Everyone who worked in an office in March 2020 remembers the week it emptied. What happened next has mostly been told from the other side of the desk: the landlord’s occupancy rate, the employer’s mandate, the badge-swipe index, the lease that did or did not get signed. Those measures say how full the buildings are. They do not say which New Yorkers are in them.
The Census Bureau asks the question from the worker’s side. Every year it asks a sample of residents how they usually got to work last week, and “worked from home” is one of the answers, next to the subway, the bus and the car. In its 2024 survey, 12.5% of the city’s resident workers, about 510,000 people, usually worked from home. In 2019 it was fewer than one in twenty. The national share is slightly higher, so New York, the city with the most office space in the country, is at home a little less than the country as a whole.
Six years on, what the worker’s side of the story shows is a settlement. Working from home has neither faded nor taken over; it has changed little since 2023. And it has settled on a particular kind of worker: before 2020 the share at home was the same at every pay level, and now it climbs with pay, with a degree and with an office job. Seven in eight of the city’s workers still travel to work, and most of the drop in commuting since 2019 is a drop in commuting to Manhattan.
Two numbers are worth holding onto, because they answer different questions: 12.5% of the city’s workers usually work from home, and, in the monthly survey described below, about a quarter reported at least some work at home in a given week of 2026. Much of the gap between them is consistent with hybrid workers, whom the census counts as commuters, alongside differences between the two surveys themselves.
That first share hides the rest of the story, so let’s open it up: who works from home now and who never could, how the commute and Manhattan changed alongside it, and where it stands now. The census microdata behind the table, the survey records themselves, run through 2024, the latest year the Census Bureau has released. A second survey, the one behind the monthly unemployment rate, carries the picture to August 2026, and the MTA’s daily counts carry it to 30 September. Every chart is interactive, and the raw data and code sit at the end.
Working from home settled at one in eight
The census has asked the same question every year, so the curve is easy to draw. Through the late 2010s the share of city workers usually at home crept up to 4.7%. In 2021, the first full survey year after the offices closed, it was 24%, one worker in four. Then it came down, quickly at first and then by just under a point from 2023 to 2024, a small but real decline. The return to the office, as far as the census can see it, happened mostly in 2022 and 2023 and then all but levelled off.
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 survey year; the black tick is the 90% margin of error. The 2020 survey used the Bureau’s experimental weights after the pandemic wrecked its response rates, and is drawn without a margin, because the Bureau warns that its replicate weights understate the uncertainty. The census counts a person as What the census means by “worked from home” The census asks how the person usually got to work last week and offers “worked from home” as one of the answers, next to car, bus, subway and the rest. So a worker at home three days and in the office two is counted at home; one in the office three days and at home two is a subway rider. It measures the usual place of work, not whether any work happened at home.That is why the census figure is lower than surveys that ask about any remote hours. We ask that other question in the Current Population Survey section. only if that was the usual way they got to work, so a hybrid worker who is in the office three days a week is a commuter here.
One thing about that curve matters for everything below. The census measures the usual place of work, not whether any work happened at home. So the one in eight is the fully remote and the mostly remote, and the hybrid worker on a three-day office week is counted with the commuters.
Working from home now follows the pay and the job
The level settled; who is in it changed, and that is the first thing the published table cannot show. In 2019, working from home had almost nothing to do with pay: sort workers by their Own earnings in 2024 dollars The worker’s own wages, salary and self-employment income over the past twelve months. The Bureau adjusts each year’s figure to that year’s dollars; we then convert every year to 2024 dollars with the New York-metro consumer price index, so a band like $60,000 to $100,000 means the same purchasing power in 2019 and 2024. Not household income. , in 2024 dollars, and the share at home was about one in twenty in every band. The remote worker of 2019 was as likely to be a freelancer or a home day-care provider as an executive.
In 2024 the same sort gives a staircase, from 10% of workers earning under $30,000 to 19% of those earning $150,000 or more. Working from home now runs with pay in a way it did not before.
The medians say the same thing. The median remote worker earns about $73,000 against about $53,000 for the median commuter. In 2019 the two were the same.
Working from home now climbs with pay
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.City workers who usually worked from home, by their own earnings, 2019 and 2024
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
Each line is one band of Own earnings in 2024 dollars The worker’s own wages, salary and self-employment income over the past twelve months. The Bureau adjusts each year’s figure to that year’s dollars; we then convert every year to 2024 dollars with the New York-metro consumer price index, so a band like $60,000 to $100,000 means the same purchasing power in 2019 and 2024. Not household income. , in 2024 dollars: the left end is the share of those workers who usually worked from home in 2019, the right end the share in 2024, blue where it rose by more than the two years’ combined margin. Hover a line for both years’ 90% margins. The filters hold the industry and the borough fixed; the same numbers as paired bars, with their margins drawn, are in the questions at the end.
The numbers behind this chart
About these numbers. Resident workers of the five boroughs at work in the survey week, by their own earnings (wages plus self-employment income, in 2024 dollars); the census counts a person as working from home only if that was their usual way to work, so a hybrid worker is a commuter here. Part-year and part-time workers sit in the lower bands. Hover for each year's 90% margin. Rows resting on fewer than 30 survey records in either year are not drawn.
| Own earnings, 2024 dollars | 2019 | 2024 |
|---|---|---|
| Under $30k | 5.5% ± 0.5% | 10% ± 0.6% |
| $30k–60k | 4.0% ± 0.4% | 9.2% ± 0.7% |
| $60k–100k | 4.5% ± 0.5% | 13% ± 0.8% |
| $100k–150k | 4.6% ± 0.8% | 17% ± 1.2% |
| $150k and over | 5.0% ± 0.7% | 19% ± 1.1% |
The gradient is mostly about what people do for a living. By occupation, the 2024 share at home runs from 30% in arts, media and entertainment down to 2.7% in food service and in police, fire and security, with computing and finance near the top and construction and cleaning near the bottom. The occupations where working from home rose most are the ones whose work can be done away from a workplace; the occupations where it barely rose are the ones whose work cannot.
Working from home rose where the work can travel
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.City workers who usually worked from home, by occupation group, 2019 and 2024
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One row per census occupation group: the ring is 2019 and the dot 2024, blue where the rise is larger than the two years’ combined 90% margin and grey where it is not. Rows are sorted by the 2024 share; hover a row for both years’ margins and record counts. Filter by borough. The same numbers as paired bars are in the questions at the end.
The numbers behind this chart
About these numbers. The census occupation codes collapsed to the Bureau's major groups; groups with fewer than 300 records in 2024 are omitted. Sorted by the 2024 share; hover for each year's 90% margin and record count. A change inside the two years' combined margin is drawn grey.
| Group | 2019 | 2024 |
|---|---|---|
| Arts, media and entertainment | 13% ± 1.6% | 30% ± 2.5% |
| Computer and math | 8.8% ± 1.7% | 26% ± 2.1% |
| Business and finance | 5.6% ± 1.1% | 23% ± 1.8% |
| Legal | 4.7% ± 1.6% | 17% ± 2.5% |
| Management | 6.5% ± 0.8% | 17% ± 1.3% |
| Architecture and engineering | 4.2% ± 1.5% | 17% ± 3.6% |
| Science | 2.8% ± 1.4% | 16% ± 3.8% |
| Community and social service | 3.6% ± 1.2% | 14% ± 3.3% |
| Health support | 3.9% ± 0.8% | 13% ± 1.7% |
| Sales | 6.5% ± 1.2% | 11% ± 1.3% |
| Office and administrative | 3.1% ± 0.5% | 11% ± 1.1% |
| Personal care and service | 7.2% ± 1.8% | 9.4% ± 2.1% |
| Production | 2.3% ± 1.0% | 8.4% ± 2.5% |
| Education | 3.1% ± 0.8% | 8.2% ± 1.1% |
| Health practitioners | 2.6% ± 0.6% | 6.9% ± 1.3% |
| Installation and repair | 4.8% ± 2.3% | 6.3% ± 2.5% |
| Cleaning and maintenance | 4.7% ± 1.3% | 5.1% ± 1.4% |
| Transportation and moving | 2.5% ± 0.9% | 4.6% ± 1.0% |
| Construction | 2.0% ± 0.6% | 4.1% ± 1.4% |
| Food service | 1.2% ± 0.5% | 2.7% ± 0.9% |
| Protective service | 1.7% ± 0.6% | 2.7% ± 1.1% |
| Arts, media and entertainment | 16% ± 2.9% | 30% ± 5.0% |
| Computer and math | 8.9% ± 3.3% | 25% ± 4.9% |
| Business and finance | 7.9% ± 2.3% | 22% ± 3.4% |
| Legal | 3.8% ± 2.2% | 16% ± 4.4% |
| Management | 7.9% ± 1.5% | 20% ± 2.5% |
| Architecture and engineering | 6.7% ± 4.2% | 23% ± 9.3% |
| Science | 3.2% ± 2.3% | 18% ± 7.5% |
| Community and social service | 5.1% ± 4.0% | 24% ± 12% |
| Health support | 5.2% ± 3.8% | 20% ± 7.1% |
| Sales | 11% ± 3.3% | 15% ± 3.5% |
| Office and administrative | 6.0% ± 2.1% | 14% ± 3.3% |
| Personal care and service | 12% ± 6.1% | 15% ± 7.8% |
| Production | 3.0% ± 2.6% | 11% ± 8.1% |
| Education | 6.1% ± 2.9% | 10% ± 2.7% |
| Health practitioners | 4.4% ± 1.9% | 10% ± 4.2% |
| Installation and repair | – | – |
| Cleaning and maintenance | 5.8% ± 4.8% | 10% ± 6.7% |
| Transportation and moving | 0.0% | 16% ± 7.5% |
| Construction | 13% ± 9.8% | 15% ± 11% |
| Food service | 4.1% ± 2.8% | 5.8% ± 4.2% |
| Protective service | 5.5% ± 3.3% | 3.2% ± 4.0% |
| Arts, media and entertainment | 11% ± 1.9% | 31% ± 3.2% |
| Computer and math | 12% ± 2.8% | 28% ± 3.7% |
| Business and finance | 4.9% ± 1.6% | 26% ± 3.1% |
| Legal | 5.3% ± 2.9% | 20% ± 4.6% |
| Management | 7.5% ± 1.8% | 18% ± 1.7% |
| Architecture and engineering | 4.3% ± 2.6% | 16% ± 4.4% |
| Science | 4.5% ± 2.8% | 19% ± 6.6% |
| Community and social service | 3.8% ± 2.0% | 17% ± 4.6% |
| Health support | 2.4% ± 0.8% | 12% ± 3.5% |
| Sales | 4.9% ± 1.4% | 12% ± 2.4% |
| Office and administrative | 2.6% ± 0.9% | 12% ± 2.0% |
| Personal care and service | 6.0% ± 2.4% | 10% ± 3.6% |
| Production | 3.6% ± 2.2% | 9.2% ± 4.9% |
| Education | 2.3% ± 1.0% | 10% ± 2.3% |
| Health practitioners | 2.7% ± 1.2% | 7.6% ± 2.5% |
| Installation and repair | 9.3% ± 6.6% | 6.6% ± 3.4% |
| Cleaning and maintenance | 3.8% ± 2.3% | 5.0% ± 2.1% |
| Transportation and moving | 5.3% ± 2.5% | 4.3% ± 1.4% |
| Construction | 1.3% ± 1.0% | 4.4% ± 1.9% |
| Food service | 1.3% ± 1.1% | 3.9% ± 1.8% |
| Protective service | 1.4% ± 1.1% | 2.0% ± 1.2% |
| Arts, media and entertainment | 12% ± 3.5% | 30% ± 4.9% |
| Computer and math | 6.1% ± 2.4% | 29% ± 4.7% |
| Business and finance | 3.8% ± 1.4% | 23% ± 3.3% |
| Legal | 6.3% ± 4.5% | 17% ± 6.0% |
| Management | 3.5% ± 1.1% | 15% ± 2.2% |
| Architecture and engineering | 2.7% ± 2.1% | 14% ± 6.0% |
| Science | 0.0% | 12% ± 6.0% |
| Community and social service | 1.4% ± 1.3% | 8.7% ± 4.0% |
| Health support | 5.7% ± 1.7% | 13% ± 2.8% |
| Sales | 5.3% ± 1.8% | 10% ± 2.2% |
| Office and administrative | 2.5% ± 0.8% | 9.3% ± 1.8% |
| Personal care and service | 4.5% ± 2.2% | 5.6% ± 2.6% |
| Production | 2.4% ± 1.9% | 9.3% ± 3.8% |
| Education | 2.8% ± 1.3% | 5.5% ± 1.5% |
| Health practitioners | 1.4% ± 0.6% | 4.5% ± 1.4% |
| Installation and repair | 4.7% ± 3.7% | 0.8% ± 1.0% |
| Cleaning and maintenance | 2.1% ± 1.3% | 4.3% ± 2.0% |
| Transportation and moving | 1.5% ± 0.8% | 3.2% ± 1.3% |
| Construction | 1.9% ± 1.2% | 2.9% ± 1.6% |
| Food service | 0.6% ± 0.4% | 1.3% ± 1.2% |
| Protective service | 1.3% ± 1.3% | 3.4% ± 2.2% |
| Arts, media and entertainment | 9.8% ± 6.4% | 21% ± 11% |
| Computer and math | 4.3% ± 4.0% | 13% ± 5.8% |
| Business and finance | 5.4% ± 2.9% | 13% ± 5.8% |
| Legal | 3.0% ± 3.9% | – |
| Management | 4.4% ± 2.4% | 10% ± 4.0% |
| Architecture and engineering | 3.1% ± 4.6% | 9.6% ± 10% |
| Science | – | – |
| Community and social service | 3.5% ± 3.3% | 8.4% ± 4.6% |
| Health support | 3.5% ± 1.9% | 11% ± 3.0% |
| Sales | 2.3% ± 1.5% | 5.8% ± 2.5% |
| Office and administrative | 1.4% ± 1.0% | 7.8% ± 2.6% |
| Personal care and service | 7.9% ± 3.7% | 8.3% ± 5.3% |
| Production | 0.0% | 4.5% ± 3.4% |
| Education | 1.7% ± 1.6% | 5.7% ± 3.0% |
| Health practitioners | 3.6% ± 2.6% | 6.2% ± 3.6% |
| Installation and repair | 0.4% ± 0.7% | 3.7% ± 3.4% |
| Cleaning and maintenance | 6.5% ± 2.8% | 4.0% ± 2.9% |
| Transportation and moving | 2.3% ± 1.4% | 1.6% ± 1.1% |
| Construction | 0.0% | 2.4% ± 2.6% |
| Food service | 0.1% ± 0.1% | 1.0% ± 1.0% |
| Protective service | 0.9% ± 0.9% | 3.0% ± 3.0% |
| Arts, media and entertainment | 2.1% ± 3.4% | 15% ± 12% |
| Computer and math | 5.7% ± 5.3% | 24% ± 8.6% |
| Business and finance | 0.7% ± 1.1% | 15% ± 7.4% |
| Legal | – | – |
| Management | 7.4% ± 5.1% | 11% ± 5.0% |
| Architecture and engineering | – | 16% ± 18% |
| Science | – | – |
| Community and social service | 7.0% ± 5.7% | 11% ± 9.4% |
| Health support | 2.4% ± 2.6% | 4.8% ± 3.4% |
| Sales | 9.5% ± 6.5% | 7.9% ± 2.7% |
| Office and administrative | 4.7% ± 2.6% | 12% ± 4.4% |
| Personal care and service | 8.6% ± 11% | 8.3% ± 4.8% |
| Production | 0.9% ± 0.9% | 1.8% ± 2.3% |
| Education | 1.3% ± 1.4% | 4.2% ± 2.8% |
| Health practitioners | 0.4% ± 0.7% | 6.2% ± 3.3% |
| Installation and repair | 1.0% ± 1.7% | 8.6% ± 7.7% |
| Cleaning and maintenance | 13% ± 13% | 3.1% ± 5.2% |
| Transportation and moving | 0.0% | 1.9% ± 1.6% |
| Construction | 0.0% | 1.0% ± 1.2% |
| Food service | 0.0% | 2.0% ± 2.5% |
| Protective service | 0.8% ± 1.4% | 1.9% ± 2.0% |
The same split shows up as a map. Manhattan residents are at home at 18% and Bronx residents at 6.7%, with the other boroughs in between: Manhattan highest, as in 2019, and the Bronx and Staten Island lowest. Part of the Bronx’s gap is that its residents hold fewer of the jobs in which working from home spread; Bronx residents in the same industries are at home less often too.
Where the city's remote workers live
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.Workers living in each of the city's 55 census neighbourhoods (PUMAs), 2024
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 a few hundred survey records from the 2024 one-year file, so the shares carry margins of two to five points; hover a neighbourhood to see its own. Working from home is the usual way to work among workers at work in the survey week, the measure used throughout; working in Manhattan is counted among those who travel to a workplace. The colour classes are equal steps with round edges, and the triangle on the legend marks the whole city.
The numbers behind this chart
About these numbers. American Community Survey 2024 one-year microdata, civilian employed residents 16 and over, by the neighbourhood (public-use microdata area, PUMA) they live in. Shares carry 90% replicate-weight margins; medians carry none. Work in Manhattan is computed among residents who do not usually work from home and report a workplace. Shapes: Census Bureau cartographic boundary file, 2020 PUMAs, clipped to the shoreline.
| Neighbourhood | Borough | Work from home | Work in Manhattan | Median earnings |
|---|---|---|---|---|
| Midtown, East Midtown, & Flatiron | Manhattan | 16% ± 3.1% | 87% ± 2.5% | $116,800 |
| Fordham, Bedford Park, & Norwood | Bronx | 7.1% ± 2.9% | 33% ± 4.9% | $35,500 |
| Pelham Parkway & Morris Park | Bronx | 8.0% ± 2.3% | 29% ± 4.7% | $44,500 |
| Crown Heights (North) | Brooklyn | 19% ± 3.9% | 48% ± 5.5% | $60,900 |
| Co-op City & Throgs Neck | Bronx | 8.0% ± 3.3% | 34% ± 4.8% | $50,800 |
| Highbridge & Concourse | Bronx | 3.0% ± 1.6% | 35% ± 5.3% | $39,000 |
| Crown Heights (South) | Brooklyn | 16% ± 4.1% | 45% ± 6.2% | $60,900 |
| Harlem | Manhattan | 22% ± 4.3% | 82% ± 4.5% | $55,800 |
| Bushwick | Brooklyn | 17% ± 3.3% | 55% ± 5.6% | $55,800 |
| Washington Heights & Inwood | Manhattan | 12% ± 3.1% | 80% ± 4.1% | $45,700 |
| Sunset Park & Windsor Terrace | Brooklyn | 13% ± 3.2% | 41% ± 3.9% | $50,800 |
| Morningside Heights & Hamilton Heights | Manhattan | 13% ± 3.7% | 82% ± 5.5% | $45,700 |
| Jackson Heights & East Elmhurst | Queens | 6.5% ± 1.9% | 39% ± 4.6% | $42,600 |
| Lower East Side & Chinatown | Manhattan | 14% ± 3.3% | 80% ± 4.5% | $66,000 |
| Upper East Side & Roosevelt Island | Manhattan | 22% ± 3.3% | 89% ± 2.8% | $115,700 |
| South Ozone Park & Howard Beach | Queens | 11% ± 2.8% | 29% ± 5.1% | $49,700 |
| Coney Island & Brighton Beach | Brooklyn | 5.1% ± 2.3% | 27% ± 4.7% | $45,700 |
| Canarsie & Flatlands | Brooklyn | 9.0% ± 2.0% | 28% ± 3.4% | $50,800 |
| Bensonhurst & Bath Beach | Brooklyn | 9.2% ± 2.0% | 27% ± 3.5% | $40,600 |
| Flatbush & Midwood | Brooklyn | 13% ± 2.5% | 41% ± 4.8% | $50,800 |
| Williamsburg & Greenpoint | Brooklyn | 20% ± 2.7% | 55% ± 4.2% | $81,200 |
| Astoria & Queensbridge | Queens | 12% ± 2.1% | 57% ± 3.8% | $67,000 |
| Elmhurst & Corona | Queens | 5.5% ± 1.9% | 45% ± 4.1% | $38,300 |
| Ocean Hill & Brownsville | Brooklyn | 6.6% ± 3.2% | 31% ± 7.3% | $35,500 |
| Morris Heights & Mount Hope | Bronx | 5.3% ± 2.6% | 35% ± 5.8% | $30,500 |
| Borough Park & Kensington | Brooklyn | 9.5% ± 2.1% | 21% ± 3.4% | $40,600 |
| Downtown Brooklyn & Fort Greene | Brooklyn | 23% ± 4.6% | 65% ± 4.2% | $101,500 |
| Queens Village, Bellerose, & Rosedale | Queens | 9.3% ± 2.2% | 25% ± 3.7% | $50,800 |
| South Shore | Staten Island | 7.4% ± 2.3% | 22% ± 3.4% | $67,000 |
| East New York & Cypress Hills | Brooklyn | 6.8% ± 2.6% | 26% ± 3.9% | $40,600 |
| Soundview & Parkchester | Bronx | 4.8% ± 1.8% | 36% ± 5.0% | $38,600 |
| Forest Hills & Rego Park | Queens | 18% ± 2.9% | 53% ± 5.3% | $73,100 |
| Mid-Island | Staten Island | 10% ± 2.1% | 20% ± 4.3% | $60,900 |
| Jamaica, St. Albans, & Hollis | Queens | 7.2% ± 1.8% | 23% ± 2.4% | $43,800 |
| East Harlem | Manhattan | 14% ± 4.1% | 81% ± 5.3% | $42,600 |
| Bedford-Stuyvesant | Brooklyn | 14% ± 2.9% | 44% ± 3.9% | $55,800 |
| Upper West Side | Manhattan | 22% ± 3.2% | 88% ± 3.1% | $106,600 |
| Sheepshead Bay & Gravesend (East) | Brooklyn | 12% ± 2.8% | 33% ± 4.5% | $50,800 |
| Bay Ridge & Dyker Heights | Brooklyn | 13% ± 2.9% | 31% ± 4.3% | $55,800 |
| Long Island City, Sunnyside, & Woodside | Queens | 15% ± 3.6% | 55% ± 4.7% | $71,100 |
| Kew Gardens, Richmond Hill, & Woodhaven | Queens | 9.2% ± 2.1% | 31% ± 4.0% | $46,200 |
| Wakefield, Williamsbridge, & Eastchester | Bronx | 6.1% ± 2.0% | 27% ± 4.0% | $48,700 |
| Chelsea & Hell's Kitchen | Manhattan | 23% ± 4.9% | 90% ± 3.6% | $101,500 |
| Melrose, Mott Haven, Longwood, & Hunts Point | Bronx | 8.8% ± 3.4% | 42% ± 5.2% | $30,500 |
| Auburndale, Bayside, & Douglaston | Queens | 11% ± 2.4% | 24% ± 4.5% | $60,900 |
| Riverdale, Kingsbridge, & Marble Hill | Bronx | 8.9% ± 2.6% | 45% ± 4.5% | $50,800 |
| Ridgewood, Maspeth, & Middle Village | Queens | 13% ± 2.2% | 34% ± 4.2% | $55,800 |
| East Flatbush | Brooklyn | 9.8% ± 3.1% | 29% ± 4.4% | $45,700 |
| Morrisania, Tremont, Belmont, & West Farms | Bronx | 8.9% ± 2.7% | 36% ± 6.0% | $30,500 |
| Financial District & Greenwich Village | Manhattan | 15% ± 3.1% | 89% ± 2.5% | $121,800 |
| Flushing, Murray Hill, & Whitestone | Queens | 9.6% ± 1.6% | 24% ± 2.7% | $40,600 |
| Park Slope & Carroll Gardens | Brooklyn | 25% ± 4.1% | 59% ± 4.5% | $101,500 |
| Fresh Meadows, Hillcrest, & Briarwood | Queens | 11% ± 2.6% | 29% ± 5.0% | $55,800 |
| North Shore | Staten Island | 5.9% ± 1.7% | 22% ± 3.7% | $50,800 |
| The Rockaways | Queens | 9.6% ± 3.2% | 18% ± 5.0% | $50,800 |
And it shows up by who people are. White workers are at home at 17% and Black workers at 8.5%, and the questions at the end have the same comparison by degree, sex, age and nativity. Hold the degree fixed, or look only inside the office-based industries, and the racial gap stays at about eight or nine points. Part of it goes with which jobs people hold; the census cannot say what accounts for the rest.
The commute that shrank is the one to Manhattan
So let’s ask the question in the title of the workers it is usually asked about. Call a worker an office-industry worker if their employer is in information, finance and insurance, real estate, professional and technical services, management of companies, or administrative and support services: the Office-based industries Information, finance and insurance, real estate, professional and technical services, management of companies, and administrative and support services: the industries whose jobs are mostly in office buildings, as the census codes the worker’s employer. It is an industry test, not a job test: a receptionist at a law firm is in, an accountant at a hospital is out. About 30% of the city’s resident workers, 1.2 million people, are in these industries. , about 30% of the city’s resident workers. In 2024, 21% of them usually work from home, about three times the 2019 share, and they are about half of everyone in the city who does.
Which means four in five office-industry workers still travel. How they travel has shifted: the subway carried 53% of them in 2019 and 44% in 2024, and most of the change in how they get to work shows up in that one column. Two-thirds of the ones who commute are going to Manhattan, as before.
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 pair of bars is the share of office-industry workers whose usual means of transportation is that mode, 2019 and 2024. Filter by borough and earnings.
Inside these industries the pay staircase flattens out, and the borough gap closes for Brooklyn and Queens; the Bronx and Staten Island stay below. Part of the citywide gradients by pay and by borough is a gradient in who holds an office-industry job.
The commutes that went missing show up in the same census file. Between 2019 and 2024 the number of resident workers stayed about the same and the number of commuters fell by about 300,000.
Most of that loss, about three quarters, is in the subway’s column. It carried 43% of the city’s workers in 2019 and 37% in 2024, while the car’s share barely moved. The census does not follow individual workers from one year to the next, so it cannot say which 2019 commuters became 2024 remote workers. What it can say is that the groups where working from home rose most were, in 2019, disproportionately subway riders bound for Manhattan.
Manhattan is where it lands. Count the people whose place of work is Manhattan and who physically travel there, from the five boroughs and from the suburbs in the census files for the region: 2.55 million in 2019 and 2.24 million in 2024, a drop of about 305,000. Commuters from beyond the study area are in neither count, so this is the change within it rather than a complete Manhattan commuter count. About three quarters of the decline is city residents, and the chart splits it by where they live.
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 pair of bars per place of residence; the count is workers whose 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. is Manhattan and who did not usually work from home. Suburban commuters from beyond the census study area are not counted.
That is the demand side of a story the office market tells from the supply side. Roughly one Manhattan-bound worker in eight is no longer counted, and roughly one Manhattan office square foot in eight stands empty. The two figures rhyme, but the vacancy rate has causes of its own, from lease timing to new supply and demolitions, and the census cannot say how much of it is remote work.
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 things stand in September 2026 we switch to two other kinds of evidence: a monthly survey that asks a broader question about working at home, and the MTA’s own daily counts.
In 2026, the level holds
Since October 2022 the Current Population Survey, the monthly survey behind the unemployment rate, has asked everyone who was at work a plain question, and since June 2024 the answers are in its public monthly files: The CPS telework question Since October 2022 the Current Population Survey, the monthly survey behind the unemployment rate, has asked everyone who was at work: “At any time last week, did you telework or work at home for pay?” and, if yes, for how many of the week’s hours; the public monthly files carry the answers from June 2024. Any hours count, so a hybrid worker says yes. Comparing the hours at home with the hours worked separates people whose whole week was at home from people who split it.About 650 New York City residents at work answer each month. There are no replicate weights in the public file, so margins here are approximate: the spread of the monthly estimates, with a floor for the months’ shared sample, since adjacent months share most of their respondents. Any hours count, so a hybrid worker says yes, and a second question asks how many of the week’s hours were at home. The five boroughs can be picked out of the public file, about 650 employed residents a month.
From January to August 2026, about 25% of employed New Yorkers at work in a given week worked at home for at least part of it: 10% for all of it and 15% for some of it. Within the noise of a sample this size, none of the three has moved since the same months of 2025. The one in ten fully at home is the closest match to the census’s one in eight usually at home.
A quarter work from home at least part of the week, month after month
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 residents employed and at work, June 2024 to August 2026, Current Population Survey
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One line per measure, one point per month: any hours at home, some but not all of them, all of them, and the same any-hours share for the whole country. Monthly samples are small and adjacent months share most of their respondents, so read the level, not the wiggles; no margins are drawn. The lines break at October 2025, when the survey was not conducted during the federal shutdown. The same months as bars are in the questions at the end.
The numbers behind this chart
About these numbers. Asked of everyone employed and at work: did you telework or work at home for pay at any time last week? 'All hours' means the hours teleworked cover every hour actually worked that week. Monthly samples of about 650 NYC residents at work; no margins are drawn. The survey was not conducted in October 2025, during the federal shutdown, so each line breaks there.
| Group | June 2024 | July 2024 | August 2024 | September 2024 | October 2024 | November 2024 | December 2024 | January 2025 | February 2025 | March 2025 | April 2025 | May 2025 | June 2025 | July 2025 | August 2025 | September 2025 | October 2025 | November 2025 | December 2025 | January 2026 | February 2026 | March 2026 | April 2026 | May 2026 | June 2026 | July 2026 | August 2026 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Any hours at home, New York City | 24% | 26% | 25% | 25% | 27% | 27% | 24% | 26% | 26% | 26% | 26% | 23% | 24% | 21% | 23% | 22% | – | 22% | 22% | 29% | 28% | 25% | 22% | 23% | 22% | 24% | 25% |
| Some but not all hours at home | 15% | 17% | 17% | 19% | 18% | 18% | 16% | 15% | 17% | 17% | 17% | 14% | 17% | 15% | 14% | 14% | – | 14% | 13% | 19% | 18% | 15% | 13% | 13% | 13% | 14% | 15% |
| All hours at home | 8.6% | 9.3% | 7.4% | 6.1% | 8.5% | 9.5% | 8.4% | 10% | 9.0% | 8.6% | 8.9% | 8.7% | 7.6% | 6.0% | 8.7% | 7.1% | – | 7.6% | 8.9% | 9.8% | 10% | 10% | 9.1% | 10% | 9.5% | 11% | 10% |
| Any hours at home, United States | 22% | 23% | 23% | 24% | 24% | 23% | 23% | 24% | 24% | 23% | 22% | 22% | 22% | 22% | 22% | 22% | – | 23% | 23% | 23% | 23% | 23% | 22% | 22% | 22% | 22% | 22% |
The gradient is the same one the census drew. Among residents at work with a bachelor’s degree, 42% worked at home for some or all of the week in 2026; among those without, 6%. The splits by occupation and by borough look the same, and no group’s share moved by more than the survey can detect.
Degree holders work from home, and in 2026 little moved
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 residents employed and at work, January to August 2025 and 2026, Current Population Survey
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One row per group: the ring is January to August 2025 and the dot the same months of 2026, blue or orange where the change is larger than the two periods’ approximate combined margin and grey where it is not. The margins come from the spread of the monthly estimates, not from replicate weights. Switch between any, all and some hours at home; the same numbers as paired bars are in the questions at the end.
The numbers behind this chart
About these numbers. Groups are the CPS's own: occupation major groups 1 and 2 are management, business, financial and professional occupations. Hover a row for the approximate margins. A change inside the two periods' combined margin is drawn grey.
| Group | January to August 2025 | January to August 2026 |
|---|---|---|
| All employed NYC residents | 24% ± 2.0% | 25% ± 2.1% |
| Bachelor's or higher | 39% ± 3.0% | 42% ± 3.3% |
| No bachelor's | 7.0% ± 1.8% | 6.1% ± 1.7% |
| Management, business and professional occupations | 42% ± 3.3% | 43% ± 3.5% |
| All other occupations | 10% ± 1.9% | 10% ± 2.0% |
| Lives in Manhattan | 40% ± 4.7% | 42% ± 5.1% |
| Lives in the other boroughs | 20% ± 2.1% | 20% ± 2.2% |
| Full-time | 26% ± 2.2% | 28% ± 2.4% |
| Private employer | 25% ± 2.2% | 25% ± 2.3% |
| Government | 14% ± 4.9% | 20% ± 5.8% |
| Self-employed, unincorporated | 29% ± 8.5% | 29% ± 8.2% |
| All employed NYC residents | 8.5% ± 1.3% | 9.9% ± 1.5% |
| Bachelor's or higher | 13% ± 2.0% | 16% ± 2.4% |
| No bachelor's | 3.7% ± 1.3% | 3.6% ± 1.3% |
| Management, business and professional occupations | 14% ± 2.3% | 16% ± 2.6% |
| All other occupations | 4.3% ± 1.3% | 4.9% ± 1.4% |
| Lives in Manhattan | 9.3% ± 2.8% | 13% ± 3.4% |
| Lives in the other boroughs | 8.2% ± 1.5% | 9.1% ± 1.6% |
| Full-time | 8.3% ± 1.4% | 11% ± 1.6% |
| Private employer | 8.4% ± 1.4% | 9.8% ± 1.6% |
| Government | 3.5% ± 2.6% | 3.7% ± 2.8% |
| Self-employed, unincorporated | 19% ± 7.2% | 21% ± 7.3% |
| All employed NYC residents | 16% ± 1.7% | 15% ± 1.7% |
| Bachelor's or higher | 27% ± 2.7% | 26% ± 2.9% |
| No bachelor's | 3.3% ± 1.3% | 2.5% ± 1.1% |
| Management, business and professional occupations | 28% ± 3.0% | 27% ± 3.2% |
| All other occupations | 5.8% ± 1.5% | 5.5% ± 1.5% |
| Lives in Manhattan | 30% ± 4.4% | 29% ± 4.7% |
| Lives in the other boroughs | 12% ± 1.7% | 11% ± 1.7% |
| Full-time | 18% ± 1.9% | 17% ± 2.0% |
| Private employer | 17% ± 1.9% | 15% ± 1.9% |
| Government | 10% ± 4.3% | 16% ± 5.4% |
| Self-employed, unincorporated | 11% ± 5.7% | 8.8% ± 5.1% |
The MTA’s turnstiles say the same thing in a different unit. Through 30 September 2026, average weekday subway ridership stood at 75% of the MTA’s 2019 baseline, while vehicle crossings at the bridges and tunnels were at their 2019 level every day of the week.
The shape of the week is the telling part. Tuesday to Thursday run near 78% of their 2019 counts, Monday at 69% and Friday at 71%, and the commuter railroads show the same midweek hump more sharply. Subway trips include school, shopping and leisure travel, so the pattern is consistent with office work concentrated on Tuesday to Thursday rather than a measurement of office attendance.
The office week, every week since 2023
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.Ridership each day as a share of the 2019 average for the same day of the week, 1 January 2023 to 30 September 2026
The interactive chart needs JavaScript; its numbers are in the table below.
How to read this, and what it cannot show
One square per day, one column per week, Monday at the top, and one band of seven rows per year; the darker the square, the higher that day stood against the 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. , in steps of five points. Switch to the Long Island Rail Road or Metro-North: Tuesday to Thursday run ahead of Monday and Friday on the commuter railroads too, in every year. Hover a day for its figure, or tab to a year for its averages by day of the week. Trips, not commuters: subway trips include school, shopping and leisure travel, so the pattern is consistent with an office week rather than a count of it. The averages by day of the week, year by year and for the buses and bridges too, are in the questions at the end.
The numbers behind this chart
About these numbers. MTA daily ridership (data.ny.gov). Each day's count is divided by the MTA's 2019 baseline for the same day of the week, recovered from its comparable-day percentages over 2024, as in the bar chart by day of the week in the questions at the end. Trips, not commuters: subway trips include school, shopping and leisure travel. Holidays show as pale days inside darker weeks. Administrative counts, no margins.
| Month | Subway, average day | Long Island Rail Road, average day | Metro-North, average day |
|---|---|---|---|
| January 2023 | 63% | 62% | 59% |
| February 2023 | 66% | 64% | 59% |
| March 2023 | 70% | 69% | 65% |
| April 2023 | 71% | 72% | 67% |
| May 2023 | 73% | 76% | 73% |
| June 2023 | 71% | 83% | 75% |
| July 2023 | 67% | 80% | 69% |
| August 2023 | 66% | 81% | 71% |
| September 2023 | 70% | 81% | 72% |
| October 2023 | 74% | 83% | 78% |
| November 2023 | 72% | 83% | 78% |
| December 2023 | 71% | 84% | 76% |
| January 2024 | 64% | 71% | 66% |
| February 2024 | 68% | 76% | 69% |
| March 2024 | 72% | 84% | 74% |
| April 2024 | 73% | 84% | 78% |
| May 2024 | 74% | 88% | 80% |
| June 2024 | 73% | 97% | 84% |
| July 2024 | 66% | 88% | 79% |
| August 2024 | 66% | 91% | 75% |
| September 2024 | 77% | 97% | 84% |
| October 2024 | 79% | 96% | 90% |
| November 2024 | 77% | 93% | 84% |
| December 2024 | 75% | 94% | 83% |
| January 2025 | 69% | 81% | 71% |
| February 2025 | 74% | 84% | 72% |
| March 2025 | 79% | 93% | 80% |
| April 2025 | 80% | 94% | 85% |
| May 2025 | 80% | 98% | 86% |
| June 2025 | 80% | 100% | 89% |
| July 2025 | 75% | 100% | 86% |
| August 2025 | 75% | 105% | 84% |
| September 2025 | 81% | 104% | 89% |
| October 2025 | 81% | 99% | 93% |
| November 2025 | 81% | 101% | 89% |
| December 2025 | 79% | 99% | 87% |
| January 2026 | 70% | 80% | 69% |
| February 2026 | 75% | 82% | 72% |
| March 2026 | 82% | 98% | 85% |
| April 2026 | 82% | 99% | 90% |
| May 2026 | 83% | 87% | 90% |
| June 2026 | 83% | 113% | 99% |
| July 2026 | 76% | 105% | 91% |
| August 2026 | 77% | 110% | 89% |
| September 2026 | 80% | 106% | 90% |
Employers’ own reports from 2025 line up with both: a survey of major Manhattan office employers in March 2025 put about three-fifths of their workers in the office on an average weekday, and a badge-swipe index in December 2025 put office occupancy at about three-fifths of its pre-pandemic level. About three-fifths in the office by those two measures, nearly four-fifths of 2019’s midweek subway trips, a quarter of residents at work doing some work at home, one in eight usually doing all of it. Those are four measurements of one arrangement.
So who still goes to the office?
Most people, and almost everyone who was never asked. Seven in eight New York City workers still travel to a job, and for the food-service worker, the nurse, the construction worker and the police officer nothing about the pandemic changed where the work is. The workers now most likely to work from home are a specific group: paid above the median, degree-holding, in an office industry, living in Manhattan or Brooklyn, disproportionately white and US-born, and, in the monthly survey, more likely to split the week than to spend all of it at home.
The city around that group looks different too, and three separate measures say so. About 305,000 fewer people from the city and its suburbs in the census study area travel to a Manhattan job than in 2019. About one Manhattan office square foot in eight stands empty. And the turnstile counts, like the employer surveys of 2025, are consistent with an office week concentrated on Tuesday to Thursday. Those measures are consistent with one another, but they are three measurements rather than one chain of cause and effect, and none of them follows the same people over time.
The more striking fact, six years on, is how settled they all look. The census share had all but levelled off by 2024; the monthly survey has been flat since mid-2024; the turnstile counts, which gained about six points of their 2019 level from 2024 to 2025, gained about two in the same months of 2026. The office did not come back and it did not go away. It found a level, and the level has a very particular set of people in it.
Cut the data your own way
The dashboard runs on the same 66,000 census records as the 2019 and 2024 comparisons above. Pick a year, a borough, an industry, an earnings band and a degree, and see the share at home, who commutes by subway, who travels to Manhattan, and what those workers earn.
Share usually working from home by occupation group, 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.
Year
The one-year census file: 2019 is the last full year before the pandemic, 2024 the latest published. Each year is its own sample of about 33,000 city workers.
Borough
Where the worker lives, from the census's public-use microdata areas, which nest inside the five boroughs. Not where the job is. We never go below the borough.
Industry
The industry of the worker's employer, grouped from the census industry codes. "Office-based industries" are information, finance and insurance, real estate, professional and technical services, management of companies, and administrative and support services. It is an industry test: a receptionist at a law firm is in, an accountant at a hospital is out.
Own earnings
The worker's own wages plus self-employment income over the past twelve months, converted to 2024 dollars with the New York-metro CPI. Not household income.
Degree
Whether the worker holds a bachelor's degree or higher.
Usually worked from home
The share whose answer to "how did you usually get to work last week" was "worked from home". A hybrid worker who was in the office most days is a commuter here. A weighted tabulation of the census records in the selection, with an approximate 90% margin beside it (see the ? on the tile).
Workers, 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 roughly 120 workers in a one-year file, so treat a small selection as an order of magnitude.
Of commuters, take the subway
Among the selected workers who did not usually work from home, the share whose usual means of transportation is the subway.
Travel to a Manhattan workplace
The share of the selected workers whose place of work is Manhattan and who did not usually work from home. For Manhattan residents this includes trips within the borough.
Earn $100,000 or more
The share of the selected workers whose own earnings, in 2024 dollars, are $100,000 or more.
Work 35 hours or more
The share whose usual hours per week are 35 or more.
By occupation group
Each bar is the share usually working from home for one of the Bureau's major occupation groups, under the filters above. Groups with fewer than 30 records in the selection are marked too few.
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 workers
Fewer than 30 census records match, so the numbers would be noise. Widen a filter.
Questions this raises
How well do these numbers match the Bureau’s own?
Closely, in every year checked, and that is the check everything else rests on. The Bureau’s 2019 table reports 4,032,673 ± 21,303 city workers and 187,030 ± 8,791 of them working from home, a share of 4.6%; rebuilding the same universe from the 2019 microdata gives 4,047,572 ± 25,555 workers and 188,639 ± 10,158 at home, 4.7% ± 0.3, 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. . For 2024 the table reports 4,070,607 ± 20,036 workers and 509,852 ± 12,449 at home; the microdata gives 4,065,043 ± 24,289 and 509,012 ± 13,017. The 2021, 2022 and 2023 tables and the five boroughs in 2019 and 2024 reproduce the same way, as does the subway’s share of workers in each year. The full comparison is in the companion files.
Why is the census figure so much lower than the surveys that say a third of people work from home?
Because it asks a different question. The census asks how the person usually got to work last week and takes one answer; someone at home two days and on the subway three is a subway rider. The Current Population Survey asks whether any paid work happened at home, and the Bureau of Labor Statistics’ time-use survey asks where work happened on a particular day; both count the hybrid worker. So the three national figures for 2024 and 2025, 13% usually at home, 22% with any hours at home, 34.5% working at home on days worked, describe one arrangement three ways. For New York the first two are 12.5% and about 25%, and the monthly survey’s 10% whose whole week is at home is the nearest thing to the census’s one in eight.
What is the remote worker like, in one profile?
Compared with a commuter, a worker who usually works from home in 2024 is far more likely to be in an office-based industry (49% against 27%), to hold a bachelor’s degree (69% against 48%), to earn $100,000 or more (37% against 25%), to be white (48% against 32%), to live in Manhattan (32% against 21%), to be self-employed (21% against 9%) and to be a woman (53% against 49%), and less likely to be Black or Hispanic (31% against 46%) or foreign-born (31% against 45%). Every one of those differences exceeds the sum of the two margins. They are about as likely to be under 35 and to work full time. In 2019 the remote worker was already whiter, more Manhattan and more often self-employed than the commuter, but earned the same and was no more likely to be paid $100,000; the pay gap is new since 2019, and the degree gap, already there, has widened.
Did people at home simply move out of the city?
Not in these figures: the census counts residents by where they live, and the number of city residents at work was about the same in 2019 and 2024, 4.05 million and 4.07 million. Someone who left for Westchester and kept a Manhattan job is counted there, not here, which is why the Manhattan commuter count above adds the suburban files. Whether the same people who were at home in 2021 are still at home in 2024 cannot be seen either; the census does not follow individuals through time.
How steep is the pay gradient, band by band?
In 2019 the share at home was 5.5% under $30,000, then 4.0%, 4.5%, 4.6% and 5.0% up through $150,000 and over. In 2024 the same sort gives 10% under $30,000, 9.2% from $30,000 to $60,000, 13% from $60,000 to $100,000, 17% from $100,000 to $150,000 and 19% above. The top band gained 14 points and the bottom band five. Inside the office-based industries the staircase flattens out: 20% at home under $30,000, then 16%, 22%, 24% and 21% up the bands.
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
Blue is 2019, orange is 2024, one pair per earnings band, with the 90% margins. The filters hold borough and industry fixed.
In which occupations did working from home rise, and in which did it not?
In 2024 the share at home is 30% in arts, media and entertainment, 26% in computing and 22% in business and finance, 17% in law and in management, 8% in teaching, 7% among doctors and nurses, 5% in cleaning, 4.1% in construction and 2.7% in food service and in police, fire and security. In 2019 every one of those groups was under 13%, and most were under 7%.
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 pair of bars per census occupation group, sorted by the 2024 share. Filter by borough.
Does the racial gap remain among workers with the same degree, or in the same industries?
It stays, at about eight or nine points. Citywide, white workers are at home at 17%, Asian workers at 12%, Hispanic workers at 9% and Black workers at 8.5%. Workers with a bachelor’s degree are at 17% against 8% without one; women slightly more than men; US-born workers at 15% against 9% for the foreign-born. Among degree holders, white workers are at home at 20% and Black workers at 12%. Inside the office-based industries it is 24% against 16%. Part of the gap goes with which jobs people hold, and the census cannot say what accounts for the rest.
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 pair of bars per group. Hispanic of any race; the other groups are non-Hispanic. Filter by borough.
Where does the drop in Manhattan commuters come from?
Mostly from the city. Of the drop of about 305,000, city residents make up about 230,000, from 1.87 million to 1.64 million, and suburban residents of the census study area about 74,000, from 681,000 to 607,000. Brooklyn sends 66,000 fewer people to Manhattan jobs each day than it did, the Bronx 38,000 fewer, Queens 33,000 fewer and New Jersey 42,000 fewer. Citywide, the number of commuters fell from 3.86 million to 3.56 million; the car’s share of workers went from 26% to 25%, walking lost about a point and the bus held about the same. Among office-industry workers the car went from 17% to 14%, and walking held about the same.
Is the empty office space the same story?
Only partly, and the census cannot say how large the part is. Manhattan’s office vacancy rate was under 8% at the start of 2020, reached 16% in April 2024, the highest since the early 1990s, and had come back to about 13% by March 2026 as leasing recovered and older buildings were taken out of the stock; it was still 13.1% at the end of June. Roughly one Manhattan-bound worker in eight is no longer counted and roughly one office square foot in eight stands empty, but the vacancy rate has causes of its own, from lease timing to new supply and demolitions.
Who worked from home in 2019, and what did they earn?
In 2019 the median worker at home, the median commuter and the median worker overall all earned about $55,000 in 2024 dollars, and about a quarter of each were paid $100,000 or more, a difference inside the margin. The 2019 remote worker was a mix: 36% were self-employed, and the share at home was highest in arts and media, computing, personal care, and, about equally, management and sales. In 2024 the self-employed are still at home at the highest rate of any class of worker, 26%, but they are now a fifth of the remote workers rather than a third; most of the rest of the growth is salaried employees of private firms, whose share at home went from 3.5% to 12%. Government employees went from 1.8% to 5.7%.
Is the Bronx gap all jobs?
Partly. Manhattan residents are at home at 18%, Brooklyn at 14%, Queens at 10%, Staten Island at 7.6% and the Bronx at 6.7%; in 2019 they ran from 7.6% in Manhattan to 2.9% in the Bronx. Bronx residents in the office-based industries are at home at 12%, and Staten Island residents at 16%, against 20 to 23% for residents of Manhattan, Brooklyn and Queens in the same industries, so a gap remains inside the industries, and a Bronx resident earning $100,000 to $150,000 is at home at 11% against 20% for a Manhattan resident in the same band. The census cannot say why. Occupation within an industry, the employer’s policy, the kind of home available for working in, and who moved where since 2020 are all candidates, and only the first of those is in the data.
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
Residence, not workplace. Filter by earnings and industry to see how much of the borough gap is the jobs its residents hold.
Why not use the five-year file?
Because it pools 2020 through 2024, and the thing being measured changed every year in that window: the pooled share at home is 17%, which is no year’s number. The one-year files are smaller, about 33,000 city workers each, so some cuts are thin and are labelled that way, but each one is a clean year. The 2020 file is drawn but flagged: the Bureau published it with experimental weights and no standard tables, and its 21% should be read as an order of magnitude, and it carries no margin, because the Bureau warns that its replicate weights understate the uncertainty.
How much should the monthly survey be trusted?
First, what else it shows. In the same months of 2025 the three shares were 24%, 8.5% and 16%, against 25%, 10% and 15% in 2026, and the city sits a little above the national figure of about 22%. In management, business and professional occupations 43% had hours at home, against 10% in all other occupations together; Manhattan residents 42%, the other four boroughs together 20%. Government employees went from 14% in 2025 to 20% in 2026, the largest change, though inside the approximate margin.
As a level with a range, not as a monthly reading. Each month’s New York City sample is about 650 employed people, so a single month’s share has a margin of several points, and the survey re-interviews the same households in consecutive months, so adjacent readings are not independent. The margins drawn on the group chart come from the spread of the monthly estimates over the period and are approximate. Two more cautions. The Bureau updates its population controls every January, which shifts the weighted counts, so shares rather than counts are compared across years. And the sample shrank after the shutdown in the autumn of 2025, which is why the September to December 2025 figures are shown but not leaned on.
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 month; switch between any hours at home, all hours, some hours, and the national figure. Monthly samples are small and adjacent months share most of their respondents, so read the level, not the wiggles. No file exists for October 2025, when the survey was not conducted during the federal shutdown.
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 pair per group; the ticks are approximate margins from the spread of the monthly estimates. Switch between any, all and some hours at home.
What do the employer surveys and badge counts measure that the census does not?
Jobs rather than residents, and days rather than usual arrangements. The Partnership’s survey asks Manhattan’s large office employers what share of their workers are in on an average weekday, which includes commuters from New Jersey and Connecticut and excludes every worker outside a Manhattan office; its spring 2025 figure of 57%, which it puts at 76% of pre-pandemic attendance, is an attendance rate, not a remote-work share; a tenth of those workers were in five days a week, three in ten on a three-day schedule and 8% fully remote. Kastle’s index counts badge swipes in the buildings it secures, relative to the same buildings before the pandemic (New York was at 59.5% in the week of 8 December 2025, a post-pandemic high), and most of the largest Manhattan portfolios are not its clients. The MTA counts trips. Each is a good measure of something the census cannot see, and none of them can say who the workers are.
Here are the MTA’s counts as averages by day of the week, year by year and 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 day of the week; pick the year, and switch between the subway, the buses, the two commuter railroads and the bridges and tunnels. Trips, not commuters, and no margins: these are 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. for that day of the week.
Where this data goes quiet
Anything below the borough. How many days a week the census’s commuter actually goes in, since the census takes one usual answer. The second job. Whether the remote worker’s employer is in Manhattan, since a person at home has their home as place of work. Reverse commuters and suburban residents beyond the census study area. Whether the same people were at home in 2021 and 2024. And the microdata is a disclosure-protected sample of the survey, with a few top-coded values, 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 anchor. Census table B08301 (means of transportation to work, workers 16 and over; the final row is “worked from home”) for the city, its five counties and the United States, from the one-year survey: 2019 from the Bureau’s sequence-format summary file, 2021 to 2023 from the table-based summary files, 2024 from the Census Reporter mirror. No one-year table exists for 2020.
Census microdata. The one-year public-use microdata samples for 2015 to 2024 covering New York, New Jersey and Connecticut, cut to a metro study area of 156 microdata areas and then to residents of the five boroughs who were at work in the reference week and therefore have a means of transportation: about 33,000 person records a year, carrying 80 replicate weights and representing about four million workers. Working from home is the usual-means code 11; commuters are everyone else. Own earnings are wages plus self-employment income, adjusted by the Bureau’s within-year factor and then to 2024 dollars with the New York-metro consumer price index, and cut into five fixed bands. Industry is collapsed to ten sectors, with sectors 51 to 56 defined as office-based; occupation to the Bureau’s major groups, the 2015 to 2017 files carrying the 2010 codes whose major-group ranges coincide to within a few codes. Place of work is at borough level, Manhattan coded 3800 in the 2010 geography used through 2021 and 4100 in the 2020 geography from 2022. A worker physically commutes to Manhattan if the place of work is Manhattan and the usual means is not working from home; because a person at home reports home as their place of work, that condition binds only for Manhattan residents. The 2020 file carries the Bureau’s experimental weights and is flagged wherever it appears.
The monthly survey. Current Population Survey basic monthly public-use files from June 2024, when the telework items, asked since October 2022, entered the public monthly files, to August 2026. The items ask everyone employed and at work whether they teleworked or worked at home for pay at any time last week and, if so, how many of the week’s hours across all jobs. Residents of the five counties are identified through the county code; the universe is employed-and-at-work, 16 and over, weighted by the final person weight. Monthly city samples run 570 to 720 people. No file exists for October 2025, when the survey was not conducted during the federal shutdown, and the sample is smaller from November 2025. “All hours at home” means the hours teleworked cover every hour actually worked that week at all jobs. The extraction reads the Bureau’s record layouts and the fixed-width files directly.
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 by day of the week. The 2019 baseline for each mode and each weekday is recovered from the MTA’s frozen 2020 to 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 by day of the week.
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 weighted means, medians weighted medians without a margin. The monthly survey’s public file has no replicate weights, so a pooled estimate from it is the mean of the monthly weighted estimates and its standard error is the larger of two figures: their standard deviation divided by the square root of the number of months, which understates the error because consecutive months share three quarters of their sample, and a floor, the binomial error with the records counted as the distinct people they are and a design effect of 1.5. Both are approximations, good for telling large changes from small ones. Every combination the chart controls can reach is precomputed on the microdata with the same estimators and a margin on each bar, and shipped as one file; cells resting on fewer than 30 survey records are labelled too thin rather than drawn. The dashboard carries an approximate margin on its headline figure, explained in its own note.
Does it reproduce? Every published count, share and borough share lands inside the margin in every year checked (2019 and 2021 to 2024), as does the subway’s share of workers, which is what checks the mode coding. The year-by-year comparison is the reproduction table.
The series. The share at home ran 4.0% in 2015, 4.7% in 2019, 24.0% in 2021, 16.2% in 2022, 13.4% in 2023 and 12.5% ± 0.3% in 2024; the 2023 to 2024 difference, 0.9 points, is beyond the combined margin of about 0.5. Commuters fell by about 303,000 between 2019 and 2024 while the number of workers stayed about the same. The subway’s share of workers fell 5.8 points, the car’s 1.0.
Who is at home. In 2019 the earnings bands were within 1.5 points of each other; in 2024 they run from 9.2% to 18.7%, and the median worker at home earns $73,098 against $52,793 for the median commuter — in 2019 both medians fell on the same heaped value, about $54,600. Inside the office-based industries the gradient largely vanishes (20.3%, 16.3%, 22.5%, 24.1% and 20.8% up the bands) and the borough gradient closes for Brooklyn and Queens but not for the Bronx, 12.2%, or Staten Island, 16.3%, against 20.4 to 22.6%. The 8.9-point difference between white and Black workers citywide is 7.9 points among degree holders and 8.7 points inside the office-based industries; the census cannot decompose it further than occupation and industry. The self-employed were the original remote workers, 18.7% in 2019 and 25.6% in 2024; private employees went from 3.5% to 12.0% and government employees from 1.8% to 5.7%.
Manhattan. The count of workers physically travelling to a Manhattan workplace fell by 305,409, or 12.0%, of which 231,081 are city residents and 74,328 suburban residents of the study area. Commuters from beyond the study area — the upper Hudson Valley, Pennsylvania and further — are not counted, so the study area’s suburban count of 607,046 in 2024 is a floor for inbound commuting, not a total.
2026 in the monthly survey. The any-hours share is 24.4% in January to August 2025 and 24.9% in the same months of 2026; the all-hours share 8.5% and 9.9%; the some-hours share 15.9% and 15.0%. Every difference is inside the approximate margins. The gradient reproduces the census’s: 41.7% of degree holders against 6.1% of everyone else, 42.7% in management, business and professional occupations against 10.4%, Manhattan residents 42.0% against 20.0%. Government employees went from 13.8% to 19.8%, inside the approximate margins.
2026 in the transit counts. In 2026 to date the subway runs at 77 to 80% of 2019 on Tuesday to Thursday, 69% on Monday, 71% on Friday and 88% at weekends: about 4.22 million trips on a midweek day against a 2019 baseline of 5.40 million, and 3.87 million on a Friday against 5.44 million. The Long Island Rail Road is at 92% on Wednesdays and Thursdays and 77% on Mondays, and above 2019 at weekends; Metro-North 85 to 89% midweek and 75% on Mondays; bridge-and-tunnel crossings 98 to 104% every day. The midweek plateau has risen from 61% in 2022 through 69% in 2023, 71% in 2024 and 76% in 2025 to 78% in 2026.
One arrangement, three definitions. Nationally the same thing is 13% in the census, which asks for the usual means of getting to work; about 22% in the monthly survey, which asks about any hours last week; and 34.5% in the time-use survey, which asks whether any work was done at home on a diary day. None of the three is wrong, and none of them is comparable to the others.
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 66,000 city resident workers in the 2019 and 2024 files, with their survey weights and the derived variables, so a weighted mean of its worked_from_home column within a year reproduces the headline share.
Continue exploring New York
Most of New York City's recent unemployment gap is people out of work six months or more
New York City's unemployment rate was 4.8% in September 2026, above the nation's; in the survey's latest year, most of the gap was spells of six months or more.
After inflation, New York City's minimum wage nearly doubled since 1989, and about half its wage-earners still make under $30 an hour, as they did then
Through September 2026, 52% of wage-earning New Yorkers earned under $30 an hour.
- Explore this dataset: the data and the code behind every number above.
- All articles · Search · RSS
SparkyData Weekly
Every week, the most surprising things New York's public data reveals, and the evidence behind them.
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 — Means of Transportation to Work (B08301), New York city, ACS 2019 1-year — 4,032,673 ± 21,303 workers, 187,030 ± 8,791 of them worked from home (4.6%); the pre-pandemic anchor, taken from the Bureau's summary file
- U.S. Census Bureau — Means of Transportation to Work (B08301), New York city and United States, ACS 2024 1-year — 4,070,607 ± 20,036 workers, 509,852 ± 12,449 worked from home (12.5%); nationally 13.3%; the five counties from the same release
- U.S. Census Bureau — the same table for 2021, 2022 and 2023 (1-year summary files) — 24.1% of city workers at home in 2021, 16.1% in 2022, 13.4% in 2023; no one-year table was published for 2020
- U.S. Census Bureau — ACS 2015–2024 1-Year PUMS (New York, New Jersey, Connecticut), person files with replicate weights — every count, share and margin labelled census; the 2020 file carries the Bureau's experimental weights
- U.S. Census Bureau and Bureau of Labor Statistics — Current Population Survey basic monthly public-use files, June 2024 to August 2026, with the telework items — the 2026 view: any telework last week, and hours, for employed residents of the five boroughs; no October 2025 file exists because the survey was not conducted during the federal shutdown
- U.S. Census Bureau — Current Population Survey Telework Items, technical documentation (January 2024) — the question wording and universe for the telework items
- 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
- 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 baseline for each day of the week is recovered
- Partnership for New York City — Return to Office Survey Results, March 2025 — employers report 57% of Manhattan office workers at the workplace on an average weekday, 76% of pre-pandemic attendance; 8% fully remote, 10% in five days a week, 30% three
- Kastle Systems — Back to Work Barometer hits all-time post-pandemic highs (week of 8 December 2025) — badge-swipe occupancy in the buildings Kastle secures: New York 59.5% of the pre-pandemic level
- Office of the NYC Comptroller — Spotlight: New York City's Office Market (May 2024) — Manhattan office vacancy under 8% at the start of 2020 and 16% in April 2024; about 600 million square feet of office space in Manhattan
- PropertyShark (Yardi Research) — NYC Office Market Report, Q1 2026 (27 May 2026) — Manhattan office vacancy 13.1% at the end of March 2026
- PropertyShark (Hubble and Yardi Research) — Q2 NYC Office Market Report (28 July 2026) (second quarter of 2026) — Manhattan's vacancy rate held at 13.1% at the end of June 2026, unchanged from March
- U.S. Bureau of Labor Statistics — American Time Use Survey, 2025 results, Table 6 (25 June 2026) — nationally, 34.5% of employed people did some or all of their work at home on days they worked in 2025; 51.4% of those with a bachelor's degree, 19.0% of high-school graduates
- 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
- 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) — 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.