University of Maryland, Robert H. Smith School of Business

December 7, 2020

Includes first systematic evidence of noncompete use, effects of banning low-wage and high-tech noncompetes

The debate over whether noncompete agreements help or hurt employees is addressed in four research papers forthcoming in top journals and co-authored by management professor Evan Starr at the University of Maryland’s Robert H. Smith School of Business.

The results, he says, point to the same conclusion: Noncompetes stifle workers.

The findings will be published as

The latter publication, Starr says, represents the most sweeping work and is the first systematic investigation of noncompetes in the United States. In it, he studies a nationally representative sample, looking at all sorts of workers.

(Cytonn Photography/Unsplash)

One of the key findings: Noncompetes are found even among low-wage workers.

“There have been anecdotes of that fact, but this is the first systematic evidence,” Starr says. “This is shocking because when you think about noncompetes, you think about tech workers and executives – you’re not thinking about doggie-daycare sitters or hairstylists or yoga instructors, but that’s the modal worker that’s bound by a noncompete. Our paper launches from that fact, and the key question for policymakers is this a good or a bad thing?”

“The argument for why they are bad is pretty clear,” Starr says. “Take the case of the low-wage worker, earning $12 an hour, who gets a better offer at a competitor to make $15 an hour. A noncompete could prevent them from making those sorts of moves that are going to enhance their social and economic mobility,” Starr says.

Collectively, Starr’s papers show that workers do better without noncompete agreements. But what about firms?

Firms may be less profitable if they have to pay workers more, Starr says, but there’s definitely a benefit for them too: Without noncompetes, firms have unfettered access to the labor market and can hire the workers they want to hire, including those from a competitor.


“It’s not really a firm versus worker issue,” Starr says. “It could be a win for both workers and firms.”

So how do companies justify using these restrictions?

Starr says past research argues that noncompetes give firms incentive to invest in workers to develop their skills and knowledge to make them more productive, and that boost will help the individual workers make more money than they could have otherwise.

“All states in the U.S. that enforce noncompete agreements require that workers have some sort of ‘protectable,’ legitimate interest, which is exactly this sort of investment: That you’re giving workers trade secrets or very specialized training.”

Policymakers have taken up the issue. In late 2019, Starr testified about the use of noncompetes before the U.S. House of Representatives and the U.S. Senate. Even earlier, in 2016, the U.S. Treasury and the Obama Administration used Starr’s findings as the basis of related reports, spurring states to reexamine noncompete policies.

In the past two years, some – including Maryland, Virginia, and Washington DC – have moved to ban noncompetes for low-wage earners. But Oregon was the trailblazer, doing so back in 2008, making it the perfect case study for Starr (in the forthcoming in Management Science paper), with co-author Michael Lipsitz, an economist for the Federal Trade Commission.

Women benefit almost twice as much

The Oregon law banned noncompete agreements for all hourly workers and for workers earning below the median income for a family of four in the United States. It also banned noncompetes for certain occupations.

“We looked broadly at hourly workers, because hourly workers comprise 67% of working population in Oregon, and it was very clear in the bill that all noncompetes are banned for hourly workers,” Starr says.

(Inspired Horizons Digital Marketing/Unsplash)

The findings were clear: Low-wage workers benefited from the ban, earning higher wages and changing jobs more.

On average, hourly workers’ wages grew by 6% five years after the ban. The positive wage effects were across the board, in all age groups and education levels. The researchers also found that workers’ job-to-job mobility rose by 17% overall, with 79% of this increase in mobility driven by an increase in moves within an industry – “exactly the kinds of moves that a noncompete would prohibit,” Starr says.

The findings show that not only are workers moving to new jobs within their industries, they are also moving up to better jobs.

And while men benefit from this ban on noncompetes, women benefit almost twice as much, suggesting that noncompetes were holding down the earnings of low-income women more than men, Starr says. “It could be that women are less likely to negotiate over these sorts of provisions, and so they are harmed more when these are being enforced.”

The researchers analyzed how companies responded to the noncompete ban and found that there was no impact on hours worked, so firms are still employing workers, on average, for the same amount of time.

“What that means is workers’ take-home pay ultimately goes up – hourly wage increases translate into actual earnings increases,” Starr says.

They also looked at whether firms changed compensation structures in response to the new law, because Oregon only banned noncompetes for hourly workers and those who make less than the median household income for a family of four. They find that workers earning just above that threshold were more likely to become salaried. “One perk of the structure of this bill is to get workers into more stable situations, even if they are still using noncompetes with them,” Starr says.

“Despite good arguments for noncompetes potentially benefiting workers, the evidence just doesn’t support that claim. The evidence suggests that these are bad for low-wage workers.”

Paralleling the Oregon study, Starr’s research for the Journal of Human Resources looks at a 2015 Hawaii law that banned noncompetes in the high-tech industry. His findings are the same. “After Hawaii banned noncompetes for tech workers, their wages rise, their job mobility rises. It’s basically the same story but in a very different set of occupations, where arguments about the investment effects of noncompetes seem more plausible.”

And, the work in the Journal of Law, Economics and Organization looks at the influence of per se unenforceable noncompete agreements. “We find that they still do matter, in fact to a similar degree as in states where they are totally enforceable. We look at why that is and find evidence that workers are proactively turning down jobs, not because of the law, but because they are scared of a lawsuit.”