Measuring the Causal Economic Effects of Scientific Research

How to estimate the economic returns of public science is a longstanding but equally challenging topic in quantitative science studies. In this paper, we exploit the staggered foundation of the SENAI Innovation Institutes (ISI) in Brazil since 2012, to estimate their effects on GDP using a dif-ference-in-differences (DiD) approach. Building on historical and institutional insights from inter-views on the foundation process, we unravel the conditions under which the parallel trends as-sumption is likely to hold. Our analysis reveals that these institutes significantly contribute to GDP per capita, with an average treatment effect of 985 BRL (approximately €160). Moreover, by rely-ing on detailed project-level data, we were able to show that the effects come almost exclusively from genuine research projects and not from the provision of scientific services, such as metrolo-gy. Finally, tentative calculations suggest that the SENAI institutes may account for about 0.66% of Brazil's overall GDP, emphasising the importance of applied science in regional economic devel-opment and providing insights into effective collaboration between research and industry.