option that would eat away at the bread and butter revenue source for the PT company. From an eocnomic perspective, it has to do with the relative own-price elasticity of various kind of trips, which has to do not only with the value they deliver. Getting to work is valuable, but elastic, as i am informed about the trip and can make other investments to avoid the cost. getting to a point i didn’t know far advance i needed to visit may be less valuable but more inelastic…)
    I am really keen to do some research on the claim above, which I currently have only anecdotally. If anyone knows of any research into why young people buy their first cars, i’d love to hear of it!

  17. Jarrett at HumanTransit.org June 3, 2009 at 4:57 am #

    Jason. I think I agree with everything you say, and have said most of it at one time or another. But I really appreciate your idea of extending the restaurant analogy to show how it can capture peak pricing (which is a real element of operating cost) and also how we can use it to think about senior/student discounting.
    The whole issue of deterring car purchases is interesting, because relatively few governments — usually just dense core cities — are willing to even state this as an objective. When I’m talking about transit as possibly reducing car ownership, I usually focus on the possibility of just having one car for a couple. Lots of people out in suburbia just can’t comprehend why anyone would choose not to own a car, but owning one car as a couple seems less alien to them.
    I haven’t looked for research on this, but I agree it’s worth studying. It’s related to the excessive supply of residential parking with units in the dense inner city, which drives up the price of housing for young buyers. So it also ties into the issue of location-efficient mortgages, where you can qualify for a larger loan if you can show you won’t have high transportation expenses (e.g. won’t own a car).

  18. Ari August 31, 2009 at 10:24 am #

    The best fare system I’ve seen (although I haven’t been everywhere) is Melbourne’s. Collection there is a bit of an issue, but they seem to have elegantly solved the problem of zones with, overlapping zones.
    The problem of zones is, of course, that taking a short trip that crosses a zone boundary can be significantly more expensive than a long trip that stays within one zone. (Caltrain seems to be rather egregious in this respect: as is proffered on its wikipedia page, it is possible to pay $2.50 for a 12 mile ride but $4.25 for a 2 mile ride.) In any case, in Melbourne, there are two zones (there used to be three), and most stations near a zone boundary can be accessed with either a Zone 1 or a Zone 2 ticket (see a train map, other maps are similarly simple). This reduces the inequity for short trips which happen to cross zone boundaries.
    The problem with flat fare structures is, of course, that it encourages people to travel further distances (some fixed auto costs, like insurance and registration fees, as well as low use fees like gas taxes, have a similar effect) when transit is least efficient. Thus, efficient transit users who travel shorter distances subsidize others. The problem, of course, is how to define the boundaries, and Melbourne’s solution seems pretty simple, easy and equitable.
    The other thing Melbourne has done well is to create a unified system of ticketing for trains, buses and trams. One ticket is good for a specified period of time and can be used on any service during that time. For any city where all services are run by the same agency (Boston, Philly) free and/or easy transfers for all customers should be a no-brainer. And in cities with multiple agencies (like the half dozen or more in San Francisco) work should be done to unify fare structures. As someone pointed out to me last time I was in San Francisco, they missed an opportunity in the planning of their mostly-unimplemented RFID payment system to unify fares on BART, Muni, Caltrain, AC Transit, Golden Gate Transit, SamTrans (and others) but did not.

  19. John Gear November 4, 2009 at 5:39 pm #

    My local transit system is imploding because the elected board has kept fares way too low for too long and have tried to operate the bus system as a social service more than as a basic part of the city infrastructure like police, sewers, fire, etc.
    Thus we have an impoverished, low frequency system that attracts few outside peak commute hours and, thus, has little public support, and thus has operating levy failures, and so on and so forth. The ultimate vicious cycle.
    I went back to the founding of the system in 1979 and plotted the minimum wage in the area vs. the fare, and found that our minimum wage was $3.00 and fares were $0.30 — for a skeletal system, which we are fast returning to.
    Today, minimum wage is $8.45 an hour but the fares are only recently up to $1.25 (from $1.00).
    I prefer entirely fareless with the cost of the system put on the property tax, as with fire, police, and other vital necessities. Absent that, fares ought to maintain a relationship to the minimum wage (which is indexed to the cost of living here) — I proposed 25% of the hourly minimum wage for the basic, one-way fare, with discounts for multiple ride passes etc.
    The board and budget committee gasped at this, but have done nothing useful about the meltdown in the bus system’s budget, which is only going to get worse as gas spikes up above $100 again soon.

  20. rhywun December 19, 2009 at 7:27 pm #

    I’ve always been interested in the phenomenon of how buying a car in an urban area where one isn’t strictly necessary affects one’s choice of trips. I have never owned one, but if I did, I imagine that I would use it much more often than strictly necessary just to get the most value out of it. Otherwise how could justify the expense? For example, I would do my grocery shopping at the huge superstore 10 miles away rather than the smaller but perfectly adequate store within walking distance as I do now. I would drive to the suburban mall rather than take the train two stops to the local shopping district. Jarrett is right that cities do little or no “advertising” of the ability to live car-free. Even here in NYC, where living without a car is absurdly easy, one still gets the feeling that those of us in the majority who don’t own a car are the strange ones.

  21. Greg Hooper January 7, 2010 at 10:30 am #

    Thanks for an interesting article and discussion. Have there been any similar discussions of fares vs. tax-supported operating costs? I’m looking at the recent 10% increase in fares here in Calgary, which puts our fares slightly above those of NY. At the same time, I suspect our transit system receives more public funding.

  22. Curt Sampson May 26, 2010 at 1:19 am #

    Hm. Anecdotal evidence only, of course, but using a smart card charged with cash here in Tokyo where we have a small-increment fare system for the trains, I don’t typically worry about how much I’m paying. Nor do I know of anybody who does. Perhaps we treat it the way drivers treat a gas tank: you have a general sense of how much any trip might cost (a “couple of dollars,” “a sawbuck”) based on approximate distance, you know how much it costs to “refill” your card when it gets low, and you mostly pay attention to how often you need to do that refill.
    Also, you didn’t mention that these days, with the web available on mobile phones, usually exact trip price is easy to get. Here we just put in start and end stations into one of several services that gives us the route, transfers, times, distance and fare. Perhaps these systems aren’t as common in North America?
    Then again, perhaps Tokyo is Just Different. Here, for example, the last train is always completely packed.

  23. Aleks Bromfield December 7, 2010 at 2:01 am #

    To me, it seems like there are only two reasonable criteria to use in designing a fare system.
    The first is profit maximization. Business (as a rule) don’t set prices based on costs; they set prices based on what the market will bear. Prices are lower when there is more supply, and higher when there is more demand. By this standard, you would expect peak fares to be higher (since demand for travel is much higher), and long-distance fares to be higher (since competing travel modes also cost more).
    The second is incentivization. By charging more during peak hours, transit agencies encourage people to switch from overused peak service to underused off-peak service, which is a win for everyone.
    In many cities, peak service is almost overflowing, while service at night and on weekends is almost empty. The logical conclusion would be that peak service should cost more, and night/weekend service should cost less, or maybe even nothing at all.