Quote of the Week: Transit Construction Costs in the U.S.

It seems like every time I read about a metro line outside the United States, except in the UK, it is way cheaper than we can do. … Alon Levy has contrasted the cost of subway construction in New York with the much lower costs in Tokyo, for example. We seem to have a system in the US that significantly inflates the cost of construction vs. the rest of the world. Many of the typical complaints as to why this might be would seem to have no merit. Other countries are heavily unionized and regulated, for example, so don’t blame organized labor. (South Korean unions are famously militant). Spain and Japan are not exactly low cost countries. And basically all new systems world are fully compliant with equivalents to the [Americans with Disabilities Act].

— Aaron Renn, the Urbanophile

I keep hearing this observation about US construction costs.  It’s totally outside my expertise, but if anyone has seen a satisfactory explanation of why US transit construction is so expensive, please link to it in a comment.

58 Responses to Quote of the Week: Transit Construction Costs in the U.S.

  1. EngineerScotty December 15, 2010 at 4:20 pm #

    Not a link–but many projects in the US are funded with federal matching funds–free money, as it were.
    Could local officials be inflating costs of projects to get the largest injection from Uncle Sam possible into the local economy?

  2. Andrew December 15, 2010 at 4:45 pm #

    This article at The Transport Politic outlines a few reasons why Besançon, France was able to build their new Tramway so inexpensively relative to other places. I suspect a large part of it comes down to culture, which this quote illustrates nicely:

    “In some places, instead of reworking a street to offer both rail and automobile traffic, the city is simply banning cars.”

    Does that ever happen in the car-culture dominated United States?

  3. Leigh Holcombe December 15, 2010 at 5:36 pm #

    In my experience, road construction in the US is as much about employment as it is about the actual work. A new highway ramp can be built over the weekend in an emergency. But under normal conditions, road contractors are running a public works program. The pace of the work is designed to provide gainful employment over time, even if that means increased cost. I assume rail construction, if it is publicly subsidized, operates the same way.

  4. Brandon December 15, 2010 at 6:14 pm #

    Ive wondered the same thing, and I seriously think its holding the US back. NYC in particular is having some serious growing pains with the price of transit expansions.

  5. Cap'n Transit December 15, 2010 at 6:20 pm #

    It seems like it’d be relatively straightforward to break the total cost down into parts, labor, administration, guaranteed profits, bond interest and so on, and compare them between a North American project, a European one and an Asian one. Are there particular segments that make up most of the difference, or is it across the board? Has anyone done this?

  6. Valkraider December 15, 2010 at 6:30 pm #

    Lawsuits? Insurance? NIMBYism? Public “process”?
    In the USA people can and will sue if they feel like their voice wasn’t heard. So we have to have 100 public input forums and the plans get changed 10000 times. Construction work has to have huge insurance policies in case *anything* happens. One person who doesn’t like the project can hold it up in court for years. We are never allowed to close roads to traffic during construction, or people complain. Every property owner has a complaint that had to be addressed. Each project has to meet ADA requirements and be built to prevent every possible stupid thing someone could do.
    I could go on and on…

  7. Jonathan Parker December 15, 2010 at 6:51 pm #

    Scotty, I don’t think so. Maybe if the federal funds were formula based like highways (everyone gets a certain amount per year), but not based on a fiercely competitive process like New Starts that creates incentives to keep costs down.
    My guess would be a combination of limited supply of designers and contractors who know how to perform the work, combined with environmental regulations that lengthen the development timeline that drive up already high wages.
    Wages in the US are higher than most other countries. I can’t verify this source, but take a look at this link which compares salaries of engineers in various countries: http://www.worldsalaries.org/engineer.shtml. If it’s even close to correct, aside from the UK no one is close to the US. How many engineering man hours does it take to take a rail project from concept to service? It adds up quickly.
    Most projects take perhaps 10 years or more to make it through the environmental regulatory process in the US, and certainly time is money. It doesn’t help when the demand for these transit projects has outstripped the supply of funds as we see now, adding further delays and increasing costs.
    And when there are fewer transit projects being built due to the above factors, the labor market has fewer incentives to respond with more planners, engineers, and contractors who know how to execute these projects. Fewer skilled professionals in the industry yields even higher wages than the already high wages that exist in the US. Every contractor out there knows how to pave a parking lot, but track ballast is a comparative niche market.
    If, due to a variety of factors (environmental, social, and economic) the US decided to engage in a transit space race (apologies to Jeff Wood), I think you’d see the growth in the costs of these projects decline over time.

  8. rhywun December 15, 2010 at 7:17 pm #

    It doesn’t help that so many American projects are driven by politics rather than demand. The motivation for most projects seems to be some combination of bringing home the bacon, setting up a lucrative jobs program, or perhaps just stroking one’s ego (this seems to drive many NYC mayors…). The root motivation is of course votes.

  9. Alon Levy December 15, 2010 at 7:26 pm #

    The US builds both road and rail infrastructure at exorbitant cost. For one example I recently came across, compare the costs of the Tappan Zee Bridge and Bay Bridge Eastern Span replacements with the cost of the Oresund Bridge-Tunnel.
    As far as I know, this is restricted to transportation infrastructure; water projects seem to be fine. But I’m not as familiar with the numbers as I am with rail and road expansions.

  10. Danny December 15, 2010 at 7:38 pm #

    Water projects aren’t so fine, but I have a feeling that has to do more with geography and our penchant for building cities in the middle of deserts.

  11. oscar December 15, 2010 at 8:13 pm #

    labor costs (other countries may be heavily unionized, but American unions have grifting down to a science)
    public input processes
    pro-car culture

  12. Cap'n Transit December 15, 2010 at 9:26 pm #

    Okay, so nobody actually knows why they cost more, we’re just conjuring up our favorite bogeymen, right?

  13. Adriana December 15, 2010 at 10:55 pm #

    I think it depends on the project – this paper plots up BRT costs per km on 44 projects worldwide.. the costliest is Boston, but considering the tunnels, that is no surprise.

  14. Inmotu.wordpress.com December 16, 2010 at 3:27 am #

    Wow, I thought that construction costs in Russia are high ($150-$170M per km of the tunnels in St. Petersburg subway) but these figures are amazing.

  15. Patrick M. December 16, 2010 at 3:32 am #

    Valkraider largely has it right to my understanding.
    I have always heard it explained that it is substantially higher bureaucratic costs. Because of extreme fear of local governments handing over favours as well as preparation to prevent court action/disruption of daily life.
    Again, this is largey second hand, but the necessity to accept and investigate other bidders in order to ensure a fair process (and their litigious recourse’s) results in a much higher end cost. It’s funny that methods meant to ensure the tax payer get value for goods are so onerous and costly they substantially drive up the end product cost.
    As to the comments by ‘oscar’
    “labor costs (other countries may be heavily unionized, but American unions have grifting down to a science)”
    Um… no. That’s just silly, go visit France. Or hell, Canada. We’re much better at ensuring fair wages than America is.
    The remainder of his post though not fleshed out, can be interpreted as fair comment if you read into the basic words.

  16. teme December 16, 2010 at 6:16 am #

    Phrase I keep hearing is unneccesarily heavy engineering. If light rail would cost as much here I’d be advocating BRT too.
    Over here in Helsinki tram line constructions costs are something like 3 to 5 million euros per km that is about the same as a busway, german style light rail has been estimated at about 8 mil. Our (very heavy) metro costs about 70 to 100 mil per km to build fully underground, underground stations being a major expense.
    Railwaytechnology.com has interesting project directory with short descriptions. At random I picked Hudson – Bergen light rail, http://www.railway-technology.com/projects/hudson/ If somebody asked me to guess what would it cost to build 20 mile long light rail, with 52 cars, depot and so on I’d say about 300 to 500 million euros. If there are lot of tunnels or similar heavy engineering and converting to mickey mouse currency I’d say max one billion dollars. The cost given is two biilion dollars.
    Now compare to Bergen, Norway light rail which just opened. It’s 10km, one third of Hudson – Bergen. Third of it is in tunnels due to terrain, which is expensive. Norway on the whole is about as wealthy and expensive as it gets, and they probably doubled the construction costs due to protectionism. Total project cost 225M€ and in say Sweden would probably have been something like half that. http://www.railway-technology.com/projects/bergen-light-rail/

  17. Tom West December 16, 2010 at 6:17 am #

    The New York subway is an exception: it’s tunnelled through solid granite. Most subway systems are tunnelled through much softer rock, so progress is quicker (and hence labour costs are lower). For example, London is on clay, and so digging tunnels for the Underground was/is easy.

  18. teme December 16, 2010 at 6:19 am #

    “Our (very heavy) metro costs about 70 to 100 mil per km to build fully underground, underground stations being a major expense.”
    Sorry, about 50 to 70. 70 to 100 is per station, not per km.

  19. Alex December 16, 2010 at 7:04 am #

    Even though many of the usual culprits are resurfacing, it is always useful to ask this question. The more we talk about it, the less engineers in the US can ignore it.
    The stereotypical American isolationist attitude may be present in the engineering profession as well. How many engineers on the Seattle light rail project have even heard of Bergen’s project, much less reviewed its cost-saving techniques? Instead it seems to me (an observer) that engineers in the US only look at projects in the US and only talk to other engineers in the US, and thereby limit their ability to learn from successes elsewhere.

  20. Jimmy December 16, 2010 at 7:11 am #

    Interesting question. Here might be some insight: http://vbn.aau.dk/files/14076659/Comparison_of_Capital_Costs.pdf

  21. Jonathon December 16, 2010 at 7:51 am #

    One thing of note is that in Toronto, projects go ahead with absolutely massive contingencies, and generally eat them up mid-way through, such as the Spadina subway extension, now cutting back on station finishes after blowing through most of a $400 million contingency fund.

  22. Jimmy December 16, 2010 at 8:01 am #

    Also, are the costs of the NYC project representative of the rest of the US? The Seattle University Link extension is priced at about $380m/km and the Washington Metro Silver Line extension is priced at around $180m/km. Granted, the Silver Line is above ground but even if you triple the cost it doesn’t even begin to approach NYC 2nd Ave. extension figures.

  23. JJJ December 16, 2010 at 10:09 am #

    Because we let them?
    There is no accountability for contractors in this country. If company a says they can build a rail line for 50 million, and then they keep coming back and asking for more and more after they win the bid….why the hell would we ever allow them to bid again?
    But we do. One of the big dig contractors, for example, is involved with the extension of the Boston green line. Why?
    If your plumber screws you over, you get a new plumber. Government doesn’t seem to operate this way, they give their business to the same flawed companies over and over again.
    If it was written in stone that a company who screws over the public on a project will never be hired again, I foresee better jobs and lower costs.

  24. EngineerScotty December 16, 2010 at 10:30 am #

    Steve Duin of The Oregonian has an interesting column this morning; concerning the role of NEPA in inflating project costs. While Duin (a metro columnist) is no expert, he notes that the financial modelling requirements of NEPA (which include finance charges in the stated costs, and requires costs to be stated in year-of-expenditure figures, added together as though they are current dollars) frequently make projects look more expensive than they actually are.
    The latter point is most interesting. While its laudable to make sure all future charges are accounted for, it seems to me that long-term expenses should be appropriately discounted.
    Duin also complains about the NEPA process itself–the expense, the time, and the opportunities it provides for NIMBYs to through wrenches in the works. Of course, the latter issue is a big reason we have NEPA in the first place–to restrain the impulse to use eminent domain to bulldoze entire neighborhoods in order to build infrastructure.

  25. Tobias December 16, 2010 at 11:05 am #

    Simple answer: the US-Dollar is overvalued. The same can be said about the Pound Sterling.

  26. Aaron M. Renn December 16, 2010 at 11:33 am #

    I think that a big issue in the US is that there are only a tiny handful of engineering firms that handle these jobs. For really complex jobs, I believe just one company is dominant. They profit from bigger price tags. We need to break up this oligopoly and bring in new entrants to shake things up. Unfortunately, professional services are discretionary contracts and highly political/influence driven. Also, the firms we’d want to enter the market are often foreign, and that’s political problematic as well.

  27. TransitPlannerMunich December 16, 2010 at 11:33 am #

    Last weekend the new metro extension in Munich was opened for operation. Length:
    about 2 km (all underground), costs: 180 million Euros in total, including all
    tunnels and stations (2 stations), art and surface reconstruction.
    Here is a link:
    Usually, here in Munich, the costs for a new tram/light rail line are below 10 million
    Euros per km, for a metro/subway (underground) below 100 million Euros per km.
    But basically Tobias probably gave the perfect explantion in his statement.

  28. Brian McCann December 16, 2010 at 1:07 pm #

    I agree that it is the NEPA process and requirements. One almost believes that this was developed by the auto and oil interests to make transit projects difficult, if not impossible. Another factor, especially in HSR or commuter rail projects is inducing a private property owner to sell or lease rail right-of-way. Without some element of compulsion, there has to be an incentive, and these add up costs.
    We also have extreme demands on transit vehicle design requirements that appear to be attempts to limit foreign entrants into the market. Yet the US has done nothing to create the railcar orders that would encourage and incentivise domestic manufacturing.

  29. EngineerScotty December 16, 2010 at 1:58 pm #

    Regarding NEPA–it applies to road projects as well (and much of NEPA was drafted in response to the abuses involved in the massive freeway construction boom last century).
    The ROW issue is another important one. Railroads are protected by Federal law from local use of eminent domain–and many railroads demand a king’s ransom to sell off their ROW, especially when the buyer is a government looking to build transit. (UPRR is a particularly egregious offender). And given that many rail lines are built on easements rather than land owned outright by the railroad, there’s often encroached-on property owners who would much rather have the line simply abandoned than converted to transit use.

  30. D December 16, 2010 at 2:51 pm #

    As part of the movement to promote alternatives to a capacity-enlarging reconstruction of a mega-interchange in Montréal – the Turcot – one of the municipal opposition parties recently published a report that touches on the particularly high cost of construction in Québec relative to elsewhere in the world and North America. Part of the back story on this is that the construction industry in Québec is suspected of being heavily corrupt, with recent scandals plaguing the current provincial government.
    The report has a great comparison of material and wage costs for Québec and other jurisdictions, and considers wages and cost of living as well; partly to argue against spending so much on an interchange (they recommend spending half the proposed budget ($3B) on a slew of transit initiatives) and partly to question why, if materials are affordable and workers are actually paid less in Québec, costs are so high.
    See pages 7 and 8 of the report (in French), which is at: http://www.projetmontreal.org/files/documents/2010_12_14_projetmontreal_document_1292375153_fr.pdf
    For more information on Turcot, I’d heavily recommend that you visit http://turcot.ca/2010/11/26/counter-proposal-for-the-turcot/ – there are also other good resources on that site, including links to the official proposal of the provincial government.

  31. Alon Levy December 16, 2010 at 9:20 pm #

    I’ve heard the granite explanation for Manhattan costs a few times before. I’m not really convinced, for two reasons: first, soft ground can be as big of a problem to tunnel in as hard ground, if there’s a threat of a cave-in.
    And second, the region’s light rail costs are out of whack as well. The Hudson-Bergen Light Rail, a project connecting existing abandoned railroad ROWs with a streetcar-mode segment in downtown Jersey City, is the most expensive at-grade LRT in the US. New Jersey Transit’s Meadowlands extension was supremely expensive as well, about 5-10 times as much as you’d expect an above-ground commuter line to cost in Europe.
    I don’t think the dollar is overvalued overall. The standard IMF and OECD calculations of purchasing power actually say it’s undervalued. The US has very high costs of public goods, and compensates by having cheaper private goods. If you try to deflate GDP based only on public goods, you’ll get nonsense figures: Spain would look like the richest European country, and the US wouldn’t look much richer than China. It’s not observed in corporate choices of where to relocate, or in living standards (jokes about US health care aside), or in the spread of consumer goods.

  32. J B December 17, 2010 at 12:26 am #

    @Tobias, if both the UK pound and the US dollar are overvalued, you still have to explain the very large discrepancy between the US and the UK.

  33. Eric December 17, 2010 at 6:30 am #

    Just a theory, but, we’re just a bigger country and used to “greenfield” development. This kind of development tends to be bundled, tweaking pro forma minutiae for profit. (Hence the higher salaries?) Because we are good at this, we resist the kind of centralized planning that would give us greater proficiency with complexity. Complex infrastructure improvements don’t allow as much tweaking for cost efficiencies. It’s a different culture of development here. We’re geared toward rapidity, not lumpy project scenarios with a lot of political foot-dragging, calisthenics and back-up involved.

  34. Tobias December 17, 2010 at 7:10 am #

    @ J B
    I don’t say that they are equally overvalued.
    @ Alan Levy
    The US are flooded with consumer goods from several countries, one in particular. Such excessive supply boosts purchasing power. But only on these goods. Domestically manufactured goods or investments in infrastructure which can’t be imported are rather expensive in comparison.
    This kind of purchasing power is rather misleading in my eyes. More important is what domestically produced goods and services does a US Dollar buy. And this seems to be little.

  35. Paul December 17, 2010 at 7:40 am #

    The cost of getting a project to the design/build stage are not usually considered part of the capital cost of a project so I don’t think onerous public consultation is driving the higher costs for North American infrastructure, although there may be something to the previously mentioned mitigation work during construction to placate businesses and residents impacted by construction.
    I have no “concrete” evidence to bring forward but one of the differences I’ve noticed between infrastructure projects in North America and other parts of the world is the difference in footprint of the facilities. From my experience, European and Asian projects tend to minimize footprint, only building what is necessary. North American projects seem to involve absolutely massive amounts of concrete to build oversized facilities (especially at transfer/terminal stations)to accommodate future volumes which may never be realized. Some of this may be tied to NFPA (not NEPA) requirements.

  36. anonymouse December 17, 2010 at 3:44 pm #

    It’s something of a vicious cycle. Projects are so expensive because there are so few of them, and there are so few of them because they’re so expensive. Projects being rare means that as much as possible gets piled on to one project: street reconstructions, bike paths, and so on. And of course the facilities are often oversized (see Millbrae for an example, with three commuter rail and three rapid transit platforms, of which only two commuter and one rapid transit platforms are ever used). That is partly a reflection of the project-oriented rather than system-oriented approach to transit planning and funding here. The other aspect of this is that much more of the work is custom, rather than standard designs, for tracks, rolling stock, structures, and so on. And there might be an aspect of shoddy project management as well, with that function being outsourced since there’s no in-house expertise (again, due to the project-oriented rather than system-oriented approach to building.)

  37. Ed Sanderson December 17, 2010 at 7:48 pm #

    I have been stopped in mid-conversation by the cry of TMI but this seems to be a case of TLI. Too Little Information because:
    1. What is the cost of the right-of-way. Does it involve buying private property? Can you use existing ROW but pay for the privilege? Is below surface ROW “free” or does some entity have to be compensated?
    2. Is geology destiny? How expensive is it to burrow through Manhattan schist? Chicago’s (and London’s) muck and mire? Or in the case of Sydney ancient soft sandstone?
    3. What is the hourly wage of the average common laborer or an uncommon expert engineer and all those in between?
    4. For above ground projects, what is the percentage of cost associated with weatherizing to withstand, say, plus 100 degrees(F) and minus 20 degrees(F)?
    5. If new or extended lines require new rolling stock, what is the cost of these new transit machines?
    6. How equal are the requirements for new stations that meet he American ADA standards with foreign ADA standards?
    I, frankly, don’t know the answers to these questions and after reading all of the above, I’m not sure that anyone does.

  38. Alon Levy December 17, 2010 at 10:45 pm #

    @Ed: 5 is actually easy to answer, with minimal Wikipedia gazing and Railway Gazette archive searching. I’ve punted on it because the system I’m most familiar with, New York, gets rolling cost for below-average prices.
    6 I don’t know for every city, but I know that in multiple cities I know of in Asia, all stations must be barrier-free, and some are 100% or nearly 100% wheelchair-accessible.
    1 is also pretty easy, in the sense that “most” of the cost (i.e. a very large majority – I’m not sure how large) of transit in the US is not ROW acquisition. The biggest-ticket item is the basic infrastructure, which consists of grading at-grade land, building the viaducts and tunnels, and digging station shells.
    4 I’m not so sure about. But the weather in the US cities we’re comparing unfavorably with the rest of the world isn’t especially severe. It’s not just Minneapolis or Phoenix that has very large costs, but also (primarily) the East Coast and California.
    3 I’m also not sure about, but you can get a first-order approximation by plotting unionization rates and per capita income. The US is slightly richer than most other developed countries, but not all, especially if one uses exchange rate conversions instead of PPP.
    2 is the only one that could be a serious variable, and when I ask New York-area engineers, they never mention problems with the rock, just labor costs or regulatory problems. If anything, I’ve heard of geological problems elsewhere, in ancient cities like Athens and Istanbul, where tunneling has uncovered archeological artifacts.
    @Tobias: first, just a nitpick, but my name isn’t Alan.
    Second, even if you look at domestic items, the US is no more expensive than Europe. In fact, most GDP is about domestic items, because of the dominance of untraded services over traded goods. Food costs are almost all domestic, housing costs are entirely domestic, etc.

  39. Ed Sanderson December 18, 2010 at 7:11 pm #

    TLI take two:
    The cost for the CTA’s Red Line extension has been estimated from $740 million to $1.4 billion yet I have not been able to verify any number based on the documents by the agency itself. The estimates may be there and I missed them so the best I can do is provide the link.

  40. Louis December 18, 2010 at 10:45 pm #

    For me, it is basically a case of doing things “perfectly” vs. doing things “well enough.” Back in the heyday of streetcar building in the united states, tracks could be built without federal funding and without being so incredibly over-engineered. This also applies to road-building projects. The case of Besançon mentioned above is a great example. Why can’t we just put some rails in a street, run some cantenary, and be done with it? Is laying some track in streets really so difficult?

  41. Aaron M. Renn December 19, 2010 at 5:05 pm #

    Alon, not sure if you’re still reading this thread, but what’s your take on airport construction? I haven’t looked at hard data, but I do get a general impression that US airport construction costs are out of line vs. the world. They strike me as ballpark average.

  42. Alon Levy December 20, 2010 at 8:27 am #

    I have no idea. It wouldn’t surprise me either way.

  43. TransitPlannerMunich December 20, 2010 at 12:04 pm #

    I doubt it has to do with “perfection”.
    Indeed I would usually not use the words “perfection” and “U.S. infrastructure” in a grammatical context. Especially I was astonished about the term “incredibly over-engineered”. As an engineer who spent a part of his university education in the U.S. but most of it in Europe it is not exactly how I would judge U.S. engineering – a lot seems to be good enough that it works somehow, but not more. Except maybe that in the U.S. it seems that bigger is better. I guess that in the U.S. a lot of money for public projects is paid for consultants, more consultants, lawyers, bank people, MBA-guys etc. – but certainly not for engineering or “over engineering”. Maybe an exception is the military technoloy. No surprise when you see how important military and military technology is in the U.S., but also this has a price tag.

  44. Rob December 20, 2010 at 11:32 pm #

    I don’t know if anyone is still following this thread, but here are my thoughts for what it’s worth.
    If you look at construction cost inflation on a graph (both highway and transit), there was a big jerk in 1991 when inflation headed skyward. So what changed suddenly then?
    One thing was clearly the introduction of environmental regulations, as others have mentioned. I don’t think it’s the NEPA analysis; there has been a ton of environmental retrofitting that’s needed to be incorporated into every project.
    But where I live, and maybe it didn’t happen everywhere, we were in the middle of a bunch of economic bubbles that made people feel a lot wealthier than we really were. At the state and local level where I live, suddenly it became fashionable to pose voter initiatives that promised immediate results by selling tons of bonds.
    My state raised gas taxes by almost 15 cents, and bonded the entire amount in order to deliver an aggressive construction program within 10 years. Not only does that double the price of every project in the program (because financing costs exceed actual construction costs), it also strained the capacity of engineering resources, so engineering talent needed to be imported at an ever-increasing price. Now that spending is near its end so layoffs are coming, and most of the remaining gas taxes (almost 3/4) will be tied up paying the financing costs for many years to come. I’ve mentioned the highway side, but rail transit has followed the same path to an even greater extent.
    One last factor could be the new risk-based cost estimating methods that have been developed to try to stay ahead of mega-inflation. Sometimes I wonder whether processes that presume huge inflation at the estimate stage just accelerate the underlying trend, as contractors follow the public estimates and plan their bids accordingly.
    In the old days governments used to fashion boring 20-year programs to deliver improvements incrementally as they could be afforded, but in the last decade we’ve preferred mega-projects that can be packaged and sold politically with the promise of immediate results. I have to believe that’s at least part of the reason costs have soared.
    I wish someone would write the definitive treatise on this. It’s amazing to me that, after two decades of project inflation nobody can suitably explain, that nobody has tackled this and wrestled it to the ground.

  45. Mike December 21, 2010 at 4:19 am #

    I’m not an expert but was under the impression that with modern techniques tunnel construction is cheaper in solid hard rock like granite if the rock itself serves as the tunnel structure without (much) reinforcement. Softer rock and sediments are much more problematic, requiring reinforcement or a full concrete tube to be built. Also, is the NYC area really primarily granite? I would have thought that the coastal plain of the Hudson River would have lots of sediments.

  46. Alon Levy December 21, 2010 at 6:05 am #

    @Mike: Manhattan is primarily granite, yes. One of the reasons it has skyscrapers is that the rock is very hard, especially in the two CBDs, making it easy to build high without deep foundations. The ground gets softer in the rest of the metro area, though.
    The Hudson River is a drowned river valley and has no coastal plain.

  47. Nicholas Barnard December 21, 2010 at 5:53 pm #

    @Rob Some projects have to be done as a mega project. I’ll cite the Seattle area’s Central Link from downtown to the airport. In actuality the start of this project was in 1990 when the city put in a bus tunnel downtown with rails and the intention of extending this further.
    Now the downtown to the station before the airport portion was built in one fell swoop, (The airport station was an added on four months after the rest of the line opened.) Its pretty hard to argue that there would be enough economic benefit to build out just a few of the stations on this line but not the others.
    Can you imagine building just half of the Second Avenue Subway line in New York City? IT doesn’t make sense.

  48. Rob December 21, 2010 at 7:04 pm #

    @Nicholas, you’re right of course, there are good uses for bonding for projects that need to be done all at once. But let’s look at the Seattle central link project a little more.
    Sound Transit’s service area comprises most of three counties, but their subarea policies require all of the costs in one subarea to be paid for from the same subarea. If there was *trust* among elected officials, they could have directed revenues from the entire service area to pay for central link, and then moved on to other portions of the region (that was one of the rationales for establishing such a large district). But since there isn’t trust like that, the Seattle subarea needed to borrow the funds instead. As the system expands, the other subareas will need to do the same when it’s their turn.
    I don’t think Central Link could have been built without some bonding, but it’s certainly possible that it could have been done with less bonding if all five subarea’s revenues could have been focused on the project – with the understanding that they would be similarly focused on the other subareas later. If that was possible, then the cost of the program could have been cut dramatically, and the system could have been finished far sooner.
    I don’t have access to numbers I’d need to figure out what could have been possible or how much money it could have saved to use less financing. I don’t have a problem with bond financing, but I think one needs to take a deep breath before taking a $billion dollar project and making it a $2billion project, just to get it done a couple of years sooner, or to make up for lack of trust between elected officials. In either case, it’s worth asking whether there are alternatives to adding billions of additional cost simply because policymakers can’t reach agreement or wait a couple more years for results.

  49. Nathanael December 21, 2010 at 10:15 pm #

    I have read several analyses excoriating the bidding system, particularly the one mandated by law in New York State, for encouraging overpricing. I don’t quite understand the analyses myself so they may be completely wrong.
    Financing costs obviously will inflate prices massively. There is also a lot of evidence that estimates are being massively padded these days (100% contingency funds?!?).
    But that doesn’t account for the actual, after-the-fact, financing-excluded prices, which *still* seem too high in the New York area at least. So perhaps it is the bidding system.

  50. Scott December 28, 2010 at 7:50 pm #

    Sorry if I sound ignorant here but I did not think that engineering and architectural costs to projects were even in the same order of magnitude as construction costs.
    If that is correct then the high cost of construction is either due to:
    1) high material costs
    2) higher labor wages
    3) re-do’s (oops)
    4) ROW acquisition (which someone already said was not that expensive)
    5) more time spent constructing over the same unit length
    6) the reported costs are incomparable (i.e. costs for different projects omit and include different “costs”)
    I’m going to have to say that #5&6 seem most likely. I’m not claiming to know though. Just something to think about.

  51. CroMagnon December 29, 2010 at 2:30 pm #

    With regards to NYC: My understanding was that large portions of the IRT were built just below the surface by removing soft earth while sitting atop the bedrock. Obviously there are exceptions, like the dip in bedrock between Midtown and Downtown.
    The other reason why NYC would be expensive is that most was done cut-and-cover and as a result is surrounded by utilities. And all utilities in NYC are underground. Most new tunnel projects use deep boring machines well below any utilities.
    I also think solid rock is better and cheaper to tunnel through than soft earth for stabilities issues and the likelyhood that more things will go wrong with soft earth. However, assuming nothing is going wrong, I’ve heard tunneling through soft earth proceeds faster than solid rock. And the cost of underground stations in solid rock requires blasting a vertical clearance.
    Nevertheless, the risks would seem to be lower with rock, which should lower costs. There is the added advantage of easily being able to tunnel directly beneath built-up property that is difficult if not dangerous on soft earth.

  52. Aaron M. Renn December 29, 2010 at 3:18 pm #

    @TransitPlannerMunich, here’s an example of US over-engineering. When the Brown Line L station renovation and expansion project was done in Chicago, these elevated stations had their railings replaced. Reportedly, the design of these railings is such that they would stop a train that de-railed, even though that should not have been a design goal since they are on the rear of the platform and none of the rest of the line has that type of derailment protection. I can’t swear this is actually the case, but I can tell you the railings are made out of extremely heavy steel that is thicker and stronger than any railings meant to protect people I’ve ever seen.

  53. Alon Levy December 29, 2010 at 4:37 pm #

    @CroMagnon: New York’s current projects are all deep bores with TBMs. It’s those projects whose costs are an order of magnitude too high, not the IRT.

  54. CroMagnon December 29, 2010 at 4:45 pm #

    So they’re bores, but are you sure they are deep? Second Ave has near continuous utility excavations along the route and certain buildings had to be vacated during certain phases of construction. Also, wouldn’t the SAS have to be close to the surface to merge with the existing system on its south end?

  55. Alon Levy December 29, 2010 at 5:37 pm #

    The bottom of SAS is 70′ deep, if I remember correctly.
    But deep-level construction doesn’t save money. On the contrary, it makes stations more expensive. In Japan they cite the necessity of deep construction as the main reason why costs are so high there; in Spain they build as close to the surface as possible, to save money.

  56. CroMagnon December 29, 2010 at 5:50 pm #

    OK, so it’s still really unclear what makes the SAS so expensive. Other US cities have done deep boring with stations at a small fraction of the cost. Was it something about navigating between other train tunnels?

  57. Alon Levy December 29, 2010 at 7:11 pm #

    If I had real answers, I wouldn’t be commenting on blogs. I remember reading on SAS but can’t find the link anymore that the MTA is working on a report on how to reduce construction costs. But it can’t be just tunneling because the Hudson-Bergen Light Rail was very expensive to construct, too: $2.2 billion for 33 km, of which 27 are in existing railroad rights-of-way and 6 are in downtown streetcar mode (1-1.5 elevated, the rest at-grade).

  58. Rob December 30, 2010 at 6:24 pm #

    I think the important question isn’t a project-specific one. Someone should be able to explain why transportation infrastructure construction inflation suddenly started skyrocketing nationally and across the board about 20 years ago, regardless of mode or profile. For one thing, I’d like to know whether that period has now ended. It’s weird to have so little light shed on such a widespread phenomenon.