Using Data Science to Examine the Link Between University R&D and Innovation

Paper Session

Friday, Jan. 6, 2017 7:30 PM – 9:30 PM

Swissotel Chicago, Zurich F
Hosted By: American Economic Association
  • Chair: Dan Black, University of Chicago

The Link between R&D and Entrepreneurship

Ron S. Jarmin
,
U.S. Census Bureau
Nikolas Zolas
,
U.S. Census Bureau
Nathan Goldschlag
,
U.S. Census Bureau
Julia Lane
,
New York University

Abstract

The reason for the secular decline in entrepreneurship is not well understood. It is evident in all sectors of the economy and almost all regions. One approach to stimulating innovation and entrepreneurship has been to increase investments in science: the U.S. federal government contributed nearly $38 billion for university-based research in Science, Technology, Engineering, and Mathematics (STEM) in 2014. However, there has historically been little evidence about the links between investments in university research and innovation - largely because surveys cannot capture the complex ways in which scientific ideas are created, transmitted and adopted.<br />
<br />
This paper examines the relationship between the funding of research teams - in terms of structure, field and type of funding - and the subsequent propensity of members of those teams to start up businesses. It also examines the subsequent survival and productivity growth of those startups. <br />
The work is now possible because of a new data infrastructure resulting from collaborations between the Census Bureau's Innovation Measurement Initiative, the National Science Foundation and the Institute for Research on Innovation and Science at the University of Michigan. The infrastructure links universe data on all people employed on research grants, their funding, and their economic and scientific activities.<br />
<br />
This paper is the first to directly trace the pathways from the bench to the workplace at a large scale, using universe data from 25 universities covering about 25% of federal university based R&D. It is the first to use universe data on workers (the LEHD data) to draw comparison groups of individuals employed both within the university and from other R&D intensive businesses. And it is the first to use universe data on business startups to compare the dynamics of university sourced entrepreneurship with other types of entrepreneurship.

Pathways to Production

Erling Barth
,
Institute for Social Research
James C. Davis
,
U.S. Census Bureau
Gerald R. Marschke
,
State University of New York-Albany and NBER
Andrew Wang
,
Harvard University
Sifan Zhou
,
Harvard University

Abstract

Science funding agencies often require researchers to demonstrate their project’s prospects for “development of a diverse, globally competitive STEM workforce,” “increase[d] partnerships between academia, industry and others” (NSF, 2016), and other goals beyond the creation of scientific knowledge. This paper attempts to measure these wider impacts of scientific research.<br />
We use the newly created Census data infrastructure that links university grant transaction data to Longitudinal Employer-Household Dynamics (LEHD) data to map employment linkages between universities and industry. First, we ask, what are the flow rates of new STEM workers—post-docs and recent doctorates—into research intensive firms, industries, and regions? startups and established firms? high- and low- productivity firms? local and out-of-state employment?<br />
Second, we estimate the impacts of and returns to university-based human capital accumulation by STEM workers. The sudden increase in science funding under the American Recovery and Reinvestment Act of 2009 (ARRA) increased demand in the academic sector for post-graduate researchers, both lengthening existing post-graduate research engagements in universities, and increasing the likelihood that recent graduates, especially doctorates, obtain post-graduate employment in universities. We estimate the impact of increased university-based research training on career paths, including the likelihood of obtaining a faculty post, and for researchers who enter industry, which firms they match to, and their wage outcomes.<br />
Third, we investigate the extent to which a firm placement depends on the history of previous placements from the same university. Such a correlation could be evidence of “hiring chains”, or of specific knowledge links between the research and teaching at a specific university and the production technology of particular firms. The hiring patterns we uncover between universities and industry reveal important features of the labor market for specialized skills, and increase our understanding of how university research contributes to the diffusion of new ideas in the economy.

Financial Advice and the Entrepreneurial Spillovers of Basic Research

Francesco D'Acunto
,
University of Maryland
Liu Yang
,
University of Maryland

Abstract

We test for the effect of informal financial advice on the establishment and subsequent performance of entrepreneurial ventures that commercialize the results of basic research. To this aim, we construct a unique data set that includes: (i) the characteristics of the faculty recipients of federally-funded grants across 10 large U.S. universities, which produce innovation that can be commercialized through the establishment of startups; (ii) the likelihood that recipients establish a non-employer venture (iLBD) or an employer venture (LEHD), as well as the job growth characteristics of these ventures; and (iii) the network of neighbors in the locations where the recipients’ reside, including the occupation titles and demographics of the neighbors (ACS/Decennial Census). We use the presence of financial-sector employees among the faculty’s network members (spouses or neighbors) to test for the effect. We compare faculty grant recipients in similar areas of research, obtaining grants of similar sizes in the same rounds of funding, and at similar stages of their academic careers, but belonging to networks with different levels of exposure to informal financial advice from family and friends. Financial advice from one’s social network is informal because advisers are not paid fees for providing their service. Therefore, the paper broadly tests for whether advice is a positive externality of one’s social networks, which is valuable to the individual entrepreneurs as well as to economic growth.

Research Funding and Subsequent Entrepreneurship: The Role of Underrepresentation

Catherine Buffington
,
U.S. Census Bureau
Ben Harris
,
U.S. Census Bureau
Fiona Feng
,
New York University
Bruce A. Weinberg
,
Ohio State University

Abstract

Federal funding affects both who does research, and the environment in which research is done. In a recent study, 6 in 10 female doctoral recipients had been supported by federal research funds, compared to 7 in 10 male doctoral recipients. Federal funding also appears to be highly correlated with the pipeline of researchers going into different fields; particularly into R&D fields and the decision to pursue postdoctoral fellowships.<br />
<br />
This paper uses rich new Census Bureau data linked to detailed information on the individuals supported by research funding to examine the effect of both the type and structure of federal funding on the outcomes of underrepresented students. It makes use of rich measures on student characteristics, including their race, gender, place of birth, marital status and presence of children. It constructs new network theoretic measures of team environment, based on the characteristics of all individuals working together on research grants. It also includes information about household and family structure in the model. It also examines two types of outcome measures - placement in R&D performing, high technology or young and small firms - as well as the propensity of underrepresented groups to start up businesses.
Discussant(s)
Beth Webster
,
Swinburne University
Kaye Husbands Fealing
,
Georgia Institute of Technology
Reinhilde Vergeulers
,
University of Leuven
Richard B. Freeman
,
Harvard University
JEL Classifications
  • C1 - Econometric and Statistical Methods and Methodology: General
  • O3 - Innovation; Research and Development; Technological Change; Intellectual Property Rights