Quick Note: Don’t Overlearn From a Case of Success

I’ve been asked in comments to my previous post about construction costs what can be done to contain them, and tempting as it is to just repeat listing good cases, in the wrong context it can do more harm than good. Whenever we are faced with a success story, it’s tempting to confuse a good system with individual competence, in both directions.

The list of conclusions given by Madrid Metro CEO Manuel Melis Maynar is a good place to start for discussing low-cost subway construction. So is Calgary Transit’s explanation for how it keeps costs low. But things are always more complex than a short list of principles, and details always matter, and those can be easily lost when trying to port cases of success. I think it’s obvious that Madrid’s EPB method is not easily ported to the harder rock of Manhattan, but the administrative factors could be problematic, too: is there enough expertise within the MTA to complete projects with an in-house staff of six? After all, in California, the small size of the in-house staff is one reason why the consultants can run circles around everyone and propose multiple billions’ worth of concrete to solve problems that good organization could fix for no money.

Of course, in the other direction, it’s easy to attribute to individual genius what is the result of good business culture. Compare, for example, the praise heaped on Steve Jobs, with the more sober description by Malcolm Gladwell of the office cultures involved in the birth of the Macintosh. But even this opposite problem can be shoehorned into the same issue.

To be more precise, in both cases, what’s really needed for optimal performance is good organization and business culture. This does not mean that individual lessons about keeping design and construction separate and choosing contractors based on more than just cost are bad, or that they shouldn’t be implemented everywhere. They should. Obtaining average performance is not difficult; that’s why a large majority of cities have it. What’s difficult is obtaining optimal cost control.


  1. aw

    For the Madrid project, you have to wonder if they got a volume discount by ordering 6 TBMs at one time.

    • Miles Bader

      Hmm, wouldn’t any rail agency doing a lot of construction inevitably get a fair amount of advantage from that?

      It always seemed to me that one of the problems with U.S. passenger-rail construction is simply that not enough of it gets done, at least in one place, so they never reach a critical mass of experience/bulk-prices/standardization/competition-among-suppliers/etc…

  2. Andre Lot

    I’ll make an incidental comment: whereas much focus on the transit blogosphere revolves around “cost per km”, I think from an engineering standpoint it is better to bring some information on stations/km as well.

    Tunnel boring strictu sensu is *relatively* comparable across countries within geological similar areas. Majority of cost overruns on subways, usually, involves stations and their immediate surroundings, not the tunnels themselves. If cities worldwide were to use sequential cut-and-cover to build stations (e.g., the method used until the 1960s, in which a whole bore is opened and stays opened during all station construction, being closed just as everything is already finished beneath), construction costs would surely drop, but I don’t think – for instance – New Yorkers would accept a 3-year long permanent disruption of their street front, with all but a small alley left for access to properties.

    This is why a long high-speed tunnel under hundreds of meter of rock and earth (increased pressure) and dealing with extensive high-pressure groundwater aquifers are still much cheaper on a distance-base cost than subways, despite the fact they are more challenging as you can’t just dig – for instance – 2km down from the Alps to study the rock formation every 100m or so like you can in an urban environment.


    @aw: I do believe this is one of the transit cost overruns factors in US. You don’t have a large number of projects going on simultaneously, creating a stable business that would attract more suppliers and firms. Then, US transit agencies are severely understaffed in critical personnel within engineering positions. There is nothing inherently wrong in using consultants and contractors, but you do need an efficient, if small, technical supervisory team that can get hold on consultants, or else they are given a free hand to design solutions that might not be optimal.

    • Stephen Smith

      I’m not sure about the exact sequencing, but New York’s got East Side Access, the 7 extension, and the Second Ave. Subway. Maybe not enough for six TBMs like in Madrid or eight (?) as proposed in the Grand Paris plan, but theoretically more than enough to bring costs below $1 billion/km.

    • anonymouse

      TBM prices are a red herring: they’re pretty cheap compared to the total construction cost. And as far as cut and cover construction goes, that’s still pretty much how it’s done in some places, at least for stations. Seattle’s Capitol Hill station is a giant open pit, and will continue to be that way until they finish building the station and cover it back up. On the LA Gold Line extension, they also dug giant pits for the stations, but once they did so, they put up walls to hold the dirt and installed decking on top so that road traffic could continue relatively unimpeded. And that extension was relatively inexpensive to build, given the tunneling and other challenges involved.

      • aw

        It’s not just the TBMs themselves, but also the competetive pressure of bidding out multiple segments for simultaneous construction, and the expertise that’s built up by the contractors. Seattle’s U-Link might have goten some effect from that; they have two tunnelling contractors with three TBMs. I believe both of the tunnel contracts came in under estimates, but that might have also been due to the sluggish economy.

        Also, the two stations are big pits that won’t be covered for a while, but they’re also not being built under arterial streets.

        • Richard Mlynarik

          It’s not just the TBMs themselves, but also the competetive pressure of bidding out multiple segments for simultaneous construction, and the expertise that’s built up by the contractors.

          Exactly right.

          There was also the very strong understanding that if you do this well there will be plenty more work where that came from. If you’re not up-front with us, you’re dead. There were technical problems, but they were addressed early and in the open and always with the aim of minimizing risk (safety and schedule, both of which can dominate cost.)

          This is the exact opposite of the organizational/managerial/economic incentives we see in the US, where the larger the fraud in budget/utility/schedule estimation the bigger thay payoff and the larger the subsequent monopoly/duopoly/cartel projects and contracts.

          The incredible amount of subsequent low-cost high-quality public works construction in Spain — subways in provincial cities for about the same as BRT street repaving in the US — attests to the mutual private and public benefits of doing more good stuff well than the single too-big-to-fail megaproject boondoggle sole-source contractor-captive earmarked there-is-no-alternative projects with which we’re all so familiar.

          • Stephen Smith

            How much did the ones in Bilbao, Valencia, etc. cost, anyway? I’m trying (not very hard) to find costs but am coming up short.

    • Alon Levy

      You’re right about stations per km, and I’m sorry for not emphasizing this more. In my defense, in the first set of data I looked at (SAS, the 7 Extension, Paris M14, Berlin U55, Tokyo’s Oedo and Fukutoshin Line, and London’s Jubilee Line Extension), there wasn’t much variance in the station spacing. Of course the most expensive project of all, ESA, has only one station, but this is because it’s a multilevel cavern, so it’s best to omit it from any normal urban rail calculation.

      As for the cut-and-cover issue, there are so many resident complaints about disruption along Second Avenue anyway that in comments people are proposing cut-and-cover as an alternative. The idea is that there is going to be disruption anyway, for station construction and for blasting, and cut-and-cover is good as a quick and dirty option: major disruption for 3 years, instead of some disruption for 8. In both cases, about half the street would be unavailable at a time, as was the case in the construction of the first subway; judging by the description of early-1900s construction, the disruption was about the same as what I saw at 72nd Street and 2nd Avenue when I lived nearby, so the main difference is whether there’s disruption along the entire street or only at station locations.

      The issue is that cut-and-cover is only feasible when the line can be done relatively shallowly. This is the case for most of SAS, which only intersects deep-level east-west tunnels and is designed to go above rather than below them, but is not the case for most other infill subway lines (including the 7 Extension), which go under older infrastructure.

      But yes, you’re right that stations are one major reason urban infrastructure is generally more expensive than tunneling through the mountains.

      • Adirondacker12800

        I’m old enough to remember the parts of the Second Avenue Subway – that are already built – being built. It not an open pit for years. They cover it up in wood and traffic flows overhead while they are building.

  3. Pingback: More on Cost Comparisons | Pedestrian Observations
  4. Pingback: Why is it so expensive to build a bridge in America? The answer: Our greedy and undemocratic political culture | Bamboo Innovator

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.