Over the weekend, venture capitalist Bijan Sabet kicked off an interesting discussion by saying that he doesn't believe in noncompete agreements
and suggesting, anecdotally, why he thinks that they do more harm than good. Venture capitalist Fred Wilson responded by disagreeing
and suggesting that noncompetes do more good than harm. This is a topic that I've become deeply familiar with recently, for some research I've been working on. My interest in the specifics of noncompetes was kicked off by a small part of David Levine and Michele Boldrin's book Against Intellectual Monopoly
, where they discuss how the lack of noncompetes helped Silicon Valley grow. This lead me to a lot of research on the topic, some of which I thought it would be worth bringing up, as the discussion has become so heated -- with almost all of it focused on anecdotal points, rather than actual research. Some of this research was for a separate project I am working on, but with so much interest in the topic, I thought it would be worth a detailed post.
Much of this discussion kicked off with AnnaLee Saxenian's 1994 book Regional Advantage
that tries to understand why Silicon Valley developed into the high tech hub it is today, while Boston's Route 128 failed to follow the same path -- even though both were considered at about the same level in the 1970s. Saxenian finds that the single biggest difference in the two regions was the ability of employees to move from firm to firm in Silicon Valley. That factor, ahead of many others, caused Silicon Valley to take off, while the lack of mobility in Boston caused its tech companies to stagnate and make them unable to compete against more nimble Silicon Valley firms. Saxenian claims that the difference in mobility was simply due to "cultural" differences between the east coast and the west coast. However, the impact was massive. The frequent job changes helped speed up the process of innovation, as ideas flowed more freely, allowing ideas to quickly change and grow and build upon other ideas leading to faster and better innovation. In contrast, employees in Boston stuck with their firms. The firms grew bigger, but slowly, and new ideas didn't flow nearly as easily. There was less direct competition from firm to firm, so firms were able to rest on their laurels rather than increasing their own pace of innovation.
Ronald Gilson found this to be interesting, and followed it up with his own research suggesting that that it had much less to do with cultural reasons and much more to do with the legal differences
between the two places, specifically: California does not enforce noncompetes, while Massachusetts does. Gilson looks at a few of the other possible explanations for the difference and shows how they're all lacking, leaving the difference in noncompetes as being the key difference between the two regions in terms of the flow of information and ideas leading to new innovations. He also explains the history of non-enforcement in California, showing that it was mostly an accident of history more than anything done on purpose.
The problem with all of this research was that none of it really showed how much more mobile employees were in California than elsewhere, so that job fell to some researchers from the Federal Reserve and the National Bureau of Economic Research, who produced some data to back up the findings of Saxenian and Gilson in their report Job Hopping in Silicon Valley
. Their data showed that, indeed, there was much greater mobility in Silicon Valley than elsewhere. Their research further backed up Gilson's suggestion that it was noncompetes that made the difference by showing that other high tech communities in California outside of Silicon Valley also showed greater job mobility -- suggesting it was a California-wide phenomenon.
Finally, to make the case even more compelling, some researchers from Harvard Business School put out some research earlier this year that not only compared the situation in Silicon Valley to Boston, but added a third natural experiment in Michigan
. You see, Michigan used to
not enforce noncompetes, but in 1985, Michigan inadvertently began allowing noncompetes to be enforced again. The research showed that immediately following the change, mobility of inventors in Michigan decreased noticeably, slowing the spread of certain ideas. Their research found that "The networks of small companies so crucial to Silicon Valley's
growth would be less likely to develop in regions that enforce noncompetes."
Noncompetes Are The DRM Of Human Capital
In order to understand how this makes sense, just think of noncompetes as the "DRM" of human capital. Just as DRM tries to restrict the spread of content, a noncompete seeks to restrict the spread of a human's ideas for a particular industry within the labor arena. Both concepts are based on the faulty assumption that doing so "protects" the original creator or company -- but in both cases this is incorrect. What it actually does is set up an artificial barrier, limiting the overall potential of a market. It may not be easy to see that from the position of the content creator or company management (or investors). It's natural to want to "protect," but it's actually quite damaging.
We're already seeing this in the recording industry, of course. The desire to protect has actually limited the market size of other avenues for the music industry to make money. It's held back the ability to use music as a promotional good to build up the overall market
for other tangible goods. In the same way, noncompetes limit the market size of the industry where those noncompetes are enforced. It holds back the ability of firms to innovate. Innovation is an ongoing process -- and the fuel of that process is the continual spread of ideas that allows multiple parties to build on those ideas, try different approaches and seek better solutions. While it may seem scary to a firm that supposedly "risks" losing some of its top employees to direct competitors, that's not necessarily the best way to look at this. What it does is force companies to keep on innovating and keep trying to come up with newer, better solutions to top those competitors. At the same time, that free flow of ideas means that the companies in the space have more fuel with which to attack the problem, rather than quarantining those ideas off in separate bins that can't be connected.
While it may seem easier to "protect" your ideas and your people, what you really end up doing is blocking off your own access to many of the ideas that you need to continue to innovate. You limit the vital mix of ideas to build not just decent products, but great products. Just as DRM has helped to destroy the record labels when competing against more nimble, more open technology -- noncompetes destroy businesses when competing against more nimble, more open technology clusters.