But scant attention has been paid to operating costs, meaning the cost of running and maintaining the subway system rather than expanding it. New York City Transit’s subway operating costs are high by both domestic and international standards—about 60 percent higher than those of the largest European systems, and 90 percent higher than in Chicago.
Urbanize.LA has covered some of these issues, since Los Angeles's urban rail operating costs are even higher than those of the New York City subway. The metrics used are operating and maintenance costs per car-mile and car-hour; this is because train maintenance needs and energy costs generally scale with the number of cars, and only the train operator is paid for the entire train.
For an idea of how New York compares to the rest of the world, the chart below looks at data from comparable transit systems in the U.S. and major cities around the world; the American data comes from the National Transit Database, whereas the European data is from the international Community of Metros (CoMET), which also includes New York City Transit.
Subway system operating costs
|Cost per mile
|Cost per hour
|Cost per mile
|Cost per hour
|Los Angeles Metro Rail
|New York City Subway
|Philadelphia SEPTA subway
|Paris Metro and RER
CoMET lists data separately for operating costs (excluding maintenance), track maintenance, and train maintenance. New York's train maintenance costs are not out of line, but in the other two categories it is far and away the most expensive.
Why are New York’s subway operating costs so high?
The reason why track maintenance is so costly is simple: the NYC subway operates 24/7. While other systems, such as London’s Underground and the Washington D.C. Metro, shut down overnight for maintenance, NYCT does not. Instead, it has to do track work with partial weekend shutdowns, with workers on one track and running trains on the adjacent track. For safety, every time a train passes on adjacent track, the track workers have to pause what they're doing.
This makes every task take much longer than it has to, which in turn drives up costs. NYCT recognizes this is a problem, which is why implemented its Fastrack program in 2011, shutting down some lines on some nights to speed up repairs.
But the bulk of costs come from operations, not maintenance—and there, New York has a different set of challenges. As a high-income city, New York naturally has to pay high wages; NYCT's total labor cost, including benefits, is about $140,000 per year per worker, compared with $108,000 in Chicago and Los Angeles.
Still, a bigger difference between New York and other cities comes from the number of workers required to move the same number of trains. In his masters' thesis, Kyle Kirschling, now a planner at NYCT, studied operating costs and productivity on New York’s subway and elevated lines going back to 1870. The number of annual revenue car-miles per subway employee in New York was 14,000 in 2010. In Chicago this number is somewhere between 14,000 and 16,000, depending on how one allocates administrative workers between rail and bus. On Tokyo’s Metro, the comparable figure is about 18,500. The trains in Tokyo are famously overcrowded, but Tokyo has high productivity per car-mile, not just per passenger.
What's more, worker productivity in New York was higher decades ago. Per Kirschling's thesis, the system got about 16,000 car-miles per worker in 1935. This figure declined steeply in the subsequent 20 years; in the 1970s and ’80s, it was just 9,000. Since the subway system's nadir in the 1980s there has been significant improvement in productivity, but the system still lags.
How did the NYC subway get here?
First, New York has two-person crews on most subway trains, consisting of a train operator and a conductor. This is uncommon worldwide; among the biggest subway networks, only Toronto and some lines in Tokyo have two-person crews. Elsewhere, crews make do with just one driver, who also operates the doors.
In New York the technology for this, called one-person train operation, or OPTO, was tested on the L train when it got signal upgrades in 2007, but the Transportation Workers' Union, which represents New York's subway workers, fought to restore two-person crews. But OPTO is ubiquitous in Europe, East Asia, and other American cities, even on lines without the advanced signaling that the L train has in New York.
Moreover, the crews are less efficient than those in other cities. A train operator in New York is driving a revenue train for only about 550 hours a year. The comparable figure in London is 720; in Helsinki, it’s 867. In Toronto, with two-person crews, the figure is about 1,400 or 1,500; in Chicago, it can be as high as 2,000.
This is because of how managers can schedule trains' revenue and non-revenue trips. Chicago, for example, achieves its high productivity because it has very little deadheading; an “L” train is carrying fare-paying passengers for almost every mile it travels, whereas a New York subway, with a driver collecting wages, spends more time in non-revenue service heading to and from the railyard.
Any rail network with high frequency at peak times and low frequency in less busy hours would have problems scheduling drivers. It would have to give workers paid idle time or rely on unpopular split shifts, in which workers work a few hours in the morning and a few in the evening, rather than eight continuous hours.
In New York, peak frequency on many subway lines is almost twice as high as off-peak frequency; for example, on the 7 line, it’s every 2.5 minutes at rush hour and every five minutes in off-peak times. In London and Toronto the ratio of peak to off-peak frequency is much lower, and in Helsinki, the Metro runs at the same frequency from opening to closing time, simplifying scheduling.
If New York ran more trains in the off-hours, operating costs would of course rise, but not in proportion to the increased service, so per-mile expenses would fall.
Finally, New York suffers from slow trains. The average speed is not so low by the standards of some older systems, such as the Berlin U-Bahn (which opened in 1902, just two years before the first NYC subways). But it is low by the standards of what the city could achieve. Average scheduled time is higher now than it was in 1906, thanks to slowdowns since the 1990s, some of which are about safety and others of which are about the pretense of safety.
The trains increasingly cannot make even their scheduled time. An internal document circulated within NYCT, leaked independently to Curbed and to the New York Daily News, attacks the agency’s management for focusing on the wrong punctuality metrics, and estimates that the average passenger on the subway experiences 1.5 minutes of delay per trip (not including 1.5 minutes of extra time spent waiting for the train on the platform). That equals about eight percent of the average trip time. If NYCT succeeds in eliminating those delays, it will boost speed by eight percent, reducing operating costs per mile.
Those delays are sometimes the result of poor dispatching, and sometimes the result of train operators going more slowly because of confusing signals; a planner at NYCT told Curbed that the L, the most automated line in the system, has the fewest delays.
How can New York increase subway efficiency?
NYCT can’t fix everything about its out-of-control operating expenses; it cannot shut down overnight, or cut wages until it couldn't get workers to fill jobs. But there are things it can do to become more efficient.
It can transition to OPTO, and increase off-peak train frequency to improve scheduling; this would also absorb conductors made redundant by OPTO. These measures would improve service for free. NYCT could also expand Fastrack to reduce track maintenance costs, and make more of an effort to reduce train delays.
If it could reduce its operating costs per mile to those of the London Underground, Chicago “L,” or Paris Metro, NYCT would no longer need direct operating subsidies and be on a stabler, more fiscally sustainable path; it would then have more money to invest in its capital plan, which is crucial in the years to come.
Alon Levy grew up in Tel Aviv and Singapore. He spent ten years in math academia and has blogged at Pedestrian Observations since 2011, covering public transit, urbanism, and development. Now based in Paris, he writes for a variety of publications, including Vox, Streetsblog, Voice of San Diego, PlanPhilly, Urbanize.LA, Railway Gazette, and the Bay City Beacon. You can find him on Twitter @alon_levy.