The World-Trade Web: Topological Properties, Dynamics, and Evolution

TL;DR

This paper analyzes the topological and dynamic properties of the global trade network over 20 years, revealing stability in network structure and a shift in link weight distribution from log-normal to power law.

Contribution

It provides a comprehensive analysis of the world trade web's topological properties, their stability over time, and the evolving distribution of trade link weights using a weighted-network approach.

Findings

Network statistics distributions have remained stable over 20 years.

The distribution of link weights is shifting from log-normal to power law.

Growth rates of network statistics follow fat-tailed distributions like Laplace.

Abstract

This paper studies the statistical properties of the web of import-export relationships among world countries using a weighted-network approach. We analyze how the distributions of the most important network statistics measuring connectivity, assortativity, clustering and centrality have co-evolved over time. We show that all node-statistic distributions and their correlation structure have remained surprisingly stable in the last 20 years -- and are likely to do so in the future. Conversely, the distribution of (positive) link weights is slowly moving from a log-normal density towards a power law. We also characterize the autoregressive properties of network-statistics dynamics. We find that network-statistics growth rates are well-proxied by fat-tailed densities like the Laplace or the asymmetric exponential-power. Finally, we find that all our results are reasonably robust to a few alternative, economically-meaningful, weighting schemes.