Everyone in New York has a commute story, and most of them are about time. The neighbour who leaves at 5:30 to be in Midtown by 7. The home health aide on two buses to the Bronx. The Staten Islander for whom “the city” is a boat and a train away. The stories are so familiar that the number underneath them rarely gets a second look.
The Census Bureau looks every year. In its 2024 survey, the average New York City resident’s trip to work took 40.6 minutes, one way, against 27 minutes for the country as a whole.
The average hides the number we’re really after. One city commuter in four, about 900,000 people, spends an hour or more each way. Nationally it is one in eleven.
You can make that number say almost anything. A transit advocate hears that the outer boroughs are stranded; a driver hears that the subway is broken; a real-estate column hears a reason to move to Westchester. The census’s own answer is plainer, and it has held for a generation: the hour is a matter of address and vehicle. It falls on people riding a subway or a bus from the outer boroughs toward Manhattan or across to another borough, and far less on low pay than you might expect.
So instead of picking a side, let’s look at who is actually spending the hour: where they live, how they travel, what they earn, and whether it has always been like this. The numbers come from the census’s own microdata, the survey records behind the published table, which run through 2024, the latest year the Census Bureau has released; the MTA’s ridership counts, to September 2026, and its own measures of bus speed and subway delay, to August, carry the story into 2026. Every chart is interactive, and the raw data and code sit at the end.
The hour lives at the edges of the city
Start with the picture the Bureau publishes. Set aside the workers who usually work from home, and the city has about 3.6 million Who counts as a commuter Residents of the five boroughs aged 16 and over who were at work in the survey week and did not usually work from home. That is the Census Bureau’s own universe for its travel-time tables. It counts people by where they live, so someone living in Queens and working in Newark is in; someone living in Newark and working in Manhattan is not. , people who actually travel to a job. Their trips pile up around the half-hour mark, and then comes a long tail: a quarter of them take an hour or more.
Put those trips on a map and the long tail has an address. Neighbourhood by neighbourhood, the share over the hour runs from about one commuter in twenty in lower and midtown Manhattan to nearly four in ten in Canarsie and Flatlands, at the far end of Brooklyn, with southeast Queens, the Rockaways and much of the Bronx close behind.
Where the hour-plus commute lives
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.Commuters living in each of the city's 55 census neighbourhoods (PUMAs), 2020-24
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
Neighbourhoods are the Census Bureau’s public-use microdata areas, which mostly follow the city’s community districts. Each rests on about one to five thousand survey records from the 2020–24 five-year file, so the shares carry margins of one to three points; hover a neighbourhood to see its own. The colour classes are equal steps with round edges, and the triangle on the legend marks the whole city. Pick a borough to rank its neighbourhoods in the side panel. The distribution of trip lengths, band by band, is in the questions at the end.
The numbers behind this chart
About these numbers. American Community Survey 2020-2024 five-year microdata: resident workers of the five boroughs who did not work from home, by the neighbourhood (public-use microdata area, PUMA) they live in. Travel time is the usual one-way trip the respondent reports; the mode is the one used for most of the distance. Every figure carries a 90% replicate-weight margin, shown on hover. Each neighbourhood's figures rest on 900 to 5,100 survey records. Shapes: Census Bureau cartographic boundary file, 2020 PUMAs, clipped to the shoreline.
| Neighbourhood | Borough | An hour or more | Average trip | Subway or bus | Work in Manhattan |
|---|---|---|---|---|---|
| Midtown, East Midtown, & Flatiron | Manhattan | 5.6% ± 1.0% | 27 min ± 1 min | 45% ± 2.2% | 87% ± 1.4% |
| Fordham, Bedford Park, & Norwood | Bronx | 36% ± 2.9% | 46 min ± 2 min | 64% ± 2.7% | 33% ± 2.8% |
| Pelham Parkway & Morris Park | Bronx | 32% ± 2.6% | 41 min ± 2 min | 45% ± 2.5% | 31% ± 2.1% |
| Crown Heights (North) | Brooklyn | 21% ± 1.9% | 41 min ± 1 min | 66% ± 2.8% | 46% ± 2.7% |
| Co-op City & Throgs Neck | Bronx | 35% ± 2.5% | 45 min ± 1 min | 40% ± 2.6% | 32% ± 2.3% |
| Highbridge & Concourse | Bronx | 28% ± 2.5% | 43 min ± 1 min | 66% ± 2.6% | 43% ± 3.2% |
| Crown Heights (South) | Brooklyn | 22% ± 2.6% | 40 min ± 1 min | 64% ± 3.0% | 40% ± 2.5% |
| Harlem | Manhattan | 16% ± 2.2% | 38 min ± 2 min | 72% ± 2.5% | 79% ± 2.6% |
| Bushwick | Brooklyn | 20% ± 2.1% | 41 min ± 1 min | 68% ± 2.6% | 49% ± 3.0% |
| Washington Heights & Inwood | Manhattan | 22% ± 2.4% | 40 min ± 1 min | 66% ± 2.4% | 76% ± 2.2% |
| Sunset Park & Windsor Terrace | Brooklyn | 26% ± 2.1% | 43 min ± 1 min | 57% ± 2.3% | 36% ± 2.2% |
| Morningside Heights & Hamilton Heights | Manhattan | 12% ± 2.1% | 33 min ± 1 min | 63% ± 2.7% | 84% ± 2.4% |
| Jackson Heights & East Elmhurst | Queens | 27% ± 1.8% | 42 min ± 1 min | 55% ± 3.0% | 41% ± 2.3% |
| Lower East Side & Chinatown | Manhattan | 9.8% ± 1.8% | 30 min ± 1 min | 54% ± 2.6% | 83% ± 1.9% |
| Upper East Side & Roosevelt Island | Manhattan | 7.2% ± 1.1% | 30 min ± 1 min | 54% ± 2.1% | 87% ± 1.5% |
| South Ozone Park & Howard Beach | Queens | 33% ± 2.4% | 45 min ± 1 min | 37% ± 2.6% | 24% ± 2.1% |
| Coney Island & Brighton Beach | Brooklyn | 33% ± 2.9% | 44 min ± 2 min | 47% ± 3.1% | 27% ± 2.7% |
| Canarsie & Flatlands | Brooklyn | 39% ± 1.9% | 48 min ± 1 min | 46% ± 2.1% | 26% ± 1.9% |
| Bensonhurst & Bath Beach | Brooklyn | 34% ± 1.6% | 44 min ± 1 min | 53% ± 1.9% | 29% ± 1.8% |
| Flatbush & Midwood | Brooklyn | 29% ± 1.9% | 44 min ± 1 min | 58% ± 2.3% | 36% ± 2.5% |
| Williamsburg & Greenpoint | Brooklyn | 12% ± 1.4% | 35 min ± 1 min | 59% ± 2.0% | 51% ± 2.1% |
| Astoria & Queensbridge | Queens | 20% ± 1.7% | 39 min ± 1 min | 67% ± 1.6% | 56% ± 1.7% |
| Elmhurst & Corona | Queens | 28% ± 2.0% | 43 min ± 1 min | 64% ± 2.1% | 42% ± 1.7% |
| Ocean Hill & Brownsville | Brooklyn | 29% ± 3.4% | 43 min ± 2 min | 64% ± 3.8% | 27% ± 3.2% |
| Morris Heights & Mount Hope | Bronx | 28% ± 3.1% | 42 min ± 2 min | 68% ± 3.5% | 39% ± 3.0% |
| Borough Park & Kensington | Brooklyn | 22% ± 1.5% | 37 min ± 1 min | 37% ± 2.1% | 21% ± 1.4% |
| Downtown Brooklyn & Fort Greene | Brooklyn | 11% ± 1.4% | 34 min ± 1 min | 64% ± 2.3% | 56% ± 1.9% |
| Queens Village, Bellerose, & Rosedale | Queens | 34% ± 2.0% | 46 min ± 1 min | 30% ± 1.8% | 22% ± 1.6% |
| South Shore | Staten Island | 33% ± 1.9% | 45 min ± 1 min | 19% ± 1.7% | 19% ± 1.8% |
| East New York & Cypress Hills | Brooklyn | 35% ± 1.9% | 46 min ± 1 min | 56% ± 2.4% | 25% ± 1.9% |
| Soundview & Parkchester | Bronx | 36% ± 2.2% | 46 min ± 1 min | 56% ± 2.5% | 36% ± 2.5% |
| Forest Hills & Rego Park | Queens | 32% ± 2.5% | 45 min ± 1 min | 56% ± 2.5% | 45% ± 2.3% |
| Mid-Island | Staten Island | 32% ± 2.1% | 43 min ± 1 min | 24% ± 2.1% | 22% ± 2.3% |
| Jamaica, St. Albans, & Hollis | Queens | 37% ± 1.5% | 47 min ± 1 min | 44% ± 1.7% | 23% ± 1.4% |
| East Harlem | Manhattan | 12% ± 1.8% | 34 min ± 1 min | 66% ± 3.1% | 82% ± 3.1% |
| Bedford-Stuyvesant | Brooklyn | 22% ± 1.9% | 41 min ± 1 min | 63% ± 2.3% | 40% ± 2.1% |
| Upper West Side | Manhattan | 7.1% ± 1.5% | 30 min ± 1 min | 65% ± 2.5% | 85% ± 1.8% |
| Sheepshead Bay & Gravesend (East) | Brooklyn | 34% ± 1.9% | 43 min ± 1 min | 44% ± 2.0% | 32% ± 2.1% |
| Bay Ridge & Dyker Heights | Brooklyn | 30% ± 2.0% | 43 min ± 1 min | 51% ± 2.3% | 33% ± 1.6% |
| Long Island City, Sunnyside, & Woodside | Queens | 16% ± 1.9% | 37 min ± 1 min | 65% ± 2.6% | 56% ± 2.2% |
| Kew Gardens, Richmond Hill, & Woodhaven | Queens | 34% ± 2.0% | 45 min ± 1 min | 48% ± 2.4% | 30% ± 2.0% |
| Wakefield, Williamsbridge, & Eastchester | Bronx | 32% ± 2.4% | 43 min ± 2 min | 46% ± 2.8% | 28% ± 1.6% |
| Chelsea & Hell's Kitchen | Manhattan | 5.5% ± 1.4% | 26 min ± 1 min | 45% ± 3.4% | 88% ± 2.0% |
| Melrose, Mott Haven, Longwood, & Hunts Point | Bronx | 24% ± 2.6% | 40 min ± 2 min | 62% ± 2.9% | 38% ± 2.7% |
| Auburndale, Bayside, & Douglaston | Queens | 27% ± 2.3% | 41 min ± 2 min | 20% ± 2.1% | 21% ± 1.8% |
| Riverdale, Kingsbridge, & Marble Hill | Bronx | 34% ± 3.0% | 44 min ± 2 min | 50% ± 3.1% | 42% ± 2.7% |
| Ridgewood, Maspeth, & Middle Village | Queens | 31% ± 1.9% | 44 min ± 1 min | 49% ± 1.6% | 33% ± 1.7% |
| East Flatbush | Brooklyn | 34% ± 2.4% | 46 min ± 1 min | 61% ± 2.1% | 33% ± 2.2% |
| Morrisania, Tremont, Belmont, & West Farms | Bronx | 29% ± 2.5% | 43 min ± 2 min | 58% ± 2.7% | 34% ± 3.0% |
| Financial District & Greenwich Village | Manhattan | 4.0% ± 0.7% | 25 min ± 1 min | 50% ± 2.6% | 88% ± 1.6% |
| Flushing, Murray Hill, & Whitestone | Queens | 31% ± 1.8% | 41 min ± 1 min | 33% ± 1.8% | 21% ± 1.3% |
| Park Slope & Carroll Gardens | Brooklyn | 11% ± 1.4% | 37 min ± 1 min | 63% ± 2.6% | 58% ± 2.3% |
| Fresh Meadows, Hillcrest, & Briarwood | Queens | 31% ± 2.5% | 42 min ± 1 min | 38% ± 2.4% | 26% ± 2.1% |
| North Shore | Staten Island | 30% ± 2.0% | 41 min ± 1 min | 25% ± 1.7% | 20% ± 1.6% |
| The Rockaways | Queens | 37% ± 3.0% | 47 min ± 2 min | 32% ± 3.2% | 21% ± 2.5% |
By borough, that is one commuter in ten in Manhattan and close to one in three in Queens, the Bronx and Staten Island, with Brooklyn in between.
Now cross that with how people travel, and the picture sharpens. Among drivers, one in five or fewer is over the hour in every borough, Manhattan included. Among subway riders it is 11% in Manhattan and 50% in the Bronx, and the chart fills in the boroughs between and beyond. The hour-plus commute is not a New York thing in general. It is one specific thing: a train or a bus from the edge of the city toward its middle.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per borough of residence; the mode filter is the census’s single The one usual mode The census asks for the single means of transportation usually used for most of the distance. A bus ride to the subway counts once, under whichever leg was longer. So “bus” here means people whose main leg is a bus, and a two-seat trip through a transfer is invisible. . Switch the measure to median minutes or the share working in Manhattan.
Where the trip goes completes the picture. More than half of all city commuters work in their own borough, and their typical trip is half an hour. Crossing into Manhattan puts four in ten over the hour, and so does crossing from one outer borough to another; for Bronx residents who work in another outer borough it is more than half. The network was built to carry people to Manhattan and back, and the trips that run around it take just as long.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per Place of work The census records where the job is at the level of a place-of-work area, which in New York City is the borough. Here it is grouped as Manhattan, the commuter’s own borough, another borough, or outside the city. Nothing finer than the borough is used. ; switch between the share of commuters, the hour-plus share and the median.
The address matters more than the paycheck
You might expect the long commute to be a low-wage problem, and it is, but not in the way the phrase suggests. Sort commuters by their Own earnings The worker’s own wages, salary and self-employment income over the past twelve months, in 2024 dollars. Not household income: a part-time earner in a well-off household sits in the lowest band. The bands are fixed dollar cuts rather than quartiles, so they can be compared across boroughs. and the hour-plus share is about a quarter in every band up to $100,000 and a little lower up to $150,000; only above that does it drop sharply, to 15%.
Look inside a single borough and even that drop mostly disappears: in Queens and in the Bronx, about three in ten commuters are over the hour in every earnings band. What the top earners have is mostly a Manhattan workplace, seven in ten of them, and a short subway ride to it. Below $150,000, then, earnings barely distinguish the hour-plus commuter. Borough and mode do. The jobs with the longest commutes, construction, cleaning and protective service, are by and large the ones that cannot be done from home, and the FAQ has that chart.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per earnings band, in 2024 dollars. Switch the measure to the share working in Manhattan to see what is different about the top band.
The same geography produces the starkest number here. Black commuters are over the hour at 31% and white commuters at 18%, with Hispanic and Asian commuters in between.
Part of that is where people live: nine in ten Black commuters live outside Manhattan, against seven in ten white commuters. But the gap mostly stays when you hold the borough fixed, and among people riding the same subway it is wider. Black commuters also take the bus, the slowest way to ride in the city, at more than twice the rate of white commuters.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per race or ethnicity group (Hispanic of any race; the others non-Hispanic). The filters hold borough, mode and earnings fixed one at a time; the next chart holds all of them at once.
Filters move one thing at a time. To ask how much of each gap remains when several characteristics are held fixed together, we fitted one survey-weighted model of the chance that a trip is over the hour, on borough, mode, place of work, earnings, race, occupation, age, sex, nativity and degree. It is a descriptive adjustment, not a causal model, and it can only hold fixed the characteristics it includes.
The results sort the findings into two piles. Mode and geography stand on their own: with the model’s other characteristics fixed, riding the subway or the bus rather than driving still adds about 20 points to the hour-plus share. Occupation mostly does not: most of its raw gaps disappear. Earnings and race sit between: the top earners’ raw advantage halves, and the raw Black–white difference of 14 points shrinks to about 6 points once where people live, how they travel and where they work are held fixed.
That attenuation is substantial, but it is not a decomposition of the gap into causes: the covariates overlap with one another and with race, and the model says only that a residual remains after standardising on the characteristics it includes. The census records no station, transfer or service frequency, so it cannot say what that residual is. Outside work on door-to-door transit access finds Black New Yorkers can reach about 16% fewer jobs within an hour than white New Yorkers, which points at the network itself.
Modelled A statistical model fitted to survey records. It is descriptive, not causal: it holds fixed what the survey records and nothing else, and its residual is not an explanation.
How to read this, and what it cannot show
Pick a factor to compare. Each bar is the percentage-point difference against the reference group named in the note; the FAQ describes the model.
The pattern has held for thirty-four years
So has it always been like this? The census has asked how long the trip takes, in the same words, since 1990, so the same count can be run on every census and survey since. In 1990, 23% of city commuters were over the hour. The share climbed to a peak of 28% in 2017, the year the subway’s reliability collapsed and the governor declared a state of emergency. After the pandemic’s odd interval it has settled at 25%, where it was in 2005.
For thirty-four years, the hour has been an outer-borough commute
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.NYC resident commuters by borough of residence, 1990 census to 2024 ACS
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One line per borough of residence and one for the whole city: the 1990 and 2000 censuses, then the annual survey from 2005, with 90% margins where the files carry replicate weights. Switch to the average trip in minutes, or hold the mode fixed; 2020 used the Bureau’s experimental pandemic weights and is drawn without a margin, because the Bureau warns that its replicate weights understate the uncertainty. The same series as bars, with working from home and the mode shares by year, is in the questions at the end.
The numbers behind this chart
About these numbers. Each year is its own sample tabulated the same way. Working-from-home and mode shares are available for the whole city only. The 2020 file uses the Bureau's experimental pandemic weights, and its replicate weights understate the uncertainty, so it carries no margin. The bar chart in the questions below has the same series with the working-from-home and mode shares.
| Group | 1990 | 2000 | 2005 | 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 (experimental weights) | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All five boroughs | 23% | 25% | 25% ± 0.7% | 26% ± 0.5% | 25% ± 0.6% | 26% ± 0.5% | 25% ± 0.5% | 24% ± 0.5% | 25% ± 0.6% | 24% ± 0.5% | 25% ± 0.5% | 25% ± 0.4% | 26% ± 0.5% | 27% ± 0.5% | 28% ± 0.5% | 27% ± 0.6% | 27% ± 0.5% | 25% | 25% ± 0.6% | 26% ± 0.6% | 25% ± 0.5% | 25% ± 0.5% |
| Manhattan | 8.6% | 9.4% | 10% ± 1.0% | 11% ± 0.9% | 9.4% ± 0.8% | 10% ± 1.0% | 9.3% ± 0.7% | 10% ± 0.7% | 10% ± 1.1% | 11% ± 1.0% | 9.6% ± 0.8% | 9.8% ± 1.0% | 11% ± 1.0% | 10% ± 0.8% | 13% ± 1.0% | 11% ± 1.0% | 9.7% ± 1.0% | 10% | 11% ± 1.3% | 10% ± 1.2% | 9.2% ± 1.0% | 9.3% ± 0.9% |
| Brooklyn | 27% | 29% | 28% ± 1.3% | 29% ± 1.1% | 29% ± 1.0% | 28% ± 1.0% | 26% ± 1.0% | 27% ± 1.1% | 27% ± 1.0% | 26% ± 0.8% | 26% ± 0.8% | 27% ± 0.8% | 27% ± 0.9% | 28% ± 0.8% | 29% ± 1.0% | 28% ± 0.9% | 29% ± 0.8% | 29% | 27% ± 1.0% | 26% ± 1.1% | 25% ± 0.9% | 26% ± 1.1% |
| Queens | 26% | 28% | 28% ± 1.4% | 30% ± 1.3% | 29% ± 1.1% | 31% ± 1.1% | 30% ± 1.0% | 28% ± 0.9% | 30% ± 0.9% | 27% ± 0.9% | 29% ± 1.0% | 29% ± 1.0% | 30% ± 0.9% | 32% ± 1.1% | 32% ± 1.1% | 33% ± 1.1% | 31% ± 1.2% | 28% | 27% ± 1.0% | 32% ± 1.0% | 31% ± 1.1% | 31% ± 1.1% |
| The Bronx | 27% | 30% | 32% ± 1.6% | 31% ± 1.6% | 31% ± 1.7% | 32% ± 1.6% | 33% ± 1.4% | 34% ± 1.5% | 32% ± 1.6% | 32% ± 1.4% | 33% ± 1.4% | 33% ± 1.4% | 35% ± 1.4% | 35% ± 1.8% | 38% ± 1.5% | 36% ± 1.7% | 37% ± 1.7% | 32% | 30% ± 2.0% | 32% ± 1.6% | 30% ± 1.7% | 32% ± 1.5% |
| Staten Island | 32% | 34% | 32% ± 2.7% | 33% ± 2.2% | 35% ± 2.5% | 34% ± 2.4% | 31% ± 2.3% | 29% ± 2.4% | 33% ± 2.5% | 32% ± 2.3% | 36% ± 2.7% | 33% ± 2.2% | 34% ± 2.4% | 36% ± 2.4% | 38% ± 2.5% | 37% ± 2.4% | 37% ± 2.2% | 27% | 29% ± 2.8% | 31% ± 2.4% | 36% ± 2.3% | 34% ± 2.5% |
Underneath the citywide line, three things moved. The subway took over a growing share of commuters until 2018 and then fell back. Working from home went from one worker in forty to one in eight, and it is now most common among Manhattan residents and the best paid, the groups with the fewest hour-plus trips, which changes who is left in the commuter count. And the mode whose hour-plus share rose most was the bus: 27% of bus riders were over the hour in 1990 and 40% are now, while the subway riders’ share edged down. That is the riders’ reported trip, not the bus’s speed, which the census does not measure. Until 2019 the Bronx was where the change landed hardest; Manhattan barely moved.
The census stops in 2024; its 2025 survey, normally published in September, has been postponed while the Bureau adapts it to a new Commerce Department rule on disclosure protection. To see where the commute stands in September 2026 we have to switch to a different kind of evidence: the MTA’s own records of how many trips are made, how fast the buses move and how late the trains run. They are administrative figures rather than a survey, they describe trips and vehicles rather than commuters, and they come with no margin of error.
In 2026, the trip is no quicker
The MTA publishes, for every day since March 2020, how many trips were made on the subway and the buses and how many vehicles crossed its bridges and tunnels, alongside its own figure for the comparable day in 2019. Through 30 September 2026, average weekday subway ridership stood at 75% of the 2019 baseline, and buses at 61%. Vehicle crossings at the bridges and tunnels were back at their 2019 level; those are vehicle counts, not commuters, so they say the traffic returned, not that the same drivers did. Weekends are a different picture, much closer to 2019; the yearly bar chart in the questions at the end has them.
Read next to the census, that is the commute of 2026. The weekday transit trip, the one we have been following, is made about a quarter less often than it was in 2019, a pattern consistent with a disproportionate loss of the shorter, remote-capable office trips, though neither the turnstile counts nor the census follow the riders themselves. Congestion pricing, which began in January 2025, coincided with a 7% rise in subway ridership, but 2026’s weekday figures are only about a point above 2025’s.
Six years back toward 2019, service by service
Administrative Counts from an agency’s own records: the calls, inspections, trips or transactions it logs. They carry no sampling margin, but they count events rather than people, and they can be cut only by what the agency records about each one.Average weekday ridership or traffic as a share of the same month in 2019, March 2020 to September 2026
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One panel per service: each month’s average weekday against the same calendar month of 2019, so a summer month is compared with a 2019 summer, and the dashed line is the 2019 level. All five panels share one scale, the faint lines are the other four services, and the panels are sorted by the latest month. Hover any panel to read the same month in all of them. The yearly figures we quoted above set 2026 to date against one The MTA’s 2019 baseline The MTA publishes each day’s ridership alongside its share of a “comparable pre-pandemic day”, the same weekday in the same week of 2019. That 2019 baseline is recovered from the MTA’s own figures (count divided by share, averaged over 2024 weekdays and weekends) and sets 2025 and 2026 counts against it. It is an administrative count of trips, not a survey of commuters, and it has no margin of error. for the whole year, so they differ a little from any single month. Bus figures for April to August 2020 are left blank: buses were boarded at the rear without fares, and the MTA’s file counts almost no riders. The yearly averages, with weekends, are in the questions at the end.
The numbers behind this chart
About these numbers. MTA daily ridership and traffic (data.ny.gov), weekdays averaged by month. Each month is set against the MTA's comparable-day baseline for the same calendar month of 2019, recovered from its 2020-2025 file as each day's count over its percentage of the comparable day, averaged over that month's 2024 weekdays, so a summer month is compared with a 2019 summer. The yearly figures we quote and the bar chart in the questions at the end use one weekday baseline for the whole year. Bridges and tunnels count vehicles. Trips, not commuters; no margins.
| Month | Subway | Bus | Long Island Rail Road | Metro-North | Bridges and tunnels |
|---|---|---|---|---|---|
| March 2020 | 51% | 54% | 47% | 58% | 63% |
| April 2020 | 8.0% | – | 2.9% | 3.4% | 32% |
| May 2020 | 10% | – | 6.3% | 4.5% | 47% |
| June 2020 | 17% | – | 14% | 7.9% | 62% |
| July 2020 | 23% | – | 21% | 12% | 72% |
| August 2020 | 26% | – | 23% | 13% | 75% |
| September 2020 | 28% | 46% | 27% | 15% | 76% |
| October 2020 | 31% | 48% | 27% | 15% | 77% |
| November 2020 | 32% | 50% | 29% | 15% | 72% |
| December 2020 | 31% | 49% | 25% | 14% | 84% |
| January 2021 | 32% | 48% | 24% | 14% | 85% |
| February 2021 | 30% | 44% | 22% | 13% | 77% |
| March 2021 | 33% | 48% | 26% | 15% | 88% |
| April 2021 | 36% | 53% | 28% | 17% | 91% |
| May 2021 | 39% | 54% | 33% | 20% | 93% |
| June 2021 | 44% | 58% | 36% | 33% | 96% |
| July 2021 | 46% | 59% | 39% | 36% | 97% |
| August 2021 | 47% | 60% | 39% | 35% | 96% |
| September 2021 | 48% | 60% | 43% | 39% | 95% |
| October 2021 | 54% | 65% | 48% | 43% | 100% |
| November 2021 | 59% | 67% | 52% | 49% | 99% |
| December 2021 | 56% | 65% | 46% | 43% | 99% |
| January 2022 | 50% | 58% | 40% | 35% | 91% |
| February 2022 | 54% | 61% | 46% | 42% | 99% |
| March 2022 | 57% | 64% | 51% | 50% | 98% |
| April 2022 | 57% | 63% | 53% | 50% | 99% |
| May 2022 | 58% | 64% | 54% | 52% | 100% |
| June 2022 | 60% | 66% | 57% | 56% | 99% |
| July 2022 | 57% | 63% | 55% | 53% | 99% |
| August 2022 | 58% | 64% | 55% | 54% | 98% |
| September 2022 | 60% | 62% | 60% | 59% | 100% |
| October 2022 | 62% | 64% | 61% | 59% | 100% |
| November 2022 | 65% | 68% | 63% | 65% | 99% |
| December 2022 | 66% | 67% | 61% | 62% | 103% |
| January 2023 | 67% | 68% | 61% | 62% | 100% |
| February 2023 | 65% | 66% | 62% | 62% | 101% |
| March 2023 | 66% | 65% | 63% | 63% | 100% |
| April 2023 | 66% | 64% | 65% | 63% | 100% |
| May 2023 | 68% | 65% | 67% | 68% | 102% |
| June 2023 | 66% | 65% | 69% | 68% | 101% |
| July 2023 | 65% | 63% | 66% | 66% | 101% |
| August 2023 | 65% | 63% | 68% | 67% | 99% |
| September 2023 | 65% | 60% | 70% | 69% | 100% |
| October 2023 | 67% | 61% | 74% | 71% | 103% |
| November 2023 | 71% | 62% | 77% | 78% | 102% |
| December 2023 | 70% | 61% | 74% | 74% | 105% |
| January 2024 | 69% | 60% | 72% | 71% | 100% |
| February 2024 | 67% | 59% | 73% | 72% | 101% |
| March 2024 | 68% | 58% | 73% | 72% | 101% |
| April 2024 | 67% | 57% | 75% | 72% | 100% |
| May 2024 | 69% | 58% | 77% | 76% | 102% |
| June 2024 | 68% | 60% | 78% | 77% | 102% |
| July 2024 | 66% | 57% | 76% | 74% | 100% |
| August 2024 | 65% | 57% | 77% | 72% | 100% |
| September 2024 | 70% | 63% | 81% | 78% | 102% |
| October 2024 | 71% | 65% | 83% | 79% | 103% |
| November 2024 | 74% | 67% | 83% | 82% | 101% |
| December 2024 | 73% | 65% | 82% | 80% | 106% |
| January 2025 | 73% | 66% | 79% | 76% | 100% |
| February 2025 | 73% | 67% | 80% | 76% | 101% |
| March 2025 | 72% | 66% | 81% | 79% | 101% |
| April 2025 | 74% | 67% | 85% | 79% | 102% |
| May 2025 | 73% | 64% | 84% | 80% | 103% |
| June 2025 | 74% | 67% | 83% | 81% | 102% |
| July 2025 | 74% | 67% | 86% | 81% | 102% |
| August 2025 | 73% | 67% | 86% | 78% | 102% |
| September 2025 | 73% | 63% | 88% | 83% | 102% |
| October 2025 | 73% | 61% | 87% | 83% | 102% |
| November 2025 | 78% | 63% | 89% | 87% | 102% |
| December 2025 | 77% | 62% | 88% | 84% | 105% |
| January 2026 | 76% | 61% | 80% | 75% | 97% |
| February 2026 | 73% | 56% | 79% | 75% | 93% |
| March 2026 | 76% | 60% | 86% | 83% | 100% |
| April 2026 | 76% | 61% | 88% | 83% | 100% |
| May 2026 | 76% | 61% | 80% | 85% | 102% |
| June 2026 | 76% | 63% | 92% | 88% | 102% |
| July 2026 | 75% | 62% | 90% | 85% | 102% |
| August 2026 | 73% | 62% | 89% | 83% | 100% |
| September 2026 | 74% | 61% | 93% | 87% | 101% |
How many trips are made is half the question. The other half is how long the ride takes, and the MTA measures that too, for the vehicle rather than for any rider. From January to August 2026 the average weekday local, limited or Select bus ran at 7.7 miles an hour, within a tenth of a mile an hour of the same months in 2019 and in 2024. Manhattan’s buses, the slowest in the city, averaged about six.
Administrative Counts from an agency’s own records: the calls, inspections, trips or transactions it logs. They carry no sampling margin, but they count events rather than people, and they can be cut only by what the agency records about each one.
How to read this, and what it cannot show
One bar per year, January to August of each year so that 2026 compares like with like: total mileage over total operating time on weekday local, limited and Select Bus Service routes, 4 a.m. to 11 p.m. Routes are assigned to a borough by their letters, and Queens routes were redesigned in the summer of 2025. The MTA withdrew its figures before 2019 while it recalculates them.
The trains tell the same story. About 15% of weekday subway trips ran more than five minutes behind schedule in 2026, as in 2019 and 2024; in 2017, the year the hour-plus share peaked, it was a fifth. So the network the hour-plus commute runs on in 2026 is much the network of 2024, when the census last asked: buses no faster, trains no more punctual, and fewer trips on both than in 2019. Nothing in these figures says the hour has got any shorter since the survey measured it.
Administrative Counts from an agency’s own records: the calls, inspections, trips or transactions it logs. They carry no sampling margin, but they count events rather than people, and they can be cut only by what the agency records about each one.
How to read this, and what it cannot show
One bar per year, January to August of each year: the MTA’s estimate of trips more than five minutes behind the timetable, waiting and riding together, for each train a rider takes, so a trip with a transfer counts twice. Weekdays 6 a.m. to 11 p.m. The figures before 2017 cover the numbered lines only and are not drawn.
So who is spending the hour?
The average commute is a number about a city. The hour-plus commute is a number about a place and a vehicle: the outer-borough resident on a subway line that runs to the end of the borough, or on a bus that averages under eight miles an hour, heading toward Manhattan or across to another borough.
That is why pay barely distinguishes the people spending the hour: what the top earners have is not a faster train but an address and a workplace that put them on a shorter one. The racial gap is the harder case: standardising on where people live, how they travel and where they work shrinks it by about half, and the characteristics in the model leave the remainder unexplained. And the map has barely changed. Across thirty-four years of investment and one pandemic the share over the hour stayed within a few points of a quarter, and in 2026 the buses run no faster and the trains no more punctually than they did in 2019.
None of that makes the long commute a story about failure; a 45-minute train ride to a Manhattan job is a bargain most metros cannot offer. But when the city talks about the commute, it usually means the average commuter’s, and the hour-plus commuter is a different person, in a different borough, on a different vehicle. The data have been saying so since 1990. If that hour is going to get shorter, it will have to get shorter out there.
Cut the data your own way
The dashboard runs on the same 125,000 census records as everything above. Pick a borough, a mode, an earnings band and a race or ethnicity, then drag the slider to decide for yourself what counts as a long trip.
Share over the threshold by mode, under the filters above
Loading the census records…
Every number here is a direct weighted tabulation of census microdata records, not a model prediction, so the cross-tabulations it shows are measured relationships within one survey. Their uncertainty is approximated: the headline tile carries an approximate 90% margin: a binomial margin on the record count, scaled by a design effect calibrated on the margins of the charts we walked through earlier (click the ? on the tile). Those charts carry the exact margins; treat any selection with a wide margin as indicative.
Borough
Where the commuter lives, from the census's public-use microdata areas, which nest inside the five boroughs. We never go below the borough.
Mode
The census asks for the one means of transportation usually used for most of the distance. A bus-to-subway trip counts once, as the subway if that was the longer leg. Car includes trucks, vans and motorcycles; taxi, ferry, light rail and "other" are grouped.
Own earnings
The worker's own wages plus self-employment income over the past twelve months, in 2024 dollars. Not household income: a $25,000 earner in a $150,000 household sits in the lowest band.
Race or ethnicity
Hispanic of any race; the other four groups are non-Hispanic. "Other" includes people reporting two or more races, American Indian and Pacific Islander.
The threshold
The headline figure uses 60 minutes, which is where the Census Bureau's own table starts its top bands. Drag to see how much the "long commute" depends on where the line is drawn. Trips are reported in minutes and heap at 30, 45, 60 and 90, so the share can jump at those marks.
Travel longer than the threshold
The share of the selected commuters whose usual one-way trip is at least the threshold, read off a histogram in five-minute steps. A weighted tabulation of the census records in the selection, with an approximate 90% margin beside it (see the ? on the tile).
Median one-way minutes
The middle commuter's usual one-way travel time, door to door, as reported. Read off the five-minute histogram, so it lands on a multiple of five.
Commuters, weighted estimate
The sum of the survey weights of the records in the selection: the Bureau's estimate of how many New York City residents these records stand for. Rounded to the nearest thousand. Each record stands for about 25 workers, so treat a small selection as an order of magnitude.
Work in Manhattan
The share whose place of work is in Manhattan, from the census's place-of-work area. For Manhattan residents this is the share who work in their own borough.
Leave home before 6 a.m.
The share whose usual departure time for work is before 6:00 a.m., from the census's time-of-departure question.
Foreign-born
The share born outside the United States, whatever their citizenship now.
By mode
Each bar is the share over the threshold for one mode, under the borough, earnings and race filters. Choosing a mode highlights its bar rather than hiding the others, so the comparison is always visible.
The ± figure
An approximate 90% margin of error on the headline share. The dashboard cannot carry the 80 replicate weights behind the charts we saw earlier, so it estimates the margin as 1.645 × √(deff × p(1 − p) ÷ n): n is the number of census records in the selection and deff is the design effect, the ratio of the survey's true sampling variance to that of a simple random sample of the same size.
The design effect is calibrated from those charts: for every chart cell, (replicate-weight margin ÷ binomial margin)², and the median is used. The note under the chart gives the effective sample size, n ÷ deff. For any cut the charts precompute, the chart's margin is the exact one.
Too few commuters
Fewer than 30 census records match, so the numbers would be noise. Widen a filter.
Questions this raises
Is New York’s commute really the longest in the country?
For a city, yes: among places the Census Bureau reports, New York City’s 2024 mean of 40.6 minutes is the highest of any large city, and the surrounding metro has led the Bureau’s ranking of large metro areas for years. Claims that drivers in particular have the longest commute in the nation, which circulate in the congestion-pricing debate, rest on older private data and rank the city second for driving. The census figure is for all modes, and New York’s is long because of transit, not despite it: drivers here have a median trip of 30 minutes, which is ordinary; subway and bus riders have a median of 45, which is not.
How well do these numbers match the Bureau’s own?
Closely, and that is the check everything else rests on. The Bureau’s 2024 table reports 3,560,755 ± 22,605 commuters and 896,483 of them at 60 minutes or more, a share of 25.2%, with a mean trip of 40.6 minutes and 12.5% of workers at home. Rebuilding the same universe from the 2024 microdata gives 3,556,031 ± 26,625 commuters, 25.3% ± 0.5 over the hour, a mean of 40.7 minutes and 12.5% at home, all inside the Margins of error from replicate weights Each PUMS record carries 80 replicate weights alongside its main weight. A statistic is recomputed 80 times, once with each replicate, and the spread of those 80 answers around the main answer gives its standard error: variance = 4/80 × Σ(θr − θ)². The Census Bureau reports 90% margins, 1.645 standard errors, and every census margin we show follows that convention.This is the Bureau’s own method (successive difference replication), so a margin here matches what the Bureau would publish for the same cut. . The five-year file, which most of the charts use, gives 3,245,001 commuters against a published 3,263,450, a difference of 0.6% that sits just outside the two margins; the microdata is a subsample of the full survey with its own weights, and a gap of that size on a count is usual. Its 25.1% ± 0.3 over the hour and 40.3-minute mean match the published 25.3% and 40.3. The full comparison, borough by borough, is in the companion files.
How long are the trips, from under fifteen minutes to over ninety?
About a third of commuters are under half an hour, the largest single band is 30 to 44 minutes, and a quarter take an hour or more, including about one commuter in fifteen at ninety minutes or more. The answers heap on round numbers, which is one reason the hour is drawn at the Bureau’s own band edge. Filter by borough or mode to watch the shape move.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
Each bar is a band of one-way Travel time to work The census asks how many minutes it usually took to get from home to work last week, door to door, one way. It is self-reported and rounded: answers pile up at 30, 45, 60 and 90 minutes. People who worked from home have no travel time and are not commuters.The Census Bureau publishes it in bands; the top two, 60 to 89 minutes and 90 or more, are what we call the hour-plus commute. ; the black tick is the 90% margin of error. The filters apply to every band.
Who is the hour-plus commuter, in one profile?
Compared with someone whose trip is under half an hour, they are far more likely to live outside Manhattan (92% against 72%), to ride the subway (53% against 17%) or the bus (18% against 6%), to be Black or Hispanic (56% against 41%), to be foreign-born (51% against 43%), and to have no degree (62% against 52%). They earn a little less, a median of about $49,000 against $53,000. They are nearly twice as likely to leave home before 6 a.m. And at five commuting days a week, an hour each way amounts to ten hours in transit, against about six for the median commuter; the census records the usual one-way trip, not days worked, so that is a scenario rather than a measurement.
Which jobs have the longest commutes?
Construction workers are over the hour at 37%, cleaners at 35% and health aides at 28%; lawyers at 16%, teachers at 19% and managers at 21%. The jobs with the longest commutes are, by and large, the ones that cannot be done from home, and mostly the lower-paid ones. Once borough, mode and workplace are held fixed in the adjusted model, most occupations land within about three points of office work; construction keeps a residual of about 6 points and legal work one of about −5.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per census occupation group. Switch the measure to median earnings: the long-commute groups are mostly the low-paid ones.
How flat is the hour-plus share across pay, band by band?
Citywide it is 26% under $30,000, 28% from $30,000 to $60,000, 27% from $60,000 to $100,000, 23% from $100,000 to $150,000 and 15% above. Inside a borough the last step mostly goes too: in Queens the share is 29 to 32% in every earnings band, and in the Bronx 29 to 33%. Seven in ten commuters earning $150,000 or more work in Manhattan, and their subway rides are short. In the adjusted model the top band’s raw 12-point advantage shrinks to 6, the $100,000 to $150,000 band keeps about 2 points, and the lower bands sit within a point of the reference.
Does the racial gap close inside a borough, or on the same mode?
It narrows inside a borough and, except on Staten Island, does not close. Citywide, Black commuters are over the hour at 31%, Hispanic commuters at 27%, Asian commuters at 26% and white commuters at 18%. In Brooklyn the Black and white shares are 32% and 19%; in Queens, 36% and 24%; in the Bronx, 32% and 28%. Among people riding the same subway the gap is wider, 45% for Black riders against 20% for white ones; among bus riders it reverses, and among drivers the two are about the same. Black commuters take the bus at 17% against 7% for white commuters.
How does each mode compare, and does the census see a two-bus trip?
Subway riders are over the hour at 34% and bus riders at 38%; drivers at 18%; cyclists at 6%; walkers at 1%. By borough, subway riders are over the hour at 11% in Manhattan, 35% in Brooklyn, 45% in Queens, 50% in the Bronx and 79% in Staten Island, where the subway is a ferry ride away to begin with; bus riders run 33 to 64%, and drivers 15 to 20% in every borough. The median Manhattan trip is 30 minutes and the median trip from the other boroughs is 40. Commuter rail, a small group of about 33,000 city residents, is over the hour for six in ten. The survey asks for the one means of transportation used for most of the distance, so a bus-to-subway trip is recorded once, as the subway if that was the longer leg; the two-fare trip is invisible as such, and “bus” means people whose main leg is a bus. Travel time is what the respondent writes down for the usual door-to-door trip, and it is heaped: far more people say 30, 45, 60 or 90 than 28, 47, 58 or 95. The 60-minute line is the Bureau’s own band edge, which is why it is used here; the dashboard slider lets you move it. Because answers heap on the hour, the patterns were checked at 90 minutes too, and they hold: 13% of Staten Island commuters, 8.5% of Bronx, 8.2% of Queens and 6.4% of Brooklyn commuters report ninety minutes or more against 2.5% of Manhattan’s; Black commuters at 9.4% against 4.5% for white commuters; bus riders at 12% and subway riders at 8.6% against 4.6% of drivers. At 45 minutes, 41% of all commuters are over the line.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per usual mode, with the borough and earnings filters. Switch the measure to see how many people use each: the subway carries four in ten commuters, the car three in ten.
Did the bus get slower?
The census cannot say. What it measures is the share of bus riders reporting an hour or more, which rose from 27% in 1990 to 40% in 2024, and a rider’s trip can lengthen for reasons other than the vehicle’s speed: longer routes, more transfers, jobs that moved. The MTA’s own operating data measure the vehicle. On the method it adopted in 2025, which it has applied from 2019 on, the average weekday local, limited or Select bus ran at 7.6 miles an hour in both 2019 and 2024 and 7.7 in 2025 and in 2026 to August, with Manhattan’s at about six and Staten Island’s at about twelve; it withdrew the earlier years while it recalculates them. The Comptroller’s figure of 8.17 miles an hour for 2024, down 0.6% over ten years, counts express buses and was computed from the MTA’s figures before that revision. Either way buses are very slow today, have not got faster since the pandemic, and the speed series does not reach back to 1990.
What is the adjusted model, and what does it hold fixed?
A survey-weighted logistic regression of whether the one-way trip is 60 minutes or more, on borough of residence, usual mode, place of work, own-earnings band, race and ethnicity, occupation group, age band, sex, nativity and degree, fitted to the 125,000 commuter records in the five-year file with their person weights. What the chart shows is the average marginal effect of each category: the average change in the predicted chance of an hour-plus trip when every commuter is switched from the reference category to that one, everything else as observed. Reference groups are Manhattan, car, own borough, $60,000 to $100,000, white, office and administrative work, age 35 to 44, male, US-born and no degree. Margins come from refitting the model with each of the 80 replicate weights. The model explains a modest share of the variation (a pseudo-R² of 0.16), which is expected when the outcome depends on things the census does not record. With the model’s other variables fixed, riding the subway rather than driving adds about 19 points to the hour-plus share, the bus about 21, and an outer-borough address adds 11 to 23. It is an adjustment on observed variables, not a causal estimate: it says how much of a raw gap survives the controls, not why. Construction keeps a residual of about 6 points and legal work one of about −5; most other occupations land within about three points of office work.
Why pool five years, and what about the pandemic?
The five-year file pools the 2020 to 2024 surveys, which is what makes cuts like Staten Island subway riders or Bronx commuters earning over $150,000 possible at all. That average contains the pandemic: 17% of its workers were at home against 12.5% in 2024 alone, which is why the working-from-home figure quoted above is the 2024 one. On the hour-plus share the two agree, 25.1% pooled against 25.3% ± 0.5 in 2024, and the borough, mode, earnings and race patterns recomputed on 2024 alone are much the same: Black commuters 31%, white 18%; subway 33%, bus 40%, car 19%. The one-year figures are in the results file.
Can 1990 and 2024 really be compared?
The questions are the same words. What differs is the survey: 1990 and 2000 are the decennial census long form, a 5% sample of housing units, and 2005 onward is the annual American Community Survey, about 1% a year with different weighting. The 1990 and 2000 points carry no margin because those files have no replicate weights; a 5% sample of the city’s workers is about 120,000 to 130,000 records and the sampling error on a citywide share is well under a point. The five boroughs are identified in every year (the microdata areas were redrawn in 2000, 2010 and 2020, always inside borough lines), so the city and borough totals line up; means-of-transportation codes were mapped to the current scheme. The 2020 survey used experimental weights after the pandemic wrecked its response rates, and the chart says so and draws it without a margin, because its replicate weights understate the uncertainty. One thing did change: before 2005 the survey asked about the week before Census Day in April, and since then it has asked every month of the year, which slightly favours the modern series with summer trips.
Microdata estimate Tabulated directly from survey microdata, with no model in between: every variable in the chart is observed in the same record, so the relationship shown is measured rather than modelled. It is still an estimate from a sample, not a count of everyone, and the ticks are 90% margins from the survey’s replicate weights.
How to read this, and what it cannot show
One bar per year: the 1990 and 2000 censuses, then the annual survey with its margins. Filter by borough and mode.
What happened in 2017?
The hour-plus share peaked at 28% and the mean trip at 42 minutes in 2017 and 2018, and the subway’s own hour-plus share peaked at 38% in 2017, leaving aside the experimental 2020 file. That is the period of the subway’s reliability crisis, the “summer of hell” and the governor’s state of emergency in June 2017. The MTA’s own delay measure, which covers every line from 2017, puts it in numbers: a fifth of weekday subway trips ran more than five minutes behind schedule from January to August 2017, with 2.7 minutes above schedule per train trip, against 15% and 1.9 minutes in the same months of 2019. The census cannot attribute a cause, and part of the rise is composition, since the subway was carrying close to its largest share of commuters. But the timing is what it is, and the Bronx, the borough most dependent on the longest lines, is where the share rose most, from 27% in 1990 to 37% in 2019; it is 32% now.
For the rest of the series: the share was 25% in 2000 and stayed near that through the early 2010s; the subway’s share of commuters rose from 37% in 1990 to 47% in 2018 before falling back to 43%; and the average trip, 37 minutes in 1990, is four minutes longer now.
Did working from home shorten everyone else’s commute?
Not on average. Working from home rose from 4.7% of workers in 2019 to 24% in 2021 and 12.5% in 2024, and it is now most common among Manhattan residents and the best paid, the groups with the shortest trips. That change in who commutes should, on its own, have raised the hour-plus share among commuters, and instead it fell from 27% to 25%. Something offset it that the census cannot identify. Less crowded trains are one candidate; a more punctual subway is not, since the MTA’s delay measure was about the same in 2019 and 2024. The census follows no one from one year to the next, so it cannot separate a change in who commutes from a change in how long the same trips take, and no attempt is made to here.
What does the MTA data add, and what can it not say?
The census asks people; the MTA counts turnstiles, fareboxes and tolls, and times its trains and buses. The ridership series runs to 30 September 2026, nearly two years past the last census year, and with the MTA’s speed and delay measures it shows the transit commute since congestion pricing began in January 2025. Its weekday subway figure rose from 69% of 2019 in 2024 to 74% in 2025 and 75% in 2026 to date, about 4.0 million trips a day against 5.4 million; weekend subway ridership is at 88% of 2019, and the Long Island Rail Road’s weekends are above it; the governor’s office reports nearly 1.3 billion subway trips in 2025, up 7% on 2024, and the citizens’ advisory committee puts that at 76.5% of 2019’s 1.7 billion. The speed and delay measures describe the vehicle against its timetable: how fast a bus covered its route, how far behind schedule a train ran, averaged over the riders on it. None of the three can say how long any one person’s door-to-door trip took, who was riding, or where they were going. So the 2026 section is about volumes and the network: how much of the traffic has returned, and whether the network got any quicker.
Here are the same counts as averages by year, weekdays and weekends, for each service:
Administrative Counts from an agency’s own records: the calls, inspections, trips or transactions it logs. They carry no sampling margin, but they count events rather than people, and they can be cut only by what the agency records about each one.
How to read this, and what it cannot show
One bar per year, through 30 September for 2026, as a share of the MTA’s The MTA’s 2019 baseline The MTA publishes each day’s ridership alongside its share of a “comparable pre-pandemic day”, the same weekday in the same week of 2019. That 2019 baseline is recovered from the MTA’s own figures (count divided by share, averaged over 2024 weekdays and weekends) and sets 2025 and 2026 counts against it. It is an administrative count of trips, not a survey of commuters, and it has no margin of error. . Switch between weekdays and weekends and between modes.
Where this data goes quiet
Anything below the borough. The second leg of a two-seat trip. Why a trip takes an hour, whether that is distance, a slow bus, a transfer or a walk at either end. Reverse commuters into the city, who live in the suburbs and are not in a city survey at all. Whether the same people were over the hour in 2019 and 2024, which would need the same workers followed through time. The census’s own view of 2025 and 2026, which the Bureau has not yet released. And the microdata is a disclosure-protected sample of the survey, with a few top-coded values and some perturbation, so every relationship seen in it is an estimate of the city’s, not the city itself.
Under the hood: data, definitions, estimators and the full results
This is the methods note for everything above. It is deliberately more technical than the rest.
The published anchors. Table B08303 (workers 16 and over who did not work at home, by travel time), B08013 (aggregate travel time, whose ratio to B08303’s total is the mean) and B08301 (workers by means of transportation), for the city and the United States from the 2024 one-year survey and for the city and its five counties from the 2020–24 five-year survey. These are both the anchors and the reproduction targets.
Census microdata. The 2020–24 five-year public-use microdata sample for a 156-area study region, cut to residents of the five boroughs who were at work in the reference week and therefore have a means of transportation: 151,386 person records carrying 80 replicate weights and representing 3.92 million workers. Commuters are those who did not work from home and reported a one-way travel time — 125,312 records representing 3,245,001 workers, which is the universe of the published table. The usual means of transportation is a single mode, grouped as subway, bus, car (including trucks, vans and motorcycles), commuter rail, walked, bicycle and other. Place of work is coded at borough level in the 2020 vintage (Manhattan 04100, Bronx 04200, Brooklyn 04300, Queens 04400, Staten Island 04500) and grouped as Manhattan, own borough, another borough, or outside the city. Own earnings are in 2024 dollars in five fixed bands; race and ethnicity is Hispanic of any race, otherwise the census’s single-race groups; occupation is collapsed to the Bureau’s major groups; departure time before 6 a.m. is code 30 and below. The 2024 one-year file, with its own replicate weights, carries the reproduction and a sensitivity check.
The 1990–2024 series. The 1990 census 5% sample, the 2000 census 5% sample and the 2005–2024 one-year files. The five boroughs are identified in every one: by the 1990 area prefixes 050 to 054, verified against the 1990 census county populations to within 1%; by the 55 area codes 03701 to 04114 shared by the 2000 and 2010 vintages, verified against the 2000 equivalency file; and by the 2020 geography from 2022. Each year’s means-of-transportation codes are mapped to the 2019 scheme and the extraction script records every mapping. The 1990 and 2000 files have no replicate weights; the 2020 one-year file carries the Bureau’s experimental weights and is flagged. The two pre-2005 surveys refer to the week before Census Day in April, the later ones to a rolling reference week through the year.
Transit. MTA daily subway, bus, Long Island Rail Road, Metro-North and bridge-and-tunnel counts from 1 March 2020 to 30 September 2026, averaged by year and day type. The 2019 baseline for each mode and day type is recovered from the MTA’s frozen 2020–2025 file, which carries the MTA’s own share of the comparable pre-pandemic day for every date: the count divided by that share, averaged over 2024 weekdays and weekends. The series counts trips, not commuters, and has no sampling margin.
Speed and punctuality. Bus speeds are the MTA’s monthly route figures from Bus Time location data: for weekday local, limited and Select Bus Service routes, total mileage over total operating time between 4 a.m. and 11 p.m., citywide and by the borough the MTA assigns each route to, with the peak at 7–9 a.m. and 4–7 p.m.; express routes are kept apart. The MTA rebuilt the figures from 2019 on with a new method in August 2025 and removed the earlier years in September 2026 while it recalculates them, so the series starts in 2019. Subway punctuality is the MTA’s customer journey metrics by line: for each weekday train trip between 6 a.m. and 11 p.m., the minutes a rider spent waiting and riding beyond the timetable, estimated from fare-card entries, a model of the routes riders take and the trains’ recorded movements; late means more than five minutes over. Line figures are combined weighted by passengers; the 2015 and 2016 figures cover the numbered lines only, so the series starts in 2017. Both are compared over January to August of each year, the months 2026 has. They measure the vehicle against its schedule, so a slower timetable is not a delay, and a trip with a transfer counts twice in the subway figures.
Estimators. For a statistic computed with the full person weight, the same statistic is recomputed with each of the 80 replicate weights and the variance is the Bureau’s successive-difference estimator, four-eightieths of the summed squared deviations of the 80 replicate estimates from the full-weight one. Margins are 90%, or 1.645 standard errors. Counts are weighted sums, shares and means weighted means, medians weighted medians without a margin. Every combination the chart controls can reach is precomputed with the same estimator and a margin on each bar; cells resting on fewer than 30 records are labelled too thin rather than drawn. The dashboard sums pre-aggregated cells and reads medians and threshold shares off a five-minute histogram; its one margin, on the headline share, is approximate, a binomial margin scaled by a design effect calibrated on the charts’ replicate margins.
Does it reproduce? Every share and mean lands inside the margin, and the one-year commuter count lands inside the published margin. The five-year count differs by 18,449, or 0.6%, slightly outside the two margins: the microdata sample is a subsample of the full survey with its own weights, a difference of this size on a five-year count is usual, and it moves no share. The comparison is the reproduction table.
Geography and mode. The five-borough gradient, 9.8% in Manhattan to 31.5% in the Bronx, is almost entirely a transit gradient. Drivers are over the hour at 14.9 to 20.2% in every borough; subway riders at 10.9% in Manhattan, 34.5% in Brooklyn, 45.5% in Queens, 50.0% in the Bronx and 79.3% in Staten Island, the last on 330 records. By place of work, 55.0% of commuters work in their own borough and 13.9% of those are over the hour; 27.1% travel to Manhattan from another borough (40.4%); 10.3% work in another outer borough (39.6%); 7.6% work outside the city (32.6%). Bronx residents working in another borough are over the hour at 55.7%, Staten Island residents working in Manhattan at 70.4%.
Earnings. The citywide share is flat at 26.0 to 27.7% below $100,000, 22.8% between $100,000 and $150,000, and 15.4% above. Within Queens and the Bronx there is no gradient at all; within Staten Island it runs upward, 26.2% to 43.6%, because the better paid there travel to Manhattan; among subway riders it runs downward, 39.2% to 13.7%, because the highest earners board in or near Manhattan — 69.4% of commuters earning $150,000 or more work there. Where there is an earnings gradient, it follows where people live and work rather than how fast they travel.
Race and ethnicity. Black commuters are over the hour at 31.4% ± 0.7%, Hispanic commuters 27.4% ± 0.5%, Asian commuters 26.4% ± 0.8% and white commuters 17.9% ± 0.4%. Holding the borough fixed narrows the Black–white gap without closing it (Brooklyn 32.3% against 19.3%, Queens 36.1% against 23.9%, the Bronx 31.6% against 27.9%); holding the mode fixed at the subway widens it, 45.1% against 20.2%. Black commuters take the bus at 17.0% against 6.9%, and 89.9% live outside Manhattan against 68.7%. This is a descriptive decomposition one variable at a time, not a model, and it does not rank the contributions of residence, mode and occupation.
The profile. Against commuters under 30 minutes, those over the hour are more often outer-borough residents (92.4% against 71.6%), subway riders (53.4% against 17.0%), bus riders (17.5% against 6.0%), Manhattan-bound (47.5% against 34.8%), Black or Hispanic (55.8% against 41.1%), foreign-born (51.2% against 43.1%), without a bachelor’s degree (62.4% against 51.7%) and out of the door before 6 a.m. (13.7% against 7.4%). Every one of those differences exceeds the sum of the two margins. Median own earnings are $49,240 against $52,793. Among occupation groups with at least 500 records the share runs from 37.4% in construction and 34.7% in cleaning and maintenance down to 15.7% in legal work and 19.2% in education and the arts.
The adjusted model. A weighted logistic regression of the probability of a 60-minute-or-longer trip on borough of residence, usual mode, place of work, earnings band, race and ethnicity, occupation group, age band, sex, nativity and degree: 50 parameters on 125,312 commuter records with person weights, fitted by iteratively reweighted least squares, pseudo-R² 0.16. What is reported is average marginal effects — for each category, the weighted mean across all commuters of the change in predicted probability when every commuter is set to that category rather than the reference (Manhattan, car, own borough, $60–100k, white, office and administrative, 35–44, male, US-born, no degree). Margins come from refitting with each of the 80 replicate weights, warm-started from the full-weight solution. Mode and geography survive: the subway is +19.0 ± 0.7 points and the bus +21.1 ± 1.1 against the car, where the raw gaps were +15.7 and +19.4, and the outer boroughs are +11.1 to +22.5 against Manhattan. Earnings partly do: the top band’s raw −11.6 points becomes −5.8 ± 0.9, the $100,000 to $150,000 band’s −4.2 becomes −2.1 ± 0.8, and the lower bands sit within a point of the reference. Occupation effects shrink to within about three points of office work except construction, +6.0, legal work, −4.8, and the arts, −3.6; a degree’s raw −7.9 becomes −2.2. The Black–white gap of +13.6 points becomes +6.0 ± 0.6, so about half is accounted for by residence, mode, workplace, earnings and occupation and about half is not. The residual is not attributed to any mechanism here, because the microdata record no station, transfer or service frequency. It is a descriptive adjustment on observed variables, not a causal estimate, and it cannot hold fixed what the census does not record. The code is src/articles/hour_plus_commute/model.py.
The series. The hour-plus share was 22.8% in 1990 and 24.5% in 2000, held near 25% through 2014, rose to 28.4% ± 0.5% in 2017 and 27.1% in 2019, and stands at 25.3% ± 0.5% in 2024; the mean trip went from 36.5 to 40.7 minutes. The subway’s share of commuters rose from 37.5% to 46.7% in 2018 and fell back to 42.6%; the car’s fell from 33.6% to 29.0%. The subway’s own hour-plus share was 35.5% in 1990, 37.5% in 2017 and 32.5% in 2024. The bus’s rose from 27.4% to 39.5% — that is the share of riders reporting an hour or more, not a measure of vehicle speed; the Comptroller’s operating data put buses at 8.17 mph in 2024, down 0.6% over ten years. The car’s rose from 12.6% to 19.0%. Working from home went from 2.5% of workers to 12.5%, and 24.0% in 2021. The 2017 peak is not attributed to a cause; it coincides with the subway’s reliability crisis of that year and with the subway’s near-peak share of commuters.
2026 in the transit counts. Weekday subway ridership rose from 68.7% of the 2019 baseline in 2024 to 73.8% in 2025 and 74.8% in 2026 to date; buses 60.3%, 64.8%, 60.8%; bridge-and-tunnel crossings 101%, 102%, 100%; weekend subway ridership 80.3%, 87.7%, 88.7%. The Governor’s office reports nearly 1.3 billion subway trips in 2025, up 7%; the citizens’ advisory committee puts 2025 at 76.5% of 2019’s 1.698 billion and attributes much of the growth to congestion pricing, which began on 5 January 2025. The census cannot see 2025 or 2026, and these counts say how many trips were made, not how long they took or who made them.
2026 in the speed and punctuality figures. Weekday local, limited and Select buses averaged 7.64 miles an hour from January to August 2019, 8.17 in 2020, 7.64 in 2024, 7.70 in 2025 and 7.73 in 2026; at rush hours 7.17, 7.23, 7.29 and 7.32 in 2019, 2024, 2025 and 2026. By borough in 2019 and 2026: Manhattan 5.93 and 6.15, Brooklyn 7.02 and 6.92, the Bronx 7.00 and 7.04, Queens 8.58 and 8.65, Staten Island 12.16 and 12.22. Weekday subway trips more than five minutes late: 20.0% in 2017, 19.1% in 2018, 15.4% in 2019, 15.6% in 2024, 14.6% in 2025 and 15.2% in 2026, with 2.72, 2.44, 1.92, 1.87, 1.75 and 1.86 minutes above schedule per train trip. Nothing in either series moved by more than a few percent between 2024 and 2026.
Rerunning it. The analysis script computes every figure and number above and the bar charts, from a fixed seed; its docstring is the long-form method. The template charts and the dashboard come from two further steps of the same code, and the build manifest lists every module of it, and of the shared library of readers and estimators it imports, with its checksum. The companion dataset is the census records themselves: the 151,000 city resident workers in the five-year file, with their survey weights and the derived variables, so a weighted mean of its hour_plus column among commuters reproduces the headline share.
Continue exploring New York
Per subway ride, major felonies fell for thirty years. Felony assault broke from the fall.
In January to May 2026, felony assault per subway entry was about twice its 2019 rate, counting unpaid entries, while the six major felonies together sit close to 2019.
Who still goes to the office
Working from home has settled rather than faded, and it has settled on the better paid.
- Explore this dataset: the data and the code behind every number above.
- All articles · Search · RSS
SparkyData Weekly
Once a week, one New York theme: what the public data says about it, and the week's new articles.
Free, weekly, and one click to leave. Your address is held by Buttondown, which sends it; see the privacy policy.
Sources
- U.S. Census Bureau — Travel Time to Work (B08303), Aggregate Travel Time to Work (B08013) and Means of Transportation to Work (B08301), New York city and United States (ACS 2024 1-year) — 3,560,755 ± 22,605 NYC commuters, 896,483 of them at 60 minutes or more; mean travel time 40.6 minutes in the city and 27.2 nationally; 4,070,607 workers, 12.5% working from home; the published numbers everything here starts from
- U.S. Census Bureau — the same tables for New York city and its five counties (ACS 2020–2024 5-year) — 3,263,450 ± 12,564 commuters; 25.3% at 60 minutes or more; mean 40.3 minutes; borough shares from 10.0% in Manhattan to 32.1% in Staten Island
- U.S. Census Bureau — ACS 2020–2024 5-Year PUMS and 2005–2024 1-Year PUMS (New York), person files with replicate weights — every count, share and margin labelled census; the one-year files give the 2005–2024 points of the series
- U.S. Census Bureau — Census 2000 5-Percent Public Use Microdata Sample, New York — the 2000 point of the series: travel time, means of transportation and place of work for a 5% sample of the city
- U.S. Census Bureau — 1990 Census Public Use Microdata Sample A (5 percent), New York — the 1990 point of the series, same questions
- Metropolitan Transportation Authority — MTA Daily Ridership and Traffic: Beginning 2020 (data.ny.gov) (retrieved 4 October 2026, counts to 30 September 2026) — daily subway, bus, railroad and bridge-and-tunnel counts through 30 September 2026; the 2025 and 2026 figures
- Metropolitan Transportation Authority — MTA Daily Ridership Data: 2020–2025 (data.ny.gov) — the MTA's share-of-comparable-pre-pandemic-day figure for each date, from which the 2019 baselines are recovered
- Metropolitan Transportation Authority — MTA Bus Speeds: Beginning 2015 (data.ny.gov) (retrieved 25 September 2026, months to August 2026; 2019 onward on the method the MTA adopted in August 2025, earlier years withdrawn while they are recalculated) — monthly mileage and operating time by route, borough, weekday or weekend and peak or off-peak; weekday local, limited and Select Bus Service routes averaged 7.64 miles an hour in January to August 2019 and 7.73 in the same months of 2026
- Metropolitan Transportation Authority — MTA Subway Customer Journey-Focused Metrics: Beginning 2015 (data.ny.gov) (retrieved 25 September 2026, months to August 2026; every line from 2017) — per line and month, weekdays: passengers, minutes above schedule on the platform and on the train, and trips more than five minutes late; 20.0% of trips more than five minutes late in January to August 2017, 15.4% in 2019 and 15.2% in 2026
- U.S. Census Bureau — American Community Survey news: 2026 updates (update of 6 August 2026) — the release date for the 2025 ACS one-year estimates is being determined while the Bureau works to a new Commerce Department order on disclosure avoidance; why the census figures here stop in 2024
- Governor of New York — Governor Hochul Highlights Record-Breaking Year of Performance and Ridership for the MTA in 2025 (2 January 2026) — nearly 1.3 billion subway trips in 2025, up 7% on 2024 and at 85% of pre-pandemic levels by the MTA's measure; bus ridership up about 8%; a record 4.65 million riders on 11 December 2025
- Permanent Citizens Advisory Committee to the MTA — Ridership Returns: Mapping Post-Congestion Pricing Ridership Trends (2026) — 1.299 billion subway rides in 2025 against 1.698 billion in 2019; weekend ridership grew faster (9.4%) than weekday (7.7%); morning-peak trips into the congestion zone up 7.1%
- Office of the NYC Comptroller — Behind Schedule: How New York City's Bus System Slow Rolls Riders (April 2025) — buses averaged 8.17 mph citywide in 2024, 6.29 mph in Manhattan, and speeds fell 0.6% over the ten years to 2024; 1.1 million daily bus riders
- Zhang, Xu, Talen, González and Li — Quantifying racial inequality in transit access across New York City, PNAS Nexus 5(5), May 2026 — within 60 minutes by transit, Black New Yorkers can reach 15.7% fewer jobs and 16.5% fewer healthcare facilities than white New Yorkers; outside evidence on the mechanism the census cannot see
- PolitiFact — Does New York City have the longest commute times? (March 2024) — the 'longest commute in the nation' claim as it circulates, checked against the Census Bureau's metro ranking
- U.S. Census Bureau — Cartographic Boundary File, 2020 Public Use Microdata Areas, New York, 1:500,000 (GENZ2020 release; the 2020 PUMA definitions used by the ACS from the 2022 files, and on every record of the 2020-2024 five-year file) — the neighbourhood shapes in the map, clipped to the shoreline
How to cite
Reference
BibTeX
Quote the version: a correction bumps it, and the corrections log says what changed. Charts can be embedded from the “Embed this chart” link under each one.