New Clustering Method Reveals Evolution of Bilateral Trade Relationships

This article introduces a novel dynamic clustering method that summarizes product-level trade data by classifying international trade dyads based on similarities in their import and export composition profiles. Using two billion observations from annual trade flows, the methodology tracks how typical trade relationships evolve over time—beginning with sparse exchanges moving towards interindustry interactions before transitioning to intra-industry specialization.

Data & Methods: The approach analyzes extensive product-level information using dynamic cluster analysis.

Key Findings: Trade patterns demonstrate a clear progression from basic exchange through diverse industrial engagement toward specialized intra-industry trade relationships.

This granular trade profile measure provides essential insights for international relations research on trade competition and global economic dynamics.