Innovative Behavior and Spatial Location - Using Patent Counts and Geographic Location to Estimate Innovative Spillins

In this paper we examine the relation between geographic location and innovative behavior. Knowledge spillins, as opposed to knowledge spillovers, are modeled as an externality which exists between geographically close economic agents and enters the representative inventor production function explicitly from neighboring regions. To proxy new innovative behavior and new knowledge generated we use counts of patent filings per county. The proposed geographic spillin is tested for the US Midwestern States of Iowa, Minnesota, Missouri, Kansas, Nebraska, South Dakota and North Dakota using a newly constructed data set and implementing spatial statistical methods. The data set is comprised of primary inventor utility patent filings per county for the years 1975-2000. The results do indeed suggest spatial interaction does occur and innovative activity in surrounding counties is an important factor in explaining new innovative behavior. Further analysis also reveals lagged patenting behavior within the county also has a significant impact on patenting activity suggesting innovative externalities exist over both space and time.