A 7,704-person study says remote staff feel best. It never measured what RTO mandates claim to buy.

The turnover finding travels through two links rather than one, and the study's own language admits it. Researchers surveyed well-being, came back a year later, and found that employees who had scored higher were less likely to have left ^[6]. Fully remote staff had scored highest ^[3]. Remote work is therefore tied to retention by way of well-being, not by a direct count of who quit out of which arrangement ^[19]. That joint is the first thing a finance team will lean on.

Scale cuts the same way. Roughly a quarter of the 7,704 worked fully remotely and about half entirely onsite ^[2], which sets something near 1,900 remote employees against about 3,850 onsite ones ^[18]. Large, and still one employer: an academic cancer center in Texas ^[1]. Jamie Dimon's five-day recall at JPMorgan covers more than 300,000 people on its own ^[10], about 39 times the study population ^[21].

The hole in the evidence is exactly where the mandates are argued. Johnson and Holladay did not track productivity ^[7]. What they did measure was well-being, connection and turnover, and on connection they found little sign that remote workers felt further from colleagues or workplace culture ^[5]. So the paper answers the collaboration half of the case Dimon and Musk have made and leaves the output half untouched ^[20]. The authors keep their own claim that narrow: the findings "challenge the idea that simply bringing people back into a building will automatically make them more engaged, connected, or likely to stay," the two told Fortune ^[8].

Then there is the motive argument, which is weaker than it sounds. A survey last year found one in five HR professionals saying their employer's in-office policy was meant to encourage people to quit ^[13], and Johnson said "clearly we have to question the motive," pointing to separate research that leaders scoring higher on narcissistic qualities are more likely to dislike remote work ^[14]. The first is admitted intent from the people writing the policies. The second is a disposition, and it says nothing about whether a given mandate produces anything. They do not carry equal weight.

The part an operator can actually use is Holladay's description of the office day: people travel in and spend it on Zoom and Teams calls, and she notes that less office space also means less rent ^[15]. That comparison is measurable at any employer, which the well-being ranking from a single cancer center is not. The authors' prescription follows from it, and it is a design brief rather than an attendance count: decide what presence is for, whether that is collaboration, mentoring, innovation, relationship-building or culture, then build the experience that produces it ^[9]. Johnson adds that flexibility may matter more than any particular arrangement, especially for younger workers still forming relationships ^[16], and that if people are asked to come in, the trip should be worth something to them ^[17].