Consider Silicon Valley, which has become the most entrepreneurial area not only in the United States but also in the world. Not only are formal alliances commonplace among the local firms, it is also, arguably, one of the most dynamic labour markets in the world. It is not unusual for your engineers or executives to switch allegiance to a competitor down the road, taking their knowledge with them. Here, the culture is one of loyalty to the profession rather than to a firm. Equally, there is nowhere else on the globe that can compete with the number of start-ups which successfully launch and, more importantly, stay in business and go on to be global leaders in the industry. But what drives this collaborative environment?
Silicon Valley’s vibrancy can’t all be pinned down to one factor. Scholars have tried linking its success to a combination of initial conditions, that kick-started a favourable dynamic decades ago, and a multitude of agglomeration benefits that continue to sustain this virtuous cycle. However, one critical characteristic of Californian is that firms simply cannot enforce non-compete agreements on their employees – as the state doesn’t recognise them legally. Contrast that with the U.S. state of Massachusetts, where courts enforce non-compete agreements – one of the reasons for the significantly lower inter-firm employee mobility and a resulting decline in the once-vibrant information technology cluster along Route 128 in the Greater Boston area.
In regions where the legal system does allow such documents , asking employees to sign a non-compete agreement should, in principle, give employers the reassurance that recruits will not leave on a whim taking sensitive knowledge to a competitor. One would think this practice would make the firms more willing to invest in expensive R&D and cutting-edge human capital. But, what if the very idea of having to sign such an agreement came in the way of your being able to attract the best talent in the first place? What if the flexibility of being able to switch employer was so important that a potential employee avoided going to work for companies in regions where non-competes were prevalent? And what if this ultimately translates into certain regions, and the firms located there, being less competitive economically?
My research paper Regional disadvantage? Employee non-compete agreements and brain drain (co-authored with Matt Marx of MIT Sloan School of Management and Lee Fleming of the University California, Berkeley) examines the issue of how allowing legal enforcement of non-compete agreements can affect a region’s ability to retain talented engineers. In initial analysis leading up to this research, we found that states enforcing these agreements do indeed face a greater extent of “brain drain”.
To establish that this correlation represents a causal relationship, our follow-up analysis relies on an interesting “natural experiment” we stumbled into. In particular, we looked at what happened when unrelated regulatory changes associated with anti-trust policy in the U.S. state of Michigan led to the region’s law inadvertently being changed to allow the imposition of non-compete agreements within the region’s firms. Our data shows this change was accompanied by a sudden increase in emigration of successful inventors – those that had previously filed one or more patents while residing in the state. Further, the most affected workers turned out to be the most creative and collaborative types, as they were more likely to emigrate to states that did not enforce non-compete agreements so that they wouldn’t feel tied in to a specific employer for their entire careers.
While there is no readymade public dataset on migration behaviour of individual engineers across regions, we used inventor records from the U.S. Patent and Trademark Office (USPTO) to construct employment histories for all inventors who had registered one or more patent with the office between 1975 and 2005. We then used this data to examine emigration patterns for the patenting inventors for the state of Michigan. Our particular focus was on comparing emigration dates prior to versus post-1985, as 1985 was the year that non-competes became enforceable. What we found was that, following the policy change, the relative risk of workers likely to leave the state of Michigan to go to other states where non-competes were not in place was twice as high as in states that continued not to enforce non-compete agreements. In particular, we observed that inventor types who had impact in their respective industries (those actively registering patents) demonstrated a relative risk of emigrating 186 percent higher than inventor types in other states who were not bound by non-competes.
Inventors who were particularly collaborative showed a relative risk of leaving the state 237 percent higher after the policy change than collaborative inventor-types in other states. This can probably be explained by their being more likely to hear about job vacancies as well as by the likelihood of being known outside their firm and receiving outside offers. From the point of view of Michigan, this would be a particular worry given that these indeed are the very people it would normally like most to retain to enhance the state’s future innovative capacity.
Our paper provides plenty of additional analysis to make a convincing case that the effect is driven by the non-compete policy change and not other factors that happen to be correlated with it. For example, while Michigan is known to be a state with an over-representation of the automotive industry (an industry which has been in decline for several decades), this was not a factor which led to the brain drain. Even removing data specific to the automotive industry, the emigration levels remained constant. Talent was clearly leaving the state because of the policy change, regardless of industry.
From our findings, there is clear evidence that talented workers feel a pull towards states where they can change employers freely and take their creativity with them. In today’s collaborative business environment, this raises some questions about whether non-competes really are the best tool for protecting intellectual property and recouping technological investments even from the perspective of the companies involved. It is important to recognise that these are a double-edged sword. Preventing employees from going to work for a competitor could not only hurt your competitor, but it could be hurting your own business too if one looks at the bigger picture. Because finding the right match between an employee and employer is a little like the dating game — some are matches are made in heaven, some are not — and it is beneficial to be able to break the relationship as easily as possible, should the need arise. If, however, your own employees and your competitor’s employees are unable to find the right match for fear of breaking a contract, the only outcome is a zero-sum game. In this situation, the degree of experimentation between employees and firms becomes that much harder, leading to an inefficient labour market.
As well as being detrimental at firm level, the consequences are felt across the industry and region. On the whole, states who choose to enforce non-competes could be doing more harm than good for their regional development by reducing labour market mobility. As my colleagues Bruno Lanvin and Paul Evans note in their recent article, talent attracts talent. The challenge, therefore, at policy level is to ensure that the best talent is attracted in the first place and that a regional disadvantage isn’t brought about by encouraging the most skilled workers to leave for more relaxed states in terms of job mobility, and where enterprise and growth are encouraged. It is no surprise that many regions have of late been debating whether they should move away from allowing policies like enforceable non-competes that might come in the way.
At the firm level, there is much to be said for being more selective in relying on non-compete agreements. You could consider other ways of protecting your intellectual property – such as relying on more specific legal provisions like patents and trade secrets. If your region’s policy does enforce these, there might still sometimes be a case for your firm using non-competes. But it is crucial that the potential downsides have also been recognised and factored in before you reach such a conclusion.
This article is republished courtesy of INSEAD Knowledge Copyright INSEAD 2014.
(Jasjit Singh is Associate Professor of Strategy at INSEAD.)