Sorry Eno, the US Really Has a Construction Cost Premium

There’s a study by Eno looking at urban rail construction costs, comparing the US to Europe. When it came out last month I was asked to post about it, and after some Patreon polling in which other posts ranked ahead, here it goes. In short: the study has some interesting analysis of the American cost premium, but suffers from some shortcomings, particularly with the comprehensiveness of the non-American data. Moreover, while most of the analysis in the body of the study is solid, the executive summary-level analysis is incorrect. Streetsblog got a quote from Eno saying there is no US premium, and on a panel at Tri-State a week ago T4A’s Beth Osborne cited the same study to say that the US isn’t so bad by European standards, which is false, and does not follow from the analysis. The reality is that the American cost premium is real and large – larger than Eno thinks, and in particular much larger than the senior managers at Eno who have been feeding these false quotes to the press think.

What’s the study?

Like our research group at Marron, Eno is comparing American urban rail construction costs per kilometer with other projects around the world. Three key differences are notable:

  1. Eno looks at light rail and not just rapid transit. We have included a smattering of projects that are called light rail but are predominantly rapid transit, such as Stadtbahns, the Green Line Extension in Boston, and surface portions of some regional rail lines (e.g. in Turkey), but the vast majority of our database is full rapid transit, mostly underground and not elevated. This means that Eno has a mostly complete database for American urban rail, which is by construction length mostly light rail and not subways, whereas we have gaps in the United States.
  2. Eno only compares the United States with other Western countries, on the grounds that they are the most similar. There is a fair amount of Canada in their database, one Australian line, and a lot of Europe, but no high-income Asia at all. Nor do they look at developing countries, or even upper-middle-income ones like Turkey.
  3. Eno’s database in Europe is incomplete. In particular, it looks by country, including lines in Britain, Spain, Italy, Germany, Austria, the Netherlands, and France, but even there it has coverage gaps, and there is no Switzerland, little Scandinavia (in particular, no ongoing Stockholm subway expansion), and no Eastern Europe.

The analysis is similar to ours, i.e. they look at average costs per km controlling for how much of the line is underground. They include one additional unit of analysis that we don’t, which is station spacing; ex ante one expects closer station spacing to correlate with higher costs, since stations are a significant chunk of the cost and this is especially notable for very expensive projects.

The main finding in the Eno study is that the US has a significant cost premium over Europe and Canada. The key here is figure 5 on takeaway 4. All costs are in millions of PPP dollars per kilometer.

Tunnel proportionMedian US costMedian non-US cost

However, the study lowballs the US premium in two distinct ways: poor regression use, and upward bias of non-US data.

Regression and costs

The quotes saying the US has no cost premium over Europe come from takeaways 2 and 3. Those are regression analyses comparing cost per km to the tunnel proportion (takeaway 3) or at-grade proportion (takeaway 2). There are two separate regression lines for each of the two takeaways, one looking at US projects and one at non-US ones. In both cases, the American regression line is well over the European-and-Canadian line for tunneled projects but the lines intersect roughly when the line goes to 0% underground. This leads to the conclusion that the US has no premium over Europe for light rail projects. Moreover, because the US has outliers in New York, the study concludes that there is no US premium outside New York. Unfortunately, these conclusions are both false.

The reason the regression lines intersect is that regression is a linear technique. The best fit line for the US construction cost per km relative to tunnel proportion has a y-intercept that is similar to the best fit line for Europe. However, visual inspection of the scattergram in takeaway 3 shows that at 0% underground, most US projects are somewhat more expensive than most European projects; this is confirmed in takeaway 4. All this means that the US has an unusually large premium for tunneled projects, driven by the fact that the highest-cost part of the US, New York, builds fully-underground subways and not els or light rail. If instead of Second Avenue Subway and the 7 extension New York had built high-cost els, for example the plans for a PATH extension to Newark Airport, then a regression line would show a large US premium for elevated projects but not so much for tunnels.

I tag this post “good/interesting studies” and not just “shoddy studies” because the inclusion of takeaway 4 makes this clear: there is a US premium for light rail, it’s just smaller than for subways, and then regression analysis can falsely make this premium disappear. This is an error, but an interesting one, and I urge people who use statistics and data science to study the difference between takeaways 2 and 3 and takeaway 4 carefully, to avoid making the same error in their own work.

Upward bias

Eno has a link to its dataset, from which one can see which projects are included. It’s notable that Eno is comprehensive within the United States, but not in Europe. Unfortunately, this introduces a bias into the data, because it’s easier to find information about expensive projects than about cheap ones. Big projects are covered in the media, especially if there are cost overruns to report. There is also a big-city premium because it’s more complicated to build line 14 of a metro system than to build line 1, and this likewise biases incomplete data because it’s easier to find what goes on in Paris than to find what goes on in a sleepy provincial town like Besançon. Yonah Freemark thankfully has good coverage of France and includes low-cost Besançon, but Eno does not – its French light rail database is heavy on Paris and has big gaps in the provinces. French Wikipedia in fact has a list, and all of the listed systems, which are provincial, have lower costs than Paris.

There is also no coverage of German tramways; we don’t have such coverage either, since there are many small projects and they’re in small cities like Bielefeld, but my understanding is that they are not very expensive. Traditionally German rail advocates held the cost of a tramway to be €10 million/km, which is clearly too low for the 2010s, but it should lower the median cost compared to the Paris-heavy, Britain-heavy Eno database.


    • Alon Levy

      It excludes some higher US numbers too, like the just-announced cost overruns on Ballard/West Seattle, currently up to $650m/km for a 30% underground line.

      • Gok (@Gok)

        Ok but of the 67 US projects from the Eno dataset, yours has 6, and they are 6 of the 8 the most expensive US projects they found.

          • Herbert

            Will there be a database focusing on light rail any time soon?

          • Herbert


            There is frequently an argument made “Let’s build light rail instead of rapid transit because it is cheaper”…

          • Henry Miller

            Alon does seem to have a bias to underground. Which to be fair is the only sane thing to do for dense downtown type areas that he also has a bias to. (elevated sometimes can sometimes work in these areas, but it is a second choice)

            Alon does like at grade rail for between city service. however he doesn’t have much use for at grade within a city. At grade within a city makes sense for suburban service (if you can get the land/right of way – not trivial), but often such lines cannot get enough ridership to be worth it even at the lower costs, while a subway downtown can get such ridership despite much higher costs (assume reasonable costs).

            There is also the time/effort limit. Alon’s team is only a few people, they can only do so much. It makes sense to limit his study to underground and get good information that try to cover everything and never finish. If you want to throw a few million dollars his way I’m sure he will be glad to start a new study on light rail type service.

          • Herbert

            Well, there was a lot of brouhaha in German media a few weeks back about the alleged negative CO2 impact of subway construction which is supposedly never recouped.

            And for the half a million people sized cities in Germany, subways don’t really make all that much sense… I think in general there’s more light rail yet to be built than subways…

          • Alon Levy

            It’s the same kind of dumb analysis as when EELV wants to slow all TGVs to 200 km/h to reduce air resistance. It sounds green but it encourages people to bolt for cars and planes.

            Likewise, U-Bahn construction has to be analyzed as reduction in car traffic, not relative to the present, but relative to a future in which people are richer and buy cars because transit isn’t any faster; this is the same mistake made by Americans who estimate modal split trends and claim that the Washington Metro or Portland light rail system hasn’t increased public transport usage because the modal split is the same now as in the 1970s, where cities that did not build such systems have seen their modal split collapse and car usage rise.

            This is also equally true of development – the carbon footprint of a concrete high-rise may look higher than that of a wooden house, but between lower heating costs and lower car usage thanks to high density, it’s much greener.

          • Herbert

            In the real world high rise buildings are usually built spread so far apart that they end up *less* dense than 19th century urban construction…

            Also, as some in the comments point out, if one *only* looks at CO2 buses and cars getting more electric actually makes the recouping less likely…

            And it’s not like subways have such a huge “bonus” over light rail in places where a subway is marginal – but both have a “bonus” over buses in terms of ridership…

          • Alon Levy

            33 km/h subways have a big bonus over 18 km/h light rail. It’s hard to quantify because it’s uncommon to replace a tramway with a subway today, but you can compare bus-to-subway and bus-to-tramway railstitutions, and bus-to-subway tends to get you way more ridership increase. Second Avenue Subway is only on quarter done and it’s supposed to gt 200,000 riders per work day (less now because of corona, but give New York a few years to recover), where the bus on First and Second Avenues got 50,000 before it opened; the ridership projection for if the entirety of SAS opens from Harlem to Lower Manhattan is 500,000. Part of it is the connection to Times Square, but that is not enough to explain a factor of 10 difference.

            And high-rise buildings are built far apart in the context of postwar residential towers-in-a-park projects like Gropiusstadt and Märkisches Viertel. But if you have organic towers-on-a-base development, you do get higher density – the Upper East and West Sides of Manhattan have the world’s highest built-up density, and near-tie for highest population density in the first world, only falling behind poor neighborhoods that achieve high density through overcrowding. The 15-story buildings on the main streets of Taipei provide plenty of density too, as do the growing high-rises of Tokyo. Commercial high-rise towers concentrate density too, and that’s a good thing – those Sony Center jobs are helping fill trains.

          • michaelrjames

            @Alon, Herbert …

            But if you have organic towers-on-a-base development, you do get higher density – the Upper East and West Sides of Manhattan have the world’s highest built-up density, and near-tie for highest population density in the first world, only falling behind poor neighborhoods that achieve high density through overcrowding.

            There is no question that going high improves the efficiency of all those things. But the reason we keep having the same old arguments on the topic is that the gain in efficiencies for all criteria–density, energy/enviro efficiency, transit-supporting–plateaus occurs at quite low heights. In fact it may plateau at very low levels, say 4 floors–which is the ‘missing middle’ range and is missing because it can’t support a few other necessary elements such as a developer’s ROI given the cost of the land. In terms of energy it gets worse once you exceed low-rise, both for the body corporate and the individual owners. For the building (body corporate) it is because elevators become the major recurring expense both in terms of power but also in maintenance and insurance. For individual owners it is largely because hi-rise these days consists of 100% curtain-wall glass and no matter how fancy the hi-tech glass or super-expensive double/triple glazing it requires more heating in winter and cooling in summer, plus the higher you go the more dependency on mechanical ventilation (usually few windows are openable).

            Cost of construction passes thru several thresholds due to engineering criteria such as the need for deeper and stronger foundations, hi-performance facades, space devoted to services, not to mention hugely expensive excavation if subterranean parking is to be provided. Or in those ultra-talls an entire floor in about every 12, is left unused (except for services–reducing vertical ducting overall) and left open to reduce wind vortex loading (and thus the need for bulkier engineering to counter building sway). Clearly at some point it no longer supports a case for building high, for which Alon’s example of a few neighbourhoods in Manhattan prove the case: ultra-talls and developments like Hudson Yards do absolutely nothing for population density and worse, nothing at all for providing affordable housing even for the above-average-income Manhattanites. Indeed those buildings remain seriously under-occupied because they are so expensive the only people who buy them are the kind of can afford to leave them empty (and they have such properties around the world). Indeed the creators of such buildings have admitted that they only work well with occupancy of about one in three occupants (who start complaining if elevators take too long to arrive, a combination of too many others competing for them and the sheer vertical spans involved; and of course the huge inefficiency of adding an extra elevator shaft).

            Alon’s claims about Manhattan are only marginally true:
            NYC Upper East Side: 219,920 (2010) on 5.23km2: 42,100/km2
            NYC Upper West Side: 214,744 (2010) on 4.9 km2: 43,800/km2
            Paris-11th 152,500 (2005) on 3.7 km2 for 41,600/km2

            The point, of course, is that while building high, and forever higher, may achieve high density but actually not by much at all; and building forever higher actually decreases effective density while worsening affordability. Paris’ 11th arrondissement is materially no less dense than those densest areas of Manhattan (and comparable in land area too) yet the only structures that exceed 6-7 floors are a few church spires and maybe the tippy-top of the olive branch held aloft by Marianne in Place de la Republique. As Alon themselves admit, those are the densest zones in the rich world so clearly in the vast majority of cases the opposite is the case: hi-rise is lower density than lower-rise dense zones as found in most Euro cities. Just to pre-empt Alon’s usual response that the 11th is nothing but a low-quality slum housing the impoverished and students with no choice, the counter is that there is no need at all for these kinds of densities and the biggest arrondissement (and most middle-class residential) in Paris, the 15th has 232,000/8.5 km2 at 27,300/km2. In reality anything above very approximately 15,000/km2 is as good as the higher densities and will have commensurately better urbanism.

            But this is not to obsess about arcane issues. Here’s the crux: when discussing the pros and cons of building typologies, we should consider the minimum required to achieve the goals which in the context of this blog are density,, affordability and supporting transit and justifying the higher cost of Alon’s fave type. On those bases it is clear that the minimum threshold is quite low, approximately 5 to 7 floors. No accident that that corresponds to Haussmannian Paris but also California’s failed (so far) SB827 and subsequent SB50 to facilitate transit-friendly and affordable housing.

          • michaelrjames

            @Alon and all P. Observers:

            Let me take the opportunity while on this topic, to bring to your notice the following:

            Knowledge Quest: Submit your example of a great high density environment
            07 Jan 2021.

            Do you know a great example of high density living environments built within the last 30 years? Share your knowledge and contribute to the creation of an open repository via Crowd Creation. To be truly exemplary, the area should include a mixture of functions (at least some of them high-rise) where the physical fabric retains a human scale at street level despite the high density.

            Submissions close soon, on the 29th Jan.
            Browsing the current submissions rather works counter to the desire of the people behind this project: too many of these are too high! Some are quite oppressive, and seemingly the authors and photographers are oblivious to this. Alas, their criteria seems to have eliminated any low-to-medium rise modern developments but that is entirely counter-productive against their objective. The academic architects and urbanists at Amsterdam University of Applied Sciences are interested in creating a resource to inform the ongoing densification of Amsterdam, but they appear to have totally loaded the dice towards serious hi-rise. I admire their concerns but they almost totally ignore density (which, blah blah … see my previous response to Alon). There are plenty of low-rise developments that exceed density requirements in Paris in ‘modern’ times, eg. dozens of ZACs since the late 70s. In the light of the mostly poor quality (urbanism-wise) of their submissions I was hoping to submit one for the area I live in; though it has exceeded it original design criteria (to keep within the existing fabric which is defined by the adjoining heritage buildings of about 6 floors) is still better than most of those existing submissions, but unfortunately my ancient laptop (newest supported OS is 10.7) has trouble with their submission webpage. So far I haven’t summoned the energy or will to attempt it on my phone.
            I encourage POers to make a submission. There are plenty of better lower-rise developments that otherwise meet the requirements, in Berlin, Asian megacities, heck even London or Brooklyn.

          • Henry Miller

            Mode share is interesting, but it isn’t a useful metric. Mode share doesn’t guide any decisions. You can use it to say my city is better if you can’t find a different way to compare your self and still want a score, but that isn’t useful.

            Total number of riders (in a time period) is useful. This tells you how busy/popular you are. This should be combined with unique riders to say how useful your system is for those who ride it – are they using it for one task, or many different ones.

            Total traffic related deaths/injuries is useful. Having lost a good friend to a traffic death this is the one I personally care about the most. This tells you how safe your city is, and when you drill down you can find where/why it is unsafe and work on solutions to make your city safe.

            Profit/loss of an individual line or the system is useful. You need to be careful to figure out what the subsidies are before measuring this. If a disabled person uses the line, and it would cost $10,000 to serve the disabled person by other means, then add $10,000 to it. Likewise if students use it, then you need to figure out what the alternative student transportation would cost, and also the marketing because students may keep using it at full price as adults. Then we need to figure out how the unprofitable line makes the whole better and so might be worth it. Still it is a useful metric because you can make decisions based on it, even if it often is used against transit advocates.

            I’m sure there are more metrics that are useful that I don’t know. Mode share isn’t one.

          • RossB

            In Seattle, the light rail system is essentially a subway. There are huge sections that are underground, a lot that is above ground, and very little on the surface. I’m sure if it was heavy rail it would have cost and operated much the same.

            This is another example of why case studies are really the best way to make a comparison. It takes longer, but otherwise you are making an apples to oranges comparison. Either way, I think that ridership should be included, and will show the other side of the coin. Often, because of cost pressure, an agency will cut corners where it actually matters, and build crap. If you are just looking at cost, things look OK, but ridership suffers. Of course there are lots of other things that determine ridership, but if your ridership per dollar costs aren’t very good, then you are either spending too much, or building something poor (or both).

            I would love to see a chart of ridership per dollar spent not only with U. S. projects, but worldwide.

          • Herbert

            I can think of a few historical trams being replaced by subways. I think the most clear one would be in Nuremberg where the old tram line to Ziegelstein was replaced by U2. U2 may be marginally faster, yes, but overall if your trips do not happen to align with the specific lines, in Nuremberg you have a lot of a transfer penalty (And I think you know better than I do that the transfer penalty is more than just the extra time it takes). To say nothing of the fact that in Nuremberg interchanges between tram and subway or between S-Bahn and Subway are usually incredibly poorly designed.

            If you say the problem is 14 km/h trams, there seems to be a simple solution: Don’t build 14 km/h trams, speed them up. But then you run into a similar issue that makes all German “full” subways have a distance between station well below 1km: While the vehicles may get faster, the people who ride them will have to walk further and ultimately their trips get longer. Of course for every mode that is not at-grade you have to add the time it takes to get down or up there…

            When Nuremberg extended Tram line 4 to “Am Wegfeld” they initially wanted the tram to stop less often than the previous buses. Ultimately they decided against it due to concerns of people living there media coverage indicates ridership even exceeded the prognoses.

            And there is another problem with subways: While there may be (there are actually no studies that got as far as calculating a benefit cost quotient) subways in Berlin yet to be built that can get federal funds, in Nuremberg it is pretty obvious that all of those have been built or will be built by 2025 (the extension to “Gebersdorf” by the way a partial replacement of a tram line shut down decades ago when U2 sorta kinda maybe got in the general vicinity). An extension of U2 to Stein is not happening any time soon.

            On the other hand, there are numerous tram projects that could be started with federal funds as soon as possible – including the 25 km extension via Erlangen to Herzogenaurach….

            And as some user in some forum likes to say “Eine Straßenbahn ist besser als keine U-Bahn”

          • Sascha Claus

            Second Avenue Subway is only on quarter done and it’s supposed to gt 200,000 riders per work day […], where the bus on First and Second Avenues got 50,000 before it opened; […]

            What was the average speed of these buses? One would expect the passenger growth to be bigger if the speed gain of the rail line compared to the railstituted bus route(s) is bigger.

          • adirondacker12800

            Second Avenue Subway is only on quarter done and it’s supposed to gt 200,000 riders per work day […], where the bus on First and Second Avenues got 50,000 before it opened
            They allow hoi polloi east of Third Avenue to walk to Lexington and use the subway there. They haven’t stopped running the buses. How many were using the buses after the Second Avenue line opened?

          • Eric2

            Re stop spacing (on any form of transit): I think a good rule of thumb is that people will generally not tolerate more than about 20 intermediate stops on their trip. Why such a uniform number? Because not only does each stop have a time penalty, but too much acceleration and deceleration becomes nauseating. I think a look at transit lines across many geographic scales will show that 20 stops to downtown is about the longest service run in practice.

            Of course if you have surface light rail which has to stop at traffic lights, the maximum is somewhat less than 20, let’s say 15. But for any reasonable light rail stop spacing, 15 stops is enough to get from Midtown to anywhere in Manhattan. So light rail could work well in Manhattan. But outside Manhattan, light rail times to Manhattan would be too long. As for light rail within an outer borough – most passengers would just be transferring to the subway. Better to improve buses and add a few strategic subway extensions in places like Utica to handle the diffuse market of people transferring to the subway.

          • Eric2

            BTW what exactly is the problem with glass towers? Perhaps the biggest geometric constraint on density is the amount of sunlight reaching each apartment, but glass walls are the perfect method to minimize waste of sunlight. As for towers vs Hausmannian, who really cares, maybe let builders build whatever potential residents want? But if you insist on Hausmannian, it can be Hausmannian with glass walls.

          • Herbert

            Glass is a terrible facade material. Among other things it creates a greenhouse effect. And if it is shaped wrong, it can literally melt stuff…

            Also, if you look at Berlin, people who have long commutes by and large use the S-Bahn. The trams are for shorter trips. So where does the U-Bahn come in?

            New York City needs an S-Bahn…

          • Sascha Claus

            Also, if you look at Berlin, people who have long commutes by and large use the S-Bahn./blockquote>Might be because the S-Bahn runs further out than the U-Bahn?

            The trams are for shorter trips. So where does the U-Bahn come in?

            In places where there is neither an existing mainline railway that could be equipped with additional (third-rail-powered) S-Bahn tracks nor tunnel through the city centre?
            The main difference is problably not in their tasks, but in governing laws (EBO vs. BOStrab) and operators (state railway vs. city transit). A main difference in tasks is between ‘proper’ U-Bahns (fully separated, 3rd-rail-powered) and Stadtbahns (that tend to call themselves U-Bahn).

          • Eric2

            Greenhouse effects are very helpful in winter. In summer, you can cool off with the excess daytime solar power we will soon have. Or perhaps make the outer side of your summer curtains be shiny and reflective.

            Glass only “melts stuff” when it forms a concave mirror – apart from a few prima donna starchitects, nobody designs their glass facades to form concave mirrors.

            In Berlin, trams are for East Berlin, U-Bahn for West Berlin. This is a historical accident, but it does mean there’s little duplication of infrastructure, and each can serve much the same market in different parts of the city. In general it seems tram stops are good every 500m while metro stops are good every km. So there is a distance range from downtown that is really too far for trams but good for metro. In east Berlin, I am guessing people in that range take the tram a few stops and transfer to S-Bahn…

            Berlin S-Bahn stop spacing seems to be every 1-1.5km in built up areas, which is comparable to the NYC subway express lines.

          • michaelrjames

            @Eric2 “Greenhouse effects are very helpful in winter. In summer, you can cool off with the excess daytime solar power we will soon have. Or perhaps make the outer side of your summer curtains be shiny and reflective.”

            Herbert is 100% correct. The problem with your greenhouse effect is that it doesn’t ever operate when it would be most needed: nighttime, which is also much longer in winter. And we (the tropics and Mediterranean Europe) solved the issue centuries ago and it wasn’t curtains which still trap most of the incident heat inside the house and block breezes, whereas shutters stop it ever getting inside as well as providing for ventilation without compromising privacy or security. Are we really going to be so stupid as to waste solar power on cooling houses instead of using centuries-old passive design? Yes, of course we will, especially in the Anglosphere. Australia is one of the most insolated places in the world, certainly the rich world, yet if we use window shutters at all we put them on the inside! Plus your developer buddies have managed to get building code changed to eliminate the previous requirement for eaves to provide shading–purely to allow them to build houses filling most of the tiny blocks, right up against the next house; and I suppose to save a tiny bit on roof structure. (There are moves to retract this regressively, monumentally stupid legislation but the developers hold huge sway over our politicians.)

          • adirondacker12800

            15 stops is enough to get from Midtown to anywhere in Manhattan.
            Why would I take a slow streetcar when there is fast subway? If my origin is an express stop and my destination is an express stop it can be even faster. You do understand that Manhattan has lots of subway lines and most of them have local and express service?

          • Eric2

            A baseline for light rail ridership is current bus ridership, which is much less than subway ridership, but still substantial.

          • Eric2

            So build high rises with glass windows and shutters.

            But more generally, I don’t get the obsession with saving a few dollars on heating/cooling costs. We are talking about units that share most of their surface area with other units, not with the outdoors. Heating/cooling costs are marginal. Sunlight is a precious commodity though. Dense housing should be designed around sunlight allocation, not heating/cooling costs.

          • michaelrjames

            @Eric2: “Dense housing should be designed around sunlight allocation, ”

            They are, and have been in Paris since the 17th century and updated/regularised by Haussmann in mid-19th century, I keep telling you that it is one of the factors in why visitors find Paris so harmonious. Its human scale, light and openness. And living in those buildings you are anchored to the ground in a way you lose with just a few more floors. These are fundamental natural relationships, and thus timeless and not really subjective though people may have differing ‘opinions’ on the issue it’s not clear such opinion is informed or merely ideological (for example, like an econocrat placing economic ‘efficiency’ above all else). It’s hi-rise that breaks all the rules.

            Re heating and cooling, primarily it is not about cost, at least not in dollar and cents. However you underestimate it because heating and cooling accounts for a huge chunk of domestic power bills.

            Philip Johnson, after building his own Glass House in New Canaan, Connecticut, had this bit of advice to offer others in the field: “Don’t build a glass house if you’re worried about saving money on heating.”

            Incidentally the vast majority of domestic solar-PV installations will not support house heating and cooling (except simple extraction fans)–if you want that you need a much bigger solar installation if you’ve got the insolation to support it, and much bigger batteries.
            Then there is the impact on overall energy generation which is huge. It’s the main factor in why NYC uses 4,696 kWh per household per year compared to Dallas household average at 16,116 kWh more than three times as much. Obviously a big factor in contributions to global warming.
            Then there is the fact that house climate control by natural ventilation is superior, however glass towers are not built for it and passive ventilation won’t be enough to counter the oven heating from all that glass. In fact I am a fan of brise soleil on high-rise, ie. vertical style, and sun-tracking ones even better to enable optimising during each day and seasonal effects. But developers of buildings–except for a few showpony buildings*–don’t want to spend the money on these thing, because after all they won’t be paying the heating/cooling bills or have to worry about all the problems created by their glass ovens.
            [*examples are Renzo Piano’s NY Times building in NYC; Norman Foster’s St Mary Axe, aka Gherkin, London]

            Everyone needs to stop building giant glass skyscrapers right now
            Towering glass structures are an environmental nightmare. And there’s a growing consensus that we should stop building them
            By SOPHIA EPSTEIN, 11 Nov 2019

            Any building playing host to hundreds of people is going to have a huge climate footprint, but the glass is particularly problematic. The sunlight has unlimited access into the building, but no way to get out. “With an all glass building, you’re fighting the environment rather than working with it,” says Simon Sturgis, who is an adviser to the government as well as chairman of the Royal Institute of British Architects sustainability group. Conventional glass skyscrapers are just tall green houses. The heat inside can’t escape because the whole structure is wrapped in a glass skin. That’s great for tomatoes, but for people it just means more air conditioning.
            The amount of energy used for cooling has more than doubled since 2000, and it will double again by 2040 if we don’t curb our reliance on air conditioning, according to the International Energy Agency. “Even in a moderate climate, cooling issues are becoming more and more severe, and the proliferation of glass buildings is accentuating that fact,” says Henrik Schoenefeldt, senior lecturer in sustainable architecture at the University of Kent.
            It’s a vicious cycle: we build a glass skyscraper, then have to cool that glass skyscraper, which uses energy, which contributes to the climate crisis, which increases the temperature. The hotter weather makes the glass building even harder to cool, but we have to keep cooling it because sweaty co-workers are not happy co-workers, and so the cycle goes on.
            This problem can’t be solved by just turning off the AC. “We don’t want people to sweat in overheated buildings, that’s not the point,” Ürge-Vorsatz says. “The point is that with responsible architecture you can keep people cool without unnecessary cooling.” We need to build better buildings.

            Glass skyscrapers need to be seriously re-examined
            Shane Reiner-Roth, May 21 2019.

            How glass skyscrapers took over the world–and why we need to stop building them
            May 17, 2019

            NYC’s ‘Green New Deal’ to ban glass, steel skyscrapers
            By Kim Slowey, April 24, 2019

            Could the era of glass skyscrapers be over?
            Published 27 May 2014

          • Eric2

            Why should anyone care about your feelings about what is “harmonious”, “anchored”, etc? Let developers build whatever they want. If there’s demand for Hausmannian buildings then developers will build them. If not, then apparently your feelings are idiosyncratic and shouldn’t drive public policy.

            The main reasons Dallas emits more CO2 than NYC are that NYC units have much smaller outside surface area (smaller units to begin with, and sharing walls with other units of equal temperature), and that cars are driven less and for shorter distances in NYC. Not because insulation is better in NYC – it’s probably worse on average, since the buildings are older.

          • michaelrjames


            You see what you just did? You totally ignored the evidence I presented; other expert people’s evidence. You poked your fingers into ears and yelled la-la-la-la …

            Why should anyone care about your feelings about what is “harmonious”, “anchored”, etc?

            Not me, but countless millions going back to the 17th century right up to now. In fact NYC in the late 19th. All of this attests to the fact that by whatever means the city had locked in some fundamentals that appealed to the way our brains are wired.

            Let developers build whatever they want.

            No thanks. NYC has gone thru that phase in the 19th century and it led to untold misery and actual plagues which is what finally brought legislative action to outlaw what they really wanted to build. Especially today when property prices are untenable and there is immense pressure to provide more housing in prime cities like NYC, SF etc. In fact it is close to your apparent ideal: those ultra-talls, or even Hudson Yards, are doing absolutely zilch to solve problems for ordinary Manhattanites but developers are making money selling to mostly non-residents who leave their trophies empty 95% of the time. London under Boris has done exactly the same thing–given the developers exactly what they want with the same useless result.

            Your responses–or lack of reasoned response–leads me back to the improbable theory that you are the second son (Eric2) of that infamous developer turned politician turned treasonous disaster. Go on Eric, wear your MAGA hat with pride ….

  1. df1982

    Streetcar lines in non-mass transit cities in the US often seem to end up pretty cheap, even by European standards (in established transit cities like DC this is of course not the case). Why is this? Shouldn’t they suffer from the same problems of cost overruns, project mismanagement, etc.? This would support Eno’s thesis that the problems really kick in when tunnelling and underground stations are involved.

    • Alon Levy

      When you say streetcar lines, you mean things that are called streetcars in the US, so the Portland Streetcar, the KC Streetcar, etc., and not light rail, right?

      • df1982

        Yes, I’m specifically talking about what Americans call streetcar rather than light rail. As I understand it, the former run exclusively on city streets (often but not always in mixed traffic), whereas the latter tend to recycle freight rail alignments in the suburbs, and run either on-street or in a tunnel in the city centre.

        Pretty much every time I see the costs for these in transit deserts like KC, Cincinnati, Detroit, El Paso, etc., I’m pleasantly surprised. And this stands in stark contrast to the experience of building tramways in places like Edinburgh, Sydney and DC (all of which are established transit cities in the Anglosphere), not to mention other rail-based projects in the US. They even stack up well against the costs of French tram systems. So why is this?

        Of course, most of them are useless for transit purposes and have pitifully low ridership, operating dinky little one-way loops around the CBD, but in theory there should be nothing stopping a city from designing a useful line AND building it to the same costs as these streetcar circulators.

        • df1982

          OK, here’s a little database drawn from Wiki (so caveats apply), kilometres are full dual track equivalents, since a lot of them are single track loops:

          Salt Lake City: $55m for 2.2km (2013), $25m per km
          Tucson: $196m for 6.3km (2014), $31.1m per km
          Atlanta: $90m for 2.15km (2015), $41.86m per km
          Dallas: $47m for 1.97km (2015), $23.86m per km
          KC: $102m for 3.5km (2015), $29.14m per km
          Cincinnati: $102m for 2.9km (2016), $35.17m per km
          Detroit: $137m for 5.3km (2016), $25.84m per km
          El Paso (heritage): $97m for 3.85km (2018), $25.19m per km
          Milwaukee: $124m for 3.4km (2018), $36.47m per km
          Oklahoma: $136m for 4.5km (2019), $30.22m per km
          KC extension (planned) $227m for 6km (2023), $37.83m per km

          So they all seem to come within the $25-$40m per km, and this generally includes stabling and rolling stock. Pretty comparable to the German and Scandinavian costs given by BindingExport below.

          • Eric2

            Is it possible that because these streetcars use little bus-sized vehicles, they are less heavy that regular LRT vehicles and require less utility relocation than a full LRT line?

          • Herbert

            You want to move utilities out of the way of tram tracks no matter their weight.

            French trams are often “an opportunity for city beautification” which can make costs opaque in both ways because on the one hand “costs that the tram didn’t cause” may end up in the tram’s accounting and on the other hand some sources will over-correct and try to explain away tram related costs as “city beautification”

          • Herbert

            Also, those U.S. style streetcars are a folly that I think hasn’t even spread to other Anglosphere countries.

            While the German GVFG has been changed recently to allow street bound track to receive federal funding, it’s still strongly incentivized by the way the benefit cost quotient values trip times to reduce shared right of way

          • df1982

            What was the real cost killer in Sydney (AU$3b for a 12km tram line, possibly a world record) was utility relocation in the CBD, which was exacerbated by the ridiculously over-engineered concrete base for the track. Maybe the American streetcars are built to lightweight standards that lessen the need for utility relocations. The fact that the costs in different cities are so analogous suggests that a cookie-cutter approach for streetcar construction has developed there.

            But probably the best explanation is the rent-seeking argument. These are all in states that are controlled either by Republicans or fiscally conservative Democrats, so the tolerance for grandiose public infrastructure projects is really low (even some of the liens were extremely controversial, e.g. Cincinnati, despite costing so little). Which means that either a line is built for the absolute bare minimum it could cost or it doesn’t get built at all.

            By contrast in DC a streetcar project can get bogged down in mismanagement, delays, corruption and gold-plating, and still find the political will to be built.

          • Alon Levy

            On the other hand, you see high light rail costs in red states like Texas and North Carolina. And with a lot of the streetcars specifically, there are no dedicated lanes, no stations, etc. – it’s like 1952 all over again. (Not that it’s stopped H Street from being more expensive, granted.)

          • michaelrjames

            @df1982 “What was the real cost killer in Sydney (AU$3b for a 12km tram line, possibly a world record) was utility relocation in the CBD, which was exacerbated by the ridiculously over-engineered concrete base for the track.”

            Are you really convinced that that was the real reason? It was the publicised reason–by government, so immediately suspect. And remember this only blew up as a cost issue after the tramway was almost complete, so it implies that the contractor was ignorant of these issues–this was specifically in George street, Sydney’s major thoroughfare in Sydney CBD, yet the construction began in George street in 2015 while the issue blew up in 2018 as it neared completion. They claim that (someone, possibly the telecoms and other utlities, or maybe the city council?) misled them about the difficulty of the relocation work. Call me a cynic but none of that flies with me. Even if it were all true, it just means that the construction company didn’t do its due diligence, for which it should not be rewarded (the govt ‘settled’ with a $576m gift to the consortium led by Acciona).
            Anyway you give me the opportunity to post again my favourite interpretations on the tramway debacle (and I don’t apologise for multiple postings of this; some people just don’t get it):

            Making heavy work of the light rail task
            Jacob Saulwick, 30 June 2018.

            One theory about what went wrong relates to the way in which the project was contracted. Faruqi (Greens member of parliament), who has a doctorate in engineering, has argued extensively there has been a hollowing out of technical know-how in the public service. The end result is more time and money trying to fix design changes. “I am hugely concerned about the deliberate de-engineering and politicisation of the public sector and the immense over-reliance on outsourcing,” says Faruqi. “This has led to a diminished capability to establish accurate scope and cost in the first place, followed by a lack of capacity to properly scrutinise design, procurement and delivery from private contractors and consultants.”

            From the same article, the usual neoliberal delusion about efficiency etc of the private sector (efficient at extracting greenmail from compliant governments on these PPPs)”

            Martin Locke, adjunct professor at the Institute of Transport and Logistics Studies at the University of Sydney, says there’s logic to the procurement model chosen by Transport for NSW. “If you have an integrated project you can outsource interface risk to the private sector and the private sector takes the responsibility for working out how to build the project, maintain the project, and operate it,” says Locke.

            Sure it does Martin, as long as it extracts an extra billion or so in funds for nothing. Of course it was no accident that this issue was left just before completion, when the govt was at maximal vulnerability (I think it was before an election).

          • Herbert

            I don’t quite understand your implication that there are some circumstances where there are certain utilities you don’t want to move out of the way of your tram. Like… What do you do when the water main under your tram bursts? Or the internet cable needs replacing? Why would you keep utilities under a tram and which would that be?

          • michaelrjames


            Not sure who your question is directed at.
            All tramways in the world will have utilities under them. The issue is whether during installation of a tramway they need to be buried/rebuilt deeper or the roadway/bed has to be reinforced to prevent damage to them, or access points (presumably at the sides) installed etc.
            We’ve covered this issue before on PO. I recall for the construction of Paris T3 on the Boulevards des Marechaux they used prefabricated single-piece trambeds that allowed rapid installation so that cross-street intersections could be done in a single overnight work session without interrupting daytime traffic on the large number of radial arterials carrying heavy traffic in/out of Paris. That could not have involved anything touching utilities which could not possibly have been done overnight. So I assume they could just dig a shallow trench in the road and lay the preformed-concrete trambed so that only a few hours later traffic could roll over it. Of course those boulevards and the cross streets may have had their utilities re-engineered yonks ago, especially those with metro just under them …

            That’s why the drama over George Street in Sydney seems so total melodrama, and probably either completely unnecessary or somehow just a cover story for the politicians and construction consortium. If df1982 knows more, I’d love to hear it.

          • df1982

            Alon, the stops on streetcar lines are often basic, but I’m not aware of any line that has done without stop infrastructure entirely, and generally you’re in favour of a minimalist approach here (e.g. Boston’s Green line extension), as long as there is adequate shelter, seating and passenger information. There’s no reason why a streetcar stop needs more facilities than a decent bus stop.

            And why would having dedicated lanes or not affect the cost of the project? Or does planning for traffic re-routing end up costing a ton?

            It’s true that light rail (in the American sense) tends to have higher costs even in the historically non-transit cities, although there it is more of a mixed bag, and it’s hard to make adequate comparisons since the lines can be underground, street-running, using existing/abandoned freight lines, or forging a new above-ground alignment, with wildly different costs as a result.

            So I would categorise the cost of projects in the US as follows:

            1. streetcars: uniformly good (except DC, but that was a totally botched project)
            2. light rail: mixed
            3. subway: uniformly bad

          • RossB

            “And why would having dedicated lanes or not affect the cost of the project? Or does planning for traffic re-routing end up costing a ton?”

            Planning is part of it, but often the city doesn’t want to lose the lane. So a road that is two lanes each direction remains two lanes each direction, with a streetcar off to the side. This means the road needs to be widened, and this can involve moving utilities (lamp posts, telephone lines, fire hydrants) and this can really add up. Sometimes it isn’t a matter of keeping two lanes each way, but having one lane each way (making the choice much harder). Sometimes there are bike lanes. Sometimes you want to run the train down the middle of the street (as a way to avoid turning traffic). Doing so means making space for center stops. If you decide to go with curbside boarding, then you have to deal with those cars making turns, which means adding special traffic lights. That’s not super expensive, I imagine, but it adds up, and you have more traffic planning to consider. Bigger streetcars also mean bigger stops, as well as special care to make sure the longer streetcar doesn’t block side traffic.

            On the other hand, if you simply run a small streetcar in the middle of mixed traffic (as you would a bus in a low density area) then it is pretty cheap. If you look at the list of streetcars you mentioned, by and large they perform poorly. So, basically, American transit projects are just a bad value. Either they pay a fortune for a very good line (e. g. Second Avenue Subway) or they pay a relatively small amount for something that gets stuck in traffic, and operates no better than a bus (e. g. just about every streetcar in America). Sort of like buying a Ford back in the ’80s — you can spend a little, or a lot, but there is no way you are going to end up with a good value. Buy a Honda.

          • Herbert

            If the streetcar gets stuck in traffic, just ban cars from that lane. Easy as.

  2. BindingExport

    About German Tram Costs

    The Lichtwiesen extension in Darmstadt did cost €20 Million for 1.3 km and is supposed to open in 2020, the “Mainzel-Bahn” in Mainz did cost €90 million for 9.3 km by 2016 but subsequent repair costs for shoddy welding raised this to €100 million. The last major extension in Frankfurt (Friedberger Landstraße) did cost €55 million for 3.5 km in 2007. Line 2 in Ulm did cost €192 million by 2016 for 9.3 km and rolling stock. Stadtbahn Nord in Mannheim did cost €90 Million for 6.4 km in 2016. Stadtbahn Nord in Heilbronn did cost €84.2 million for 3.6 km in 2014.

    The best comparison would be completely new built system in Scandinavia
    A very low-cost example for a completely new built light rail is Tampere in Finnland with €330 million for 16.3 km, rolling stock and depot. Lunds spårväg did cost SEK 1.5 billion for 5.5 km, rolling stock and depot. Odense Letbane did cost DKK 3.3 billion for 14.4 km including rolling stock and depot. Aarhus Letbane did cost DKK 2.4 billion for 12.2 km including rolling stock and depot.

  3. Herbert

    The 25 km is quoted as costing roughly 3*10^8 € which is seen as “in line with expectations and experiences in other cities”. However, it does include a difficult river crossing in a sensitive spot.

    And it’s also not built yet, so that figure may balloon

  4. Kamoro

    So if I’m understanding their use of linear regression correctly, it would be like measuring sunlight levels between 6am and 2pm, drawing a line of best fit, and incorrectly concluding that sunlight peaks at 2pm?

    • Tom M

      I think rather what Alon is describing is that there are two variables in the data. One of the variables (tunnel proportion) being larger in magnitude than the other (US or nonUS project). The larger variable is this obscuring the smaller magnitude dependent on a cursory review. Looking at the scatter gram it does appear that the US projects are not (roughly) equally distributed above and below the regression line as you move left to right along the x axis. Plotting the regression errors I suspect would show an uneven distribution of the errors. One of the key assumptions behind linear regression (for small sample sizes) is that the errors should be normally distributed.

      I’m no stats expert and I’m sure that there are people on the blog here who will quickly correct/clarify the above!

  5. RossB

    Part of the problem with comparing projects is that you are often comparing apples and oranges. There are ways to save money, but they often involve crap. For example, a light rail line that sits in traffic, or a stop that is so far away from anything that no one uses it. The opposite is true as well — giant mezzanines and other amenities that no one cares about. All of these are common in the United States. It would be interesting to see a chart that looked at ridership per dollar spent, especially since that is relatively easy (rider time saved requires more analysis). My guess is that the U. S. ends up at the bottom. We either building stupid rail lines, or spending too much (or both). New York, by the way, would end up looking less like an outlier, and more like a typical U. S. city. Yes, they get a lot of riders, but they have to spend a fortune to get them.

    I realize that there are a lot of things that determine transit ridership (density is a big one). But I think it is commonly accepted (based on numerous studies) that if you run the trains and buses more often, you get higher ridership. So why don’t we? America sucks at building railways. That much is clear. Maybe for now the best thing we can do is just run more buses. Run them so often that it becomes clear that some bus routes really should be rail (just like Vancouver knows damn well that UBC/Broadway should have been rail a long time ago). Maybe then we can stop building crap, and hope that folks will build those essential pieces at a decent price.

    One final note, to get a real comparison, you really have to dig into the details, and you’ve done it. I have no doubt, based on what you’ve written, that we have a major transit construction problem in the United States. Unfortunately, I don’t think it is the worst problem with transit in this country — I think we build too much crap.

    • Henry Miller

      I’ve been saying for a while that we should start by running more buses because as you say it becomes quickly obvious where the train is needed.

      However there is one important detail that makes this not quite as simple as it sounds: water. All major cities are built on/near water because that is the only cheap way to get freight long distances, and so anywhere there is a good port there will be a city to take on (and ship out) the goods. (there are a few exceptions to this rule, but not enough to worry about). Crossing water is expensive, so there is always a lack of roads. A train bridge should be cheaper – in part because a 1-track bridge (and good scheduling) can handle as many people as a rather large car bridge, and you can get more than double that with a two-track bridge if you need to.

      Thus any city with significant water should look closely at can they ease their current traffic problems around water crossings with a train. Even if the train doesn’t make sense otherwise, it is cheaper than the car bridge it is preventing, and because you intentionally run it far any current bridge there are a lot of trips that it is faster than a car even in the least busy traffic times. This needs some study, but it is a point to consider. Beware though, freeway bridges tend to be built near enough to downtown that you often cannot get a lot out of this, so don’t oversell it.

      When there are roads that more or less do what you need for transit, then by all means start running a good bus system to prove that people will ride good service. Then use the real numbers to show what the real cost of a train is. just don’t let the system become too expensive. A train should be cheaper than a bus in the long run not more expensive because of all the additions.

      • Herbert

        A good real world example of this would be Erlangen. The river Regnitz with its wide floodplains only has two crossings near downtown that are open for cars. One to the north of the core city (“Dechsendorfer Damm”) and one to the south (“Büchenbacher Damm”). The former has two lanes in total whereas the latter has four. Recently one of those four lanes was made into a bus lane.

        The plans for the call for a new bridge between the existing crossings only open to trams and buses. This would obviously give a huge boost to transit as it would suddenly become the fastest option along loads of corridors, but alas the floodplains are beloved by the citizenry and ecologically important and a socdem mayor from the seventies stopped then existing plans for a car bridge on environmental grounds…

    • Herbert

      Taking an erstwhile car lane and making it a transit lane is cheap in terms of dollars.

      Mixed traffic isn’t a question of money, but one of politics

      • Henry Miller

        Politics are important. Sometimes you need to compromise which means you agree to a transit plus car bridge just to get something done. Sometimes you can get the train only option through without the car bridge because the cost is affordable. You need to figure this part out yourself and your city. There are times to compromise, and times to stand your ground. I’m not good at politics.

        • Herbert

          “Taking stuff away from cars” is not popular when it is done, but after.

          That’s the crux of it…

      • RossB

        Not always. Sometimes the road is too narrow, and there is no easy alternative, making it a tough choice for a transit mall.

    • RossB

      I understand what you are saying, and I agree. I’m from Seattle, and not only do we have lots of water, but we have lots of big hills that form natural obstacles, and freeways (that form unnatural obstacles). Without a doubt the ability to avoid obstacles — to create a system that enables trips that are faster than driving at noon — plays a huge part in its popularity. But even in Seattle, ridership on the train is similar to what it was with the buses. Here is a transit map of Seattle: On it you can see that the train goes from downtown to the University of Washington, via Capitol Hill. This is by far the most popular section, even though the train also goes south, to the airport (well off of the map). Everyone knew it was going to be popular, just because of the areas it served. So it really didn’t matter if you looked at the bus ridership, or just the neighborhoods (and what they connect) — it was clear this was going to be success.

      But consider my original point. There is only one stop between downtown and the university. While this stop (Capitol Hill) is a great stop, there should be more. The neighborhood of First Hill has good density ( and major hospitals, as well as office towers ( It is essentially part of greater downtown, but cut off from it by a steep hill and an obnoxious freeway. The original plans for the light rail line included First Hill, but because of soil concerns, it was abandoned. Thus they kept the costs low, but they also lowered the quality. If you look at just the costs, including First Hill would probably make the project look worse. But if you include the ridership, it performs better.

      This is common in the United States. DART has very poor stations, but low costs. I’m sure those making the trip have a faster ride than before, but very few people are making that ride.

      But that isn’t the most common failure; it is doing the opposite of what you recommend. Using freeway right of way can keep costs low, but results in very low ridership. You just aren’t going to get a lot of riders if the new train takes longer than the old express bus (running in HOV lanes) by connecting you to park and ride lots along the way. Likewise, a streetcar stuck in traffic isn’t going to get a lot more riders than a bus (if any).

      You really need all of it. You need good stops (connecting urban destinations) and speed improvements. There is an art to it, but whether you look at existing (slow) transit patterns, or just the neighborhoods themselves, you can fairly easily predict what will get a lot of riders, and what won’t. As a Seattle resident, the latest news (that Alon mentioned) suggests that a borderline project (West Seattle to Ballard light rail) will not only be a lot more expensive than originally planned, but it will be a lot worse. My guess is that they abandon good station placement, and go with cheap (and very poor) station placement, while championing the fact that they kept costs (relatively) low. They have done it before (several times) which is why the transit system isn’t as good as it could be.

      • Gok (@Gok)

        There is an argument that in sprawling cities like LA, Dallas, Denver, Houston, etc there really are no good places to put stations, so you might as well put them somewhere cheap today and hope that development shows up around the stations eventually. Driving can be made worse in the future cheaply but station construction costs are only going to go up.

        • RossB

          “There is an argument that in sprawling cities like LA, Dallas, Denver, Houston, etc there really are no good places to put stations, so you might as well put them somewhere cheap today and hope that development shows up around the stations eventually.”

          Except that isn’t the case with DART. There are good potential stations, they just weren’t built. Major universities and medical centers as well as the highest density neighborhoods were skipped, despite being the largest light rail system in the country — a whopping 93 miles. Maybe you can’t build a great line in Dallas — but you can at least try. Otherwise, why bother? Seriously — what is the point? You are far better off just improving bus frequency, especially since the train system, inevitably — despite the big investment — will have poor frequency as well. Of course it will. With ridership per mile of less than 1,000, you can’t afford to run the trains often.

          The only reason to build something that bad is if you don’t know what you are doing, or just don’t care. If you feel like it is about symbolism, not substance (“Hey, look at us, we’ve got a big mass transit system!”). While Dallas may be one of the worst examples of this mindset, they are not alone. Of course the United States suffers from bad density, and other weaknesses. But that is no excuse for spending a fortune on projects that carry so few riders.

          • Richard Mlynarik

            The only reason to build something that bad is if you don’t know what you are doing, or just don’t care.

            Endless and incontrovertible evidence is that it’s not “you just don’t care”, it’s that they — America’s Finest Transportation Planning Professionals — uniformly and without exception do the exact opposite of “caring [about the public interest]”, every single time, without exception.

            The worst projects are built in the worst places along the worst routes in the wost manner at the worst costs, every single time, in the USA. Every single time. It’s almost as if this isn’t by accident.

          • adirondacker12800

            You are assuming they wanted it to be good. Build stuff that is designed to be bad, when it’s bad the people think Real Americans(tm) drive every where can point to it and say “see almost no one uses it, lets drive”

          • RossB

            @Richard — I don’t think you can blame the professionals. My guess is the vast majority of professionals are shaking their heads, and wondering what the hell these cities are doing. Except, of course for the folks who are being paid, and they basically just shrug, and build whatever they want them to build (like a tattoo artist who asks “Do you really want that on your back?”).

            I can tell you, for a fact, that professionals had nothing to do with this: Every single decision was made by politicians who know nothing about transit, and have more important jobs. I realize that sounds like I’m insulting politicians — I’m not. My mom was a politician, and I respect what they do. But if you are the mayor of a big city, do you really have the time to read up about transit systems? Of course not. If you have a board full of mayors, then this problem is amplified. The big problem is that they don’t back up, and look at the big picture. Rather than asking “What is the problem we are trying to solve, and what is the most cost effective way to solve it?”, they say “Let’s build a subway! Where should we put it?”. They then compound the problem by assuming the route is intuitive. My guess is a big part of the problem is thinking that it is similar to a freeway. It’s not. Subways work best when they connect to destinations within short walking distance. You want people getting on and off at each stop. Riders have to stop at every stop, which means that longer distance lines only make a lot of sense if you have lots of destinations along the way. Freeways, on the other hand, don’t have to serve a neighborhood — the driver can just drive that extra mile. It doesn’t matter how far you are, or how many stops there are; if anything, a freeway is best if the distances are relatively long. This explains the mindset with so many American mass transit systems — you had people without a solid understanding of mass transit mimicking a freeway system. This is why DART is so huge, and manages to cover so little of the city. It also is, in a nutshell, why Seattle is spending over $50 billion building a system 120 miles long, yet most of the high density neighborhoods won’t have service, and the vast majority of transit trips will still be on (fairly slow, urban) buses.

            Then you have the political aspect. The public doesn’t know any better. Mayors can brag about bringing mass transit to their city (which really should have decent bus service first). Never mind the fact that the stations won’t be anywhere near where people live — if you think of it like a freeway ramp, then it sounds good. You have fancy ribbon cuttings, and the naive belief that it will restore your city to glory, when it fact, it will merely be a big albatross around your neck, while you struggle to pay for things that are far more important (bus service, good schools, social services).

            I’m no expert on how cities do this, but my understanding is that Seattle is nothing special. This is common. I know of no example where they actually do what I would do, under the same situation. I would hire a few firms, and have them come up with rough ideas for the best way to spend the money. Then I would have a public discussion as to the trade-offs. I have no idea exactly what we would build, but there is no way it would be anything like Seattle is building, or what Dallas built, or Denver …

          • Herbert

            Running more trains is a cheaper way to attract ridership than adding new infrastructure… If they are running 20 minute or worse frequencies at rush hour, are they even trying?

      • df1982

        Australian metro projects are also suffering from a bad case of station aversion syndrome. The planned west metro in Sydney has sections with 5-7km gaps between stations, even though it is passing through some of the most densely populated parts of the city. This is partly to keep costs down (which are ballooning out of control anyway), and partly due to an unrealistic expectation that a metro line should be able to compete in journey time with existing limited stop commuter trains, when really it should be about opening up new sources of ridership.

        • Eric2

          Just looked this line up – does it really serve only suburbs that are already served by existing rail lines, while ending 3km outside downtown for no particular reason? Absolutely nuts at any construction cost!

          • michaelrjames


            Yes, more or less. But the terminus is Central Station which is at the top end of the CBD ie. the CBD is generally defined as the area from the station down to Circular Quay; that’s also the route of George Street that has the new tramway (it does a turnaround at Circular Quay). The justification for Metro West is that the existing rail, with its classic double-deck trains, can’t run at high enough frequency but of course that begs the question of why one wouldn’t simply upgrade its signalling system. The real answer is as I have described previously on Alon’s blog for the NW-Metro (first phase completed last year) and the SW-Metro that will pass thru Central, go under the harbour and meet up with NW-Metro at Chatswood in North Sydney (effectively making one thru line). The conservative government has made the new Metro lines incompatible with the existing duplex trains deliberately to exclude the transport unions or government itself. They have already given the concession to run NW-Metro to HK’s MTRC. It is part of a broader plan to privatise these new Metro lines as a separate and independent system to the historic Metro, though they no longer talk about it because privatisation is deeply unpopular in Australia.

          • Eric2

            Well, the capacity thing is just a lie, the entire corridor has 4+ tracks already.

          • michaelrjames

            Oops. At some point the route of the Metro West line changed (there’s a map in the linked article below): it now terminates at the “CBD” (which probably means the Town Hall station, about halfway between Central station (at the top of the CBD) and Circular Quay at the bottom/harbour end of the CBD). But now it doesn’t have a stop at Central! That’s a bit like having RER B & D stop at Chatelet while passing thru but not stopping at Gare du Nord. The reason is that the line comes into the CBD somewhat to the north-west of Central so it would involve a dogleg, and thus be more expensive both because of the station and the extra tunnel length.
            They’ve hardly really started building this thing and the cost inflation is already happening–in $3bn increments!

            Warning that Sydney’s biggest rail project risks costing $27b and opening late
            By Matt O’Sullivan, 08 Feb 2021

            The NSW government was warned last year that the cost of building its flagship rail line [Metro West] between central Sydney and Parramatta risked ballooning to almost $27 billion – nearly $3 billion more than earlier internal estimates – and opening three years late.
            The documents show the estimated cost of underground stations alone for Metro West range from about $640 million for the CBD and $600 million for North Strathfield, to about $390 million for the Bays Precinct at Rozelle and $350 million for Five Dock. The figures include so-called escalation costs.

      • Henry

        To be fair, the straight line would also bias towards Westlake -> Capitol Hill. Going to First Hill would require a diversion south, which ultimately wouldn’t mean too much in terms of travel time, but the issue is if it’s gonna be a part of the spine, it is a diversion. IMO the worse sin is that the second line intentionally does not have a First Hill stop. It is entirely possible to have routed the second downtown line in First Hill to avoid the original location with the soil problems.

        As far as express bus, while the general point is true (light rail only > bus if bus is faster), at least in the Seattle case we’ve reached the limits of that strategy. Express bus travel time and unreliability has shot up with the general growth of the metro area, and there were issues with driver recruitment pre-pandemic to run even more services. IIRC East Link is projected to have comparable travel times to the express bus’s faster running times, but it will do this at all hours of the day. And it certainly isn’t helped by the fact that the road network is notoriously constrained (e.g., there’s only two lanes under the Convention Center, and there is only one through I-5 lane from north through south; if you want to travel through Seattle on another lane, you’re going to have to merge, and probably quite a lot)

        • RossB

          “To be fair, the straight line would also bias towards Westlake -> Capitol Hill. Going to First Hill would require a diversion south, which ultimately wouldn’t mean too much in terms of travel time, but the issue is if it’s gonna be a part of the spine, it is a diversion.”

          OK, now you are just being silly. If you are going to have one line through downtown, this is the only order that makes sense, which is why it was the original plan going all the way back to the failed Forward Thrust Plan proposed back in the late 60s. Yes, the train took a minor detour south before heading north, but at that point the line is largely going East, to get to the UW (which you could also call a diversion if you are being pedantic). It doesn’t matter, because the stop would be one of the most popular in the system (likely top 5, in a system that will have over 70 stations). It would be extremely popular from every direction, including those who would be “diverted”. Thousands upon thousands of riders from the north would ride to that station, which in turn would mean ridership would go way up. Now those riders will simply drive, call a cab, or in some cases, ride the bus (thereby weakening the bus system). Give the agency some credit. They didn’t skip the station because it somehow made the spine curvy, they skipped it because they were worried about cost overruns, due to the soil.

          You are right though, they did it twice. If you are going to run a second line through downtown, and you somehow missed First Hill the first time, then without a doubt the line should include it with the new line. This is just common, standard practice — maximize coverage downtown. But again, they didn’t include it. This time, though, it wasn’t even considered. They didn’t want to spend the extra money. Again, it was always about the money.

          This was a design that just about very transit expert would tell you is flawed. But it keeps costs down, which is my point. If you only focus on costs, then these decisions look pretty reasonable. They avoid the station at First Hill (and the problematic soil surrounding it) which in turn lowers the overall cost of the system, as well as the cost per mile, or cost per station. Wonderful! You could save even more money if you just did away with all the pesky stations, and ran the train on an empty field!

          That is why, at a minimum, it is important to look at ridership, along with costs. If you get poor ridership for the money, you failed. Either your design was terrible, or your costs were too high. You would have been better off just improving the bus service — which, I can tell you, sucks if you are in America, and nowhere near the saturation point. Ridership per dollar spent is just the minimum. If you pass that test (and only a handful of American systems do) then you start looking at other measures, like ridership increase per dollar spent as well as time saved per rider. Both require more work (the former with adjustments to outside forces, the latter by looking at detailed stop data) but even rough estimates will show that American mass transit systems are usually too expensive, poorly designed, or both. Again, the poor design is often driven by high costs (but not always). For Sound Transit, it is a mix.

          Side Note: While there are flaws with East Link, the basic idea is solid. Like the main line, a lot of ridership can not be easily replicated by express buses. You can run express buses from downtown Bellevue to downtown Seattle, or Redmond to downtown Seattle, but you also need Redmond to downtown Bellevue, which should also include the Microsoft campus in between. I’m sure there are some that would argue it could be done better (or cheaper) with buses, but I’m not one of them. There are two many trips between stations to easily replicate something that East Link will provide.

          Extensions to Everett and Tacoma, on the other hand, or a completely different story. These largely just connect Park and Ride lots close to the freeway, a terrible plan for a very expensive system. At some point a subway system should end, and express buses should terminate there. That point is debatable, but clearly well beyond where Link will end.

          Oh, and on another (very trivial note) you have the geography for Seattle messed up. I-5 runs north, I-90 runs east. So express buses that will be replaced by East Link run on I-90 (in HOV lanes) where they entered a busway, and then a transit tunnel through downtown (or at least they did, until recently). So the express buses were actually quite fast, all times of day. It is connecting the stops within Bellevue/Redmond that take a while with a bus (but will be fast with a train).

  6. Pingback: Friday’s Headlines As We Look to the Future – Streetsblog USA
  7. Pingback: The Week Observed, January 22, 2021 | City Observatory
  8. Pingback: More on Eno and Construction Costs | Pedestrian Observations

Leave a Reply

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

You are commenting using your 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.