Subway cost per mile nearly 9 times higher than for light rail, says study

Recent study results of subway-LRT investment costs posted on the Light Rail Now (LRN) site provide a sobering reality check on the cost issue.

Projects examined were an assortment of “recent urban rail projects (all from the 2000s), either completed or well under construction and fully budgeted.” The report lists 24 subway and 13 LRT projects included in the analysis (in the USA and worldwide).
LRN’s study found median investment cost to be nearly 9 times higher for subway construction projects than for in-street LRT.

“Basically, for a given amount of available funding, you can construct 7 to 9 times as much surface LRT as subway. Put another way: For available resources, you can have a far more comprehensive
rail system with surface LRT, many times the size of a system relying on subway construction.”

For years, Austinai??i??s civic leaders and official urban rail planners have been trying to figure out how to raise more than $500 million in local and federal funding for a new-start surface urban rail project ai??i?? 5-6 miles of light rail transit (LRT) running in relatively lower-cost surface alignments. Suddenly, at least some official interest has turned to ai??i??.building a subway instead?

This seems to be influenced mainly by two factors:

ai??? An aversion or reluctance to shift urban public thoroughfare space away from apparently precious motor vehicle traffic and re-allocate it to public transit (rail in this case), and

ai??? Insistent claims by several subway proponents (disputed by professionals and advocates of LRT) that subway construction costs are nearly the same as, or only slightly more than, surface LRT.

Projects examined were an assortment of ai???recent urban rail projects (all from the 2000s), either completed or well under construction and fully The report lists 24 subway and 13 LRT projects included in the analysis.

ai??? Only ai???full subway projects (entirely or nearly totally underground)ai??? were examined in the study, including subway portions of LRT projects.

ai??? Only surface LRT projects exclusively, or nearly totally, in street alignments were included (ai???to compare the most difficult, highest-cost type of surface construction with subway constructionai???).

Summarizing the study results, LRN underscores the huge cost disparity between subway and in-street LRT construction, and the implications for a long-term rail expansion policy:

ai??i??for recent U.S. projects, subway construction has a median cost nearly seven times that of in-street LRT construction. Worldwide, the differential is nearly 9:1. And thats only comparing in-street LRT construction, not accounting for the possibility of, say, transitioning into an available railway alignment outside the city center, with far lower installation cost.

What this means is that, even if your community can somehow afford the initial financial commitment (even with federal assistance), expansion of your system will be severely attenuated. Basically, for a given amount of available funding, you can construct 7 to 9 times as much surface LRT as subway. Put another way: For available resources, you can have a far more comprehensive rail system with surface LRT, many times the size of a system relying on subway construction.


12 Responses to “Subway cost per mile nearly 9 times higher than for light rail, says study”
  1. Rico says:

    Grade separation is more expensive than not. That article is a great example of cherry picking facts though. Compare the above article with its cherry picked choices with this, That gives much more detailed info on ALL of the North American projects….oops so much for 9 times as expensive. The average North American LRT project is less than 3 times the average North American Heavy rail project.

    Regardless general averages are useless. The Evergreen line is less than the average North American LRT cost. That does not mean it would have been more expensive to build at grade, only that the factors governing the project costs such as the cost of tunnelling or elevated on that corridor were good. For that matter the fact they are spending several billion (that is with a b) dollars on a subway station in New York does not mean a subway station anywhere else will cost that much. The same thing means just because Barcelona can build fully underground subway lines for less than 60million per km does not mean we can.

    Zwei replies: Here is what several civil engineers told me years ago, subway construction costs about twice of that of elevated construction and elevated construction costs up to tens time of that of at-grade construction. The trick TransLink has learned is to grossly over engineer LRT projects, so the price is only a few percentage points of elevated construction and this has fooled everyone including BC’s Auditor General’s Office.

    In BC, LRT costs about twice as much to build than it should, so the costs for an economy LRT in Surrey are about bang on.

    What I sense is fear and loathing with the SkyTrian set because the truth is emerging that not all is what it seems with Translink.

  2. Richard says:


    If you take a look at the detailed estimates for Surrey LRT, you will note that they are not directly comparable to your estimates. They include the 30 year lifecycle capital costs which includes more vehicles for additional capacity. It also includes BRT from Newtown to White Rock. It is also in the year of expenditure $$ and includes inflation. Specifically:
    - $59 million for BRT construction
    - $30 million for BRT initial vehicles and OMC
    - $19 million for BRT expansion vehicles and OMC
    - $177 million for LRT expansion vehicles and OMC
    - $187 million for Real Inflation (Inflation over Consumer Price Index)*.

    - It also includes $97 million for right of way costs that may or may not be in your estimate

    Cost of vehicles and the OMC are $307 million almost twice your estimate.

    I suspect their estimates probably have a contingency of 20%-30% based on what they did. This is to be expected as there has been no actual design work done.

    Add this all up, and it is much closer to your estimate.
    Page 73

    Zwei replies: Actually the costs come from low cost LRT lines recently built and I believe very accurate.

  3. Richard says:

    That should be I suspect their estimates probably have a contingency of 20%-30% based on what they did in the UBC Line study.

    Zwei replies: TransLink has been known to use 50% contingencies. With absolutely no hands on experience with modern LRT I would suspect that TransLinks estimates could be as much as 100% too high. Economy is not in TransLink’s lexicon.

  4. Richard says:

    FYI The contigency on the Broadway subway is over $500 million.

    Zwei replies: Just demonstrates TransLink’s absurd planning. No credible transit planner would have such a high contingency for their plans.

  5. Richard says:

    Well, it’s 20% on $2.5 billion. As it is mostly tunnel and they never know what the conditions will be until they do geotechnical samples, it’s playing it on the safe side.

  6. eric chris says:

    Contingency on a large engineering project is 10%. Any more and the engineering firm risks losing the job, unless of course, the firm is SNCL bribing transportation officials to win the project. In other words, the real cost of the Broadway side show is $5 billion and the accountants at TransLink got sloppy when they published the $500 million contingency for the subway on Broadway.

    Thanks for the information. It looks like TransLink is playing the oldest trick in the book, low ball the project to obtain the tax funding and then cut it short until more funding can be squeezed out of taxpayers to finish it (or open cut like SNCL did on Cambie Street to save $400 million while business went bankrupt. The spineless COV engineers who allowed it to happen were rewarded with promotions and are now the proponents for the Broadway subway). In the meantime, the swindlers at TransLink are laughing all the way to the bank and are driving to work, heroes.

    Good news, TransLink is hiring and you can apply for a job. I can’t figure out the reason for TransLink hiring a civil engineer who knows little about SCADA; the proper engineer is someone with a mechanical engineering degree or electrical engineering degree, but then nothing about TransLink makes much sense.

    Still you’ll notice from the job description that the job at TransLink doesn’t require any engineering knowledge and the role is to supervise other engineers – likely SNCL engineers. Key to the “civil engineering” job at TransLink are negotiation skills and conflict resolution skills – engineering competence to design the system isn’t mentioned. Couldn’t TransLink just hire the electrical or mechanical engineer directly to do the design rather than hire an engineer to watch over the engineer who TransLink is going to be paying to do the design at SNCL?

    “About TransLink

    The South Coast British Columbia Transportation Authority (TransLink) is one of Canada’s most innovative transportation authorities and our responsibilities continue to grow. It is our unique task to move people and goods in an integrated road and transportation network, while preserving & enhancing Metro Vancouver’s quality of life.


    Responsible for defining and managing work programs to meet multiple objectives on several concurrent projects including managing resource requirements by allocating and reviewing work tasks. Included in this role is to liaise at a senior level with internal and external stakeholders and effectively manage internal and external communications. The scope of the role also involves defining and managing project risks, determining, justifying and managing consultant assignments to ensure overall on-time, on-budget delivery.


    Professional Engineer Designation is required. This includes four years of previous experience while qualifying as a Professional Engineer.
    Advanced courses in project management, contract negotiations and conflict resolution.

    Minimum of ten years’ experience in transportation related project management.
    Proven ability to build strong business relationships with internal and external stakeholders.
    Strong analytical thinking and problem solving skills, including ability to address conflicting issues with multiple stakeholders, and excellent verbal and written communication skills.
    Experience in road/bridge operations, field services, monitoring of construction and/or rehabilitation works, preparation of contract documents and contract administration on major projects are assets.
    Strong interpersonal and negotiation skills, leadership ability and commercial acumen are required.

    Other Requirements:
    Solid systems background such as automatic train control, train signaling, SCADA, communications and life safety systems

    Other Information
    Salary range starting at $101,091 per annum.”

  7. Rico says:

    Even I know the amount of contingency depends on the stage of design. 10% contingency is for a fully designed project, contingencies are typically 20 or 30% for projects that are early in the design process. I assumed you had enough experience to know that….

    I know now, who I will not be relying on for an estimate to paint a porch, eh Rico…!
    the Cardinal

  8. Richard says:


    These are not construction estimates. They No design work has been done at all. Just estimates based on the unit costs. The point is to just compare the options so a decision can be made on which one to go with. Once design work has been done and it is put out to tender, obviously the contigency would be lower.
    You’re like a drowning man Richard.. grasping for a smart response and how very wrong you are.
    the Cardinal

  9. eric chris says:

    @rico and richards,

    How long have you two worked EPCM?

    What are they, fake estimates? Projects like the sky train are cost plus and not lump sum. A contingency of more than 10% is robbery. Boring a tunnel is based on the known distance traveled and the level of uncertainty is low.

    We are not talking about a feasibility study for an unknown design where it is the first of its kind and the vendor has not been chosen. In the case of sky train, you have two other recent lines here and know the costs. Also, the vendor is Bombardier which isn’t designing a new train with new technology.

    To be sure, the tram line or LRT line on the other hand involves many unknowns (utility cost relocation) and the uncertainty is high. Still engineering firms don’t generally bid lump sum. The cost is “cost to build the project plus an extra margin for profit (10% at least) and the contingency is low.

    Finally, were is that link, rico, see my last comment which you avoided?


    I just told you that I can’t find the link and you keep being evasive. Provide the link, SVP,or quit making inane comments on this site.

    If you give a group of monkeys a computer, they will arrive at some result. What did the group of monkeys at TransLink arrive at for the transit demand on Broadway? What is the maximum transit demand on Broadway according to TransLink in terms of pph?

    There is no theoretical relationship between the number of times that people board buses and the actual transit demand. It is unique for each route and must be obtained empirically. Even the COV engineers don’t know what the transit demand is on Broadway.

    How can the COV “engineers” say that only sky train can handle the transit demand which they have never made public?

  10. Rico says:


    Maybe I should get you to paint my porch. Who knows since I have not told you where it is or how big it is or what I want done maybe you will give me a really good price.

  11. Rico says:


    How come you can’t use google. I just spent 15minutes (my computer was glitchy today) using the same info I gave you to dig up the info (which is actually for 2041 not 2030, so much for my memory). Be nice if you held Zei to the same standard since he has still not posted any links or sources (I know I am truly shocked…..). Anyway here you go.

    Zwei replies: Rico you just don’t get it, Translink’s plans are not worth the paper they are printed on, no matter how many links they use. Links do not equal facts. End of debate.

  12. eric chris says:

    Thanks for wasting my time. Much appreciated.

    How much did the garbage in the link cost TransLink, bye the way? Did TransLink really have to pay SNC Lavalin and Steer Davies Gleave for this arts level “study” in the link provided by you?

    I asked for the present ridership along Broadway as a function of distance from Commercial Drive to UBC Q(x), where Q is the number of people using the buses on Broadway as a function of distance along Broadway, X. Your link on page 11 doesn’t indicate this. On page 11, your “Googled” link merely states the capacity of the buses for a number of transit options (lacking any depth of thought) to UBC:

    You’ve proven that when you are caught pulling numbers out of the air, you’d rather lie or try to hide it than admit it. Good job. You’re TransLink material for sure.