Do Non-Compete Agreements Really Impede Innovation?

I spent my entire practice career in California, representing clients who were, by and large, free to move from company to company due California's ban on covenants not to compete. At the time, the question of inevitable disclosure was unsettled in California; the law allowed for injunctions to stop "threatened" misappropriation, but the question was whether merely being in the same business is a sufficient threat. Most people though no, and the courts eventually went there.

But employees were not always free to do as they wished. I represented many employees against their awful former employers who (wrongfully) claimed that the employee had exited with trade secrets. And I represented many companies against their awful former employees who (wrongfully) claimed they had exited cleanly without trade secrets. For this reason, I've always bristled at the notion that California is a place where everyone happily moves from company to company and nobody cares or legitimately worries about trade secrets because it's all good.

But through this all, I always thought employee mobility was a good thing. I still do. And so do a lot of other people. Indeed, a sort of orthodoxy has arisen where you simply can't be honest, pro-innovation, and pro-noncompete all at the same time. It's as if an entire strand of the literature on how non-competes can improve investments in human capital training got lost in the mix, replaced by theory that says the contrary. This theory is supported by evidence from empirical studies on the matter. It seems crystal clear.

Not so, according to Jonathan Barnett (USC) and Ted Sichelman (San Diego). A draft of their paper, Revisiting Labor Mobility in Innovation is now up on SSRN. In it, they take on the orthodoxy:

It is now widely asserted that legal regimes that enforce contractual and other limitations on labor mobility deter technological innovation. First, recent empirical studies purport to show relationships between bans on enforcing noncompete agreements, increased employee movement, and increased innovation. We find that these studies misconstrue legal differences across states and otherwise are flawed, incomplete, or limited in applicability. Second, scholars have largely adopted the view that California’s policy against noncompetes promoted Silicon Valley as the world’s leading technology center. By contrast, Massachusetts’ enforcement of noncompetes purportedly stunted innovation in the Route 128 region near Boston. We show that this account is incomplete. During the rise of Silicon Valley, California noncompete law did not as vigorously preclude noncompetes as today and firms could substantially mimic noncompetes through contractual and other instruments. Rather, fundamental technological and economic factors more persuasively account for the rise of Silicon Valley and the Boston area has remained a significant innovation center. There is little compelling ground for the view that barring noncompetes and other limitations on employee mobility promotes innovation.

I believe this is an important paper, even if you are not inclined to agree with its conclusions. I'll discuss more after the jump.

The paper proceeds in three basic sections. The first is a discussion of the transition in thinking about noncompetes over time. It presents a short review of the older literature and shows how that transitioned to today's thinking, including identification of some of the key papers that started the transition.

It then launches into critique of what the paper calls the "weak thesis," that noncompetes depress labor mobility and further questions the "strong thesis," that labor mobility enhances innovation. In doing so, nothing is safe. Statements about the laws of states are examined closely. Assumptions about how those laws are/were applied are examined. The paper even goes after the bedrock of all worker mobility scholarship, Saxenian's book and Gilson's article on Route 128 and Silicon Valley. The analysis takes a critical look at the statements of those books, examining the companies, practices, norms, and legal differences that were claimed to give rise to different innovation paths. The article concludes that technological and economic factors explain any differences, and that--with apologies to Monty Python--Route 128 isn't quite dead yet, and in fact is pretty robust.

Finally, the paper takes on all of the empirical studies, questioning both data but also categorizations used for differences-in-differences analysis. These studies, they argue, don't necessarily give the wrong answer, but neither do we know if they give the right one. More and better (and different?) studies are necessary, especially to support the strong thesis.

The article is very effective in its criticism, even if I quibble with a few of the arguments. Those who favor mobility should take a hard look at the critiques and consider responses. I like this paper, but not necessarily because it questions noncompete agreements.  As I said above, I like the law in California. I think it's the better law, and I think it's the law other states should adopt. And I even think it leads to better human capital and innovation outcomes. But I believe these things to be true. I don't know them. I like this paper because it takes a hard look at views we take for granted.

Further, the cranky scholar in me believes that the more time that passes where most people believe something, the harder we should look at the reasons why we have that belief. To repeat a quote often attributed Richard Feynman (though I can't confirm* that he said it): "I would rather have questions that can't be answered than answers that can't be questioned." This is why I like Beauchamp on the first patent explosion and Katznelson & Howes on the Wright Brothers. It's why I try to read source materials whenever I am doing work, rather than reading someone else's view of source materials. It's even why we should continue to aggressively peer review climate change studies, even if one believes (as I do) that it exists and is caused by humans.

*According to the discussion on Wikipedia, the quote is likely more accurately: “You see, one thing is, I can live with doubt, and uncertainty, and not knowing. I think it’s much more interesting to live, not knowing, than to have answers which might be wrong.”