The office is free to the company and expensive to the worker, and the mandate depends on nobody adding up the second figure. A return-to-office rule does not appear on any budget line as a cost, yet it imposes one: the daily hours in transit, the fare or the fuel, the childcare that has to stretch to cover the commute at both ends. The company books the collaboration it hopes to gain. The employee pays for it in time that is never reimbursed and rarely even named.
What makes this a tax rather than a shared expense is that its incidence is uneven. It falls hardest on the people with the longest trips and the heaviest obligations at home, and lightest on those who live close and have few dependents. A single rule, applied identically to everyone, produces radically different bills depending on a worker's address and household. That is the signature of a tax that is regressive in practice even when it looks uniform on paper.
The evidence names the payers
The randomized hybrid trial at Trip.com is unusually clear about who values relief from the commute, because it broke the retention effect down by group. Overall, two days a week at home cut voluntary resignations by about a third. But the reduction was far larger in specific subgroups: around 40% among non-managers, 54% among women, and 52% among employees with long commutes. The people who most wanted to stay when the commute was cut are the same people a mandate would tax most heavily when it is reimposed.
This is not a story about preference in the abstract. It is a story about a real burden, measured through the door people walk out of. When the commute goes back on, the workers with the longest trips and the most care responsibilities face the steepest increase in unpaid daily cost, and their elevated willingness to leave under those conditions is exactly what the retention data captures. A policy that treats the commute as free is, in effect, choosing which employees to charge the most.
The second tax: fragmented time
There is a subtler levy on top of the literal one. The commute does not only consume the hours it occupies; it fragments the day around them. Concentrated work needs uninterrupted blocks, and Microsoft's own telemetry describes an office day already broken into fragments, with workers interrupted roughly every two minutes and hundreds of interruptions accumulating daily. Bolting a two-hour round trip onto a day that already resists focus does not add collaboration; it subtracts the quiet edges of the day, early morning and evening, where much real work actually got done.
So the worker pays twice: once in transit, and once in the loss of the flexible margins that made the commute survivable. Neither charge shows up in the productivity case for the mandate, because neither is borne by the party making the decision. The asymmetry is the whole mechanism. The company decides; the employee pays; the ledger where the payment would appear does not exist.
Pricing the invisible
The fix is not necessarily to abolish the office. It is to stop pretending the commute is costless and to price it into the decision. A firm that genuinely needs people in a room for certain work can ask for that specific presence and acknowledge what it is asking them to spend. A firm that imposes a blanket mandate while claiming neutrality is running a transfer it refuses to record.
The cleaner test is to ask who bears the cost of every option and to count it honestly. Hybrid arrangements survive that test in the data because they leave the heaviest-taxed workers a way to stay. The blanket mandate fails it, not because presence is worthless, but because it collects its bill from the people least able to afford it and books the revenue as culture. A cost that large should at least be visible to the person deciding to impose it.
Sources
- Bloom, Han & Liang (2024). Hybrid working from home improves retention without damaging performance (Nature). RCT: two-day hybrid cut resignations ~a third overall, with larger effects among non-managers (40%), women (54%), and long-commute employees (52%). https://www.nature.com/articles/s41586-024-07500-2
- Microsoft WorkLab (2025). Work Trend Index: The Year the Frontier Firm Is Born. Microsoft 365 telemetry: workers interrupted roughly every 2 minutes, ~275 interruptions/day. https://www.microsoft.com/en-us/worklab/work-trend-index/2025-the-year-the-frontier-firm-is-born