Brendan McDermid / Reuters

The New York City Subway Is Beyond Repair

Last weekend, Peter Wayner advocated for a radical overhaul of the city’s current subway system, proposing, instead, a network of subterranean highways filled with hoverboards, scooters, and autonomous vehicles.

After reading “The New York City Subway Is Beyond Repair,” I felt compelled to respond to what I see as basic inaccuracies that undermine the piece as a whole.

I have a degree in infrastructure engineering and am an engineer-in-training in the field, but the inaccuracies in question are not nearly so arcane as to require such credentials.

The respective capacities of free-flowing vehicular lanes and subway transit are well established. Generously, a freeway lane might carry 2,000 vehicles per hour, which—again, generously—might each carry somewhere between one and two travelers, on average. This gives us a high-end estimate of moving 2,000 to 4,000 people per hour in a single direction.

An MTA subway track, such as Mr. Wayner effectively proposes to replace with a single lane of traffic, is capable of carrying in excess of 30,000 people per hour. This is not a small difference and makes one wonder whether the author has considered it.

This is the simplest critique, as it relies on simple math, but the challenges with Mr. Wayner’s proposal are legion. For the geometry alone, there are a number of difficulties with using passenger vehicles rather than trains.

Vastly larger stations would be required to accommodate all the spaces for cars picking up and dropping off pedestrians. In order to prevent delays for vehicles not stopping, additional bypass tunnels would need to be excavated at every station. To permit safe operation, much of the signaling equipment that Mr. Wayner wanted to rip out would instead need to be replaced with much more sophisticated and expensive Intelligent Transportation Systems (ITS) to coordinate a far larger number of vehicles.

These are all massive challenges, with huge price tags, that would, again, result in a tremendous decrease in capacity for the system as a whole. It’s a pleasant fantasy to believe that innovation and a Silicon Valley mindset are all that’s necessary to solve one of America’s most intractable infrastructure challenges. The truth, much less attractive, is that it requires massive and consistent funding, collaboration across a range of stakeholders, and time.

Patrick Zerr
North Vancouver, Canada

Peter Wayner proposed that we should replace trains in the New York City subway with autonomous cars and hoverboards. He sounds like a modern-day Robert Moses, obsessed with automobiles as a replacement for public transit, proclaiming that cars will be more convenient and efficient for us all. Robert Moses built a system of roads on which only cars and trucks may travel—which are and have always been notoriously congested, especially during the rush hours. Subway riders must tolerate stops at stations they will not get off at because other passengers might be boarding or exiting the train, but automobile passengers must tolerate stops for traffic jams that serve no purpose for anyone.

The New York City subway is not broken beyond repair; for all its faults, for all the mismanagement, the subway remains the lifeline of this city. Mr. Wayner suggests that modern technology—autonomous cars, personal transit devices—can replace trains, but it is not as if trains have not benefited from modern technology as well. Computers, artificial intelligence, robotics—all these things are being applied to railroads, improving efficiency and reducing costs for both passenger and freight systems. There is a lot of potential in the subway; unlike most other metro systems that are double-tracked, the New York City system has numerous triple—and quadruple—tracked lines, which are currently used to allow express trains to pass local trains but which could be used to even greater effect with more modern control systems (for example, to allow a super-express service that skips more stops).

It is also important to remember that the subway system provides service to neighborhoods that are currently underserved by taxis, and which would almost certainly be underserved by autonomous cars operated by competing, for-profit companies. The most profitable places to serve will be in the city center where there are always people waiting to ride the vehicles; but the people most in need of subway service tend to live far from the city center, and to pick them up the vehicles will be forced to make long and unprofitable trips without passengers, just like subway trains. The reason the government took over the subway and commuter railroads was to maintain a vital but unprofitable service.

Benjamin Kreuter
Jersey City, N.J.

Several readers responded on Facebook:

Isaac Brumer wrote: For all its problems, the NYC transit system is not “beyond repair.” It safely serves millions of people every day, 24/7. How will those people get around while you’re ripping out the tracks, repairing the tunnels, then retrofitting them for the transportation system you’ve dreamed up? And all transportation systems need costly maintenance over time. Does the author believe the new system will be maintenance-free?

Andrea Abarca Coutts wrote: What ever happened to all the flying cars and buses I was promised by movies and TV? Eight years old me definitely thought we’d be hovering around cities by now.

Several readers responded on Twitter:

Peter Wayner replies:

Mr. Zerr and Mr. Kreuter make the same mistake that many do by assuming that the autonomous vehicles will flow like human-driven cars. Consider as a thought experiment a line of hoverboards a mile long with 10 feet of empty space behind each one. That’s 528 people. If they move 15 miles an hour, the tunnel will carry 7,290 people per hour.

Hoverboards are slim and we can slice the tunnel into three, four, or maybe five lanes and carry 23,760, 31,680 or 39,600 people per hour. Thinner lanes are a big advantage because a mishap or planned maintenance in one slim lane wouldn’t block everything. That speaks to Mr. Brumer’s point.

Interested readers can repeat the same experiment with two lanes of five-foot-wide autonomous cars spaced 20 feet apart, carrying four passengers and going 30 miles an hour. These can offer airbags and other safety features missing from trains.

These are just two hypothetical models that could deliver the same throughput as the one train—when it can follow its official rush hour schedule of  approximately sixteen trains per hour.

Why can these autonomous vehicles compete with one train that can carry 2,000 people? Local trains need large gaps because the stops take so long. Autonomous vehicles will stop only at their destination, when they zip out of the flow. They can use the large gap.

As to Mr. Kreuter’s point about the outer boroughs, the competing fleets may still be owned by the city or heavily regulated. We can have both plenty of choices and a differential pricing model with room to help whomever the politicians favor.

I agree that we need to sweat many details, but there’s plenty of opportunity. The hoverboards might zip over to the next line or carry the passenger to the street. The corridors and roads upstairs are fair game. They could give the handicapped more time to board than the subway at rush hour.

There’s also plenty of room. The platforms are 600 feet long and Disney loads their rides in much less space.  

And there’s also plenty of budget. We’re already being asked to shoulder a $19 billion bill. We can either redesign the signaling and everything else for the last generation or aim for the future.