Spillovers in the production of knowledge: A meta-regression analysis

• Published in: Research policy 2018-05, Vol.47 (4), p.750-767
• Main Authors: Neves, Pedro Cunha, Sequeira, Tiago Neves
• Format: Article
• Language: English
• Subjects: Knowledge production function; National innovative capacity; R&D; Research policy; Spillover effect
• ISSN: 0048-7333; 1873-7625
• DOI: 10.1016/j.respol.2018.02.004

•We develop a meta-analysis on the spillover effect in the production of knowledge.•Meta-regressions are estimated using hierarchical linear models.•The average spillover is inferior but close to one and there is no publication bias.•Spillover estimates vary across studies’ samples and knowledge production functions.•The results have important implications for related policy, theory and empirical research. The production of knowledge was subjected to quantitative analysis in the second half of the twentieth century, following Arrow (1962). The determinants of knowledge and the externalities present in the innovation process were discussed with immediate policy influence. In particular, the presence and strength of the spillover of the pool of past knowledge has encouraged high subsidization of R&D in the most developed countries. We survey the empirical literature on the spillover effect in the production of knowledge and implement a meta-analytic regression. We discover that the average spillover effect is less than but close to one and is highly significant. We also find that the spillover effect tends to be greater when the estimation of knowledge production accounts for foreign inputs, and it tends to be lower when the estimation includes only rich economies, regional data are used, and the pool of knowledge is not the patent stock.