Category: Construction Costs

California HSR: Where Now?

California is going ahead with construction of the Central Valley segment, and has just publicly released an email saying it will solicit bids in 3 months, totaling about $6 billion from Bakersfield to just south of Merced, a distance of about 200-210 kilometers. The alignment bypasses some small towns but not all, avoiding some of the scope creep that happened in the years leading up to the Business Plan, which required more elevated segments; however, some towns will still require many grade separations and viaducts, and so will Fresno and Bakersfield.

The HSR Authority has just released environmental impact reports for the Bakersfield-Fresno and Fresno-Merced that point out to higher costs: the sum of the two cheapest alternatives is $10 billion, in 2010 dollars, for 300 km; although the cost per km is not much higher, the Fresno-Bakersfield segment is much more expensive, whereas the extra bits included in the EIR but not the bid request are the cheapest.

There is some additional room for value engineering, especially in Fresno, where the currently preferred alternative calls for viaducts, but the potential for cost saving is not that great, especially relative to the $6 billion estimate; projects run over budget much more frequently than they come under. The main interest here is not the cost overrun: the current stage, the bidding, is the one most prone to overruns, and no matter what, we will know in three months what the projected cost is. The interest is the breakdown of costs, which, as expected, are primarily infrastructure and tracks, including grading and grade separations. The cost overruns come from scope creep, with more elevated segments than originally expected (but, due to value engineering, less than expected in 2009).

At any case, there is money to proceed, at least from Bakersfield to Fresno – there is $6.3 billion available, half from federal spending (which has been spared in the latest austerity plan) and half from Proposition 1A’s matching funds. There is another almost $6 billion locked in Prop 1A, but it has to be matched 50:50. Matching funds will almost certainly materialize, if not from the federal government then from foreign governments anxious to pay California to buy their products (for example, Japan’s ambassador to the US offered half the money, and Japan expects China and Korea to offer funding as well). It should be enough to build an initial operable segment, though probably not to build from Los Angeles to San Francisco.

The question is then how to prioritize. The gold standard here should be building all the way from Sylmar to San Jose and electrifying the legacy lines at the two ends. At the Bay Area end, the Caltrain FRA waiver ensures this wouldn’t cause regulatory problems, and while it would limit initial capacity, it would not increase travel time by more than a few minutes. At the Los Angeles Basin end, it would require Metrolink or HSR to seek a waiver, along the lines Caltrain has already gotten; the speed reduction, while still not very large, would be larger, because the travel time simulations assume higher operating speed in the LA Basin, and there will be fewer speed limits due to curves.

Unfortunately, while cutting the initial segment to San Jose-Sylmar will save a large number of billions of dollars in urban grade separations, it may not save enough, though it’s fairly close if one believes the 2009 Business Plan numbers. If California has half the money from foreign sources, then matched with Prop 1A and existing federal money, it has a total of $24 billion, which is not enough. The question then boils down to where to go first from the Central Valley – south or north. North would involve going over Pacheco Pass to San Jose (or, better yet, over Altamont Pass to Livermore and thence Redwood City). South would involve going south to Sylmar, either through Palmdale or directly through Tejon Pass, which carries I-5; although Palmdale is the preferred alternative, the HSR Authority is looking at Tejon again. For a slide show using the existing preferred options, see here. Either alone should be doable with the money available under such a circumstance, which is about $18 billion.

I claim that the southern option is the better one – in fact, that LA-Bakersfield is more important than Bakersfield-Fresno. The reason is, first, a pure numbers game: LA is much larger than anything else in California. And second, Tejon is where the existing legacy transit options are the worst: Amtrak can’t go between Palmdale and Bakersfield at all because the Tehachapi Loop is at capacity, ensuring that a mixed legacy-high speed operation in the mold of the initial TGV runs is not possible even under reformed FRA regulations.

Northern options suffer from different problems. The Pacheco option’s problem is that it uses Pacheco, and is therefore inadequate at linking the Bay Area to Sacramento. This means nothing further can be done until enough money materializes to connect to the Los Angeles Basin. The Altamont option’s problem is that the Phase 0 option connects to Livermore and requires a transfer; connecting to Redwood City is possible, but requires all of the most expensive elements of Altamont, especially crossing the Bay in the vicinity of the Dumbarton Bridge.

Once the southern option is selected, the question is how far to go. Bakersfield-Sylmar is expensive, and although it’s easily doable given 50% foreign funding, lower levels of funding may not suffice. Bakersfield-Palmdale is much easier, and could be done on existing Prop 1A money if it were not required a 50:50 match; however, Palmdale is not in the LA Basin, and the legacy rail line to LA is curvy and steep. Express Metrolink trains do Palmdale-LA in 1:28, versus 0:27 projected for HSR. Higher cant deficiency and acceleration with electrification could cut the travel time somewhat, but not enough to make HSR competitive for travel from LA to the Central Valley. Travel from LA to the Bay Area is another issue, but a situation in which it’s possible to build all the way to San Jose is one in which there’s money to build to Sylmar.

The alternative is to use Tejon and connect to the legacy line in Santa Clarita. It’s more expensive because Tejon is one big crossing whereas the Palmdale route involves two smaller crossings, one to Bakersfield and one to the LA Basin. It should still be affordable, though I have no detailed segment-by-segment breakdown of the Tejon route’s cost. The advantage is that Santa Clarita is much closer to Los Angeles than Palmdale, and the legacy Metrolink route to Palmdale is fairly straight south of Santa Clarita; even now, express trains travel to LA in 42 minutes, half an hour slower than full HSR buildout rather than an hour as with Palmdale, and there’s more potential for an increase in speed.

That said, the debate is most likely academic – Tejon vs. Palmdale is most likely going to be decided primarily on a revisited look at the costs, with other issues (LA County power brokers prefer Palmdale, Tejon is shorter) not much more than tiebreakers. In addition, a situation in which Prop 1A money could be released for the crossing is one in which matching funds have materialized, making the full Bakersfield-LA route realistic with the available money. The primary lesson is that there should be enough money to build a realistic initial operable segment, not going all the way from LA to San Francisco but still serving a fair number of intercity travelers.

Quick Note: Barcelona Rail Tunnel

Barcelona’s rail tunnel connecting the existing high-speed rail station, Sants, with city center, has just been completed. The tunnel’s total length is 5.8 km. As for cost:

The tunnel has cost over €179·3m to build, including extensive measures to protect historic buildings such as Gaudí’s Sagrada Familia from any settlement.

I believe this sets a new modern-day record for low construction costs – about $40 million per km – certainly in cases of inner-urban construction. It balances out the city’s Line 9 boondoggle, which has run so many times over budget it’s now a full $180 million per km.

The Mother of All Interest Conflicts

Best industry practice for cutting transportation capital costs, found in Madrid, is to separate design from construction and keep the project management in-house. The FTA’s practice is different:

Parsons Brinckerhoff said Wednesday it has been awarded a contract by the Federal Transit Administration (FTA) to develop and document a transit asset management framework and implementation guide that will support the FTA’s State of Good Repair and Asset Management Programs.

The FTA estimates there is a nationwide backlog of $50 billion to $80 billion in deferred maintenance and replacement needs, the vast majority of which are rail-related.

PB is going to decide what projects are necessary and how to build them, and will also be able to bid on design and construction. Naturally, the numbers it will come up with are going to be favorable to its private interest; the common interest is not profitable for the company.

This is especially egregious in state of good repair (SOGR) money, which is often a series of rent-seeking scams. Agencies do not impartially judge how much money they need for maintenance and then ask for it. Instead, they massage the numbers based on whether the political mood is such that they could get more or less money. In 2005, the Amtrak board fired President David Gunn for insisting on competing SOGR before attempting to move to profitability; by 2009, when the stimulus provided plenty of money, Amtrak suddenly remembered it had deferred maintenance and came up with the $10 billion NEC Master Plan, essentially SOGR plus a few small upgrades.

A few agencies, such as New York City Transit, treat SOGR seriously (this was thanks to Richard Ravitch and David Gunn) and push for it even when the politicians want something different; most just use it as an excuse to justify high capital costs without anything to show for it. Look again at Amtrak, which even as it cries poverty about SOGR is trying to portray its finances as very good, for example listing a farebox recovery ratio that, unlike the practice at peer national railroads, excludes depreciation and interest. Heads Amtrak is profitable and competent and should get what it wants, tails it has a backlog of deferred maintenance and needs more money.

This is more a political than technical problem, but normal political advocacy is not going to help. Politicians can get credit for massive overhauls or new infrastructure involving ribbon cuttings; they won’t get credit for adding to the design and management budget, no matter how much money it will save in the long run.

Therefore, politicians who care more about being seen as fiscally conservative than about saving money force agencies to cut their in-house expertise. Instead, agencies outsource everything to consultants; this can work sometimes, but the people who would oversee them have been cut, so that there’s nobody in charge who’s loyal to the interests of the agency or the public. As a result, nobody in the US knows anything about good practices for rail infrastructure construction except people with the mother of all conflicts of interest, and nobody knows anything about rolling stock except New York City Transit, which designs rolling stock in-house or buys designs and prototypes separately from revenue equipment.

The agencies have bought into this system, since they share in the overly expensive designs and must defend them. Madrid doesn’t separate design from construction just because of interest conflict issues; the reason stated by Madrid Metro CEO Manuel Melis Maynar is that changes are unavoidable, and a construction crew uninvolved with the original design would be less stubborn about sticking to the blueprint. Since such separation does not exist in the US, and on the contrary the people currently in charge are used to the system so much that they bring up design/build contracts as an improvement, agency inertia is directed toward making the agency even less competent.

California HSR is perhaps the worst example of this. The HSR Authority consists of nine politicians, overseeing a skeletal crew of professionals (I believe there are only six engineers/planners). Unsurprisingly, the Legislative Analyst’s Office (LAO) Peer Review Group wrote a peer-review report accusing the HSRA of having no expertise in project management or even in negotiating a good PPP so that the private sector could do it. Even more unsurprisingly, hiring more staff to bolster an agency that’s currently incompetent is risky and nobody wants to be responsible for either potential delays or spending good money after bad, despite the possibility of large cost savings in the medium and long runs.

Quick Note: Road Boondoggles

With all the focus on poorly done transit investment on this blog, it’s sometimes easy to forget that the primary source of US transportation waste is still roads. Consider for example the following projects proposed in Southern California, not all funded:

– $1 billion fully funded for adding one carpool lane in one direction for 10 miles to the 405 through Sepulveda Pass; since the 405 will have to be closed for two days, this is locally dubbed Carmageddon. This is about $60 million per unidirectional lane-km, which is to my knowledge a record for above-ground highways.

– $3 billion proposed for 4.5 miles of twin tunnels to complete a gap in the 710, of which $780 million is funded by Measure R, which generally funded transit projects. The cost, $400 million per km, is not high by global tunnel standards, but compared with the opportunity cost of building transit in the area, it’s enormous.

– $4.1 billion for widening the 5 from 8 lanes to 12-14 for 27 miles, not yet funded. It’s about $18 million per unidirectional lane-km, a figure that’s cropped up elsewhere in the US and should be compared with about $15-80 million per double track-km for light rail, which has about eight or ten times the capacity per unidirectional track or lane.

Those projects are cheaper than the Big Dig or the Bay Bridge Eastern Span replacement, but also provide much less – two are routine widenings, and one is a minor tunnel. The point is that even small upgrades to road capacity cost as much as a major transit project.

The US road network has been a money sink going back to the first federal-aid highway act, in 1917. The reference here is 20th Century Sprawl, by Owen Gutfreund, who describes how motorist lobbies complained about license fees, fuel taxes, and other fees since the 1910s, and created road lockboxes for the revenue generated. Even though gas taxes were treated differently from cigarette and alcohol taxes, which do not go toward funding tobacco and distilleries, they were still not enough to pay for roads. In fact the only paid for about half the cost of highways, and there was a huge subsidy from gas tax-ineligible urban roads to the national and state roads.

The situation today is hardly different. Although proportionally there’s much less cross-subsidy than in the 1930s, due to the growth of suburbs connected by Interstates or other gas tax-eligible numbered roads, roads’ financial performance is still low. Under the fiction that local streets are paid by the tooth fairy, US roads are $75 billion a year in the hole: as of 2008, all gas tax and toll receipts are $122 billion, including the portion diverted to non-highway purposes, whereas total receipts to be spent on gas tax-eligible highways are $197 billion, including $4.3 billion spent on collection expenses. That’s 62% cost recovery.

It gets worse when one does a total lifecycle cost analysis and does not deed all local gas tax money to state highways: in Texas, the best-performing highways have 50% cost recovery, and most have much less. In Maryland, one transit advocate computed a 20% cost recovery for state highways, based on an analysis that treats most of the gas tax as just a sales tax on gasoline; but even if one considers the gas tax to be a user fee for roads, the extra money only raises cost recovery to 32%. Even tollways frequently lose money when interest on capital is included, and in one case even when interest is not included.

In other words, the entirety of the US road program is one giant money hole, of proportions that far exceed even the worst transit projects. I talk less about it because the best industry practice is to toll the roads and build far less of them rather than to control costs; there’s a good way to build a subway, but not to build 14-lane freeways.

Boosters’ Romanticism

One would expect that boosters of unbridled growth, such as Thomas Friedman, Richard Florida, Ed Glaeser, and countless proponents of urban growth would constantly look to the future and deprecate the past. They certainly deprecate attempts to recreate the past. But do they? Despite unabashed pro-Americanism, they crow about the fast growth of China. Glaeser looks back to an era of great infrastructure spending on water works in turn-of-the-century America. Infrastructurist and urbanophile bloggers look back to Daniel Burnham and early-20th century public works (though the Infrastructurist and Urbanophile themselves are very self-conscious and are more thoughtful in their boosterism).

Instead of writing about history as a series of epics, let us examine it with the same critical eye we examine the present. This means looking at historical paths not taken, much as we should examine alternatives for projects today; this also means looking at costs and benefits. In most cases, the inspirational projects of the past tend to not look very good under the microscope.

For a concrete example, consider the Interstate system. Examples of writings on infrastructure that take its greatness for granted are numerous, even on Streetsblog as far as job creation is concerned. But in reality, it was an epic disaster for most involved. The original 1954 estimate for the cost, enshrined in the 1956 act creating the network, was $25 billion; by 1958 it had already climbed to $40 billion, and the final cost was $114 billion. The construction required demolishing thousands of dwellings in each city the highways went through. Even burying the highways does not help: the scar of Boston’s Central Artery is still there despite the Big Dig, because amidst cost overruns they dropped the option of building above the tunnel.

The utter failure of the USA’s road-building program goes further back. As explained by Owen Gutfreund in his book 20th Century Sprawl, urban streets, on which it was illegal to spend gas tax money until the late 1930s, subsidized the early highways and rural roads; overall, roads only covered about half their capital costs through gas taxes. Tollways faced intense opposition from the AAA and the auto and tire industries. Instead an entire bureaucracy was created to ram roads through, paving the way to the large-scale neighborhood destruction of the 1950s. Tellingly, New York and San Francisco, the first two major cities to have freeway revolts, had a smaller population decline through 1980 than the other major non-Sunbelt cities, and are now the only two to have since surpassed their 1950 population peaks.

Transit investment in that era was no better. New York’s major project in the 1920s and 1930s was the construction of the IND, competing with the existing privately-run IRT and BMT networks. The new lines generally did not add transportation options. The Crosstown and Queens Boulevard Lines added service, but did not connect to existing IRT or BMT stops; to this day, the G train has no transfer to non-IND lines in Downtown Brooklyn, and only one, difficult transfer in Queens, which opened just a month ago. The remainder simply paralleled existing elevated or subway lines, which were subsequently torn down.

Part of it was the general opposition to elevated rail in that era, coupled with fascination with both subways and elevated highways. But only part: one IND line, the Sixth Avenue Line, required building new track alongside and later below the existing Hudson Tubes (now PATH), dooming previous plans to extend them to Grand Central for greater regional connectivity. On top of it, the difficulty of building next to an active subway created massive cost escalation, dooming future expansion plans that would add new service.

Although both of the above examples are from the middle of the 20th century, previous infrastructure investment was not much better. It’s a commonplace that New York’s first subway line was built in four years, versus ten for just one phase of Second Avenue Subway. It’s less widely known that ground broke on the subway in 1900 only after multiple decades of political bickering, route changes, and scandals; a short underground demonstration line using pneumatic tube technology had opened in 1869.

Even before then, Britain had undergone a pair of Railway Manias, one in the 1830s and one in the 1840s (thanks to Danny in the comments for the link). Relative to GDP, the latter mania dwarfed both the 1990s’ tech bubble and the 2000s’ housing bubble. Costs ran over estimates by a factor of 2 or more, and ridership underperformed estimates. Although by the end of the Victorian era the lines had surpassed the mid-19th century predictions and were profitable, the investment was too fast, and ruined many investors.

Nobody romanticizes the present, because its problems are apparent to all. Some people romanticize the future; those are the boosters, for whom every problem with growth has a simple solution. But even those can easily slip and romanticize the past, whose main actors have since become national heroes and whose main battles have turned into epic legends. Obama and Bloomberg are controversial; Eisenhower and LaGuardia are heroes.

High Costs Should not be an Excuse to Downgrade Projects

In an environment of high construction costs, there’s an impulse to downgrade projects: build light rail instead of subways, BRT instead of rail, commuter rail on existing tracks instead of greenfield light rail, shared-lane buses and streetcars instead of ones running in dedicated lanes. Some of those downgrades have already gotten flak individually from transit supporters, of which Jeff Wood’s recent article about commuter rail and Mike Dahmus‘s repeated attacks on BRT and the Austin commuter rail are good examples.

I do not think anyone has made the following point connecting those projects: the same causes that lead to incompetence in running one mode will lead to incompetence in running all other modes. Regardless of the mode chosen, a project in the US can expect to cost several times as much as a comparable European projects. (As a single exception, FRA-compliant commuter rail can be expected to be especially bad, because there the regulations and operating traditions are especially bad.) With very few exceptions, building BRT in a corridor that begs for rail, suburban transit in a city that needs urban transit, peak-only commuter rail, and other apparent cost savers will incur the same cost escalations as in every other mode.

In particular, downgrading service will not save any money, and going to the FRA will actually raise costs. This affects both the choice of technology and the choice of how to use it: American light rail lines keep the per-km costs reasonable by building out to exurbia, creating ersatz commuter rail with low ridership. This is epitomized in Dallas, whose light rail is setting records for low per-km ridership, and whose plans for the next decade are projected to cost $2.4 billion for 60,000 additional weekday riders, i.e. $40,000 per rider. In contrast, Houston’s urban Main Street Line cost $300 million for 34,000 riders, which is about $10,000 per rider in today’s money, the lowest per-rider cost in the US in the last 15 or so years. And Houston is unusual; more common is the Portland Milwaukie light rail extension, projected at $55,000 per rider and $110 million per km.

If we start looking abroad, we see the same pattern. When European LRT is more expensive, as for example in Nice, it’s because it’s very high-ridership urban infill. And Nice is an exceptionally expensive case; Lyon’s trams are cheaper. Few European light rail lines go over $10,000 per rider, and on Yonah Freemark’s list of recent and planned projects in Paris, a few lines are below $5,000.

Something similar is true for bus transit. Despite Jaime Lerner’s admonition that “Creativity starts when you cut a zero from your budget,” American cities have failed to create good BRT under budget constraints. The Los Angeles Orange Line is expensive for the ridership it has ($15,000/rider in construction, with the high operating costs of a bus) and has mediocre signal priority. Under a budget constraint, Los Angeles still built something inferior to the Blue Line, or even the expensive-to-build, cheap-to-operate Red Line subway.

As an aside, this also holds for the costs of transit versus highways. In the rest of the developed world, prudent cities invest most or all of their transportation money into mass transit, and try to restrain traffic. This should also be true in the US, where subways and light rail are expensive, but so are highway projects: see the 8-times-over-budget Bay Bridge Eastern Span replacement, the Big Dig, and the proposed Tappan Zee Bridge replacement, and compare them to the more complex Øresund Bridge-Tunnel connecting Denmark and Sweden.

At worst, the high costs of transportation in the US imply that government should spend its money elsewhere – on health or education, or perhaps tax cuts. Even then I’m personally skeptical about the efficiency of the marginal dollar: American health care is infamously expensive, tax expenditures are byzantine and in such cases as the mortgage tax credit create the wrong sort of incentives, and so on.

Second Avenue Subway Phase 1 is by far the most expensive urban rail project in the world today, but its per-rider cost is only $25,000, high by European and Japanese standards but lower than any other rail line proposed or under construction in the US today. It would not be approved in today’s pennypinching climate, and even ten years ago it was funded only thanks to legislative blackmail by Assembly Speaker Sheldon Silver, whose district would be served by Phase 3. Of course at normal cost it would be very cheap, just as at normal cost everything else in the US would become much more affordable, but it is still more cost-effective than seemingly cheap commuter lines.

The upshot is that from the perspective of transit planning, high costs should not deter anyone. Other than the special rule that FRA-compliant commuter rail is practically never justifiable, the relative merits of projects are about the same in the US as in all other developed countries. Agencies all over the world have to choose between a subway, five trams, and twenty busways. In an environment of high costs, it still make sense to draw plans as if the costs are normal, and when the costs are not normal, build more slowly and start with the most cost-effective lines. If agencies and activists behave as if there’s no money for good transit, they will only get bad transit.

US Rail Construction Costs

Update 2017/7/1: this is the most linked-to post of mine about construction costs, even though the dataset here is relatively small. You can see links to more posts, with more datapoints, on my static construction cost page. The long and the short of it is that in non-English-speaking developed countries, the typical range for urban subways is $100-300 million per km, with a few outliers in both directions.

This is a placeholder post, in which I’m just going to summarize the costs of projects in the US and the rest of the world. I will focus on subway tunnels, but also put some above-ground rail for comparison. No average is included – all I’m doing at this stage is eyeballing numbers. As far as possible, numbers are inflated or deflated from the midpoint of construction to 2010, and exclude rolling stock. The PPP exchange rate is €1 = $1.25, $1 = ¥100. For now, only dense infill subways are included.

East Side Access: $8.4 billion; excluding preexisting tunnels, this consists of 2 km of new tunnel in Manhattan and a new connection in Queens. So this is about $4 billion per km. Update 2011/6/21: the link here stopped working. Here‘s a slightly older link, saying the cost is $8.1 billion.

Second Avenue Subway Phase 1: $4.9-5.7 billion in 2007-17 for about 3 km of new tunnel. This is $1.7 billion per km.

7 Extension: $2.1 billion in 2007-12 for 1.6 km of new tunnel. Note that this has only one station, an unusually sparse spacing for a dense urban area. This is $1.3 billion per km.

Crossrail: £15 billion in 2008-18 for a line of more than 100 km, of which the primary component is 22 km of new tunnel under Central London and Heathrow Airport. Due to the extensiveness of the London Underground network, this is the most complex project on the list. The cost per unit of tunnel is about $1 billion per km, making this the only outside New York to cross the 1 billion line.

Central Subway: $1.58 billion in 2010-6 for 2.7 km of light rail tunnel. This project is only on this list because it has to cross under the double-decked subway (Muni and BART) under Market; the standards, including station size, are for light rail. This is about $500 million per km.

Jubilee Line Extension: £3.5 billion in 1993-9 for 15.9 km of route, of which about 80% is underground. The line went over budget by 66%, crosses under the entire London Underground network, and crosses under the Thames four times. This is about $450 million per km.

Amsterdam North-South Line: €3.1 billion in 2009 money for 9.5 km of new tunnel. The project has run over budget by a factor of more than two, leading to accusations of boondoggle and remarks that the project should not have been built. This is $410 million per km.

Toei Oedo Line: ¥1.4 trillion (Japan has no inflation, so year of construction does not matter) for 40.7 km of new tunnel. While the stations are normal subway stations, the subway tunnels are of smaller than normal diameter due to the use of linear induction technology. This is $350 million per km. A short subway extension of the Mita Line cost nearly $500 million per km, but the information about it is on a Toei factsheet that’s been scrubbed from the net.

Tokyo Metro Fukutoshin Line: ¥250 billion for 8.9 km of new track. This is $280 million per km. Tokyo Metro has claimed future lines will be $500 million per km as a reason to not build future extensions.

Berlin U55: €320 million for 1.8 km of tunnel in 1996-2009. While this line does not cross or connect to any older subway, it is in the center of the city, and thus qualifies as urban infill. This is $250 million per km.

Paris Metro Line 14: €1.13 billion in 1998-2003 for 9 km. This line crosses under the Seine and had construction problems due to catacombs. This is $230 million per km.

Circle MRT Line: S$10 billion for 35.7 km, to be opened in full next year. This includes a 50% cost overrun, and a substantial delay coming from a highway collapse in 2004 that killed four workers. Because the exchange rate, including PPP, has changed considerably in the last ten years, I’m not inflating, and instead using the present rate, making this the least certain conversion on the list. This is $220 million per km.

Copenhagen Circle Line: DKK21.3 billion in 2010 for 15.5 km. At today’s exchange rate, this is $4 billion and $260 million per km in exchange rate, but the Danish currency is severely overvalued, and in PPP (judging by the ratio of PPP to exchange rate GDP per capita) this is $170 million per km.

Durchmesserlinie: CHF1.82 billion in 2007-13 for 9.6 km of new commuter tunnel under the city, relieving the existing tunnel. This is $215 million per km in exchange rate, but the Swiss franc is severely overvalued, and the PPP value is only $136 million per km.

Barcelona L9/10: €6.5 billion in 2006-14 for 47.8 km. This line is fully automated and is nearly 100% underground, and has gone over budget by a factor of more than three. This is $170 million per km.

Naples Metro Line 6: €533 million in 2007-12 for 5 km of fully underground metro. This is  $130 million per km.

Milan Metro Line 5: €500 million for 5.6 km of fully underground driverless metro. This is about $110 million per km.

Seoul AREX: 4.2 trillion won ($4.2 billion) for 61 km of line, about 60% underground, linking Seoul with Incheon Airport. This is a combined commuter and express line, and even all-stop trains only make 10 stations, by far the sparsest spacing on this list. This is about $110 million per km of tunnel – realistically a little less since the above-ground segments are greenfield.

Seoul in general gives its tunneling construction cost as $100 million per km in context of a proposal of an extension of the Sin Bundang Line that assumes a much lower budget, only $40 million/km.

Madrid gives the construction costs of its 1999-2003 expansion as €42 million/km, including rolling stock; translated to today’s dollars, this is $65 million per km. But those projects were not all infill and not all fully underground.

Observe from the low costs of Italian subways that corruption alone cannot explain high American and British costs. High Japanese costs can be explained by strong property rights protections and a process that favors NIMBYism; Paul Barter‘s thesis quotes sources arguing that the high costs of land acquisition in Japan are a reason why its cities never engaged in American-style urban renewal or massive freeway building.

Observe also that developing countries’ PPP costs aren’t very low: Beijing’s subway extensions cost about $150 million per km – see e.g. here and apply a PPP exchange rate of about $1 = 3.8RMB. The labor costs in developing countries are lower, but so is labor productivity.

Observe finally that Bent Flyvbjerg, known primarily for his work on megaproject construction cost overruns and strategic misrepresentation, wrote a paper on comparative US and European construction costs, which understated the conclusion that American costs are higher. The reason for his understating the conclusion is that the American projects examined are quite old, from the 1980s, and many have large above-ground parts.

Although the US projects included are only in New York and San Francisco, both high-cost cities, similarly high costs occur in other cities, just the projects are above ground. Portland’s light rail Milwaukie extension and Washington’s predominantly above ground Silver Line both have cost ranges of about $100-150 million per km, enough for a full subway in many European cities. Los Angeles’s Subway to the Sea is budgeted at $6 billion for the full Wilshire route to Santa Monica, i.e. $300 million per km; this is not really infill since it extends the subway out, but the neighborhoods served are quite dense, so it might qualify.

For some links of outward extensions abroad, see Brussels ($60 million/km) (deleted because the link is wrong and I can’t find the right one) the future plans in Paris ($100-200 million, with one line at $50 million; 2018 update: see updated costs around $200-250 million here and here) and Seoul’s upgraded Gyeongchun Line ($33 million).