Figure 6.1 Developing countries depend heavily on a few major economies as export
destinations
South Asia
16.0
Middle East, North Africa,
Afghanistan, and Pakistan
6.8
5.7
46.2
9.5
Latin America and the
Caribbean
2.1
39.9
12.7
45.3
Sub-Saharan Africa
5.8
Europe and Central Asia
3.9
East Asia and Pacific
8.4
18.7
10
3.6
7.5
10.0
19.2
20
30
40
50
Share in exports of developing countries (%)
United States
11.5
21.2
41.4
17.9
0
9.0
India
European Union
60
70
China
Source: WDR 2026 team, based on data of the CEPII-BACI (Centre d’études prospectives et d’informations internationales/
Center for Prospective Studies and International Information–Base pour l’Analyse du Commerce International/Database
for International Trade Analysis) Dataset, CEPII, https://www.cepii.fr/DATA_DOWNLOAD/baci/doc/baci_webpage.html.
Note: The bars use 2024 data. Developing countries are defined as low-income, lower-middle-income, and upper-middleincome countries based on the World Bank fiscal year 2027 income classification. (Refer to Metreau et al. 2026; World
Bank Country and Lending Groups [dashboard], World Bank, https://datahelpdesk.worldbank.org/knowledgebase
/articles/906519-world-bank-country-and-lending-groups.) China and India are excluded from the export share data
even though they belong to the regions of East Asia and Pacific and South Asia, respectively. This is done to avoid mixing
their role as both a source and a destination country. The United States and all European Union member countries are
classified as high-income countries and are therefore also excluded from the export share data.
Trade dependencies, as well as dependencies
in other domains such as security and financial
systems, can reinforce one another. Crucially,
the power arising from such imbalances need
not be exercised or even articulated to be effective. Where one country is materially dependent on another across many domains, the
mere existence of such asymmetric dependencies can shape the technology choices countries
make.10 The choices developing countries faced
in the late 1990s regarding technologies related
to genetic modification provide an instructive
parallel. States with closer ties to the European
Union were more likely to ratify the Cartagena
Protocol on Biosafety early, whereas those closer
204
to the United States were slower to support it or
did not support it at all.11
In the case of AI, similar dynamics surrounding
dependencies can be observed in the signing
of the Pax Silica Declaration. Launched by the
United States in 2025, the Pax Silica Initiative
spans the entire AI stack and aims to reduce coercive dependencies, secure global technology supply chains, address AI supply chain opportunities
and vulnerabilities, and explore joint investment,
as well as protect sensitive technologies and
build trusted digital infrastructure.12 At the time
of writing, signatories to the declaration include
Australia, Finland, Greece, India, Israel, Japan,
World Development Report 2026