The Ontario Auditor General released her 2018 annual report on December 5. Many topics were examined by the AG, but two related to Metrolinx bear examination by anyone concerned with the future of transit planning and management with more responsibility shifting to the provincial level.

• Metrolinx – GO Station Selection
• Metrolinx – LRT Construction and Infrastructure Planning

This article deals with the station selection process and the controversial recommendations for new stations at Kirby and at Lawrence East. I have written about this process and related issues before:

• Metrolinx Updates: New Stations Business Cases
• Metrolinx New Stations Report: The Details
• A Few Questions For Metrolinx
• A Few Questions For Metrolinx II

Updated: Links to Articles & Interviews

The Auditor General appeared on Metro Morning on December 6 speaking about, among other things, the cost of policy changes regarding LRT lines, and the evaluation of potential stations.

Former Minister of Transportation, Steven Del Duca, wrote an opinion piece in the Toronto Star claiming “I wasn’t meddling, I was building transit”. This is rich considering the effort Metrolinx went to in revising its evaluation of new stations.

Del Duca was notorious during his Ministry as using Metrolinx as an unending source of profile-building photo ops. He uses the Relief Line as an example of his intervention to get the project going despite early reluctance at the City and TTC level. This is a convenient rewriting of history and, in particular, of the huge difference between an RL ending at Danforth, and the one later evaluated by Metrolinx running north to Sheppard. The RL became popular and scored well once its extent and projected demand produced a significant dent on the Yonge line so that the Richmond Hill subway might be feasible.

A Few Thoughts About the Metrolinx Board

Although the Metrolinx Board meets in public from time to time, the legislation governing this body allows most issues to be debated and decided in private. There is no reason that this will change for the better. The chronologies set out by the Auditor General reveal situations where the Board was advised privately as issues evolved and met publicly only for the formality and patina of respectability conferred by their “approval” of matters already decided.

Throughout the station evaluation process, Metrolinx revised both published analysis and supporting documentation. This obscured the net economic costs estimated in the original business cases, making the results of the business-case analysis—both on Metrolinx’s website and in the published report to the Board—much less clear and transparent. [p. 315]

What is unclear is whether the Board actively participated in directions to staff that would lead to the rewriting of reports and recommendations, or merely chose to avert their eyes from the mechanics of political sausage making.

In any event, the process detailed by the AG throws into question everything that Metrolinx has done. Can anyone trust an organization whose professional opinion is so pliable, and which will defend recommendations, threadbare though they may be, so strongly? This is not just an issue for Metrolinx but for many public agencies involved in transportation planning notably the City of Toronto and the TTC.

To its credit, Metrolinx is developing a standard methodology for Business Case Analysis and will publish this in April 2019. However, the problem remains of just how well this will protect against hidden interference from politicians and their friends.

Metrolinx Business Cases

For many years, Metrolinx has used a methodology to evaluate projects that purports to establish the worth of a scheme, which could be negative, such as a new transit line or a significant change to existing facilities. The framework includes multiple factors examining projects from different points of view.

The Strategic Case looks at how a scheme works within the network and the wider public goals of supporting regional development. Factors include:

• Ridership projections
• Revenue and Operating Costs
• Population and employment served
• Travel time changes
• The reach of a new/revised service
• Effects on greenhouse gas emissions from trips shifted to transit

The Economic and Financial Cases review a proposal from two different monetary viewpoints.

• The Economic Case measures benefits such as auto operating cost savings, reduced emissions and air pollution, travel time savings, health benefits and reduction of accidents.
• The Financial Case looks at the cost and revenue estimates to produce a net operating cost as well as a “net financial impact” stating the total revenue over the study period minus the capital and operating costs.

The Deliverability and Operations Case concerns the implementation plan, procurement, operations, maintenance and risk management.

These factors overlap and the calculation machinery includes many assumptions such as future population and employment patterns, fare structures, operating and capital costs, trip diversion rates to transit, and the value of various benefits both to transit riders and society in general. Many of these are not published at a level of detail that would permit an outsider to understand how each factor behaves, and there is considerable leeway to affect the outcome by “twirling the dials” on factors readers cannot easily review.

A big issue with these analyses has been the question of how benefits are valued. For example, if a new transit service attracts people out of their cars, then this reduces the operating cost of those vehicles and produces environmental benefits, but it can also reduce travel time both for new riders and those on existing services. The values assigned to these and other benefits do not accrue to Metrolinx, but to the wider population. These savings, whether they be tangible (lower driving costs) or intangible (the value of time saved) are used to offset the hard costs of actually building and operating a service. While there may be an overall balance, the savings do not pay the bills which must rely on future revenue and subsidy.

A major contribution on the “benefit” side of the analysis is almost always the travel time savings for riders. For example, in the recent GO Expansion BCA, this is the overwhelming contribution to “value” in the analysis. Any factor that increases travel speed affects this measure, and in the case of stations “less is more” is the rule. Fewer stations make for faster trips and that translates to a higher modelled benefit. This has been at the heart of Metrolinx analyses for years and drives a pressure for wider station spacing even on urban lines like the Crosstown project. Adding a station to any route triggers a requirement to find an offset elsewhere such as a stimulus to riding that will drive up total rides even if they are all a bit slower.

A further problem with Metrolinx analyses is that the time period for comparison of costs and effects has grown to a 60-year horizon with the effect that far-distant benefits are shown as potentially offsetting short to medium term costs. This requires assumptions about the future of the transit system, the economy and regional development far beyond a period where anyone can reasonably know what will happen. In an effort to temper this, Metrolinx performs sensitivity analyses by changing factors to see what the effect would be. For example, if a more conservative set of assumptions goes into the model, what happens to the benefits, or does the proposal even fall into negative territory? How “successful” does Metrolinx and the region have to be in order to achieve its goals?

Needless to say, with such a timeframe, most of the readers, let alone authors, of these studies will be long gone before we could challenge their long term validity. The more subtle problem is that showing such long term benefits tends to paper over the fact that in the short to medium term, new facilities (particularly those requiring large capital investments) will not achieve anything near profitability and this shortfall must be financed. I will turn to this in more detail in a review of the Metrolinx GO Expansion BCA in a future article.

Kirby and Lawrence East Stations

The Auditor General does not mince words with her findings:

We found that the Minister of Transportation and the City of Toronto influenced Metrolinx’s decision-making process leading up to the selection of the two stations. As a consequence, Metrolinx inappropriately changed its recommendations on the Kirby and Lawrence East stations. Metrolinx’s initial business cases concluded that the stations’ costs and disadvantages significantly outweighed their benefits. Metrolinx overrode that conclusion and recommended its Board approve them because the Minister of Transportation and the City of Toronto had made it clear they wanted these stations. [p. 296]

The Auditor General reports that both of these stations originally fell into the “Not Recommended” category, and this is the position management originally took to the Minister of Transportation and to the Metrolinx Board. In the case of Kirby, the Minister wanted this station and issued a press release saying it would be included in Metrolinx plans even though management had recommended against this. Lawrence East was a different situation where the City of Toronto wanted this station as part of the SmartTrack scheme.

In both cases, the evaluation methodology was changed to shift the stations into a category that was recommended, albeit with very poor performance expected. Care had to be taken to craft changes to the methodology so that only these two would slip through without elevating other poor-performing stations to a marginally acceptable level.

The AG makes the point that a mechanism exists for a Ministerial directive that Metrolinx include stations that do not make the cut in its analyses, but that this would be done with the clear responsibility in the Minister’s hands. It should not be Metrolinx’s job to cook the books so that a favoured project looks better than it really is and provide cover for a political decision.

A written directive to Metrolinx from the Minister to add the Kirby and Lawrence East stations would have demonstrated greater transparency and accountability in that it would have signalled clear ownership of the decision. The public would have benefited from knowing that a government policy decision was overriding the results of Metrolinx’s business-case analysis. Instead, the Ministry of Transportation went so far as to issue news releases announcing the Kirby and Lawrence East stations before the Board had even met to make its final recommendations. [p. 296/8]

This is a fine idea in the abstract, but the reality of government is that politicians do not want to appear to “waste money”. A staff report saying “your pet station isn’t worth building” is hardly the sort of ringing endorsement Ministers want to have. This is not just a provincial problem, but a general issue across governments where a negative evaluation of any project will never see the light of day, barring a “brown envelope” leak to the media. Positive staff reports are routinely used to give projects an aura they do not deserve, a stamp of respectability from “professionals” who claim to have carefully reviewed and endorsed them.

At the City of Toronto, the experience of a former TTC Chief General Manager who opined that the Scarborough Subway Extension was not a good idea led to a new verb in civic discourse: to be “Webstered”. He retired somewhat sooner than expected.

The process is quite clear:

Metrolinx’s response to the influence was to make the Kirby and Lawrence East evaluation results look better. Metrolinx’s 2016 original business-case analyses of the Kirby and Lawrence East stations noted that both stations were expected to result in a net loss of GO ridership, a net increase in vehicle use (driving) in the region and an overall decrease in fare revenue. The business-case analyses did note positively that the stations aligned with municipal land-use policy, which slightly improved their evaluation results, but they still concluded overall that these stations were “low-performing” and “should not be considered further during the next ten years.” However, the Metrolinx Board Chair and Chief Executive Officer guided the process whereby the Metrolinx Board ultimately supported the decision to add these two stations.

…

Metrolinx removed Kirby and Lawrence East from the original list of “not recommended” stations and put them into a new category it created of “low” performing stations. It put the remaining “not recommended” stations into another new category it created of “very-low” performing stations. These new categories were used in Metrolinx’s June 28, 2016, report to the Board, which recommended building all but the “very-low” performing stations. In other words, Metrolinx made the Kirby and Lawrence East stations appear to have better evaluation results than the “very-low” performing stations to ensure the Board would approve building them. [p. 298]

When the AG talks about the CEO’s role, it should be clear that this process includes the term of Phil Verster who became CEO on August 24, 2017. In a conversation at a recent Metrolinx Town Hall, Verster claimed that the Kirby Station issue was “before my time”. This is true only to the extent that Ministerial interest in Kirby Station dates from mid-2016 (see chronology on p. 297), but the work to massage the recommendations and defend them to the media and public extended into 2018.

The publicly available information included in the June 2016 staff report to the Board of Directors to justify the approval did not highlight important details, especially that Metrolinx planning staff believed the Kirby and Lawrence East GO stations should not be considered for the next 10 years because of the significant delays and potential ridership loss they were expected to cause. Metrolinx’s updated analysis of the new stations, published in February 2018, presented a best-case scenario that assumed future changes to the GO system that, to varying degrees, are not certain to be fully implemented as planned when the stations are completed. The reanalysis also evaluated the stations using assumptions (such as auto-operating cost savings; growth in the value of time) that are not in line with Metrolinx’s current practices for evaluations of this kind. [p. 299]

One problem with Metrolinx’s multiple factor evaluations has been that the relative importance of each factor is unclear.

Metrolinx appropriately gathered comprehensive information for selecting new GO stations; however, it did not have a rigorous process for weighing all costs and benefits against established criteria. The information Metrolinx gathered on the Kirby and Lawrence East stations from January to June 2016 showed that the costs from an economic perspective significantly outweighed the benefits. Despite this, Metrolinx recommended the Kirby and Lawrence East GO stations in June 2016, on the basis of undefined “strategic considerations.” With such a vague process for selecting stations, any decision can be justified. [p. 305]

…

For the 2031 forecast year, the Kirby GO station was expected to result in:

• ridership loss of over 57,000 trips in that year;
• additional car travel of almost 40,000 kilometres per day (for commuters who switch from GO transit to driving); and
• an annual loss of over $900,000 in fare revenue.

The analysis estimated that these forecasted results would translate into a net economic cost to the GTHA of $478 million over 60 years. [p. 306]

…

The one strategic criterion the station did not meet was to improve transit service and increase ridership. As the analysis indicated, the area around Kirby GO station is not currently serviced by frequent local transit and is not close to key destinations, and travel time delays would translate to overall ridership loss. [p. 307]

An argument can well be made that Metrolinx overvalues the effect of additional station stops on ridership. The effect is probably a non-linear one where small changes will not discourage riders, but larger changes would. There is a general problem with the evolution of GO Transit in that its role is changing to serve more close-to-downtown locations as this will affect travel time for long-distance riders. Metrolinx plans to offset this with a combination of express trains and electrification to reduce the penalty of station stops. However, the original forecast made no assumptions about these possibilities and showed a negative value leading to a “not recommended” status for Kirby.

The corruption of the process began in June 2016.

The Minister of Transportation was the MPP representing the Vaughan riding, where a Kirby station would be located. On June 9, 2016, the Metrolinx CEO briefed him in person on the station selection status. The Metrolinx CEO let the Minister know that neither Kirby nor Highway 7–Concord (another station in the City of Vaughan) were included as recommended stations. The Metrolinx CEO stated in an email later that day to the Metrolinx Board Chair that he interpreted the Minister to be “disappointed” by the news. The Metrolinx CEO further informed the Board Chair that he was discussing an “alternative analysis” with Metrolinx’s Chief Planning Officer. [p. 308]

The first attempt to improve the numbers involved the use of express trains to ensure that riders from beyond Kirby Station would not be affected by its addition to the line. This did not solve the problem, as the CEO noted in an email.

Unfortunately, while [express train service] did “improve” the business case, both stations still perform relatively poorly. Based on this, staff would suggest that both stations be put into the “future consideration” category. I have the impression this will be looked at unfavourably at this point. I am going to think overnight if I have any other ideas. If we cannot develop a technical rationale, we may receive some direction on one or both of these. [p. 309]

Subsequently, events at the political level dictated Metrolinx’s actions, including those of its supposedly independent Board.

On June 15, 2016, the Metrolinx Board held a special meeting before a scheduled public Board meeting scheduled for June 28. The Metrolinx Board Chair explained in an email to other Board members that the purpose for the meeting was as follows:

Before our June 28 public board meeting, the Minister and Mayor Tory want to make an announcement about the Smart Track stations Mayor Tory will be recommending to Council. They want this to be a positive announcement reflecting City-Province-Mx [Metrolinx] cooperation. We did not want the Minister doing so without the input of the board in advance. To permit the joint announcement and preserve confidentiality, we agreed to this special meeting. We will then revisit the same issues in public session on June 28 but by then, it would be too late to do other than approve the staff report. Thus the real substantive meeting is this one on Wednesday [June 15].

…

On June 16, 2016, the Ministry of Transportation asked Metrolinx to review draft news releases announcing new stations. Four of the news releases announced stations that Metrolinx was planning not to recommend: Kirby and Lawrence East, as well as Highway 7–Concord and Park Lawn. [p. 309]

Staff got to work figuring out how to justify new stations. The first change was to move the dividing line between “recommended” and “contingent” stations from an economic cost of $250 million to $300 million. This allowed the Don Yard project to shift into the “contingent” group. However, this tactic would not work forever because the line would have to shift to above Kirby’s cost of $478 million, and this would bring other unwanted stations into the “contingent” category.

An internal review document of the business cases stated that the cut-off point for station selection seemed “to be set arbitrarily” and some “valid basis” for their inclusion needed to be provided. [p. 310]

To get around that problem, Metrolinx split the “not recommended” group into “low-ranking” and “very-low-ranking”.

Metrolinx did not post the [June 2016] Summary Report on its website until September 2017. When it did, it posted an edited version of the Summary Report provided by the consultants. These edits included changing the consultants’ group name of “Recommended” stations to “Best Performing,” and “Not Recommended” to “Low Performing.” Metrolinx’s renaming of the groups and removal of the word “recommended” made the results of the consultants’ analysis less clear to the reader and obscured the negative evaluation of the Kirby and Lawrence East stations arrived at by the consultants. [p. 316]

The role of political influence appears with the delicate term “sensitivities”:

Kirby and Lawrence East were the only two stations in the “low-ranking” group. Metrolinx defined low-ranking stations as “sites with poor economic performance but advantaged by strategic factors or sensitivities.” [p. 310]

Those “sensitivities” no doubt included the political aims of the Minister of Transportation and the Mayor of Toronto.

In March 2018, Metrolinx published its Draft Business Case Guidance, which states that business cases are only one of five inputs Metrolinx considers in decision-making. […] Metrolinx considers public engagement, policies and other investments, emergent trends and conditions, and capacity to deliver in addition to business cases. Based on our review of the process which led to the approval of Kirby and Lawrence East stations, a sixth input—stakeholder influence—was also an important input in Metrolinx’s decision-making.

Repeatedly adding further “strategic considerations” to the decision-making process makes it possible to justify any decision. Similarly, putting so much priority on these vague strategic considerations—and less weight on net economic costs—makes the decision-making process seem arbitrary. This is especially concerning because it resulted in Metrolinx choosing just those two stations that the Minister and the City of Toronto influenced it to choose.

Metrolinx’s Board Chair recognized this in a June 13, 2016, email to other Board members. At this point, Metrolinx was expecting to recommend just the 10 stations and not Kirby or Lawrence East.

The Chair wrote:

[T]here will be disappointed local communities both in Toronto and across the GTHA which will be very disappointed not to have achieved a station. The Minister will be bearing the political burden of explaining these outcomes which is why staff have worked so hard to be principled and evidence-based in reaching their conclusions. Absent that, our conclusions could be seen as arbitrary and essentially political which could open a Pandora’s box of new demands across the region. [pp. 310-311]

This is precisely the problem with the naïve idea that a government organization can maintain a veneer of professionalism under political pressure. When the “political burden” of justifying a decision that goes against professional recommendations lands on a politician, they will quickly find a way out. If the organization tasked with providing advice does not or will not say what is expected, there are always willing consultants whose opinions can be procured to support any position.

A year later, Metrolinx was still brewing up a new analysis to support political demands, and incorporated substantial changes including a blatantly distorted value in the economic analysis. These are summarized in three figures from the AG’s report.

“Fare integration” has been the holy grail of much planning at Metrolinx over the years both as a political imperative (reducing cross-border penalties) and as a way to boost ridership. The fundamental political problem has always been that either subsidies must rise to offset the foregone revenue, or somebody’s fares must rise so that other’s fares will fall. The latter was a zero-sum exercise that hobbled Metrolinx planning until a point where the need for some added subsidy was recognized. This was announced in the short-lived budget of the Wynne government, but there is no reason to think it will survive into the Ford era. Even the proposed “integration” was only for 50% off the full adult fare with less for seniors and nothing for passholders who constitute the largest block of TTC riders.

Cheaper fares attract more riders, and any part of the cost/benefit analysis will be affected through related metrics including time savings, reduced auto use, and so on. It is not clear whether the financial impact includes the lost revenue through reduced fares which would only be partly offset by more people riding.

Express service reappeared as part of Metrolinx’ plans in the February 2018 report. Previously it had been rejected as part of service designs for most corridors because of track limitations (the ability of express trains to bypass locals) and the problem that “local” stations would have infrequent service. This bumps into the service levels and ridership claimed for SmartTrack where a train every 15 minutes will not provide “rapid transit” service within Toronto thanks to longer transfer times. There is some debate about whether a mixed express and local operation is possible on all of the SmartTrack corridor because of right-of-way limitations for passing tracks. The service design still shown by Metrolinx for the Stouffville corridor has a mix of express and local trains with 20 and 15 minute peak headways respectively. Notable by their absence from the map are the SmartTrack stations including Lawrence East.

Express service does not currently exist on the Barrie and Stouffville lines. When Metrolinx looked at implementing it on the Stouffville line in 2016, it concluded that significant infrastructure costs, major property acquisition requirements and unacceptable community impacts constituted “fatal flaws” to its implementation. Metrolinx told us that it has since focused on how to reduce the significant infrastructure costs of express service for the Barrie and Stouffville lines, although its February 2018 updated station analysis does not include any information on this planning work. Metrolinx informed us it is planning to require the contractor it procures for the station work to achieve express service, and it is exploring options such as constructing short “passing tracks” to enable express trains to bypass non-express trains. Nevertheless, an achievable and sufficiently cost-effective express-service solution has not yet been finalized. [p. 314]

It is a mystery how the contractor for the new stations construction could possibly be responsible for solving the infrastructure problems of express operation. GO is already in the process of expanding trackage in the Barrie and Stouffville corridors, and anything beyond what is now underway would be an add-on that might not be compatible with what is being built. Moreover, the ability to use passing tracks is closely tied to service frequencies, stop spacing and schedules so that express trains overtake locals at precisely the correct time to pass them without being delayed. As local stops are added to a route, this becomes more difficult until a point where a dedicated express track is essential.

The Benefits Case Analysis for the GO Expansion program shows considerably better service on the Stouffville corridor with trains inbound from Lincolnville every 20 minutes running express south of Unionville overlaid on a local service every 7.5 minutes from Unionville south. It is unclear how this could be operated, especially with the additional SmartTrack stations (also missing from the Expansion BCA), without a dedicated express track.

One cannot help thinking that Metrolinx continues to engage in wishful thinking to achieve political objectives.

Level boarding would eliminate the step up to GO train entrances from the platform and speed loading. This would save on stop service time, but it requires changes to all platforms with potential operational impacts especially where corridors are shared with freight trains, not to mention for express GO trips that would pass the platforms at speed.

Metrolinx’s 2018 business case for level boarding found that it poses many challenges, such as modifications to existing trains and stations, and will take many years. [p. 314]

The business case document for level boarding might well have been available to the AG, but it is not posted on the Metrolinx site (notably it is not on the benefits case page).

In estimating savings to motorists who would no longer drive to work, Metrolinx used the fully allocated cost of car ownership rather than the marginal cost of operation. This presumes that a commuter could reduce their car ownership, but that is only possible if they have reliable, convenient “last mile” access to transit service. A parking garage at Kirby Station does not qualify, and the need for a car, if only for a park-and-ride trip, would remain.

According to Metrolinx’s March 2018 Draft Business Case Guidance document, the $0.66/km rate is no longer considered appropriate when there is no evidence that new GO riders will completely give up their vehicles. Although Metrolinx is undertaking further research in this area, currently the extent to which transit users give up their cars as a result of a new transit investment is unclear. [p. 314]

The value of time saved has always been a contentious issue in Metrolinx analyses because it forms one of the largest paybacks against the investment in infrastructure and better service. By escalating the value of time, the analysis offset the discounting of future costs and revenues so that “time saved” in the future was worth more proportionately than it would otherwise be. This is particularly important with a 60-year time horizon where, normally, future costs and savings would have only a small effect on the net present value.

The same December 2014 memo from the Ministry of Transportation cited in Figure 12 stated that Metrolinx should use a 0% value-of-time growth rate because a growth rate of 1.6% could have a “significant impact on the [economic value] of each project and a potentially significant impact on the ranking or prioritization of a group of projects.” The memo also noted that organizations in other jurisdictions, including Transport Canada and the U.S. Transportation Research Board, do not assume time grows in value when they assess the economics of transportation projects. [p. 314]

December 2014 is well before any of the work under discussion here had begun.

Collectively these changes made a big difference in the “performance” of new station proposals, but the problems do not stop here.

What the report lacked was “sensitivity analyses,” which would have presented a range of estimates about the economic benefits of the stations if, for example, any of the initiatives were not implemented or were implemented differently than assumed under the best-case scenario. Metrolinx did undertake such sensitivity analyses internally, assessing how the estimated benefits of each station changed with the addition or removal of each initiative. However, it did not include a range of possible benefits in the report published for stakeholders and the public.

Similarly, the 2018 Reanalysis Report did not include sensitivity analyses for different assumptions about vehicle-operating costs and the value of time, presenting only one scenario, which maximized the stations’ economic benefits. [p. 317]

The Auditor General’s report is unquestionably clear that there was an abuse of process at Metrolinx right into 2018 to convert a negative recommendation to a positive one, and that this was done at least with the knowledge if not the active participation of the Board.

Metrolinx plans to improve its processes, but this comes just at a time when political meddling will grow substantially with a government that knows what it wants to do and will not brook criticism.