A few months ago, the Huffington Post reported on a surprising piece of the standard employment contract at Jimmy John's: A "non-competition" clause, preventing employees from working at any other shop within three miles that sells sandwiches for two years after they leave. It's convenient for the company, if it keeps sandwich-makers from taking their talents -- and inside knowledge -- to a rival. It's not so great for workers, who have much more limited options if they want to quit.

But wait: Is that really a danger at a place like Jimmy John's? Non-competes are normally reserved for executives or at tech firms, to keep people from bringing trade secrets to another company. Lately, though, they've been popping up in more and more low-wage sectors too, like maids and nail stylists, prompting a spate of lawsuits.

And now, we have actual data on how prevalent non-competes are across the labor force: One in four workers have signed such agreements in their lifetime, according to a new working paper, and 12.3 percent of them are bound by one right now. Here's a breakdown of the data on non-competes by profession:


(Evan Starr, JJ Prescott, and Norman Bishara)

The research, by economists Evan Starr at the University of Illinois and Norman Bishara and JJ Prescott of the University of Michigan, took the form of a large-scale online survey. Predictably, it showed that including non-compete clauses in hiring contracts is most widespread in information fields like engineering and architecture, where 30.1 percent of workers had signed one. Non-competes are also more common at higher levels of salary and education, which comports with the theory that the clauses are benign for employees with bargaining power, since they wouldn't sign them if they weren't in their best interest.

Starr found in a previous paper that in states with stronger enforcement of non-compete agreements, employees who've signed them tend to receive more training, which suggests that non-competes can serve as an incentive to invest in people.  That training is a good thing for employees' earning potential, although that's not very useful if they can't leave for a better offer.

But non-competes are also not uncommon for comparatively low-skill occupations, like personal care and services, where 11.8 percent of employees have signed one, and installation and repair at 10.5 percent. The prevalence of non-competes is actually lowest among food service workers, at about 3 percent).

And overall, only about 10 percent of workers who've signed a non-compete ever try to argue over it, with most assuming that it's either not negotiable or that doing so would cause tension with an employer.

There's a lot we still don't know about non-compete agreements, like whether they've become more common over time. And there's a lot we're about to learn, when Starr and his co-authors finish their analyses of how non-competes affect employment mobility. In the mean time, we know that the practice is remarkably widespread -- in places where they're probably not really necessary.