This week
Governments Keep Counting Who Shows Up
The federal government's return to office just got its first report card, and attendance is the only grade on it.
"Only 12 percent of the Federal workforce teleworked during the reporting period," Office of Personnel Management director Scott Kupor wrote at the top of this month's telework report to Congress, "a 73 percent decrease from FY 2024." That report covers a single pay period in September, while past reports covered the full fiscal year. But OPM didn't just narrow the reporting period. It also stopped measuring just about everything the policy might affect. The fiscal 2023 edition tracked eight kinds of outcomes, from recruitment and retention to commuter miles and real estate, and credited telework with "a total cost savings of more than $230 million." This year's gave agencies four boxes to tick, and the only outcome it names is emergency preparedness. Asked about productivity, an OPM spokesperson told Federal News Network that measuring performance is the agencies' job.
Washington isn't alone, and this isn't a partisan habit. Gavin Newsom has had about 108,000 California state workers in the office four days a week since July, after the state auditor estimated that three days a week of telework could save as much as $225 million a year in office space. Tim Walz requires most Minnesota state employees in person at least half the time so "collaboration can happen more quickly." Canada's hybrid federal public servants went to four days in July, British Columbia's incoming premier says his "preference" is to bring public servants back, and New South Wales has required half-time attendance since 2024. The reasons rhyme: collaboration, culture, accountability, service. Researchers who mapped Canada's policies found that the announcements "do not always explain why governments have chosen one approach over another or whether they have considered the wider effects."
One government that checked changed its mind. Texas Governor Greg Abbott banned remote work for state employees in 2025, then walked it back after a multi-agency survey found telework didn't curb productivity and helped reduce turnover. Victoria, in Australia, tried to go the other way: a bill giving workers a legal right to work from home two days a week passed the lower house, then lapsed this month when Parliament's term ended before the upper house voted.
Washington, meanwhile, stopped asking. On October 29, OPM published a final rule ending the Federal Employee Viewpoint Survey, the governmentwide survey it had run for more than 20 years, after cancelling the 2025 edition outright. Agencies will run their own surveys around ten required questions, and the list is worth reading. Workload, job satisfaction and whether employees would recommend their agency are no longer required. Two of the ten are now about poor performers: whether "steps are taken to deal with a poor performer who cannot or will not improve," and whether the poor performers in your unit stay, improve, quit or get "removed or transferred."
A survey is a statement of what an employer wants to know. The questions that were cut asked how the job was treating the people doing it. The ones that were added are noticeably more cynical.
That old survey was also how outsiders could see what a policy like this did. When the Government Accountability Office warned in January that the Social Security Administration was at risk of "losing many staff in the near term," its evidence included the 2024 survey: almost half of the Social Security respondents who planned to leave within a year said their work unit's telework or remote-work options influenced that decision. None of the ten new questions asks about telework or plans to leave.
I don't think that's an accident. This is the country's largest employer ordering its people back to the office. The results could meaningfully shape how other employers approach workplace flexibility, but of course... not if you conveniently stop collecting them.
The remote years weren't measured well either. GAO found Social Security had not evaluated its telework program to understand its effects, and when California's auditor asked departments whether telework had changed their productivity, 29 said they didn't know. That answer deserves more attention than another photograph of a full office. Kupor himself wrote that if the government could "credibly demonstrate" it was "a well-oiled, maximally efficient provider of services," then "we might have a better leg to stand on in terms of advocating for a different policy." That's a fair standard. But the office should have to meet that burden of proof too.
Attendance is the easiest number any employer can move: mandate it and it rises by the next pay period. Whether the work got better is slower and messier to find out, but it's also the only question that really matters.
What an employer chooses to count tells you what it's trying to prove. When the only number left is attendance, the proof on offer is obedience.