Return to Office Statistics (2026)

Office attendance in the ten largest US metros was 55.4% of its pre-pandemic level in the week ending 30 September 2026, according to Kastle Systems' badge-swipe data. That is within a point of the record set in December 2025 and about where it has been for a year. Roughly 26% of paid working days are done from home, two and a half times the 2019 share. Of employees who can do their jobs remotely, 52% are hybrid, 26% fully remote and 22% fully on site.

The return-to-office argument is conducted with numbers from different sources that measure different things: badge swipes, foot traffic, household surveys, employer policy audits and engagement panels. The pages that rank mix them freely and present five-year-old survey results as current. This page keeps the measures apart, gives the sample and date for each, and covers how much of the week is in the office, what employers now require, what mandates do to attrition and engagement, what workers say they would do, and what the research on productivity has found.

How full are offices, and how is it measured?

Three series track office attendance, and they are not interchangeable.

Kastle's 55.4% means swipes are a little over half of what they were in early 2020. Its record is 56.3%, set in the week before Christmas 2025. Placer.ai's series runs higher because it counts anyone entering an office building, including visitors and workers in buildings without badge systems, and it shows nationwide visits a quarter below 2019 rather than nearly half. Kastle also publishes a peak-day figure: on Tuesdays, occupancy in its ten metros runs above 60%, and on Fridays below 40%. The weekly average hides a week that is full in the middle and empty at both ends.

How much of the week is remote

The Survey of Working Arrangements and Attitudes, run monthly by Barrero, Bloom and Davis since May 2020, is the longest-running measure of the share of paid days worked from home. Its October 2026 update from 38,070 respondents puts the share at about 26%, against roughly 7% in 2019 and 61% at the May 2020 peak. The share has been between 25% and 28% since early 2023, which is the clearest evidence that the post-pandemic arrangement has settled. Among workers, 12% are fully remote, 26% hybrid and 62% fully on site, a split that covers everyone including the two thirds of jobs that cannot be done remotely.

The BLS American Time Use Survey offers an independent check. In 2025, 35% of employed people did some or all of their work at home on days they worked, up from 24% in 2019. Among workers with a bachelor's degree or higher, it was 51%. The two surveys agree that remote work roughly doubled or tripled depending on the measure and then stopped moving.

Among remote-capable employees only, Gallup's hybrid work indicator for May 2026 shows 52% hybrid, 26% exclusively remote and 22% fully on site. In 2019 the same group was 60% on site. Hybrid employees average 2.6 days a week in the office.

What employers require

The Flex Index, which audits published policies at US firms, found in its third-quarter 2025 report on 9,132 companies that 34% required full-time office attendance, 42% had a structured hybrid policy with a set number of days, and the rest were fully flexible or remote. Among structured hybrid firms, the average requirement was 2.87 days a week and 67% required three days. Larger firms are stricter: companies with more than 25,000 employees were the most likely to require full time in the office, and the share of such firms doing so rose through 2025 as Amazon, AT&T, JPMorgan and several federal agencies moved to five days.

Whether more firms will follow is contested. ResumeBuilder's October 2025 survey of 978 business leaders found 30% said their company would require five days in the office in 2026, up from the share that did so in 2025. The Flex Index data show the five-day share rising slowly rather than jumping, and the SWAA share of days worked from home has not fallen. Published policy and actual attendance are different things, which is where the next section comes in.

Monitoring and enforcement

CBRE's 2025 Americas Office Occupier Sentiment Survey of 184 corporate real estate executives found 69% of companies monitor attendance, mostly through badge data, but only 37% enforce their policy with consequences. The gap between stated and actual attendance was the most common complaint, and the share of firms planning to tie attendance to performance reviews was rising. Our page on whether an employer can track your location covers the badge and Wi-Fi methods used.

The friction shows up in the data as "coffee badging," swiping in for a few hours to register attendance and leaving. Owl Labs' 2025 State of Hybrid Work survey of 2,000 US full-time workers found a majority of hybrid workers admitted to it. The question wording is self-report and the sample is an online panel, so treat it as indicative.

What mandates do to attrition

Two studies measure the effect of mandates on who leaves.

Ding and Ma, University of Pittsburgh, 2024. An analysis of 137 S&P 500 firms that issued return-to-office mandates between 2020 and 2023, using LinkedIn employment histories and Glassdoor reviews. Mandates were followed by no measurable change in firm financial performance or stock value, a significant drop in employee satisfaction with culture and work-life balance, and no improvement in productivity. A follow-up paper with Van Dijcke and others on 54 firms found abnormal turnover rose 13% to 14% after mandates, with departures concentrated among senior, skilled and female employees, and that time to fill vacancies rose by about 23%.

Bloom, Han and Liang, Nature, 2024. A randomized trial at Trip.com of 1,612 engineers, marketing and finance employees, half assigned to hybrid work with two days at home and half to five days in the office. Over six months, quit rates in the hybrid group were 33% lower, with the effect strongest among non-managers, women and long commuters. There was no difference in performance reviews or promotions over two years. The paper also found managers' initial expectation that hybrid would hurt productivity reversed after the trial. Our employee turnover statistics page puts those figures against the national quits data.

Workers say the same thing in surveys. Owl Labs' 2025 survey found 40% of workers would start looking for a new job if flexibility were removed, and in Gallup's data the share of fully remote employees who say they would be extremely likely to look elsewhere if required to be on site five days is higher than for any other group.

Engagement, wellbeing and preference

Gallup's May 2026 indicator gives engagement among remote-capable US employees as 37% for exclusively remote, 34% for hybrid and 29% for fully on site. Wellbeing runs the other way in one respect: remote workers report more daily loneliness and less connection to the mission. Hybrid employees score highest on the combination of engagement and wellbeing, which is why Gallup's recommendation has settled on two to three office days.

Preferences have been stable. In Gallup's November 2025 data, 61% of remote-capable employees preferred hybrid, 33% fully remote and 6% fully on site. Nine in ten want some remote flexibility. The preferred number of office days is two to three. Among the fully remote, the most-cited benefits were no commute, flexibility and better wellbeing; among hybrid workers, the most-cited costs of remote days were weaker collaboration and fewer connections.

The one group that benefits from more office time in the data is the youngest. The Conference Board's 2025 job satisfaction survey found workers aged 18 to 24 reported higher satisfaction when on site two to three days a week than when remote four days or more, a pattern that did not hold for older workers. Our Gen Z workplace statistics page covers that.

What the research says about productivity

The evidence on productivity depends on the arrangement. Fully remote work has produced mixed results: Bloom's 2015 call-center trial at Ctrip found a 13% performance gain at home, while a 2023 study of data-entry workers in Chennai found an 18% productivity penalty for those randomly assigned to work from home, driven partly by who chose remote work in the first place. Hybrid work, in the Trip.com trial, had no effect on performance or promotion in either direction, with the gain coming entirely through retention.

Managers and employees still disagree about it. In Gallup's 2023 data from 18,871 workers, 52% of hybrid employees named productivity as a benefit of hybrid work, against 31% of managers. The Pittsburgh study found no productivity gain after mandates. The honest summary is that hybrid arrangements have not been shown to reduce measured output, mandates have not been shown to increase it, and the measurable effects of mandates are on turnover and satisfaction. Our remote work productivity statistics page covers the studies in more depth.

Reading these figures carefully

Kastle's 55% is ten metros, not the country, and badge data undercount people in buildings without Kastle systems. SWAA is an online panel weighted to the working population, and its "share of paid days" is a different thing from the "share of workers" that BLS reports. Flex Index audits stated policy, not attendance. The two studies that come closest to cause and effect, the Trip.com trial and the Pittsburgh S&P 500 analysis, measure attrition, not output. Pew's remote work figures, which many pages still carry, are from 2023 and 2025 and could not be re-verified for this page. When a return-to-office number appears without its source and measure, it is not telling you much.

Key takeaways

• Office keycard swipes in the ten largest US metros were 55.4% of the early-2020 level in late September 2026, near the 56.3% record set in December 2025.
• About 26% of paid working days are done from home, against 7% in 2019 and 61% in May 2020. The share has been flat since 2023.
• Among remote-capable US employees, 52% are hybrid, 26% fully remote and 22% fully on site. Hybrid employees average 2.6 office days.
• 34% of US firms require full-time attendance and 42% have structured hybrid policies averaging 2.87 days, per Flex Index data on 9,132 companies.
• 69% of companies monitor attendance but only 37% enforce their policy, per CBRE's 2025 occupier survey.
• Mandates at 54 S&P 500 firms were followed by 13% to 14% higher abnormal turnover and 23% longer time to fill roles. A randomized trial found hybrid work cut quits by 33% with no effect on performance.
• Engagement is 37% among exclusively remote employees, 34% hybrid and 29% on site. 61% of remote-capable workers prefer hybrid and 6% prefer fully on site.
• Badge data, foot traffic, household surveys and policy audits measure different things and should not be placed on one trend line.