The Agglomeration of American Research and Development Labs

Abstract

We employ a unique data set to examine the spatial clustering of about 1,700 private research and development (R&D) labs in California and across the Northeast corridor of the United States. Using these data, which contain the R&D labs’ complete addresses, we are able to more precisely locate innovative activity than with patent data, which only contain zip codes for inventors’ residential addresses. We avoid the problems of scale and borders associated with using fixed spatial boundaries, such as zip codes, by developing a new point pattern procedure. Our multiscale core-cluster approach identifies the location and size of significant R&D clusters at various scales, such as a half mile, one mile, five miles, and more. Our analysis identifies four major clusters in the Northeast corridor (one each in Boston, New York–Northern New Jersey, Philadelphia–Wilmington, and Washington, D.C.) and three major clusters in California (one each in the Bay Area, Los Angeles, and San Diego).

Keywords: spatial clustering, geographic concentration, R&D labs, innovation

JEL Classification: O31, R12