International Sovereignty Index (ISI) — EU-27 Empirical Results 2024 (v1.0)

This paper reports the first empirical application of the International Sovereignty Index (ISI) to the 27 member states of the European Union using vintage-2024 data under methodology v1.0. The ISI composite, defined as the unweighted arithmetic mean of six normalised axes (financial, energy, technology, defence, critical inputs, logistics), ranges from 0.236 (France, rank 27) to 0.517 (Malta, rank 1) with a population mean of 0.344 and population standard deviation of 0.070. One country is classified as highly concentrated, 23 as moderately concentrated, and three as mildly concentrated. A full variance decomposition with formal proof shows that defence (35.1%) and logistics (34.3%) together account for 69.4% of cross-country composite variance, with the own-variance component contributing 64.3% and cross-covariance 35.7%. A defence-dominance stress test demonstrates that the ranking retains structural coherence (ρ ≥ 0.654) even when both highest-variance axes are excluded. Leave-one-axis-out analysis with both Spearman ρ and Kendall τ confirms that removing the defence axis produces the largest rank disruption (ρ = 0.833, τ = 0.675). Dirichlet weight perturbation (10 000 draws, αj = 10) yields a mean rank-order correlation of ρ = 0.979, with the Netherlands exhibiting the highest rank volatility (σ = 3.07). Rank elasticity testing (±10% perturbation) classifies 13 countries as rank-sensitive and 14 as rank-locked. Only two of 15 pairwise axis correlations are significant at α = 0.05 after Bonferroni correction: energy–technology (t = 3.12) and critical inputs–logistics (t = 3.28); the financial axis is orthogonal to all others (|t| < 1.15). PCA loadings identify three interpretable components: goods-trade concentration (PC1, 31.7%), defence–logistics (PC2, 26.4%), and financial-specific (PC3, 18.1%). A formal structural typology classifies countries by dominance gap into single-axis dominated, dual-axis concentrated, and balanced profiles. Compensability analysis identifies the Netherlands (gap = 0.624, CV = 0.880) as the most internally compensated country. Classification compression diagnostics (η2 = 0.483) confirm that the three-tier scheme is inherently constrained by the unimodal distribution. The paper provides radar profiles for eight selected countries, geographic correlates, defence structural analysis, full axis-level rankings, and a complete reproducibility protocol.