In the 1980s, Semco gave every new father two weeks of paid paternity leave. As Ricardo Semler has told it, not one father took it. The policy was real, the money was real, and the fear was more real than both: what would the boss think? Out of that came the rule he has repeated since, that a benefit nobody dares to use is not a benefit, it is decoration.
Four decades later the decoration industry has scaled magnificently. Unlimited PTO that produces fewer vacation days than the fixed allowance it replaced. Mental-health days announced in all-hands and punished in performance calibration. Parental leave printed in the handbook at sixteen weeks and taken, by anyone with ambitions, at four. Sabbaticals that exist on the careers page and in no living memory. Each one photographs well in recruiting material, costs little in practice (the fear does the cost containment), and generates its own private tax: employees now perform gratitude for compensation they cannot safely collect.
The mechanism is the manager's eyebrow
What keeps a written entitlement unused is almost never a written rule. It is inference from evidence: who took the leave and was quietly moved off the big account; whose promotion arrived the cycle after they skipped their vacation; what the director said, jokingly, about a colleague who logged off at five. Employees are superb statisticians of their own environment. They do not read the handbook; they read outcomes.
This is why usage, not generosity, is the only honest metric of a benefits program. A company that wants to know what it actually offers should publish internally, by level and by team, the percentage of entitled leave actually taken, parental leave taken in full by men and by women, vacation days used versus accrued. Where usage is low, the policy is decoration and the decoration has an author: some manager whose eyebrow is doing what the policy pretends not to.
Semco's answer to the eyebrow was structural, not rhetorical. Vacations became compulsory: thirty days, out of the office, laptop locked away. Semler's reasoning was that adults negotiating to keep working through their vacation were not displaying loyalty but fear, and that fear is the cheapest management tool ever invented. A right that can be waived under pressure will be waived under pressure; the only way to make it real was to remove the option of surrendering it. The company took the eyebrow out of the transaction.
Why this is a generational-contract clause
The decoration problem sits at the center of the renegotiated employment deal because benefits are where the written and unwritten contracts visibly diverge. The written contract says sixteen weeks; the unwritten one says four, and everyone under thirty has watched the difference get enforced on someone they know. A generation raised on receipts responds predictably: they discount the stated benefits to their observed usage value, and they price the gap as dishonesty.
The discount shows up in negotiation behavior that employers misread as cynicism. Young workers increasingly prefer terms that cannot be quietly clawed back by culture: salary over bonus potential, fixed vacation over unlimited, a four-day week over flexibility theater, written remote terms over a manager's mood. This is not entitlement. It is the correct pricing of promises by a cohort that has seen how unenforceable ones perform.
There is decent evidence about what happens when time off is structural rather than performative. A six-month trial of the four-day week with no pay cut, covering 2,896 employees across 141 organizations in six countries, found reduced burnout and improved job satisfaction, mental health and physical health, with larger individual reductions in hours producing larger gains (Fan, Schor and colleagues, Nature Human Behaviour, 2025). The design detail matters: the day off was not a favor to be requested under an eyebrow. It was the schedule. Nobody has to dare to use Friday.
The audit
Any organization can run Semco's test on itself this quarter. Pick the three benefits most prominent in your recruiting. Pull actual usage. For each, ask: would an ambitious 28-year-old on the fast track use this in full, this year, without a private calculation about what it costs them? If the answer is no, the benefit does not exist; a fear does, wearing the benefit's name. Then do the only thing that has ever worked: change the default so the entitlement executes itself, and let the first person to use it in full be someone senior, visibly, twice. Policy is what happens to the person who believes the handbook. Everything else is decoration.
Sources
- Fan, Schor, Kelly & Gu (2025). Work time reduction via a four-day workweek: effects on workers' well-being. Nature Human Behaviour. Six-month trial, 2,896 employees in 141 organizations across 6 countries: reduced burnout, improved satisfaction and mental/physical health, with dose-response by individual hours reduced; quasi-experimental, not an RCT. https://www.nature.com/articles/s41562-025-02259-6
- American Psychological Association (2023). 2023 Work in America Survey. Monitored/scrutinized work environments correlate with higher reported tension and stress (56% vs ~40%), consistent with fear, not policy text, governing usage of workplace entitlements. https://www.apa.org/pubs/reports/work-in-america/2023-work-america-ai-monitoring