Where are African resources exported to?

The short answer

Of the destinations visible in open data, the European Union takes around 7.8 billion euro of crude materials a year from sub-Saharan Africa. South Africa alone supplies about half of it. After that the list is long and thin: Cote d'Ivoire, Kenya, Guinea, Liberia, Cameroon, Gabon, Ghana, each a few hundred million, each concentrated in one or two commodities.

That is one destination. China, India, the Gulf and intra-African trade are absent from this picture — not because they are small, but because the data that would show them sits behind a key. What follows is what can be shown, followed by what cannot.

How concentrated are these exports? Three answers

This page used to stop here. With only the EU in view, every African exporter appeared perfectly concentrated on one destination, so the standard concentration index returned the same number for all of them — an artefact of the data, not a fact about the world, and it was withheld for that reason.

The IMF now supplies each African reporter against every counterpart, so the question is answerable. It is answered three ways below, because concentration is not one quantity and the three disagree in ways worth reading.

The first finding is about China. For the Congos, Angola and Zambia, one destination takes between 40 and 55 per cent of exports, and for three of the four that destination is China. This is the fact the EU-only view could not see at all, and it is the answer to the question as it was actually asked.

The second finding is that the three measures disagree, and Botswana shows why. On the Herfindahl index Botswana looks more concentrated than Namibia — 1,681 against 1,127. On normalised entropy the ranking flips: Botswana 0.72, Namibia 0.53. Neither is wrong. Botswana sells to 18 destinations and spreads evenly across them; Namibia sells to 152 and leans hard on a few. HHI punishes having few partners at all; entropy asks whether you use the partners you have. Gini agrees with entropy here — 0.67 against 0.93 — because it measures how lopsided the distribution is rather than how big the leader is.

Which one should a policy argument use? It depends what the argument is. If the worry is a single buyer exercising leverage, the HHI and the top share are the right instruments. If the worry is whether alternative buyers exist to absorb a diverted shipment, entropy is closer to the question, because substitution comes from the tail and the HHI barely sees the tail. Published together, they invite the reader to notice that the choice of measure is itself an argument.

And now the comparison that changes the reading

The measures are computed for every reporting economy, not only African ones, and putting them side by side is uncomfortable for the framing the question invites.

Canada sends 78 per cent of its goods exports to one country and scores an HHI of 6,071 — nearly twice DR Congo's 3,298. Mexico is higher still. South Africa, at 486, is as diversified as Germany. On this measure the most export-concentrated economies in the world are not African; two of them are among the richest.

That does not make the concentration equivalent. Canada's dependence on the United States runs through a land border, a free trade agreement and half a century of integrated production; DR Congo's on China runs through a commodity whose price it does not set. Concentration measures the shape of the distribution, not the terms of the relationship, and reading vulnerability straight off an index skips the part that actually determines it.

But the number is the number, and it is worth knowing that the intuition behind the question — that African exporters are unusually dependent — is not what the index says. That is the sort of correction a data platform is for.

What is missing, precisely

• One destination. Eurostat reports what the EU imports. It cannot report what China imports. Answering the question as asked needs a source where African countries are the reporters: UN Comtrade or CEPII BACI, both of which need a key or a registration this platform does not hold.
• Commodity groups, not commodities. SITC 2 and 4 lump cocoa with copper ore. "Resources" in the question probably means minerals and fuels specifically, and separating them needs product-level detail at HS-6.
• Gross value, not ownership. A shipment recorded as leaving Guinea says nothing about who owns the mine or where the profit lands. That question needs ownership data — the GLEIF parent relationships in the register are the honest starting point, and they are not loaded.
• No intra-African trade. Trade between African countries is invisible here entirely, which matters for any argument about regional processing and value retention.

What would answer it properly

One key, essentially. UN Comtrade with African reporters would turn this from a one-destination picture into a real network: every African exporter against every buyer, at product level, and then the concentration measure means something. CEPII BACI would do the same and reconciles the mirror statistics that Comtrade leaves contradictory. Both are recorded in the source register with what blocks them.

Until then the honest headline is narrower than the question: this is where African raw materials go when they go to Europe.

Source register → · On gross against value added →