David R. Henderson  

The Case for Non-Compete Agreements

India's Assault on Money... Trump on Obamacare...

I sometimes learn from people's comments. So I want to hat tip a recent commenter on one of my four pieces on the Council of Economic Advisers' report on alleged monopsony in the labor market. In Part II, I said that non-compete agreements for workers who don't have company secrets seem to "suck." Well, at least I said "seem." My late professor and mentor Armen Alchian would have tut-tutted if he had seen that. One of the most important things I learned from Armen is that when you see a practice that is adopted voluntarily but that doesn't seem to make sense, don't immediately assume that it doesn't make sense. Instead, ask under what conditions it might make sense.

I forgot that momentarily.

But commenter Aaron McNay caught my error, pointing out:

Both employers and employees would like to be able to train the employees if the cost of doing so is less than the gains in productivity. However, there is a potential collective action problem here. What happens if the employer provides the training, but the employee then moves onto another job? The employer bears the burden of the training costs, but does not receive any of the benefits. As a result, the employer does not provide the training, and a mutually beneficial trade is not made.

By preventing the employee from being able to move, a non-compete agreement eliminates the collective action problem. The employer provides the training and the employer and employee are made better off. There are more complications to this than presented here, of course. However, I still think it is an issue that is ignored too often when talking about these type of agreements.

Jason Furman and Alan Krueger followed up on the CEA report with an op/ed in the Wall Street Journal.

That led me to write this letter, which was published today:

Jason Furman and Alan B. Krueger ("Why Aren't Americans Getting Raises? Blame the Monopsony," November 4) claim "There is no reason why employers would require fast-food workers and retail salespeople to sign a noncompete clause--other than to restrict competition and weaken worker bargaining power."

That's false. There's potentially a very good reason for employers to require such clauses: so that they can get a payback on the training they provide.

A dilemma that faces any employer who provides training to his employees is how to collect on the training. Imagine that the worker produces output worth $10 per hour and is paid $10 per hour. Imagine also that the employer can provide training that costs $1 per hour but increases the employee's productivity by $1.50 per hour. It makes sense to do that, right?

But what if the employee, with that training, can get $11.50 per hour elsewhere that reflects his higher productivity. He may well leave. Then the employer who provided the training lost on the deal. Solution: a non-compete agreement so the employee can't work for a competitor in the same industry that finds that training useful.

The WSJ editor changed the last sentence in a way that's not quite faithful to my view, to make it sound as if I advocate such contracts. I don't know enough to advocate them. Rather, I'm suggesting that employers see non-compete agreements as a solution.

Comments and Sharing

CATEGORIES: Labor Market

COMMENTS (15 to date)
Toby writes:

There is a much simpler solution to address the problem directly than a non-compete agreement: have employees reimburse the employer for the training if they move within x years of working there.

I had a deal when I was working and studying at the same time. My employer offered to pay for my tuition and my books and would forgive this "loan" over a number of years.

The only reason I can think right now of that justifies a non-compete clause is if the employee would not be able to pay back the training and wouldn't be able to do so in the forseeable future. That or if the cost of the training is difficult to quantify (then again how would you determine how long the employee would not be allowed to compete?).

I don't think that this makes much sense though.

David R. Henderson writes:

Beware of single solutions. Different situations require different responses.

Michael Byrnes writes:

If this were the rationale for non-competes, they would be void upon the worker being laid off, no? But such agreements often are not.

BTW, does California (where there are no noncompetes) have a problem with training of such workers?

Kevin writes:

I don't agree. You could for example, enforce a contract that quitting before a certain time requires the employee to reimburse the company $X for training costs. These contracts exist, and they're a much better approach than non-competes for low skilled workers.

Sebastian H writes:

This would make sense if non-compete agreements expired after working somewhere for a medium length of time. I've never seen one that actually functioned like that though.

john hare writes:

Thank you David.

I keep an eye out for things that might help my company. I had never considered a non-compete for certain classes of employee training due to the problems pointed out here.

I am going to give some thought to the possibility that offering serious training for the work we do could be done with a couple of short term restrictions. As it is, training is strictly OJT and has been in my local industry for the decades that I have been in it. The college classes that I took on my own dollar way back provided a background that helped me be very competitive with people with far more field experience.

I may see a way now to invest the money, and more importantly the time to train $10.00 an hour people for $20.00 an hour skills. May not work, but I've chased that dog around the house before.

Steven Crowley writes:

In the absence of such agreements, the trained employee may leave, but the employer may gain a new employee trained by someone else. In the mass market, this seems a wash to me, without the added overhead of an agreement process.

GM writes:

I wonder if the corollary to this is vertical relationships in product markets. Practices such as exclusive dealings or minimum retail prices that appear to be anticompetitive but keep competitors from benefiting from costly information provided to customers. One must dig beneath the surface to understand how the practice enhances value to buyers and sellers perhaps.

AS writes:

How often are non-competes actually enforced? Other mechanisms that work: 401k vesting, end of year bonuses, and stock options -- all become forfeit when you leave.

James writes:

McNay's scenario is one in which the noncompete agreement benefits both parties, but there is no reason to suppose this the actual reason for noncompete clauses.

Here is another possibility: Employers benefit from a noncompete agreement because these agreements restrict the applicant pool available to competitors and make the threat of termination more dire to employees who prefer to remain in the industry. Potential employees go along with the noncompete because they are afraid questioning a noncompete agreement will alarm potential employers that they might not plan on sticking around.

Mike L writes:

I work in California and have started several high tech companies in the state. Non-compete agreements are not enforceable in California, the state considers such agreements an illegal restraint of trade. Obviously this has not prevented companies in California from both protecting trade secrets and training and retaining valued employees. In place of non-compete agreements most companies use a non-disclosure agreement to protect trade secrets. These are vigorously and successfully enforced in California. And when it comes to investing in training an employee most business owners I know make that investment because it is less expensive then trying to higher a new employee with the necessary skills. And once an employee is trained and preforming the new job then you're going to be paying the higher wage that other employees are getting for doing the same work, both within the company and outside. In my opinion non-compete agreements are completely unnecessary.

George Berger writes:

One thing this discussion has overlooked. Employers will want to protect their investment in general training that's transferable to other employers. They may not want to do that because they will lose their investment if the worker moves. However providing specific training which cannot be transferred to other firms will usually NOT require any such protection. It's the type of training provided that matters here.

Aaron McNay writes:


I am perfectly happy to admit that this may not be why there are non-compete agreements for low-skilled workers. My complaint was that this appears to at least be a possible reason why these agreements exist in low-skilled jobs and that the CEA report does not even really mention this as a possibility.

I would be more than willing to say that this is not what is occurring if there was evidence that this is not the case. Unfortunately, we don't know if there is, at least based on what was provided in the CEA report.

Does anyone know if there has been work done examining this issue? The only paper I know of that begins examining non-competes and job training is here:

David R. Henderson writes:

@George Berger,
Good point, but I didn’t overlook it. My discussion was explicitly about general training. That’s why this sentence was in my WSJ letter: "But what if the employee, with that training, can get $11.50 per hour elsewhere that reflects his higher productivity."

David Guetta writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring this comment and your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Comments for this entry have been closed
Return to top