Business · 7 min read
What Research Finds About Noncompete Agreements and Worker Mobility
What U.S. evidence shows about mobility and earnings, the limits of natural experiments, and the legal debate after the federal rule was vacated.

What Research Finds About Noncompete Agreements and Worker Mobility
A noncompete agreement restricts a worker from joining or starting a competing business for a stated period, area, or line of work after employment ends. U.S. evidence consistently associates enforceability with lower job mobility and, in many settings, lower earnings. Studies of policy changes in Oregon and Hawaii strengthen the causal case because they compare affected workers with plausible control groups before and after restrictions changed.
The evidence does not show that every clause has the same effect or that employers have no legitimate confidential information to protect. It does show that noncompetes extend beyond senior employees with trade secrets, are often presented with limited bargaining, and can affect workers who never signed one by reducing outside options in a labor market.
What a noncompete restricts
Noncompetes are one member of a family of post-employment restraints. Nondisclosure agreements prohibit use or disclosure of defined information. Nonsolicitation clauses restrict contact with clients or coworkers. Trade-secret law supplies remedies for misappropriation. A noncompete goes further by barring specified competitive work even if no secret is taken.
Legal rules differ by U.S. state. California has long treated most employment noncompetes as void; other states enforce clauses if courts find duration, geography, and protected interest reasonable; some set earnings thresholds or notice requirements. This variation gives researchers comparisons, but state policy is not randomly assigned. Industries and firms choose locations, and legal culture correlates with other economic conditions.
Enforceability and signing are also distinct. A worker may comply with a clause that a court would reject because litigation is expensive and the law is uncertain. Employers may decline to hire someone after seeing a contract. Conversely, a signed clause may never alter behavior. Studies therefore measure contracts, legal enforceability, job moves, wages, and employer practices separately.
Prevalence and bargaining conditions
Evan Starr, J. J. Prescott, and Norman Bishara surveyed 11,505 U.S. labor-force participants. Their 2021 Journal of Law and Economics article estimated that about 18 percent were currently bound and 38 percent had agreed to at least one noncompete in the past. The clauses were more common in high-skill, high-paying work but also appeared in low-wage jobs and in states that did not enforce them.
Only about one in ten employees in that survey reported bargaining over the clause. Roughly one-third of signers received it after accepting the job offer. Timing matters because advance notice gives a candidate a chance to compare offers or negotiate compensation; a first-day form creates a different choice after the worker has declined alternatives or moved.
The survey is nationally broad but observational and self-reported. Workers may misremember language or confuse a noncompete with another restriction. The authors compared responses with contract text in a subsample and used detailed definitions, but prevalence estimates still have uncertainty. More importantly, the fact that clauses coexist with wages cannot by itself show which caused the other: firms use them more often for employees who already differ in pay and access to information.
Policy changes provide stronger evidence
Michael Lipsitz and Starr studied Oregon's 2008 decision to make noncompetes unenforceable for hourly workers. Using neighboring and comparison workers over time, they estimated that the change raised hourly wages by roughly 2 to 3 percent on average and increased job-to-job mobility. The design is stronger than a simple state comparison because it uses a defined policy break, though it still assumes Oregon's affected workers would otherwise have followed the comparison trend.
A study by Natarajan Balasubramanian and four colleagues examined Hawaii's 2015 ban for technology workers. Relative to other workers and states, mobility among Hawaii technology workers rose about 11 percent and new-hire wages about 4 percent. The law's narrow occupational scope creates a useful natural experiment. Its limits are equally narrow: Hawaii is a geographically isolated labor market, the covered group was specific, and employers or workers may have adjusted along margins the data did not capture.
Larger panel studies compare changes in state enforceability over time. In a 2025 Journal of Political Economy article, Matthew Johnson, Kurt Lavetti, and Lipsitz found that stronger enforceability reduced earnings and mobility, with larger effects among occupations more likely to sign. The analysis supports a bargaining mechanism: outside offers matter for earnings even when workers stay with their employer, so restricting credible departure can weaken wage growth.
Spillovers complicate estimates focused only on signers. Starr, Justin Frake, and Rajshree Agarwal found fewer offers, lower mobility, and lower wages for unconstrained workers in state-industry cells with greater noncompete use and enforceability. If firms recruit less from rivals, the number of openings falls for others too. That result is observational and depends on measured state-industry variation, but it explains why individual consent cannot capture the entire market effect.
The employer-investment argument and alternatives
Employers argue that a credible restriction can support investment in training, customer relationships, and shared knowledge. Without it, a competitor could hire a trained worker and avoid the cost. Economic theory allows this possibility: if a noncompete increases joint productivity and the worker bargains over the gain, compensation or training could rise.
Some empirical studies find more training associated with enforceability or signing. Association is difficult to interpret because jobs with complex knowledge offer both more training and more clauses. Even when the training effect is causal, welfare depends on whether its value exceeds losses from reduced matching, entrepreneurship, and bargaining. A worker trained for one firm but prevented from taking a better-matched job may be more productive and still earn less than under a different arrangement.
Less restrictive tools can protect narrower interests. Trade-secret law targets misuse of secret information. Nondisclosure and nonsolicitation clauses address particular conduct, although they too can be drafted broadly. Fixed-term contracts, garden leave, deferred compensation, and repayment agreements distribute costs differently. Whether an alternative is effective depends on enforcement, clarity, and the kind of asset at risk.
The comparison also varies by occupation. A negotiated restriction for a seller of a business is not equivalent to a standard form for a sandwich worker. Physicians raise continuity-of-care and local-access questions. Senior executives may receive compensation tied to a clause. Evidence averaged across workers should not erase these contractual differences.
A fragmented legal debate after the federal rule
The Federal Trade Commission adopted a nationwide rule in April 2024 that would have barred most worker noncompetes. A federal district court vacated the rule on August 20, 2024, concluding that the agency lacked statutory authority. The rule did not take effect. In September 2025, the FTC voted to dismiss its appeals and accept the vacatur. As of August 2026, the agency's rule page states that the rule is not enforceable.
That sequence did not legalize all noncompetes. State statutes and common law continue to govern most contracts, and federal antitrust enforcement can challenge particular conduct. In 2025 the FTC pursued a case involving noncompetes at a pet-cremation company and emphasized case-by-case enforcement. The policy debate has therefore shifted back toward state legislation, targeted federal cases, and possible congressional action.
The failed rule also illustrates the difference between evidence and authority. Economic findings can inform whether a restriction harms competition, but they do not decide which agency may issue a categorical rule. Conversely, vacating the rule on statutory grounds does not reverse the empirical estimates. Legal validity and economic effect require different evidence.
What the evidence supports
Across surveys, state comparisons, and policy changes, the central result is consistent: broader noncompete enforceability reduces worker movement, and many studies find lower wages or earnings. Oregon and Hawaii offer especially informative bounded cases. Spillover evidence suggests the effect can extend beyond the person who signs.
Uncertainty remains about the size of national effects, the value of employer investment induced by clauses, and which tailored restrictions could preserve that value at lower cost. Results from hourly workers or one technology market should not be universalized. The current evidence supports skepticism toward routine, broad noncompetes. It does not eliminate the need to compare occupations, notice, bargaining, compensation, duration, and less restrictive protections in a specific policy.
Sources
6 sources, 5 of them with a link you can follow.
- Starr, Prescott, and Bishara, “Noncompete Agreements in the U.S. Labor Force,” Journal of Law and Economics 64 (2021), doi:10.1086/712206.
- Lipsitz and Starr, “Low-Wage Workers and the Enforceability of Noncompete Agreements,” Management Science 68 (2022), doi:10.1287/mnsc.2020.3918.
- Balasubramanian et al., “Locked In? The Enforceability of Covenants Not to Compete and the Careers of High-Tech Workers,” Journal of Human Resources 57 (2022), doi:10.3368/jhr.monopsony.1218-9931R1.
- Johnson, Lavetti, and Lipsitz, “The Labor Market Effects of Legal Restrictions on Worker Mobility,” Journal of Political Economy 133 (2025), doi:10.1086/736217.
- Federal Trade Commission, “Noncompete Rule” status page and September 5, 2025 vacatur statement, accessed August 2026.
- Evan Starr, “The Economics of Noncompete Clauses,” Journal of Economic Perspectives 40 (2026), doi:10.1257/jep.20251457.
