Every organization owns at least one monument to a departed leader: the values poster from the 2012 offsite, the leadership institute with the founder's name, the policy nobody uses. Monuments are how late careers try to outlive themselves, and they share one property: they require no one to behave differently. That is why they are easy to build, and why they don't work.
The test of what a leader actually left behind is brutally simple: watch what people do, under pressure, when the leader is gone and no rule compels them. Whatever shows up in that moment is the legacy. Everything else is signage.
The paternity leave that nobody took
Semco learned this the empirical way. In the 1980s, as Ricardo Semler has told it, the company gave two weeks of paid paternity leave to all new fathers. Not one took it, out of fear of what the boss would think. The lesson Semler drew became a design rule: a benefit no one dares to use is not a benefit; it is decoration. He points to unlimited PTO as the modern analogue.
The story is usually told as a point about benefits. It is really a point about legacy. The leave policy was, in miniature, a monument: a declared value ("we support fathers"), formally enacted, costing nothing to announce, and behaviorally inert. The organization's actual operating value, "visible sacrifice is what gets rewarded here", was never written anywhere and governed everything. Leaders who spend their final years writing values into documents are building paternity-leave policies. The values that persist are the ones people can be observed using without flinching.
Why behavior is the only durable medium
The research on how workplaces learn gives this a mechanism. Edmondson's foundational 1999 study of 51 work teams found that psychological safety, the shared belief that the team is safe for interpersonal risk, predicted learning behavior, and that learning behavior in turn mediated the link to performance. The chain matters: the belief alone produced nothing; it worked entirely through what people actually did, the questions asked, the errors admitted, the help sought. Culture reaches results only through behavior, and behavior spreads the way the paternity-leave fear spread: by watching what happens to the people who go first.
This is why declared legacies die at the first counterexample. One promotion of a bully outweighs a decade of civility posters, because the organization reads consequences, not documents. And it is why behavioral legacies survive their authors: a team that has practiced admitting errors for ten years keeps doing it under the next leader, not out of loyalty but because the habit has proven cheaper than its alternative, colleague by colleague, incident by incident.
There is even a payoff for the person doing the transmitting. Ghosh and Reio's 2013 meta-analysis found that mentors themselves report higher job satisfaction and organizational commitment than non-mentors. Spending the last years teaching behavior is not a sacrifice made for the institution; it is, on the evidence, one of the more satisfying ways to spend them.
The late-career audit
For a leader in the final stretch, the reframe converts legacy-building from construction into practice, and it suggests an audit with three questions.
What do people do because I demonstrated it? Not endorsed, demonstrated. The behaviors copied from a leader are the ones performed in view, repeatedly, at visible cost: the apology issued upward, the deal declined on principle, the vacation actually taken. A leader who wants a behavior to persist has to be seen paying for it.
What happens to the people who act on my stated values? This is the paternity-leave question. If the stated value is candor and the last three candid people plateaued, the legacy under construction is silence, whatever the plaque will say.
Who can now do this without me? The only legacy metric that compounds. Every judgment call the leader still monopolizes is a behavior that dies at the farewell party; every one transferred, practiced, and survived-in is permanent.
Monuments ask to be remembered. Behavior doesn't need to be; it just continues. The leader who leaves behind a hundred people who ask the hard question in meetings, take the leave, open the books, and teach the next hundred, has built the only structure that maintains itself. Nobody will name it after her. That is how she'll know it worked.
Sources
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams (Administrative Science Quarterly 44, DOI 10.2307/2666999). Field study of 51 teams: psychological safety predicts learning behavior, which mediates the relationship to team performance; beliefs affect results only through enacted behavior. https://journals.sagepub.com/doi/10.2307/2666999
- Ghosh & Reio (2013). Career benefits associated with mentoring for mentors: a meta-analysis (Journal of Vocational Behavior, DOI 10.1016/j.jvb.2013.03.011). Mentors report higher job satisfaction and organizational commitment than non-mentors; teaching pays the teacher. https://doi.org/10.1016/j.jvb.2013.03.011