The defining legal fact of free labor is not the wage; it is the exit. A worker who may leave, and take her skills to whoever values them most, is bargaining even when she says nothing: her alternatives negotiate for her. The non-compete clause operates directly on that mechanism. It does not protect any specific secret (other instruments do that); it restricts where a person may work next, converting departure from a right into a breach. Medieval labor law had a name for the underlying idea: the worker belongs, for a time, to the estate.
The scale is not marginal. When the U.S. Federal Trade Commission issued its rule banning most non-competes in April 2024, it estimated that about 30 million American workers, nearly one in five, were bound by one (FTC, 2024). The rule never took effect: a federal court in Texas blocked it in August 2024, holding the agency had exceeded its authority, and enforcement returned to state law, a patchwork running from California's long-standing refusal to enforce non-competes to broad enforceability elsewhere (NPR, 2024). The interesting evidence sits inside that patchwork: the most inventive labor market in the world, Silicon Valley, grew up in the one large state where the clause is unenforceable, where engineers have always been free to walk across the street. Whatever non-competes protect, California's history suggests it is not innovation.
What the clause actually purchases
Employers defend non-competes as protection for trade secrets and training investments. But the law already provides narrower tools for both: confidentiality agreements, trade-secret statutes, repayment clauses for funded training. What the non-compete adds beyond these is precisely its breadth: it encumbers the employee's general human capital, the skills that are hers, by restricting the market in which she may sell them. The practical effect is a discount on her bargaining power while employed. A worker who cannot credibly leave for a competitor negotiates her raise against a weaker alternative, which is why the clause is valuable to employers even when it is never litigated, and why it spread far beyond engineers with secrets to workers with none.
Seen from the feudal angle, the clause reverses the modern settlement. The firm's claim runs not against a rival (who signed nothing) but against the worker's future self, in territory and time where the employment relationship no longer exists. It is jurisdiction extended beyond the walls.
The employer who funded the exits
The instructive counter-example is an employer who treated departures as something to finance rather than forbid. As Ricardo Semler has told it, Semco's satellite program helped employees who wanted to leave do exactly that: they took equipment with them, became independent suppliers, and sold their services back to Semco, and to Semco's competitors. About half the company eventually went. The bridge out was built plank by plank, with the company holding the other end.
Notice what this design assumes. It assumes the relationship survives the employment; that a former employee serving your competitor is not a betrayal but a market; that people who know the exit is real, funded and honorable will deal with the firm as partners rather than dependents. The non-compete assumes the opposite on every point: the relationship is containment, the departing worker is a threat, and the way to keep value is to fence it. One design produces alumni and suppliers. The other produces, at best, employees who stay because the fence is expensive to climb, which is the least valuable form of staying there is.
Exit is the guarantor of everything else
There is a reason the right to leave deserves its central place in any feudal reading of tech employment. Every other protection an employee has, from bargaining over pay to refusing bad instructions, is ultimately backed by the credibility of departure. Vesting schedules tax exit; all-inclusive campuses raise its lifestyle cost; non-competes attack it legally. Each device is defensible in isolation; stacked together they describe a program: make leaving expensive, and staying will take care of itself.
The FTC's blocked rule means the legal question will be settled slowly, state by state, employer by employer. The cultural question can be settled faster. A worker reading a contract can ask one question that organizes all the fine print: does this document treat my departure as a right to be respected or a risk to be priced? And an employer confident in its own value proposition can make the strongest recruiting statement available in a suspicious age, the one Semco made with equipment and purchase orders: the door is open, and we will help you carry your things. People stay longer in places they are visibly free to leave.
Sources
- Federal Trade Commission (2024). FTC Announces Rule Banning Noncompetes. The agency estimated ~30 million U.S. workers, nearly one in five, are bound by non-compete clauses. https://www.ftc.gov/news-events/news/press-releases/2024/04/ftc-announces-rule-banning-noncompetes
- NPR (2024). Federal judge tosses U.S. ban on noncompetes. A Texas federal court blocked the FTC rule in August 2024, leaving enforcement to state law, from California's ban to broad enforceability elsewhere. https://www.npr.org/2024/08/21/g-s1-18376/federal-judge-tosses-ftc-noncompetes-ban
- Semler (1989). Managing Without Managers (Harvard Business Review). Primary account of Semco's participative model, context for the satellite supplier program. https://hbr.org/1989/09/managing-without-managers