Academics as a Natural Haven for Open Science and Public-Domain Resources: How Far Can We Stray?
I would like to start with the observation that has been made at this symposium, which is that, until recently, academic institutions have been regarded as a safe haven or safe harbor for open science. I would like to briefly review the rationale for that point of view. Then I would like to ask, to what extent is it possible to export the norms of openness away from the ivory tower to a corporate setting? Finally, I would like to address the issue of how the academic sector tries to accommodate both private and public sponsors of research.
The argument for why open science is such a good fit for academics hinges on two perspectives. The first is the idea that scientists derive great satisfaction from posing questions and solving problems, and to maximize that satisfaction, they should operate in an open environment in which they can share their ideas with colleagues and base their solutions on information they receive from colleagues and students. A second rationale for open science in academia, which is a bit less transparent but equally important, is that open science is really the glue that holds together the academic job market for scientists.
The argument that Paul David put forth in a very convincing fashion is that, by their very nature, scientists work in specialized areas that laypersons—in particular, university administrators and sponsors—cannot know much about.1 As such, there is an information gap between the people who are paying the bills and employing the scientists and the scientists themselves. This presents special problems then for how to evaluate the scientists’ output, what scientists to hire, what scientists to retain, and so forth.
The ingenious solution that openness allows is that other scientists working in the area, the peers, can be used as an information source to evaluate the operative science. Of course, this requires that there be full and open disclosure and that the information banks or the public domain upon which peer review takes place is as complete and current as possible. The scientists themselves derive benefit from peer review. Not only do they get feedback they also derive personal recognition from their peers and establish a professional reputation, and they signal their value to the marketplace. In addition, the information value of peer review is one that sponsors can use in making resource decisions about which scientific programs to promote and which scientists should get grants. Therefore, the open-science norm can be seen as a social equilibrium held together by a number of self-reinforcing factors.
One question we might ask is, does this paradigm work well outside of academics? When I sought to answer that, I came up with two examples. One of them is quite current—the open-software movement. The open-
See Chapter 4 of these Proceedings, “‘Open Science’ Economics and the Logic of the Public Domain in Research: A Primer,” by Paul David.
software movement, I would argue, has many attributes that are quite similar to the academic setting. It started from academic origins in the 1960s and 1970s as collaboration between university and private foundation scientists. It benefited from some visionary leadership on the part of Richard Stallman and Linus Trevault, who had a vision of promoting software in an open-source way; that is, the stock of knowledge regarding software should be allowed to pass unfettered or unrestricted to downstream users and developers. Not only was the spiritual leadership and the institutional infrastructure in place, but there was also a very important contractual relationship established, the so-called general public license, which provided the legal mechanism for having this knowledge pass downstream for use without any restrictions.
Aside from this, there are a number of supporting factors in the open-source movement, many of which have academic types of characteristics. For instance, the developers of software often benefit directly from extending and modifying the software. Often the developers are working with the users. With openness it is possible for the developers of software to delegate to the users decisions as to how to improve the software. After all, it is the users who ought to have a much better idea of which aspects of the software code should be developed and changed. As a result of this interaction between users and developers, there is an interesting and valuable cross fertilization of ideas.
We see the same needs for gratification among programmers. They enjoy sharing their latest challenges in solving software problems, just as academic scientists derive similar benefits. This is important because, like academic scientists, there are probably relatively few people who can appreciate the efforts that programmers put forth and the value of their products. Thus openness—the ability to share new ideas and to document new approaches— allows the workers in this industry to gain personal fulfillment. It also allows programmers, like academics, to signal their abilities in the marketplace so that potential employers recognize their capabilities and value.
One interesting question for economists is to ask, “Would there ever be any corporate support for open software?” They already exist in the marketplace. There are firms like Red Hat, who derive benefit from supporting open software because they can sell complementary products, they can manage the software in such a way to make it more accessible, and they can provide instruction manuals and information to move the software from the developer to the user. There are other firms such as IBM who find it in their best interest to support open software because they can use that open software with their own proprietary hardware and software products to offer a much improved product to the marketplace. So there is a combination of the open product and the proprietary product to produce an even better product. The existence of side-by-side open software and proprietary software brings into question whether open software will survive and, if so, in what segment of the market.
Open software has some inherent advantages over proprietary software. As I mentioned before, it gives rise to greater progress and cross fertilization because it is open and because users of the software give direct feedback to the developers as to the kind of products that they want. A second major advantage, which I have already mentioned, is that it is more pleasurable for a programmer to work in an open-software type of environment. He gets more feedback, he gets more direct gratification from sharing his exploits with his peers, and he is also more mobile in the marketplace because his strengths and capabilities can be signaled more directly to other employers.
There are, of course, advantages as well for proprietary software. One is that proprietary software companies can directly recoup revenues from the sale of their product. They do not need to develop complementary products, as is the case in open software. Proprietary software is more likely to appeal to a general audience because the processes of developing the software, explaining how it is used, and making it more user friendly are activities that a proprietary software manufacturer can afford to undertake as the costs can be recouped from doing so. Open software, however, tends to be more difficult, less accessible, and is a product that in reality is confined mostly to information technology professionals.
Taking all this together, what we would predict, and I think what we are seeing so far, is that there will be a continuing role for open software. Most likely, it will be in the information technology professional segment of the market, but proprietary software will probably exist side by side with open software in a different market segment—the more general user segment. This certainly is an interesting example of how some of the norms of openness we find in academics may survive in nonacademic or corporate types of settings.
Another example I will talk about briefly is the experience of the early days of Silicon Valley in the 1970s and 1980s. During that period, Silicon Valley was an exceptional region in the sense that there was an unprecedented era of technological progress and innovation. It was characterized by free exchange of information whereby
scientists and researchers at different firms were allowed, and in some cases even encouraged, to share information with their colleagues at other firms, either informally over coffee or after work, or in formal settings at conferences and seminars. The information exchange among firms resulted in a considerable amount of cross licensing of technologies and ideas.
This unusual degree of openness was supported by several factors. One was an academiclike preference for disclosure. Scientists and engineers were an integral part of the commercial success of the firms in this fledgling industry. In addition, scientists and engineers had a very strong allegiance to their professional academic affiliations, probably more so than the allegiance that they had to these young upstart firms that were very small, sometimes with just 50-100 employees. As a result, given the importance of the scientists and engineers in this industry, they demanded and received as an employment condition the right to operate in an open environment where they were free to exchange ideas with colleagues at other firms. They were relatively free even to move from one firm to the next and take some of their intellectual ideas and property with them. Although the firms in this industry perhaps did not like that, there was little they could do if they wanted to attract and maintain the very best scientific and engineering talent available in Silicon Valley at that time.
There were additional factors that gave rise to openness. One was that Silicon Valley benefited from what economists would call agglomeration economies, meaning that there was a large concentration of companies in Silicon Valley working on similar or complementary products. Coincident with that was a very large, well-trained workforce of scientists and engineers concentrated in the valley at nearby schools in the San Francisco Bay area such as Berkeley and Stanford. This meant that information flows between firms and between colleagues at different firms were very easy Also, it was easy for scientists and engineers to move from one firm to another in the valley. All of these factors conspired to produce, at least for some period of time, a very open environment in a corporate type of setting.
What conclusions can we draw from this? The first conclusion is somewhat reassuring. One can find examples where degrees of the open norm do exist outside of academia. The second is that in each of these cases in which openness did survive the corporate settings had some striking similarities with academic settings. These were all settings in which the primary players in the industry were the people doing the scientific and engineering work. These workers demanded—and, because of their importance received—special treatment, in being allowed to operate in an open environment. They were sometimes even encouraged to exchange ideas at professional meetings and to consult with colleagues at different firms. Given the special circumstances of these examples, there also is a negative message I think which comes from this. I would not expect that the norms of openness are likely to overcome the proprietary norms of most corporate settings, except in exceptional cases where the nature of the industry is such that it resembles an academic setting.
I now would like to address the issue of how universities can accommodate both public and private sources of research support. The description of the scientific open norm is somewhat more of an idea than a reality. In reality, academic institutions have for some time been facing increasing pressure to privatize. There are a number of explanations for this. The most fundamental one is that public funds are in short supply. Not only do we rely on public funds to support such admirable goals as research in public goods, but there are a whole host of other public goods such as education, welfare reform, and national defense that likewise are deserving of funding and are competing for scarce federal dollars. In addition, relying on public funding, while in principle seems like a good way to solve a lot of the proprietary concerns that one incurs when relying on private funding, nonetheless, is not a costless activity. Some analysts estimate the cost of raising one additional dollar of tax revenue ranges between 30 and 80 cents. This reflects the distortionary impact taxation has on individual employment and investment decisions. So again, the vehicle of relying on public funding in some cases is a fairly expensive and costly way to go. It should not surprise one that universities face a growing gap between their research desires and the quantity of public support for research. Universities need to fill this gap. Increasing reliance on private funding is the obvious solution.
Other factors, including the passage of the Bayh-Dole Act have created the infrastructure for universities to transfer intellectual property to the commercial sector. Coincidentally technology shocks in biomedical and computer science research have enhanced the commercial value of university research. This has induced universities to seek and obtain greater private research funding.
In addition, one can argue that the presence of private funding is likely to be disruptive to the fragile social equilibrium that I described earlier, which supports the norm of openness in academics. A number of speakers have described already how that process would unwind. However, I would like to point out that, although there are some obvious challenges to academic institutions to accommodate various sources of funding, there is also a silver lining. Private corporations, such as biotechnology firms who wish to establish a capacity to apply basic research for their own business, may find it worthwhile to strike alliances with universities and to establish their own research groups. To the extent that they wish to do so, they may be forced, if they want to get into that line of business, to accept some of the norms of openness that go along with academic research. They may also find that accepting these norms can be beneficial to them. It allows them to commit to do scientifically objective research. It gives them an advantage in attracting the best scientists and in establishing a reputation for being a leading technology firm. These opportunities aside, some very challenging steps remain for academic institutions to take to successfully accommodate different funding sources.
In concluding, I make three suggestions regarding strategies universities might undertake. First I suggest that universities carefully manage the portfolio of private and public research they undertake. Here, they could learn a lesson from accounting firms. Accounting firms have learned recently that packaging auditing and consulting services at the same time to the same clients is a bad idea. This situation positions the accounting firm in a huge conflict of interest. It is difficult for the auditing arm of an accounting firm to issue an honest statement about the financial health of a client, knowing that in doing so, it may risk losing the lucrative management consulting business that it has with that client. We have seen in recent months some of the abuses that can occur. Accounting firms have learned that it makes sense to sell off the management consulting activity to an independent firm thereby breaking up these two activities. Why? Because these two activities when grouped together just do not mesh. They present such perverse incentives that one could not possibly expect one firm to perform these contradictory activities in a satisfactory way.
This principle applies to the university as well. Universities should discipline themselves to reject private research that would enlist their advocacy or that would restrict their ability to disclose research findings. They should separate out research that hinders the universities from providing public education and research. One should apply the same principle to individual faculty. Faculty should not be asked to undertake multiple tasks, which inherently conflict and interfere with each other.
A second, and related, suggestion is universities should adopt job-related compensation. Universities undertaking research from different sponsors are going to ask their faculties to engage in various activities. It makes sense to compensate a faculty member based on his performance on the tasks he undertakes. Fine-tuning compensation to the particular job each faculty performs allows the university to target salary and resources to the most valuable areas. It also allows the university to compensate faculty according to whether their research is privately or publicly sponsored.
My final suggestion concerns the transfer of research and technology to the corporate sector. I recommend the university tailor revenue-sharing arrangements to suit the type of research transfer it undertakes. Given the different research output the university may transfer, it is unlikely that one arrangement, such as exclusive licensing, will fit all applications. Instead the university should develop a menu of transfer mechanisms conditioned on two important factors. One would be the corporate sponsor’s requirements for cost recovery and exclusive access to research findings. The second factor would be the opportunity costs to other researchers of having incomplete or delayed access to the research findings. Transfer agreements should reflect these factors in computing compensation for transfer of research findings to corporate sponsors. Transfers permitting greater circulation of research to the public domain should be performed at lower cost to the sponsor.
Some of the symposium speakers have suggested one cannot expect universities acting alone to be faithful agents for the public good. If this is true, one might establish a standard for sharing arrangements. This would prevent a “race to the bottom” where universities offer overly attractive transfers to compete against others for private funding. Standards should, however, provide universities enough flexibility to tailor transfers to different types of research.