Nobody Actually Works From Home

Thirty-six per cent of employed Australians usually work from home. That figure gets quoted constantly, and nearly everyone reads it wrong.

It comes from the ABS Working Arrangements survey, published in December 2025 and covering August of that year. It's the most recent proper national read we have, and the detail underneath the headline tells a very different story from the headline itself.

Start with scale. Australia has roughly 14.4 million people at work, out of a population of about 27 million. Of those 14.4 million, only about 356,000 work from home five days a week or more. That's 2.5 per cent of everyone who holds a job.

The far more common arrangement is one or two days a week, covering about 11.7 per cent of employed people. Another 6.6 per cent do three or four days.

Five years on from the thing that supposedly changed everything, the typical experience of working from home is a Tuesday and a Thursday.

Not quite the arrangement it was sold as

The assumption underneath most of the commentary is that people work from home to claw time back. The same survey points somewhere less comfortable.

Forty-two per cent of people who worked from home also worked overtime, against 26 per cent of those who didn't. And when people were asked their main reason for working from home at all, the second most common answer, given by about one in five, was to catch up on work after hours.

So, a decent chunk of working from home isn't replacing time in the office. It's extra work happening at the kitchen table after dinner.

Worth saying the conversation has already moved on. Plenty of the discussion now is about four day weeks and compressed hours, a genuinely separate debate with its own evidence and its own problems. We're busy arguing about the next thing while the current one is still widely misunderstood.

Most people were never in the running

Productivity Commission modelling put the share of Australian jobs ever capable of being done from home at around 35 per cent. It also found that even if every person who could work from home did so on a hybrid basis, only about 13 per cent of all hours worked in the economy would happen remotely.

Access lines up almost exactly with the advantages people already had. Melbourne Institute research using HILDA data describes the likelihood of working from home in manual and trade occupations as negligible. Among casual workers it's 4.3 per cent. Back in the ABS survey, 39.5 per cent of workers with paid leave entitlements usually work from home, against 10.7 per cent of those without.

Split it by occupation and it's starker again. In August 2025, 59 per cent of managers and professionals usually worked from home. For every other occupation group, 21 per cent.

It even shows up geographically. Canberra sits at 47.4 per cent, the highest in the country. New South Wales 38.5, Victoria 37.8, Queensland 34.5, Western Australia 31.3, the Northern Territory 22.7. That spread isn't about culture or attitude. It tracks what each place does for a living.

Working from home didn't change how Australia works. It changed how one identifiable slice of Australia works, and that slice already had the degrees, the salaries and the job security.

The clearest place to see the split

Most industries sit cleanly on one side of that line. Engineering sits on both, which makes it a useful place to watch the gap open in real time.

Indeed publishes a tracker of how many Australian job ads mention remote or hybrid arrangements, broken down by occupational category and running back to 2019. In February 2020, 6.2 per cent of civil engineering ads did. By August 2026 it was 31.9 per cent. That puts civil engineering sixth out of 44 categories, ahead of law, finance and marketing, at roughly 2.4 times the national average of 13.1 per cent.

Site-based construction went from 1.2 per cent to 6.0 per cent across the same period.

Same qualification. Same profession. Often the same employer. One end now advertises flexibility at five times the rate of the other.

Nobody engineered that outcome. Offices shut in 2020 and a chunk of that work never needed to come back. Site work ran the whole way through, because you can't pour concrete from the spare room and a survey crew can't attend over Teams. It was a by-product of what was physically possible, and five years on it's set.

Every industry with a physical component has a version of this line running through it. Engineering just happens to have it running through a single qualification.

It has a price, and people are paying it

CEDA ran the numbers in 2025 using HILDA data, which follows the same households year after year. Since 2020, Australians working hybrid or fully remote earn 5.8 per cent less than otherwise comparable people who don't work from home. About $4,400 a year.

They're not being dudded without noticing. The same analysis valued the time handed back at roughly $5,300 a year, through about 3.4 fewer hours commuting each week. People are trading salary for hours and, on those numbers, coming out in front.

American research by Barrero, Bloom and Davis arrives at a similar place from the other direction, valuing the option of two or three days at home at around 8 per cent of pay, and rising with education and earnings.

Call it 6 to 8 per cent of salary. Which means the two thirds of the workforce who were never eligible aren't just missing a convenience. They're missing something the market has put a number on.

Does it actually work

Fair question, and the Australian evidence is more settled than the volume of argument suggests.

The Productivity Commission examined this while looking into the country's productivity slump. Its conclusion was that the increased propensity to work from home did not appear to have caused significant changes to labour productivity. Most of the studies it reviewed land hybrid work somewhere between neutral and mildly positive. Fully remote is the weaker case, particularly for collaborative work.

Which is worth holding against the opening number. The arrangement the evidence is least comfortable with, five days a week at home, is the one almost nobody actually has.

The boardroom hasn't entirely caught up. A survey of 1,127 Australian company directors in early 2025 found around 60 per cent believed flexible work had a positive effect on recruitment, retention and staff wellbeing. About 40 per cent thought it was a negative for innovation.

So, the people running companies broadly accept it helps them attract and keep staff. The hesitation sits with what it does to the work itself, which is the part the evidence supports least.

Having it and using it are different things

Here's where it gets awkward for employers who reckon they've already solved this.

Flexible arrangements have been formally available across most of professional Australia for years. The large engineering design consultancies reported through Consult Australia in 2021-22 that around 91 per cent of their staff felt they had the flexibility they needed. Go back further and Engineers Australia was reporting in 2012 that flexible arrangements were broadly available while take-up stayed very low, with that reluctance evident since before a 2007 survey.

That's the better part of two decades of a policy sitting there and not being used much. Which suggests the constraint was never the policy.

Availability and usage are two different measurements. Most organisations only ever track the first one.

The rules are starting to catch up

In August 2026 a Fair Work Commission Full Bench decided to put a working from home provision into the Clerks Private Sector Award. First time a modern award has carried one. It doesn't create a blanket entitlement; it sets up an express right to request and a written agreement process. Submissions on the draft closed in September and it isn't in force yet.

Separately, since June 2023 the right to request flexible working under the Fair Work Act has had actual teeth. The Commission can conciliate, arbitrate, and order an employer to grant a request. Employers have lost cases on exactly that.

Award terms have a long history of spreading past the award that introduced them. So, the thing worth watching isn't whether working from home survives. It's that a benefit currently held by about a third of workers, concentrated among the best paid, is starting to harden into something enforceable.

Which raises the obvious question about the other two thirds, and nothing currently on the table answers it.

What the number actually says

Working from home in Australia is smaller than the noise around it suggests, far more unevenly spread than most people assume, and not especially restful.

It reaches about a third of workers, mostly for a day or two a week. It clusters among people with degrees, secure jobs and professional titles. It comes with more overtime, not less. Roughly two thirds of the workforce could never get near it. And it's worth somewhere between 6 and 8 per cent of salary to the people who have it.

The figure everyone quotes gets treated as a measure of how Australia works now. It isn't. It's a measure of who got to choose.