Recently, I reported on a staff presentation to the February 10, 2016, Metrolinx Board meeting in which it was quite clear that the provincial agency is moving inexorably toward some form of fare-by-distance not just for its GO Trains, but for all rail modes that operate on their own right-of-way including subways and LRT. Bus Rapid Transit, although a common part of The Big Move as a “solution” in the 905, is not included as “rapid transit” in their proposals.
Metrolinx has now published the background technical paper to this presentation. This is the second installment in the fare integration review. Part 1 came out in September 2015.
For background information, see:
- Metrolinx Fare Integration page
- Metrolinx Fare Integration Backgrounders (September 2015)
- Metrolinx Fare Integration Study: Heading to a Foregone Conclusion? (September 2015)
- Metrolinx Fare Integration Survey (November 2015)
In previous articles, I noted a strong inclination by Metrolinx toward distance or zone-based fares, and this continues into the most recent paper. What is also troubling, however, is the difference between the backgrounders which have some detail about the workings and effects of a new fare structure, and the staff summary reports where this is all treated as a complex issue to be reported on at some future date. “Have patience”, the Board is told by staff, even though answers to some of their questions can be found in the detailed reports.
A chart from the September 2015 report is worth looking at again.
It is self-evident that any tests of fare effects, of the various sub-markets which might see higher or lower fares, and of revenue distribution models, could not possibly be conducted without an actual model with real numbers of riders and real fares attached to each type of trip they would make. However, even by February 2016, the staff presentation studiously avoids specifics and gives no hint of the degree by which fares might rise or fall. A related issue is that the tests used a “revenue neutral scenario”, one in which any reduction in one part if the revenue pie must be counterbalanced by an increase somewhere else.
This is the “we don’t want to pay more subsidy” view of fare integration where the “winners” are those who now face multiple fares, typically to use the TTC plus any other GTHA system, while those who now pay single fares within a system will be the “losers”. This fact has been hidden from public view throughout the exercise. We hear a lot about “fair” fares, but absent specifics, there is no way to decide which elements of “fairness” are included in the new formula.
An element that emerged from the 2015 study and now forms an important part of the 2016 update is the concept of “service class”. This is a simplistic division of all transit services into three groups and three trip lengths. The groups such as “local” and “rapid” transit are defined by mode (bus or rail), not by the actual quality of service they provide. Indeed one might argue that the higher speed of “rapid” transit can be offset by the severe overcrowding and unreliability of the service. For the purpose of the fare model, only the nominal speed and presumed comfort of rail services counts. This is very much a GO-centric view of transit service.
Even GO is having its problems with comfort and crowding. For some time, the goals for service quality included the idea that 80% of peak period riders would get a seat. GO never attained this and the metric sat at 66% the last time it was reported, September 2015. This goal appears to have quietly fallen off the service targets GO seeks to hit for the obvious reason that it is not attainable without massive expansion of service beyond the point where latent demand always backfills any new capacity GO provides.
The overall fare integration study is expected to stretch into late 2016 according to the chart below. It is hard to ignore the absence of “rider effects” in the topics listed here.
However, a more detailed chart (Fig. 1.4) clearly shows that a “Preferred Option” would be recommended by Spring/Summer with the implication that any chance to debate and influence the choice has a limited window of the next few months. The Metrolinx Board is not scheduled to meet again until June 2016. Will they be faced with a fait accompli where staff have made all the decisions and the Board’s role is merely to nod in agreement? When will local Councils and transit agency boards have their chance to be heard? In the rush to get a new fare structure, will Metrolinx management do an end run around meaningful consultation at both the general public and political levels?
The January 2016 Report
The January report starts off with the findings of the previous round, the September 2015 Stage 1 report.
That report concluded that a common fare structure across multiple “service” types should not be pursued further, and moreover than fares based on travel time should also be omitted. In a network ranging all the way from truly local buses pushing their way through traffic up to commuter rail lines stretching far beyond the City of Toronto, it is self-evident that a one-size-fits-all fare model will not work. However, this opens up the discussion of distinct “service” types within urban areas, notably the inclusion of subways and LRT as a “rapid transit” class separate from the surrounding bus network. Ironically, Bus Rapid Transit, although a major part of The Big Move, is lumped into the “local” bus service class despite its higher speed compared to truly “local” operations.
The most important part of any fare study is to understand how the transit market behaves today and how it might evolve in the future. Although this information was not included in the staff report, there is a considerable breakdown of existing travel information in the backgrounder.
To develop an understanding of how transit customers travel in the GTHA and the customer travel impacts of fares, sense making analysis was conducted. This included both quantitative analysis of travel behaviour along with stakeholder engagement through a Technical Advisory Committee (TAC) that included representation of all municipal (transit) service providers (MSPs) in the GTHA.
This approach combined descriptive statistics (included in transportation data sets) with stakeholder insights, issues, and opportunities from an operational lens to ensure the sense making analysis included a wide array of factors. [p. 7]
Quite clearly, the sort of detailed breakdown of new fare structures and effects on riders already exists because this analysis would be impossible without it. However, the details are not in any published report. This begs the question of how any public consultation can answer the question “how will my transit costs change”.
For the purpose of analysis, the transit market was subdivided.
The caveat about “distance bands” is important because it will appear later in the discussion of zones. True fare-by-distance would be free of zones and only use measured travel, but it is the most complex to implement and operate, especially for high turnover local services.
The subdivision of service types is based on the following distinctions.
The criterion that “rapid transit” be 90% or more on a separated right-of-way might explain the absence of BRT from this class because it is common for bus routes to have a combination of mixed traffic and separate right-of-way. However, this shows how arbitrary subdivisions can skew a policy. For example, if a bus route goes from 75% to 90% “separated” thanks to the opening of a new section of BRT roadway, should the fares go up to reflect a different “service class”? Subway and commuter rail services do not face this problem because they are, by definition, totally segregated, but LRT, like BRT, can fall into a medium range of segregation from traffic.
The report notes:
The data used in this analysis was drawn from the 2011 TTS data and therefore predates the opening of dedicated bus rights of way in York Region and Mississauga. As fare structure design advances, the appropriate classification of these bus services for the purposes of applying fares is an area of ongoing examination.
Specific services not clearly positioned in the service typology will need to be considered at a later stage of analysis. These include: express bus services, rural services, and specialized services. [p. 11] [“TTS” is the “Transportation Tomorrow Survey” conducted every five years by the University of Toronto on behalf of government agencies in the GTHA.]
A related policy question will be the potential for a higher fare where “rapid transit” replaces bus routes. This is not a component of current public debates about LRT services, not to mention future subway extensions.
As commuter rail services evolve, trips on the GO network will increasingly depend on feeder services and co-fare arrangements (or some equivalent).
The role of regional services is expected to evolve as RER is implemented. More medium distance trips are intended to be served by RER as frequencies increase.
In addition, the existing mode split for those accessing regional stations is unlikely to be sustainable as ridership grows faster than park and ride capacity can be affordably or feasibly supplied. Attractive feeder local and RT service is critical to RER’s success, which has significant implications for fare structure design. [p. 11]
The market subdivision problem is more complex than the staff summary implies.
Having defined the markets and services, the report turns to the actual distribution of trips.
An important point in this and following charts is the definition of “Downtown Toronto”. The study uses the city’s Planning District 1.
From the chart above, it is quite clear that travel within PD1 (“downtown”) is a comparatively small part of the overall transit market, geographically speaking. Trips to and from downtown, and trips between areas outside of PD1 are a huge majority of travel. In the former case, the subway (“rapid transit”) is either the sole mode or is part of a local+rapid transit trip. A very limited amount of travel occurs on the GO regional system within Toronto.
In the case of trips that are not to or within downtown, local service takes a stronger role, but rapid transit is still an important component. The big difference lies in the proportion of trips that do not use rapid transit for the simple reason that this network does not serve large parts of Toronto.
A further subdivision gives the breakdown in numeric form (click to expand).
This breakdown reveals important facts about the travel market:
- Of the total GTHA travel, slightly less than half, 46.4%, of trips involve “downtown” defined as PD1.
- Within Toronto, almost as many trips are “non-downtown”, 31.9%, as are within or to/from “downtown”, 33.3%.
- Travel from the 905 to locations within Toronto accounts for 20.5% of all trips of which about 40% are to areas outside of “downtown”.
- Of the 1,209,500 trips within Toronto, only 285,800 (23.6%) are “rapid transit” only and the remainder are either entirely on “local” (i.e. surface) transit (449,100) or use both modes (474,600). This shows the vital role of surface transit in the TTC network, one which is repeatedly overlooked in transit planning.
- The role of transfers between modes is quite evident in the numbers for compound trips, but transfers within a mode (e.g. bus to bus) do not show up here.
Slightly less than 10% of all trips involve a double fare with the TTC either from a local 905 system or from GO Transit. This proportion could grow were it not for the absence of a co-fare or some other integration mechanism.
Another important issue is the distance travelled in various types of journeys.
It is self-evident that trips within downtown cannot be long because the geographic area itself is small. Trips between downtown and the rest of Toronto tend to be longer, especially if GO is involved. (Note that these are trip length averages, not a count of trips.) That said, a 10km radius from downtown (Queen & Yonge) takes in roughly the old City of Toronto, also known as TTC Zone 1. It is also self-evident that there must be a lot of local+rapid transit trips to downtown that are longer than 10km or the subway would be nowhere near as busy in the former suburbs as it actually is.
This distribution of trip lengths is an important factor. If we are to believe these charts, on average, trips inside Toronto, but outside of downtown tend to lie in the 5-10km range, especially for users of surface routes. This is not the typical experience of a rider originating in the outer 416 and making a long journey to just about anywhere. Any fare policy must take this into account, not simply the relatively large number of short trips taken in the older, denser part of the city on surface routes.
Also intriguing is the much long average for trips by cross-border riders. Comparing the lengths of the averages, one might argue that only charging them a double fare is still a bargain, except for those who have atypically short trips, compared with the riders who travel within Toronto.
This shows up in when the average cost/kilometre is calculated.
The cost per kilometre is much lower for cross-boundary trips than it is for trips within the City of Toronto, and especially for short trips downtown. To be fair, the cost of the trips is all set at the cash fare, but many short-trip riders are going to use passes and their actual cost per trip will be much lower. The study has not taken that factor into account.
Next, the study looks at trip lengths by service class.
The upper part of the chart is interesting because it shows the distribution of trip lengths, not the averages. Trips that use only surface routes (“local”) drop off quickly after a peak at under 5km. This shows how a very large number of “quick hops” can mask the presence of longer trips in the average. “Rapid transit” (which means subway in the 2011 stats) falls off quickly after about 14km (why there is a peak at this distance I do not know, although this could represent a large demand with a common trip length such as Finch Station to central downtown).
Another intriguing number is the indication that the vast majority of trips have no competition in their market. In other words, riders do not have the option of making the trip in another way via a competing, possibly cheaper, service. This is a direct result of Toronto’s network design and transfer policy.
From earlier charts we know that there are many riders who use more than one type of service in a typical trip, and so the subdivision of trip length by individual services (rather than the combinations) could make the “market” appear to have distinct chunks when in fact these service classes are symbiotic. For example, the hordes arriving at Finch Station do not simply materialize there, but come on local TTC and York Region services, GO buses, and by private auto.
Indeed, autos are a “missing mode” in this analysis, and we do not know how long the trip link providing access to transit is for, say, auto+regional or auto+rapid transit travel. This is an important component when considering how or whether local transit can take over the “last mile” function.
The report reviews “local” trips noting how many of them are 7km or less. If a zone (or more accurately the distance over which a single zone’s fare would apply) were slightly larger than 7km, then it would encompass the lion’s share of “local” travel just like a flat fare system. The corollary here is that, at least for “local” trips, it may not be worth the extra complexity of zones because most trips would only pay for one zone’s travel anyhow. The longer trips would be carried at the same price because it was simpler to do so.
However, this ignores the role of the rapid transit system as an integral and growing part of much travel. Having a flat fare for surface modes discriminates against riders for whom an all-surface trip is either impractical (time considerations) or impossible (network design). The problem will only get worse as “rapid transit” spreads beyond the Toronto subway system. This shows the inherent folly of attempting to treat surface and rapid transit operations separately in an integrated network.
This leaves us with a flawed model going into the discussion of actual fare structures.
Fare Structure Design
This section of the report brings us to the same territory covered in the staff presentation, namely the principles for a tariff design and alternative schemes for reconciliation of fares by service class. As already described here and in the previous article, some problems directly arise from the attempt to treat “rapid transit” and “local” classes as distinct entities in a network where they are designed to operated as an integrated whole. A related problem is the concept that each tier, each type of service is inherently “better” and more valuable than the one below it.
However, service quality is not just a question of vehicle speed, but of overall trip comfort and other sources of delay such as the inability to board crowded vehicles at peak times. Such delays can contribute a substantial portion of the overall trip time slowing the effective travel speed, while overcrowding can reduce the presumed “premium” of faster rapid transit or regional service to cattle-car conditions. Some of these are locational and time-of-day effects and it is impossible to account for them all in a fare structure. Riders will make their own selections by avoiding the worst periods to the degree that they can.
As a quick review, here are the five principles behind the would-be fare structures:
- Fare continuity ensures customers will pay a comparable fare for services that provide travel over the same distance market so that customers do not pay more because a transit service type is not available. [p. 32]
- Fare structures should be developed based on how far customers travel and the parameters and should be aligned with service type constraints. [p. 34]
- For the fare strategy to enable use of the complete GTHA transit network, it should not penalize trips that require the use of multiple service types. [p. 36]
- When travel times vary between two services that provide transport over similar journeys, fares should be lower for slower service types than for faster service types in order to ensure overall generalized costs are comparable. [p. 37]
- Fares should use small increments and gradual changes to ensure pricing is not a disincentive for passengers to access transit services. [p. 38]
Some of these principles contradict each other, especially the fourth one which argues, in brief, that a bus trip should be cheaper than a subway trip which should be cheaper than a GO trip. This directly opposes the premise in the first two points, and probably in part the third.
The fifth principle implies granularity in any zone or distance system so that large jumps as one crosses from zone to zone do not deter travel.
All of this is consolidated as per the table below.
The principle of “continuity” – that fares for different service types would be the same where they overlap in trip lengths served – is a rather large piece of duct tape covering some basic inconsistencies arising from arbitrary segmentation of the market. Note that this overlap is not necessarily geographic, but simply by trip length. This begs the question of what would happen to “regional” fares once “rapid transit” becomes more prevalent in the 905. Following the model, “mid-length” journeys should be comparably priced, but this might force a downward shift in GO fares if a region (say, Mississauga) gets an LRT line.
One might construct a logically consistent fare model with a “continuity” principle based on today’s network, but as new services in various classes are added, what gyrations must the tariff card go through?
The design tree above has several “features” its creators might not appreciate:
- If local fares use zones with a 7km radius, then we will have effectively reinstated the old “Zone 1” used by the TTC before Toronto fares were integrated four decades ago. Additional zones would cover the suburban areas (probably three would be needed with some spillover into the 905. It is unclear how this structure would eliminate the cross-border complaints. Indeed, it would recreate a border that is only a distant memory for most riders.
- It is unclear just how the “short trip” equivalence between rapid transit and local fares would work, and specifically how they would avoid large increments at the boundary between a “short” and a “medium” length rapid transit trip.
- Regional fares are already “by distance” nominally, but in fact are considerably cheaper for the longest trips on the network. Reconciling GO fares to actual distances would involve some upheaval in the current tariff.
Now we come to the actual fare models, and these were reviewed in my article a few days ago. I will not reiterate the descriptions, and leave this to the reader for review [pp. 40-51]
The basic problem with all of these proposals is that no dollar values are attached, nor are there worked examples of how various origin-destination pairs would price out in each model. This comes in the “Next Steps” even though some type of definitive modelling must have been run to produce the charts and conclusions in the Stage 1 report last September.
What is troubling is that there does not appear to be a feedback mechanism whereby someone might say “hey, this model really doesn’t work, and we need a new one”. The generic models are locked in before we actually understand how they would behave.
Each concept identified in this study will undergo further development. This concept refinement process will include:
- Calibrating each concept with priced fares and co-fares to model the impact on ridership and revenue for various scenarios, creating one or more options from each concept prepared for full business case evaluation
- Scoping/prototyping potential methods for delivering the options or implementing the options in the GTHA, including cost estimates.
This work will consider further data analysis in the GTHA as well as a review from other jurisdictions that have successfully applied fare by distance and zone structures, or have differentiated fares by service type. [p. 52]
Among the factors to be considered is a “Barrier Analysis”:
Assessment of how the structure removes fare barriers (cost, complexity, captivity) including assessment of changes to fares for each market/sub market, representative origins and destinations, and trip lengths. [p. 53]
There appears to be no concept that the new structure might create barriers, and that this is equally deserving of review.
Another factor is “Equity Analysis”:
Review and clarification of potential impact of fare structure options on economically disadvantaged communities
- Change in fare for economically disadvantaged communities
- Change in fare to access low income employment centres [p. 53]
An important issue in “equity” analysis is that one cannot simply throw a few discounts in the direction of the most severely disadvantaged, but Metrolinx and their political masters must acknowledge that the working poor are a large group in their own right for whom a significant fare increase in the name of “regional fairness” will not sit well. This is particularly important for long trips and for “trip chaining” which could benefit from time-limited transfer privileges (a.k.a. “the two hour transfer”).
Metrolinx, together with the City of Toronto and TTC, is about to launch a new rounds of public consultation session. Without the ability to discuss specifics of implementation such as zone sizes and price points, “consultation” will be difficult on a meaningful basis.
“Fare integration” has for years been code for “your fares will go down”, but that is not necessarily what the new options will bring. SmartTrack has been sold on the basis that riders will have “TTC” fares on trains running over the GO corridors, but “TTC” fares tomorrow may work quite differently from these fares today.
Toronto politicians have been silent on this issue most likely because the implications of Metrolinx’ work are only now apparent. The many consultation sessions should smoke out Councillors’ positions on what they consider to be reasonable, fair fares for their constituents.