Abstract/Summary

Canadian firms invest considerably less in research and development as a proportion of GDP than in many other OECD countries. As a result of low private sector research intensity, universities represent a comparatively large proportion of Canadian research. In order to improve university technology transfer and the absorptive capacity of its industries, Canada offers among the highest rates of government support for universityindustry research collaboration (UIRC) in the world. The government granting agencies that administer these subsidies seek to identify and fund the UIRC projects that have the greatest likelihood of commercialisation. There is considerable debate among practitioners about what UIRC characteristics are associated with commercial outcomes. Although the mechanisms of effective research collaboration and university technology transfer have been well studied, the academic literature on this specific problem is surprisingly sparse considering its growing importance to policy makers. This study examined the relationship between the characteristics of various stakeholders in UIRCs, namely academic researchers, universities, firms and governments, and commercial outcomes from UIRC. Specifically, three hypotheses were developed and tested based on the unique context of Canada’s national innovation system and the extant literature. First, Hypothesis 1 built upon the concept explored in previous studies of how economic behaviour is “embedded” in social networks to posit that researchers who are less embedded within academia will have a higher likelihood of commercial outcomes from UIRCs. Next, motivated by the growing body of literature that has found government subsidies help to stimulate greater private sector research expenditures, Hypothesis 2 suggests that UIRCs with higher cash and in-kind contributions by firms will have a higher likelihood of commercial outcomes. Finally, Hypothesis 3 proposes that UIRCs in industry sectors with higher research intensity will have a higher likelihood of commercial outcomes, since prior studies have demonstrated that research intensity improves firm absorptive capacity. A novel dataset was developed from the historical records of the Ontario Centres of Excellence, a government granting agency in the province of Ontario, and from other public sources. The UIRC project was the unit of observation and the size of the sample was 682 observations. The dependent variables represented whether or not the UIRC project achieved a commercial outcome. Five independent variables were used to test the hypotheses using binomial and multinomial Logit regression, and 19 additional control variables were included in the model. Hypothesis 1 was tested using a novel categorisation of researchers based on their level of embeddedness in academia, and found that it was significantly associated with commercial outcomes. However, the results suggested that the directionality of the relationship was opposite to what was hypothesised. Additional testing confirmed that, contrary to what was hypothesised, more embedded researchers have a higher likelihood of commercial outcomes. Therefore, the findings may shed further light on mixed results from previous studies by exploring the commercialisation behaviour of certain categories of embedded researchers. Hypothesis 2 was tested using separate measures of firm cash and in-kind contributions to UIRCs. The results found that in-kind contributions are significantly associated with commercialisation, but that cash contributions are not. In fact, in-kind contributions are positively associated with licensing outcomes in particular. This study is among the first to link firm contributions to UIRCs with their commercial outcomes at a project level. Preliminary testing indicated support for Hypothesis 3 by finding a significant and positive association between industry sector and commercialisation. Additional testing found further evidence that UIRCs in industry sectors with higher research intensity had a higher likelihood of commercialisation. The findings suggest important industry differences in UIRC commercialisation patterns. However, the absence of data on UIRCs in life sciences, which represent approximately 50 of all university technology transfer activity, is an important limitation on their generalisability. The study makes four recommendations to policy makers and government granting agencies based on its findings: 1. Develop awareness and education programs that encourage older, more career advanced and high-quality researchers to become involved in UIRC and commercialisation. 2. Design UIRC support programs and selection criteria to encourage in-kind contributions by firms. 3. Concentrate efforts on developing world class research capabilities and commercialisation infrastructure at a small number of large universities. 4. Focus on supporting research collaboration between universities and the most research intensive industries to maximise the likelihood of commercialisation.