Recently, the TTC Board, a largely dormant entity during the era of Rob Ford and Karen Stintz, decided to take a more active role in oversight of the organization by striking new subcommittees. In addition to the already existing Audit Committee, there is now a Budget Committee with members Rick Byers, Councillor Shelley Carroll, Councillor John Campbell, Councillor/Chair Josh Colle, and Councillor Joe Mihevc.

Agenda materials for this Committee are available on the TTC’s website, but on a different location from the those for full Board meetings. The Budget Committee met on June 17, 2015, and its agenda included two reports related to ridership and service levels:

• Ridership Forecasting Process
• Service Standards and Initiatives

These are intended as introductory overviews for Commissioners who are not steeped in the details of TTC planning or policy, and they give a general sense of management’s focus as the TTC enters the 2016 round of budgets.

Note: The Service Standards document has most of its content in portrait format even though the pages are primarily in landscape. Save the document, open it with Adobe Reader, and then rotate the displayed images.

Ridership Forecasts

Ridership forecasts are done at the system level to estimate the number of trips people will take on the TTC in the next budget year, and also to estimate the effect of any proposed change in fare structure on both ridership and revenue.

Many factors affect ridership, and some cannot be predicted in advance, notably the weather. The particularly cold 2014-15 winter will cause the TTC to miss its original target for 2015 ridership. Although there will be more riding in 2015 than 2014, there will not be as much as expected.

This situation hurts revenue, although that can be offset by deferral of service improvements, a tricky decision in a year when such improvements are a central part of TTC and Council policy.

A particular problem arises when there is a shortfall because all of this falls against the subsidy requirement. In good years, the TTC does better than expected and money “left over” (i.e. unused subsidy) that it never draws from the City – this goes into the City’s operating “surplus”. A vital point here is that it is not new revenue, but rather money that was not spent – like owning a car and being pleasantly surprised that the cost of gas was less than you expected.

A shortfall in TTC revenue, currently projected at about $10.2-million or 0.85% for 2015 in the recent CEO’s Report [see p. 25], will be offset with lower expenses. However, if this amount fell to the City to absorb, it would represent a 2.1% increase in the subsidy required. Tempers would flare and the budget hawks would call out TTC management for their flagrant inability to budget. This is the other side of the two-edged sword of high farebox recovery. Even if Council is parsimonious about subsidies, there is always the farebox to fall back on at budget time. But at the end of the year, an unexpected deficit is magnified because it is weighed only against the lower level of subsidy (an expense from the City’s viewpoint) rather than against overall revenue (the rider’s viewpoint).

For many years, the graph below has appeared in TTC budget presentations to illustrate the relationship between ridership and employment (as a surrogate for economic activity). This held true for many years including the huge drop from the 1990s recession. However, employment initially bounced back faster than ridership, but then flattened out. Ridership, however, continued to grow showing that it is affected by more than the job market. A related TTC statistic is the growth in off-peak demand, riding that is not necessarily linked to jobs, and riding that may be preferentially attracted to transit simply because there is some capacity available for such riders. Very large, fast changes in employment will almost certainly accompany major economic shifts, and these will show up in riding. A longer steady-state of employment, or modest growth, may be only one factor in transit usage.

When TTC fares change, the effect on riding will be from a combination of factors, and these combine to produce an “elasticity” measure to estimate change. For example, if fares go up by 1% and ridership goes down by 0.3%, then the elasticity is -0.30. This is not a linear relationship because for small increases, the change in cost does not outweigh the benefits of using transit – for example, the marginal cost of switching to or increasing the use of a car may be prohibitive. However, a larger jump in fares, particularly one accompanied by budget-induced service cuts, sends the message that the value of transit is declining. Riders are particularly sensitive to service quality, and polling routinely shows that lower fares are not their primary concern, better service always tops the list.

Although the transit industry standard for fare elasticity is -0.30, the TTC’s actual experience is a much lower factor, -0.10. For 2015, the weighted average fare increase across all forms was 3.4% leading to a projected riding loss of 0.34% or 1.8m rides. An important factor here that is not explored is that the lion’s share of the increase was absorbed by Metropass holders who have a strong incentive to continuing TTC use at the same level. Therefore it is possible that the low elasticity is achieved by placing the largest burden of a fare increase on those least likely to change their riding behaviour, and the -0.10 factor might not hold if this were done across the board. (See also the details of the 2015 adjusted projection below.)

On an historical basis, ridership has generally run slightly ahead of estimates.

This puts Council in a position where it may falsely expect the TTC goose to keep laying, if not golden eggs, at least a modest clutch of silver than can be used to offset City financial woes. This money may well come from an overshoot in the actual versus budgeted fare revenue which, in turn, could arise from an unnecessarily large fare increase. That’s the sort of thing that is easy to say in retrospect, and the TTC is better off politically to be in a surplus than a deficit position at budget time, especially when service improvements are the order of the day.

For 2015’s Actual vs Budget, and for 2016’s projection, the ridership looks like this:

Note that no change is forecast due to a fare increase or adjustment in the Metropass trip multiple (the ratio between the value of a Metropass and a token). This projection, therefore, is subject to modification based on whatever additional revenue will be needed to balance the Operating Budget. It is worth noting that the estimated effect of the fare increase in 2015 has been adjusted upward from 2m to 3m rides, and so the elasticity might not be as low as -0.10 after all. A 3m ridership loss in 2015 would imply an elasticity of at least -0.15, and the TTC should take care not to underestimate this effect in their 2016 projections.

Service Standards

The presentation about service begins with the premise that the right service, its amount (frequency and capacity), time and place are important. Mobility and service quality should be maximized. These are noble goals, but they are subject to affordability both by riders (through fares) and taxpayers (on whose behalf politicians cry ceaselessly for lower taxes and better “efficiency” in the delivery of public services including transit.

What is rarely asked or discussed is the premise of what transit service should be, what is our goal for its overall quality, and then how much are we prepared to pay for it. Almost always, the discussion starts with the premise of cuts, and must be wrestled back to a break-even position just as a starting point. That approach will always pre-empt discussion of ideals, although this constraint rarely applies to pet rapid transit projects.

Service decisions are, so they say, applied on a well-informed basis [see p. 3]:

Decision rules, standards for service resources:

• objective, data-, fact-based:
  • ridership
  • origin-destination data
  • travel behaviour research
• grounded in business logic, principles
• transparent, quantifiable, reproducible
• applied consistently, fairly, equitably

Yes, this sounds very nice, but there is one additional factor that shows up regularly: “budget availability”. In other words, if we can’t afford it, we don’t get it. That’s a reasonable budgetary approach, the sort of thing households are told to do every day, but it begs the questions “what are we not getting” and “what more could we have and at what cost”. If your budget forces you to eat mac and cheese rather than steak, you know it, and you probably even know how much more you would need to afford steak at least a few days a week. The same is not true of transit service. Decisions to run less service, or to not run service at all, occur in a black box usually as much a political exercise (“hit this target”) than one of policy (“what do we really need”).

Once upon a time, the TTC reviewed requests for new and improved services every year, and evaluated the requests against a set of criteria. This has not happened since 2008. Since then, the standards have been tweaked first to impose service cuts, and then to fine tune even those to a less draconian approach, but all “subject to budget”.

The standards have many components, and one or more will apply depending on the type of change under study.

System Structure, Design and Coverage [pp. 4-6]

Stay with a grid network to maximize travel options, but focus on the subway for fast trips. These goals can be contradictory when major subway terminals distort routes and service patterns to support a primary commuting pattern to the core area at the expense of trips that might not even require subway travel.

Take the fastest route – don’t spend time twirling through sidestreets – and don’t duplicate an existing route. Avoid transfers as these annoy riders.

That last standard is rather amusing, both because the entire TTC network is based on transfers, and because of the special fondness for transfer-free rides by those who advocate subway extensions. The point obviously is that transfers have their place, and the question really is when is it better to have a transfer, and when a direct route.

Routes are designed to provide service within a ten minute walk, and a new route request that would be within a five minute walk of an existing service is not entertained. For overnight services, the target is a walk of 15 minutes or less. Exceptions include physical or geographic barriers, although one might expect these to have an effect on walking times. For example, walking across Grenadier Pond is generally frowned upon even when it is frozen, and walking across the Don Valley requires tightrope or hiking skills possessed by only a minority of transit riders. Rail corridors and labyrinthine street layouts are more mundane examples.

Return on Investment and Productivity Standards [pp. 8-9]

Any new service is priced for its marginal operating cost and revenue. How many more vehicles and vehicle hours will be needed. How many new trips might be attracted. What is the net loss (there is almost never a net profit or the route would already exist) for this change?

The TTC has a formula saying that 0.23 new riders should be generated for each net dollar of new cost. This is equivalent to saying that each new rider should cost no more than $4.35. The problem with this formula is that the TTC claims it is “dimensionless”, that is the value stays the same regardless of changes in the underlying factors. This would be true only if the value of a new rider (e.g. the fare) actually went up proportionately to the cost of service keeping the ratio between them constant.

In fact, the ratio is not constant for at least two reasons. First, the marginal cost of operation is affected by many factors, while the fare level is as often set as a matter of policy related to the total subsidy plus fare budget the City wishes to provide. Indeed, if fares are allowed to rise relative to costs (the farebox recovery rate goes up), then the net cost per new rider goes down.

This standard originated at a time when there was a “maximum permissible subsidy” of five times the average subsidy. For 2015, the average subsidy is 89 cents (actually it is lower, but that is yet another problem), and five times this value would be $4.45. Amazing! The numbers almost match! However, in 2015, the “operating subsidy” actually includes at least 6 cents of capital-from-current funding related to the purchase of new buses (more if we include the cost of leasing garage space for them), and so five times the average subsidy is only $4.15. It’s not much of a change, but it shows how the .23 factor is sensitive to what went into the budget in the first place, and that’s before we consider inflation. If, for example, fares are frozen but marginal operating costs go up, more riders are needed to balance the cost of new service, but the formula does not take this into account. If the TTC is going to start evaluating service changes again with this criterion, the underlying methodology needs review.

As for productivity, this has been subject to more than a little gerrymandering in recent years. The formula is supposed to be that a route (or segment or period of operation) should generate at least 15 “boardings” per revenue vehicle hour. (A “boarding” could be an actual fare paid, or a transfer, or someone getting on with a pass. The average trip on the TTC includes about two boardings, but the subway is counted as one regardless of route-to-route transfers.)

In cases where there is a longer walk to transit (600m or 8 minutes, or the brisk rate of 4.5km/h), the standard can drop to 10 boardings per vehicle hour. Recently for the all-day, everyday service, this standard was reduced to 9/hour. This number is driven by how much service the TTC can afford to add, not by a detailed study establishing that “9” is a magical number.

Buried in all of this, of course, is the question of counting existing riders and estimating new ones, the subject of some concern during the era of service cuts.

A 10 boardings/hour standard also applies in cases of adding periods to service (for example, adding evening service to a route that only runs during the daytime), and to requests for additional trips at the beginning and end of the day (usually very early morning or late at night).

It is clear that elements of this standard have evolved out of sync with each other.

Effect on Customers [pp. 10-20]

Transit riders react to components of their trip different ways with actual movement in a vehicle being the least “costly” to their sense of trip value because they are actually going somewhere. In order to calculate a measure of a change, the TTC determines the effective value or “inconvenience” added or lost for all riders by the following formula.

• 1 minute of travel time = 1 minute’s value
• 1 minute waiting = 1.5 minutes
• 1 minute walking = 2.5 minutes
• 1 transfer = 10 minutes

These factors are for average riders, and they will affect some (such as seniors or those with mobility limitations) more seriously. Also the perceived inconvenience of a transfer will be subject to conditions such as physical ease of the connection and the degree to which the transfer introduces annoyances such as a long wait for an infrequent or unpredictable connection. Some riders will benefit from a change, and the calculated value (perceived cost) of their trips will go down. Other riders could face difficulties such as a diversion from their usual route and hence extra travel time (an extra “cost” albeit at the 1:1 rate of in vehicle time). Stir all this together, and you get a number indicating the weighted “value” of a change.

One problem here is that this is very sensitive to the TTC’s service design. If a change is implemented with an added vehicle, this is an extra cost, but it might avoid a degradation in overall service (e.g. preservation of existing headways rather than “stretching” an existing route to cover new territory). Some proposals are weighed down by being in a part of the network where adding new service is comparatively expensive. Other factors such as access or simple the weight of political lobbying must come into play.

A special case is the introduction of express services on existing “local” routes. To be truly “express”, a service must by definition not stop very often, and ideally such stops should not involve a lot of turnover in the passenger load. If the demand pattern is such many passengers use only a few stops, they are going to a common destination, and there is an opportunity for “line haul” non-stop service, then there is certainly a market for an express operation. However, the riders who remain will not have as frequent service as they did in the all-local configuration, and their needs have to be balanced in the evaluation.

If one proposes to spend more so that express service does not cannibalize the local service (or at least as badly as a zero-sum service design might), then there is more headroom to operate express buses. This is precisely what is happening with new express services to be added later in 2015. However, there is no formula yet to determine the extra subsidy a route might receive under this initiative.

There is a maximum waiting time of 30 minutes for bus and streetcar routes, and 5 minutes for the subway. In practice, the entire streetcar system will be part of the 10-minute network (something for which there is no standard yet defined) in the fairly near future, and so the 30 minute rule only applies to overnight and early Sunday streetcar services. The subway standard is interesting because both 3 Scarborough and 4 Sheppard already exceed the five-minute rule.

Crowding Standards

Finally, there are crowding standards, and these are the most notorious for budget related shortfalls in service. The full table of crowding standards for 2014 is on the TTC’s website.

In this version, peak standards are about 10% worse than they were during the “Ridership Growth Strategy” era of David Miller, and these have not yet been restored to “RGS” levels because the TTC has no spare vehicles, let alone the budget to operate them. It is not yet clear what program the TTC might launch to move back to RGS loading standards over coming years, and current changes only go half way, to a 5% improvement (down to 47-50 from 50-53), not the full 10%.

For offpeak service, the standard for routes with a vehicle less frequently than every 10 minutes is a seated load. For more frequent services, the standard is a seated load plus 20% which, on buses, brings the offpeak standard close to the peak standard, a condition of little surprise to regular riders. These standards need review especially for bus services where the amount of standee space on low-floor buses is less, proportionately, than it was on the high-floor vehicles when the standard was created.

An important consideration about crowding is that jamming in more riders can be counterproductive. If riders cannot move through the vehicle, enter and leave with some degree of ease, stop service times go up, and “friction” between riders is more than just a question of squeezing past someone to get to the door. Overcrowded vehicles take longer to make their journeys, and this affects service quality and the cost of operation.

What we do know [p. 35] is that 5% of peak bus service is overcrowded, and this affects 14 routes. If the newly improved peak crowding standard is considered, then 15% of service is overcrowded representing 34 routes. This will not be fixed overnight, and the bus fleet (and complement of staff to drive and maintain it) will have to grow. Will Council actually fund this?

A policy debate about loading standards would be an excellent one for the Commission and for Council, including an estimate of the costs and benefits of various levels of change. As always, the tug-of-war between budget hawks and those who would improve service is “don’t tell us we can’t afford it, tell us what it will cost, and then we will make an informed decision”.

One way of looking at the network is to map the portion of the city that is within a certain walking distance of transit. This is shown on a series of maps [beginning at  page 31 of the pdf]. What is missing here are maps for Sundays when the network has been particularly sparse, or a comparison of the before/after access patterns for restoration of off-peak services.

Wrap-Up

In the near future, service additions will be driven more by broad policy initiatives such as the restoration of off-peak services, improved crowding standards, addition of express buses and, eventually, opening of new rapid transit routes. We are unlikely to see a systemic review of network “efficiency” for at least a few years while the effects of these improvements work their way through the system. However, this does not mean that we can ignore the “standards” because they can bite the moment someone asks “why are we spending so much” or “why is that bus empty”.

The details explained above will have worn out all but the most dedicated readers, and the typical politician would have called for the end of the presentation long before now. However, we need to  understand the machinery, and its underlying philosophy. Without them, transit planning becomes little more than hacking your way blindly through a swamp.

The standards need to be consistent. The costs and benefits of achieving them must be understood and accepted as part of an overall transit plan.

On one subject, the standards are silent – to what extent should transit service operate “unprofitably” as an adjunct to development, to provide service to a new or growing area and establish a “transit habit”? For many years, this has been a small issue in Toronto, but a big one in the 905 where car-oriented travel is the dominant form. However, Toronto has new residential territories (e.g. the Waterfront), and major rejuvenation in others. Should transit be “just enough” and always play catch-up with growth, or should it lead growth? That was the idea in transit some years ago, and Toronto needs to revisit this crucial question.

[The presentation includes sections on WheelTrans and on various service initiatives. I will turn to these topics when more information about the 2016 Operating Budget is available in July.]