Recent Research: How much does place matter for scientific output?

With federal research budgets under pressure and other nations accelerating their scientific investments, the U.S. faces critical choices about where to direct limited funds. Does concentrating resources in leading institutions best sustain discovery, or would broader geographic diffusion strengthen the nation’s overall research ecosystem? And, while discoveries are usually attributed to individual scientists or teams, the role played by research institutions in shaping where and how breakthroughs occur is not as well understood.  An academic research team recently explored these questions. 

In their NBER working paper Where Discovery Happens: Research Institutions and Fundamental Knowledge in the Life-Sciences, Amitabh Chandra and Connie Xu investigate how strongly the environment provided by universities and research centers (rather than just a prestigious name) determines the generation of fundamental discoveries that later drive biomedical innovation. Using publications in top life-science journals as a proxy for discoveries, they document geographic and institutional concentration and link this to downstream technological impact via patent citations. To estimate institutional causality, they compare annual research output before and after a scientist moves from one institution to another to isolate how much productivity changes with the institutional environment. The authors state that this empirical strategy allows them to isolate the institutional impact on a researcher’s output. 

The authors found that when scientists move to institutions that already host highly productive, influential scholars, the relocating researcher’s publication output rises significantly. Star faculty seem to function as magnets and multipliers, attracting top students and attention from journals and funders. By controlling proximity to star faculty, the authors determined that about two-thirds of this institutional effect is driven by the presence of these scientists rather than the institution itself. They also found that basic research papers produced at these institutions are more likely to be cited in patents, strengthening the link between discovery and downstream innovation. 

While Chandra & Xu stress institutional and star-driven productivity effects, other studies (Tomkins, Zhang, & Heavlin 2017, PNAS; Clauset, Arbesman, & Larremore 2015, Science Advances) show perception and evaluation biases tied to prestige also shape who gets published, cited, and funded. This new research doesn’t negate the earlier findings but suggests that both real productivity spillovers and prestige effects in evaluation systems are at work.   

If institutional investments yield more papers that are later cited in patents, as the authors contend their research finds, they then argue that when deciding between similar scientists, funders should favor those based at more productive institutions, since the institutional environment itself amplifies output. Similarly, institutional funding for shared research infrastructure may be more effective at boosting discovery than awards to individual researchers. The authors also state that policies aimed at recruiting star faculty, such as endowed chairs and enhanced startup packages, could have disproportionate, positive effects on knowledge creation for their institutions. 

Chandra and Xu acknowledge that concentrating funding in a few institutions may maximize output, but it may reinforce geographic inequality. They suggest a balanced portfolio pairing support for top centers with efforts to build capacity at promising institutions may strengthen the overall research ecosystem.