Recently, President Biden signed an executive order that, among other things, directs the Federal Trade Commission to ban or limit noncompete agreements. Noncompetes preclude departing employees from working for a competitor, typically in the same industry for a period of one to two years and sometimes within a limited geographic region. For instance, a design engineer who resigns from Ford may be prevented by a noncompete from working at GM for a year. Noncompetes in various forms have been used for centuries and today are most prevalent in professional, scientific, and technical occupations.
Motivating the president’s executive order are the many policymakers, economists, and legal scholars who have recently argued that noncompetes depress wages, stymie innovation, and slow economic growth. Following this school of thought, there is virtually nothing good about noncompetes. For example, FTC Chair Lina Khan contends that such clauses “deter workers from switching employers, weakening workers’ credible threat of exit, and diminishing their bargaining power.” To support these arguments, policymakers and scholars repeatedly point to multiple studies allegedly showing that noncompetes inhibit innovation by reducing employee mobility.
Unfortunately, these arguments and studies are seriously flawed.
As we show in an exhaustive critique published in the University of Chicago Law Review, all of the major economic studies claiming negative effects on innovation and economic growth from noncompetes have significant errors or are incomplete. Invariably, these studies are produced by economists and business school professors whose interpretations of state law are over-simplified or contain serious errors. The upshot is that these studies’ results are not a reliable basis for making major policy changes that would displace most states’ tailored enforcement regimes with a sweeping national ban of noncompetes.
For instance, one leading study that purports to show the negative effects of a Michigan law enacted in 1985 that permitted noncompetes wrongly assumes that the law was retroactive, applying to existing employment agreements. As a result, the authors mistakenly infer that a nearly immediate drop in relative employee movement was caused by the 1985 law. Other errors in major studies include wrongly categorizing states that enforce noncompetes as not enforcing them (or vice-versa), overlooking important exceptions to restrictions on enforcing noncompetes, ignoring legal substitutes for noncompetes (such as vesting stock over time), improperly measuring the strength of noncompete enforcement, and misunderstanding inter-state “choice of law” rules.
These studies’ largely unqualified conclusions that noncompetes almost uniformly have adverse welfare effects run counter to older and more nuanced views among economists that noncompetes give rise to offsetting effects on innovation, employee training, and economic growth.
A noncompete can yield economic gains by providing an incentive for employers to invest in training their employees and investing in R&D. If competitors could “free ride” off each other’s investments in training and R&D, there would be a disincentive to make those investments in the first place. Balanced against these pro-competitive interests are the costs of noncompetes attributable to diminished employee mobility, which may result in the reduced circulation of ideas and, possibly, depressed wages. Consistent with this tradeoff, other empirical studies find that noncompetes decrease employee movement to a certain extent but incentivize employers’ investment in certain forms of R&D, training, and capital investment.
These complex policy tradeoffs are old news for the common law, which has traditionally allowed noncompetes, but subject to “reasonableness” requirements that significantly limit their reach and scope. This approach in various forms remains the law in all but a handful of states.
A well-known exception is California, which has substantially limited noncompete enforcement. Opponents of noncompetes often point to Silicon Valley’s success as a reason for restricting noncompetes. However, we show that other reasons are more likely to account for this success. These include a diverse technology ecosystem, local venture capital and educational resources, cultural attitudes towards risk, desirable weather, and to a significant degree, luck. Moreover, states that have traditionally enforced noncompetes (subject to reasonableness limitations) have cultivated leading technology clusters, such as biotech in Boston, medical devices in Minneapolis, and information technology in Austin.
In contrast to tech workers, we agree that the reasons for limiting the post-employment options of lower-wage workers — for instance, at sandwich shops — are weak, and there is little prospect of economic harm from restricting noncompetes in these situations.
In contrast, there is strong reason to believe that nationwide restrictions on noncompetes in knowledge-intensive industries could result in substantial economic losses from diminished investment in R&D and employee training.
At the very least, before making sweeping national changes that would supplant states’ nuanced approaches to this long-standing contractual provision, President Biden and the policymakers and scholars who advocated for his executive order should re-examine the faulty studies motivating their views. Otherwise, as the adage goes, bad facts will make bad law.
Jonathan M. Barnett is a professor at the University of Southern California, Gould School of Law. Ted Sichelman is a professor at the University of San Diego School of Law, where he directs the school’s Center for Intellectual Property Law & Markets.