Trends in and influence of regional federally funded research and development in the US

Abstract

Over the last few decades, although US gross domestic spending on Research and Development (R&D) as a percentage of GDP has risen from around 2.27% in 1981 to 2.74% in 2016, federal funding for R&D has fallen steadily, from 1.19% to o.81% over the same period. These changes reflect a broader shift in the US from a government-driven R&D model to a business-driven model. Towards the goal of identifying the regional economic impacts of federally funded R&D, I first build on previous work to develop a method to obtain federal funding for R&D at granular geographic levels using Natural Language Processing (NLP) methods to automatically classify open data on federal contracts and grants as R&D or non-R&D awards. This method results in a 95% accuracy rate in classifying federal awards, and covers 56% of US federal R&D obligations made in the year 2016. As underreporting issues in the data source are addressed, this method will yield higher coverage rates, thus creating a unique dataset that affords opportunities to study the regional impacts of federally funded R&D. Next, I adapt Hausman, N. (2012). University Innovation, Local Economic Growth, and Entrepreneurship to identify the employment-generation effects of federally funded university R&D and compare impacts of overall R&D funding to the employment-generation arising from R&D funding provided to specific academic disciplines. I find that the employment-generation effects of federally funded computer science R&D are significant and much more pronounced than the corresponding effects of overall federally funded university R&D.