Skip to main content

Learning to Compete: Are Spatial Networks of Firms Random?

, , and



We present a new approach for the empirical investigation of agglomeration patterns. We examine the clustering of manufacturing firms by identifying patterns of spatial network formation that deviate from randomly generated networks. Using firm-level panel data from Vietnam we calculate transitivity, a measure to determine the strength of clustering of manufacturing firms. We then test whether the observed clustering of firms is greater than that of a randomly generated network. 

Our findings suggest that the extent of clustering is over and above that which can be attributed to the legal and regulatory framework, economic zoning, or population patterns.

Read the full paper » (PDF)

Learn more about the Learning to Compete project »


Carol Newman

Associate Professor at the Department of Economics, Trinity College Dublin; Non-resident Senior Research Fellow at UNU-WIDER


Get daily updates from Brookings