Are remote jobs going away? What the federal data says

No. In March 2026, 22.6 percent of workers teleworked or worked at home for pay, and BLS noted the rate had ranged from 21.5 percent to 23.0 percent over the preceding twelve months (Current Population Survey, published April 16, 2026). That is a flat line. The most recent reading, August 2026, was 21.6 percent - the low end of the same band, not a break from it.

What has changed is not the amount of remote work but its distribution. It has settled into a short list of occupations and industries, and outside that list it is close to nonexistent. So the question worth asking is not whether the rate is falling - it is not - but which side of that list your occupation is on. In August 2026 the two sides were 55 percentage points apart.

Where the telework rate has gone since 2022

BLS began asking the telework question in the monthly Current Population Survey in October 2022. Four readings from that series:

How to read this. The rise from 17.9 percent to 23.0 percent happened in the first eighteen months of the series. One wording change sits inside that window: BLS rewrote the introduction to the telework questions in December 2023 to drop the reference to the pandemic, while leaving the questions themselves unchanged (Telework trends, March 25, 2025). Since then the number has moved inside a band about a point and a half wide. These are not seasonally adjusted, so a March figure and an August figure are not strictly comparable; the drop from 22.6 to 21.6 is smaller than it looks, and BLS itself describes the recent record as a range rather than a trend.

That matters for the way the question usually gets asked. "Are remote jobs going away" implies a slope. Four years of federal data show a step up, then a shelf.

Which occupations remote work actually lives in

In August 2026, the telework rate by occupation group ran from 60.8 percent down to 5.8 percent - a spread of 55 percentage points, in a year when the national rate moved by a point and a half.

Telework rate by occupation group, August 2026

Share of people at work who teleworked or worked at home for pay. Not seasonally adjusted.

Computer and mathematical		60.8%34.9% teleworked all hours
Business and financial operations		52.8%26.5% all hours
Management, professional, and related		35.9%
Service occupations		5.8%

A Software Developer sits at the top of that chart. BLS puts 1,905,400 people in software developer, quality assurance analyst and tester jobs in 2025, with a median wage of $134,040 (May 2025) and projected growth of 10 percent, or 185,400 jobs, from 2025 to 2035. Table A-44 adds the detail that makes the point: in computer and mathematical occupations, 46.8 percent of all hours worked in August 2026 were teleworked, against 15.0 percent across all occupations. Remote work in this field is not an accommodation a few people negotiated. It is close to half the work.

What this number is not: an occupation group is not a job title. "Computer and mathematical occupations" holds network administrators and computer support specialists alongside developers, and the group rate is an average across all of them.

Who inside those occupations actually gets it

Two further cuts of the August 2026 data narrow the picture. By education, 40.6 percent of workers with an advanced degree teleworked, against 5.0 percent of those without a high school diploma (Table A-41). By age the gap is sharper, and it gets far less attention: in March 2026, 6.7 percent of workers aged 16 to 24 teleworked, compared with 25.1 percent of those aged 25 to 54 and 24.1 percent of those 55 and over (TED, April 16, 2026).

Hold onto that second comparison if you are early in a career. A remote-heavy occupation does not hand out remote work at the entry level. A junior Software Developer is entering a field where 46.8 percent of all hours were teleworked in August 2026 (Table A-44), while across the whole workforce only 6.7 percent of workers aged 16 to 24 teleworked at all. The two figures are measured on different populations and do not divide into one another, but the age one is what describes the person doing the applying. The federal data does not say why that gap exists. It records where people worked, not what their employer required of them.

One more cut, because it runs against the usual assumption: 25.9 percent of workers with children under 18 teleworked in August 2026, against 19.9 percent of those without. Six points. Real, and much smaller than the education gap.

Is it remote if you telework two days a week?

The survey does not decide that for you, but it splits the count. In August 2026, 10.6 percent of people at work teleworked all their hours and 11.0 percent teleworked some (Table A-41). Roughly half of everyone who teleworks at all is fully remote, and fully remote work is about one worker in ten.

Does a remote-heavy field mean fully remote, or just some days?

Table A-42 splits every occupation group into people who teleworked all their hours and people who teleworked some of them, and the ratio between those two does not follow the headline rate.

All four rows are Table A-42, August 2026. The final column is arithmetic on the two to its left, not a BLS series; nationally it is 49 percent.

Read down that last column and the ranking inverts. Service occupations barely telework, but the small number who do are the most likely to do it entirely from home. The likeliest reason is that service work doable from home at all has no on-site half, though the survey does not ask. Management, professional and related occupations, the largest group in the table, are the most hybrid of the four: fewer than half of their teleworkers were remote for all their hours.

For a Software Developer this is the column that decides whether a job is portable. Roughly three in five teleworking computer and mathematical workers were fully remote in August 2026. In business and financial operations, where a desk-based Claims Adjuster would be counted, it was one in two. A field can be remote-heavy and still be a commute.

What this number is not: the survey records one reference week, so a worker who happened to spend that week entirely at home counts as all-hours whether or not it is their usual schedule.

Why one federal survey says 35 percent

Because it asks a different question. The American Time Use Survey, whose 2025 results were released on June 25, 2026, found that 35 percent of employed people did some or all of their work at home on days they worked, while 70 percent did some or all of it at their workplace. Those add to more than 100 because the same person can appear in both. The education split below is reported for workers aged 25 and over, which is how ATUS publishes it.

Work done at home on days worked, 2025

35%All employed people

51%Bachelor's degree or higher, age 25+

19%High school diploma and no college, age 25+

The ATUS figure is higher because it counts any work performed at home on a day the person worked, including a Claims Adjuster writing up an hour of notes after dinner. The CPS telework question is narrower. Neither is wrong; they measure different things, and quoting one against the other is how "remote work is collapsing" and "remote work is everywhere" both get published in the same week.

When an occupation group's rate misleads you

Take a Claims Adjuster. Adjusters are counted in business and financial operations occupations, which posted a 52.8 percent telework rate in August 2026 - behind only computer and mathematical occupations among the groups in the chart above. Insurance carriers sit in financial activities, which BLS reports as the industry with the highest telework rate of all: 52.5 percent in June 2026, ahead of information at 47.2 percent and professional and business services at 41.8 percent.

Read those two numbers and you would expect claims adjusting to be a remote field. Then read what the Occupational Outlook Handbook says the work is: adjusters "often work outside the office, inspecting properties on which insurance claims have been made, such as damaged automobiles and buildings," and they work irregular hours to meet clients in the evening and at weekends.

Software Developer and Claims Adjuster, side by side

Telework rates are for the whole occupation group, August 2026. Pay and employment are BLS OOH, May 2025 wages and 2025 employment.

	Software Developer	Claims Adjuster
Occupation group	Computer and mathematical	Business and financial operations
Group telework rate, Aug 2026	60.8%	52.8%
Median annual pay, May 2025	$134,040	$78,020
Employment, 2025	1,905,400	389,700
Projected change, 2025–35	+10% (+185,400)	−6% (−21,800)
What the OOH says about location	Typically work in offices with team members	Often work outside the office, inspecting damaged property

Here is the position worth arguing with: for most people, the occupation you are searching decides whether you find remote work, and no amount of search technique will move that. A group rate of 52.8 percent is an average across financial analysts, accountants, HR specialists and adjusters, and the desk-bound half of that list is carrying it. A field adjuster does not get a remote job by writing a better resume. They get one by moving into desk claims, subrogation or claims examining - titles inside the same insurance company that do not require standing in a driveway with a clipboard.

Two caveats keep that honest. First, the group rate is real for someone already in a remote-capable role inside it. Second, "−6 percent over ten years" is a projection about the occupation, not a statement about how hard it is to get hired next month.

Which industries the remote work is in

Industry rates, June 2026, from the Current Population Survey:

Financial activities splits almost evenly between people who teleworked all their hours (27.1 percent) and some of them (25.4 percent), with 47.5 percent not teleworking at all. That last figure is the one people forget: in the most remote-friendly industry in the country, nearly half of workers were on site.

The distance from financial activities to leisure and hospitality is 44.8 percentage points, and it is not a policy gap that a change of corporate mood could close. It is a description of the work. A Claims Adjuster's file can be examined from a laptop; a hotel front desk cannot be staffed from one. Any forecast about remote work that does not begin with that distinction is a forecast about a mood.

What the telework data does not measure

Two things, and both matter more than the headline rate.

It does not count postings. The CPS asks people who already have jobs whether they teleworked during the survey reference week. It measures the stock of remote workers, not the flow of new remote openings - and BLS publishes no series on the share of advertised jobs that offer remote work. A company can hold its remote headcount steady while advertising every new opening as on-site, and the telework rate would not move for years.

Table A-44 shows why the headline rate is the wrong thing to watch first. In August 2026 the average work week was 38.4 hours and 15.0 percent of all hours worked were teleworked. A Software Developer moved from five remote days a week to two moves that hours figure immediately and the telework rate not at all, because they are still counted as someone who teleworked. If remote work erodes, it shows up in hours before it shows up in headcount.

It does not capture employer policy. A worker who teleworks on Tuesdays because their manager allows it is counted the same as one hired into a distributed team, and neither is asked whether the arrangement is under review. When you read that a large employer ordered everyone back four days a week, that is a real event which the national rate is not built to show, because one employer moves a small share of a national count that covers everyone at work.

So the correct answer to "are remote jobs going away" is narrower than it sounds. Remote work is not going away. Whether remote postings in your field are shrinking is a separate question, and no federal series answers it.

If you are applying for this role

Three specifics.

Search the occupation, not the arrangement. If you are in a field whose group rate is under 10 percent, "remote" as a filter returns almost nothing, and the productive move is a title change inside the same industry - an adjuster searching examiner, subrogation specialist or claims analyst rather than remote adjuster.

Expect the split. About half of teleworkers work all their hours from home; the other half do not. A posting labeled remote in a 52 percent industry is as likely to mean two days at home as five, so read the location line and the schedule line before the salary line.

Ask which one it is, in the screen. The question that gets a straight answer is "how many days a week is this role expected on site, and has that changed in the last year?" Both halves matter. The second one is what the telework rate cannot tell you and a hiring manager can.

If the constraint is how many of these you can get through in a week rather than which ones to send, HeroApply fills the application form from your resume so the filtering is where you spend the time.