Somewhere around month fourteen, you get the offer. Better title, twenty percent more pay, fifteen minutes closer to home. You mention it to your manager, not as a threat but as a conversation, and then someone from HR sends you the document you signed on your first day. You read the clause you skimmed back then: you may not work for a competitor, in your field, within a defined geography, for two years after leaving. The other company's lawyer advises them to walk. You stay. Your wage stays too.
That sequence, repeated across millions of workers in industries from fast food to software, is how non-compete agreements suppress wages. Not through conspiracy. Not through any conduct that classical antitrust law was designed to catch, but through a structural feature of the labor market that operates quietly, contract by contract.
The mechanism nobody drew on the whiteboard
The core logic of wage competition is simple: if workers can move to whoever pays most, employers must bid against each other. Non-competes break that bidding. They don't set wages directly, which is why they escaped early antitrust scrutiny. They work by narrowing the realistic pool of employers a worker can credibly threaten to join.
Consider a mid-career software engineer, call her Priya, who specializes in financial compliance tooling. Her skills are genuinely portable, but the companies that would value them most are, almost by definition, her current employer's competitors. A non-compete covering that vertical for eighteen months doesn't just prevent her from leaving; it prevents the offer from being made in the first place. The competitor's recruiter, knowing the clause exists, moves to someone else. Priya's employer never has to raise her salary to keep her. The market signal that would have forced a pay increase never arrives.
Economists call this monopsony: a condition where a single buyer, or a small group of buyers acting in parallel, faces workers with nowhere else to go. Non-competes manufacture monopsony artificially, and that matters enormously. Antitrust law built around Sherman Act concepts was designed to catch coordination, the smoke-filled room, the price-fixing ring. Parallel behavior that emerges from identical contract templates, spread by the same handful of law firms across an entire industry, looks nothing like that. It produces the same wage-suppressing result. The law, searching for a conspiracy, finds only paperwork.
Why competition law kept missing it
Antitrust enforcement, historically, asked a specific question: is there an agreement between competitors that harms consumers? Labor markets sat awkwardly in that frame. Workers are not consumers in the conventional sense, and the harm they suffer doesn't show up as higher prices on a shelf. It shows up as a slower wage trajectory over a decade, which is nearly invisible in any individual case and essentially undetectable to a regulator scanning for conduct.
The courts reinforced this blind spot by treating non-competes as a question of contract law, not competition law. The relevant doctrine, rooted in English common law going back to the Mitchel v Reynolds case of 1711, asked whether a restraint was reasonable in scope and duration. That is a narrow, employer-friendly inquiry. It does not ask what happens to wages across an industry when forty percent of workers in a sector are bound by similar clauses simultaneously, any more than a coroner's report on one patient explains an epidemic.
Academic economists eventually started building the empirical case that competition regulators had been asking the wrong question entirely. Research drawing on state-level variation, particularly comparing outcomes in California, which has long refused to enforce non-competes, against states with permissive enforcement, found persistent wage gaps that tracked enforcement regimes rather than industry mix or educational attainment. The gaps weren't enormous in any single year. They compounded. A worker in a non-compete-heavy state who changed jobs four times in a decade might accumulate wage growth measurably below an otherwise identical peer who simply had the freedom to move.
The mechanism also worked on workers who were never personally covered. If your colleagues are locked in, the employer's need to compete for your loyalty is reduced. The suppression radiates outward, silently, to people who never signed a word of it.
The low-wage puzzle that this explains
For a long time, the presence of non-competes in low-wage jobs looked like an anomaly. Why would a sandwich chain bind a part-time worker with a clause preventing them from working at another sandwich chain? The worker has no trade secrets. There is no proprietary formula at stake.
The answer, once stated plainly, is almost embarrassingly obvious. The clause isn't protecting information. It's reducing turnover costs and holding down the wage floor. If workers at one chain cannot freely take a job at the chain across the street, the chains have no need to outbid each other on hourly pay. The non-compete functions as a soft wage ceiling, imposed without any communication between the employers at all. Jimmy John's, the sandwich franchise, drew significant attention when its non-compete language surfaced publicly, covering workers earning close to minimum wage and prohibiting them from working at any business that derived more than ten percent of revenue from sandwiches within three miles. It became a case study precisely because it illustrated so plainly that these clauses had migrated far from their original justification in protecting legitimate business interests.
That migration is not a technicality. It is the whole problem. Courts applying the reasonableness test were poorly positioned to weigh aggregate wage effects against individual employer interests. A judge looking at one contract can assess whether a two-year clause is excessive for a specific role. No judge in a contract dispute has the tools, or the mandate, to measure the cumulative suppression across an industry. The instrument and the harm simply don't match.
What people get wrong about the "just negotiate" objection
The most common dismissal of this critique is that workers with real bargaining power simply negotiate non-competes out of their contracts, or negotiate compensation for accepting them, and that sophisticated employees know how to do this. There's a grain of truth here. It misunderstands, though, how these clauses actually spread.
Non-competes are almost always presented at the point of hire, after the candidate has accepted the offer verbally, turned down other prospects, and has the least possible bargaining power. Research on how workers actually encounter these clauses consistently finds that most people see them for the first time on their start date, as part of a stack of onboarding paperwork. Refusing to sign typically means losing the job. Negotiating modifications requires knowing the clause is negotiable, having legal counsel, and being willing to risk the offer, which most workers, rationally, are not.
Found yourself in that position? If you are reading this before signing anything, you are already ahead of most people who have sat in that chair.
The deeper problem with the "just negotiate" framing is that it treats wage suppression as a series of individual transactions, each amenable to personal remedy, when it is in fact a structural condition of the labor market. Even if every worker had perfect information and a good lawyer, the systemic effect on wages across an industry would persist so long as enforcement remained permissive. Individual outcomes and aggregate harm are different things, and competition law, focused on individual conduct, was a poor instrument for addressing the second. Conflating the two is a category error that has served employers well for a very long time.
The real lesson of non-competes and wages is not that any particular clause is unfair to any particular worker, though many plainly are. It is that a legal instrument designed for one purpose, protecting genuine trade secrets and customer relationships, was allowed to drift into general use without anyone pausing to ask what it does to labor markets at scale. By the time economists had the data to answer that question clearly, millions of wage trajectories had already bent in the wrong direction. History has a habit of moving faster than the law that is supposed to govern it.