Concentration, Scope, and Complexity in Greece–Japan Merchandise Trade: A Section- and Product-Level Diagnostic

Economic complexity theory predicts a positive correlation between export sophistication and export value, but this is rarely examined within a specific bilateral trading relationship, where trade fragmentation and value-chain restructuring play out. This research examines the structure of Greek goods exports to Japan in 2023 by relating export scale to complexity and scope in 356 HS4 items with a Product Complexity Index (PCI) available (trade value USD 732.4 million) and in the 20 Harmonized System sections that contain them. The regression of export scale on complexity and scope uses heteroskedasticity-robust OLS, with section fixed effects at the product level, complemented by wild and wild-cluster bootstrap inference, a Freedman–Lane permutation test, equivalence testing, influence diagnostics, a specification curve and power analysis. Export scale shows extreme concentration (Herfindahl index 0.785; Gini coefficient 0.911; Foodstuffs make up 88.5% of value), and the value-weighted mean PCI (−0.455) is substantially lower than the line-weighted mean (−0.101) and the unweighted mean across sections (−0.300): the value of trade lies in the less sophisticated part of the basket. This is the paper’s central structural result, and by construction it implies a negative covariance between export value and sophistication. At the section level the association between complexity and export value cannot be estimated with precision (β = −0.618, HC1 SE 0.562, p = 0.271; the smallest coefficient detectable with 80% power corresponds to a roughly seven-fold difference in exports). At the product level, which preserves the 39.9% of complexity variance lost in averaging, the coefficient is −0.736 (cluster-robust SE 0.256; wild cluster bootstrap p = 0.046); it is robust to alternative estimators and to excluding the dominant sector, although it weakens when very small product lines are excluded and is not found when lines are weighted by export value. The negative association between sophistication and export value is therefore present within sections as well as between them, and the within-section association persists when the dominant section is excluded. The research gives a reproducible two-level diagnostic of bilateral export structure, shows that a section-level cross-section of this size is not precise enough to estimate the complexity–export association, and offers policymakers a measure of ex ante structural exposure that is distinct from resilience, which can only be assessed with a longitudinal design.