Identifying Microgeographies Using Hierarchical Cluster Analysis: Startup Agglomeration and Venture Investment In U.S. Cities

Abstract

This paper advances a new technique for identifying, delineating, and analyzing microgeographies. It applies this technique to locate and measure agglomerations of high-growth, high-tech (HGHT) startup activity within 205 U.S. cities. Using data from 1995 to 2018 on venture-backed companies, I estimate the effect of startup agglomeration economies on ecosystem growth. I find that a one standard deviation increase in measures of startup agglomeration is associated with around a 12% increase in the next period's venture investment, and that optimal HGHT startup density appears to be around one every 2.5 hectares. I also simulate innovation districts for Houston, Texas. I find that an optimally sized and centrally located innovation district could increase Houston's venture investment by as much 15%, while the currently proposed 'Innovation Corridor' could reduce it by as much as 5%.