Who Invents?: Evidence from the Japan-U.S. inventor survey

Author Name John P. WALSH  (Georgia Institute of Technology) /NAGAOKA Sadao  (Research Counselor and Faculty Fellow, RIETI)
Creation Date/NO. July 2009 09-E-034
Download / Links


Human resources are increasingly seen as a key to innovation competitiveness, and there is a need for detailed, systematic data on the demographics of inventors, their motivations, and their careers. To gain systematic data on who invents, we collected detailed information on a sample of inventors in the U.S. and Japan (the RIETI-Georgia Tech inventor survey). The data come from a unique set of matched surveys of U.S. and Japanese inventors of triadic patents, i.e., patents from patent families with granted patents in the U.S. and applications filed in Japan and in the EPO, with data from over 1900 responses from the U.S. and over 3600 responses from Japan.

Based on these survey data, we compare the profiles, motivations, mobility and performance of inventors in the U.S. and Japan. Overall, we find some important similarities between inventors in the U.S. and Japan. The distribution across functional affiliations within the firm, by gender, by educational fields and their motivations, are all quite similar. In particular, in both countries we find inventors emphasizing task motivations over pecuniary motivations. Firm-centered motivation (e.g., generating value for my firm) is also an important reason for inventing and this reason is relatively more important in the U.S. than Japan. Their distribution across types of organizations is quite similar. The percent of university inventors is nearly the same in the two countries, and the distribution of these inventors across technology classes is also quite similar. However, the percent from very small firms is significantly higher in the U.S.

There are a few important differences. American inventors are much more likely to have a Ph.D. American inventors are older (even controlling for differences in the share of the inventors with Ph.D.s). The modal Japanese inventor has his first invention in his 20s, while for the U.S., the mode is the early 30s, and we also find many more American inventors over age 55 at the time of their triadic patent invention. In both countries, older inventors tend to produce higher value patents. American inventors are also much more mobile (although Japanese inventors with Ph.D.s also have high rates of mobility, mainly in the form of secondments). In the U.S., mobility tends to decline with age, while in Japan, mobility is higher for older inventors (likely due to the differences in retirement ages in the two countries). In both countries, mobility is associated with greater access to outside information.

Finally, we find that foreign-born inventors are very important in the U.S. (we did not collect data on country of origin for Japan). Overall, these results suggest that inventor characteristics may be important for firm performance, and that institutional differences may affect the profile of inventors in each country, although the inventors of the two countries are very similar in many respects. Future work will examine how these cross-national differences in inventor profiles affect innovation in each country.