A construction worker inside the 86th Street cavern of the Second Avenue Subway tunnel in 2014
A construction worker inside the 86th Street cavern of the Second Avenue Subway tunnel in 2014 Bebeto Matthews/AP

It’s not just the Second Avenue Subway: Nearly all urban rail projects in the U.S. cost much more than their European counterparts.

In late December, The New York Times published a bombshell article by Brian Rosenthal about high construction costs on the New York City subway. Doing painstaking investigative work building on a set of numbers I blogged about in 2011, Rosenthal showed how, at $2.6 billion per mile, New York’s Second Avenue Subway broke records for its costs, and that all of the reasons subway officials offered were excuses. The article documented poor contracting practices, bad management, and union featherbedding.

Unfortunately, what Rosenthal portrays as a New York affliction is in fact a nationwide problem. The construction costs of American rail transit are a multiple of such costs in peer countries, for both subways and light-rail lines. Many U.S. lines that would be easy to justify economically if they cost as much as in France or Sweden are marginal at current American costs. Every city that hopes to expand its transit network should pay closer attention to best industry practices abroad if it wants its investments to be cost-effective.

How much does rail cost?

The approximate range of underground rail construction costs in continental Europe and Japan is between $100 million per mile, at the lowest end, and $1 billion at the highest. Most subway lines cluster in the range of $200 million to $500 million per mile; in Amsterdam, a six-mile subway line cost 3.1 billion Euros, or about $4 billion, after severe cost overruns, delays, and damage to nearby buildings. The Second Avenue Subway is unique even in the U.S. for its exceptionally high cost, but elsewhere, the picture is grim by European standards:






San Francisco Central Subway


$1.57 billion

1.7 miles

$920 million

Los Angeles Regional Connector


$1.75 billion

1.9 miles

$920 million

Los Angeles Purple Line Phases 1-2


$5.2 billion

6.5 miles

$800 million

BART to San Jose (proposed)

83% Underground

$4.7 billion

6 miles

$780 million

Seattle U-Link


$1.8 billion

3 miles

$600 million

Honolulu Area Rapid Transit


$10 billion

20 miles

$500 million

Boston Green Line Extension


$2.3 billion

4.7 miles

$490 million

Washington Metro Silver Line Phase 2

Freeway median

$2.8 billion

11.5 miles

$240 million

Atlanta I-20 East Heavy Rail

Freeway median

$3.2 billion

19.2 miles

$170 million

It is hard to find exact international comparisons for subway lines running in the medians of freeways. However, it should not be much more difficult to construct such lines than to build light rail. Indeed, in the study for Atlanta’s I-20 East extension, there are options for light rail and bus rapid transit, and the light rail option is only slightly cheaper than the heavy rail option, at $140 million per mile.

Lists of light-rail lines built in France in recent years can be found on French Wikipedia and Yonah Freemark’s The Transport Politic, and in articles in French media. The cheaper lines cost about $40 million per mile, the more expensive ones about $100 million.

In the United States, most recent and in-progress light-rail lines cost more than $100 million per mile. Two light-rail extensions in Minneapolis, the Blue Line Extension and the Southwest LRT, cost $120 million and $130 million per mile, respectively. Dallas’ Orange Line light rail, 14 miles long, cost somewhere between $1.3 billion and $1.8 billion. Portland’s Orange Line cost about $200 million per mile. Houston’s Green and Purple Lines together cost $1.3 billion for about 10 miles of light rail.

The first Orange Line train carrying passengers across the new Tilikum Crossing bridge in Portland in September 2015 (Timothy J. Gonzalez/AP)

Only a handful of American lines come in cheaper than $100 million per mile, the upper limit for French light rail: For example, a light-rail extension in suburban Phoenix was recently built for $65 million per mile.

Spending $900 million per mile on a subway, or $150 million per mile on light rail, may not seem so outrageous compared to New York’s $2.6 billion per-mile subway. But there’s no reason why building subways and light rail in sprawling cities should be as expensive as it is. In Manhattan, the population density and utilities around Second Avenue make it comparable to Paris, London, Berlin, and other older cities with large subway systems that have complex tunnel networks. But San Jose has wide roads and little in the way of older infrastructure, so it is better to compare it with suburban tunnels in Europe, which come near the low end of construction costs. Among the American subway tunnels recently opened or under construction, only the Central Subway in San Francisco, passing deep under the two-level Market Street Subway, has a reason to be costly.

In Stockholm, construction costs are generally low, but among its ongoing subway extensions, the most suburban one (Barkarby) is the cheapest at about $150 million per mile; an extension underwater (to Nacka) is about $220 million per mile, and a deep commuter rail tunnel under the city center is about $400 million per mile. Based on the difference in costs within Stockholm, it is likely that if other American cities built a project as complex as the Second Avenue Subway, they would find themselves facing the same costs as the Second Avenue Subway.

Cost per rider

The cost-per-mile metric is useful in measuring a city’s ability to build infrastructure efficiently. However, by itself, it does not reveal whether rail investments are socially or economically justified. In order to decide how cost-effective investment is, cities must compare costs with benefits—in this case, ridership.

In New York, the Second Avenue Subway, which is only two miles long, is estimated to eventually see 200,000 trips per day, and so far, it seems to be on track to achieving its ridership projection. This is a high figure, achievable only in very dense neighborhoods. The Second Avenue Subway serves segments of the Upper East Side that have among the highest population-density levels in the developed world. With such high intensity of usage, the new line is projected to cost $25,000 per daily rider. The cost-per-rider range for European subways seems to be $10,000 to $25,000, so the Second Avenue Subway is not even an outlier: Its extreme cost is balanced by very high ridership.

But in other American cities, there is no hope of balancing high cost with ridership in the same way. Even relatively old cities such as Boston and San Francisco have extensive suburban sprawl, and no neighborhood as dense as the Upper East Side.

Whereas the Second Avenue Subway will cost about $25,000 per rider if it meets projections, Los Angeles’ Purple Line is expected to carry 150,000 riders across its three phases (and some of this ridership comes from preexisting rather than new stations), which will come out to about $45,000 per rider. Boston’s Green Line Extension is projected to get 52,000 daily riders, making it about $45,000 per rider as well. But many American light-rail lines have per-rider costs approaching $100,000.

A worker on the construction site of the T3 tramway line from Porte de la Chapelle to Porte d’Asnieres in Paris in 2016. (Benoit Tessier/Reuters)

By contrast, Parisian tramways, which combine very high ridership and reasonable (though not low) costs, are almost all cheaper than $10,000 per rider, with many around $5,000. Parisian Metro extensions are more expensive, clustering around $10,000 per rider; these balance higher investment costs with lower operating costs.

The Second Avenue Subway would be unbelievably cheap to build in Europe. Other American lines would not be so cheap, but they could still be justifiable, with costs per rider that fall within the European range. The budgets that American cities currently devote to the odd line could fund entire networks of light rail and subway lines.

Why are costs so high?

In his exposé, Rosenthal talked about labor problems: severe overstaffing, with some workers doing jobs that are no longer necessary, and wages well into six figures. These issues are not exclusive to New York. A chapter in a textbook about megaproject construction states that “as a result of existing union agreements covering the eastern seaboard area of the United States, underground construction employs approximately four times the number of personnel as in similar jobs in Asia, Australia, or Europe.” European subway construction uses union labor, just like American construction, but the work rules that have accumulated over the decades permit higher productivity and fewer workers doing each task.

But the problems are not just to do with labor. Working alongside active commuter or intercity rail lines brings a slew of new requirements and regulations. The Green Line Extension is notable for running in a trench next to a commuter rail line, and several Boston-area rail activists told me that a key reason for its high cost is that construction is not permitted to disturb commuter rail service. Another light rail line running alongside commuter rail and Amtrak, San Diego’s Mid-Coast Trolley extension, costs $2.1 billion for 11 miles; by contrast, other light rail projects on San Diego’s wishlist are projected to cost less, about $100 million per mile in 2010 dollars.

In California, the quality of contractors is to blame for some problems. The San Diego Union Tribune ran an article in 2013 focusing on Tutor Perini, which has a pattern of submitting low bids to win public construction contracts but then running over because of change orders and courtroom battles, contributing to an average 40-percent cost overrun in the San Francisco Bay Area.

California’s contractor problem does not seem to be a national one. Jaime Garmendia, a manager at Boston’s MBTA, said, “We would cease to do business with that contractor in a heartbeat.” He put the Green Line Extension’s high cost down to poor project management. Recent changes to the Green Line Extension support that diagnosis. After the transit agency fired the contractor and hired a new general manager to oversee the project, it reduced the budget to $2.3 billion from a previous $3 billion. Of the $2.3 billion price tag, only $1.1 billion is for actual construction, whereas much of the rest is sunk cost is due to the botched earlier version of the extension.

What’s the solution?

Nearly all American urban rail projects cost much more than their European counterparts do. The cheaper ones cost twice as much, and the more expensive ones about seven times as much. The reasons for this are varied: Some arise from the character of the project (unique regulations related to commuter rail), some from state rules (California’s problem with contractors), and some from poor project management (Boston’s Green Line Extension).

Many of these problems are solvable upon a careful review of what it takes to build a new subway or light-rail line. Boston’s experience in limiting its cost overrun is instructive, and could prove to be a model for other cities. Modernizing work rules to avoid having many redundant employees, as described in detail in Rosenthal's investigation, could also provide savings. If cities overstaff tunnel construction by a factor of four, they could hire the same number of workers to build four times as much tunnel and devote the cost savings to further infrastructure improvements.

The United States is not doomed to ineffective transit investment. Since many of these problems affect some cities but not others, there is scope for cities to learn best practices from each other. Additionally, the U.S. can look at what other rich countries (such as France and Sweden) do and copy what works. There is no iron curtain between the United States and Western Europe—learning good rail-project management and procurement is well within the grasp of a large urban transit agency. All it takes is the understanding that things are better abroad, and a willingness to learn and do better.

About the Author

Most Popular

  1. A woman stares out at crowds from behind a screen, reflecting on a post-pandemic world where exposure with others feels scary.

    What Our Post-Pandemic Behavior Might Look Like

    After each epidemic and disaster, our social norms and behaviors change. As researchers begin to study coronavirus’s impacts, history offers clues.

  2. Maps

    The Three Personalities of America, Mapped

    People in different regions of the U.S. have measurably different psychological profiles.

  3. illustration of a late-1800s bathroom

    How Infectious Disease Defined the American Bathroom

    Cholera and tuberculosis outbreaks transformed the design and technology of the home bathroom. Will Covid-19 inspire a new wave of hygiene innovation?

  4. photo: an open-plan office

    Even the Pandemic Can’t Kill the Open-Plan Office

    Even before coronavirus, many workers hated the open-plan office. Now that shared work spaces are a public health risk, employers are rethinking office design.

  5. Maps

    Your Maps of Life Under Lockdown

    Stressful commutes, unexpected routines, and emergent wildlife appear in your homemade maps of life during the coronavirus pandemic.