Methodology

How we measure trade dependencies

This atlas applies the five-criteria, EU-scale methodology of Mejean and Rousseaux (2024) at the level of individual economies, following Fontagné and Rousseaux (2026), and introduces reverse dependency indicators measuring the leverage an economy exercises over others.

1. Why measure trade dependencies?

Global value chains (GVCs) have undergone significant hyper-globalisation over the past three decades, leading to highly dispersed geographical production. This fragmentation has increased trade gains but also heightened exposure to supply shocks: disruptions at any concentrated GVC node can cascade downstream through industries and consumers (di Giovanni et al., 2020; Bonadio et al., 2021).

The Covid-19 pandemic and the Russian invasion of Ukraine demonstrated how heavily economies rely on a few concentrated suppliers for critical goods — pharmaceuticals, semiconductors, rare earths, chemicals. Identifying these vulnerabilities is a prerequisite for designing effective resilience policies. The current context of US tariffs on Europe adds further urgency: it is essential to identify not only our dependencies on third countries but also the leverage we hold over them.

2. Identifying import dependencies

A product is classified as a critical import dependency of country j if it simultaneously satisfies all relevant criteria. We implement three nested lists.

2.1 European Commission criteria (List 1: EC)

The European Commission's bottom-up approach identifies a product k as a potential vulnerability for economy j if it meets three simultaneous criteria, computed at the HS6 product level over the reference period:

Notation used throughout this section: j = the importing (reference) economy · k = a product, defined at the HS6 level · M_jk = total imports of k by j · M_extra,jk = imports of k by j sourced from outside the reference zone (non‑EU27 for EU member states; rest of world otherwise) · X_jk = total exports of k by j · HHI(·) = Herfindahl–Hirschman Index of supplier concentration, computed on extra‑zone source shares.

EC Criterion 1

Import concentration

HHI(M_extra,jk) > 0.4

The Herfindahl–Hirschman Index of j's extra‑zone import sources for product k — the sum of squared market shares of each supplying economy — exceeds 0.4. A value above this threshold signals that a small number of suppliers dominate the market, so a single disruption can affect most of j's supply.

EC Criterion 2

Extra-zone import significance

M_extra,jk / M_jk > 0.5

Extra‑zone imports account for more than 50% of j's total imports of k. This rules out products where j mostly sources from within its own trade zone (e.g. other EU27 members), which is considered a lower-risk dependency.

EC Criterion 3

Import-to-export substitutability

M_extra,jk / X_jk > 1

Extra‑zone imports of k exceed j's own exports of k. If j exports more of a good than it imports from outside the zone, it could in principle redirect its own export production to cover domestic needs — so this criterion excludes goods where that substitution margin exists.

A product is classified on List 1 (EC) only if all three criteria hold simultaneously for a given j and k.

2.2 Absorption criterion (List 2: EC + Absorption)

Mejean and Rousseaux (2024) add a fourth criterion measuring domestic absorption: the degree to which extra‑zone imports fill domestic demand net of intra‑zone flows.

MR Criterion 4

Domestic absorption constraint

M_extra,jk / Absorption_jk > 0.5

where Absorption_jk = Production_jk + M_jk − X_jk is j's apparent domestic consumption of k — what it produces plus what it imports, minus what it re-exports. Extra‑zone imports covering more than 50% of this effective demand implies limited scope for substituting them with domestic production. Production_jk requires product-level manufacturing data (Prodcom, EU27 + UK, or an equivalent source), which is why this criterion is only computable for economies with such data.

List 2 (EC + Absorption) is the subset of List 1 that additionally satisfies Criterion 4.

2.3 Stickiness criterion (List 3: EC + Absorption + Stickiness)

A fifth criterion captures supplier stickiness — the difficulty of switching suppliers post-shock — using the relationship-stickiness metric of Martin, Mejean, and Parenti (2024).

MR Criterion 5

Supplier stickiness

rs_k > Q75(rs)

where rs_k is product k's relationship-stickiness score (Martin et al., 2024) and Q75(rs) is the 75th percentile of that score across the full HS6 distribution of 5,567 products. Products above this threshold exhibit unusually low ex-post substitutability between suppliers — switching costs, relationship-specific investments, or search frictions make it materially harder to replace a supplier even after a shock.

List 3 (EC + Absorption + Stickiness) is the subset of List 2 that additionally satisfies Criterion 5. For economies without Prodcom-equivalent data, Criterion 5 is instead applied directly on top of List 1 (EC), skipping the absorption filter.

~5,381

Total HS6 products

→

~380

EC criteria (3)

→

~230

+ Absorption (4)

→

~50

+ Stickiness (5)

Indicative counts for EU27 pool 2015–2019 from Mejean and Rousseaux (2024).

3. Reverse dependency indicators

For each source economy s, we compute the mirror image: the set of products for which s is a critical supplier to other economies j. Using X_s→j,k for economy s's exports of product k to economy j, and share_s,jk for s's share of j's extra‑zone imports of k, a product k is a reverse dependency of s on j if it simultaneously satisfies:

Reverse Criterion 1

Concentrated market + source presence

HHI(M_extra,jk) > 0.4 and share_s,jk > 0

The import market for k in j is concentrated (as in EC Criterion 1), and s is an active supplier within that concentrated market.

Reverse Criterion 2

Source dominance in j's imports

X_s→j,k / M_jk > 0.5

Economy s alone supplies more than half of j's total imports of product k — the mirror image of EC Criterion 2.

Reverse Criterion 3

Export power

X_s→j,k / X_jk > 1

Economy s's exports of k to j exceed j's own total exports of k, ruling out easy redirection of j's domestic production to replace s — the mirror image of EC Criterion 3.

4. Aggregate indicators

4.1 Vulnerability Index (VI)

For economy j, the Vulnerability Index aggregates critical import values weighted by each product's position in global value chains (upstreamness, Antràs and Chor, 2022), normalised by GDP:

VI_j = Σ_k ( M_extra,jk × upstreamness_k ) / GDP_j
// sum taken over k in j's critical import list; extra-zone import value in EUR 2019 constant
// GDP in EUR billion (World Bank + Eurostat for EU27)

Upstreamness ≈ 1 indicates a final good; upstreamness ≫ 1 indicates an upstream intermediate input whose disruption propagates through many downstream sectors.

4.2 Leverage Index (LI)

For source economy s, the Leverage Index aggregates the critical export values it supplies to each importing economy j, weighted by upstreamness and normalised by j's GDP:

LI_s→j = Σ_k ( X_s→j,k × upstreamness_k ) / GDP_j
// measures the share of j's GDP over which s holds critical export leverage

The total leverage of s over the rest of the world is LI_s = Σ_j LI_s→j / GDP_s, i.e. leverage held summed across all importing economies j, normalised by s's own GDP.

4.3 Dependency Balance Score (DBS)

DBS_j = VI_j / LI_j
// DBS < 1 → economy j exports more leverage than it imports risk (net powerful)
// DBS > 1 → economy j is more vulnerable than it is powerful (net vulnerable)
// DBS ≈ 1 → balanced

The DBS is the main indicator on the atlas homepage map. Economies coloured blue (DBS < 1) hold net leverage over others; economies coloured red (DBS > 1) are net importers of critical risk.

5. Data sources

Dataset	Coverage	Use
BACI CEPII HS17 V202601	2015–2024, HS6 ~5,000 products, ~200 countries	All trade flows (import/export values)
Eurostat PRODCOM	EU27 countries, 2015–2023	Domestic production for absorption criterion
UK Prodcom (ONS)	United Kingdom, 2015–2023	UK domestic production
Martin, Mejean, Parenti (2024)	5,567 HS6 products	Relationship stickiness (rs) for criterion 5
Antràs & Chor (2022)	~5,000 HS6 products	Upstreamness scores for VI/LI weighting
World Bank WDI (NY.GDP.MKTP.CD)	~200 countries, annual	GDP for normalisation (non-EU)
Eurostat nama_10_gdp	EU27 aggregate, annual	GDP for normalisation (EU27)
ECB (EXR.A.USD.EUR.SP00.A)	Annual, 2015–2024	USD → EUR conversion
Eurostat STS (sts_inpp_m)	EU27, monthly	PPI deflator (base 2019)

6. Scope and limitations

The absorption criterion (criterion 4) requires domestic production data and is therefore only available for EU27 member states and the United Kingdom (Prodcom). For all other economies (US, China, Japan, etc.), only the three EC criteria are applied, and stickiness is computed directly on EC products without the absorption filter.

All trade values are in EUR 2019 constant prices, deflated using the EU27 PPI from Eurostat.
The "extra-zone" definition varies by country: for EU member states it means non-EU27; for all other economies it means rest of world.
The pooled 2015–2019 period is used as the baseline (pre-shock), reflecting persistent structural dependencies rather than annual fluctuations.
Production data for countries outside EU/UK would significantly improve coverage. If you have production data for additional countries, please get in touch.