The fight over return to office has stopped moving in one direction. Workers still push for choice. Some executives still push for face time. Yet the momentum behind sweeping mandates has faded, and the costs of mistaking presence for power are getting clearer. In a Stanford and Atlanta Fed survey of more than a thousand US executives in February 2025, only 12 percent with hybrid or fully remote staff planned a new return-to-office mandate in the next year. Even if enacted, those plans would shave less than half a percentage point off the share of paid days worked from home, from 21.2 percent to 20.8 percent (Bloom, Davis and Barrero, 2025). Workers are not meekly lining up either. In the December 2024 SWAA survey, just 44 percent said they would comply with a fully on-site policy. The rest said they would quit or look elsewhere (Bloom, Davis and Barrero, 2025).
Presence as a brittle power move
Treating office attendance as a proxy for control carries a talent risk, and it is not hypothetical. A 2024 working paper by Ding and Ma tracks career histories across 260 million resumes and uses distributional synthetic controls to isolate the effect of return-to-office mandates at Microsoft, SpaceX, and Apple. They find a disproportionate exit of senior employees after mandates, with a dose–response pattern that scales with rigidity: Apple’s one day a week mandate showed a seniority shift below 4 percent, Microsoft’s roughly 50 percent presence mandate showed more than 4 percent, and SpaceX’s full-time presence requirement was higher. At Microsoft, average tenure in the top deciles dropped by about two months after its April 2022 mandate, and the share of employees with 0 to 3 years at the firm rose by 6 percent. The authors flag this as a threat to productivity and innovation. This is a working paper and not yet peer reviewed, but the design is unusually strong for a policy question that is hard to randomize (Ding and Ma, 2024).
The message is consistent with the survey evidence on compliance. Workers with options do not treat physical presence as an end in itself. They vote with their feet. When only 12 percent of executives plan to tighten the screws and nearly half of workers say they would not comply with a full return, mandating presence looks less like leadership and more like a bet against the labor market’s memory (Bloom, Davis and Barrero, 2025).
What hierarchy confuses about control
Classic management equates supervision with performance. That model works best when tasks are known and environments are stable. It falters when work is interdependent, cross-functional, and novel. Lee and Edmondson synthesize decades of research and define self-managing organizations as those that decentralize authority formally and systematically, eliminating the manager–subordinate relationship as the basic unit of control (Lee and Edmondson, 2017). Their review argues that managerial hierarchy is good at reliable execution of established tasks, but it tends toward rigidity and status protection. Those features inhibit complex problem solving across boundaries and can stifle people’s development. The office, historically, amplifies those status differences. The most visible person often gets the floor.
Learning requires something different. Edmondson’s seminal field study of 51 manufacturing teams defines team psychological safety as a shared belief that the team is safe for interpersonal risk taking. In those teams, psychological safety was associated with learning behaviors, and learning mediated the link between safety and performance (Edmondson, 1999). Visibility did not drive performance. The social conditions for speaking up, asking for help, and surfacing errors did. That is the opposite of presence as power. If the office is used to reinforce rank and silence dissent, it subtracts from the very conditions that make teams smarter.
The broader engagement picture reinforces the point. Gallup’s 2025 report finds global employee engagement at 20 percent, with an estimated cost of about 10 trillion dollars in lost productivity, or 9 percent of global GDP. Manager engagement has fallen nine points since 2022, and the largest annual drop was between 2024 and 2025, from 27 to 22 percent. In best-practice organizations, by contrast, 79 percent of managers are engaged, which suggests design choices, not fate, drive the gap (Gallup, 2025). If firms turn middle managers into attendance police, they should expect more disengagement, not less performance.
Redesigning the office to serve power, not symbolize it
The office still matters. It is a tool, not a talisman. The aim is to decouple voice, decision rights, and recognition from proximity. That calls for three design moves.
First, make presence purposeful. Treat the office as a commons for activities that gain from co-location, such as conflict-heavy decisions, onboarding, and deep design work. Use clear team agreements on when and why to meet in person. The survey evidence shows that blanket mandates are unlikely to move the national needle on remote days and risk noncompliance. Targeted, transparent norms are more credible with adults who have choices (Bloom, Davis and Barrero, 2025). This is a design recommendation, not a randomized result.
Second, distribute authority where the information lives. Lee and Edmondson describe organizations that eliminate the routine boss–subordinate tie and formalize decentralized control. You do not have to go all the way to adopt the logic. Push decisions to the edges, make decision rights explicit, and publish who decides what. Presence then supports coordination instead of signaling rank (Lee and Edmondson, 2017).
Third, measure the conditions for learning, not just the count of badges swiped. Psychological safety predicts learning behaviors, which support performance. Managers can run lightweight pulses on whether people feel safe to challenge a plan or admit a mistake, and leaders can role model fallibility in office meetings to lower the status barrier. This is an application of Edmondson’s construct to a hybrid reality, not a claim about causal effects of office design in her 1999 study (Edmondson, 1999).
Practical guardrails help. Do not tie promotion odds to hours on site. Publish meeting notes so remote colleagues are not second-class. Rotate facilitation and who speaks first so status does not track seating charts. When mandates are considered, stress test the likely response among senior talent. The Ding and Ma results imply that stricter mandates correlate with larger seniority losses, particularly where the market offers outside options, although the evidence is from a working paper and specific firms (Ding and Ma, 2024).
The power shift is already here. Executives are not broadly tightening presence rules, workers have learned to exit bad deals, and engagement data warn that using the office to reassert top-down control will backfire. Presence can still create value, but only when it serves distributed authority and genuine voice. The practical implication is simple. Redesign who holds power and how decisions get made, then decide where people need to be. Get that order wrong, and the office becomes a very expensive symbol.
Sources
- Bloom, Davis and Barrero (2025). Stanford Report — Return to office? Not everybody is doing it. Shows that only 12 percent of executives planned new mandates and that even full compliance would barely dent remote days, with most workers unwilling to accept full RTO. https://news.stanford.edu/stories/2025/03/return-to-office-not-everybody-is-doing-it
- Ding and Ma (2024). Return-to-Office Mandates and Brain Drain. Finds disproportionate senior exits after mandates at Microsoft, SpaceX, and Apple, scaling with rigidity, using distributional synthetic controls on 260 million resumes; working paper caveat. https://arxiv.org/pdf/2405.04352
- Lee and Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Defines self-managing organizations and argues that managerial hierarchy favors reliability but inhibits complex cross-boundary problem solving, reinforcing status. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams. Introduces team psychological safety and shows its association with learning behaviors that mediate performance in a field study. https://journals.sagepub.com/doi/10.2307/2666999
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20 percent with a 10 trillion dollar productivity cost, and a sharp decline in manager engagement, underscoring design challenges in control and presence. https://www.gallup.com/workplace/659610/state-global-workplace.aspx