The median American worker has now been with their employer 3.9 years, the lowest since 2002 (U.S. Bureau of Labor Statistics, January 2024). Workers aged 55 to 64 hold median tenure of 9.6 years; those 25 to 34, 2.7. The usual reading is generational decline: the old stayed, the young hop. The number is real; the reading gets the causality backwards.
Tenure was never loyalty. It was a financial instrument. Under the old deal, years-at-firm was an appreciating asset: it purchased seniority protections, a defined-benefit pension that vested with time, internal promotion ladders that outsiders could not board, and an implicit severance of dignity. Staying was not devotion; it was compounding. When employers dismantled those payoffs (pensions to portable accounts, ladders to flat structures, protection to at-will restructuring), they converted tenure into a non-earning asset. Workers responded the way holders of any non-earning asset respond: they moved into liquidity, holding skills, reputation and market options instead of years.
Who exited first
The record on sequencing matters, because the moral story depends on it. Defined-benefit pensions were being closed and frozen across the private sector decades before anyone accused Gen Z of disloyalty. Layoffs became a routine tool of ordinary management, not an emergency measure, while the language of family persisted in onboarding decks. The party that stopped paying for tenure was the party that had priced it. What looks like a generational abandonment of loyalty is a lagging indicator of an employer abandonment that came first, and the lag was exactly one generation: the time it took workers to watch their parents' loyalty get marked to zero.
Even now, the tenure that survives clusters where the old payoffs survive: public-sector workers, who retain something like the old deal, show median tenure of 6.2 years against 3.5 in the private sector. Loyalty did not die. Its price is simply still posted in only a few places.
Liquidity has its own logic, and its own costs
Once tenure is illiquid and unpaid, the rational career becomes a portfolio: acquire transferable skills, mark them to market regularly (often by moving, since external offers reprice faster than internal reviews), and never hold a concentrated position in one employer's promises. Employers experience the result as churn and mercenary behavior. It is the exact discipline they taught, applied back.
But honest accounting includes what liquidity costs. Firms lose institutional memory and the deep product knowledge that only accrues in years; workers lose the compounding relationships and slack that made some workplaces humane; and both lose the ability to make investments that pay off past the expected departure date. The evidence on forced returns to the office shows how carelessly firms now spend the tenure they still have: analysis of 260 million résumés found that strict return-to-office mandates at major tech firms triggered disproportionate exits of senior employees, the longest-tenured people, toward direct competitors (Ding and Ma, 2024). Firms holding the scarcest asset in the modern labor market, veterans who stay, liquidated it for attendance.
The third option Semco built
The loyalty-versus-liquidity frame hides a third design, which Ricardo Semler describes from Semco's satellite program. Employees who wanted out could take their equipment, become independent suppliers, and sell back to Semco and to its competitors; half the company eventually went, as he tells it, across a bridge built plank by plank with the company holding the other end. The insight is that the exit and the relationship are separable. A departure can end the employment and keep the network, the trust, the flow of work. Tenure inside the walls ended; the compounding continued outside them.
Modern equivalents are cheap to build and rarely built: alumni networks treated as supplier and customer pipelines rather than nostalgia lists, boomerang hiring as policy rather than exception, contracts that let a departing engineer keep maintaining the system as a vendor. Each converts departure from default into refinancing.
The generational contract has already been rewritten on this clause; the only question is whether employers read the new text. Workers will hold what pays. If years at your firm pay, in money, protection, or genuinely compounding roles, people will hold years again; the public sector proves the demand exists. If years do not pay, workers will hold options, and no amount of loyalty rhetoric will move the portfolio. Tenure lost its meaning when it lost its yield. Restore the yield, or stop mourning the meaning.
Sources
- U.S. Bureau of Labor Statistics (2024). Employee Tenure in 2024. Median tenure 3.9 years (lowest since 2002); 55-64: 9.6 years vs 25-34: 2.7; public sector 6.2 vs private 3.5. https://www.bls.gov/news.release/archives/tenure_09262024.htm
- Ding & Ma (2024). Return-to-Office Mandates and Brain Drain. Distributional synthetic controls over 260M résumés: RTO mandates at Microsoft, SpaceX and Apple caused disproportionate senior-employee departures, scaling with mandate strictness; working paper, not peer-reviewed. https://arxiv.org/pdf/2405.04352