The contribution of statistical network models to the study of clusters and their evolution

Abstract

This paper presents a systemic review of the contributions that stochastic actor-oriented models (SAOMs) and exponential random graph models (ERGMs) have made to the study of industrial clusters and agglomeration processes. Results show that ERGMs and SAOMs are especially popular to study network evolution, proximity dynamics and multiplexity. The paper concludes that although these models have advanced the field by enabling empirical testing of a number of theories, they often operationalize the same theory in completely different ways, making it difficult to draw conclusions that can be generalized beyond the particular case studies on which each paper is based. The paper ends with suggestions of ways to address this problem.