University-Industry Cooperation

What about the contribution of universities in Japan, the world's second-largest Internet market? Japan's recent blockbuster contribution to digital information technology, NTT DoCoMo's i-mode protocol to make mobile phones Internet compatible, was developed almost entirely in the laboratories of NTT and associated companies. Satoshi Koike, President and CEO of Netyear Group Corporation, and Joichi Ito, CEO of Neoteny, both say their companies are not backing any university-based entrepreneurs or, with one exception mentioned below, any companies based upon ideas originating in universities. Neoteny and Netyear are two of the leading Internet business incubators in Japan that have associated capital funds.

Indeed, with the exception of the work of professor Jun Murai, the "Internet samurai" who heads the Widely Integrated Distributed Environments (WIDE) Project at Keio University's Shonan Fujisawa Campus (SFC), most Internet-savvy persons are hard pressed to think of examples of discoveries originating in Japanese universities that have become widely used in the Internet or related businesses. The WIDE Project encompasses, among other things, Japan's first computer network; Japan's first indigenous commercial Internet service provider, which began operation in 1992 (Internet Initiative Japan -- IIJ -- actually a commercial spin-off from the non-commercial WIDE Project); Internet-based education; Internet Protocol version 6 (IPv6), which is becoming the world's new Internet protocol; and an Internet car that, if successful, will integrate global positioning, mobile Internet, and privacy technologies. (See sidebar) Other SFC projects are also helping companies advance in Internet communications. One of these, based upon a venture company founded by a recent SFC graduate, has developed software and pulled together component manufacturers to enable the first launch of mobile phone email service in Japan.

Certainly there are other Japanese university researchers besides those at SFC who are doing important Internet related research. However, it is beyond the scope of this article to survey such research. But assuming the truth of the impressions that, with the exception of Keio's SFC, Japanese universities are making relatively minor contributions to commercial IT, we should ask: Why is this so? What are the prospects for improvement? In particular, what can be learned from existing examples of successful cooperation? How does this affect investors who are interested in supporting promising Japanese Internet entrepreneurs, particularly investors contemplating cooperating with university researchers in the hope of accessing untapped talent and information? These are the issues the remainder of this article addresses.

ROOTS OF THE PROBLEM

Insufficient Manpower
Inadequate human resources is more serious in Japan than in the US. Until the mid-1980s, the Ministry of Education (the Monbusho) rigidly controlled the number of slots for faculty, graduate students, and undergraduate students, even at the subdepartment level in National as well as private universities. Quotas for faculty and students in each university remained high in subjects such as maritime and mining engineering, where demand for graduates had long been declining, while it was hard to increase positions for faculty and students in new fields such as computer science. Even today, faculty appointments in national universities, which receive the lions share of R&D funds, must be centrally approved by the Monbusho. Furthermore, inter-department rivalries have hindered the expansion of departments dedicated to software engineering, computer science and even applied mathematics. Instead, interested students and faculty often have to enter these fields as extensions of research in departments of electrical engineering.

Until recently, Japanese government support for computer-related R&D tended to focus on hardware rather than software development. One exception was MITI's Software Industrial Generalization and Maintenance Aids (SIGMA) Project in the late 1980s to promote technology transfer to industry and standard interfaces to enable software interoperability. However, SIGMA produced few significant technological or commercial results. Only in the past three years has the government begun to promote Internet projects. But compared to the US, where a significant proportion of government research grants are used to employ and train young researchers, it is more difficult to use Japanese government grants for such purposes. As a consequence, most Japanese software engineers are trained on the job in hardware manufacturing or end-user companies, where they acquire skills that often are narrowly tailored to a particular company's needs.

Concerning the other side of the spectrum of skills necessary for IT entrepreneurial success, there are few business schools in Japan. Furthermore, most large Japanese companies do not know how to make good use of a business education. In sharp contrast to the majority of new graduates of many prestigious US business schools who are forming their own companies, for most graduates with Japanese MBAs, the degree is little more than a ticket to transition from an engineering to a management career track in large companies where lifetime employment and seniority based promotions are still largely the norm. Keio University's business school is one of the very few with a US-style program and a significant emphasis on entrepreneurship.

"It is not just a numbers problem, however" says Joichi Ito. "Universities don't teach the right things. Students are not taught to think on their own. There is no interdisciplinary perspective among most graduates and thus little sharing of ideas between technical and business graduates." A common refrain from heads of IT businesses is, "Professors have no business sense or experience. What they say is irrelevant to Internet businesses. Professors and students lack entrepreneurship and lack the incentive to be entrepreneurs."

Low Incentives for Cooperation Stymie Entrepreneurship
Two interrelated factors, one concerning career incentives and the other mechanisms for university-industry cooperation, may partially explain the pervasive lack of entrepreneurship and business sense among professors and students. The first rung on the academic career ladder is a research assistantship within the "kouza" of a particular professor. A typical kouza consists of one professor and his disciples, one associate professor, perhaps one assistant professor, and finally one or two research assistants. Promotion depends upon keeping in the professor's good graces and waiting until a vacancy opens up higher in the hierarchy. If anyone gets off the ladder to spend a year or so in a company, it is nearly impossible to get back on, although the Law to Strengthen Industrial Technology enacted in April 2000 may change this. Furthermore, salaries and a minimum level of R&D funding are guaranteed to all persons in the kouza. Until the new law, there were strict limitations on outside sources of income for national university faculty members. Thus, economic incentives to do any sort of entrepreneurial activities (e.g., management, consulting, or off-hours R&D for a startup company) were low, at least for faculty in national universities.

As for the second factor, university-industry cooperation in Japan is difficult for a variety of reasons. The two officially sanctioned mechanisms for companies to support research in national universities in exchange for data or IP rights are "Commissioned Research Agreements" and "Joint Research Agreements." Under these agreements, corporate funds must be disbursed through the Ministry of Finance, which takes 30 percent overhead and then, often after a delay of several months, passes the money to the university. Until the April 2000 law, all money had to be disbursed on a fiscal year basis (i.e., on April 1) and spent by the following March 31 (see sidebar). However, use of such funds for travel is still limited, and only a low stipend can be paid to graduate students and post docs working on company-supported research. Concerning control of inventions that may arise under the research, the most the company can obtain is co-ownership with the university of patent rights. This means that if the company subsequently wishes to transfer its patent rights to a third party, it needs permission of both the university and the Monbusho. For Internet venture companies whose business strategy may include merger or partial buy-out by another company, this may significantly limit the value of such patent rights.

For all the above reasons, companies and faculty usually prefer that industry support come in the form of donations. Officially regarded as charitable gifts, donations are in fact the principal way industry supports research in Japanese universities in return for data, IP, access to new graduates and the goodwill of the professor. However, it is illegal to formally link donations to a promise by a national university professor that he will transfer IP to the company. (In Japan, professors retain the right to apply for patents on their inventions arising under donations, and they are free to transfer this right to companies.) Also, it is illegal to link donations to a detailed cooperative research plan. Furthermore, donations above 5 million yen (about $50,000) annually require special permission from the university president. Therefore most donations are kept below this relatively low level. As in the case of Commissioned and Joint Research funds, donations to national universities can be used to pay only a meager stipend for graduate students or post-docs, and they cannot be used to augment salaries of tenure track (i.e., full-time) faculty.

Japanese Universities Playing Catch-Up in Encouraging Entrepreneurship: the Role of TLOs
US universities encourage entrepreneurship in ways that most Japanese universities as yet cannot. US universities have the right to patent, license, and collect royalties on almost all inventions made by their faculty and students. Therefore, they have strong incentives to find the best possible companies to develop new technologies so that the university and inventors will benefit through high royalties and follow-up collaborative research with the companies. Most major US universities have technology licensing offices (TLOs) responsible for carrying out these technology transfer activities. TLOs in major US universities have the incentives and resources to educate faculty about the benefits of IP protection and technology transfer and to encourage entrepreneurship. They also have the resources to negotiate favorable development and royalty commitments. Finally, often with the backing of alumni and local governments, an increasing number of universities are mobilizing capital and providing incubator services for faculty with promising technologies who want to form their own companies. Thus, paradoxically, the fact that universities rather than individual inventors hold IP rights may encourage entrepreneurship.

In contrast to the many obstacles to corporate support of Japanese university research, it is easy for companies to support research in US universities. There are few restrictions on using corporate funds to employ graduate students or post-docs or even to pay tenured professors' salaries. In return, companies can receive data and, in most cases, exclusive IP rights to inventions arising under their support. This is one reason that incentives for companies and university researchers to collaborate are higher in the US than in Japan, and this contributes to the lack of entrepreneurial spirit among faculty and students in Japanese national universities.

National universities account for most of the "top flight" academic R&D centers in Japan, for example the University of Tokyo, Tokyo Institute of Technology, and Kyoto, Osaka, Tohoku, Tsukuba, Nagoya, Kyushu and Hokkaido Universities.

However, Keio, Waseda, and Sophia Universities also have significant research capabilities. But they are private universities, and therefore most of the restrictions associated with Commissioned and Joint Research Agreements and donations do not apply. It is no surprise that Jun Murai's WIDE Project, which until recently has been funded almost entirely by private companies rather than government grants, is being carried out in association with a private university. It would be nearly impossible to mobilize the resources for such a project in a national university.

In almost all Japanese universities, faculty members retain the right to patent and transfer their inventions. However, most have neither the resources nor inclination to find (or create) companies that are committed to developing their discoveries. Most academic discoveries are either never transferred to companies or are transferred essentially for free to companies with which the inventors have longstanding relationships based upon donations or hiring of students. Lacking obligations to develop these discoveries or to pay royalties, companies often ignore such discoveries unless they are obvious "home runs." Japanese national universities have no independent legal status, and thus are not able to manage inventions by their faculty or students. The 1998 Technology Transfer Law partly got around this problem by authorizing universities to establish semi-independent technology transfer entities (aka TLOs) that can sell or license inventions and distribute royalties to inventors and the university. However, academic inventors are not obligated to assign their inventions to TLOs and can continue to transfer their inventions directly to companies. Anecdotal evidence suggests that inventors often turn to TLOs only when an invention has no takers (i.e., for less attractive discoveries).

No Japanese TLO is close to covering costs with license revenues. As of April 2000, the government had authorized 14 TLOs (see sidebar). A few of the larger ones have filed patents on over 50 inventions. However, as of February 2000, no TLO had issued over three licenses. It appears that information- and communications-related inventions account for 10 to 20 percent of TLO patent filings. There has been at least one software license.

Japanese TLOs realize that, barring infusions of additional government money, survival depends upon being able to convince many faculty to transfer their inventions to the TLOs. TLO officials assert that younger faculty members (especially those familiar with how technology transfer works in the US and who have not yet established longstanding ties with particular companies) are willing to transfer most of their discoveries to TLOs. Also, some TLOs are trying to structure cooperative research relationships between faculty and companies and to set up venture companies to develop promising inventions of entrepreneurial faculty.

Social Factors Undermining Entrepreneurship
The cost of failure for someone willing to swim against the current and join a venture company is high in Japan. In a nation where lifetime employment and seniority-based promotions are still the norm, subsequent employment opportunities are limited if the venture fails. Also, failure is regarded not as a valuable learning experience but as a sign of ineptitude or moral turpitude. Family members are likely to be ostracized. Credit ratings are ruined. Obtaining a housing mortgage or renting an apartment becomes impossible. Attitudes towards lifetime employment in big companies are changing, especially among younger people. But persons who are willing to turn their back on this ideal still constitute a small minority. Therefore, the April 2000 Law to Strengthen Industrial Technology is important, because it permits national university faculty to have an official management role in venture companies formed to develop their discoveries and to be appropriately compensated by such companies.

Silver Linings and Guidelines for Investors
Why is Keio's SFC succeeding? Since its founding at the dawn of the Meiji reformations, Keio University has been oriented towards business rather than government and administration. From the time it was established in 1990, SFC has encouraged students and faculty to develop links with business. Also, the two main faculties at SFC -- Policy Management and Environmental Information, and the Graduate School of Media and Governance -- are heavily oriented towards information technology. To hear Nobuo Saito, dean of the Faculty of Environmental Information, and professor Murai talk, the basic goal of SFC is to integrate hard-core information science with business, humanities, and the social sciences. Forty percent of the faculty have computer backgrounds. Thus, SFC was well positioned to ride the Internet wave of the 1990s.

SFC also has never had department or kouza systems. Faculty vacancies are filled according to individual qualifications and perceived social needs for practical research. As noted above, it is much easier for companies to support research at Keio campuses than in national universities.

Still there are rigidities at SFC. There is hesitancy to increase research manpower by using industry funds to support graduate students, partly due to a nationwide tradition of graduate students being self-funded and partly out of concern that year to year fluctuations in industry support may make it difficult to guarantee stipends. Also, SFC's orientation is still toward big companies. SFC would like to establish a fund based largely upon alumni contributions to provide seed money and incubator services to SFC spin-offs. However, some in the university's central administration object that this would detract from the university's main mission of teaching and education.

Can SFC's success be emulated by other universities, particularly national universities? The answer is probably a qualified "yes." However, the system for companies to support research in national universities is still too rigid, although the recent reforms in the Commissioned/Joint Research systems are significant steps in the right direction (see sidebar). Furthermore, the TLO system of technology licensing is still too weak to make licensing or cooperative research the cornerstone of a successful system of university-industry collaboration. Thus, the WIDE model of large-scale cooperative research with established companies probably will not work in national universities. However, the new law that sanctions faculty playing active roles in venture companies seems to be a true breakthrough. Now the formation of venture companies, free of funding and employment restrictions, seems to be a viable way to develop promising university technologies. However, the success of this model depends upon adequate numbers of skilled entrepreneurial people being willing to join venture companies. "The fundamental challenge," says Sachio Senmoto, CEO of entrepreneurial telco eAccess, "is lack of entrepreneurship in society as a whole."

One point to consider for investors willing to seek out (latent) entrepreneurs with good ideas in universities: a growing number of organizations can help investors find such persons, although their networks are far from comprehensive. Among these are the TLOs, many of which serve more than one university. Some private companies are becoming technology brokers. For example, Recruit recently helped a professor of engineering at Yokohama national university license software to JR. Venture capital and business incubation companies can provide assistance by leveraging capital, evaluating technologies, advising on business plans, marketing, etc. Sometimes it may be necessary to work with companies that have already received rights to a professor's inventions. For example, Neoteny is working with a major electronics company that received rights to Internet security-related patents from a university inventor. As is often the case, the large company did not know what to do with the technology. Now Neoteny is trying to promote the development of these technologies, perhaps by a third party. One strength of Japanese university-based venture companies, particularly those involving a respected professor, is that many researchers are often willing to collaborate with the company even if they are not located in the professor's university. In effect, the company becomes the commercialization vehicle for an extended network of researchers, many of whom do not have an equity stake in the company. Under the April law, venture companies should be able to compensate members of such an extended network for venture-related work, so long as it is after hours.

The bottom line is that the SFC experience shows that Japanese universities can make significant contributions to IT. However, further progress will depend on changes in the mind-sets of students, faculty, and companies; additional administrative and intellectual property reforms; and the ability of companies to find creative ways to mobilize university researchers to do commercially relevant work. Venture companies can help bring about these entrepreneurial changes, but they will never flourish as long as the system of lifetime employment remains strong.