the canada line is brilliant, actually

Every transit enthusiast in Vancouver seems to think that the Canada Line is an underbuilt, over capacity, and poorly designed transit line. ¹ Unfortunately, I simply cannot agree with this sentiment. The Canada Line is a brilliantly designed, perfectly sized transit line that has built-in latent capacity sufficient for sustained growth beyond 2050. Creating an “overbuilt” Canada Line would not only have been politically and financially impossible but would have been fiscally irresponsible. The claim that the Canada line is “underbuilt” relies on incorrect assumptions about the actual design capacity of the Canada line, ignores the political climate of the time—which would have made an “overbuilt” line impossible—and ignores the most powerful argument for reducing present-day capital expenditure: the actual cost of a retrofit in present value terms.

Why make big train when small train does big train work?

Many people look at the 40-50m platforms and crowded trains of the Canada line today and assume that the line is operating near its maximum capacity, when this isn’t actually true. Every platform on the line is designed to be extended to 50m, which will accommodate the use of a 3-car train with 50% more capacity* than the existing rolling stock. While the Canada Line’s platforms feel short, a 3-car Canada Line train actually has slightly more capacity (501 people vs 498 people) than a 76m 6-car Mark I trainset, and slightly less than a 4-car Mark III trainset (532 people) or 4-car Mark II trainset (528 people). Headways are designed to be reduced to as low as 120 seconds per train, which allows the line to hit a theoretical maximum capacity of 15000 passengers per hour per direction (pphpd). ² This is the same as what Toronto’s Eglinton Crosstown achieves with 90m platforms. ³ The fact is, most people significantly underestimate the built-in expandability of the Canada line, and wrongly assume that the line is bursting at the seams when in reality it is only half awake.

If “underbuilt” means “ridership projections above design capacity,” then Canada Line simply doesn’t fit the bill. TransLink’s 2024 performance statistics show that peak hour ridership for the Canada line along its busiest segment reached 4600 pphpd in 2024. ⁴ According to TransLink’s Transport 2050 backgrounder—which already factors in new developments and transfer ridership—the Canada Line is projected to reach about 112% of the current capacity of 8600 pphpd during peak hours in 2050, ⁵ for a total ridership demand of about 9600 pphpd—36% less than the design capacity of 15000 pphpd. ⁶ Even assuming ridership continues to increase past 2050 at an average of 3% per year—an extremely high sustained ridership growth rate—the line will not reach its design capacity until at least 2065. A line that reaches its design capacity 55 years after opening cannot, under any common definition of the word, be considered underbuilt—it was simply built in a city that eventually outgrew it. Ultimately, the idea that the Canada Line is “underbuilt” falls apart once you realize the amount of latent capacity buffer that remains in the system, as shown by the actual design specs and ridership data of the line.

the real capacity ceiling

The capacity argument also ignores a key fact: Like the Expo Line, there are ways to increase capacity beyond what the original designers planned for without major redesigns or retrofits. While the maximum capacity of the Expo Line is commonly cited as 25000 pphpd, the original design capacity of the line was only 19400 pphpd ⁷ using 6-car Mark I trains and a minimum headway of 93 seconds. It is only due to the fact that we have new, wider rolling stock that we can actually claim 25000 pphpd is possible in practice. The most obvious capacity improvement the Canada line can make is to simply push down the headways. While the Canada line is designed for normal operation with 2-minute headways, it uses the same SelTrac S40 CBTC signalling system as the Expo line ⁸ and, per the schedule 6 amended and restated proposal extracts document, is capable of running trains with a minimum headway of just 75 seconds ⁹. Assuming some operational buffer, at 90-second headways, you actually beat the design capacity of the original Expo Line by moving 20000 pphpd. Combining this with a three-train YVR airport–Lansdowne (short turn)–Richmond Brighouse rotation, you can better serve commuters in Richmond and Vancouver during rush hour while avoiding the single-track segment bottleneck that critics cite as constraining the headways. Redesigning trains with more standing room and internal space, like what the Expo Line eventually did, can also squeeze out extra capacity.

The stations themselves may feel small, but fundamentally, they are sized to handle expected passenger flows when the line operates at 15000 pphpd—this is explicitly stated in the schedule 6 document ⁹. There is no evidence that the stations lack enough circulation capacity to hit this ridership target. Stations can also be incrementally upgraded as needed. Current projects include new escalators at busy stations to improve passenger flow and an upgraded entrance for Oakridge 41st, since the mall is undergoing a massive redevelopment. ^{10 11} Larger retrofits are also feasible in the long term, as seen by Vancouver’s $60 million expansion of Commercial Broadway and Toronto’s $137.5 million expansion at Union. ^{12 13} These expansion projects show that targeted station infrastructure upgrades are an effective and historically proven long-term plan to address capacity and passenger flow concerns at the station level when demand patterns change.

The numbers don’t lie: the Canada Line is operating far under its design capacity, and even farther under the theoretical limit of its infrastructure. It is not underbuilt; it’s just so efficient that it only needs 50m platforms to do what the Expo Line used to do with 80m.

politics and games

Proponents of the hypothetical overbuilt Canada line also ignore a key fact: The political realities of the time would have killed a line any more expensive or complex than what actually got built. In fact, TransLink’s board had already voted down the proposed Canada Line (then called the RAV Line) twice before reluctantly agreeing on it, provided the cost did not exceed $1.35 billion. The SNC-Lavalin proposal that was selected already exceeded the existing committed funding by $210 million ¹⁴. Bombardier’s RAVxpress bid, with 80m platforms, no cut and cover, and compatible technology, came in at $600 million more than the competing Serco/SNC-Lavalin bid, requiring $2 billion dollars of public contribution funding that simply did not exist ^{15 16}. Although no official figures exist, the extra cost of building 80m platforms and eliminating the single-tracked segments likely constituted the majority of the cost delta, likely $400-$600 million in 2004, almost certainly prompting an automatic rejection by TransLink’s board. The competing Bombardier proposal was rejected for this very reason—the cost premium for an “overbuilt” line was simply not justifiable.

The Canada line is brilliant because it exists, and it works. An overbuilt line would have sat as drawings at the back of a filing cabinet, never to materialize as concrete in the ground. The fact that the Canada Line was a resounding success in terms of ridership and finances directly motivated the region to invest in other major transit projects and double down on the automated light metro philosophy, seen in the Evergreen extension and the under-construction Broadway Subway. Critics of the Canada Line fail to realize that it’s much better to have a 19km Canada Line and a 10km Evergreen extension than an overbuilt 19km Canada Line and no Evergreen extension—or worst of all, no Canada Line at all.

The procurement of the Canada Line also hinged on a hard deadline: the 2010 Vancouver Winter Olympics. If the line was designed with complex interchanges, deep mining, or large stations like the ill-fated Eglinton Crosstown, there is a significant chance that the November 2009 opening target would have been missed and the region would have faced significant delays and cost overruns—factors which absolutely obliterate the realized benefit-cost ratio (BCR) of a transit line. The fact that the Canada Line was designed with pragmatism as its core principle allowed it be delivered on the agreed project budget and timeline ¹⁷—nearly unheard of for modern transit projects. What planners understood was that justifying delays for large platforms and longer trains—which would have sat half empty for the next 50 years—would have been politically and financially impossible for the line. The Canada line is brilliant because it is designed to meet the present it created, not for a future that never comes.

money talks

Perhaps the most compelling argument for the Canada line as it stands today is the time value of money. What critics of the Canada Line don’t understand is that money is not static; if we are trying to decide which option is economically superior: an extra $500 million in 2004 to “overbuild,” or a multi billion dollar retrofit in ~2060, the question is not which price tag is larger, but what the 2060 retrofit cost is actually worth in 2004 dollars. To establish some reasonable baselines, we can look at the recently completed $62 million Capstan station ¹⁸ and the underground stations on the Eglinton Crosstown—arguably with more complex excavation and geography than Vancouver—estimated at $150 million to $200 million apiece. ¹⁹ These prices are also in line with recent station expansion projects such as Commercial Broadway (adding a new 80m platform, pedestrian bridge, and concourse for $60 million) and Toronto’s Union Station ($137.5 million for a new 140m platform constructed underground in one of the busiest subway stations in Canada). ^{12 13} We will use similar prices to estimate the cost of retrofitting a station—$200 million for an underground station and $62 million for an elevated station. For the at-grade stations, we will assume the cost of a retrofit is $10 million. Since the Canada Line has 6 elevated, 8 underground stations, and 2 at-grade stations, a retrofit to add 90m platforms and 5-car trains would cost $1.992 billion today or $1.31 billion 2004 real dollars. Since construction price increases more than the Consumer Price Index (CPI), we will assume that the CPI averages around 2% and construction inflation outpaces general inflation by 1% (3% construction price index) between 2004 and 2060. Therefore, our retrofit price tag comes out to $2.29 billion 2004 real dollars in 2060. Assuming a 5% real discount rate—similar to the discount rates used in BCA cases in BC ²⁰—we can see that the massive $2.29 billion retrofit price tag is actually only a $149 million price tag in 2004 present-value terms. To account for varying discount rates, retrofit dates, and construction price indices, we perform a sensitivity analysis.

Table 1: 2050 Retrofit (high ridership case) Sensitivity Analysis (Values in Millions of 2004 CAD)

Table 2: 2060 Retrofit (base ridership case) Sensitivity Analysis (Values in Millions of 2004 CAD)

Table 3: 2070 Retrofit (low ridership case) Sensitivity Analysis (Values in Millions of 2004 CAD)

Only when you assume a very low discount rate, a very high construction inflation rate, and a significantly sooner than anticipated retrofit does the math support the view that overbuilding in 2004 was the economically superior choice. Intuitively, it seems right to assume that “building it right the first time” is cheaper than retrofitting later, but the math actually supports the idea that “right size now” and retrofits later is an economically justified strategy under the normal cost and discount rate assumptions.

When the time comes to finally retrofit the line to accommodate more passengers, the majority of the stations on the Canada Line are actually fairly easy to expand from a construction perspective, at least compared to many other metro systems that have done major retrofits. The elevated stations and at-grade stations simply require a bit more structural decking and concrete, with little to no guideway work. The underground stations have service equipment and emergency exits, but are all fairly shallow cut-and-cover boxes—a blessing of the line’s value engineering. Some of the extended platforms may not be perfectly level or straight, but for most passengers, this is not a major issue. For passengers with mobility needs, we can have dedicated boarding areas and platform screen doors to ensure safety. Yes, the Canada Line platforms will be difficult to extend when the time comes, but it is not impossible; under reasonable cost and discount rate assumptions, you can justify a very large budget for this capital works project before overbuilding the line in 2004 becomes the more economically prudent option.

We also have a powerful cost-saving lever at our disposal: selective door operation. By keeping some doors closed at less busy stations with shorter platforms, we get the capacity of a longer train with very minimal infrastructure cost. This is not just some hypothetical: the Docklands Light Railway uses selective door opening at specific stations ²¹ every day, without major issues. Doing this in Vancouver, particularly for less busy underground stations, would reduce the cost of a future retrofit by billions.

Retrofitting also has one major advantage compared to overbuilding—when you retrofit, you only pay for capacity you know you will use, but when you overbuild, you pay for capacity based on projections that may or may not materialize. Overbuilding from the start to pay for empty platform space, like the Sheppard subway, would have been plain fiscal irresponsibility, taking away from other important investments in the region. The Canada line is brilliant, specifically because it is not overbuilt. It is just right—it doesn’t pay for anything it doesn’t need in the short term, and defers major upgrades far into the future, where the present value of those retrofits is much smaller than the reduced construction costs today.

It’s how you use it

When you squeeze yourself onto a Canada Line train, it might feel crowded—that’s not a failure of design. Remember that the Canada line is designed to be expanded—the capacity is just locked behind a few knockout panels and a new fleet order. When the line finally reaches its actual capacity ceiling, smart software, selective door opening, and retrofits will be able to extend the line’s lifespan for less than the cost of overbuilding the line in 2004, in present value terms. The numbers don’t lie: the Canada line is a brilliantly engineered, perfectly sized, and remarkably efficient transit line whose designers knew that transit projects are not built in vacuums—they have real budgets, real design constraints, and real construction timelines. The Canada Line is brilliant because it exists; its resounding ridership success and value engineering directly led to the development and funding of other major transit projects in the region, fueling development and growth for decades to come.

* I’ve seen online discussions claim the Canada line middle car will be 10m long, however this claim is almost certainly incorrect. The city of Vancouver indicates ² that a third car can increase the line’s capacity by 50%, so a 10m half-size car is unreasonably small. Technical documentation from Serco/SNC-Lavalin ⁹ indicates that the third car is a normal-sized car with the same minimum of 37 seats and capacity of 167 passengers, assuming 4 pax/m^2 standing density. If I were to guess, commentators simply inferred that 50m platforms equate to a 50m train, when this is not supported by technical documentation. The Canada Line seems to be designed to let the train overhang the platform slightly, like the Mark V on the Expo Line currently does, meaning the total length of a 3-car train is likely around 55-60m.