A Few Questions About Presto

This article has been broken off from the February 2012 Metrolinx Meeting report given its size and the fact that much information here does not actually come from that meeting.  Comments related to Presto have been shifted to this thread.

In the original article, I reported a series of questions posed to Metrolinx about its Presto card and fare structure.  I have now received responses, and these have been interpolated into the text.

On January 9, 2012, following the last Metrolinx Board meeting, I posed a number of questions to clarify some of the statements and discussions.  These were answered during the week of February 13.  In the text below, the Metrolinx response is given “as is” as a quoted block.

Question:  How will the fare increase be administered for passholders? Will they pay for February at the old rate including the current, higher discount?

Monthly pass holders will still have the opportunity to buy their February monthly passes at the old discount rate. Starting February 18, however, monthly pass holders will continue to receive a discount, but at a lower rate –15% off the cost of 40 single-ride adult tickets and 30% off the cost of 40 single-ride student tickets.

Question:  If someone prepaid for a pass, they would get the entire month at the “old” rate. However, the way Presto works is that it deducts full fares from your account balance until you get to the maximum for the month.  Will a Presto rider will pay fares at the old rate up to the cutover date, then at the new rate until they hit the maximum fares for the month where the free loyalty bonus trips cut in? Will the cost for February will be a blended value of the old and new rates.

Yes. PRESTO customers will pay their fares at the old rate until February 17. Starting February 18, PRESTO customers will pay their fares at the new rate for each ride.

Question:  How long will old multi-ride tickets will continue to be accepted?  This was answered in a report on the February agenda — July 31, 2012.

Question:  The entire discussion of the rationale for or against a tiered fare increase and, if so, how it should be structured, was, at the very least, confusing and at times almost evasive. Some of the statements made during the Board meeting and information in the presentation/report simply did not line up.  For example, it was clear that the question of true fare by distance uncovered a lack of full understanding of your cost base and how it would vary depending on service level and/or ridership.

In reply to a question, Bruce McCuaig stated that the majority of costs are fixed in nature and not related to the distance travelled. From the presentation, we know that diesel fuel is a high area of cost pressure and from the financial statements, we know that it’s about 1/3 of total costs.

Some of our costs, such as the cost of diesel fuel, do vary based on the distance travelled. As a result, we have implemented a three-tiered fare increase starting February 18.

Comment:  The increases are proportionally lower for long trips than for short ones, although not quite as blatantly as in some previous increases.  This continues a pattern whereby “equity” in fares based on distance travelled declines, and GO is favouring the long-haul commuter over those closer to the 416.  This has implications for a future world in which GO becomes a regional rapid transit system, but may not be appropriately priced as an alternative to local systems.

Question:  The fact that a train is there at all is as much a policy decision as it is one driven by ridership. The Kitchener-Waterloo trains are not full when they leave Guelph, and the KW passengers could be served by a shorter train.

GO Transit determines the train length for train trips on a corridor by the maximum load along that corridor, rather than ridership numbers at a particular station.

Question:  The problem then becomes how much of the cost of running 12-car trains all the way to KW should really be borne by riders coming from that far out?

This fare increase is about ensuring a fair and balanced payment system. Introducing three different increments of increase means that passengers who travel shorter distances will pay less of an increase than passengers who travel longer distances. These rates support the fare-by-distance system that GO’s fares are based on.

Comment:  But fares are not entirely based on distance travelled, and as GO itself acknowledges, many of its costs are not a function of trip length.  “Distance” is one measure that can be used to apportion fares, but it is not the only one, and no method will produce a “fair” result especially when service exists to support a policy (minimum frequency of trains, availability of service as an incentive not to drive, etc.)

If we regard the existence of a train in the schedule as a fixed cost, then there is little difference in carrying a passenger from KW or from Weston because you have to operate the capacity for the Weston passenger all the way to KW, at least on the trains that travel that far.

I will be the first to admit that this is a rather nasty line of questioning because it delves into the whole underpinning of GO’s fare and cost structure.  Fare-by-distance is a superficially “fair” way of charging for service, but the cost of providing that service on either a marginal or fully-allocated basis does not necessarily translate to a uniform value for each unit of “service” consumed.  A simple example is that the cost of a train is shared by fewer and fewer riders the further out one goes, and therefore the cost per rider goes up quite steeply as the occupancy of the train declines.

Question:  If for policy/political reasons, GO starts running all-day service to, say, Georgetown, should the marginal losses of this be borne by riders on that line, by the GO riders in general through fares, or by increased subsidy? In particular, should the government/agency get kudos for service improvements that have lower cost recovery only to tell riders that they must pay for these through higher fares?

GO Transit sets fares based on the distance travelled by a customer and to cover the cost of operations across the system, rather than for a particular service. These costs include bus and rail operations, corridor operations, and peak/off-peak service.

Comment:  Running trains all day both ways will inevitably result in lower utilization of capacity and an increased cost per trip.  There is a cost to the policy of running better service (as we have seen on the TTC).  Will riders face increased fares to maintain GO’s farebox recovery ratio when less-productive services are introduced?

Another issue in cost allocation is that since GO contracts out a lot of work, the cost drivers at the contract level may be different from those one might see if the functions were in house.  These factors may not be linked to variables directly affected by passenger volumes or distances travelled, but rather to equipment usage.  These values are almost certainly confidential and dis-aggregating them in a public report could prove troublesome.

Question:  In the meeting, it was claimed that the cost per passenger is $10, with a $2 (20%) subsidy per passenger. However, the fare presentation pegs the average fare at roughly $6.50 (two different numbers are cited, one in the report, another in the presentation foils). Where does the other $1.50 come from? That’s a lot of “other revenue”.

GO Transit has two major sources of operating revenue – customer fares, which made up about 82 per cent of our operating costs for fiscal year 2010-2011, and a Provincial subsidy. Our cost-recovery ratio – one of the best in the world at 82 percent – helps us to minimize subsidy requirements from the Provincial government, something that benefits the Province and taxpayers as well.

It should be noted that the sundry revenues also helps us to keep down the average passenger fare, which is about $6.50. These sundry revenues include advertising, track fees on rights-of-way owned by Metrolinx, rental fees, and proceeds from the divestment of assets, such as the sale of old locomotives.

Comment: This answer perpetuates the same inconsistencies that we heard at the meeting.  First off, if fares are 82% of operating costs, and the cost per passenger is $10, then the average fare must be $8.20.  If sundry revenues contribute, then they would bridge the difference between the average fare and the average revenue per passenger (which are not the same thing).

For example, on the TTC, sundry revenues account for a few percent of total revenue, and this leads to ongoing confusion between their 70% cost recovery rate for the system and the fact that riders only pay about two-thirds of the cost of running the system.  Moreover, sale of used equipment is cited as operating revenue, but it is a non-recurring item that should be booked against capital accounts, not operations.

In the covering report for the Metrolinx financial statements for 2009-2010, the non-fare revenues are described as 3.8% of total revenue, and they have a value of about $11.5-million.  This has been declining since a high of about $13.5m in 2007-2008.  It is worth noting that the chart correctly does not include gains or losses related to the sale of capital assets.  It is still unclear how we get from a reported average fare of $6.50 to a cost per passenger of $10 and a farebox recovery rate of 82%.

Metrolinx reports its finances on a consolidated basis and it is not possible to see the contribution of each component (GO, etc) to the overall picture, nor the cross-subsidization, if any, between divisions.  When GO was a separate agency, its financial statements did not break out sundry revenue as a separate item, and so there is no basis for historical comparison.

Any model of cost allocation and fares should not treat occasional revenues as part of the ongoing base.  This is especially dangerous if the amounts involved are more than a few percent of total revenues because the disappearance of one-time revenue could have a significant effect on subsidy requirements, fares or the ability to operate service.

What is troubling is that an obvious inconsistency in revenue and cost reporting was not caught by staff or by the Board, or why it has taken a month (and counting) to get an answer to what should be a straightforward question.

The basic problem here, as we saw quite clearly during the January Board meeting, is that Metrolinx staff either do not understand how their costs and revenues work, and how these interact with the fare structure, or they are deliberately obtuse in answering questions.  There is no perfect way to allocate costs and to set fares relative to service consumption, but Metrolinx clouds the issue with what appear to be policy-based fares (artificially low costs for long trips).

Question:  In the original comment thread, the matter arose of how Presto discounts would apply for riders who do not travel between one set of stations on the system.  On the GO website, the operation of discounts is described thus:

With the built-in loyalty program, adults will receive fare discounts similar to the discounts currently offered with GO adult 10-ride tickets and monthly passes.

If you take the same GO trip each time you travel within a calendar month, your GO fare* will be:

  • Rides 1 – 35, 7.5%** off the single adult GO fare
  • Rides 36 – 40, 87.5%** off the single adult GO fare
  • Rides 41+, 100% off the single adult GO fare

If you do not take the exact same trip each time you travel on GO within a calendar month, your first 35 rides on GO will be 7.5% off the single adult GO fare. For rides 36 and onwards, your discount will be based on the value of the rides you’ve taken that month.

What is not clear is the way in which the discount is calculated for someone who has a primary trip pattern, but who also make occasional trips elsewhere. This would be roughly the equivalent of a monthly pass holder (primary trip) plus occasional riding on the rest of the system. How does the calculation work?

1. For Rides 1 to 35 in a month, a customer will always receive a 7.5% discount for each ride regardless of the origin and destination travelled.

2. For Rides 36 to 40 in a month, the discount will depend on the cumulative amount that the customer spent in the first 35 rides. If the customer spent more (as in travelled occasionally farther than his/her predominant origin and destination), he/she will receive a higher discount that will vary marginally from the posted 87.5% for each ride, depending on travel usage. If the customer spent less, he/she will receive a lower discount that will vary marginally from the posted 87.5% for each ride, depending on travel usage.

3. For Rides 41+ in a month, the same logic as #2 will apply but with a deeper discount, which is based on the cumulative amount the customer spent in Rides 1 to 40.

Comment: This does not clearly answer the question, but appears to say that the discount is calculated against the primary trip pair and then applied to all riding.  A worked example would have been helpful.

As GO moves to serve more than just the weekday in-out commuter pattern, those extra trips will be important, and people need to know how they will be charged for them. No doubt GO is wrestling with this as part of figuring out how to get rid of monthly passes as we now know them.  Once the need to sell a pass for a specific trip (say from Brampton to Union) disappears, and the network becomes an all-day system to enable trips between many points, the concept of counting trips between specific locations and linking discounts to this falls apart.

On the TTC it will be even more challenging because of the question of what constitutes one “trip” for the purpose of charging a fare and counting up to the threshold where the price falls off, eventually to zero.  How will a heavy TTC user be charged for regular or occasional trips on GO or on one of the other 905 local systems?

Once GO figures out a tariff for cross-system fares and frequent usage between a variety of points on its network, it must then turn to how the revenue will be allocated to the participating systems.

About Steve

Steve thanks you for reading this article, even if you don't agree with it.
This entry was posted in Fares & Fare Collection, Transit. Bookmark the permalink.
RSS for Comments

13 Responses to A Few Questions About Presto

  1. Phil Piltch says:

    Looks like I’m finally going to have to get a Presto card if I want multi-trip tickets. I must say I do like the convenience of taking GO downtown though given my normal preference to bike most places I’m not the typical “regular” user GO is thinking of with Presto. I currently get 10-ride tickets for travel between Scarborough and Union (about $4.03 per trip) . So Presto would offer me the same cost?

    I have also frequently used GO as a means of avoiding riding endless kilometres of suburbs to get to quieter country roads, such a taking it to Oakville and riding from there (only about 10 k north on Sixth Line to actual undeveloped countryside). I’m assuming the single-ride paper tickets will still be issued as well as the day pass for unlimited travel between two stations.

    Phil

    Steve: As the report says, they are only getting rid of the 2-ride and 10-ride tickets at this time. It isn’t clear how Presto will replace the 10-trip ticket because of the way it charges to build up to a monthly total. According to Presto fare info on GO’s website, the first 35 trips one takes in a month get a 7.5% discount off of the single fare. This is roughly equal to the discount now offered on 10-trip tickets (there is no discount on 2-trip tickets). The problem arises for riders who have more than one O-D pair in their travel because the loyalty program only kicks in if all of your trips are between the same two points.

    Adults:
    With the built-in loyalty program, adults will receive fare discounts similar to the discounts currently offered with GO adult 10-ride tickets and monthly passes.

    If you take the same GO trip each time you travel within a calendar month, your GO fare* will be:

    • Rides 1 – 35, 7.5%** off the single adult GO fare
    • Rides 36 – 40, 87.5%** off the single adult GO fare
    • Rides 41+, 100% off the single adult GO fare

    If you do not take the exact same trip each time you travel on GO within a calendar month, your first 35 rides on GO will be 7.5% off the single adult GO fare. For rides 36 and onwards, your discount will be based on the value of the rides you’ve taken that month.

    It is unclear just what the discount will be for riders who have the temerity to travel between multiple points on the system. I will ask about this.

  2. Tom West says:

    For passholders, I can answer your question: the fare changes apply to *tickets* sold after Feb 18th, not for travel after that date. Feb monthly passes were at the old prices. (Similarly, 10-rides bought before Feb 18th could still be used after at no extra cost.)

    Steve: That’s the logical way to do things, but there is still the question of the charge to Presto users. It’s amazing that such straightforward info wasn’t included in the planned fare change’s announcement. It all had the feeling of a very hurried affair considering that a special Board Meeting was called to approve it.

    Before the recent changes, GO’s fares were 9.5c/km + $3.85 (rounded to the nearest 5c), with a minimum of $4.20. (Where distance is between the centroids of the zones, NOT the stations). All the fare increases in recent years have changed the latter two elements, leaving the distance-related element unchanged. If GO had wanted to increase fares for longer-distance travellers by a greater amount, they should have just increased the per km charge.

    The fairest fare system would be if the cost of a train/bus travelling between two stations/stops was split between those riding on it at the time…. (you could actually do this if every used smartcards)…. but people tend not to like that idea. If you don’t have such a system, then some riders will always subsidise others.

    Steve: What was noticeable about the change was that it had the effect of minimizing the increase for riders on their new long-haul trips in from KW immediately after the service began operation. The problem with fiddling the fare system is that once you built in an inequity, it is very hard to re-adjust later.

  3. Karl Junkin says:

    The problem with fiddling the fare system is that once you built in an inequity, it is very hard to re-adjust later.

    True, but it is also hard to readjust when your subsidy is suddenly withdrawn by a government looking to cut spending, as we’ve seen all too recently – we can say this is a bad thing until we’re blue in the face, but it still happens and probably will continue to happen going forward, at least for a few years yet until the investment strategy actually starts to kick in (if it ever actually gets to the floor of the legislature).

  4. Robert Wightman says:

    The problem with determining a fair fare is how to assign those costs which are determined by distance and those that are fixed regardless of distance travelled. This is also a function of service philosophy. Do you run a very frequent service to the end of the line that makes all stops and has enough capacity at the inner end to carry all the short haul riders or do you run zone express service which stops at the first 3 stops then goes express to Union, followed by another zone express which stops at the 3rd, 4th and 5th stops then runs express to Union and so on.

    The first type of service provides a very frequent service to everyone at the expense of increased travel time and the need for more cars to carry the same number of passengers. The second type of service provides a faster service for most passenger to Union but at a reduced frequency. It also makes someone who is not going to Union transfer at least once. Also required, at least closer to Union, is an extra track for the express to pass a train running local. If you are going to run all day 2 way service then 3 tracks are needed for part of the line.

    Staten Island Rapid Transit does this with only 2 tracks but they use transit, not railway, operating rules. Since the ferries arrive on a 15 or 20 minute headway, they only dispatch trains after a boat arrives. The first train out serves only the farthest group of station then a minute later another train serves the second farthest group etc. All the trains but the one from the end come back express to wait for the next boat.

    Costs that are fixed include the home station and (95% of the time) Union. These costs should be the same for everyone but It might be nice to provide a discount for those who do not drive. Capital and operating costs for equipment, cleaning, lay over and maintenance facilities, human resources, signalling and policing are mainly fixed and not mileage dependent.

    Mileage related costs include fuel, wear and tear, wheelage cost to CN and CP for trains that run over their tracks, and crew costs. Do you charge the person who gets on at Bramalea the cost for hauling an empty coach all the way from Kitchener so they could get a seat or do you charge the person in Kitchener the cost because the coach has to go that far so they could have a service. It is not an easy choice to make and arguments can be put for both sides. Metrolinx has to explain how they figure out the 2 costs and the total cost. Transparency is the new in word.

    I believe that a zone express service on the busier lines makes more sense but it does reduce the effective frequency while increasing the speed.

  5. Phil Piltch asked, “I currently get 10-ride tickets for travel between Scarborough and Union (about $4.03 per trip) . So Presto would offer me the same cost?”

    As Steve mentioned, the first 35 trips per month offer a 7.5% discount, which makes the cost per trip roughly equal to the cost per trip of using the 10-ride ticket.

    The really nice thing about Presto is that you don’t have to fork out $40.30 to pay for ten rides up front. The minimum you can add online is $20 at a time, so you can spread out your cash flow, especially if you only use it now and again. Just remember that it can take up to 48 hours for an online load to reach your card (and I have seen 5 days occasionally!). Adding money in person at a GO Station is immediate, and you may be able to add even less than $20.

  6. Ka-Ming Lin says:

    re: having the temerity to travel between multiple points

    What beautiful phrasing.

    On rare occassions, I drive/car pool to work, so sometimes a monthly pass makes sense, and sometimes a month of 10-rides do. Presto helps a lot in this case since I don’t to figure out in advance if my month is going to be over 33 rides.

    My normal ride is between Kennedy and Clarkson, but on some rare occassions I go to Agincourt instead. Under the old system this was no problem since they are both the same zone and the pricing was the same, so I just got tickets/passes to Agincourt even though I usually cancelled in/boarded at Kennedy.

    Also, on some rare occassions (resulting from the very limited service that Kennedy Station gets) I start/finish my trip at Union, which was also no problem with passes. The one time I recall doing that on a cancelled 10-ride and losing out on the dollar-ish difference in fare (plus the cost of a TTC token) was when they announced via PA speaker that my train was delayed by over an hour. The announcement was made a few minutes after the train was supposed to have arrived. There is a display unit there, presumably to provide updates on train delays, but it’s been sitting unconnected and unused for about a year now.

    Speaking of one year, IIRC the Presto rollout hit Kennedy early-mid 2011, so we’re getting about one year to switch over.

    Finally, if fares are at least partly based on the fixed cost of providing the service and the marginal cost of a single added rider is zero, then are those of us who use stations regularly bypassed by express trains and only have access to a (small) portion of the services going to start getting discounts?

  7. Jonathan says:

    The situation at Kennedy station is a particularly good illustration of how the lack of fare integration and GO/TTC co-operation leads us to spend billions on increasing capacity that could otherwise be provided more cheaply. I’ll grant that many people don’t want to go to Union, but the fact that a GO fare from Kennedy is going to be $4.65 (or $4.30 with discount) compared to a marginal cost of $0.00 for a TTC Metropass holder is a huge distortion affecting which option most people choose. Every TTC rider has to transfer here anyway, and most will transfer again to the YUS subway so for many it would be a no brainer to take GO rather than the subway milk run. If GO had runs that only went from Kennedy to Bloor station (and vice versa) on which TTC transfers were accepted, it would take a lot of pressure off of both subway lines and cost far less than the capital cost of increasing subway capacity.

  8. Robert Wightman says:

    Jonathan says:
    February 17, 2012 at 10:47 am

    “The situation at Kennedy station is a particularly good illustration of how the lack of fare integration and GO/TTC co-operation leads us to spend billions on increasing capacity that could otherwise be provided more cheaply. I’ll grant that many people don’t want to go to Union, but the fact that a GO fare from Kennedy is going to be $4.65 (or $4.30 with discount) compared to a marginal cost of $0.00 for a TTC Metropass holder is a huge distortion affecting which option most people choose. Every TTC rider has to transfer here anyway, and most will transfer again to the YUS subway so for many it would be a no brainer to take GO rather than the subway milk run. If GO had runs that only went from Kennedy to Bloor station (and vice versa) on which TTC transfers were accepted, it would take a lot of pressure off of both subway lines and cost far less than the capital cost of increasing subway capacity.”

    Unfortunately GO was never intended to provide a service to compete with inner city TTC service. The other problem is GO does not have the line capacity to handle extra trains between Scarborough and Union on the Lakeshore East line nor does Union have the platform capacity to handle extra trains or passengers. The Weston sub is supposed to get a lot of extra tracks but the ARL is going to eat up a lot of this extra capacity.

    It would require 9 trains to run a 10 minute shuttle service which is the lowest 2 way headway you can operate on a 2 track line using railway rules. This would require double tracking the Stouffville line from Scarborough to Kennedy and the addition of at least 1 if not 2 tracks from Scarborough to Union and the use of a least 2 full platform at Union. This would provide a capacity of 12 000 passenger per hour from Kennedy to Union and from Bloor to Unions but would not help divert anyone who got on the Danforth west of Kennedy. The walk from Dundas West is not a short or pleasant walk either.

    This is a lot of capacity to give up at Union for passenger who are only riding for 15 to 20 minutes. There will be too much demand for people coming into Toronto from outside of the 416 for this platform capacity to give it to short haul riders. The other question is how many passengers at Kennedy want to go that far south in the downtown who would benefit from this service? If you were getting off north of Dundas you would be better off to take the subway.

    It would be much better to spend the money that this would cost building the DRL which would serve a lot more riders and provide a better distribution within the downtown.

  9. The issue that I have with Presto is that I have to accept some blind faith that the appropriate amount is taken off. With the 10 trip ticket which I normally get, I know immediately if the trip is taken off the ticket because I can see the line marking it off the ticket. Also, how will I know how much is left on the card? With a ticket I know how many trips I have left simply by looking at the ticket – I can’t do that with the Presto card.

  10. TorontoStreetcars wrote,

    “The issue that I have with Presto is that I have to accept some blind faith that the appropriate amount is taken off.”

    Only if you are visually blind.

    When one taps a Presto card at the terminal, the display shows the amount taken off, the time remaining (for a time-based transfer system), and the balance remaining on the card. When on a time-based system, if tapping within the remaining time from a previous tap, the display clearly shows that $0.00 is deducted, along with the remaining time and balance.

    This information is displayed only for a few seconds so that the terminal will be ready for the next passenger.

  11. Andrew says:

    If we regard the existence of a train in the schedule as a fixed cost, then there is little difference in carrying a passenger from KW or from Weston because you have to operate the capacity for the Weston passenger all the way to KW, at least on the trains that travel that far.

    If the trains become frequent enough that significant numbers of people start using them to commute between Guelph and Kitchener, this problem will disappear. There is never going to be more than a tiny market for commuting from Toronto to Kitchener because a 2 hour each way commute is ridiculous. If the same seat is used by someone commuting from Toronto to Brampton, Brampton to Guelph and Guelph to Kitchener then you are using seats more efficiently.

    Steve: Exactly. As GO becomes more like a “subway”, it is impossible to assign a cost to a specific passenger’s trip because there are too many variables, and the capacity operated is a function both of policy and scheduling constraints. At that point, you know the system costs “X” to operate, and the question is simply of carving that amount up in some broadly agreeable way between passengers. Trying to justify specifics of one fare over another gets into deep water quickly because any formula one proposes can be countered by an alternative philosophy of cost allocation.

  12. At some point GO and Metrolinx and TTC are going to have to sit down and have those discussions about service, fares, and new routes.

    Personally, I’d love to see all-day GO service running in the inner GTA (say, from Clarkson or Port Credit, Erindale or Cooksville, Brampton or Bramalea, etc… call it “Linx” or “Lynx” or “GO Toronto” or whatever) while GO still runs the peak hour commuter services.

    Another question … is it possible to expand on the daytime bus services, improving service frequencies around the inner GTA, before we have to look at expanded train services. The Mississauga Transitway (to be used by GO & MiWay) will be opening in a few years, and it would be nice it there was a frequent service along Eglinton to reach the Eglinton Cross-town line.

    Another thing worth looking at would be smarter HOV lanes on the QEW, 401, 403, 407, 409, 410, 427, Gardiner and DVP … which would help buses move faster and ensure more reliable services.

    Even putting flexible speed limits on our expressways, which would do a lot to reduce stop & go traffic, would help improve bus services.

    Cheers, Moaz

  13. Sean says:

    Andrew, that 2 hour commute will become less the 1hr 30-40minutes once the track work is done by VIA Rail this year and next.

Comments are closed.