Two Ways a Non-Compete Can Change a Career
This chapter examines the two economic mechanisms behind a non-compete agreement (NCA): protecting a firm’s investment in firm-specific human capital, or narrowing a worker’s outside options and bargaining power. The distinction matters because the same signed clause can produce very different career paths. By the end, the reader should be able to predict what each mechanism would look like at the moment a worker changes jobs and over the years that follow, without treating either explanation as established merely because it sounds plausible.
The useful test is temporal. If a worker is giving up future mobility in exchange for something, what should appear immediately? If the agreement helps the employer invest in knowledge that is valuable inside the firm, what should appear later? And if the restriction mainly makes alternative employers less reachable, how might the cost accumulate? These questions turn a broad argument about non-competes into observable predictions about wages and job movement.
When the firm is protecting knowledge it helped create
The investment defense begins with a problem of specificity. Some knowledge is portable: a worker can use it in many firms, and another employer can value it without knowing the details of the worker’s current workplace. Other knowledge is firm-specific. It may involve a company’s processes, relationships, or way of combining tasks. Its value is greatest inside the firm that helped create it, and it may be difficult to protect through an ordinary promise to pay for training later.
The concern is not simply that a worker becomes more capable. It is that the firm bears a cost today while the resulting value may be used elsewhere tomorrow. If the worker can immediately take the relevant knowledge to a competitor, the employer may fear that the investment will benefit a rival. An NCA can be proposed as a way to limit that immediate transfer or to make the employer more confident that the investment will remain attached to the employment relationship for a time. In the paper’s framing, this is an incomplete-contracting argument: the agreement is meant to solve a problem that wages and ordinary promises do not fully solve.
That defense does not require the restriction to be good for the worker in every respect. It claims that the firm’s investment could create value that otherwise would not exist. A worker might accept the constraint if the job offers unusually valuable training, a role that builds specialized experience, or compensation that recognizes the surrendered option. The central question is whether the value received is enough to offset the future mobility the worker gives up.
This is why an upfront wage difference matters. In a competitive version of the investment story, a worker who is asked to accept a restriction should have some reason to do so. One possible reason is a transition compensating differential: a positive wage jump at the move into the NCA job. The payment would not prove that the arrangement is fair, but its presence would fit the idea that the worker is being compensated for accepting a limitation.
The timing is crucial. A positive wage at a single point can be a reward for many things, including ability, job quality, or a favorable match. A wage jump right around the transition is more specifically relevant to the proposed exchange. It asks whether the worker appears to receive a premium at the moment the restriction begins, rather than merely having a higher wage in a job that was different from the start.
The investment account also makes a prediction about the later path. If the firm’s protected investment improves the worker’s productive capability, later wage growth might be faster. The worker may become more valuable through experience that is difficult to obtain elsewhere, or the firm may have more reason to reward that accumulated contribution. The Chicago Fed paper describes a positive post-transition gradient as consistent with human-capital investment and faster wage growth. “Consistent with” is doing important work here: the wage pattern can fit the mechanism without directly observing the training expenditure, the knowledge created, or the negotiation that links them.
The investment story therefore has a recognizable shape:
- The worker enters a job with an NCA.
- The restriction helps the firm feel more secure about investing in firm-specific human capital.
- The worker receives some immediate compensation or another sufficiently valuable benefit.
- The worker’s later wage path improves as the investment becomes productive.
That sequence is a hypothesis, not a universal lifecycle. It is more plausible when the job actually involves meaningful firm-specific learning and when the employer’s investment is material enough to need protection. It is less persuasive if the restriction is doing no visible work for training or knowledge creation. But the available study does not directly measure those job-level ingredients, so the mechanism must remain an interpretation rather than a finding that the data independently prove.
A hypothetical example makes the logic concrete. Imagine an engineer entering a specialized role in which the employer spends substantial time teaching an internal process that has little value outside the firm until the worker has mastered a larger system. The employer may view an NCA as protection for that investment. The worker may view the same arrangement as a trade: less freedom to move immediately, but access to training and a possible path to greater responsibility. Whether that trade is attractive depends on the actual compensation, the portability of the resulting skills, and the worker’s alternatives. The example illustrates the mechanism; it is not a reported case from the study.
The investment account can also coexist with unequal bargaining power. A worker may receive useful training and still be underpaid relative to the value created. The existence of firm-specific investment does not automatically settle who should capture the return. It only explains why the employer might want a restriction and why a later positive wage path would be relevant evidence. The worker’s career judgment still depends on the balance between learning, compensation, autonomy, stability, and the value of remaining free to move.
When the restriction weakens the next negotiation
The second mechanism starts from the worker’s outside option. An outside option is not every imaginable job; it is a plausible alternative that the worker could pursue and credibly accept. It gives the worker leverage because an employer must consider what happens if the worker leaves. If an NCA makes a competing job appear unavailable, risky, or costly to pursue, the worker may have fewer credible alternatives even before anyone tests the clause in court.
The effect can begin at the job search stage. A worker who believes they cannot join a competitor may not apply. A worker who does apply may discount the value of the opportunity because of uncertainty about legal consequences. Or a worker may stay in a job while negotiating from a weaker position because an outside offer is less usable. These are mechanisms through which a restriction can matter in practice; the paper does not directly observe each belief or decision, so they should not be read as measured intermediate outcomes.
Once the outside option is narrower, the economic consequence need not appear as an immediate wage cut. A worker can start with the same wage as a comparable unrestricted worker and still face a weaker path later. Wage growth often depends on the ability to move, compare offers, signal one’s value to another employer, or credibly threaten to leave. If those routes become less available, the worker may have less leverage at each later decision point. Small differences can accumulate into a slower trajectory even if no single pay decision looks dramatic.
This is the central contrast with the investment defense. Under the outside-option account, the contract’s cost is a dynamic loss of bargaining power rather than a price paid once at entry. A positive transition premium is not required. Indeed, the absence of a clear premium would be troubling for a story in which workers are competitively compensated for giving up mobility. A negative later wage gradient, by contrast, would fit the possibility that the restriction’s effect grows as the worker encounters future choices.
The practical channel does not depend on a judge declaring the agreement enforceable. The Chicago Fed paper notes that workers may comply with NCAs because they are uncertain about their legal status or fear litigation, including in settings where enforceability is weak or absent. Uncertainty changes the perceived cost of an action. If the cost of finding out is high enough, a worker may behave as though the restriction is stronger than a later legal analysis would say.
That does not mean uncertainty always harms the worker or that every worker responds in the same way. A worker with several strong alternatives, specialized credentials, or access to professional advice may treat the clause differently from a worker with few nearby employers and limited resources. The point is narrower: the practical constraint can be a function of perception and risk, not only of formal legal strength. This is one reason incidence-based evidence can capture a channel that enforceability-based studies may miss.
A second hypothetical illustrates the alternative path. Consider a worker moving into a job where the employer offers little firm-specific training, while the worker’s skills could otherwise be used by several local employers. If an NCA makes those moves feel uncertain, the restriction may not protect a major investment. It may mainly reduce the number of credible next steps. The worker can still earn a reasonable starting wage, yet have fewer ways to turn experience into a raise or a better match. Again, this is a mechanism example, not an observation of a particular person in the data.
The two pathways can coexist in one workplace. An employer may invest in some genuinely firm-specific capabilities and also gain leverage from limiting the worker’s alternatives. The question is which effect dominates for a particular group of workers and over which horizon. That is why the study’s education split is not a decorative detail. It offers a way to ask whether the same kind of agreement operates differently in labor markets with different work content, investment opportunities, and outside options.
Education itself is not the mechanism. In the paper, it divides the sample into workers with less than a four-year college degree and workers with a four-year degree or more. Those groups may differ in the kinds of jobs they hold, the extent of firm-specific investment, the number of plausible employers, and the portability of their skills. But the study does not establish that education alone causes the divergence. It uses the groups to reveal heterogeneity that a single average effect would hide.
The empirical predictions are now clear enough to test. At the transition, a positive coefficient would be compatible with an upfront compensating differential, while an estimate near zero would leave that compensation story unsupported but would not prove that no immediate benefit exists. After the transition, a positive wage gradient would be consistent with an investment account; a negative gradient would be consistent with a loss of outside options and bargaining power. Before the transition, the absence of differential trends would make it less likely that the two groups were already moving along different wage paths for reasons visible in the design.
Those predictions also define the limits of the argument. Neither a positive nor a negative trajectory identifies a single mechanism by itself. The data can show how wages evolve around a change in NCA status; they do not directly record a firm’s training budget, a worker’s legal knowledge, every rejected application, or the precise negotiation behind a raise. The mechanism-first model is valuable because it tells us what to look for and what would count against a simple story. It is not a license to convert a plausible explanation into a measured fact.
The next step is therefore methodological but not merely technical: follow workers through actual job transitions and compare them with workers whose NCA status does not change. Only that comparison can tell us whether the observed wage path bends when the restriction changes, rather than simply reflecting who takes which jobs.