Industry-Academia Collaborations for Open Innovation in Japan: OECD's latest survey as seen in cases from the United States and Europe

NOT for quotation

Date November 1, 2016
Speaker Mario CERVANTES (Senior Economist and Head of TIP Secretariat, Science and Technology Policy Division, OECD Directorate for Science, Technology and Innovation)
Commentator Anders KARLSSON (Vice President, Strategic Alliances, Global Academic Relations, ELSEVIER)
Moderator WATANABE Masashi (Deputy Director, Industry-University Collaboration Office, Industrial Science and Technology Policy and Environment Bureau, METI)


History of Japanese policy on industry-academia collaboration

Mario CERVANTES's Photo


Looking at the 20-year history of industry-academia collaboration promotion policy in Japan, the first decade focused on the development of the domestic systems including the technology licensing organization act. During the second decade, the policy has shifted to the establishment of industry-academia collaboration centers as strongholds in universities.

Japanese universities have fundamentally changed following the two periods. In 2004, national universities became incorporated. In 2006, the amendment of the Basic Act on Education meant that contribution to society, including university-industry collaboration, is placed as one of the principle missions of universities along with education and research.

During the second decade, from 2008 to 2013, METI supported facilities to be utilized as strongholds in major regions with the specific end goal of bridging leading technologies in the region for practical development across industry and academia. It was expected that such strongholds would provide venues for developing viable human resources related to such technologies.

Industry-academia collaboration quickly expanded from 2003 to 2014. The number of relevant research projects doubled during the period. But the ratio of corporate research and development (R&D) investment in domestic universities in Japan is less than 1% of the total corporate R&D budget. The expense of each industry-academia collaboration remains small, less than $30,000, compa to some other countries averaging around $1,000,000. There is also a large gap in the number of commercialized products based on universities' inventions compared to other countries.

Industry-academia collaboration in Japan today

Development of domestic systems took place and strongholds were established. Now, we are moving to a new policy stage. The starting point is a Keidanren (Japan Business Federation) report titled "Towards Strengthening of Industry-Academia-Government Joint Researches" which was released in February 2016. The report describes the need for promotion of full-scale, top-level industry-academia-government collaboration along with reform of domestic universities and research institutes.

Clear policy and targets were subsequently introduced in a 2016 report which includes tripling companies' investment in universities and national research entities by 2025. The related ministers, METI and MEXT, are formulating guidelines in a working group compiling prescriptions for and thoughts on the issues of universities and national research and development corporations seen from industry. We established the council of industry-academia-government dialogue for the promotion of innovation in July 2016, focusing on approaches for full-scale collaboration among the sectors. We will try to finalize the guidelines by November 2016.


Japan is reforming its universities, in part to promote strategic public-private partnerships. Many types of cooperation exist, such as staff mobility, that create linkages in R&D projects between companies and universities, sometimes involving governments. In Japan, the government has sponsored new innovation centers.

It is difficult to measure this cooperation within the Organisation for Co-operation and Development (OECD). Many studies focus on R&D and personnel funding, and Japan does not rank very highly under either measure. However, we may be failing to capture important factors. In the United Kingdom, for example, university professors are very engaged in the community, providing scientific advice and sitting on company boards; much of this cooperation is not being captured. I will focus on one type of cooperation: industry-government-university partnerships.

First, we need to put this into context. In Japan, most R&D is performed by the private sector. In the public sector, Japan exists between a university-based system and other countries, in which government labs are more important for R&D than universities.

Firms that engage in R&D cooperate much more. In Japan, close to 60% of firms that engage in R&D also cooperate with other companies or with universities or government labs. However, many small firms do not engage in R&D, and it is more difficult to involve them.

Larger firms have more experience working with universities than small firms. This is one challenge for policymakers: how to increase innovation among small firms.

The market provides the biggest source of external information. Many companies receive feedback from users. This includes big data being generated by smartphones or search engines that provide feedback on user preferences. Suppliers also provide information. By contrast, universities provide less information. Innovation occurs downstream and is more about competition. Because of this, market information is more important than university information. Industry depends on knowledge of public research.

In Japan, most research in health technologies, e.g., medical devices and drugs, is done by the public sector. The number of companies citing universities patents is increasing throughout the OECD and also in Japan.

Tax breaks are not the only means of encouraging cooperation. Increasing research quality is more important than just giving tax breaks to companies that work with universities.

Partnership policy

It is important that these partnerships focus on the specific challenges we face in our respective countries (e.g., competitiveness challenges for Japan with the rise of China and other countries, aging in Japan and Europe, etc.). Some governments have budget problems. They already give money to universities and tax breaks to companies. How can they increase the impact of all this support by promoting cooperation?

In Europe, regional funding for innovation is moving toward trying to reduce fragmentation, and countries work together to support each other and capitalize on their respective strengths. Smart specialization means reducing duplication. It's a national development strategy in Europe.

We then also have the lead market strategy. The internet is part of a global strategy to reduce fragmentation, but only in sectors that have a global market; photonics in Europe, for example. A great deal of effort is being made to cooperate in achieving critical mass to set global standards and develop export markets. These are some of the reasons why partnerships are much more important now. They reduce duplication and reduce budgets while improving competitiveness in global markets.

The most traditional government cooperation policy is just to increase cooperation with the goal of increasing the rate of innovation. We want companies to engage in more R&D and thereby produce more innovations. This approach pays little attention to the direction of innovation. Many industrial policies targeting specific sectors have been pursued. These then shifted to more horizontal policies geared to improving competitiveness across the board. Now, soft industrial policy is returning, targeting broad areas such as health or energy in an attempt to promote these areas through industry and partnership support. This is a new trend in strategic partnerships seeking to accelerate innovation in these specific areas. They evolve not just in industry and universities but also with government support. High-risk technology projects require a much more strategic approach to partnerships.

Partnership actors

Many different actors are involved in this: small and large firms, universities, and institutes are all trying to involve the whole value chain of research and innovation. There is a focus on users. Autonomous vehicle programs, for example, require working with companies, research, standards, etc. To go beyond market failures that justify government support, this is more about failures in demand articulation but also technology lock-in. These partnerships aim to break some of the lock-ins. Partnership is about the future and thinking about how to explore and capture value in new areas that are going to be important.

The difference between partnerships and other types of cooperation in that this is a top-down initiative. The selection of the actors is a bottom-up process, but there is top-down steering. Intellectual property (IP) is also very important for this cooperation. For universities, it has been a learning process. Many universities had not protected their IP in the past, and they went to the opposite extreme: trying to protect too much, patenting everything, restrictive licensing, etc. More universities are now trying to be more adaptive by moving into the middle ground and trying to organize and take a more flexible approach to IP, because what works for MIT or Stanford may not work in your country or university, so it has to be more adaptive.

The ability of partners to enter and exit a partnership, depending of course on the contract, is another important issue, as is financing. For these partnerships, when does public support materialize? The role of equity financing must also be considered. Universities can invest in some equity now. Government agencies in turn can bring their operational focus and introduce different financial instruments, e.g., milestone payments, to nurture partnerships.

Partnership examples

Some examples of strategic public-private partnerships focus on grand challenges. One example is America Makes, the U.S. National Additive Manufacturing Innovation Institute. The government dedicated federal funds for this, but the universities partner with companies to compete for funding. The government is funding nine of these centers in the United States to invest in these new manufacturing technologies, and is involving different departments: Department of Defense, Department of Commerce, etc. This is a strategic goal-oriented partnership bringing in some small companies, universities, and even a community college.

Another example at the institutional level is Austria's Christian Doppler Laboratory and associated entities. These organizations perform basic research that companies strongly demand. The institutes are funded by the region and the government around projects with companies. There is very inclusive stakeholder engagement, with company personnel, academics, and ministry officials all on board with the mission of fostering partnerships in these areas. 50% of a partnership is financed by the public and 50% by the commercial partners. If small and mid-sized enterprises (SMEs) are involved, the public share increases.

Funding and intellectual property

In some countries, funding is now being converted into loans. This is happening in the United Kingdom, for example. That is another way of helping to ensure that the government gets something back from this. If the project is unsuccessful, the loans are a loss.

Intellectual property rights (IPR) is also important. How copyrights, trademarks, trade secrets, etc. are organized among the various partners is very important, and there are many models. For people working on these on a daily basis, the message for the government is to have good IPR management plans for these projects. The details have to be designed according to the particular institute or area. Of course, evaluation is very important.

Good practices

Another example is the consortia for knowledge and information in the maritime sector in the Netherlands. The Netherlands is strong in this sector (e.g., shipping, transportation) and is also using this as a resource for its own economy, so there is a focus on the entire maritime economic value.

We have observed several major good practices. Project selection is very important. Well-designed partnerships are not very easy to create but are very important. From a longer-term perspective, an economic approach should be taken to clarifying the issues surrounding ownership and control and then taking on the value chain. Financial transparency is important. For universities, some of the financing challenges relate to unexpected research overhead and uncalculated surprises. Effective governance, with strong and clear arrangements especially in partnerships with many actors, is important. The sharing of IP is very important upstream, but incentivizing is very important downstream. Clear assignment is desirable. With regard to human resources, how do you incentivize researchers to get involved in partnerships? One focus has been getting Ph.D. students involved, and not just professors. Internationalization is another important component. Evaluation and impact are also very important. There are lots of new methodologies. The European Evaluators Network shares practices and methodologies for evaluating projects. There is a case-by-case application of the rules.

For the OECD, these partnerships are a major shift from a horizontal approach to a more targeted and strategic one. The three major main factors are project design, contractual design, and incentivizing firms to bear the market risks while the research risks are borne by the public sector.

Commentary by Anders KARLSSON

We have learned a great deal about the various good practices for public-private partnerships. One of the key issues here from METI's side is getting a stronger industry commitment. Full-scale projects are desired.

Since I represent an industry, I will explain why we are here. We are increasingly involved in collaboration with academia. We need access to talent and to be at the forefront of technology, and as the world is increasingly data-driven, we are quite involved in working with metrics and data. We have a lot of startup collaborations. In the digital economy, it is important for industries to avert disruption. How important are universities for growth? Why does less collaboration take place than expected? Which model works? I will talk about these and make some personal comments as well.

To take the case of MIT, its graduates have contributed to companies, the regional economy, and the larger economy. The impact of industry-academia collaboration is often discussed only in terms of direct impact, but there is also an indirect impact that is actually stronger than the direct impact in many cases. It's the case of having a culture in which students can be useful for companies. In Japan, that might mean instilling a more entrepreneurial culture. MIT is one of the key success cases globally in this respect. Most universities probably want to do something similar.

In the Japanese context, we quite often see risk minimization in many discussions. Cross-sector collaborations imply risks. You are setting up new partnerships with people you haven't worked with. Departments may not collaborate enough within a university. This may lead to things taking time. The professors have been quite strong in Japan's case. A strong local economy is necessary. Remember how much of Japan's GDP is export? It is less than 20%. This means that Japan may have naturally had a domestic focus, but if Japan wants to attract global companies for industry collaborations—which should be one of the objectives—it will need global experience from the universities it collaborates with and not just academic excellence.

Venture funding is not directly related to public-private partnerships, but many of the startups we have been speaking to here in Japan received their first funding from outside Japan. Also, from a company perspective, if we go in or other companies go in, there are expectations that the university should also commit financial resources. There are expectations of a quid pro quo. METI has its programs, e.g., via the New Energy and Industrial Technology Development Organization (NEDO) and with MEXT Center of Innovation and some on-campus labs exist, but maybe due to geography or a lack of space, not many research parks exist. Many other countries have research parks which have generated impressive amounts of industry funding. In one case, BMW contributed $75 million. In Japan, there are Tsukuba Science City and the Nanotech Innovation Arena. One thing to look at perhaps is the issue of "red tape," i.e., complicated procedures. Why should things take the time that they have in some cases?

Elsevier is increasingly a data science company. We have strategic collaborations with University College London and Carnegie Mellon University. Actually, just on the day of this talk, we set up a data science collaboration with the University of Amsterdam. Companies go in because they want capability development. Long-term joint investments should be made. We also want executive-level access. Interestingly, I believe Japan has had very strong industry collaboration that is not reflected in the statistics. Many Japanese professors have very good industry connections, and through them, the best students are connected to the industry.

With regard to the exchange of knowledge, looking into our data on mobility of researchers, movements are more domestic than international in Japan. This should be considered when determining how public-private partnerships will work, because the United Kingdom, for example, has a global pool. I think what is important here is: What does that mean on the skill side? And what does it mean for global competitiveness in innovation?

Open innovation and open science are examples of how Elsevier is active. As a simple example, using our system, all campuses in the state of Texas are networked with all of their expertise and equipment in order to attract companies and other players who want to collaborate with the University of Texas system. This is part of open science. In Japan, Keio University is also doing this, and there are others. One challenge we have seen is that there are many more universities who have these types of systems but they are not open. In the case of the United States, it is open to bring in the expertise, whereas in Japan, some hesitation remain.

What does the balance of the funding system (i.e., government vs. private investments) mean in terms of the types of collaboration we can see? There is not a one-size-fits-all model. What would be best for Japan?

Japan Science and Technology Agency President Michinari Hamaguchi has a new "Mirai Project" which is attempting to generate stronger collaborations. I heard that your committee is between MEXT and METI. That's really good. Inter-ministerial inter-funding agency collaborations are probably going to help a lot in crafting these policies.

In the European context, the concept of the knowledge triangle has been important. This model attempts to discuss the linkage between higher education industry-minded students and the research system. The triangle is composed of research, education, and innovation. I believe this should be discussed in Japan. One of the prime reasons for collaborating with universities is the access to talent.

ICT can also play a role in lowering barriers to collaboration. How can open science in a major context change the playing field overall? In Europe, this is a very strong policy drive. In Japan, in the basic science and technology innovation plan, open science comes in as one part of open innovation. How to best reap the benefits of what ICT allows you to do is something that would be interesting to discuss further.

To summarize, we see very interesting ambitions in Japan to have a stronger commitment to full-scale projects. The knowledge applied from OECD and the good practice suggestions are truly excellent.


Q1. What is your opinion of Dr. Karlsson's comments?

Collaboration is the way forward. It's essential from a business point of view. Dr. Karlsson reminded us that no one size fits all. We have to work within different ecosystems. The OECD offers principles, and the devil is in the details—who the actors are and their agendas, etc.

One point Dr. Karlsson mentioned is that we are not measuring very well. Mobility indicators are interesting, but with today's technology, we could do a network analysis of researchers in Japan and the projects they are working on. With this information, we could identify strengths that could be scaled up, and what strengths need greater support.

Global partnerships are important. The Koreans and the Chinese are moving in this direction more rapidly than Japan, engaging with universities abroad and so on. Japan should watch this trend carefully and not miss it. Japan does have some partnerships with top U.S. universities, but they should be broadened. Perhaps a study should be conducted on successes and possible improvements.

Venture capital is also important. It's a problem for many countries, but I think now the issue is to raise awareness in the venture community of the potential of these partnerships. They may not be paying enough attention to universities in Japan. Having the forums and engagements that happen in other countries would be a good awareness-raising initiative that could bring venture capitalists to universities.

Q2. Your points about developing an effective incentive scheme are very important. The current problem in Japanese academia, not just in the social sciences but also in science and engineering, is that for a long time we have had a very big multi-storied silo. We are having a hard time breaking that. We are having trouble reconciling this academic mindset with the needs of industry, which are based on a shorter economic cycle. It is impressive that the Americans and Europeans talk about silos. I don't see them as such a huge problem in your systems, but you have already noticed that they are a phenomenon to overcome. How can we break out of these silos?

One experience we have seen in many countries is to award some share of research funding on a competitive project basis. These are generally more short-term funding. This helps reduce or break down some of the silos. Long-term projects also deepen research, but it creates empires of professors and their departments.

There will be a need to collaborate across departments and so forth. Japan is putting forward the challenge of Society 5.0, and it will be very important to have strong involvement both in shorter and long-term projects from the arts, humanities and social sciences as well. Otherwise, there is a risk of being too technology-focused. This technological focus may not solve the challenges you are addressing in Society 5.0. It's very much up to funding agencies to provide the necessary incentives. Of course, silos exist and things have not been easy. In Sweden, the attitudes of professors changed greatly in the early 2000s before the IT boom. Before that, almost no professors collaborated with industry. During the IT boom, most professors collaborated. Funding incentives partly changed the mindset. It has not been easy there, either.

We have also seen success where the executive leadership of a university makes it a priority. With big data open sciences, universities are creating chief information officers to create open data plans for entire university departments to make this more efficient for the research process and also in terms of supporting different departments and strengthening the abilities of the social sciences to use data could also impact the public research.

Q3. Your point about the motivation of companies was interesting. What about on the university side?

One motivation is competition for talents. Universities are competing for professors and students by providing them with favorable conditions. Autonomy is important for long-term research and cooperation. Competition is important because attracting talent will get more funding.

Reward mechanisms are important. For instance, is it well recognized that you have industry collaboration in the academic promotion system? It is greatly appreciated at some universities in some countries. Many of the most successful groups also address mid-term industrially relevant projects alongside their basic research. If reward mechanisms exist and they are recognized on the university side, they make a strong incentive.

The role of media and communications is also important. They need to highlight startups and entrepreneurs.

Q4. I have a technical question for Dr. Karlsson. I would like to know how the inflow and outflow figures were calculated on your slide.

It's calculated based on the affiliations of researchers during the 15-year time period we have data from. The movement from academia to industry can be followed by affiliation.

Q5. Regarding motivation, in the case of the usual public-private partnership, the goal is for the government or public sector to be customers or sponsors of the private sector, so there is some incentive for us to collaborate. In the case of industry-academia collaboration, a university would be our sponsor. What is the main objective of calling a public-private partnership like this a "collaboration"?

Companies cooperate with universities because of talent, that is, access to talent that they can later hire but also competence in specific areas of research that are important to the companies' longer-term objectives. Excellence in research will attract companies.

*This summary was compiled by RIETI Editorial staff.