Workers have done quite well in recent years, at least according to official statistics: Some 2.7 million jobs were created last year, and the unemployment rate never went above a quite low 4%.

Still, in the four decades before 2020, these worker gains were anemic, and economists have been looking into the many factors. But here’s one possibility: noncompete agreements. They’re illegal for a number of positions in some states, but legal in many others.

Marketplace’s senior economics contributor Chris Farrell in St. Paul has been reviewing the literature. He spoke with “Marketplace Morning Report” host David Brancaccio. The following is an edited transcript of their conversation.

David Brancaccio: I once had a noncompete clause in a minimum-wage job. These are about, you can’t leave and work for a competitor for a period of time, right?

Chris Farrell: Right. And these restrictive noncompetes — it varies a lot by state. And there really isn’t a comprehensive data set on noncompetes. But the estimates range between 18% and 40% of the American workforce is impacted by these agreements. And they typically involve high-skill, high-wage professionals like financiers, software engineers. But they have become surprisingly widespread in lower-wage industries — that’s the experience that you had. Hair stylists, office cleaners. This is according to a report by the Federal Reserve Bank of Minneapolis.

Brancaccio: We’ve reported on the Federal Trade Commission, in Washington, is actively working to try to regulate many or most of these noncompetes out of existence.

Farrell: Yes. And so California, North Dakota, Oklahoma — they essentially have bans on noncompetes. Colorado and Minnesota have now joined the movement. And other states haven’t gone quite that far, David, but it seems like most are creating new restrictions.

Brancaccio: What do we know about the impact that noncompete limits have on workers?

Farrell: OK, so there’s a recent study by three economists, “The Labor Market Effects of Legal Restrictions on Worker Mobility,” and they find that increases in noncompete enforceability, that decreased workers’ earnings and mobility from 1991 to 2014.

Brancaccio: And did they detect a big hit?

Farrell: It is significant. I mean, they calculate that if a state moves from the 25th percentile to the 75th percentile — in terms of enforcing noncompetes — workers see a 1.7% average decline in annual earnings. And they say making noncompetes unenforceable nationwide could boost average earnings among all workers by 3.2% to 14.2%.

Brancaccio: Now, employers, businesses worry about trade secrets and the training that they invested in their workers. And many have been lobbying hard — or suing — to keep the noncompete arrangements.

Farrell: Yeah, it does still appear that the tide is turning. And, see what you think of this idea, David: Besides helping workers, an additional factor that is providing momentum to the anti-noncompete initiatives is the suspicion that it could be a real boon to innovation. There’s a series of fascinating studies that were done years ago looking into high-tech innovation. And these studies have argued that California’s long-term ban on noncompetes, that helped Silicon Valley become the epicenter of high-tech entrepreneurship. In comparison, take a competitor: Boston’s high-tech corridor, Route 128. Well, it fell short early on, perhaps because Massachusetts allowed businesses to limit employee mobility.