Research Universities and Regional High-Tech Firm Start-Up and Exit
De Silva, Dakshina G., McComb, Robert, Economic Inquiry
During the decade of the 1990s, a significant level of firm start-up activity was observed in the newly emerging high-technology industries. This activity tended to be concentrated in relatively few locations such as Silicon Valley in California, the Route 128 corridor in Boston, and the 1-35 corridor in Texas. As the regional employment dynamism and relatively high incomes associated with these new technology firms were widely coveted by regional policy makers, regional economic development interests focused on initiating or attracting high-tech industrial "clusters" by looking to exploit the presence of correlates. (1) Chief among these correlates has been the presence of a research university or institution, or a broader research complex.
In this article, we seek to estimate the effect of federally funded R&D in universities and related research complexes in Texas on the likelihood of high-technology firm entry and survival. By restricting the analysis to Texas, we control for state-specific conditions across counties that influence the variables of interest and gain fully disclosed access to a highly detailed industry data set at the 6-digit level of the North American Industrial Classification System (NAICS).
Previous researchers have considered the question of the effect of university research on the regional economy. We are, however, unaware of any previously published paper that analyzes hazard rates of firms in terms of geographic proximity to knowledge generators.
As Nelson (1986), Jaffe (1989), Acs, Audretsch, and Feldman (1992, 1994), Acs, Anselin, and Verga (2002) and Fischer and Varga (2003) point out, it is quite plausible that the presence of a research university can make locally specific contributions to the level of commercial innovation in its region. (2) The university provides geographically specific access to resources such as libraries, faculty, and a ready pool of graduates at all levels. Research universities and institutions conduct basic research, that is, create knowledge, with the purpose of diffusing the knowledge they create. New knowledge that spills over most readily into the locality should result in localized private sector innovation. Moreover, universities increasingly seek means by which to facilitate faculty start-ups and to enhance access to university resources to support regional entrepreneurs. (3) Although universities can be the sources of direct spin-offs in the form of startups, this impact seems to be moderate and relatively recent. The Association of University Technology Managers reports that 462 new high-technology companies based on academic discovery were formed in 2004 by 191 institutions, up 23.5% from 2003. (4) This only represents an average of somewhat more than two start-ups per institution. In Texas alone in 2004, there were 787 start-ups in the high-technology activities.
If there is a geographic component to diffusion of knowledge, rapid innovation of new knowledge will enhance the economic value of geographic proximity to the knowledge production location. Moreover, given the publicly funded nature of university research, spillovers may be relatively more available from universities than private sector firms conducting similar R&D. Jaffe (1989) and Jaffe, Trajtenberg, and Henderson (1993) find evidence of localized knowledge spillovers from universities. In particular, they find that the presence of a university positively affects the local or regional level of patent activity. Anselin, Varga, and Acs (2000) find evidence that university spillovers are specific to certain industries. For example, their results suggest the strong presence of spillovers in the case of electronics but not in drugs and pharmaceuticals (at the 2-digit SIC level). Mansfield (1995) also finds evidence that the level of university R&D expenditures and quality of relevant faculty are important to industrial innovation in technology industries (at the statewide level). …