Updated June 19, 2017 at 3:30 pm: The TTC has clarified that the hourly costs shown for various routes are net costs, not gross costs, and this addresses my concern that some of these values were understated. The text of this article has been updated where appropriate.

Updated June 20, 2017 at 10:30 am: A section has been added on gross operating costs (the TTC study includes only net costs) to illustrate how these vary from route to route.

The TTC Board recently approved an Express Bus Network Study that proposes several new and enhanced express routes in Toronto. The premise of the study – improving the bus network’s attractiveness and convenience to riders – is a good one addressing the basic function of any transit system. However, thanks to the TTC’s severe constraints on capital and operating funding, the actual implementation of these proposals drags out for the better part of a decade. Bus planning in Toronto is converging with new subways for a lengthy gestation period.

There are three types of “express” route in Toronto:

• The (mainly) peak period services usually signified by an “E” suffix on the route number. Typically, buses will run express over part of a route stopping only at major transfer points or destinations, and will continue as local service on the outer part of the route. These services are useful for riders who would otherwise face a long stop-and-go trip on a local bus for their entire journey. By carrying the long-distance riders express for part of the trip, the cost of route operations is reduced from that of an all-local service and provides a more attractive service overall.
• The Rocket services (route numbers in the 18x-19x series) operate for most of the day on weekdays and weekends, and provide a more limited stop service, end-to-end, than the “E” branches. Some are designed around major endpoints such as the 192 Airport Rocket from Kipling Station to Pearson Airport, while others more closely resemble the stopping pattern of “E” services. Unlike the “E” branches, the Rockets do not necessarily duplicate the route of a local service.
• The five Downtown Premium Express services (route numbers 14x) charge an extra fare for the privilege of avoiding the crowded Yonge subway and the 501 Queen car.

The TTC proposes an interim classification of the first two of these as Tier 1 (Rockets) and Tier 2 (“E” branches) with the intent of coming up with some sort of branding that could be used to market them. Some cities have special bus services with their own names such as Hamilton’s B Line Express and Vancouver’s 99 B-Line. Given that there already is an established name for Tier 1 with a strong Toronto reference, it is hard to understand why a new brand is required. As for the Tier 2 services, riders are well acquainted with the “E” convention (broken in rare cases such as 60F Steeles West).

The first recommendation of this study is that a marketing effort is required to brand these services. That says a lot about where the TTC’s focus has been in recent years – selling the “pizzazz”, to quote a former TTC Chair, while the system gradually declines thanks to penny-pinching by two successive City administrations.

Summary

This is a long article. Here are the highlights:

• Growth of express services is limited as much by the political question of transit funding as it is by planning and resource constraints within the TTC.
• The TTC’s bus fleet plan should be thoroughly reviewed to determine how more buses can be made available for service sooner than 2019/20 when McNicoll Garage opens.
• TTC budgets should reflect a return to full streetcar service on the streetcar lines in 2018 and the redeployment of replacement buses back to the bus network.
• Express bus routes that were added in 2016 have performed better than expected showing that these are popular services and should be expanded as soon as possible.
• New Rocket and express routes are proposed in two waves, one for 2019-21 and the second with a tentative date of 2026.
• Costs and revenues allocated to existing and proposed routes should be verified.
  • Update (text deleted): The costs shown for some routes appear to be in error if the methodology for costing used by the TTC is to be believed. (This issue has been referred to the TTC for comment.)
  • Revenue allocation in a flat fare system can distort the benefit of a route, and the measure of value should be based on usage not on a misleading allocation of fare revenue.
• Express buses provide a means of carrying riders on a route with a mix of short and long haul demand more efficiently and attractively than an all-local service.
  • Riders on express services travel further than on local services taking advantage of the faster trip between major points on the route. There are fewer riders per bus kilometre because there is less turnover of the passenger load on express services.
  • Express buses cost more per ride than local buses because of their lower turnover (i.e. more riders), but the overall route cost is lower with a mix of services.
  • A proposal by one member of the TTC Board to charge extra for express routes would be counterproductive.
• The Premium Express buses to downtown operate at a very high cost per passenger, although this needs to be verified in light of the issue with cost calculations. Demand on these routes is relatively light, and they contribute only trivially to reducing demand on parallel services. The TTC proposes to leave them in place at least until 2021, but this should be reviewed even though removing the services will be politically challenging.
• Transit Signal Priority is not just an issue for an Express network, but for transit in general. It should be pursued on major routes whether or not they include express operations.
• Route supervision will be essential to maintaining reliable service not just on express routes, but on the transit system overall.
• The staff proposal to “brand” services continues the TTC’s focus on marketing when what is needed is service. The “Rocket” name is already well-established as a service type in Toronto, as are the “E” express branches of various routes. In a few cases, the “wrong” name is associated with a service (some “rockets” are really just “E” services in terms of their service patterns), but this does not justify a complete rebranding exercise.

Constraints on Adding Service

In the short term, the TTC claims that no new services can be added to the network because of the shortage of vehicles and the lack of garage space. Both of these statements should be challenged.

• As the Flexity streetcar fleet builds up, the need to operate buses over major routes will evaporate. Indeed, the current situation with 65 buses on 501 Queen is due more to the convergence of many construction projects than it is to a streetcar shortage, although it gives the TTC a convenient way to reduce the total need for streetcars.
• Construction delays due to the TYSSE are largely out of the way and that line will open in late 2017. The net requirement for buses should go down in 2018 for services in that corridor thanks to the end of construction and shift of riders onto the subway.
• Construction delays due to the Crosstown LRT will probably last into 2019 when structural work at stations will reach a point that road lane occupancies will no longer be required.
• The TTC’s decision to increase its spare ratio for buses took a chunk of service off of the road. However, as the oldest and least reliable parts of the fleet are retired thanks to a very large bus purchase now in progress, it should be possible to reduce the spare ratio, at least for the short term.

A substantial portion of the bus purchase trumpeted by Mayor Tory to improve service actually went to increasing the spare ratio, not to service on the street.

The TTC’s Bus Fleet Plan tells the story:

The total number of buses planned for service when Mayor Tory was elected in 2014 was 1565, and the planned number for 2017 is 1560. Meanwhile the provision for spare buses has risen from 188 to 265.

One big problem for anyone touting the purchase of new vehicles is that any service increase requires more operators, and that is a cost pressure the TTC Board and Council fight against constantly. Whether there will be enough streetcars to restore full operation on Queen and other routes early in 2018 remains to be seen. If there are, the TTC faces the added cost of actually using its buses on bus routes, not as streetcar replacements.

The garage situation is not much better.

Even after the opening of McNicoll Garage, the system will still be over capacity by 176 buses, close to the equivalent of another garage, in 2026. There are no plans in the Capital Budget for such a garage. The overcrowding situation at garages will continue for the foreseeable future, and will likely remain one of the standard excuses for limits on service improvements.

To put it another way, one must ask whether the scope of any proposals such as the Express Bus Plan is dictated not by actual system need, but by political constraints on capital, facilities and operating subsidies.

Performance of Recently Added Express Routes

In March 2016, the TTC added four new Rocket services and one new Express branch (counting both Finch Rockets as one route).

Results for the Rocket services were generally better than expected with ridership coming in above the expectations. The only shortfall was the 188 Kipling South, but this is not surprising when one considers that the travel time saving (4 minutes one way) is less than the headway of the express service (7’30”) making it not worth the wait unless it is in sight when a local 44 Kipling South is available.

There is no question that riders like and will use express buses (whatever they are called) because their trip is faster. However, from a planning point of view, the implementation of a new route is not as simple as just rejigging the available vehicles. Many routes have strong local demand, and the more stops that are removed to create an express service, the more riders are excluded from using them. A good example was on Jane Street where service between the 35 Jane and 195 Jane Rocket was rebalanced to better match local and express demand.

The proportion of express riders varies from route to route (even between route segments) depending on local demand patterns. In some cases, a new express service requires a net addition of buses, especially if substantial net new riding is expected, and this runs headlong into the TTC’s budget constraints. Attracting new riders costs money, a fact that many members of the TTC Board and Council simply cannot understand, much less accept.

All Door Boarding

An important factor in getting the best utilization out of transit vehicles is the full use of internal space. This is especially important on routes with a high turnover of riders at stops along the way. Space is needed for people to move to and from the door, and bottleneck caused by passenger flow and vehicle design can reduce the effective capacity of a bus.

The TTC proposes to move to all door boarding on its Tier 1 (Rocket) services at a future date, but not now. This change will require fare inspectors for the bus network, and they will be much more dispersed across the city than they are now on the dense streetcar network. As a trial, the TTC proposes to move to all door loading on 192 Airport Rocket in 2018, but this is an unusual route with few stops, one of which is already a paid area (Kipling Station). This may not be an ideal pilot for a wider rollout.

New Rocket Services (Tier 1, All Day Service)

The study proposes that by 2021, Rocket services would operate on three new routes:

• Dufferin from Dufferin Loop to Wilson Station
• Steeles West from Finch Station to Pioneer Village Station
• Steeles East from Finch Station to Markham Road (with peak period service to Morningside Heights)

Although the map of proposals includes what is now called the 190 STC Rocket, it does not appear in the tables of Tier 1 and 2 services in the study [see pp. 38-40].

The existing 196 York University Rocket will cease operation when the subway opens in December 2017.

New Express Services (Tier 2, Peak and Some Midday Service)

The study also proposes a first group of new express services by 2021. Some services that are now styled as “Rockets” would be reclassified because their hours of service would not qualify them as Tier 1 routes (e.g. 186 Wilson Rocket and 188 Kipling South Rocket).

• Islington from Islington Station to Steeles
• Weston from Keele Station to Finch
• Markham Road from Warden Station to Centennial College Progress Campus
• Sheppard West from Yonge to Weston Road (extended from 84E Yonge to Sheppard West Station service planned for TYSSE opening)
• Lawrence West from Lawrence Station to Pearson Airport

Some services overlap on this map such as the UTSC Rocket and the Scarborough Express to Meadowvale, but only one type of service is shown.

The resources required for these services are detailed in the study. The plan is based on the acquisition of 32 new 60-foot articulated buses plus spares. Provision for this will be included in the 2018-2027 Capital Budget. Note that deliveries would be in 2020 and 2021.

Further Enhancements by 2026

In a true leap of planning faith, a second wave of improvements is proposed for the mid-2020s following opening of:

• Line 5 Eglinton Crosstown LRT
• Line 7 Finch West LRT
• Line 2 Scarborough Extension

There is also the question of how Line 6 Sheppard LRT might fit into the overall scheme, and of course there is political pressure for extending Line 4 Sheppard east to STC and west to Downsview.

Updated: According to @ttcdesign on Twitter, Finch will be line 6 because it will open before Sheppard East whose status is uncertain. The Express Bus report refers to Finch West as Line 7.

The study proposes several possible services:

• Rocket service on Eglinton West from Mount Dennis to Pearson Airport
• Enhanced Rocket service from Scarborough Centre Station to UTSC
• Rocket service on Bathurst from Bloor to Steeles
• New express services on Royal York (Lake Shore to Weston), Kennedy (Kennedy Stn. to Steeles) and Warden (Warden Stn. to Steeles)
• Extension of the 54E Lawrence East service west and south from Lawrence East Station (which will be abandoned) to Science Centre Station at Don Mills & Eglinton
• Extension of the Markham Road express service from Centennial College to Steeles

The resulting map would look like this:

Nine years is a long time to wait for new services particularly when they do not involve the long lead time of subway planning and construction. Such is the state the TTC finds itself in thanks to politicians whose only interest is in big ticket, photo-op projects, preferably those paid for by other levels of government who can be blamed when they don’t materialize on time.

In theory, even at a few percent a year, TTC ridership will have grown by close to 20% when this map gets filled in, and even more if the system could actually succeed in becoming more attractive. The resources needed for express bus routes will be small change beside the overall demand on the TTC’s bus fleet and network.

Route Costs and Revenues

The TTC faces constant political pressure to justify the money it spends on service and to produce some measure of the financial worth of a route. This can produce contortions such as tradeoffs between operating subsidies and the imputed value of time savings for riders. Sometimes the wider goods, such as reduced traffic congestion and the environmental benefit of shifting trips to transit, are invoked, although not in this study.

The fundamental problem is that it costs money to move people around, and the further they go, the more this costs. Calculating that cost is one part of any evaluation. The TTC uses three factors to allocate the cost of bus operations. For 2015, these were:

• Bus hours (primarily operator wages, benefits and service management overhead): $92.30 per service hour
• Bus kilometres (fuel and some running maintenance): $1.88 per bus kilometre
• Vehicles (maintenance that is performed based on the calendar, not on vehicle use, plus garaging costs): $150 per bus per day

These values do not include any capital costs for vehicles or facilities.

The trickier part of the equation is to calculate revenue for each route, and in practice this cannot be done accurately in a flat fare system. Assumptions are required, and they can produce anomalous results. For the purpose of this study, the TTC allocated $1.30 per boarding to regular fare services, and $2.00 per boarding for premium express routes. (A “boarding” is one passenger boarding one vehicle, and every transfer in a trip, if any, generates another boarding.)

The problem with this scheme is that it assumes all trips have the same number of boardings. (Another way to say this in planning lingo is to assume that the ratio of unlinked trips to linked trips is constant across the system.) This is not true, and the result is to overallocate revenue to trips with more than the system average ratio, and underallocate it for those below that line. The “worst case” situation is a “one seat ride” where there is no transfer, and the formula allocates considerably less revenue to the route used than it actually deserves to get. Conversely, trips using more than two routes (roughly the average for all trips) allocate more revenue than the rider actually pays. The effect is to divert, on paper, revenue from routes that have many “local” trips that do not involve transfers.

A further complication is the presence of flat-fare media (passes) whose users pay less per trip than those travelling on single fare media such as cash, tokens and Presto (unless they have a pass loaded onto their card).

Adding or removing a service not only changes transit’s attractiveness for riders, it can trigger peculiarities in revenue allocation. For example, if a trip that formerly required a transfer becomes a one-seat ride, the allocated revenue goes down because there are fewer boardings even though nothing has changed in the rider’s journey.

Revenue figures cited for routes should be regarded with caution because of these effects.

The much more important question for any route is this: are people using it? The cost of carrying riders varies immensely from one part of the system to another, but each part contributes to the overall goal of moving people. Inherently unreliable and potentially biased revenue allocations should not be used to create a “profitability” score.

As part of this study, the TTC surveyed other cities to determine which factors were used to warrant an express service:

• Travel speed improvement
• Ridership/capacity (sufficient demand for a separate service, potential for optimization of route operation)
• Strategic land use, both existing and potential, that transit could reinforce
• Average trip distance
• Financial standards

It is ironic to note that the first two criteria were used by most agencies surveyed, and the proportion fell to only a third for trip distance and financial standards. Another way of looking at this is that making rides faster, optimizing operations and supporting land use goals were more important than financial targets. By implication, those goals justify the investment. This is a basic fact of transit economics that is, to be kind, overlooked in many of Toronto’s debates.

The Savings of Express Services

Where both a local and an express service co-exist, the express service can hive off riders from the locals so that, in total fewer buses and bus hours are required to move passengers. For example, if the express service runs 10% faster, it requires 10% fewer buses and operators. (The vehicle mileage remains the same, however, because the 10% fewer buses run 10% further per hour.) This produces savings which vary from route to route depending on the speed benefit gained. The Rocket services show a better improvement because their speed relative to local services is higher thanks to the limited number of stops enroute. The “Premium Express” buses are quite another matter, but I will discuss these separately.

The two tables below present these savings from different points of view. Table 2 gives the saving in operating cost of the express services over what they would have cost operated as locals. The second table (unnumbered in the study) blends in the cost of local services in each corridor to show the level of saving for the corridor, rather than just for the express operation.

Demand on Express Services

The demand on the express services varies considerably from route to route, and is worst on the Premium Expresses.

Note that many of these routes also operate during off-peak periods, but their ridership is not included here.

A high boardings/hour value can reflect several different factors:

• If a route has strong bi-directional demand, the buses pick up riders in the “counter peak” direction, not just as a service taking riders to (or from) one point such as a university campus or subway station.
• If a route has enough stops, demand patterns will exists not just for the endpoints, but between stops along the route. This allows a bus to turn over its load and handle more boardings than would be possible if everyone wanted to go to the same destination.
• If a route is short, the turnover rate necessarily goes up because one boarding does not require as many bus hours as on a longer route.
• If a route operates comparatively quickly, fewer service hours required to carry a load over a trip, and the boardings per hour go up.

The important point here is that the performance of any specific route, or the median of them all, does not necessarily indicate how any new service might operate, nor should it be seen as a target to which services should aim. Each route’s characteristics affect its operating cost. The same situation applies to local services where the cost per boarding varies immensely from route to route, and very short routes typically have very low costs per boarding. It is simply impossible for a rider to consume much time on a short route before being forced to get off.

The cost per boarding for express services is actually fairly high even though provision of expresses reduces total operating cost. The reason for this is that the expresses draw the relatively expensive riders, those who are travelling longer distances and benefit from the express service, while the local trips remain behind on the local route at a comparatively low cost per boarding because there are more short trips, turnover and boardings per service hour on the locals. Again, this is an example of the dangers in looking at only selected parts of an operation and concluding that express buses are “more expensive” and might even warrant a higher fare.

Update: The costs shown in the table below are net of allocated revenue. Routes with a high boardings/hour value will be allocated more revenue/hour and have a correspondingly lower net cost/hour.

Update (text deleted): The validity of these numbers bears further scrutiny as we will see below.

The cost per service hour varies considerably from route to route, and a great deal of the difference comes from operational issues that are not directly due to “express” service per se. For example, if the proportion of hours required to get buses between a garage and a route are higher compared to hours spend actually serving the route, then those “deadhead” hours count against costs but provide not direct service. This can be affected by the distance from a garage to a route as well as by the number of hours the service lasts.

For example, many of the Rocket buses stay out all day and their garage trips are a comparatively small part of their total hours. Express buses on the other hand tend to be peak only, and can run up a substantial cost just getting to and from service. Premium Express buses may travel a long way from their garage to reach their route, make only one trip, and then return to the garage. The cost to serve that one trip is very high per hour because there is so much deadhead mileage.

Update (text deleted): These values do not make sense because the cost per service hour is well below the allocated cost/hour of service ($92.30) for several routes, for the system-wide and rocket medians. This should not be possible. The very low cost for 195 Jane Rocket is particularly suspect.

Gross Operating Costs (Section added June 20, 2017)

Although the study does not include a table of gross operating costs, these can be calculated from the available data:

• Net cost per peak hour
• Boardings per peak hour
• Allocated revenue per boarding

A few points of note in this table:

• The gross cost per hour will vary between routes for three reasons.
  • The faster a route operates, the more mileage buses will accumulate per hour, and this will increase the contribution of the distance-based element to the total. Fewer buses are needed to provide service on faster routes, but their allocated operating cost per hour is higher due to the distance component.
  • The proportion of deadhead mileage for each route depends on the distance from the garage and on the length of the peak period. If a bus operates only a single trip, then a higher proportion of its total mileage and driving time is taken enroute to and from the route increasing the effective hourly cost of actual in service time.
  • The fixed cost of vehicles is allocated against in service time, and the fewer hours a vehicle operates, the greater the contribution of this factor.
• The revenue allocated to premium fare boardings is only $0.70 more than for standard fare services even though the premium is at least $2.00 per trip (via Metropass sticker), or a full extra fare (cash, ticket, token). This could understate the revenue actually due to these routes, especially if their customers are primarily “one seat” rides where the entire fare should be allocated to the boarding, not just a partial value. The gross costs would not change, but the per hour and per boarding net costs would be lower.
• The net cost per hour of the 141 Mt. Pleasant Premium Express is particularly high.

The issue here is that assumptions built into cost and revenue allocations can affect the standing of various transit services. If these assumptions skew the purported “profitability” of routes, this can lead to misguided policies.

Trip Lengths and Travel Time Savings

The table below contains interesting comparisons of the express services with their corresponding local routes.

• The average trip length on expresses, where there is a local for comparison, is generally longer showing that the express service is attracting those riders for whom the speedier service is a real benefit. Where the change is low, this suggests that riders get on whichever bus shows up first rather than waiting for the express.
• The travel time savings, as a percentage, vary depending on the speed of the express and local services.
• The bi-directional load factor shows the degree to which each route serves destinations in both directions such as a subway-to-campus link. This shows the importance of looking at service design and stop patterns without overemphasizing a specific destination. Not every rider wants to get to a subway station for a trip to downtown.

Premium Express Routes

The Premium Express Routes are historically poor performers with low demand and high cost, although this has not routinely been reported as part of the occasional system stats. These routes exist as much thanks to political intervention in their respective origins as they do from any planning principles or standards to which other routes are held. Any attempt to kill them off and reallocate their vehicles is met with a call to “just give them more time” to reach reasonable performance goals.

The Avenue Road and Mt. Pleasant expresses nominally provide subway relief, although their infrequent service an limited number of trips barely makes a dent. With 380 boardings per day (AM and PM peaks combined), these services represent roughly one subway car’s worth of demand. The study considered the possibility of attracting more riders by changing to a standard fare, but the market for these services is limited to those who are directly served by stops in the origin area, not by transfer traffic from subway feeders. One must ask why a rider would get off a bus on Eglinton at Mt. Pleasant to connect with a half-hourly service to the core when, despite the challenges, they can simply stay on board to Eglinton Station. Even with more service, these routes offer no advantage in trip time especially when the need to transfer at outdoor locations is considered. (The construction effects of the Crosstown stations are an extra challenge.)

The Don Mills, Beach and Humber Bay services exist at the specific behest of local Councillors who, at the time they were implemented, were members of the TTC Board. This is not to say that they do not provide a service to their riders, but one cannot regard them as anything more than bandaids to fix problems in specific neighbourhoods. The idea of a point-to-point service into the core is only remotely feasible when there is an uncongested corridor for truly “express” operation. In the case of the Beach and Humber Bay services, the “problem” being addressed is the poor service on Queen over many years, and resources might have been better focused on that route’s problems than on serving a portion of that route’s demand with a premium service.

TTC staff take a cautious approach to removal of these routes and recommend that they be reviewed in light of future changes to the network that could deal with their problems:

• In the case of the Beach and Humber Bay routes, future capacity improvements with new streetcars on Queen, plus a future Waterfront West LRT taking a quicker route into the core, could replace these bus services.
• In the case of Avenue Road and Mt. Pleasant, improvements to service on the Yonge line could actually provide space for riders to board at Eglinton inbound. The report cites the 20% improvement in subway capacity possible with automatic train control, although the TTC does not actually own enough trains to operate this level of service, and a good deal of any capacity increase is likely to be backfilled by latent demand further north.

Such a review would not take place until 2021, and so existing Premium Express riders need not fear for the imminent demise of their service.

Transit Priority Measures

As part of the Express Bus network, the TTC seeks to substantially expand transit signal priority to many routes that do not now have it (those not outlined in yellow on the map below). A more important question, however, is why this should have to wait until the new services are implemented. The report recommends:

4. [That the Board] direct TTC staff to work with City of Toronto Transportation Planning and Transportation Services staff to implement transit priority measures including transit signal priority and queue jump lanes for existing express routes and report back jointly on progress.

TSP is a benefit that all transit routes, whether they have express services or not, can benefit from. For routes with mixed express and local service, there is the question of whether TSP can or should be implemented only for express services, or for all service on a route taking into account the different stopping patterns. The map indicates that TSP is already in place on Jane and on Finch which have express service, and these routes could provide some preliminary information about benefits and challenges today.

The TTC has long wanted to have queue jump lanes at certain locations where buses could scoot around the queue for right turns. Like TSP, these are physical changes that do not have to await the rollout of the new express network.

Route Supervision

The issue of service reliability is discussed frequently on this site, and I could not resist including this paragraph from the study.

4.4.5 Route Supervision and Monitoring

Effective route supervision is fundamental to the provision of good transit service. Route Supervisors undertake many types of work, all geared to ensure good service reliability including monitoring/adjusting transit service and responding to emergencies and customer service issues. On-street supervision is essential to providing good customer service because it allows direct observation of crowding and scheduling issues, with real-time adjustments as well as development of longer-term solutions. Route supervision will be vital to the initial roll out of many of the new express services. The VISION CAD/AVL system, which will be in place, will further enhance the route supervisor’s abilities to manage these services.

VISION CAD/AVL is the TTC’s new system for Computer Aided Dispatching / Automated Vehicle Location that is in the preliminary stage of implementation. It will replace a three-decade old system that provided substantially less functionality for route supervision and management. The real challenge for the TTC will be to adjust its “culture” to reflect the abilities the new system will provide.