Business Networks in US Manufacturing

• Activity Number: 540
• Type: Invited
• Date/Time: Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Tapabrata Maiti*+ and Arnab Bhattacharjee
• Companies: Michigan State University and Heriot-Watt University
• Keywords: Econometric models ; Regression ; spatial ; social network

Abstract

The network structure of business firms is a subject of considerable importance and interest. In economics of industrial organization, there is the question as to whether supply-side or demand-side linkages drive interactions between firms. In the management sciences, an important concern is to understand whether cohesive networks of socially embedded ties or sparse networks rich in structural holes characterize the network structure of firms. Beyond the above kinds of structural connections, it could be the case that inter-firm networks are driven primarily by spatial strong dependence - the heterogeneous impacts of unobserved latent factors. Using COMPUSTAT data, this work examines the above issues through analyses of business networks of US firms where the network structure (adjacency matrix) is not known a priori. Two methodological approaches are taken. First, the emerging literature in spatial econometrics focusing on unknown and potentially endogenous network structure, and second, the current literature on social network models. The estimated networks provide partial answers to the above questions.