Title: Identifying important nodes in input-output networks
Authors: Alexander Semenov - University of Florida and Saint Petersburg State University (United States) [presenting]
Abstract: The economic system may be represented as a complex network, consisting of different sectors linked by input-output relations. In such a network, shocks occurring in one of the sectors may result in wide disruptions of other sectors' production. Previous research in network science has established multiple ways to measure the importance of a node or an edge in a network, such as degree centrality, closeness centrality, PageRank, or betweenness centrality. Shocks occurring in the most important sectors may lead to more pronounced impacts. Several attempts have been made to analyze the topology of the networks constructed from input-output tables; however, there has been little analysis of the effect of disruptions in the important sectors of the other parts of the network. We construct networks from published data on input-output relations, and study how shocks occurring in the most important sectors of the economy, quantified according to network science-based measures, affect production of other sectors. Next, we propose an optimization problem aimed at finding the most influential sectors and explore how effects of important nodes differ over networks constructed from different economies.