Geographical spillovers and university research: A spatial econometric perspective

Using data for U.S. metropolitan statistical areas, an earlier study of agg regate local geographic research spillovers generated by universities (Anse lin et al.1997) was extended to a sectorally disaggregated level. These fin dings suggest the existence of significant sectoral variation with respect to local university effects on innovation. Apparent differences were found across sectors with respect to the "mix" of applied local knowledge inputs in general, and the extent to which university research plays a role in inn ovation in particular. The main conclusion is that local university spillov ers seem to be specific to certain industries, such that at the two-digit S IC level, no university spillover effects are at work in the Drugs and Chem icals (SIC28) and in the Machinery (SIC35) sectors. On the contrary, very s trong and significant university research spillovers are evidenced in the E lectronics (SIC36) and the Instruments (SIC38) industries. These spillovers extend beyond the boundary of the MSA within a 75-mile range from the cent ral city.