Three essays on the economics of science policy: The impact of funding, collaboration and research chair

This thesis studies the determinants that influence the number of citations, the effect of having a research collaboration with top-funded scientists on scientific productivity, and the effect of holding a research chair on scientific productivity. Based on a review study by Bornmann and Daniel (2008), one can argue that non-scientific factors determining the decision to cite do not significantly alter the role of citation as a measure of research impact. Assuming that the number of citations is a good measure for research impact and, in turn, for a certain kind of quality, we showed that the number of articles and the visibility of a researcher, the impact factor of the journal, the size of the research team, and the institutional setting of the university are the important determinants of citation counts. However, we have found that there is no significant effect of public funding and gender in most of the domains examined. The point that funding amount is not a significant determinant of citation counts does not necessarily contradict the positive effect of funding on scientific productivity.;We also developed a theoretical model and proposed some hypotheses about the effect of collaboration with top-funded scientists on scientific productivity. We then validated the hypotheses with empirical analysis and showed that such collaboration has a positive effect on scientific productivity. This significant effect may exist through different channels: transfer of tacit knowledge, more scientific publications, economy of scale in knowledge production because of better research equipment, and expanded research network. The results also verified the positive effect of funding, the positive effect of networking (measured by number of co-authors), the inverted U-shaped effect of age, and the fewer number of publications by women compared to men.;Finally, we made a distinction between different attributes of research chairs and their effect on scientific productivity. One of the important questions is to find out whether a research chair still has better scientific productivity (compared to non-chair holders) after controlling for the research funds available to the researchers. To investigate that question, we employed a matching technique to identify pairs of scientists (chair and non-chair holders) of the same gender, funding and research field. After such matching, we found that the effect of the Canada research chair program on scientific productivity remains significant and positive, while the effect of industrial chairs and the chairs appointed by the Canadian federal granting councils (NSERC and CIHR) become non-significant. This finding highlights the effectiveness of our matching technique methodology; because before matching, holding any type of chair had a positive and significant effect on scientific productivity.;This finding highlights the special attributes of the Canada research chair program, which are not replicated in other chairs. Those specific attributes may significantly push scientific productivity. For example, Canada research chairs are generally associated with some degree of prestige or higher visibility to recruit talented students or to have research collaboration with top scientists in the field. In addition, the Canada research chair program has a firm and efficient method of allocation (which is explained in the thesis). This approach institutionally synchronizes different chairs in universities and research fields. The fact that other types of research chairs, once matched with equivalent scientists, do not have an impact on scientific output in terms of quantity does not imply that these chair holders are lesser scientists, but that they are devoting part of their time to other endeavours of a more practical nature. Hence universities are maintaining a balance between the pursuit of pure scientific knowledge and its application to socioeconomic benefits. By solely studying scientific articles, we are missing a great deal of the university professors' activities. Although not trivial, future research should aim to cast a wider net on outputs, outcomes and impacts of university research.