Case study — Travel · 11 min read · 5 questions
Why American public transit sucks
The thesis
American transit does not have a funding problem, and everyone inside it knows that. Federal money for urban mass transit has climbed to nearly 30% of all federal transportation spending — $226 billion in the past decade. At the prices other countries pay, that is the Beijing–Shanghai high-speed line, the Taipei Metro, the Doha Metro and the Riyadh Metro combined. America bought none of them. It bought payroll.
Labor is 24% of operating expenses at Tokyo Metro and 27% at Transport for London, against 59% in New York. American transit workers are the best-paid on earth relative to their own cities, at a 114% premium against 38% in Europe. Overtime is picked by seniority rather than merit, pensions are computed on final-years pay so veterans load up before retiring, and firing someone for theft takes years — if they do not simply retire mid-hearing with a full pension.
None of this is about unions existing, because unions exist everywhere. France has militant unions and a functioning metro; Hong Kong has no recognized union at all. What America lacks is a funder willing to say no. In Berlin, Paris or Tokyo, an agency posting New York's numbers would see management replaced and subsidy withheld pending a recovery plan. In New York the losses are absorbed, so budgets are treated as suggestions and the tab stays open.
How do you think about this? 5 strategy questions this case raises and does not answer.
Read the comments, or add yours →
The statistics
By the numbers — swipe or use arrows
Figures from agency annual reports and National Transit Database filings for the American operators, published accounts for Tokyo Metro, JR East, JR West, MTR, Taipei Metro, SMRT, Transport for London, RATP and their European peers, Federal Highway Administration disbursement data, and OECD general government spending statistics
Key takeaways
The gap is not marginal, it is categorical. Tokyo Metro runs a 15% average operating margin and Hong Kong's MTR 26%, against −182% in New York, −274% in Chicago, −336% in Boston, −377% in San Francisco and −451% in Washington.
Labor is the variable that separates them. It accounts for 24% of operating expenses at Tokyo Metro, 26% in Kansai and 27% at Transport for London, against 59% in New York and San Francisco, 54% in Chicago and 53% in Philadelphia.
Put per dollar earned, the difference is stark. American agencies pay $2.96 to workers for every $1 of revenue generated, against $0.23 in Tokyo, $0.27 in Hong Kong and $0.50 in Paris — and Los Angeles alone pays $5.45.
American transit workers are the best paid in the world relative to their own cities. Compensation runs 114% above the local full-time average in the United States, against 52% across Asia and 38% across Europe — with Washington at 173%, LIRR at 142% and New Jersey at 137%.
The agencies respond to losses by increasing both sides of the cost. Headcount grows 2–5% and wages 2–7% every year for over a decade — New York adding 6.0% to staff and 6.5% to pay annually, Washington 7.2% to pay — against 0.2% headcount growth and 0.6% pay growth at Tokyo Metro.
Productivity is where it all lands. Tokyo moves 210,711 passengers per worker per year, Taipei 146,056 and London 103,587, against 27,250 in New York, 25,041 in Chicago and 10,016 in Los Angeles.
The headcount comparison is absurd on its face. New York employs 74,800 people to carry 1.6 billion riders; Tokyo employs 11,390 to carry 2.4 billion.
Money is not the constraint. $226 billion of federal funding went into urban mass transit over the past decade — against completed overseas systems costing $49 billion for Beijing–Shanghai high-speed rail, $36 billion each for the Taipei and Doha metros, $23 billion for Riyadh and $57 billion to modernize the London Underground.
Highways are not profitable either, and nobody argues they should be. Highway revenues from fuel taxes and tolls covered $136 billion against $265 billion of disbursements in 2022 — a gap that has widened every year and attracts none of the scrutiny transit does.
That has actually reversed. Transit is now 29% of federal transportation spending against 65% for highways, and mass transit funding has risen from $5 billion in 2000 to $28 billion in 2024 — the money arrived and the outcomes did not.
London shows what happens when the subsidy stops. After operating support was withdrawn in 2018, median worker compensation fell 1% a year for nine years and TfL became the only leading European operator whose staff shrank every year since 2017 — with labor now at 27% of expenses.
The waste has a texture beyond the aggregates. Documented cases include a foreman paid nearly half a million dollars while at home during claimed shifts, a janitor taking home $270,000 found napping to inflate overtime, and a customer service representative claiming 100-hour weeks to steal a quarter of a million dollars.
Overtime is the mechanism, not a symptom. At the MTA seniority rather than merit decides who gets first pick of overtime shifts, and because pensions are calculated on the average of an employee's final three years, veterans load up on it before retiring — producing retirees drawing $200,000 a year.
Discipline barely functions. The MTA cannot fire without just cause, cases run months or years, and a worker who files for retirement before a hearing closes the investigation and keeps a full pension — which is exactly what happened in several documented theft cases.
Hiring more people is the standard answer to overtime, and it has never once worked. The MTA has hired roughly 1,000 workers a year for a decade and still spends billions on overtime, while Chicago's CTA is currently promising politicians it will hire its way out of the same hole — where operating losses have already tripled from hundreds of millions to billions.
The documented cases are not isolated lapses. Three electricians and a repairman pooled swipe cards so one person clocked everybody in and out; cleaners billed hundreds of thousands in overtime from home; a manager had sex with a subordinate on the clock. None of them, and none of their supervisors, were fired.
Common questions
Why does American public transit lose so much money?
Labor cost, primarily. It accounts for 59% of operating expenses in New York and San Francisco against 24% at Tokyo Metro and 27% at Transport for London, and American agencies pay $2.96 to workers for every $1 of revenue against $0.23 in Tokyo. Compensation runs 114% above the local full-time average in the US, against 52% in Asia and 38% in Europe — and agencies add 2–5% to headcount and 2–7% to wages annually regardless of results.
Is public transit profitable anywhere?
Yes, and on ordinary fare revenue rather than exotic financial engineering. Tokyo Metro averages a 15% operating margin and takes 90% of its revenue from tickets. Hong Kong's MTR runs 26%. Taipei Metro and MTR both earn over 80% of revenue from fares. The honest caveat is that not every Asian operator is profitable on transit alone — JR West runs −2.0% on transit itself despite a 5.4% overall margin.
How much does the New York subway lose?
About $8.5 billion in 2024, the largest operating loss of any transit agency in the comparison, at a −182% operating margin. It employs 74,800 people to carry roughly 1.6 billion riders a year. Tokyo employs 11,390 to carry 2.4 billion.
Is American transit underfunded?
Not by the numbers. Federal funding for urban mass transit has risen from a couple of percentage points of transportation spending to nearly 30%, from $5 billion in 2000 to $28 billion in 2024, and $226 billion went in over the past decade. At overseas prices that is more than the cost of the Beijing–Shanghai high-speed line, the Taipei Metro, the Doha Metro and the Riyadh Metro combined.
Why is Tokyo Metro profitable?
Productivity and restraint, not a different business model. It moves 210,711 passengers per worker per year against 27,250 in New York, keeps labor at 24% of operating expenses, has grown headcount 0.2% a year since 2017 and median pay 0.6% a year. Its workers still earn 25% more than the median full-time employee in Tokyo. Ninety percent of its revenue is ticket sales.
Do transit agencies overpay their workers?
Relative to their own cities, American agencies pay more than anywhere else in the world — 173% above the local average at Washington's Metro, 142% at the Long Island Rail Road and 137% in New Jersey, against an Asian average of 52% and a European average of 38%. The pattern that matters is not the level but the trajectory: raises of 2–7% a year alongside headcount growth of 2–5%, sustained for over a decade through mounting losses.
Are highways profitable in America?
No, and the asymmetry in how that is discussed is the point. Highway revenues from fuel taxes, vehicle taxes and tolls came to $136 billion in 2022 against $265 billion of disbursements — a gap that has widened nearly every year since 2010. Roads are not expected to break even, and transit is held to a standard nothing else in American transportation is held to.
Why do Asian cities have better public transit?
Not because they spend more of their economy on government. Singapore's public spending is 20% of GDP, Taiwan's 22%, Hong Kong's 13% — all well below America's 36%. The difference is in how the agencies are run: labor at roughly a quarter of expenses rather than well over half, headcount that grows under 1% a year, and 100,000 to 200,000 passengers moved per worker rather than 10,000 to 27,000.
Did car companies destroy American public transit?
They contributed materially. Automakers lobbied for the 1956 Interstate Highway Act and bought and dismantled existing streetcar networks across American cities, leaving millions with no commuting alternative to car ownership. As late as 1967, urban mass transit was 1% of federal transportation spending against 97% for highways. But that history explains the starting position, not the present — transit is now 29% of federal transportation spending, and the performance gap persists.
What happened when London cut transit subsidies?
Costs came down and workers absorbed it. Operating subsidies of around £700 million a year were withdrawn in 2018 on the argument that Transport for London should stand on its own after billions in modernization. Median worker compensation has since fallen 1% a year for nine years, TfL became the only leading European operator to shrink its staff every year since 2017, and labor is now 27% of operating expenses — on a par with the Japanese private operators. The UK resumed subsidies in 2020.
What do Asian transit systems do differently to stay sustainable?
Four things, none of them exotic. They hold headcount almost flat — Tokyo Metro has grown staff 0.2% a year since 2017 and JR East has been shrinking its transit workforce by about 5% a year. They move vastly more people per employee: 210,711 a year in Tokyo and 146,056 in Taipei against 27,250 in New York. They keep labor to roughly a quarter of operating expenses rather than well over half. And they still earn most of their money the ordinary way — Tokyo Metro takes 90% of revenue from tickets, Taipei and Hong Kong over 80% — so the property-development folklore explains less than people assume.
How do Asian and European agencies control labor costs when American ones cannot?
Mostly by facing a real budget. In Europe and Asia, subsidies come attached to contracts, performance targets and oversight; if an agency posted New York's numbers, management would be replaced, the subsidy adjusted down and further funding withheld pending a recovery plan. London is the clean experiment: operating support of around £700 million a year was withdrawn in 2018, and despite four years of escalating strikes, median compensation has fallen 1% a year, staff shrank every year since 2017, and labor is now 27% of expenses. American agencies pile up losses, get bailed out without consequence, and add 2–5% to headcount and 2–7% to wages annually regardless.
Why are American transit labor costs so high when France and Germany also have strong unions?
Because union strength is not the variable — what unions bargain for, and what constrains them, is. France has 35-hour weeks and genuinely militant unions, but transit strikes are legally limited to short actions announced weeks ahead and barred during rush hour, so service never fully stops and the leverage is capped; RATP's unions spend it on pensions, jobs and resisting automation rather than on open-ended wage escalation. Japanese and German unions bargain cooperatively — Berlin has struck twice since 2008, both times to block layoffs. Hong Kong cuts the other way and proves the point: no recognized union, no collective bargaining law, and workers have still received 6% raises a year for nine years. What is distinctive in the United States is the surrounding structure — overtime picked by seniority, pensions computed on final-years pay, rigid job classifications, near-impossible dismissal, and no funder willing to withhold money — which turns ordinary bargaining into an open tab.
Are American transit unions corrupt?
The documented pattern is less dramatic than corruption and more corrosive: abuse that is technically permitted and effectively unpunishable. A foreman drew nearly half a million dollars while at home during shifts he claimed to be working and faced no punishment. A janitor took home $270,000 and was found napping in a storage area to run up overtime — he is still employed a decade later. Three electricians and a repairman pooled swipe cards so one of them clocked everyone in. Cleaners billed hundreds of thousands in overtime from home. In none of these cases were the workers or their supervisors fired, because the MTA cannot dismiss without just cause and a case can take years — long enough for the employee to file for retirement, which closes the investigation and preserves the pension.
Why doesn't hiring more workers reduce overtime?
Because overtime is not a staffing gap, it is a compensation strategy. Pensions are calculated on the average of an employee's final three years, so senior staff have a direct financial interest in accumulating as much overtime as possible before retiring — producing retirees drawing $200,000 a year. Seniority rather than merit decides who gets first pick of those shifts, so the people with the strongest incentive also have first claim. The MTA has hired around 1,000 people a year for a decade and still spends billions on overtime. Chicago's CTA is currently making the same promise to politicians, and its losses have already tripled.
Could automation fix American transit?
Not on its own, and not soon. Moving thousands of people through a century-old system involves enough edge cases and enough safety exposure that automation is genuinely hard everywhere, not only in America. But the more binding obstacle is that the parties who would have to implement it are the ones it threatens: RATP's unions in Paris spend their leverage specifically on resisting automation and outsourcing, and American work rules lock tasks into rigid classifications that exist to protect headcount. Technology does not resolve a problem whose cause is that nobody can be told no.
Discussion
Federal money for urban mass transit reached $226 billion over a decade — enough, at other countries' prices, to buy four major metro systems outright. America bought payroll instead. How does a system arrive there without anyone deciding to?
No answers yet — be the firstLabor is 24% of operating expenses at Tokyo Metro and 27% at Transport for London, against 59% in New York. Which of those numbers is the anomaly, and what would it take to move it?
No answers yet — be the firstOvertime is allocated by seniority rather than merit and pensions are computed on final-years pay, so veterans load up at the end. Design a better rule — then name who would fight it and why they might be right.
No answers yet — be the firstIf the money is there and the outcome is not, what is the actual constraint? Test at least two candidates before settling.
No answers yet — be the firstYou are given control of a large American transit agency and a mandate to fix it. What do you do in year one, knowing you cannot cut wages?
No answers yet — be the first
Related case studies
TravelWhy car rental is such a bad businessMargins as thin as an airline, and the companies make more money selling the cars than renting them out.8 min read
TravelWhy Asian airlines are so much betterAsian carriers deliver a service Americans barely recognize. American airlines earn more doing far less.8 min read
TravelWhy Chinese EVs are so far aheadCheaper, faster and better equipped than anything the West sells, and cheap labor has nothing to do with it.8 min read