Saturday, December 21, 2013

An Interview with Orly Lobel-- Talent Wants to be Free

Guest Blogger

Kiel Brennan-Marquez
Orly Lobel 

[Yale ISP fellow Kiel Brennan-Marquez interviews University of San Diego Law Professor Orly Lobel  about her new book, Talent Wants to Be Free: Why We Should Learn to Love Leaks, Raids, and Free Riding (Yale University 2013)]

KBM: By way of introducing the book, I'd be curious to hear about your adversaries. Here is how I understood the central claim of Talents Wants To Be Free: value is destroyed when we constrict the fluid movement of highly skilled labor, and when we create disincentives for enterprise and entrepreneurship outside of formal employment channels. This made a lot of sense to me. In fact, it made so much sense, I must say that I found myself wondering who actually takes the opposing view. I guess an incorrigible formalist might say that non-compete clauses, IP assignment clauses, and the like should be respected on traditional contract grounds. And I can certainly imagine a spokesperson for the private sector combating some of your ideas. But among serious academics, taking a public interest approach to the question of how best to govern the highly skilled labor economy, do you encounter a lot of resistance? If so, from whom? On what grounds?

OL: I am happy you find the arguments persuasive! But the reality is that not only big business spokespersons but also quite dominant voices in the academic world truly believe that robust human capital constraints are essential to incentivize firms to invest in R&D and innovation. It's true that William Landes and Richard Posner have admitted about non-competes that "it is not even clear that enforcing employee covenants not to compete generates social benefits in excess of its social costs." And it's also true that we have had as much conversation about what I call "under the radar IP expansion" as we've had about the scope of patents and copyright where we have deep and ongoing scholarly debates. But we've had for a while quite an unchallenged he traditional analysis assumes that absent contractual protections over knowledge, skills and ideas, employers would under-investment in research, development, and employee training because of the increased risk that these employees might leave and use that training as competitors. We see this analysis going strong, from earlier writings such as Harlan M. Blake, Employee Agreements Not To Compete, 73 Harv. L. Rev. 625, 627-29 (1960) through the 1980s with Paul H. Rubin & Peter Shedd, Human Capital and Covenants Not to Compete, 10 J. Legal Stud. 93, 100 (1981) and into the 21st century, for example Mark A. Glick et al., The Law and Economics of Post-Employment Covenants: A Unified Framework, 11 Geo. Mason L. Rev. 357, 357-60 (2002). What is exciting at this moment though is that especially in other fields, in economics and business schools, psychology and organizational behavior, there is a critical mass of evidence to suggest a much richer story and to examine anew how we regulate human capital.

KBM: Very interesting. I can't claim any deep familiarity with this literature, though I did look over the articles you cite. Conceptually, it's hard for me to see how the under-investment thesis gets off the ground, given the intrinsic relationship between employee training and employee performance. If a firm doesn't train its employees, the employees won't—presumably—produce value for the firm. If that's true, the fact that a firm is scared of "talent flight" can't really provide grounds for not training employees. It might cause other dynamics to change—for example, we might expect compensation to increase as time goes on—and reasonable observers may disagree about the desirability of those changes. But a firm that says, "I don't think we should help Jane get better at the tasks she does for us, given that she might end up doing those tasks for another firm," seems implausible on its face. Instead, it seems like the natural result of taking mobility-controls off the table would be a value transfer from capital to labor, as firms fight to retain employees; and because covenants not to compete (and similar controls) have a quasi-monopolistic effect, this value transfer would represent a restoration of surplus back to the "consumer," which in this case would be labor. In other words—regrettable as it might be to hear—I still find your position natural, and your opponents' position silly.

OL: So you are asking me to continue to play devil’s advocate to myself. Which is good, because it pushes me to distinguish my position from more black/white—free/control—approaches.   I do think there exists a tipping point in which no limits on competitive actions, no limits on what insiders can do and use from whatever they’ve learned at a firm will reduce corporate incentives to invest in training and research. In that sense, I am perhaps less radical than is implied in your question: I believe that we already had a pretty good balance historically when we had a narrow definition of trade secrets, patents, and unfair competition and that beyond that we are doing more harm than good by expanding waves of control. And I concede that non-compete enforcement can create some confidence that will impact certain choices about whether to further train or not, but the costs will ultimately outweigh those potential contextual benefits. In other words, I agree with what you write above: most training will occur no matter what, just as in the context of debates about defining trade secrets cautiously, pricing and customer service will occur no matter what, and thereby, anything that is intrinsic to production and business competition loses the justification of restriction [and really, we are having the same debates across all innovation policy fields: what kinds of creativity and invention will we see under narrower scope of protection in copyright and patent law?].
Still, there are some investments where managerial decisions and risk assessment come in: Becker’s distinction between general and specialized skills remains valid. Yes, we will teach Jane how to get better at the tasks at hand, but perhaps there will be one point of production, which is the most secretive and key to our competitive advantage, which we will not teach her. We will minimize exposure and have only few people, say the founder and her partner, whose interests are most aligned with the corporate interests, know that piece of the puzzle. This can inevitably create inefficiencies—firms vertically integrating less people than they would otherwise be optimal, creating artificial boundaries among insiders, Chinese walls between R&D teams and other departments. But the effects, as you are suggesting above, are clearly smaller than the orthodox economic theory suggests, because competition inherently requires training and the incentives to innovate are far more complex than the traditional model has recognized. The distributional effects that you describe in the end of your question—from capital to labor—are real. The empirical evidence shows that in places where mobility is higher, like Silicon Valley and California at large, salaries are higher than in places in which there is a slower tournament for talent. This is the logic of the new antitrust case, the class action that was just certified against high-tech giants which had agreed to not recruit each other’s employees. The workers in the class action are (exactly as you wrote above) the “consumers” who were harmed by the do-not-hire agreements, anti-competitive agreements that  depressed wages by reducing competition for labor.

KBM: Switching gears, your reference to Stewart Brand's famous quip, "information wants to be free," reminded of McKenzie Wark's book from a few years back, The Hacker Manifesto, which opens with the memorable riff, "Information wants to be free (but is everywhere in chains)." Wark sought to develop a Marxist account of the digital information economy, one in which "the hacker class," which creates new ideas, is at war with "the vectoralist class," which controls the infrastructure by which ideas are transmitted. I bring this up because for Wark, class dynamics are very simple, whereas for you, those same dynamics seem more complicated. At times, your account makes it sound like allowing highly skilled labor to move about freely is a winning proposition for everyone—big companies only think it's in their interest to constrict what their workers can do, when in fact, these companies stand to gain as much as society writ large by implementing your proposed reforms. Other times, this seems highly implausible—you're identifying areas where the interests of big companies are overtly antagonistic to the public interest. Which is it? The book occasionally feels like a leftist-wolf disguised as a technocratic-sheep.

OL: This is a key question. It's absolutely true that one of the messages of the book is that there is a far greater zone of win-win than we normally assume for these context. Rather than a Marxist, labor versus management, frame, which, as an employment law scholar I recognize as oftentimes hugely distorting, the book points to many instances of commonality. It looks at some of the best practices of industry leaders and shows how businesses have learned to strategically gain from loss, how they can realize that over time, in repeat games, the occasional loss is outweighed by the positive effects of mobility that spread throughout the industry. At the same time, I care about policy and I think that this is an area in which corporate culture and smart laws are mutually reinforcing. In chapter 9 I discuss why despite all the empirical evidence that shows that businesses are more likely to succeed in a region that supports talent mobility, it is improbable that given the availability of constraints through contracts. I conceptualize the talent wars and the sort of ammunition that firms will use to retain their talent as a prisoner's dilemma: firms would all be better off if some of the strongest arms, unconventional weapons, if you will—such as non-competes which absolutely restrict movement to a competitor—were off the table, the industry would thrive, the talent pool would become richer, the region would grow stronger, and each competitor would be able to recruit in times of need. But given that at any given moment there is a poacher firm and a firm that is poached for talent, competitors engage in a talent control arms race. There is an undeniable element of loss at that moment for the company who see one of its best athletes jump ship (although the loss is usually exaggerated, leading firms to overlook ongoing opportunities that come maintain strong ties with former employees), and so companies can find themselves trapped in a sub-optimal equilibrium that holds back the industry and offers perverse incentives to each actor.

And yes, as I describe in the book, there is a particular cost to independent entrepreneurs within regions that are more controlling. Beyond the reduced movement of talent, the mobility that does exist is patterned: in these regions employees will be far more likely to move to another established employer who can promise to indemnify them than go off on their own and fulfill their dream as inventors. So incumbents indeed have an advantage under the more controlling human capital regimes. There are other distributional effects as well. For example, once we understand human capital controls as adding significant rigidity to labor market searches, we can see how these can disproportionately effect people who are already less mobile. For example, people of color compensate for having a lessened degree of professional ties to begin with. Women, who on average are more constrained geographically because of dual career issues, and can't relocate as easily to avoid breaching a non-compete.

KBM: The second point, about the disparately negative impact of talent-restrictive measures, is very well taken. As for the first point, there still seems to be an elision afoot as you transition from (1) the observation that "businesses are more likely to succeed in a region that supports talent mobility," to (2) the proposition that individual businesses stand to benefit from the enforced labor mobility that you advocate. The first is a truth of the aggregate; the second, a claim about the viewpoint of an individual actor. It still seems like a big business with lots of negotiating power, the kind of business that would be most likely to secure highly restrictive covenants from its employees—due to the desirability of the relevant jobs, the desperation of the relevant labor pool, or whatever else—could read your book, accept your statistical arguments about labor mobility improving the business climate writ large, and still regard your proposals as deeply antithetical to its interests. This isn't a critique. I'm just curious where the battle-lines actually fall. 

OL: I don’t disagree. I do discuss some of these questions in the book in game theory terms, suggesting that businesses could have an optimal win-win equilibrium which they can’t reach on their own without policy which removes the inefficient restrictions off the table. But I also show that these questions of business interests are more complex. Ron Gilson and I have recently had some very useful discussions about this aspect and we are planning to continue the discussion (perhaps at the Yale Innovation Beyond IP conference in the Spring). As Ron would helpfully describe it, even though there is a real prisoner’s dilemma story here – where an industry develops more rapidly when everyone is allowed to raid competitor talent more freely, these are multi-party PD games, the models of which are fragile, and the time frames from each firm’s perspective may be, again in Ron’s terms, rationally myopic. In other words, a company may care more about short term survival or keeping its dominant position in the market than the long term benefits of these repeat games. And, perhaps most directly responsive to your question above about big businesses, there will be firms who either correctly or incorrectly predict that despite the statistical gain of freer human capital regimes, they will be more on the giving (losing) side than the on the receiving end of this repeat game, again, even if they know that the net overall effects on all businesses is positive under a more open regime. I think most firms will be incorrect in this assessment, but I am not denying that there will be losers and, as Schumpeter would say, constructive destruction, in any market. Not all survive, though more (and better) survive under regimes that support competitive and mobility. The bottom line is that businesses will be driven by their self-interested predictions and individual risk assessment whether or not they are good at these predictions and assessments. This is why for most regions and industries there is a role for policy to solve both the coordination problems and the asymmetrical negative effects human capital restrictions have on entrepreneurship, start-ups, newer and smaller companies.

KBM: For my final question, I’ll keep it more within the conventional bounds of legal academia. You make a compelling case that jurisdictions that refuse to enforce non-compete clauses, like California, are recognized as more "innovation-friendly." And the market has responded accordingly. As I understand it, you want to use these jurisdictions as a model for everywhere else—every state should become like California. But I wondered if there might be value in jurisdictional multiplicity, akin to the "states as laboratories" theory of federalism. In other words, why not have some states that are more innovation-friendly, and other states that are more—I'm not sure of the right phrase; "stability-focused"?

OL: Jurisdictional variation has of course some advantages. For me as a researcher, an obvious one is creating a natural experiment. This is part of what has been absolutely thrilling in writing the book. across so many campuses and disciplines, researchers are picking up on these jurisdictional variances to conduct rigorous empirical work. This also why the fields of non-competes laws, trade secrets, and human capital laws, including doctrines about patent and copyright assignment in the employment context, are particularly interesting because unlike core patent and copyright laws, they are largely state law. So states as laboratories is indeed a reality in this context. I've thought about your question quite a bit - about whether there are different regimes that would better serve regions and industries at different stages in their development. The short answer is that of course it is true that there are different lines to be drawn for the optimal balance between control/openness, just like with other line drawing questions in the IP world, for different industries, depending on the type of innovation that happens in that industry (low-tech/high-tech; long-term/short-term etc) and the structure of competition in the industry/region (concentrated, networked, etc). So yes, we could think of different states adopting different regimes. And they have. BUT - strikingly, at this point, we have enough empirical evidence, which I discuss in the book, a lot of which is quite recent, to suggest that states would be unwise to stick with some of their strongest talent controls -- because the positive effects of weakening controls are not only with regard to start-ups and entrepreneurship levels, and not only limited to high innovation industries, but in fact go beyond that with robust findings about job creation and economic growth more broadly.

Kiel Brennan-Marquez is a fellow at the Yale Information Society Project. You can reach him be e-mail at kiel.brennan-marquez at

Orly Lobel is the Herzog Professor of Law at University of San Diego. You can reach her by email at lobel at

Older Posts
Newer Posts