RIETI Policy Symposium

Innovation Process and Performance: Findings and lessons from inventors surveys in Japan, the U.S., and Europe


  • Time and Date:
    09:45 - 18:00, Friday January 11, 2008
  • Venue:
    Otemachi Sankei Plaza
    1-7-2 Otemachi, Chiyoda-ku, Tokyo (Tokyo Sankei Bldg. 3F)
  • Language:
    Japanese/English (with simultaneous interpretation)

Summary of Proceedings

Part 1: Keynote Speech and Presentations (2)

Alfonso Gambardella Keynote Speech

Based on the inventors survey conducted in Europe, Professor Gambardella's speech dealt with policy implications.

(1) European survey was the precursor of the Japanese and U.S. surveys

  • The European inventors survey was conducted in 2003-2004 and targeted patents with priority dates in 1993-1997. Major European nations including Germany, France, Italy, Spain, the Netherlands, and the UK participated in the survey, which obtained 9,000 responses from 27,000 questionnaires sent out.
  • The survey consisted of five parts: (1) inventor's personal information, (2) inventor's education, (3) inventor's employment and mobility, (4) the innovation process, and (5) the value of the patent.
  • Pilot surveys were conducted prior to the main survey. We considered whether the survey should target inventors or managers, and decided that inventors were the appropriate target for a large-scale survey.
  • Confirming addresses for inventors was time-consuming due to changes of job or residence.

(2) Fundamental results of the Japanese, U.S., and European surveys are very similar.

  • The inventor community has common attributes worldwide. In Europe, as in other regions, more than two-thirds of inventors work for major corporations. The proportion of inventors from universities and public research institutes is low in all three regions. Academic backgrounds show the same trends, and the proportion of doctorates is high in the pharmaceutical field. Inventor mobility in Europe is similar to the U.S. The typical European inventor is 45 years old, with tertiary education and employed in an established firm.

(3) Policy lessons from survey results

  • Importance of large firms
    Survey results showed that more than two-thirds of inventors work for large firms. There is a possibility that policy to date has overestimated the contribution of universities and small- and medium-sized enterprises (SMEs) to patenting.
  • Policies to increase the supply of potential investors
    The proportion of female inventors is low. Policies making it easier for women to build career paths would increase the number of potential inventors.
  • Policies to motivate inventors
    As in Japan and the U.S., the satisfaction gained from solving technical problems is the biggest motivator for European inventors. Policy should preserve this inventor ethos. Doubts remain over the effectiveness of the invention compensation system used in Germany.
  • Knowledge spillover within and outside the organization
    Knowledge spillover most frequently occurs within an organization. Results showed that even if organizations are in close geographical proximity, knowledge spillover to other companies is not great. They also showed that more highly educated inventors have broader networks. Thus rather than creating a large number of Silicon Valley-type clusters, creating a large number of Mercedes Benz-type companies in each area may be more effective for knowledge spillover. More emphasis should be placed on local formation of human capital than on local spillover.
  • Innovation policy should be about demand
    Suppliers, customers, and product users are inventors' most important knowledge sources. In devising innovation policy it is important to first think about demand.
  • Patent markets increase utilization of patents
    In many cases unused patents are not necessarily of low quality. Large firms have a higher number of sleeping patents than SMEs, and transaction costs often create a barrier. Compared to larger firms, SMEs utilize patents efficiently. A European patent market is necessary for greater utilization of patents held by large firms, and startups from large firms should also be encouraged.
  • There is no shortcut to innovation
    Forty percent of inventions are expected outcomes of research projects, another 40% are expected by-products, and the remaining 20% are serendipitous. Thus compared to innovation inputs, outputs have a high degree of uncertainty. The classic method for boosting innovation is to increase investment in R&D and in R&D personnel.

Hideo Owan Presentation

Using the results of the Japanese inventor survey, Professor Owan analyzed and reported on the relationship between inventor motivation and patent value, paying particular attention to the effects of factors including (1) differences between exploration of new capabilities and exploitation of existing knowledge and (2) the endogeneity of companies introducing monetary rewards for inventors.

Questions on inventor motivation
Using a five-point scale, respondents were asked to rate the importance of each of the following seven factors: (1) Satisfaction from contributing to progress of science and technology, (2) Satisfaction from solving challenging scientific problems, (3) Enhancing the performance of their organizations, (4) Career advancement and better work opportunities, (5) Reputation and prestige, (6) Improved research conditions such as bigger budgets, (7) Monetary reward.

Survey results
Inventors were categorized by age, academic background, and firm size. In almost all categories the main motivators were the interests in solving technical problems (item 2 above) and contributing to science and technology (item 1 above), and the latter, in particular, was termed "taste for science." Monetary reward was the lowest-ranked motivator.

Result of regression analysis

  • "Taste for science" has an extremely high correlation with the number of patents and the value the patent. Thus researchers' taste for science is a key factor in determining their productivity, since there are two possible explanations for this correlation: (a) early access to cutting-edge science and technology by researchers with a strong taste for science has a positive effect on their R&D productivity (productivity effect) or (b) there is a high correlation between taste for science and research ability (screening effect). The impact of (b) is regarded as particularly high.
  • Monetary reward has a certain effect. This effect does not differ greatly between exploration (commercialization of potentially useful scientific or technological discoveries or exploration of new technical seeds) and exploitation (solving technological issues important to the business of the firm).

The following comments were made in response to the presentation.

Professor Hall:
Motivation for individual researcher effort is a separate issue from the question of whether corporate R&D efforts are inadequate. It is not correct to argue that high researcher motivation can offset inadequate corporate investment in R&D.

Professor Fujita:
This is very interesting research and I look forward to even more in-depth studies in the future. I hope such research will play a useful role in establishing more efficient reward systems by clarifying the effect of monetary rewards. I also hope that we may be able to stimulate a taste for science throughout society as a whole, in the same way that Japan has a "taste for baseball."

Questions and Answers

The following question was received from the floor.
Q: Professor Gambardella's keynote speech showed that knowledge spills over more frequently to distant units of one's own company than it does to people in other companies nearby, but isn't this simply a result of the total number of people, in other words the fact that the population is larger?
A (Professor Gambardella): The question we posed was, "who do you talk to in the course of invention?" The results showed that respondents most often talk to people in their company, in other words people in the same building or facility. However, the number of people in the same building or facility should represent the smallest population. Thus the size of population does not decide the results.

Q (Professor Nelson): According to research conducted by the Massachusetts Institute of Technology (MIT), communication patterns exist within projects. Sometimes the roles of internal interaction and external interaction are divided among project participants. Did you research the communication patterns for projects?
A (Professor Gambardella): We didn't especially conduct that kind of research, but we did identify a relationship between academic background and communication patterns. We determined that academic background affects the scope of potential interaction, with Ph.D.s able to communicate at the international level and those with university degrees restricted to the local level.

Q (Regarding Professor Owan's presentation): Evaluation of the different effects of motivation in exploration and exploitation and the differences in inventors' "taste for science" are very interesting research topics, since they would enable more targeted policies and strategy. What kind of research are you planning for the future?
A (Professor Owan): I intend to conduct further analysis using the survey results once I have made clearer definitions of exploration and exploitation. I would also like to analyze the effect of corporate strategy choices on exploration and exploitation activities. And I want to shed further light on the relationship between corporate diversification and these two activities as well as why exploration is said to be greater in the U.S. than in Japan.