ACKNOWLEDGMENTS: Dr. Kneller's research was assisted by a grant from the Abe Fellowship Program of the Social Science Research Council of the American Council of Learned Societies with funds provided by the Japan Foundation Center for Global Partnership; and also by the cooperation of many Japanese scholars and government and business officials who kindly provided information for this study.

BIOSKETCH: Dr. Kneller (J.D. Harvard, M.D. Mayo, M.P.H. Johns Hopkins) worked in the U.S. National Institutes of Health from 1988 to 1997, first in cancer epidemiology, then in science policy at the Fogarty International Center and finally in technology transfer at the National Cancer Institute's Technology Development and Commercialization Branch. In 1997 he received an Abe Fellowship to study technology transfer in Japan. In 1998, he became a professor in the Research Center for Advanced Science and Technology (RCAST) of the University of Tokyo where he continues research on technology transfer. Dr. Kneller can be reached by e-mail at: kneller@ip.rcast.u-tokyo.ac.jp

---

NOTE: U.S. dollar equivalents of Japanese yen amounts are estimated using the Organization for Economic Cooperative Development's (OECD's) "purchasing power parity indices" listed in Reference 5. For 1996, this index was 177 yen/dollar (NSB 1998).

---

Japanese professors usually retain rights to their inventions, but some publicly-funded inventions must be transferred to the nation and thereby face obstacles to commercialization. This leads professors to pass discoveries informally, without documentation or development or royalty commitments, to companies that have given donations and employed their students. The companies rarely develop such inventions. A new law to encourage the establishment of technology transfer offices (TTOs) may fail because incentives to switch from the informal system are unclear and some TTOs charge high membership fees, discouraging their use and limiting the pool of potential (particularly small) licensees. Unless universities can require faculty to assign inventions to their TTOs, Japan's best hope may be to eliminate compulsory transfers to government and empower inventors to be effective transfer agents who can insist on contractual development commitments.

There is widespread consensus in Japan today that relations between Japanese universities and industry should be improved in order to promote the development of high technology products and industries. Legislation to promote university-industry technology transfer was enacted on 6 May 1998 (1998 Technology Transfer Law).

This paper first clarifies how companies have traditionally obtain intellectual property ("IP") rights to university discoveries, particularly how laws and procedures dating from the 1970s which still govern technology transfer today create incentives for companies, university administrators, and university researchers to transfer technology in ways that are less formal and transparent than the manner in which U.S. university technologies are transferred under the "Bayh-Dole" Regulations (37 CFR section 401.1-401.16). The paper then analyzes how the new legislation will affect university-industry technology transfer. Finally it examines how technology transfer practices interact with other factors to influence the overall nature of university-industry cooperation and the development of high technology industries in Japan. This analysis should be of interest to persons concerned with science policy as well as companies and individual investors interested in cooperating with Japanese university researchers.

The U.S. and Japan lead the world in government support for non-military research and development (R&D) according to official statistics.^{( 1)} The level of this support and that for university R&D are shown in the following table:

Table 1: Indices of Government Support for R&D in Japan and the U.S.

Source: NSB 1998.

III. Overview of Technology Transfer in the U.S.

In the U.S. prior to 1980, there was no uniform policy regarding IP rights to university discoveries made with U.S. Government support. Policies differed according to the laws or policies applying to each funding agency. In general, IP rights to such discoveries vested in the U.S. Government, unless the funding agency waived its rights (Eisenberg 1996). Agencies tended to license discoveries on a nonexclusive basis. In 1980, fewer than 250 patents were issued to universities per year (COGR 1993), only a fraction of which were for inventions made with Government support (Latker 1977).

1980 amendments to U.S. patent law (Public Law 96-517) and the implementing regulations issued in 1987 (commonly known respectively as the Bayh-Dole Act and the Bayh-Dole Regulations) allow companies and non-profit organizations, including universities, to claim world-wide patent rights on inventions made under U.S. Government grants and contracts. The Bayh-Dole Regulations also require universities to establish procedures to ensure that university employees inform them of inventions soon after they are made. Because nearly all U.S. universities require their employees to assign to them rights in employment related inventions (Etzkowitz 1994), the Bayh-Dole Act and Regulations allow universities to take full control of IP rights in inventions arising from Government support. Therefore, universities have strong incentives to ensure that discoveries by their employees are transferred to companies that will effectively develop and commercialize them. Technology licensing offices (TLOs) have become the focal point for technology transfer activities in U.S. universities.

For a variety of reasons, U.S. universities generally license rather than assign rights to companies (see Kneller 1998). In other words, they retain ultimate ownership of IP created by their employees. Thus, licenses are the cornerstone of the transfer of IP rights from U.S. universities to industry. This is true even in the case of "sponsored research agreements" under which a company agrees to support a particular line of university research, usually in return for scientific information and options to license inventions that may emerge from the sponsored research. It is also true concerning the formation of "start-up" or "spin-off" companies to which universities usually exclusively license the key technologies the start-ups will develop. In 1995, over 5000 licenses from U.S. universities and non-profit medical research institutions were generating over $480 million in license royalties. Direct corporate support for R&D in these institutions was over $1.6 billion, largely through sponsored research agreements. Over 190 start-up companies were formed bringing to over 1400 the total number formed since 1980. (AUTM 1998)

This summary of the U.S. technology transfer system provides a basis for comparison with the Japanese technology transfer system.

Although Japan has 565 universities, only 385 offer graduate courses and only 274 have Ph.D. programs (NSF 1997a). Even among these 274 institutions, research excellence is heavily concentrated in a small number of universities (Irvine 1990; Yamamoto 1996; Kei-i 1984).^{( 2)} With a few exceptions (see Yamamoto 1996; Kei-i 1984), most highly regarded research universities are National Universities, which are funded and controlled by the Ministry of Education, Science, Sports and Culture (Monbusho or MOE) and whose faculty members are government civil servants. National Universities have no independent legal status apart from Monbusho, and thus have no legal basis to own intellectual or other property. Most of the National Universities have graduate schools, and they account for approximately 75 percent of total R&D expenditures (Irvine 1990). ^{( 3)} Therefore, the laws and procedures governing technology transfer from National Universities determine how most Japanese university discoveries are transferred to the private sector.

Formal technology transfer procedures for Japanese National Universities are set forth in a number of official "Notifications" issued by Monbusho and internal rules that individual National Universities have enacted to implement the Notifications. There is no overarching Diet legislation setting forth technology transfer procedures.

The most important single Notification underlying the Japanese technology transfer system is Monbusho Notification 117 of 1978 titled, "The Handling of Patents Related to Inventions by Faculty of National Universities" (Study Committee 1997). It sets forth the basic principle that IP rights to inventions made in National Universities belong to the individual inventors. However, it qualifies this principle by stating that inventions arising either under:

1. special funding from the Government for a project specified as being for the purpose of developing practical applications, or

2. a project specified as being for the purpose of developing practical applications, and which utilizes special research facilities and equipment (such as nuclear power research facilities and particle accelerators) that are established for use in Government sponsored research

should be classified as "National Inventions," and the Nation should inherit rights to these inventions. As described in subsection D below, central government agencies patent and license National Inventions.

In order to determine whether National University inventions should belong to the Nation, Monbusho Notification No. 117 of 1978 states that National Universities should establish internal regulations under which university inventors report their inventions to university presidents. University "Invention Committees" will then review the inventions and recommend to the university presidents whether particular inventions should belong to the Nation. The type of funding under which an invention is made is an important determinant of whether it will be classified as a National Invention. The principal categories of R&D funding in National Universities are listed below, and Table 2 summarizes the amount of funding under each category: ^{( 4)}

Standard Research and Educational Allowances for Faculty Members: These cover general laboratory operating expenses including supplies, equipment such as computers and laboratory instruments, utilities, upkeep, and general maintenance, but not salaries. Monbusho gives National Universities funds for Standard Allowances in amounts determined primarily by the number of laboratories and the number of tenure track researchers in each university (NSF 1997a; Tamura 1989).^{( 5)}

Grants-in-Aid for Scientific Research: Grants-in-Aid are proposal-based competitively evaluated Monbusho awards to individual researchers or groups of researchers. There are several classes of Grants-in-Aid ranging from special priority-area research projects that may involve several laboratories, to small scale research grants to supplement doctoral or post-doctoral fellowships. (NSF 1995)

Donations (including Endowments): These are donations from private organizations, foundations and individuals to support scientific research and educational activities. They can be targeted to a particular laboratory. Unlike other R&D funds, they can be rolled over to the next fiscal year and used flexibly to purchase equipment, support travel, hire temporary secretaries or technicians, etc. Faculty members must disclose donations to their university administrators who then report these data to Monbusho headquarters.

Commissioned Research Funds: Under Commissioned Research Agreements, universities can accept funds from companies, governmental organizations other than Monbusho, and other outside organizations to carry out contract-type research. Special Monbusho "investment funds" for application oriented research are also awarded under Commissioned Research Agreements.^{( 6)}

Funds for Joint Research with the Private Sector: Under Joint Research agreements, National Universities can accept funds from companies to conduct collaborative research in a defined area. Usually such projects involve the exchange of researchers.

Bench Fees for Commissioned Researchers: National Universities can accept technicians and researchers employed in companies and other outside organizations to do graduate or post-graduate level research. Their employers usually pay a research or "bench fee."

Table 2: Support under Various Funding Categories for R&D in Japanese National Universities in 1996

* The amount of Grants-in-Aid includes grants to local and private universities. Thus, the amount and percentage for National University R&D should be less than shown in the table.

Sources: Monbusho 1996; Monbusho Advisory Com. 1998; Committee 1996.

D. Patenting and Licensing of National Inventions by JSPS and JST

A decision by a university Invention Committee that a particular invention is a National Invention, is in effect a decision that the government should apply for a patent. Until April 1999, the Japan Society for the Promotion of Science (JSPS), an affiliate of Monbusho, handled most patent applications on behalf of National Universities. The Japan Science and Technology Corporation (JST, an affiliate of the Science and Technology Agency, STA) licenses most National Inventions. In April 1999, JST assumed responsibility from JSPS for patenting most university National Inventions. The offices of both JSPS and JST are located in the Tokyo area. JST can consider requests from companies for "preferential licenses" under which a period of conditional exclusivity is negotiated, although the general principle is that National Inventions belong to the public and should be made available non-exclusively to whomever wishes to use them. Royalties from National Inventions belong to the Nation and are transferred to a common National University general revenue account under the control of the Ministry of Finance (MOF). JSPS and JST can use some of these funds for administrative expenses. Neither individual inventors nor their universities share in the royalty stream.

As indicated by Table 3, the amount of patenting activity by JSPS and licensing activity by JST has been modest.

Table 3: JSPS and JST Patenting and Licensing of University Inventions

* Most were non-exclusive, non-preferential licenses according to JST.

   Sources: JSPS Web site: http://www.jsps.go.jp/j-home.htm.

Table 4 summarizes the decisions of National University Invention Committees for two recent years. As suggested by the last row of that table , the number of university inventions reported to university presidents and reviewed by university Invention Committees is small, especially in comparison to the nearly 10,000 inventions reported to U.S. universities and medical research institutes in 1996 (AUTM 1998).

Table 4: Disposition of Inventions Reviewed by Invention Committees in all 99 Japanese National Universities in 1995 and 1996

Sources: (Monbusho Advisory Com. 1998; Monbusho undated)

It is clear that the actual number of university inventions exceeds the numbers reviewed by university Invention Committees. A study by the Japan Bioindustry Association (JBA) estimated that 38% of Japanese patent applications in the field of genetic engineering (on average 874 of 2327 applications annually from 1987-97) listed at least one university professor as an inventor. Out of a sample of 252 such applications, none were National Inventions filed by JSPS and at least 72% were filed by private companies. (The remainder were filed by individual inventors, non-profit organizations and government agencies other than Monbusho.) However, only 40 (16%) had become issued patents. In the case of 155 (62%), the applicants had not even requested the Japanese Patent Office (JPO) to review the applications. (JBA 1998) Furthermore, a study by the JPO showed that Japanese companies neither develop, use nor license approximately two-thirds of the technologies for which they actually obtain patents, usually because they are not interested in the technologies but nevertheless want to block other companies from using them (JPO 1996). These and other anecdotal studies^{( 7)} suggest that many university discoveries are transferred to the private sector informally and unaccounted for in any normal statistics, and that the vast majority of technologies transferred in this manner are not developed.

Informal transfers occur in many ways. Professors consult with companies. Corporate researchers working in university laboratories communicate research results back to their companies. Graduates find employment in companies. Sometimes the transfer of IP rights occurs without any written agreement. Sometimes a short document known as an "obo-e-gaki" serves as an assignment agreement. If a company believes a university researcher's discovery is important, it may at its discretion decide to apply for a patent. Some productive faculty members do not know how many of their discoveries have been patented by companies.

University inventors rarely bargain for development commitments in exchange for transferring their discoveries to companies, and discussions concerning personal royalties are postponed until the companies decide to apply for patents. However, companies often give Donations to individual professors' laboratories to maintain an amicable relationship and to help ensure the professors' cooperation in providing access to new technologies and promising graduates. Donations tend to be under 1 million yen ($6000), because division directors in many corporations can authorize Donations up to this amount, whereas executive board approval is usually needed for larger Donations. Monbusho needs to approve Donations above 5 million yen ($30,000) a year. Professors often say that the value of the research they perform for companies far exceeds the amounts they receive in Donations.

Donors are prohibited from making advance contracts to receive IP rights (Monbusho 1996). Nevertheless, professors and companies often perceive Donations to be ideal forms or research support, provided both parties trust and cooperate with each other. As noted above, Donation funding is more flexible than other forms of support. Until Japanese tax authorities recently began to scrutinize the contractual aspects of Donations, companies could automatically deduct from taxable income the entire amount of their Donations as if they were purely charitable gifts. In contrast, corporate support for Commissioned or Joint Research has been subject to scrutiny by Japanese tax authorities, who can challenge the deductibility of specific expenditures in support of such research. Perhaps most importantly, inventions arising under Donations do not face the delays and uncertainties associated with reporting to university presidents and review by Invention Committees, because there is a presumption, implicit in Notification 117 and the internal rules of many National Universities, that such inventions do not belong to the Nation. The low levels of total funding under Commissioned and Joint Research compared with Donations (see Table 2) are consistent with university researchers and companies favoring Donations as a means for companies to support university R&D in return for IP rights to university discoveries and access to university researchers.

Commissioned and Cooperative Research agreements are the only formal mechanisms for companies to fund university research in return for IP rights to discoveries made with their support. The Monbusho notifications governing the transfer of IP rights under these agreements date from 1970 and 1983, respectively, and are complex (Notifications 260 of 1970, 195 of 1983, 172 of 1984 and 186 of 1996 in Study Committee 1997). To summarize:

• These agreements can be negotiated directly between National Universities and companies. Uncomplicated agreements can be concluded in one or two months. However, the beginning of research projects must coincide with the start of the Japanese fiscal year on April 1. In other words, if a company and a National University decide in July to begin a Commissioned or Joint research project, the company cannot begin disbursement of funds until the following April. In addition, corporate funds must be disbursed on an annual basis through MOF, which takes 30% as overhead. Occasionally, MOF takes several months to disburse funds. Funds for one fiscal year cannot be rolled over to future years.

• It is imperative that Joint Research inventions be reviewed by University Invention Committees, and there is a strong presumption that the same is true with respect to Commissioned Research inventions.

• There is a strong presumption that Commissioned and Joint Research inventions by university researchers belong to the Nation.

• However, it is possible for the sponsoring company to contract in advance with the university to jointly apply for patents (i.e., to have joint ownership). Under Japanese patent law (section 73), the consent of all co-owners is necessary to assign or license patent rights.

• If there is no pre-existing contract, the University President can approve a 10-year extendible preferential license to the sponsor or another company specifically designated by the sponsor. The President can revoke this license if the invention is not developed or if the license causes severe public harm. Any sublicenses must be approved by the President and, in the case of Commissioned research, any government agencies that supported the research.

Contracts to jointly apply for patents give the sponsor stronger IP rights than preferential licenses and, not surprisingly, are much more commonly used (Invention Com. 1998).

If the decisions of the University of Tokyo's Invention Committee are representative of those of National Universities as a whole, then Commissioned and Joint Research inventions account for over 70% of the approximately 50 inventions that are designated annually to be National Inventions (see Table 4 and Invention Com. 1998).

University Invention Committees are often criticized for meeting only once a year in many universities. However, many persons believe it may be better for the Invention Committees to be inactive, if being more active might result in the Committees deciding that more inventions should be National Inventions, thereby diminishing prospects for commercialization. In addition, the science and engineering professors who constitute most Committee members may be reluctant to antagonize their peers by ruling that an invention should be a National Invention.

In summary, an "official" technology transfer system, focused on the need to identify inventions that should belong to the Nation and that should be patented and licensed (usually non-exclusively) by central Government agencies, co-exists with a system that allows university researchers to transfer rights in their discoveries directly to companies. The boundary between these two systems is not clear. But because the criteria for classifying inventions as National Inventions emphasize "practical applications," the inventions that are vulnerable to being classified as National Inventions are also likely to be of interest to industry. However, because of the remoteness of JSPS and JST from the scientists and engineers who understand the details and utility of their discoveries, the delays inherent in the official procedures, and the barriers to companies receiving the strong exclusive or preferential license rights, the likelihood that discoveries classified as National Inventions will be developed and commercialized by private companies is small. Therefore, university researchers, university administrators and companies, all have considerable incentives to take advantage of the unclear boundary between the two systems and, whenever possible, to bypass official technology transfer procedures and let university inventors transfer their discoveries informally to the private sector. Available evidence suggests that this is precisely what has been occurring.

This "informal" or "bypass" system of technology transfer can be very efficient if there is a good match between the inventor and company, i.e., if the company is willing and able to develop the inventor's discoveries. Transaction costs are low, at least for universities. No funds need to be paid to support the operations of a TLO. Marketing and licensing have already been done by the inventor and company. However, the informal system lacks a mechanism to select the most appropriate development partner among potentially interested companies, relying instead on preexisting relationships between individual researchers and companies. It lacks a mechanism to ensure that the company will actually develop and market the technology. It provides no mechanism to educate university researchers about the commercial value of their research and the importance of protecting IP rights, even though recent surveys indicate that awareness of these issues is low (Arai 1998). There is no mechanism to collect data on technology transfers as they occur, much less on the outcome of such transfers. Finally, because official procedures are bypassed, companies may be concerned whether they have obtained legally enforceable rights to university discoveries.^{( 8)}

Except when private universities receive R&D support from the Japanese Government for specific projects, they are generally free to set their own technology transfer policies. Recently, Nihon, Ritsumeikan, and Tokai Universities established technology transfer offices and required their employees to assign to them any work-related inventions. Other private universities, however, still adhere to the principle that IP rights remain with the inventor. As in National Universities, most transfers still occur informally with no reporting of inventions or licensing/assignment agreements, except in the cases of the three institutions mentioned above. Private universities have more flexibility in receiving research funds from industry. Nevertheless, it appears that most industry support still occurs as Donations directed to a particular professor's laboratory.

In theory, inventions arising under Monbusho Grants-in-Aid to private universities may belong to the Nation. At least in the case of small Grants-in-Aid, this possibility seems to be overlooked by all parties concerned. Aside from the three universities that require assignment of inventions by faculty, few if any other private universities have invention committees. However, when private universities receive government support under Commissioned Research agreements, any industrially applicable inventions should become National Inventions to be patented and licensed according to the rules of the particular funding agency.^{( 9)}

In May 1998, the Japanese Diet enacted legislation drafted by the Ministry of International Trade and Industry (MITI) and Monbusho to improve university-industry technology transfer (Technology Transfer Law 1998). This legislation authorizes the establishment of "technology transfer offices" TTOs with authority to patent and transfer university inventions. The TTOs can be legally independent private corporations or publicly chartered corporations, thus enabling National Universities to avoid complications that would arise if they directly collected royalties from private companies. Private universities may establish TTOs within their official structures.

Article 4 permits TTOs whose operating plans have been approved by MITI and Monbusho to patent university inventions, assign or license these inventions, and transfer royalties to their universities and the individual inventors. Heretofore, National Universities did not have the authority to manage IP or to share in royalties. However, the law leaves unchanged most of the technology transfer procedures described in Section IV above. In particular, it does not alter the basic principle that inventors retain the right to patent their inventions. It gives inventors the option of assigning their inventions to TTOs but does not require them to do so. Furthermore, it does not change the criteria for deciding which inventions should belong to the Nation nor the responsibilities of the National University Invention Committees. Absent special permission from Monbusho, the TTOs do not have authority to transfer National Inventions (Article 12). So far, no TTO has received permission under Article 12 to transfer National Inventions, other than on a case-by-case basis. This is because MOF and Monbusho maintain that the National Properties Law (sections 2, 3 & 21) and the Financial Law (section 9), require the TTOs to compensate the government for the value of each National Invention that the government transfers to a TTO, before the TTO transfers the invention to industry. Therefore, the uncertainties and inefficiencies associated with the National Invention system still exist.

The law also enables MITI to provide approved TTOs, whether affiliated with National or private universities, up to 20 million yen ($120,000) per year for five years to supplement their budgets. However, annual operating costs are considerably higher, especially since TTOs are tending to concentrate their resources on patenting technologies (which requires the services of expensive patent attorneys) and attempts to estimate the value of their technologies, rather than on marketing. In addition, royalty income as of April 1999 has been low--only one license had been issued by an approved TTO.

Many Japanese universities are now establishing TTOs, although as of April 1999, only four had been approved by MITI and Monbusho. The following descriptions illustrate how the TTOs of three major National Universities, the Tokyo Institute of Technology, Tohoku University and the University of Tokyo, are trying to cope with their cash flow problems:

Two of these TTOs were established as independent for-profit companies free from direct supervision by Monbusho. The other plans to constitute itself as a non-profit organization within the overall structure of its associated university. The two for-profit TTOs are encouraging science and engineering faculty members to each purchase at least several hundred dollars worth of non-voting TTO stock. A portion of any future royalties (approximately 10%) will be set aside for distribution to stockholders, although most faculty recognize their stock purchases to be contributions to help their TTOs rather than financial investments. In the case of one of these TTOs, faculty contributions will be matched by contributions of local banks interested in promoting the local economy.

All three TTOs are also establishing a system of privileged access based on payment of membership fees. One plans to charge individuals 50,000 yen ($300) and corporations 100,000 yen ($600) for annual membership. It hopes to have approximately 200 corporate members. Another plans to fix membership charges at between 1-2 million yen ($6,000-12,000) and to have 50 to 100 members. Membership charges for small businesses will be reduced by an as yet undecided amount. Both of these TTOs want to encourage technology diffusion to local industries. The third TTO is hoping that 20 companies will each pay 5 million yen ($30,000) in annual fees. In the case of all three TTOs, membership confers the right to see lists of available technologies, sign confidentiality agreements to obtain detailed information about these technologies, and have first rights of refusal in licensing. After 6 to 9 months, if no member takes a license, non-members will be able to sign confidentiality agreements and negotiate licenses. The two TTOs charging lower membership fees say that any company can become a member and they are reaching out to faculty in all science and engineering departments to urge that all inventions be assigned to the TTOs. The TTO charging $30,000 membership fees welcomes applications from any company, but certain companies from a cross-section of high technology industries are being encouraged to become members. It is hoped that one member might serve as a marketer of technologies to non-member companies, although the other members may insist on limitations on such outside marketing. In the case of all three TTOs, licensing royalties will be used to cover the TTOs' operating and patent prosecution costs, with any remaining funds to be allocated between the inventor, the inventor's department, the university as a whole, and profits for distribution among shareholders. As of the end of 1998, none of these TTOs had reached its membership goals.

The recently formed TTO of a major private university has entered into a cooperative relationship with an investment subsidiary of a major bank. This bank subsidiary will review new technologies that inventors assign or bring to the attention of the TTO and select those technologies that it intends to develop or license to other companies. The bank subsidiary will then receive assignments to these technologies from either the TTO or the individual inventors. In return, the TTO will receive a percentage of the royalties that the bank subsidiary earns from the transfer of these technologies to third parties. The TTO itself will take responsibility for transferring inventions not selected by the subsidiary. Any funds remaining after paying TTO expenses will be shared between the inventors and the university. Although this arrangement reduces TTO staff requirements and patent prosecution costs, it substantially diminishes the TTO's control over the university's technologies.

Besides the need for financial support, TTOs also appear to be seeking partnerships with private companies for expertise on evaluating the commercial potential of their technologies. Ironically, many Japanese investment and venture capital firms say that they, themselves, lack such expertise. It is curious that many TTOs appear to be eschewing a strategy of active marketing to as many interested companies as possible, so that "value" can be better approximated by competitive bidding or by running royalties based on a fixed percentage of sales linked to strong development commitments and realistic business plans. ^{( 10)}

There are essentially four paths for transferring IP rights to Japanese university inventions to industry:

Although the number of such inventions is very small (see Tables 3 and 4 and the last sentence of IV(F)), the possibility that inventions will be classified as National Inventions hangs like a Sword of Damocles over university-industry technology transfers. This is because many inventions can theoretically be classified as National Inventions, especially if they arose under Grant-in-Aid support, the main source of project-specific R&D funds in National Universities. Classification as National Inventions means that such technologies face substantial legal and bureaucratic obstacles to development and commercialization by private companies. The possibility that many inventions could be classified as National Inventions probably has discouraged inventors and companies from entering into transparent contractual agreements. This contributes to a culture of informal, undocumented transfers under which inventors are in a weak bargaining position. They have no means to ensure that their inventions will be developed, and universities lack data on what technologies they create and the fate of these discoveries. A Monbusho official responsible for university-industry cooperation stated publicly in early 1998 that the process of deciding which inventions are National Inventions and the system of patenting and licensing such inventions are without public benefit.

In 1996, the number of inventions arising under Commissioned and Joint Research Agreements that were ruled to be National Inventions was probably about 40. This number will likely increase as the number of Commissioned and Joint Research Agreements increases. Although these inventions do not face the overwhelming barriers to commercialization faced by other National Inventions, there are many reasons why companies and professors prefer to cooperate through donations rather than these types of agreements unless high levels of corporate support are necessary.

TTOs will not diminish transaction costs associated with having to obtain Government permission to patent or license inventions, nor will they give universities control over the inventions made by their employees. Therefore, the two principal factors underlying the success of the U.S. Bayh-Dole system are not present in the new TTO system.^{( 11)} As in the U.S., TTOs will have incentives to concentrate scarce resources on licensing a few promising inventions while ignoring inventions with less immediate commercial potential. In Japan, however, these financial constraints are exacerbated, because university inventors can still transfer their inventions directly to companies. This raises the possibility that inventors will transfer their most commercially attractive discoveries to companies with which the have long standing relationships, leaving their less attractive inventions to the TTOs.

The membership fee priority system that most TTOs are adopting threatens to exacerbate weak TTO finances and undercut their fundamental technology transfer mission. If membership fees are high, companies will have strong incentives to deal directly with individual scientists in order to avoid paying the fees and the royalties that the TTO will charge. Inventors that have longstanding relationships with companies that are not TTO members may also want to bypass the TTOs. The fact that the TTO that charges 5 million yen membership fees was far short of its membership goals as of April 1999 suggests these concerns are warrented.

More fundamentally, the basic goal of technology transfer is to find the best companies to develop university inventions and to ensure that such companies do their best to follow through with development commitments. High membership fees restrict the pool of potential developers and may compromise this fundamental goal. High membership fees may be particularly burdensome for small companies, and thus may perpetuate the tendency for university technologies to flow to large companies, opposite to trends in U.S. universities. If a subsidiary of a large financial institution that is a core member of a "kei-retsu" group of companies is given a monopoly on attractive university inventions, that subsidiary may be more likely to license to companies within the same kei-retsu group,^{( 12)} even though these companies might not be the best candidates to develop certain technologies. Finally, if a limited number of companies have privileged access to a university's technologies, they may be able to obtain these at lower royalty rates than if the TTOs marketed them to an unrestricted number of companies.

Nevertheless, the new TTOS should increase awareness of the importance of technology transfer among university researchers and encourage some universities to develop expertise in managing their technology. Whether the system will be sustainable depends upon the willingness of inventors to transfer their inventions to their TTOs, the interest of companies in TTO inventions, and the effective use by the TTOs of provisions in their license agreements that require companies to develop the technologies they receive (see Note 11).

Although it is risky to speculate on the number of patent applications filed by companies on inventions transferred directly from university inventors, the anecdotal surveys mentioned above suggest that this number probably exceeds 1000 annually. In 1996, U.S. and Canadian universities and major non-profit research institutions filed over 3200 patent applications and negotiated over 2700 licenses of which 53% were exclusive licenses (AUTM 1998). Considering that (1) the population of Japan is roughly half that of the U.S. and (2) Government support for biomedical university R&D in Japan is much less than in the U.S. (Kneller in Branscomb in press) while biomedical inventions account for the majority of patents and approximately two-thirds of active licenses by most U.S. universities (Mowery in Branscomb in press; AUTM 1998); the number of patentable technologies transferred by this path probably is comparable to the number of inventions transferred by U.S. universities to industry.

If the "inventors retain rights" path of technology transfer is deficient, the reason is probably not lack of contact between university researchers and industry nor an inability of industry to obtain IP rights to university discoveries. Rather the nature of university-industry contacts and the lack of incentives for Japanese companies to use effectively the technologies they receive from universities are probably at issue. As discussed in Section IV(E) above, the vast majority of patentable discoveries passed by inventors to industry probably are not being used or developed.

Since the Inventors Retain Rights path will probably remain the most frequently used system of technology transfer for the near future, Japan's best strategy may be to try to improve this system. This might be done through outreach to faculty members and providing them with information and support services to enable them to transfer their discoveries more effectively. The goal would be to empower individual university inventors to be more effective technology transfer agents; in other words; to change the "inventors retain rights" path from an "informal," "bypass" system to a system which enables inventors to act as effective technology transfer agents. This would require more contractually based transfers. Although this would entail adjustments to traditional relations between university faculty and industry, it would increase the transparency of technology transfers and make collection of data on such transfers considerably easier. However, it would be important to clearly limit the criteria for classifying inventions as National Inventions and to ensure that, except in rare cases, such classifications be made prospectively at the beginning of research projects rather than retrospectively once inventions have occurred and their commercial potential is apparent.

Changes in other areas besides IP rights may also be necessary to improve university-industry cooperation in Japan. The preceding discussion assumes that faculty members are willing to bargain for royalties, increased research support and firm development commitments at the time they transfer their rights to companies. However, it appears university inventors rarely do so, even in private universities where restrictions on cooperation with industry are less than in National Universities.

One reason professors do not bargain may be that the transfer of IP rights is just one aspect of multifaceted long term relationships with companies. Professors depend upon companies to employ their graduate students. Corporate contacts provide interesting information, ideas for research with practical applications, and annual Donations usually not tied to any particular invention or line of research. Under these circumstances, it may seem ungrateful or counterproductive for professors to bargain over transfer terms when promising inventions do arise.

Another reason is that increased royalties or contributions from industry may be of limited usefulness in view of difficulties in recruiting young researchers to work on industry-sponsored research projects. Lifetime employment, with promotion opportunities and retirement benefits depending greatly upon length of service within a particular organization, is still the norm in Japan (Aoki 1990). It is generally not considered to be a career-enhancing step for a promising new M.S. or Ph.D. recipient to spend a couple of years working on a temporary university research project (Coleman in press). Moreover, the number of doctoral, postdoctoral and tenure track positions in Japanese universities are limited (Notes 2 and 5). Companies still prefer to hire graduates with M.S. rather than doctoral degrees whom they will train in-house to suit their needs (Coleman in press). Ph.D. students usually do not receive research assistantships or stipends to defray tuition and living expenses, and most National University faculty and administrators interviewed for this study believe it is not permissible to use corporate research support for such purposes. In fact, however, it appears that it is permissible to use Donations and corporate Commissioned and Joint Research funds for stipends and research assistantships for Ph.D. students and post-doctoral researchers, although there are limits on the hourly pay such persons can receive and on the amount of time Ph.D. students can spend per week on industry-supported projects that are not their own thesis research.

National University researchers who dare to form their own companies must resign from their university positions if they hold management positions in the companies. The infrastructure of individual donor networks and venture capital companies to support such start-up companies is not as developed as in the U.S. Japanese venture capital companies, most of which are subsidiaries of large financial institutions or major companies, tend to fund research that is at a more advanced stage than do their U.S. counterparts and to direct much of their investment overseas. It is common to hear officials of Japanese venture capital companies say that the quality of Japanese university research is inferior to that in the U.S., and to decry a lack of entrepreneurial spirit among Japanese university researchers. However, the Japanese financial services industry itself lacks the expertise that exists in the U.S. to evaluate new technologies and to assist in managing start-up companies. If a Japanese scientist-entrepreneur persists against these financial obstacles and lack of job security and then his company fails, the personal cost is much higher than in the U.S. His credit record is ruined, it is hard for him to find employment, his family members face ostracism and he often must move because of failure to make mortgage payments.

Therefore, whether because of limited numbers of researchers, greater concern about job security, real or perceived labor immobility, limited access to capital and management advice, or the greater cost of failure; very few university researchers form companies. As of 1995 in all Japan, approximately 89 new companies had been formed that had significant R&D activities related to biotechnology based pharmaceuticals. Only four of these were started independently rather than as off-shoots from large companies or as spin-offs from government organized research consortia. (Kano in Branscomb in press) In the U.S., small companies backed by venture play an important role in developing university inventions to the point where large companies are willing to pursue final development and commercialization (Atkinson 1994; Kortum 1998). Small companies account for over half the licensees of university technologies (AUTM 1998; GAO 1998). Without such companies, technology transfer from Japanese universities to industry may be more difficult than in the U.S., no matter what rules govern the ownership and transfer of IP rights.

In order to encourage entrepreneurial dedication to the development of potentially valuable discoveries, some large Japanese companies are allowing some scientists and engineers to form their own "in-house" start-up companies capitalized by funds from the parent companies. Thus, the possibility exists that Japanese universities will never become more integrated into a national or international technology development system, but instead will be bypassed as the private sector develops unique ways to exploit the capabilities of talented scientists and engineers interested in fundamental research topics of importance to industry. On the other hand, there may soon be a substantial pool of young scientists and engineers who cannot find lifetime employment in major companies or universities and who are willing to work as post doctoral researchers or join small companies. As mentioned in Note 2, the government is funding the training of more Ph.D. level researchers. At the same time, at least in the pharmaceutical industry, hiring of new university science and engineering graduates has fallen dramatically.

In this situation, there might be opportunities for American or European individual investors or venture capital companies to work with Japanese counterparts to develop promising early stage technologies. Under the "inventors retain rights" path of technology transfer (and under the Commissioned/Joint Research path provided close relations are maintained with university administrators), the transfer of IP rights to foreign venture capital companies should be manageable. Whether the new TTOs can play a constructive role in this process depends upon their openness to working with small companies and independent Japanese and foreign investors.

Aoki, Masahiko. 1990. "Toward an Economic Model of the Japanese Firm". Journal of Economic Literature 28:1-27.

Arai, Hidehiko. 1998. Actor Analyses on Cases of Successful Transfer of Technology from Universities and National Institutions. Tokyo: Second Theory-Oriented Research Group, National Institute of Science and Technology Policy, Japan Science and Technology Agency (in Japanese).

Asian Technology Information Program (ATIP). 1998. Report: ATIP98.084: Tohoku University Venture Business Laboratory (VBL) (distributed by Internet by ATIP, 12 Oct. 1998).

Association of University Technology Managers, Inc. (AUTM). 1998. AUTM Licensing Survey FY 1996 (AUTM).

Atkinson, Stephen. 1994. "University-Affiliated Venture Capital Funds." Health Affairs, Summer 1994:159-175.

Branscomb, Louis; Kodama, Fumio; Florida, Richard and Pechter, Kenneth (eds.). Industrializing Knowledge: University-Industry Linkages in Japan and the United States. Cambridge: MIT Press (in press).

Coleman, Samuel. (In press) Japanese Science: Hierarchy, Mobility and Control. London: Routledge.

Committee 1996: Full citation: Committee for the Study of Scientific Research Funding (Ed.). 1996. List of Approved Projects for Monbusho's Grants-in-Aid for Scientific Research, Part 1 or 2. Tokyo: Gyousei (in Japanese).

Council on Governmental Relations (COGR). 1993. University Technology Transfer: Questions and Answers. Washington D.C.: COGR.

Eisenberg, Rebecca 1996. "Public research and private development: patents and technology transfer in government-sponsored research", Virginia Law Review 83: 1663-1727.

Etzkowitz, Henry. 1994. "Knowledge as Property: The Massachusetts Institute of Technology and the Debate over Academic Patent Policy", Minerva 32:383-421 (for a history of the assignment requirement).

Government Accounting Office (GAO). 1998. Technology Transfer: Administration of the Bayh-Dole Act by Research Universities (GAO/RCED 98-126) Washington DC: GAO.

Invention Com. 1998. Full citation: University of Tokyo's Invention Committee. 1998. Japanese Patents Belonging to the Nation: Applications and Acquisitions (in Japanese).

Irvine, John et al. 1990. Investing in the Future: an International Comparison of Government Funding of Academic and Related Research--Report of a study sponsored by the UK Advisory Board for Research Councils and the US National Science Foundation. Hants, England: Elgar .

.

Japan Bioindustry Association (JBA). 1997. The Usefulness of University Research for the Promotion of Our Country's Bioindustry (in Japanese).

Japan Patent Office (JPO). 1996. Survey on the Appropriate Promotion of Patents: Fact Finding Survey of Unused Patents: Information Report: Summary (Industrial Technology Laboratory. Tokyo: Japan Technomart Foundation (in Japanese); and Japanese Patent Office. 1996. Outline of Results of Status Survey on Unused Patents (English summary of above).

Japan Science and Technology Corporation (JST). 1997. JST 1996-1997 (English brochure).

Japanese Patent Law (various cited articles). At http://www.jpo-miti.go.jp/index.htm (in Japanese and English).

Kei-i, Tominaga. 1984. A Study Evaluating Universities (Tokyo: Tokyo University Press, Tokyo) (in Japanese).

Kneller, Robert. 1998. "University-Industry Technology Transfer in the United States". Tokugikon (a publication of the Japan Patent Office) 198:15-46 (in English and Japanese).

Kortum, Samuel & Lerner, Josh. 1998. Does Venture Capital Spur Innovation? Working Paper No. 6846. National Bureau of Economic Research.

Latker, Norman (Patent Counsel, U.S. Department of Health Education and Welfare). 1977. Statement before the Subcommittee on Science, Research and Technology of the U.S. House of Representatives, 26 May 1977.

Miyashita, Kenichi & Russell, David. 1996. Keiretsu: Inside the Hidden Japanese Conglomerates. New York: McGraw-Hill.

Monbusho. 1996. Full citation: Science and International Affairs Bureau of the Ministry of Education, Science, Sports and Culture (Monbusho). 1996. Research Cooperation Between Universities and Industry in Japan. Tokyo: Monbusho (in English and Japanese).

Monbusho undated. Pertaining to the Handling of Research Results (Patents) Arising in Universities (undated Monbusho data sheet) (in Japanese).

Monbusho Advisory Com. 1998. Full citation: Report of the Monbusho Advisory Committee for the Promotion of University-Industry Cooperation: Aiming at the Construction of a New Technology Transfer System Related to Patents (in Japanese).

National Science Board (NSB). 1998. Science and Engineering Indicators: 1998. Washington D.C.: National Science Foundation (Appendix Tables 2-30, 4-2, 4-41, 4-45, 4-46 & 5-8).

National Science Board (NSB). 1996. Science and Engineering Indicators: 1996. Washington D.C.: National Science Foundation (Appendix Tables 2-30, 2-31 & 5-5.

National Science Foundation (NSF) Tokyo Regional Office. 1998. Report Memo 98-01: The 1997 Monbusho White Paper.

National Science Foundation (NSF) Tokyo Regional Office, 1997a. Report Memo 97-11: Japanese Government Organization for Science and Technology,

National Science Foundation (NSF) Tokyo Regional Office. 1997b. Report Memo 97-06: Japanese Government Science and Technology Budget - Fiscal Year 1997.

National Science Foundation (NSF) Tokyo Regional Office. 1995. Report Memo 95-18: Japanese Ministry of Education's Grants-in-Aid for Scientific Research.

Nelsen, Lita. 1998. "The Rise of Intellectual Property Protection in the American University". Science 279:1460-1461.

Nikkei Industrial Times. 1997. "University Industry Cooperation: New Industry Creation, Part 1 of 3" (24 December 1997, page 4) (in Japanese).

Normile, Dennis. 1995. "Japan Expands Graduate, Postdoc Slots". Science 269:1335-36.

Public Law (PL) 96-517. 1980. Amending Title 35 of the U.S. Code by adding Chapter 18, Sections 200-212.

Rule 53. 1978. "Tokyo Institute of Technology Invention Rules" in Vol. 7: Cooperative Research (approved 8 September 1978) (in Japanese).

Rule 58. 1978. Tohoku University Rules for Handling Faculty Inventions (approved 21 November 1978 ) (in Japanese).

Study Committee. 1997. Full citation: The Study Committee for Handling Outside Funding for National Universities (Ed.). 1997. Administrative Handbook for Research Cooperation Between Universities and Industry , 2d Ed. Tokyo: Gyou-sei. (in Japanese). For earlier versions of these notifications see: The Study Committee for Handling Outside Funding for National Universities (Ed.). 1992. Administrative Handbook for Research Cooperation Between Universities and Industry . Tokyo: Gyou-sei. (in Japanese).

Tamura, Kenzi. 1989. "Funding Research in Japan". Science and Public Affairs 4: 31-41 (London, Royal Society, 1989).

Technology Transfer Law. 1998 Full title: Law on the Promotion of the Transfer to Private Enterprises of Research Results Pertaining to Technologies Arising in Universities. 1998 No. 52. Ministry of International Trade and Industry. Promulgated 6 May 1998 (in Japanese).

University of Tokyo Invention Rules. 1979. (especially section 4) In Vol. 4: Commissioned Research and the Like; and Regulations Implementing the University of Tokyo Invention Rules (especially section 2). (in Japanese).

Yamamoto, Shinichi. 1996. "The System of Mass Higher Education and Research Universities". Research Reports, No. 91. Chapter 8, pages 271-284 (Tokyo: Broadcast Education Development Center, Tokyo), (in Japanese).

---

1. It has been suggested that Japanese Government statistics overstate the amount of Government support for university R&D, compared to the way expenditures are generally accounted for in other industrialized countries (Ohtawa in Branscomb in press; Irvine 1990).

---

2. Most Japanese science and engineering graduates take jobs in industry after receiving either a B.S. or M.S. degree. The "kouza" system outlined in Note_5 provides only limited tenure track opportunities for Ph.D. holders, and as described at the end of this paper Japanese industry still prefers to hire M.S. rather than Ph.D. graduates. In 1995, only 5,453 persons received Ph.D.s in science and engineering from Japanese universities, compared with 26,207 in the U.S (NSB 1998). Approximately one third of the Japanese Ph.D.s are "thesis" doctorates awarded by Japanese universities to persons who do not enroll as Ph.D. candidates but instead submit for evaluation the results of research performed at their place of work (NSB 1996). The Japanese Government is increasing the number of doctoral and postdoctoral researchers, awarding 2300 doctoral and 3800 postdoctoral fellowships in 1996, while planning to increase the number of awards to 3600 and 6400, respectively, in the year 2000 (Normile 1995; NSF 1997a).

---

3. Prominent among the 99 National Universities are the seven former "Imperial Universities": the University of Tokyo and Kyoto, Osaka, Kyushu, Nagoya, Tohoku and Hokkaido Universities; along with the Tokyo Institute of Technology and the University of Tsukuba (formerly the Tokyo University of Education) (Yamamoto 1996; Kei-i 1984).

---

4. Separate Monbusho accounts, distinct from any of the listed categories of R&D support, cover salaries, bonuses, and health and retirement benefits for full time Government employees (i.e., tenured faculty and full-time support personnel); scholarships and stipends for graduate students and post-doctoral researchers; and building and construction costs.

    Separate Monbusho accounts also cover special projects to enhance university-industry cooperation, such as the establishment in 1996 and 1997 of 24 fully equipped "Venture Business Laboratories" ("VBLs") in universities to provide space for cooperative research between university and industry researchers. The initial 1995 budget to establish these VBLs was 25 billion (about $150 million) (ATIP 1998). Inventions made in VBLs are not National Inventions simply by virtue of being made in these laboratories.

---

5. A typical laboratory in a Japanese National University consists of a single "kouza", or tenured group of civil servant academics, usually consisting of one professor, one associate professor and one or two research associates; plus support personnel.

---

6. One of the principal Monbusho programs using investment funds is called "Research for the Future". It provides large-scale 5-year competitive grants to groups of university researchers for application-oriented research. 20.6 billion yen ($120 million) was budgeted for 204 projects under this program in Japanese fiscal year 1997. (NSF 1997b; NSF 1998)

---

7. A 1997 article in the Nikkei Industrial Times stated that over a recent five year period 317 of 359 inventions (88%) made in the graduate engineering research department of Tohoku University were assigned informally to companies, and that 185 of 197 inventions (94%) made in the University of Tokyo's Institute for Industrial Technology were similarly passed informally to companies.

---

8. At least one Japanese company that serves as an intermediary between foreign venture capital and Japanese universities routinely requests that all technologies that are candidates for licensing by its clients be reviewed by the appropriate university Invention Committees. Although in other circumstances, the inventors might not bring the inventions to the attention of the Committees, it insists on this step to diminish the possibility that an invention that is licensed to one of its clients might later be ruled to be a National Invention.

---

9. Inventions made under Monbusho "special investment funds" (see Note_6) are National Inventions. In a recent case, a private university inventor, unaware of this stipulation, "transferred" his rights to a discovery made under such funds a company with which he had been collaborating. JSPS invalidated the transfer. The company was forced to write a letter of apology. But then the company negotiated a co-ownership agreement with JSPS under which it agreed to pay royalties to the government. Neither the inventor nor the university obtained a share of the royalties.

---

10. A common practice of U.S. TLOs is to require prospective exclusive licensees to submit business development plans for the inventions they want to license and then to include "benchmark" or "due diligence" clauses in the license agreements. Such clauses enable the TLO to revoke the license if the licensee is not meeting the development milestones set forth in these clauses. Japanese universities appear to have very little experience using such clauses. Some Japanese TTOs say they intend to use them in most exclusive licenses. Others say they are afraid that licensees will object. The author knows of no case of individual inventors including such clauses in "obo-e-gaki" or other agreements they negotiate directly with companies.

---

11. Although there is an ongoing debate in the U.S. about the effectiveness of the Bayh-Dole system (see Mowery in Branscomb in press; Nelsen 1998), eliminating these transaction costs and thereby giving universities strong incentives to ensure that a large number of promising technologies are commercially developed is the principle benefit most often attributed to this system. In other words, the Bayh-Dole Regulations positioned U.S. university TLOs to be effective technology transfer and innovation agents (Kano in Branscomb in press).

---

12. "Kei-retsu" is a term to describe associations between large numbers of companies based primarily upon a common major bank which provides financial support to the various companies in its kei-retsu in return for equity in these companies. Cross holding of stock is also common among companies within a particular kei-retsu (Miyashita 1996).