This is a free article from The Transit Brief, a regular newsletter and blog about transportation infrastructure, cities and more If you enjoy the content — consider liking, if you would like some exclusive posts and want to help support me, please consider a paid
Cost comparisons of transit projects are useful for understanding the differences in prices paid for the same technology — fully accessible metro — in different jurisdictions. However, these comparisons have a major gap, they do not consider the hard value of differing capacity, which can vary by a multiple of over four times in some project comparisons. They also fail to incorporate the soft value of civic pride from grand structures, reduced externalities by burying more of a project, and other amenities which money can buy. In this post, we discuss this problem, and consider some possible solutions, including a refined cost-per-kilometre metric which includes transport capacity. This metric shows that Anglosphere metro projects like Toronto’s Ontario line — frequently cited as costing 10 times more than best practice in other cities may cost closer to 20 times more when considering actual carrying capacity — the whole point of a mass transit project. MotivationThanks to researchers and advocates like Eric Goldwyn, Elif Ensari, Alon Levy, Marco Chitti, Jonathan English, Reece Martin, Jedwin Mok, and the Transit Costs Project (TCP), there has been increasingly strong visibility on the extremely high unit costs of constructing a transit project in the English-speaking world. The TCP was not the first to raise the issue: publicly doing unit costs comparisons for transit projects was historically common from the Subway to LRT wars of Toronto to the Hong Kong XRL debates of the 2010s. The early activist attempts of normalizing project costs to unit length (per km) are laudable, but the arguments arising from these exercises were unfocused. As a result, the applicability of the cost implications from these early attempts had limited impact and salience in the overall discussion. It didn’t help that the broader political discourse surrounding the debates of both respective comparisons in Toronto and Hong Kong was extremely toxic and drowned out any opportunities for refinement for the more technically-oriented transit project unit costs debate. However, a larger reason these initial comparisons failed to gain traction is that directly comparing unit costs between constructing different modes of transport or alignment types is technically dubious. It is more of an exercise in performative austerity — read: “how little money we can get away from spending regardless of what gets built.” Comparing different LRT, metro, and high-speed rail projects with various different alignments strictly on per kilometre cost offers a shallow contribution to the discourse. In contrast to older, half-hearted failed attempts, the Transit Costs Project brought structure to the conversation by focusing unit comparisons on only fully grade-separated mass transit projects. The methodology is sound:
The GapWhen looking into the detailed analysis done in the case studies of the TCP, costs are commonly prefaced in the context of scope. For example, in the Boston Green Line Extension case study, TCP notes that the project is an LRT running at-grade in an existing right-of-way, so the conversations on costs should take this into account. A hypothetical “sane non-Anglosphere” cost of the Boston Green Line Extension project should be much lower than, say, an underground city centre metro project in say, Paris. While the TCP is great, there has been misuse of its global project dataset in the transit advocacy zeitgeist that developed around it. Similar to what happened in Toronto and Hong Kong in the 2010s, some fairly sloppy direct unit cost comparisons between the various different metro projects around the world have been made using the data. One cannot simply just take PPP-adjusted per-kilometre cost of one project in one part of the world and compare it to another assuming projects are comparable in scale and scope. Costs for an elevated light metro project with an underground heavy rail project should not be directly compared without any caveats that note this context. The TCP’s methodology used the dataset as a broad global comparison and successfully found a statistically significant order of magnitude difference between the Anglosphere and the rest of the world that simply cannot be explained by scope of construction or labour costs. However, a comparison between two specific projects requires a conversation with much more precision than per kilometre unit costs. There remains a lot of nuance that can be brought into the conversation of transit costs when delving deeper into a direct project to project comparison. We can expand on the groundbreaking work of the TCP, to account for value created from scale and scope. My own consulting work has noted that the value enabled by scale (more transport capacity) and ambition (which increases complexity) matters a lot for proponents of a transit project, not generally true when discussing whether we can get the unit costs down by 10%. A metro line running underground with direct alignments, visually more palatable to the public, which has better travel times and/or roomy, efficient underground interchange stations can bring pride and notoriety to a city or country, and enhance quality-of-life, not to mention functionality. These things are structurally more expensive. It is important to ask if the same project built elevated on a slow snaking alignment, with long interchanges to other lines and barebones stations offer the same perceived value even if the unit kilometre cost of the latter design is lower than the former. An incrementally more expensive project should bring equal or more proportional incremental value to the society it serves. It also helps if the increase in cost from more features or an improved design is modest enough to make it easy to justify. If you can build grand, high-capacity designs at a reasonable cost, it would be easier to “defend” the externalized return in benefits from such “frivolous” grandeur form skeptical opponents (at least enough to politically tip the scales in your favour). The internet “hype” from Shenzhen’s grand Gangxia North metro station? Priceless. Yet this famous massive interchange hall is the transfer station between three extremely high-capacity metro lines (Line 10, 11, and 14) but built with fairly average unit construction costs globally. One part of the equation for value then is a sort of “soft” value, that consists of architectural design, passenger comfort, civic pride, and externality minimization. Differences in ValueIt is important to understand and compare the “value” (and its perceived value to the public) a transit capital expansion project’s design can bring into the construction cost conversation. Different projects can perform different roles and should be designed and evaluated as such, meaning a strict unit cost comparison is not necessarily appropriate. Intuitively, this makes sense: Toronto built median suburban tramways in the middle of suburban arterials (colloquially called “stroads”) for the same PPP inflation-adjusted cost as Guangzhou builds some of highest capacity city centre underground metro lines on the planet. The other part of the equation for value is hard value, that consists of raw long-term people moving capacity. The unit PPP-adjusted per kilometre costs of both projects above are roughly the same, but which project profile is providing more value to the respective societies here? A slow, low-capacity suburban tram or a city centre ring line running very high capacity trains?
-Jedwin Mok in The English-world is TERRIBLE at building transit. This is why. Thus, we need to shift the transit construction cost conversation from raw unit costs to one centred on the project’s value-add to society. Most constituents are not mad at the high costs of transportation infrastructure projects; they are frustrated at the perceived limited value and scope these transit projects bring with their sky high price tags. If you are going to pay gold-plated prices, the project should look gold-plated. A recent comment on The Transit Brief suggests that the cost literally plating the rails of many of these projects with gold would be significantly less than those we have incurred from cost disease. While it is hard to quantify “value”, again, what is the value of “hype” from Gangxia North Station? Does it even count? One way we can quantify value is to tie it to capacity. In an ideal world, capacity is the base raison d’etre for all rail transit projects. Rail Transit relies on “massification” — big rail infrastructure is only economically viable when high capacities are required. This is in contrast to dispersed modes of transport like e-scooters or personal cars which are flexible but inefficient overall and which will lead to problems if a lot of people want to go somewhere in the same general direction at the same time. Massive economies of scale are required to offset the high upfront investment costs a rail transit project incurs. In cities, space is a premium, thus road transport can never fully compete with rail transit. Roads may be cheaper to build and flexible to operate, but from a capacity per capital dollar spent perspective, rail transit is the clear winner. Rail transit solves the urban transport geometry problem and as such capacity is the heart of rail transit’s value proposition, something which should be considered when talking about costs. Don’t miss a single post. Become a subscriber today for the latest on public transport in Canada, and beyond. CapacityI live in East Asia, a region of the world that has successful (and at many times even profitable) rail transit. Metro lines in the region routinely need to convey well over 50,000 people per hour per direction. This sectional demand is not currently found anywhere outside of modern East Asia except maybe rail systems in Delhi or São Paulo, which means projects in East Asia ideally should have large trains and stations to handle high passenger flows generally not found in other places. The increase in scope requirements from such extreme capacity requirements are not reflected by simply cross checking at raw unit costs as popularized with the TCP. One of the biggest value propositions of rail transit is capacity: comparisons between projects should be normalized for how much rated capacity you are getting for those hard-earned tax dollars. Another aspect of normalizing for capacity is that projects are not punished for future proofing, longer platforms for bigger trains in the future is a benefit. Smaller cities in the world don’t need massive East Asian megacity capacity and ideally would build light metro lines. I expect such light metro lines to have capacity requirements that lead to structurally smaller designs, which also structurally lead to being cheaper to build overall compared to a typical East Asian metro project. This is important as these cities often do not have the fiscal capacity to build a transit network with capacity suitable for an East Asian megacity and should not need to. When building a transit line, 200 million per km is a very different financial calculus when contrasting between a 1 million population provincial city like Mersin, Türkiye or a growing 25 million+ population globalized East Asian megacity like Shanghai with financial firepower and national clout. A metro project in the former is a one-off “once in a lifetime” city altering project that consumes a small minority of the city budget for years to come, while the latter would just be another metro project in a line of many and is a line item in a spreadsheet. Larger capacity trains ultimately lead to bigger stations and higher costs. So naturally, projects with high capacity designs will have higher unit costs. Accounting capacity into the construction cost conversation allows for better perspective that also scales with the context of the city it serves. What might this framework look like? I can give you an example. The TCP has identified Spain, Italy, Turkey and South Korea as global high performers in metro construction unit cost efficiency. However, a closer inspection will reveal a schism. Despite the similar low costs, (mostly) Asian systems have examples of projects that are much higher capacity than those in Europe. The differences can be explained by the social-cultural-urban geography of the European and Asian countries. The Asian low cost countries have megacities like Istanbul (famously half in Asia and half in Europe) and Seoul while the European low cost ones do not. As an example, compare Milan Metro M5 and Madrid Line 11 with Istanbul Metro M5/M10 and Seoul Metropolitan Railway Shinbundang Line. Below are the relevant project parameters: It is important to note that I have made a bunch of assumptions on maximum achievable train frequency not current scheduled frequency. These assumptions are based on real world throughputs have been achieved with similar sized trains with similar signalling technology (eg Hitachi Italy Automated Metro). All lines are crosstown fully underground metro lines with city center and suburban segments. However, the European cities of Milan and Madrid support much less capacious trains and by extension are much lower capacity than examples in the Asian cities of Istanbul and Seoul. Showing that unit cost is not everything and that there is more underneath the surface in this game of project appraisal. One way to normalize capacity is to calculate the maximum amount of train floor area that can be moved on the line. Using the following equation: A New MetricWe can then look at the cost of a line per kilometre per sectional capacity. This is the cost of a line per kilometre, divided by the lines sectional capacity. We calculate a baseline sectional capacity by taking Toronto’s Ontario line’s approximate numbers - 40 trains per hour, 100 meter long trains, 3 meter wide trains - they are nice, round, and middle of the road. This gives us 12000 square meters per hour. We can then get the cost paid per unit of capacity on a metro line. For the Ontario line — the high end — this is approximately 125,000 $/km/m2, by comparison some of the best performers globally cost just a few thousand $/km/m2. The nice thing about normalizing this way is that the train layout — whether it is biased towards longitudinal seats or more transverse doesn’t matter. We look at the floor area which gives a good sense of what the upper bound on capacity would be if seats were rearranged (as sometimes happens on very busy lines). This also lets us get around the issue of different jurisdictions quoting different capacities based on different crowding standards. You can see the large Asian systems achieve much higher capacity designs — with more capacity potential — per dollar spent. In other words they are more cost-efficient at building. This extreme plays out with metro projects in jurisdictions with very poor cost control like North America’s HART and Ontario Line and Hong Kong’s MTR lines. Doing this exercise demonstrates that those jurisdictions are paying a lot for very little hourly sectional capacity regardless of alignment choice. Hong Kong may have horrible cost control, but the city is building some of the most complex and highest capacity projects on Earth. In my Toronto Tragedy article, one thing to note is that while metro construction costs in Toronto are high down-scoping a metro project to a light metro or light rail project is only a solution to reduce “sticker shock”. However, they treat both these alternatives as providing equal value in speed, comfort and capacity. This is incorrect, sacrifices have been made in the design that erode the value of the project. Similarly, a common counterargument for advocates of light metro or more “lean” metro construction are the lower costs yielded from smaller stations. If capacity is an issue from a high demand corridor the cost delta of building said project as a light metro may be sufficient to fund most of another parallel light metro line. You can reach a similar screenline capacity potential as a single traditional metro line with a pair of light metros. However, a pair of light metros also provide more coverage. You might be able to designate one light metro to be an express regional metro and the other for local trips. This sounds great but, the analysis of the Ontario Line, a project which opted for reduced capacity and structural scale to reduce cost shows that the savings never materialized. Dividing a big number from a project afflicted with Anglo-world cost disease in half, still creates a big number. While doing these “optimizations” to projects you are losing value and capacity potential in an effort to reduce the cost. Meanwhile the rest of the world is multiplying and dividing with a small number as they optimize the scope of their metro design to fit their context, which sometimes means spending more money for a more lavish or capacious design. Just like the analysis with low cost countries, a closer analysis of the high-cost jurisdictions above revealed a similar schism in regards to capacity potential. East Asian Hong Kong is known for very poor cost control but tends to build pretty capacious “high value” projects. You can see it with comparisons between NYC 2nd Ave Subway and Hong Kong West Island Line. Both are complex city centre lines that need massive capacity to serve heavy passenger flows under a dense urban setting. Hong Kong has a lower cost per capacity which makes sense, the West Island Line’s per kilometre construction cost is lower than the 2nd Ave Subway and both projects have similar design capacities. You can then swing the comparison to Toronto’s Ontario Line which still has similar per kilometre unit costs with the Hong Kong West Island Line. However, the West Island Line demolishes the Ontario Line, being twice as cost-effective in capacity potential, really showing how important building for value is. The sell was for “lean” metro design using smaller trains and stations which clearly never materialized with the Ontario Line. Changing comparisons to New York, Toronto has the same eye-watering 150,000 $/km/m2 cost per capacity potential as the 2nd Ave Subway. However, the Ontario Line is arguably a simpler project with at-grade and elevated running it should be even cheaper to build, but it really isn’t. In Hong Kong, it is worth noting that the West Rail Line and Ma On Shan Medium Capacity System have proper scaling potential capacity cost effectiveness (both with ~20,000 $/km/m2). Both are more suburban mostly elevated ex-KCR lines. A doubling of capacity with double the size of stations and trains should approximately double costs, showing how projects should be properly scaling in cost and value. So then, significant real cost differentials exist both within the high cost Anglosphere, and the rest of the world groupings when capacity is considered. In the high cost world, the differential between the Ontario line and the KCR west rail line in Hong Kong (which is also a blend of above and below ground) is nearly 8 times. In our low cost examples, Milan M5 has 4 times the cost per capacity of Istanbul M5. Notably, the cost per unit capacity of Milan M5 is 50% higher than the KCR west rail line, though M5 is entirely underground. Keeping abreast of the transit networks across all of Canada’s big cities takes a lot of time and commitment. The Transit Brief lives and dies by your support. To receive new posts and support my work, consider becoming a free or paid subscriber. Thank you. If you enjoyed this post, consider sharing “The Transit Brief”. Made with love, in Canada - 2026. Invite your friends and earn rewards
If you enjoy The Transit Brief, share it with your friends and earn rewards when they subscribe.
|
- Home
- _Home Aviation
- __Home Airbus
- __Home Army
- __Home Drones
- __Home Helicopters
- _Home Watches
- __Omega
- __Home IWC
- __Vacheron Constantin
- __Jaeger Lecoultre
- __Home Rolex
- __Audemars Pigue
- _Home Moto
- __Home Suzuki Moto
- __Home Yamaha Moto
- __Home RZR Moto
- _Home Yachts
- __Fraser
- __Ferretti
- __Ferrari Yachts
- __Majesty
- __Benetti
- __Azimut
- __Edmiston
- __Dream Group
- __Feadship
- __Heesen
- __HMY Yachts
- SUV
- Sports
- Sedan
- Rally
- Convertible
- Coupe
- Hatchback
- Luxury









0 Comments