What an FTC Noncompete Ban Could Mean for Workers and Businesses

In January, the U.S. Federal Trade Commission proposed a ban on noncompete clauses in employment contracts. In addition to barring new noncompete agreements with employees and independent contractors, the rule would require employers to rescind existing ones. To understand what a federal ban could mean for workers and businesses, why it’s facing opposition, and how employers can prepare, we spoke with Evan Starr, an associate professor of management who studies noncompetes at the University of Maryland’s Robert H. Smith School of Business.

This interview has been edited for length and clarity.

MIT Sloan Management Review: How widespread are noncompete clauses, and who is subject to them?

Evan Starr: Surveys show that between 50% and 60% of firms use noncompetes for at least some workers, and about 30% of firms use them for all workers. An estimated 18% to 30% of workers are bound by noncompete agreements.

Executive roles are by far the most common place in which we see noncompete agreements. For other professionals, such as technical workers like engineers and doctors, it’s around 50%. But you see them in every corner of the labor market. The typical worker bound by one is paid by the hour and makes at the median about $14 an hour. I’ve seen noncompete agreements in volunteer contracts and nonprofit organizations, too.

When and how did noncompetes become common in the U.S.?

Starr: Noncompete agreements date back to the 1400s, when master craftsmen would train apprentices, who signed agreements that they wouldn’t become competitors. That eventually became English common law, which treated these on a case-by-case basis, balancing the need for this restriction against the harm that it did to a worker and society.

Most U.S. states followed this model, except California, Oklahoma, and North Dakota, all of which made noncompetes unenforceable in the late 1800s. Since then, there have been mostly tweaks across states until recently. In the past few years, there has been a raft of legislation aimed at reforming noncompete law at the state level, mostly to carve out low-wage workers who may be subject to these agreements.

Who counts as a low-wage worker is a mostly arbitrary distinction. Policy makers have taken enormously different approaches, banning noncompetes for workers making under $13 an hour to, recently in Washington, D.C., those making under $150,000 a year.

And what would you say has been the effect of these noncompete agreements on workers?

Starr: The best evidence we have comes from several natural experiments.

In 2008, Oregon became the first state to ban noncompete agreements for hourly workers and workers earning under the median income for a family of four. I have a paper on this, and we found that wages for hourly workers in Oregon were about 5% higher than for hourly workers in other states five years later, and their job mobility also increased about 11% over time.These laws don’t have an immediate effect because they typically aren’t retroactive; they apply only to new contracts.

Hawaii banned noncompetes in 2015 for tech workers only. That set up another natural experiment, allowing us to compare tech workers to other workers in Hawaii and tech workers in Hawaii to tech workers in other states. We found similar results: Wages among new hires grew about 4% and job mobility rose by about 11%.

Both cases suggest that these agreements hold workers’ wages down and prevent them from taking better jobs.

Is there any research on how noncompetes affect employers?

Starr: There are several studies documenting that when noncompete agreements are enforced or are more likely to be enforced, we see reductions in entrepreneurship, particularly among women, and new firms tend to grow a little bit slower, as they likely have trouble hiring.

There’s a paper by Hyo Kang and Lee Fleming on a 1996 policy in Florida, which became the vanguard of noncompete enforcement. Its statute said that when the state considers enforcing a noncompete agreement, it won’t consider the harm done to the worker. Kang and Fleming found that large firms benefited from that policy. Fewer small firms entered the state, and small firms made up a smaller share of employment. So it seems that noncompete agreements tend to benefit large incumbents over small firms and new entrants.

Do we know whether workers who are bound by noncompetes in one state move to another state to get out of them?

Starr: Absolutely. In the 1980s, Michigan changed its law to start enforcing noncompetes. Studies by Matt Marx et al. examined how policies like this cause workers to move across state lines. In particular, these studies found that the mobility of inventors in Michigan slowed relative to other states and that they were more likely to leave the state.

We have a follow-up paper that similarly found that in states that more vigorously enforced noncompete agreements, tech workers were more likely to leave the state for jobs elsewhere in the same industry.

What do you expect will be the effect of banning noncompetes at the federal level, as the FTC proposes?

Starr: The evidence suggests that workers will largely benefit by being able to take their skills to other employers and receive higher pay. Employers will benefit to the extent that they’re able to hire workers they couldn’t previously. Instead of having to hire, for example, a recent college grad who needs training, they may be able to hire somebody more senior. New businesses will be able to hire more easily.

It’s harder to say what will happen to innovation at larger firms. These firms are concerned that workers might take trade secrets they spent a lot of money developing and go across the street and sell them to competitors or start competitors themselves. You might worry that this harms firms’ incentives to develop these trade secrets.

But that argument may be somewhat overblown. The many companies founded in Silicon Valley had a lot of industry-specific knowledge, which generally would have been grounds to enforce noncompete agreements, but California’s long-standing ban on them allowed founders to start new firms, hire workers, and share ideas. The Silicon Valley example counters the argument that you need to enforce noncompetes to have innovation. In fact, maybe what really spurs innovation is workers having incentives to develop valuable knowledge for themselves with the ability to exploit it, which is suggested in a recent paper by Zhaozhao He that the FTC relies on.

The second major reason I think the concern is overblown is that there are other tools firms have besides noncompetes, such as trade-secret laws, NDAs [nondisclosure agreements], and agreements not to solicit clients or coworkers.

Is there a concern that companies might invest less in training their employees if noncompetes are banned?

Starr: That’s possible. One reason why firms invest in training is they’re unable to hire experienced workers who are bound by noncompetes, so they have to hire workers from outside the industry and then train them. It’s not clear what the optimal amount of training is, but the goal is not to maximize training expenditures. The goal is to maximize welfare. There’s a reason we don’t have patents last 100 years. You do want information to spill over to some extent.

Has the proliferation of noncompete agreements emerged as a kind of arms race, where one company might start putting these restrictions on their employees and so other companies in the industry would want to do the same in defense?

Starr: Absolutely. As soon as you try to hire from a competitor and get embroiled in a lawsuit over a noncompete agreement, or you get raided by a competitor, you realize that if you’re not using noncompetes already, you ought to think about it. There’s an article by Ronald Gilson published in New York University Law Review in 1999 that basically says that while it’s privately optimal for every firm to use noncompete agreements, if every firm does this, then you’re going to get a stagnant labor market where nobody’s moving, ideas are not flowing, new firms are not being formed, and there’s less competition, which is not good for workers or consumers.

Gilson argues that this is a prisoner’s dilemma, in which everyone is doing the thing that is privately optimal but it’s not socially optimal. So unless there is a ban on noncompete agreements, firms are going to engage in this behavior.

While so much of the evidence points toward the benefits of banning noncompete agreements, there is still some resistance to the notion of banning them at the federal level. The U.S. Chamber of Commerce said it would fight such a ban in court. What’s the rationale behind the opposition?

Starr: There are two reasons I see. One is that noncompetes have been under state purview for some 200 years, so states are not going to want their authority usurped. But while it might seem like a federal ban is overturning precedent, the FTC could have regulated noncompete agreements after they came into being in 1914. For whatever reason, it left it to the states until now.

One of the challenges with states setting their own policies here is that if you’re a multistate employer, it is difficult to navigate all these state laws with different exclusions for noncompete agreements. For workers, if you move across state lines, how do you know which state’s law is going to apply? There is value to uniformity. Whether the FTC is the right vehicle for that uniformity is a different question.

The second point is that the other tools firms have to protect their legitimate interests are harder to enforce. To enforce an NDA, for example, you need to prove in court that somebody violated the terms of the NDA, which can be difficult if that person was secretive. In contrast, it’s easy to prove that somebody violated a noncompete agreement. You just observe that they went from firm A to firm B.

What measures should companies be taking ahead of an expected ban on noncompete agreements? Do you have any suggestions?

Starr: One suggestion is that if you’re using noncompete agreements for every worker at your firm — and we know a lot of firms do this — then revisit that, because the FTC is probably going to ban them for at least some workers.

The ban as currently proposed will apply to all workers. But there is a push to carve out high-wage workers. In that case, start thinking about your low-wage workers and figure out which ones don’t need to have a noncompete agreement, because then you’re more likely to be compliant with any particular ban.

If this ban comes into play, the other thing to start thinking about is, who are those workers who would be great fits at your firm? If you’re a startup or a firm with a new idea but you currently can’t hire people with the expertise you need because they’re bound by noncompetes, you may want to start identifying the workers you might want to try to hire once this ban comes into effect.