Abstract This study develops a pathway-based view of university knowledge commercialization by comparing university startup formation and patent licensing. Integrating insights from regional innovation systems, entrepreneurial ecosystems, and innovation intermediation, the study examines whether the effectiveness of university intermediation depends on the commercialization pathway and the regional innovation environment. Using panel data on university technology transfer in Japan, the analysis estimates Poisson models with university and year fixed effects and assesses robustness using fixed-effects negative binomial models. Regional innovation agglomeration is measured as technology-specific private-sector patenting intensity normalized by prefectural economic size or population. The results reveal meaningful pathway asymmetry. Basic research is consistently associated with subsequent startup formation, whereas its association with patent licensing is weaker and not robust across specifications. Complementarity between university-facilitated external linkages and regional agglomeration is more clearly observed for startup formation, especially in science-based fields and when agglomeration is measured relative to the local economic base. By contrast, there is no robust evidence that university-based TLOs become more effective for patent licensing under stronger regional agglomeration. Patent licensing is, however, more clearly associated with local joint research ties, consistent with the possibility that collaborative ties help reduce uncertainty surrounding the commercial applicability of basic-research outcomes. These findings suggest that university intermediation is route-specific and context-dependent rather than uniformly effective.
Building similarity graph...
Analyzing shared references across papers
Loading...
Nobuya Fukugawa (Thu,) studied this question.
synapsesocial.com/papers/69fed19ab9154b0b8287904e — DOI: https://doi.org/10.1007/s11192-026-05648-4
Nobuya Fukugawa
Tohoku University
Scientometrics
Tohoku University
Building similarity graph...
Analyzing shared references across papers
Loading...