Identifying Key Sectors in the Regional Economy: A Network Analysis Approach Using Input-Output Data

TL;DR

This paper introduces a network analysis method using random-walk based centrality measures to identify key sectors in regional economies, overcoming limitations of traditional measures and accounting for recursive ties.

Contribution

It develops novel centrality measures suitable for dense economic networks and provides an R package for their application to input-output data.

Findings

New centrality measures better capture sectoral influence.

Method effectively accounts for recursive ties in dense networks.

Provides tools for practical implementation with IMPLAN data.

Abstract

By applying network analysis techniques to large input-output system, we identify key sectors in the local/regional economy. We overcome the limitations of traditional measures of centrality by using random-walk based measures, as an extension of Blochl et al. (2011). These are more appropriate to analyze very dense networks, i.e. those in which most nodes are connected to all other nodes. These measures also allow for the presence of recursive ties (loops), since these are common in economic systems (depending to the level of aggregation, most firms buy from and sell to other firms in the same industrial sector). The centrality measures we present are well suited for capturing sectoral effects missing from the usual output and employment multipliers. We also develop an R package (xtranat) for the processing of data from IMPLAN(R) models and for computing the newly developed measures.