Association for Public Policy Analysis & Management

Despite large public investments, evidence on the marginal product of investments in research is limited. We use data on 137,499 principal investigators (PIs) over 31 years supported by the National Institutes of Health (NIH), whose funding typically depends on 27 institutes and centers that specialize in specific diseases or body systems. We interact attachment to institutes and centers (measured in a researcher’s initial year) with politically-determined funding trends for each institute and center to obtain plausibly exogenous variation in funding. We estimate that increases in funding increase productivity and allow for non-linearities, and lagged effects. We also use rich measures of the quality of research to estimate how funding amounts are related to the quantity and quality of output. Finally, using rich administrative data from the UMETRICS project, we investigate how PIs spend research funding, which begin to shed light on the nature of the scientific production function.