This article is the third section of my critique of the December 2013 review of the Metrolinx Big Move Plan written by Michael Schabas for the Neptis Foundation. It should be read in conjunction with Part I and Part II.
6. Subway Schemes
Schabas begins by talking about TTC’s “innovation” in subway construction. This is something of a misnomer. Certainly the free transfer between subway and surface routes at half of the stations on the original Yonge line greatly simplified passenger flow, but this was a direct offshoot of the fact that Toronto had always had free transfers between its routes (subject to a set of rules that may finally disappear with the implementation of the PRESTO fare card). Subway-surface interfaces go back to the early 20th century, notably in Boston.
The 75-foot subway cars had lightweight aluminum bodies and could accelerate faster than their predecessors, the lumbering red “G” cars, but the TTC only used their higher performance capabilities for a short time. The first series of “H” cars (the ones purchased to open the BD subway) had motors that failed at high speed due to internal vibrations. Moreover, higher speeds required better standards of track maintenance. Even though faster trains could cut operating costs and speed journeys, the TTC has never reverted to “high rate” operation even though the “H1” cars have long been retired.
This review is an odd way to start a commentary on subway proposals because it misses the point that “innovation” is relatively easy when a system makes a generational change – a new line, a major fleet replacement – but is more challenging in an environment where just keeping the lights on day-by-day is uncertain. The real problem for the TTC is that it has operated in that constrained environment for so long that an entire generation of management knows nothing else. Coupled with a long-held sense that Toronto need not look elsewhere for inspiration, this is a recipe for stagnation.
The Spadina Vaughan Extension
Schabas discusses the ridership and economics of the Vaughan subway extension, and proposes that some of the investment value be recaptured through “smart pricing”. He misses the basic political point that York Region came into the project on the basis that the TTC would take all of the revenues and most of the operating costs, and that the Toronto fare zone would be extended to Vaughan. This may not make economic sense, but it is the political arrangement we have inherited. It is likely that York Region expects similar treatment for the Richmond Hill subway, and their citizens look forward to a cheap ride all the way to downtown largely subsidized by Toronto taxpayers.
Appendix M of the Environmental Assessment for the Spadina extension includes ridership estimates and a breakdown of their location on the route. The daily forecast is for 100k passengers of which
- 24k would use Steeles West Station,
- 56k would use York University Station,
- 10k would use each of Finch West and Sheppard West.
(At the time of the study, the extension to Vaughan was not yet in the mix.)
The projected southbound am peak demand at the existing Downsview terminus would be 12-15k depending on the demand model used, and this translates to somewhere around 7.5k for the peak hour. Many of these will be existing transit users and so will not contribute additional fare revenue as Schabas notes.
An important point here is that travel to York University will not contribute to peak demand. That university creates its own “problem” for the fare structure because all York Region buses now terminating at York will be banished to Vaughan Centre Station. This means that York Region students will have to pay a separate TTC fare to ride the subway a few stops south to the university. This group is an obvious target for changes to “cross border” fare structures.
As in other cases, Schabas proposes higher fares than are currently planned for riders from Vaughan to Toronto, and includes the imputed value of diverting road users to transit as an economic prop to the subway’s financial picture.
A factor he misses here is that the new subway is expected to divert a few thousand riders off of the Yonge line’s peak contributing on a small scale to Bloor-Yonge relief.
Schabas concocts a business case for the extension with the following argument:
- The TTC will need 6 to 10 more trains to provide service at a cost of $120m. In fact, the cost of these trains is about $160m (see TTC’s recent report on new subway cars). Moreover, when the Spadina extension was priced and contracted, the TTC had not yet planned to move to automated operation. This was not included in the signalling specification, and it became an “add on” cost to the project. In any event, Schabas predicts a net operating cost of $25m/year, or $575m on a net present value basis over a 30-year span.
- This cost would be offset, at least on paper, by the NPV of imputed savings to road users which cannot be captured as revenue, and by proposed higher fares that are unlikely to be implemented. In practice, the extra cost of operating the subway will fall on the TTC’s cost base with the lion’s share going to Toronto subsidy payments, fare increases or service cuts elsewhere in the system.
This is the sort of financial hocus-pocus behind far too much transit “planning”.
The Richmond Hill Extension
Schabas begins by talking about relief of demand on the Yonge line to make room for new riders from Richmond Hill which, as he has previous argued, would be provided mainly by GO Transit. He proposes the elimination of four stations leaving only Steeles and a single Richmond Hill Station on the initial line to reduce its cost by, say, $800m. Only 10% of the projected ridership would be lost. About 1/3 of the riders on the extension would be net new to transit.
There is no discussion of the effect frequent service on GO Transit would have on subway demand. Obviously this would be sensitive to two factors: how much of the projected demand actually originates in areas served by GO, and how would the fare structure work against attracting riders away from the subway. Riders on the extension who originate at Richmond Hill are candidates for a frequent GO service, but many who would arrive at Steeles Station would be much harder to divert, in part because many would already have paid a TTC fare and could use the subway at no extra cost.
Schabas proposes a new subway yard under the Ontario Hydro lands at Langstaff, although this option had already been rejected by Hydro when his report was written. The TTC has two alternate schemes for storing additional trains involving inline trackage either north of Finch Station (there is already a three-track section north of Finch that could be extended to Cummer) or at Richmond Hill. All of this has been publicly discussed in TTC reports.
Once again, Schabas disguises a fare increase under the guise of “smart pricing” and presumes a higher fare for travellers from north of Steeles Avenue than the subway advocates in York Region expect to pay. As with the Spadina extension, he includes this revenue and imputed savings to road users through reduced congestion as offsets to the new operating costs the line will incur.
It gets tiresome to read of how magically new lines will pay for themselves out of revenue we will not see and “savings” we cannot recoup, while Toronto taxpayers and transit riders foot the bill for suburban expansion.
Schabas talks about operational improvements from unattended operation (about which more later in this article) including the “potential to inject trains” from sidings [Page 51]. There is one small problem: the TTC does not have many sidings available in which spare trains could sit, and what tracks exist are normally kept clear for short turns. The TTC already positions “gap trains” at a few locations on the YUS as peak relief, albeit with the cost of a crew and the dedication of peak period fleet to possibly a single service relief trip.
Adding capacity through the day will require more trains. With the Vaughan extension, TTC will need about 750 cars to operate the current service. [Page 52]
Off-peak capacity does not require more trains because, by definition, off-peak service is not as frequent as peak service. As for the fleet size, the actual YUS fleet including the Vaughan extension is 70 trainsets, and a further 10 are now on order for future service improvements. Presuming a target of 1/3 capacity increase with ATO, the number of trainsets needed is 93 (558 cars), or 13 more than current plans.
When Schabas talks about the benefits of “modernization” on subway ridership, he is double-counting growth unless all of the additional riding will come during the off-peak when there would be capacity to absorb demand beyond what is already projected.
The Downtown Relief Line
The DRL is a major component of The Big Move that Schabas rejects outright. In doing so, he fundamentally misunderstands what a DRL might do and how it would fit into the transit network. In this he is not alone.
A frequent remark about any new subway line is that it must be surrounded by density to be successful. This is transparently false as we can see from parts of the YUS and a good deal of the BD line. A similar comment might be made about expressways whose purpose is not just to serve end-points of trips, but to carry travellers who assemble and disperse via the local street system. Density is certainly nice to have, but there is also the basic fact that if many people want to travel from “A” to “Z”, capacity is required at “b … c… d …” to make this possible.
A related issue about any new subway line is that it should provide some benefit beyond simple redirection of traffic around Bloor-Yonge. To that end, a route serving Don Mills & Eglinton, Thorncliffe & Flemingdon Parks, East York’s high rise corridor, and the developing lands near the Don and eastern waterfront will do far more than offload a critical junction. It would provide an alternate route from downtown to the northeast linking many neighbourhoods that are geographically isolated.
If one looks only for “solutions” at Bloor-Yonge, then one produces only “answers” addressing one part of a larger potential in the network. There are some parallels to the “aha” moment when someone recognizes what GO Transit could do if only it were the equivalent of a “regional subway”, not just a commuter line for park-and-ride travel.
Once again, we bump into the City/Metrolinx conundrum of local versus regional travel. There are strong local reasons for building a DRL or “Don Mills” subway as I have dubbed it, but Metrolinx’ fixation on the regional can downplay, even denigrate, the city building rationale for this new line.
A bit of history is necessary. Once upon a time, a proposed alignment for what would then be a streetcar subway on Queen followed the rail corridor southwest from Carlaw & Gerrard to Queen. The idea of going straight down Pape to Eastern arose from the planned use of ICTS technology, and that scheme needed a yard. Property was available on Eastern (now part of the film studio district), and so the line was redrawn to access it. (I should mention that our provincial friends would have built an elevated structure across the front of Union Station after the line emerged from the rail corridor at Bay Street. The drawings did not show how the ICTS station would affect the street.)
Other schemes for a Queen subway have include turning north from Queen onto Roncesvalles and thence to the BD subway at Dundas West. There would be, at best, one intervening station even though the demand along Roncesvalles is scattered along its length, not at one point. A route up the Weston rail corridor would be much better, but we have just spent a small fortune rebuilding it for the UPX which, sadly, is not designed to provide local service despite chewing up lots of real estate enroute. But I am getting ahead of myself.
On again, off again support for a DRL arises partly from an old city pretense that core area development could be strangled if no new transit capacity was built to serve the area. We got the Sheppard line instead, but GO didn’t hear the message and proceeded to fuel increased travel to the core as its network expanded. The other big problem a DRL faces is sticker shock, and this isn’t helped by the repeated quotation of a price for the entire Dundas West to Don Mills line as if it were applicable only to the Pape to Union leg. Indeed, focus on that short piece contributes to the perception that this is a line for “downtown” and of no benefit to the suburbanites whose taxes will pay for it.
(Shall we pause briefly here while other aggrieved municipalities mourn the waste of their tax dollars far from home?)
Schabas notes that travel time savings to downtown via a DRL will depend on where people are headed. Strangely, he does not make a similar observation for his own alternative scheme using Main and Danforth stations.
One problem with looking only at moderate term ridership on the DRL (and the associated diversion of traffic from Bloor-Yonge) is that this ignores the value of providing better service to current riders with room for future growth on the subway network that would otherwise run out of capacity again. Even with all the changes the TTC proposes for the YUS (and assuming that they are all even technically feasible), they will use up a great deal of new capacity simply through growth in demand on this major corridor. Another route to serve the medium distance demand in the city is essential, just as GO improvements are needed for the long haul trips.
A DRL would not interchange with GO at the CNE, but further east at Spadina/Bathurst where Metrolinx has already talked about a satellite station. The Weston lines are physically remote from GO CNE, and moreover are now in the Strachan Avenue underpass where it would be very tricky to retrofit a new station with some sort of walkway south to the CNE.
An interchange has also been proposed east of the Don River in the context of planned redevelopment on the Unilever site by Great Gulf. Indeed, Schabas claims that development is “unlikely” along the DRL. Possibly he does not know of Great Gulf’s plans, or of interest in redevelopment of a large block of property at Gerrard Square. If one presumes that the line stops at Danforth, the possibilities in Thorncliffe Park and at Don Mills & Eglinton don’t even come into the discussion.
For a critique of a plan that should include land use considerations, these are not trivial oversights and they suggest a rather superficial understanding of the situation “on the ground” in Toronto today.
The Scarborough Subway
As things stand today, the subway vs LRT debate is moot pending the outcome of municipal and provincial elections. A new Mayor could have different ideas, and more definitive subway cost estimates may yet derail the idea that Toronto can “afford” to build this line.
We conclude that the expenditure of almost $2 billion will bring little benefit to existing transit riders or to the region. Much greater benefits, at lower cost, could be achieved by modernizing the Scarborough RT, and integrating it with the GO Rail system, which can offer the required capacity and much faster journey times, not just to downtown Toronto but across the GTHA. [Page 50]
We may disagree about technology, but we appear to be in agreement on the general principle of better GO service and a “modernized” RT. I will leave other Scarborough-related issues to the Transit City discussions below.
TTC Labour [An aside]
Schabas takes more than a few misinformed potshots at labour costs on the TTC. With respect to the proposed yard on the north end of the Richmond Hill extension, he states:
Currently, TTC must run about 10 trains empty each morning and evening from Wilson to Finch, a distance of 30 km. These trains are apparently crewed with two employees, even though there are no passengers and no need to open or close the doors. [Page 45]
There are two fundamental points Schabas chose to ignore. Crews do not materialize out of thin air and staff who are based at Wilson Division would have to get from there to Finch Station early in the day. Their shifts would still start at Wilson even though they would not perform any on-train duties immediately. Second, and more to the point, the TTC has already announced that it plans to move to one-man crews as part of the conversion to ATO on the Yonge line.
In an earlier section (one dealing with EMUs on GO Transit), Schabas states:
Toronto’s subway cars are actually EMUs, although TTC does not vary the train length during the day. One reason is the high crew cost; TTC has a policy of operating with a two-person crew on every train; one drives, while the other closes the doors. [Footnote 22, Page 31]
Years ago, the TTC found that the labour effort needed to break up trains after the PM peak was quite substantial and, given the round-trip time on both routes, the process would extend well into the evening. This was possible in the days of the original Yonge subway when the route was short, and Davisville Station was designed to simplify cutting cars off of trains and running them out of service. This is not practical with the subway routes and yards as they now exist.
Running short trains between the peak periods is impractical because one would no sooner have shortened trains after the AM peak than one would be adding cars back in for the PM.
The size of the crew has nothing to do with whether cars could be cut off of the train. When the TTC did cut trains, the guard would be positioned at the cutoff point, and the shortened train would pull out of Davisville Station while a yard operator drove the remaining cars into the yard. If there had been no guard, the operation would have been identical (provided that the door controls would allow the single driver to open only the cars that would remain in service).
The point is moot because the TTC has moved to unit trains that cannot be shortened, a point Schabas does not mention. I cannot help feeling that Schabas’ gratuitous attacks on labour are motivated in part by an agenda for automated, unmanned operation.
Schabas notes that:
Most western European transit systems have adopted one-person operation, with the driver controlling the doors. [Page 50]
The statement implies that the TTC is not moving in this direction when in fact that is precisely its intent. Later, he states:
Converting the TTC subway to UTO could save about $200 million per year … [Page 51]
This is a gross overstatement of the available savings. The total TTC operating budget is about $1.6 billion, and subway operators do not account for 1/8 of the budget. The actual breakdown of operating staff is:
- 605 subway operators (12.2%) of which 335 are assigned to YUS/Sheppard and 270 to BD/SRT
- 575 streetcar operators (11.6%)
- 3,778 bus operators (76.2%)
- 4,958 total operators
- 10,700 total operating budget workforce
In other words, the subway operators account for only 1/8 of the workforce who drive vehicles, and this group accounts for slightly less than half of the total operating budget complement. [Source: TTC 2012 Operating Budget background papers]
On a pro-rata basis, this puts the cost of subway operators at about $100m annually (1/8 of 1/2 of $1.6b), and the TTC already plans to save possibly $30m of this by going to one-man crews on YUS.
There will be offsetting costs for roving customer service and/or security personnel, but this has not been factored into the comparison.
I have no problem with a philosophical discussion of manned versus unmanned trains, or of the alternative service arrangements that might follow such a change, but this should occur without inflated claims about the amount of money actually on the table.
Schabas drags out the “Sunshine List” of public servants who make more than $100k per year. [Full disclosure: I was on that list for a few years before retiring in 2009, and this happened purely from inflation in my salary.] He notes that 1,300 at TTC make more than $100k (oh, the horror!) but it turns out that there are only 191 operators and 21 station collectors. They worked a lot of overtime to get to that level, and it is unclear what relevance this has to the overall critique of the Metrolinx plan beyond a not-too-subtle slap at unionized labour.
Pension benefits, although considerable, do not scale with pay because overtime is above the pensionable income cap.
Updated April 25, 2014: The preceding statement was incorrect.
Many evening and weekend shifts will be longer than 8 hours and attract overtime, but this is a better use of staff than having separate employees, benefits and all, for the left over bits and pieces of work. In the 2008 Collective Agreement, the evening shift premium amounted to 3% over the top operator pay rate, slightly higher for junior staff. This is not exactly a “break the bank” premium for off-hours work.
Schabas suggests that the Sheppard line operate with two-car trains on weekends. Long ago, TTC stopped practice of breaking trainsets because of extra hostling costs. With two car trains there would be no backup in case of failure whereas with a pair, the failing units can be isolated as trailers. The cost of running the cars back and forth is small compared to the fixed costs of the subway infrastructure.
The claimed benefits of outsourcing should be placed in a local context especially in the UK where Schabas lives and works. There are times I wonder where folks across the pond are still fighting Thatcher’s battles against the trade unions, and exporting that philosophy to us here in the colonies.
This is an area completely outside of remit of a Big Move review, and brings an ideological bent to the discussion. One might claim that lower labour costs could lead to better service, but in Toronto (not unique in this regard) any cost reduction is an excuse to reduce subsidies, not to improve service.
In a strange remark, Schabas implies that the policy of setting service levels by GO for private operators is a bad thing. Why?
GO has outsourced the operation and maintenance of trains to Bombardier, but overall service features are still set by GO. By comparison, in Hong Kong, Stockholm, Shenzhen, and several French cities, companies compete to operate subway, light rail, and bus lines to which they bring commercial innovation as well as operating efficiency. [Page 91]
It is unclear whether Schabas intends that we simply turn “public transit” over to private operators and hope that the service they provide will be suitable. If nothing else, this could preclude Ministers from announcing new services because the private operators, with their “commercial innovation” and “operating efficiency”, might disagree.
There is a consistent line through this report that implies that if only we could somehow reduce labour costs, many of transit’s funding problems would vanish. If we must premise the improvement of transit on widespread changes in wages which, by the way, are arbitrated as an “essential service”, we risk losing sight of the real need for better transit everywhere as an essential part of society.
Is Schabas purporting that the only way to improve transit is to slash wages? That hardly constitutes a “review” of a regional plan.
7. Transit City LRT Schemes
The context for LRT and other technologies in Toronto needs to be understood to make sense of where we are today. I covered this in the introductory article, but some points are worth repeating.
Around 1980, TTC began construction of an LRT line from the subway terminal at Kennedy Station, running north and east to the new shopping mall at Scarborough Centre. A further extension to Malvern was planned.
But even then, labour costs were recognized as a serious impediment to reliable, frequent, and affordable service. The Province decided that the solution was Automated Light Rapid Transit (ALRT), and invested about $100 million developing this technology at the UTDC research centre in Kingston. The plan was to build it for Ontario cities, but also to export it around the world. In 1982, TTC agreed to use ALRT for the Scarborough RT. [Page 54]
The TTC did not “agree” to use ALRT (aka “Skytrain” or the Scarborough “RT”). They were forced to change the Scarborough line from LRT to what was then called ICTS (Intermediate Capacity Transit System) by provincial demand. Continued subsidy from Queen’s Park was dependent on adopting ICTS technology.
At the time, Ontario was marketing ICTS to Vancouver and needed to prove that the technology would be embraced in their own back yard. As part of this deal, Vancouver wanted a guarantee against failure of an unproven technology, and the then-minority PC government introduced a bill allowing Ontario to backstop the Skytrain project. When the Liberal opposition threatened to derail passage, this was deemed a “confidence” issue that would force an election if it were defeated. An out of control provincial project was desperate to prove its worth. [I will forgive readers if this sounds familiar.]
On another track, the proposed “GO ALRT” system was based on more conventional equipment akin to LRT and subway cars. The Ministry of Transportation wanted 120km/h capability for routes with widely-spaced stations, and the spec was carried over into the CLRV (the streetcars now serving Toronto). This made for a needlessly heavy, overbuilt city streetcar in place of the PCC derivative originally planned. (The PCC was a lightweight car developed specifically for on street, mixed traffic operation in the 1920s with technical improvements continuing through the 1940s.)
For many decades, “LRT” was not spoken of in Toronto as an option in part because of early misrepresentations that there was no “intermediate” technology (the “I” of ICTS) between buses and subways, and in part because subways were something people understood. They also meant big dollars for engineering and construction activity, something not lost on the NDP’s Rae government during the recession of the 1990s. The Eglinton and Sheppard subways were part of a plan inherited by the NDP from the Peterson government.
The “Mark I” RT cars used in Scarborough have many problems not the least of which is their age. They are noisy in part because they are lightly built (anything that might contribute to excess weight like sound barriers was dropped because of power limitations on the early version of the propulsion system). There was also a mistaken idea that wheels used only for support, not for traction and braking, would not wear thanks to a fundamental misunderstanding of how steel wheels behave on track. Wheels bounce wherever there is the slightest imperfection (including track joins) and this creates a self-perpetuating effect of rail corrugation that makes for a very noisy ride. The TTC trues the wheels and grinds the rail from time to time in residential areas, but only to a minimal degree necessary.
The terms “ALRT” and “RT” have combined in Scarborough residents’ minds to mean the same thing as “LRT” with the added problem that “LRT” is presented as “streetcars”, not as the true rapid transit it could easily be in the RT corridor.
That is the starting point for the Transit City debate, one that has been further poisoned by Rob Ford’s deliberate misrepresentation of LRT to anyone who will listen.
Having failed to find the funds to finish either the Scarborough RT or Sheppard subways, TTC was now reverting to a cheaper technology. It seems the idea was to spread rail more cheaply, and more widely, so it would get support from all parts of Toronto. [Page 55]
This was not a TTC idea, although they did background work on it. Transit City originated in David Miller’s office, and I had a hand in it.
The basic premise was that ridership was growing on major suburban corridors that would exceed the capability of buses especially with mixed traffic operations. (Congestion in the suburbs is actually worse than downtown, and the problem is growing). LRT would extend more reliable, higher capacity transit to less well-off neighbourhoods and improve their travel options.
In parallel, the city’s Official Plan included the concept of “Avenues” where redevelopment at medium density with street-focused buildings would transform old neighbourhoods be they declining industrial districts or first-generation suburbs with large parking lots surrounding malls. Schabas appears to echo the “latte sipping downtowner” moniker as the perceived proponent of this scheme and dismisses out of hand the thought that the suburbs could evolve into this form. This is a chicken-and-egg situation: if we continue to design transit and rebuild suburbs to a 1950s model, then we will not see “urban” forms and walkable communities.
There is high irony in that midrise growth (and not a little highrise too) is happening along the old streetcar network where there is capacity for transit growth, albeit constrained by competing demands for road space. People want to live “downtown” provided that they can afford it, something aided by frequent nearby transit service.
The Eglinton line is very expensive because there is no alternative to tunnel in central section. We are not going to build an elevated structure along a 5-lane street with old built-to-the-sidewalk form. This is not suburban Vancouver, and even that city put a good chunk of Skytrain underground with elevated and at grade construction mainly where there was available right-of-way and new development that could be integrated with station structures.
The Sheppard/Finch LRT proposal, as modified by Metrolinx, is expensive because it includes a segment from Don Mills to Finch Station compared to the original Transit City. This was a McGuinty contribution to the map that made little sense at the time, but was intended to placate those who wanted a “continuous route” across the top of Toronto.
Sheppard’s LRT was to be an eastern extension from Don Mills, and the initial phase of the Finch LRT would run only west from Keele (Finch West Station). If you “analyze” the wrong “solution”, you will predictably get the wrong “answer”.
There is no question that Metrolinx Benefit Case Analysis figures are suspect, but these were not part of original Transit City proposal. Indeed, if anything, Metrolinx weakened Transit City by lopping chunks off of the network and removing the very connections to neighbourhoods and destinations like the airport that were integral to the plan.
Unfortunately, Schabas appears to have an anti-LRT predilection at least in part as a justification for proposing an alternative technology he favours, ALRT.
Comparing LRT and ALRT costs in Toronto to Vancouver, Schabas contrasts the price of the Evergreen Skytrain extension with Scarborough proposals. What he misses is that that the original ALRT proposal included provision for a new maintenance facility, and this cost was not, initially, trimmed out of the LRT scheme even though it would share the carhouse at Conlins Road with the Sheppard LRT line. The Evergreen line does not require a new maintenance facility, and this difference inflates the relative per/km cost of Scarborough proposals.
Moreover, the cost of a facility depends on work it will do – a carhouse for storage, servicing and minor repairs vs a shops for major maintenance. In the original SRT proposal, the new facility was a standalone full-capability shops for a fleet of Mark II Skytrain cars.
Another source of extra cost on the Scarborough LRT plan was the demand by local residents that the connection track at Sheppard (that would see only a handful of movements each day) go underground, and that the proposed Malvern extension be buried to preserve what the community perceives as parkland. In a world where one can simply build whatever structures one wants, this would not happen, but one cannot attack “high” Toronto costs where decisions about structures are political responses to local concerns. Now we are building a subway to keep Scarborough happy, some argue, at a cost vastly greater than the LRT line that might have been.
Schabas claims [page 60] that neither the TTC nor Metrolinx has contacted Bombardier regarding a price for more Mark 1 cars. Unless he can cite specific proof, this is false as far as I know. The TTC also looked at buying and refurbishing cars from Vancouver, but they wanted equivalent replacement cost to buy more Mark IIs, and the limited remaining lifespan of the Mark I’s made this a totally unworkable proposal.
The Scarborough RT to subway connection at Kennedy is massive annoyance to travellers because:
- in the outbound direction, it is a roundabout path that (for escalator users) begins at the extreme east end of Kennedy Station, followed by a walk back a good distance to complete the journey up to the RT level, and
- in both directions, there is a good probability that one of the escalators will be out of service.
Any type of transfer between LRT and subway at Kennedy or at Don Mills on the Sheppard line is seen in the context of existing links rather than the greatly simplified arrangements proposed for both locations. Indeed Schabas himself argues for a modified location of an RT station at Kennedy to shorten the vertical distance between lines.
On Eglinton, Schabas cites Metrolinx cost estimates of:
- $100m/km for tunnels
- $150-200m each for stations
- $700m for everything else to a total of $4.9b
Schabas proposes that stations at Oakwood, Chaplin, Avenue Road and Mt. Pleasant could be dropped from the line. His unfamiliarity with local issues shows here:
- Oakwood has been kept in the plan by political fiat and in part because redevelopment was planned based on the station’s future existence.
- Chaplin was originally going to be an extraction location for the tunnel boring machines, but this has changed although the station remains in the plan.
- If Avenue Road and Chaplin were both dropped, there would be a 2km gap in stations in an established residential part of the city not unlike Danforth Avenue (but with tonier demographics).
- Mt. Pleasant already has higher density buildings on 3 of 4 quadrants (the one remaining is a school), and stops either side of Mt. Pleasant are well used on local bus routes.
- For a final irony, Schabas would keep Bayview which is almost all low-rise and construction is limited by underground conditions as a hapless builder found with the first and, as yet, only medium rise building on the corner.
He questions vehicle costs claiming that they might be more cheaply sourced elsewhere. In fact, the Siemens bid was 1.5 times the value of Bombardier’s and they would have incurred the extra cost of setting up a production facility somewhere in southern Ontario. In a way, Siemens was lucky that they did not get trapped into ramping up for an order whose delivery dates drift into the future and whose very existence is threatened by local and provincial politics.
Schabas forgets that Metrolinx took over part of the TTC’s LRV order which had originally included the Transit City cars. There was no point in duplicating a process that had already taken years to complete. He also forgets that at one time an untendered “extension” of the Scarborough RT to the airport (!) was proposed in place of the Transit City LRT line. It seems that “ALRT” often needs a helping hand to win out against other, more established technologies.
Schabas makes comparisons between Transit City and the Metrolinx goals:
The Transit City schemes will certainly improve local mobility, but will do little to further Metrolinx’s regional transport objectives. [Page 62]
That is precisely what Transit City was intended to do. The fault lies with Metrolinx in producing a network that includes both local and regional components, but with an expectation that regional goals take precedence.
Metrolinx did major damage to the LRT concept when they took control of the Transit City projects because of their “regional” focus. They love the idea of “Mobility Hubs” around major transit stations, but a local, linear development such as the city’s Avenues design is beyond their understanding or interest.
Because LRT will run along main streets like Sheppard and Finch, many passengers in neighbourhoods like Malvern will need to change to a local bus or walk some distance to complete their trip home. [Page 62]
How would a subway or ALRT line be any different? The “last mile” problem faced by transit systems depends on good local services and walkable neighbourhoods, and the problem is accentuated when there are fewer, widely-spaced stations serving a “regional” demand.
This type of statement is designed to make LRT look bad, but without fairly comparing modes. Unless we are going to rehouse all residents in high density buildings clustered at stations, local buses are common to any technology. Indeed, one Scarborough Councillor, in downplaying the importance of more LRT versus fewer subway stations, was proud to claim that everyone will take the bus to get to the rapid transit line anyhow.
The big issue along streets where LRT might be built is the effect on traffic, lane and turn arrangements, and the perception that “streetcars” will bring chaos. However, this is common to any surface mode that takes lanes including Bus Rapid Transit (BRT) and to the construction of elevated structures as well. Advocates of “els” love to show slender guideways (this sales pitch goes back decades including one memorable public meeting where the then Minister described them as “flimsy”), but the requirements for stations are rarely discussed except where they have been integrated in adjacent buildings or cross streets at 90 degrees (as in a junction between an old rail corridor and a street) with minimal effect on the street below.
Stations require at least twice the width of a guideway, not to mention vertical access by stair, escalator and elevator that must land on the sidewalk, not in the middle of a wide roadway. These are not trivial structures, and on narrower roads such as the central section of Eglinton, they would completely span the street and consume some roadspace just for their supports and access paths.
TTC’s plans make sense only if Eglinton Avenue is substantially redeveloped, at much higher densities, a prospect that is at best, distant. [Page 62]
Redevelopment is already planned for the Golden Mile (Victoria Park/Pharmacy) and other areas have or will see infill development.
If the LRT is replaced by something with stations every 2km, as Schabas proposes, this will be twice the existing arterial grid and will leave considerable amounts of property far from rapid transit.
An important part of Transit City was to stay on the surface with comparatively low construction costs as much as possible. The Eglinton line running from the airport to Kennedy Station would have been only one third underground, and the average cost/km would be correspondingly lower. A fundamental point about LRT is that extensions into less tightly-built areas should not cost a fortune to build and operate.
Schabas quotes a TTC cost estimate for subway to LRT conversion on Sheppard of $600m. [Page 67] This might seem a lot, but it could mean a through LRT service from eastern Scarborough to Yonge would actually cost less than the proposed Scarborough LRT upgrade (about $1.6b, although I suspect the Sheppard price is based on an earlier date). He does not pursue this idea in part, I suspect, because he has his own scheme for an ALRT line. (See the “Scarborough Wye” below.)
Again speaking of the Scarborough conversion and the preservation of ALRT:
But curves can be rebuilt, and all transit cars are built to order. It is not clear why the TTC believes the line needs to be rebuilt at all, when Vancouver is operating and even extending a line of similar age and technology, carrying many more passengers, without serious problems. [Page 68]
Vancouver does not have snow and ice (most of the time). In Toronto, the weather is a major incentive to change power collection and propulsion technology both of which are responsible for service shutdowns in the winter. Also, Skytrain was built to more generous specifications and could easily accommodate a new fleet. (As an historical note, the small tunnel at Ellesmere was downsized from original plans so that larger LRT cars would not fit at Queen’s Park’s orders to prevent a change of heart on the technology.)
One alternative would be for TTC to get Bombardier, or another manufacturer, to supply 40 replacement cars, for a price of about $4 million each or $160 million. For perhaps $100 million, TTC could remove the curve at Kennedy, and build a new station on a north-south alignment alongside the GO platform. If GO is eventually upgraded into a regional metro, most Scarborough RT passengers will want to transfer onto GO
anyway, because it offers a much faster service to downtown. We suggest that TTC is trying to improve the wrong interchange. [Page 69]
40 replacement cars would not provide the level of service needed on the new line given the known backlog in demand. At a unit cost of $4m, they are in striking distance of the cost of LFLRVs. As for the “wrong interchange”, I would argue that the real aim for GO should be to intercept riders further north and east rather than expecting them to first congregate at Kennedy where there is an attractive subway line available at no extra cost.
Schabas suggests that Metrolinx should consider ALRT and cites a number of cities where the technology is used. This list purports:
… [ALRT] has been applied very successfully in more than 20 cities around the world. [Page 70]
In his long list [footnote 89] he includes not just ALRT but metro (subway) and automated LRT (almost an oxymoron) systems in many cities, not a few of which are only airport shuttles. Readers may not know the details of every system, but they do know that ALRT or its cousins are not the preponderant technology in New York, London, Paris or Tokyo.
There is a high cost of segregating transit from other traffic, and Transit City took the view that this could not be justified for the projected demand in most cases. (As part of the ongoing debate about LRT, it has been fascinating to hear on one hand claims that LRT is “too much” for expected demands, but that if we build rail transit we should “build for the future” of runaway intensification.)
Schabas claims that a 5 minute saving in travel time would translate to 10 fewer trains. This is only valid if the trains operate on a 60 second headway. In any event, how much does 5 minutes represent out of a total trip? As he himself points out in talking about GO Transit, access times to stations are an important component of an overall journey.
Is it too late to change course, he asks?
After reviewing this report, Metrolinx staff told us that they had been reluctant to challenge the Transit City schemes because these projects already had strong political support when the organization was created in 2007. They argued against re-opening design issues relating to the Eglinton Crosstown, for which some contracts have been awarded, fearing the result might be complete cancellation of the project. We disagree.
Over the past half-century, Toronto has built three expensive rail lines (Spadina, Scarborough, Sheppard) that have attracted few new riders and had little impact on road congestion. After each scheme was completed, Toronto lost its appetite for new rail investment for a decade. If $5 billion is spent building an Eglinton LRT that is slow and attracts few new riders, the same thing could happen again. Metrolinx’s own BCA for the Eglinton Crosstown, which has finally been released, shows that a fully grade-separated line is likely to give a far better return on investment. Metrolinx needs to adhere to its own guiding principles, and insist that money is spent “where it matters most”.
Metrolinx managers were also cautious about considering the “Scarborough Wye” proposal (see Chapter 11 in this report), fearing that yet another proposal could “cloud the debate” over subway and LRT options. If the debate is “cloudy,” it is because policy makers have not been presented with coherent and consistent information, or attractive choices. We think the “Scarborough Wye” is worthy of investigation, with the potential to give greater benefits at lower costs than either the subway or LRT schemes. Metrolinx needs to consider all promising solutions to the need to extend transit into Scarborough. [Page 73]
As I said before, there is a fundamental difference of opinion on the purpose of the Transit City network between the City/TTC and Metrolinx. A model with fewer stops and faster end-to-end service serves one vision while one with a surface alignment and more stops serves another.
Transit City was highjacked by Queen’s Park on basis of 100% funding even though most routes clearly are not “regional” by their definition, especially when their ends were lopped off to save money. Queen’s Park concocted an accounting scheme to make provincial funding vanish from the books and to prepare ground for a possible transfer of transit construction and operations to a private partner. In other words, Transit City was warped by the political and economic agenda at Queen’s Park into something it was not intended to be.
It is quite easy from a false premise to say the network has to be changed because it does not suit an alternate agenda.
As for “three expensive rail lines”, this statement is very misleading:
- Spadina is carrying well for what it does. Demand cannot rise to full subway scale without overloading the University line which must have capacity for transfer riders from the BD subway. There is a possibility that as demand on Spadina grows with extension to Vaughan and full service to Downsview, space left over at St. George may not be sufficient. More generally, no outer end of a subway line can possibly carry full loads without swamping the existing inner part of the network. We see this already on the Yonge line south from Finch.
- Scarborough is carrying at maximum capacity for the fleet size, a size that was constrained by the very high price charged TTC for cars (>$2m/vehicle). TTC essentially made up for losses elsewhere in UTDC while Vancouver got cars much more cheaply.
- Sheppard was a political subway, a sop to Mel Lastman to support amalgamation. It ends at Don Mills because that’s all $1b would buy. Sheppard should have been an LRT line with an underground central section, but “Real cities don’t use streetcars” said Mel, an attitude that continues today.
The whole discussion of Transit City is replete with assumptions and ad hoc arguments that have as much to do with justifying an alternative plan for ALRT as they do with the original goals of the plan. Schabas undermines arguments that might have some merit with errors of fact and an insentivity to local conditions and history.
A Skewed Comparison: Eglinton LRT vs the Modified ALRT Scheme
I do not want to spend time delving into every cost comparison in this report because there are so many errors in the main body, but a review of the purported comparison of schemes for Eglinton illustrates the sort of problem that can arise when a huge report that almost nobody will read cover-to-cover shows up claiming to have “the solution” to our problems.
The information in question is in Section A9 starting on page 124. Readers whose eyes glaze over with this sort of thing should skip ahead, but bogus financial claims are the nub of more government cock-ups than anyone can count. It is important to get this stuff right.
The table begins with the basic data for the LRT lines including Eglinton and the capital costs of $4.9b for infrastructure and $336m for vehicles.
Incremental operating cost savings (explained in a note on page 125) are claimed to come from “about 60% of the Eglinton West and Eglinton East routes” for a saving of $570m NPV. With an NPV factor of 23 which converts annual to life cycle costs, this gives a purported annual saving of $24.8m. In fact other routes operate on Eglinton, and 34 Eglinton only went beyond the limits of the LRT line in peak periods (east of Kennedy) during 2012, the last year for which the TTC has published cost-by-route data. Even so, when all of this is corrected, the annual cost of services that the LRT would replace comes out in the same ballpark, about $26m, and so I won’t quibble.
The projection uses a daily ridership in the corridor of 176,000 with a bump of 25% by 2023, or 44,000 new riders. These riders are presumed to contribute new revenue to the system (although some may simply divert from other routes) and to generate offset benefits both for themselves (shorter trip times) and for motorists (reduced congestion). Only net new transit users will contribute these soft benefits, but the calculation presumes that they were all former auto users. These factors are important because collectively they make the imputed value of new riders critical to the overall calculation.
In the “modified crosstown” costing, we see that the total infrastructure cost is about the same (presumably because there are fewer and shorter stations as Schabas urges). Vehicles costs are the same for LRT and ALRT at $4.2m each, but there are more ALRT cars included. Note what this means: the ALRT would be more expensive than LRT if it were built to the same station layout, and only by eliminating stops in the name of “regional” services do the numbers come out roughly equal.
Life gets interesting with the imputed value of higher speed which gives a 25% uplift to the entire ridership, or an added 44,000 rides per day. What this ignores is that trips from the west end of the line to Eglinton West Station would be unaffected by proposed changes, and even trips coming to Eglinton and Yonge would only save a few minutes enroute. From the east, only riders who would start from Don Mills and points east will save significant travel time with the change to an elevated system, and some of this could be negated by longer access times to stations. There is no attempt to subdivide the line and assign marginal values to the benefits of each group of riders. Most certainly, they are not all riding from Kennedy to Mount Dennis.
In other words, the projected ridership and benefits for the modified line are artificially inflated while the capital cost is reduced by making virtue of fewer stops and longer access times. This has a cascading effect through the values assigned to more, shorter rides, and to the magnitude of imputed road user benefits.
The difference between the two schemes is almost entirely accounted for in soft costs and benefits, not in the capital costs and future operating costs. This makes the analysis very sensitive to ridership projections.
It could be argued that all things being equal on the hard costs, we should pursue the option with greater soft benefits. However, these are not fully worked out, especially the effect of removing stations from the line. Moreover, the embrace of ALRT technology precludes further surface extensions as conventional LRT, and that is a future cost that must also be considered.
8. 905 Region BRT and LRT Schemes
After trashing Transit City, Schabas spends precisely one and a half pages talking about the Mississauga and Durham LRT and BRT schemes. It is odd that he never turns his gaze northward to York Region and VIVA, but that may be enough of a fait accompli that it’s not worth his attention. Similarly, there is no discussion of the Hamilton LRT.
This shows even more clearly that his real aim is to demolish the credibility of Toronto’s LRT network.
The Hurontario LRT is 1/4 of the cost of Eglinton. This is no surprise considering it is entirely on the surface and does not have to deal with a constrained road corridor. It shows precisely the advantage of a surface LRT for holding down costs.
11. The Scarborough Wye
[Sections 11 is included here because of its relevance to the subway and LRT discussions above.]
Schabas repeats his claim about Vancouver’s Skytrain that “most is elevated” but does not give the context – old railway corridors, industrial areas, wide arterials and parking lots. Notably the line is underground where it would be obtrusive, and Vancouver had the good fortune of an existing, abandoned railway tunnel linking the north and south sides of downtown that was repurposed for Skytrain.
What he proposes for Scarborough is a “Y” shaped ALRT system linking Don Mills Station, Malvern and Kennedy Station with Scarborough Town Centre as a focal point. This would include conversion of the existing Sheppard Subway to ALRT.
The proposed route avoids on street construction on Sheppard by slipping south to the 401 corridor. This emphasizes the “regional” view of transit and destroys any idea that Sheppard East might redevelop along a rapid transit corridor. The route to Malvern presumes an elevated in an area where intrusion on community has already been rejected for LRT at grade, let alone on a structure.
This scheme could be described as a subway extension, as it would indeed be an extension of the Sheppard line and would offer the quality of service that is associated with a subway. Of course, if the Markham GO line is upgraded to offer a “metro” service, there may be no need to rebuild the Scarborough line south of Ellesmere. The Scarborough line could simply be joined into the Sheppard Subway, offering a direct “subway” from Sheppard-Yonge to Malvern via Scarborough Centre, with interchange to the Markham GO line at a new station east of Kennedy and north of Ellesmere Road. With travel times of about 30 minutes to Union Station, GO would offer much faster connections to downtown, and indeed through to emerging employment centres elsewhere in the GTHA, and at much lower cost than extending the subway. [Page 96]
Clearly the primary intent is to move people quickly to the central subway, not to stimulate local development. I agree that subway extension is not a good choice for this purpose versus GO, but that presumes that the main reason we would build anything in Scarborough is to move people onto the existing subway.
The main benefit cited is a shorter ride for people now served by buses for part of their journey (including the 190 Scarborough Rocket and bus feeders to the RT and subway). However, this should be balanced against an analysis of an LRT configuration that would provide greater coverage within Scarborough than an express ALRT line.
[This series will conclude with a fourth article dealing principally with integration of transit systems and funding.]
[This article was edited on April 25, 2014 to remove typos, make some minor stylistic changes and to note an error related to pension benefits in my original text.]