University Knowledge Spillovers, Geographic Proximity and Innovation : An Analysis of Patent Filings Across U.S. Counties

The literature suggests that knowledge production intensity depends upon the geographic proximity of knowledge and information sources. On the other hand, there have been arguments that the rapid development of communication technologies would reduce the importance of proximity to the production of knowledge. That is, distance would not matter. In this paper, we propose a measure for university knowledge spillovers as a gauge for the importance of proximity to knowledge creation. Using a cross-section of 3110 United States counties in recent years - 2009 to 2011 - we empirically test the relationship between knowledge spillovers and innovative activities as measured by patent filings. We use university research and development (R&D) expenditures in STEM fields as the foundation for our metric of knowledge spillovers and a decay function to reflect the diminishing influence of those expenditures as the distance from the university increases. Using patent counts to measure the local technological innovation as our dependent variable, we test knowledge spillovers as spherical scores of university R&D expenditures as our explanatory variables of interest using three distance thresholds - 50, 100 and 250 miles. We also control for the private R&D intensity and regional social and economic characteristics, including a county's employment density, educational attainment, proprietorship rates and venture capital inflows. We find that the influence of university knowledge spillovers can reach as far as 250 miles, but evidence that spillovers decline with distance is not unambiguous. Research intensive counties exhibit different spillover effects from counties without research anchor institutions. We find evidence that counties with higher concentrations of R&D expenditures and concentrations of science and engineering programs will exhibit higher patent rates