A $100 Billion Lesson In Why Building Public Transportation Is So Expensive in the US

There's a plan to spend $100 billion fixing the Northeast Corridor by 2035. Similar countries build entire new rail corridors with dozens of new stations for a fraction of that cost. Why can't the U.S.?
Bloomberg / Contributor via Getty
Screen Shot 2021-02-24 at 3
Moveable explores the future of transportation, infrastructure, energy, and cities.

In 2015, the Paris region began construction on the Grand Paris Express, a 125-mile expansion of the city’s rail system that will transform the city and its suburbs. Most of it—80 percent—will be in tunnels. It will have 68 new stations, four new lines, and two extensions of existing lines. It will be completed, if all goes according to plan, in 2030, with the first line opening in 2025. It will cost some $38 billion. And this is somewhat of a scandal in France because it is much more than originally planned. 


Meanwhile, in the U.S., there is a plan to rehabilitate the country’s busiest rail corridor, called the Northeast Corridor, from Washington, D.C. to Boston. It will add zero new stations. It will add zero new lines. It will extend zero existing lines. It will convert a grand total of 100 miles to high speed track, meaning the majority of the corridor will still be conventional slow rail. It will be completed by 2035. And, according to a project cost breakdown released in October, it will cost an estimated $101.8 billion, some three times more than the Grand Paris Express. 

This is not a case of France building new train lines remarkably cheaply. According to Eric Goldwyn, an NYU fellow and member of the Transit Costs project which has compiled a massive dataset of transit costs around the world. It is about the global median in terms of construction costs. The problem is the U.S. 

The U.S. continually breaks its own records for most expensive track miles on Earth, stretching the realm of credulity on costs for projects big and small. The problem has existed for decades, but it has been allowed to metastasize to such an extent that it is difficult to even fathom where the money is going. For $100 billion, the Northeast Corridor Commission, a kind of meta-agency to coordinate efforts along the corridor, says Acela riders will have a “nearly 30 minute” faster trip on the main section of the corridor from DC to New York, with a similar time savings from New York to Boston. By comparison, a new 170-mile high-speed rail line connecting Turin, Italy and Lyon, France with a 35.7-mile underground tunnel—the longest rail tunnel in the world upon completion in 2032—is estimated to cost about $26 billion, or a quarter the cost.


There are several reasons why U.S. transit projects, generally speaking, cost so much more than in the rest of the world. As the NYU research group found after years of extensive study, it is a complicated web of poor contractor and consultant management, poorly coordinated work between different entities, viewing infrastructure projects as job programs, and over-designing projects from the start, among other problems. These findings have been broadly confirmed by other research groups

But the problems start at the very beginning. From the outset, no one seems to care about costs. And nowhere is this more true than on the Northeast Corridor. On the hundreds of projects planned up and down the corridor, there appears to be no sense of tradeoffs, that spending more on one project makes it harder to find money for another. Every project is assessed solely on the basis of whether it needs to be done, and if so, it is deemed worth whatever it costs. Nor is there any overarching goal in terms of travel times or service improvements.

The end result is what Alon Levy, another NYU fellow and member of the Transit Costs project, calls a $100 billion “staple job,” putting together a bunch of expensive, unrelated projects and calling it a plan, wasting tens of billions of dollars in the process.


“Do you remember that scene in The Dark Knight with the Joker and the piles of cash?” said Levy. “Yeah, it is essentially that.” For those who haven’t seen it, the Joker lights the money on fire.


Of the $100 billion wish list, one-third of it, or $37.5 billion, is the Gateway project. 

The Gateway project is a massive program to rehabilitate and expand the bridges and tunnels that link New York City with New Jersey. It is currently the biggest bottleneck in the Northeast corridor. It also relies on century-old infrastructure that is deteriorating, especially the two tunnels underneath the Hudson River. 

Nobody disputes that the project is necessary, particularly the tunnel replacement, which would also rehabilitate the existing tunnels once the new ones are built. But the cost of the Gateway program is difficult to fathom. Consider just the tunnel portion of the project, known as the Hudson Tunnel Project. It will create nine new track-miles with two 4.5-mile tunnels and is slated to cost $16 billion. About $2 billion goes to rehabbing the old tunnels—which, it should be said, is on its own a horrendously expensive project for rehabbing two rail tunnels—so that’s $14 billion for the new tunnel construction, or $1.55 billion per mile. 

Elsewhere, Denmark and Germany are constructing a train and road tunnel underneath 11 miles of the Baltic Sea with two road tubes, one emergency tube, and two rail tubes. It will be the longest immersed tube tunnel in the world. And it will cost about $6 billion less than the Gateway tunnels.


I asked Stephen Sigmund, chief of public outreach for the Gateway Program, why the Gateway tunnels cost so much. “The answer is, one, we’re totally different than other countries,” Sigmund said. “But much more important than that is, what’s the cost of not doing this?”

Levy said we don’t even have to compare the Hudson Tunnel Project to other countries to question why it costs so much. We just have to look across the river. Levy said that the tunneling costs of Second Avenue Subway in Manhattan, then a world record, was about $340 million per mile, after stripping out building the stations themselves, because the Hudson Tunnel Project doesn’t involve any stations. Gateway will be about five times that. 

“So this is literally an internal comparison where we know all of the bullshit that has happened with procurement, labor, and everything else,” Levy said. “And it’s still something like [five times more expensive]. And I have no idea why.”

It is worth emphasizing that nobody I spoke to for this article knows why the Hudson Tunnel Project costs so much. In their detailed case studies, the NYU research team took months to study individual projects, requiring cooperation from the agencies building it. A key tool for those investigations are itemized cost lists. Without knowing where the money goes, it’s impossible to say where it is wasted. And there is no public itemized cost list for the Hudson Tunnel Project.


I asked Sigmund repeatedly for a cost breakdown on the tunnel project. I never got one. Levy said this is typical for U.S. rail projects, where cost breakdowns are often shrouded in secrecy and regarded as trade secrets. “There is a remarkable lack of sunlight on everything touching mainland rail in the United States,” Levy said.

Sigmund emphasized that even with the $16 billion price tag, the project pays for itself because the cost of doing nothing is even more. No doubt the project is necessary and worth doing, but if the Hudson River Tunnel project could merely cost as much per track-mile as the historically expensive Second Avenue Subway tunneling project, or the tunneling of the world’s longest underwater tunnel in the Baltic Sea, there would be enough money left over to pay for Second Avenue Subway Phase II.


But it is not just about the tunnels. The Gateway Program is slated for another $20 billion in improvements. What does that buy us?

Of that money, $11 billion is earmarked for a Penn Station expansion of dubious benefit, one that hinges it's worth—other than as a real estate giveaway—as accommodating more trains even though similarly-sized train stations around the world operate busier schedules with fewer platforms simply by managing the trains in the terminal better. Even if it was a good project, it is still amazingly expensive. A massive new urban development project in Barcelona which includes a new 18-track two-level train station with a connection to buses and the Metro as well as a 2.3-mile High Line-type linear park is slated to cost $2.1 billion, about a fifth of the Penn Station renovation.


There is also $1.9 billion earmarked for a New Jersey Transit storage yard, or a giant train parking lot. Other countries avoid paying billions of dollars for train storage by, firstly, keeping them circulating during operating hours and, secondly, parking them in yards far from cities where the land is cheap. 

There is some $5 billion for replacing the Portal North and South Bridges which, like the Hudson River Tunnels, are old and need to be replaced but the costs are unreal. The Portal North Bridge, for example, is a 960 feet long and will cost $2.2 billion to replace, including the elevated approaches.

It is easy to pick on the Gateway program for insane costs, but this is not a Gateway problem. It is the same story up and down the Northeast corridor. A century-old bridge in Greenwich, Connecticut called the Cos Cob Bridge that is 1,063 feet long will cost $1.1 billion to replace. That’s $5.5 billion per mile. There are three other bridges in Connecticut with similarly baffling replacement costs. There’s a $2.2 billion project to renovate Boston’s South Station that I, for the life of me, cannot explain what it actually does in a way that comes close to justifying $2.2 billion. There’s $11.7 billion earmarked to rebuild Washington’s Union Station to, as best as I can tell, make it easier to access local buses while adding underground parking, a goal that is not worth spending $1 billion much less almost $12 billion. 


Even small, basic projects, like raising platforms for ADA accessibility at underused train stations, like in Aberdeen, Maryland, are slated to cost some $26 million when, according to Levy, doing the exact same work at train stations in other parts of the world costs maybe $10 million at most.

I could go on. There are 307 projects on the Northeast Corridor project list. 


How does this happen? How do we get $100 billion in projects that achieves, by the Northeast Corridor Commission’s own numbers, a time savings return on investment of a tenth of a percent?

I asked Mitch Warren, the executive director of the Northeast Corridor Commission, how the project list was put together. He said all of the railroads and agencies that make up the Northeast corridor—commuter railroads, Amtrak, freight railroads—send the commission a list of the infrastructure projects it wants to do and how much it would cost. The commission then puts them together and crunches some numbers to find out what all these plans mean for riders. While the commission does work to resolve disputes and encourage work to be done in a coordinated fashion, Warren said “the basic information for the plan comes from all of our members, the Amtrak and the different transit agencies, commuter agencies, and along the corridor.”

That is not the way to go about planning something, NYU’s Goldwyn said. “You don’t start with, well what are all the problems I need to fix?” He said the Northeast Corridor’s approach amounts to “I have all these problems, and yeah, if I connect Boston and New York in a faster amount of time, that’s secondary.” 


Goldwyn said the better approach is to do what the Japanese do: “You start with what are the two things I’m trying to connect? And then: What is service going to look like? Is it every hour? What speed? And then you pick your technology.” This is also how projects are generally planned in Germany, Switzerland, and elsewhere in Europe. 

“I’m not saying let that bridge collapse,” Goldwyn said, “But what I am saying is when you do that planning process, you do want to start with, like, what are the goals?”

I asked Warren if he thinks that would be a better approach. He said, “Well, I think it's certainly the owners of the infrastructure who are the ones out there every day, addressing the infrastructure issues that come up and operating the trains. So you know, they have a level of expertise that I think is important to any plan that would be created since they're the ones operating the trains and fixing the tracks. So it's critical that they have an important role and putting the plan together.”

But because there are so many different agencies, jurisdictions, and authorities involved, this means an incoherent, expensive, and ultimately underwhelming plan for the entire corridor is not just an accident, but a virtual guarantee. With no one setting goals, there is nothing for these agencies to be trying to achieve other than to fix what’s broken.

As a result, for the people building and funding these projects, the goals are to rebuild a crumbling rail system, whatever it costs, because it is too important to fail. Nobody I spoke to for this story wants the Northeast corridor to fall apart. They want it to get better. But there is an undeniable insanity to spending $100 billion just to slowly rehabilitate the infrastructure we already have when other countries similar to us are spending much less to make their trains much better.

To be perfectly clear, this $100 billion is a wish list. Some of these projects will never happen, although the bipartisan infrastructure law makes it more likely many of them will. But costs almost never go down after they’re added to a wish list. Generally, the arrow only moves in one direction. And for most other countries in the world, $100 billion would be enough money to build hundreds of miles of new train lines, hundreds of new stations, or something approaching a new high-speed rail network. For the U.S., we are simply hoping $100 billion is enough to shave a few minutes off a train trip while stopping the bridges from falling into a narrow body of water.

Update: A previous version of this article stated that the Hudson River tunnels will be “inactive” while they are rehabilitated. The original wording was needlessly confusing as to whether the tunnels are currently inactive. The two existing tunnels will be shut down one at a time while the rehabilitation work is done. We have removed the word “inactive” to avoid this confusion.