Silicon Valley is the most innovative region in the world. But how did an area of fruit orchards get this way? How can other cities — say, Boston with its Route 128 — replicate this creative energy?
A research paper from Wharton takes a step closer to answering that question by looking at trade secrecy laws and what’s called the inevitable disclosure doctrine. They found that employer-friendly trade secrecy regulations, such as delaying workers from moving to a competitor, do dampen innovation. But not for the reason that current academic literature thinks.
The paper, “Trade Secrets and Innovation: Evidence from the ‘Inevitable Disclosure’ Doctrine,” was authored by Wharton management professors David Hsu and Iwan Barankay and doctoral candidate Andrea Contigiani. Their paper recently appeared in the Strategic Management Journal. Hsu and Contigiani joined Knowledge at Wharton recently to discuss their findings in more detail. (Listen to the podcast using the player above.)
An edited transcript of the conversation follows.
Knowledge at Wharton: Tell us about your research. What’s it all about, and why did you choose the topic of trade secrecy and innovation?
Andrea Contigiani: We were very interested in what drives innovation activities and innovation performance. So far, much of the [academic] literature in this area has focused on patents. Patents are really a key tool that companies use to protect their invention, and governments use them to generate and incentivize innovation.
Of course, the literature has studied this in great detail. But on the other hand, there is another important aspect of IP [intellectual property] law, which is trade secrecy. That may play an important role in driving innovation, so we decided to look at this slightly less understood aspect of the IP environment.
David Hsu: Just to build on that, what’s really important to recognize is when you ask managers – ‘How do you protect your innovative activity? How do you protect all of your investments in research and development?’ — they will first and foremost talk about trade secrets (such as Coke’s secret formula), rather than patents. But the problem with trade secrets [for the purposes of academic study] is that we can never observe them unless there’s litigation or a lawsuit.
Patents, on the other hand, are a very different institution. The idea is that we have a system, much like the rest of the world, in which we give incentives for inventors to disclose their novel ideas. In exchange, we give them a monopoly for — in the United States — 20 years from the time that they disclose their invention.
In the face of this mismatch in observability — researchers can observe patents and study them, while trade secrets are very important in practice but hard to study because they’re a secret — we decided to look at the impact of a shift in the legal regime associated with trade secrecy law and how that might affect something that we can actually observe and study, which is the impact on inventive activity patenting. It’s this situation in which ordinarily we would not be able to observe, but we find some very surprising results with our study.
Knowledge at Wharton: Let’s first define what you mean by trade secrets. I understand that the legal definition has changed over time. Can you get into that a little bit more?
Contigiani: The notion of trade secrecy has evolved a little bit. The one that we have right now is the notion suggested by the Uniform Trade Secrets Act, which was introduced in 1979, and then over time has been adopted by many states in the U.S. Based on that, a trade secret is a piece of information [that gives a company competitive advantage], basically things like formulas, patents, programs, devices and so on. It has two aspects. One, it derives economic value — actual or potential — from being known to outsiders. Second, it is the subject of efforts that are reasonable under the circumstances to maintain its secrecy. This is the legal definition that we have right now.
“While companies may prefer [to keep employees from jumping to a rival], this could backfire in terms of less innovation.” –Andrea Contigiani
There is a set of legal tools that protect trade secrecy, and in the study we look at one specific tool, which is the inevitable disclosure doctrine or IDD. That’s essentially the idea that when you move from company A to company B, just by doing the work required at company B, you might disclose some important information about company A even if you didn’t mean to do it.
Spurred by a watershed case in 1995 in Illinois — PepsiCo vs. Redmond — this doctrine of inevitable disclosure has been tested across states. And this is the empirical setting that we use in our study.
Hsu: Going back to the definition of trade secrets — it’s very broad, as compared to patents. These are things like your customer lists, the way you go about competing in the marketplace, etc. In the popular press, we often think about things like Kentucky Fried Chicken’s secret recipe or a restaurant’s “secret sauce.” That is actually protected by trade secrecy, and the nice aspect of trade secrecy is that there is no fixed duration of protection (unlike patenting).
Now, that doesn’t prevent others or competitors from trying to discover your trade secrets. If you accidentally disclose it, or if your competitors backward-engineer it, that is not illegal. But if your competitors discover it through illegal means, etc., this is not fine. And that’s where we have the legal system come in. By the way, this is very different from formal patent protection, which requires different standards for protection.
Now what’s also very important about trade secrecy law is that it is adjudicated at the state level, not at the federal level (like patents). States have independence in the way that they adopt this particular doctrine or any trade secret law. We actually exploit that fact, because different states come in with their adoption or rejection of this inevitable disclosure doctrine.
The inevitable disclosure doctrine basically says because you as an individual cannot forget your experience or what you know from one employer, even if you don’t have a noncompete clause or any other contractual agreement in place that will prevent you from disclosing information to your new employer, [the company can prevent you from moving to a competitor]. That’s not for all time, but that is for a fixed period of time in a domain of knowledge.
The upshot of this for employers is that if this legal doctrine is in effect in your state, it gives more rights to employers. They can say, “No, you cannot move [to a competitor]. And I’m going to invoke the courts to prevent you from moving.” That is quite a distinct legal regime that we exploit to better understand the relationship between trade secrecy law and individual innovation outcomes. What happens when trade secrecy protection is more employer-friendly and less employee-friendly? What happens to the incentives and innovative profiles of technical staff who live under these different regimes?
Knowledge at Wharton: What did you find out, and how do these types of secrecy laws affect innovation?
Contigiani: Basically, the question that we tried to address here is how the adoption of the inevitable disclosure doctrine across states and over time affects inventors’ innovative activity [as measured by the number and quality of patents issued].
A set of theories would suggest that because companies, after IDD, are less likely to lose their employees, they might invest more in them, and that should lead to more innovation. (Other theories, however, suggest the opposite effect.) A big explanation for innovation that we have in the literature is this idea of recombination of ideas. Recombination happens if people are able to move around [and share ideas]. If people are not able to move around, there is less recombination, and there’s less innovation. That would suggest a negative relationship.
“We find in our setting that stifling mobility can dampen innovation incentives.” –David Hsu
Then there is another theory that we happened to find support for, which is this idea of labor market signaling. If people are not able to move within their domain [or job market, they are less motivated to develop innovation because they won’t be rewarded for it by getting a better job at a competitor — or dangle the higher job offer to the current employer to get a raise or promotion.] This idea also would suggest a negative effect, because essentially there would be this change in the incentive structure of the inventors that might lead to less innovation.
When we go to the data, what we find is that overall, IDD leads to substantially less innovation — that’s the main result. When we try to understand what drives that, we find some evidence regarding the labor market signaling approach. And so it does seem that inventors change their incentives, and by doing that produce less innovation.
Hsu: Remember the IDD, when it is in effect, gives more rights to employers. So you as the employee may want to move to another employer, but with this doctrine in place, that’s going to give more rights for the employer to say no or have some legal recourse to prevent you from moving.
Under that type of legal regime, because you’re more likely as the employee to stick around, the employer may say, “Look, I know you’re going to be sticking around more. Let me invest more in you.” Under that regime, you might expect theoretically that is going to lead to more innovative activity. We don’t find support for that explanation.
There’s theory that suggests the opposite prediction — if you give more rights to the employers, that’s going to dampen innovation. Why? Because innovation requires individuals to circulate their ideas, to recombine those ideas. To the extent that that is dampened, you may not have as much innovative output. That would be one explanation for finding the negative result [that trade secrecy laws favoring companies decrease innovation.]
But [lack of idea circulation is not] the root cause [of a decline in innovation] in this case. Instead, what we find is that now that the employer has more rights to stop you, as the employee, from moving around in the labor market — that realization may give the employee less incentive to innovate because you’re more beholden. You can’t use the labor market to get a better deal in the company for yourself. Think about your own behavior when you try to negotiate better employment terms. You might say, “Look, you guys are not valuing me enough. Look at the outside labor market — they’re saying that I’m worth so much, and they’re willing to give me this better deal.”
Suppose that channel is shut down, and you know that you are going to be more beholden to your current employer. You may be discouraged. You may say, “Well, I’m not going to try as hard because I can’t activate that external labor market for my benefit, so therefore I’m going to put less effort in.” And this is the explanation or the mechanism that we find support for in our empirical work. … The reason for that negative outcome is because these employees are not able … to activate the external labor market to get a better deal for themselves.
Knowledge at Wharton: But given that, what are some of the practical implications of your research? How can people use what you’ve found to actually make a positive change that will not dampen innovation?
Contigiani: In principle, companies might prefer a regime or a situation where they can keep their talent more easily. But at the same time, we find these results that suggest that their human capital becomes a little less productive in terms of innovation in this setting. So while companies may prefer [to keep employees from jumping to a rival], this could backfire in terms of less innovation. This is something that companies might want to keep in mind in terms of location choice, for example, or in terms of deciding their broader strategy.
“We are able to put our finger [on one catalyst] … to [answer] the trillion-dollar question — how do you create the next Silicon Valley?” –David Hsu
Obviously, this would also have implications for policy because policy is generally concerned about driving innovation. While lawmakers in general — or the system — might try to protect companies, they need to be aware that they’re also hurting their innovation performance.
Hsu: It’s very paradoxical in that the first instinct for companies and employers is, ‘Let’s keep our employees under a tight wrap, and that is going to hopefully lead to better outcomes for the company.’ But one big insight of our research that we couldn’t have predicted ex-ante is that paradoxically, in situations in which you allocate more rights to the employer to rein in the labor market mobility of their technical staff, that actually backfires.
This, I think, is consistent with the experience of states like California versus Massachusetts. There was a very famous early study that tried to compare Silicon Valley in California versus Route 128 in Boston. In both locations there are great universities, great scientific staff. But how come there is a disparity in the development of these entrepreneurial ecosystems? California takes off — it used to be an apple orchard there in Silicon Valley — [but Massachusetts was less successful].
In Boston and the greater Boston area, you’ve got great human capital, great universities and good companies — but why the divergence in performance? One legal infrastructure explanation for that divergence is because in California, you had a situation in which noncompete clauses were not really enforced, so the employees could really circulate around. They could find the best match. They could use competing offers to really find the best employment situation for themselves. And that kind of ecosystem really promotes innovation.
What we’re finding is a bit of a parallel. What’s very interesting about the inevitable disclosure doctrine is that even if you don’t have a noncompete or any other contractual agreement in place between the employer and the employee, this is a doctrine that can, as a legal remedy, prevent employees from circulating around. While the earlier literature isolated the labor mobility mechanism for leading to better innovation results, we find in our setting that stifling mobility can dampen innovation incentives. So, this actually does resonate with that earlier literature. I think the implications for not only companies and private managers, as well as perhaps policymakers, is that this is not something that you can easily architect.
What we have to do, probably, is to balance the interests of employees and employers because innovative economies are a very complicated thing. What we put our finger on in the study is an interesting empirical context in which we are able to really pinpoint innovation outcomes and try to isolate the mechanism that leads to that result.
Knowledge at Wharton: Many cities are trying to become the next Silicon Valley with varying degrees of success. For policymakers out there, what advice would you give them so their cities can get a step closer to becoming the next Silicon Valley?
Contigiani: One way would be to try to balance this situation a little better in terms of how high is the cost of moving across companies. The most direct implication would be trying to manage the legal tools in place in a way that keeps the cost low enough, similar to the California situation.
But of course at the same time, that creates other kinds of imbalances, and so the overall situation is probably a little more complex.
Hsu: Think about what is necessary, what is sufficient for an innovative economy. Our paper is meant for an academic audience, but I think that we are able to put our finger [on one catalyst within] a well-controlled setting to [answer] the trillion-dollar question — how do you create the next Silicon Valley? Everyone would like a very innovative economy.
There are so many different moving pieces, usually, when it comes to architecting an innovative economy. Instead of trying to tackle all those different jumble of variables that could lead to innovation, we try to hold everything constant — except for one thing. The one thing that we really try to isolate is the allocation of trade secrecy legal rights in this tussle between employers and employees.
“While lawmakers in general — or the system — might try to protect companies, … they’re also hurting their innovation performance.” –Andrea Contigiani
Now, it’s likely that in order to become an innovative economy, you have to have all those other pre-conditions, right? We talked about Silicon Valley versus Route 128, [both with] great human capital, great universities, great science, great companies there that could train people, great venture capitalists and investors. All these things come in, and we’re still really, in our scholarship work, trying to understand this huge question of, what is necessary? What is sufficient? What if you took out one factor, how important is that one factor? I don’t want to try to pretend that our academic study here solves all those things. Instead, we’re playing the long game in academia. We’re trying to be very rigorous about our approach to nailing one puzzle piece and, hopefully, that can be informative to policymakers and to company managers.
But I think that’s actually a good way to transition to probably the next question that you might think about asking, which is where do we go from here?
There’s been a lot of work that has tried to understand what are the human capital requirements and financial capital requirements by geography to really make our region, our city, our zip code highly innovative? The field in general is trying to move beyond just regional outcomes to isolate down to the individual level, and that is the direction of my own future research. What happens if two research teams, living in different locations, actually come up with the same scientific discovery? What is the likelihood that scientific team A tries to commercialize their discovery, while scientific team B, who came up with the exact same discovery, did not? Through this new research, we’re trying to localize the necessary versus sufficient conditions that are needed to translate scientific advance to commercial outcomes that can benefit both companies and society.
Contigiani: I’ll quickly mention another avenue that I’m really interested in, which is looking into this idea of mobility. What trade secrecy and specifically IDD does is essentially make mobility harder — it’s essentially a cost to mobility. There is another stream of work that looks at how transportation costs affect mobility and therefore innovation. So I am looking at how trade secrecy and transportation costs, by modifying different costs to mobility, jointly affect innovation.
