Do universities generate spatial spillovers? Evidence from US counties between 1930 and 2010

This paper explores the impact of new universities established in the USA between 1931 and 1980 on population density, GDP and market size measured from 1930 to 2010. The analysis is based on differences in differences on counties selected through propensity score matching, as well as an instrumental variable approach. The evidence suggests that counties hosting a university for the first time grew by between 1% and 3% annually on top of the general trends of population density and GDP growth, and that this effect expanded to neighboring counties. Controlling for research intensity and interstate road infrastructure shows that the potential gains from these new universities were severely constrained by the ease of access, which eventually resulted in higher congestion costs. These results point to a situation where new universities create spillover effects that eventually fade away if not accompanied by additional investments.