7%
Potential long-run loss to global output under deep fragmentation, per the IMF, about $7.4 trillion
When trade became a geopolitical weapon, Africa’s choices changed.
In the same season of 2025, two letters travelled between Africa and Washington. One arrived in Maseru: a notice that goods from Lesotho, a mountain kingdom of roughly two million people, would face a fifty percent tariff, the steepest rate applied to any nation on earth. The other left Kinshasa: a handwritten appeal from a president under military pressure, offering the United States access to the cobalt and copper beneath his country in exchange for security. Same trading system, same year, opposite ends of the same weapon.
For most of two generations, economists taught that trade rewarded efficiency. A country made what it made cheaply, sold it abroad, and bought what others made cheaply in return. David Ricardo called the underlying logic comparative advantage, and the institutions built after 1945, the GATT and later the World Trade Organization, encoded it into rules meant to keep politics out of commerce. The promise was that goods would follow prices, not flags.
That order did not end with a treaty. It ended with a tariff schedule. On 2 April 2025, in a ceremony he named Liberation Day, the American president declared a national emergency over the trade deficit and, invoking emergency economic powers, imposed a 10% duty on nearly all imports with far higher rates on dozens of trading partners. The rates were not calibrated to any foreign tariff. External analysts quickly showed they tracked one number: the size of America’s bilateral goods deficit with each country, divided and halved. A formula that had nothing to do with what anyone charged the United States, and everything to do with leverage.
Trade stopped being a market and became an instrument. The question this poses for Africa is unavoidable: what happens to a continent when access to markets, capital and technology is governed by national security rather than comparative advantage?
The story runs deeper than one administration or one set of tariffs, several of which were later suspended, exempted or struck down in the courts. It describes a structural shift that predates 2025 and will outlast it. The pandemic, the war in Ukraine, the semiconductor contest and the scramble for the minerals behind artificial intelligence have each pushed governments to prize resilience over price. The International Monetary Fund has a name for it: geoeconomic fragmentation, the policy-driven reversal of decades of integration. Its economists estimate a deeply fragmented world could shed as much as seven percent of global output over time, close to seven and a half trillion dollars.
Africa did not choose this contest. It is being conscripted into it. And as the two letters of 2025 show, the continent enters holding the weapon at both ends: acutely exposed where it sells finished goods into distant markets, quietly indispensable where it sits on the minerals the great powers now regard as matters of survival. That dual position is the subject of this brief.
The world did not stop trading. Measured crudely, globalisation looks intact: the ratio of goods trade to global output has hovered between 41 and 48% since the financial crisis. But beneath that stable surface, the direction of trade is bending along political lines. American goods imports from China fell 29.7% between 2024 and 2025, a $130 billion drop, on US Census Bureau trade data reported by the Office of the US Trade Representative. China’s share of US goods imports had already fallen from 21.6% in 2018 to 13.4% by 2024, and kept sliding toward 7% through 2025. Flows are being redrawn not by cost but by alliance.
Fragmentation bites harder now than during the Cold War for one reason: scale. As the IMF’s Gita Gopinath has argued, the same cut inflicts far more pain on a world this interwoven, and the pain lands unevenly. The Fund’s modelling finds low-income countries would lose more than four percent of GDP permanently under a severe split into blocs, a pandemic-scale shock that does not reverse.
One line in that modelling matters more to African capitals than the headline loss. In a fragmented system, the countries that suffer least are not the richest but the least substitutable and the least dependent. A supplier the great powers cannot easily replace holds leverage. A supplier whose sales run through a single revocable channel holds none. Comparative advantage measured how efficiently you produced. The new order measures something colder: how hard you are to do without, and how badly you need the other side.
Non-aligned states have so far cushioned the system, acting as connectors that keep goods moving between blocs that no longer trade directly. That role has value, and Africa is well placed to play it. But the connectors are only useful while the walls are low. As the weapon is used more freely, the space to sit comfortably between camps narrows, and every government is pressed to answer a question it once could ignore: whose side is your supply chain on?
The Liberation Day schedule read, for African officials, like a document written about them without their participation. Lesotho, singled out from the podium as a country “nobody has heard of,” drew 50%. Madagascar, one of the world’s poorest nations, drew 47. Across Southern Africa the rates clustered high: Mauritius at 40, Botswana at 37, Angola at 32, South Africa at 30. As the CNN tally noted, several of the hardest-hit economies rank among the 26 poorest on the planet, together responsible for half a percent of global output yet home to nearly 40% of the world’s poor.
The tiles below show the reciprocal rates announced against African economies. The pattern is not one of hostility so much as indifference: rates rose with the bilateral trade surplus a country happened to run with the United States, which is why textile and apparel exporters were punished while raw-commodity shippers, whose goods already entered at low duties, were largely spared.
A tariff is a shock. The withdrawal of a preference is a structural change, and Africa absorbed both at once. On 30 September 2025, the African Growth and Opportunity Act, the 25-year arrangement that granted 32 sub-Saharan economies duty-free access to the American market for thousands of products, expires without renewal. In that moment the United States becomes the only major economy with no formal trade programme in sub-Saharan Africa. Whether Congress restores it, and on what terms, is now an open question. The lapse is real. So is the lesson.
The numbers explain why the lapse frightened trade ministries from Nairobi to Maseru. Kenya grew apparel exports to the United States from about $50 million at the programme’s start to roughly $500 million; one Nairobi factory owner said that without the preference there was “zero chance” of competing with Asia. In Lesotho, AGOA apparel supports 30,000 to 40,000 jobs, most held by women, where roughly a third of exports depend on it. For South Africa, the largest beneficiary, vehicles and parts made up 64% of AGOA-eligible exports. UNCTAD calculated that without renewal, Kenya’s trade-weighted US tariff would nearly triple, from 10 to 28%.
One qualification travels with those numbers, and it resizes the shock. Eligibility was never the same as use. AGOA covered 32 economies, but the benefits pooled in a handful: on the US International Trade Commission’s 2023 review, over three-quarters of non-crude imports under the programme between 2014 and 2021 came from just five countries, South Africa, Kenya, Lesotho, Madagascar and Ethiopia, and the Center for American Progress puts the average utilisation rate for non-oil goods at barely 20%. Many eligible states scarcely touched the preference at all; Cote d’Ivoire, on Carnegie’s reckoning, claimed AGOA on 3.7% of its exports to the United States across two decades, because most of what it ships already entered duty-free under normal terms. So the lapse was not a uniform continental blow. It fell lightly on the many who never used the programme and brutally on the few, above all the apparel exporters of Lesotho, Kenya and Madagascar, whose factories were built around it.
Set beside the countries that policymakers reach for as alternatives, the ranking looks almost designed to redirect production. A shirt sewn in Lesotho met a fifty percent wall. The same shirt sewn in Vietnam met twenty. Mexico, inside a rules-based agreement, met effectively zero on qualifying goods. For a buyer deciding where to place next season’s order, the tariff schedule was a map to the exit, and it did not point toward Africa. When trade becomes a weapon, the damage lands not on the least efficient producer but on the most replaceable and most dependent.
The chart above holds the uncomfortable finding. The rate did not rise with sophistication; it rose with the trade surplus a country ran into the American market, and the surplus that drew fire was the apparel one. Lesotho, Madagascar and Mauritius earned their steep rates for having built garment industries, while Egypt, Kenya and Morocco, more complex by any measure, drew the baseline 10%. So the schedule was not quite a tax on complexity. It was a tax on selling manufactured goods into the United States at scale, and it fell heaviest on the one industrial path open to the poorest, the needle and the thread African policy has spent 40 years trying to encourage.
The second letter, the one from Kinshasa, carries the weapon’s reverse edge. In February 2025, as a rebel offensive seized cities in the eastern Democratic Republic of Congo, President Tshisekedi wrote to Washington offering American access to Congolese minerals in exchange for security. The United States brokered the Washington Accords between the DRC and Rwanda in June 2025, and is pressing for a follow-on partnership that would grant American buyers preferential access, establish a coordinated minerals reserve, and ask Kinshasa to consult Washington before changing its cobalt export policy.
The urgency comes down to one fact of geology. The minerals that make electric vehicles, missiles and artificial-intelligence data centres run overwhelmingly through African soil, and their supply is dangerously concentrated. The DRC holds roughly 55% of the world’s cobalt reserves and mines about three-quarters of what the world digs each year, on the US Geological Survey’s Mineral Commodity Summaries of January 2025, and the gap between those two figures is not pedantry: reserves are potential, production is the tap you can turn. The processing that converts ore into usable metal is more concentrated still, held not in Africa but in China.
This is the paradox at the heart of Africa’s position. The continent is not peripheral to the twenty-first-century economy; it is foundational to it. Yet holding the reserve is not the same as holding the leverage. The DRC ships cobalt hydroxide abroad to be refined, where the profit accrues and the product is finished. Its indispensability is real but hostage to a single processing partner and, until recently, a single route. Washington understands this, which is why its minerals push runs alongside the Lobito Corridor, a rail line steering Congolese copper and cobalt west toward the Atlantic and away from Chinese-controlled channels.
A reserve you cannot refine is a resource. A resource you can refine, and route, and withhold, is leverage. The distance between the two is the whole of industrial policy.
The DRC has been here before, and the record is the sharpest teacher. In 2008 it signed the arrangement known as Sicomines, a minerals-for-infrastructure exchange in which a consortium led by Sinohydro and China Railway would build roads and hospitals and repay itself out of a copper and cobalt mine. Congo’s state miner, Gécamines, took 32% against the Chinese partners’ 68%, and when President Tshisekedi reopened the terms in 2024, calling the original unequal, that split survived intact. What grew was the size of the promise, raised to as much as $7 billion in infrastructure, alongside one revealing condition: the annual payments now depend on the copper price holding above $8,000 a tonne. Below that line, by the reckoning of the Congo Is Not for Sale coalition, Kinshasa receives less, or in a poor year nothing at all. A country that supplies the mineral has indexed its roads to a price set on an exchange in London. That is what a Fulcrum looks like from the inside: leverage that is real, and spent on another party’s terms.
The contest is acquiring a logic that points toward institutions rather than markets. What is taking shape is a preferential bloc, with a strategic stockpile behind it, price floors coordinated across members, and admission offered to resource-rich states willing to align with Washington. This is trade recast as statecraft. A market clears at a price; a club admits members, and the party holding the door sets the terms. Africa holds the entry ticket for cobalt, copper, manganese and the rare earths. The unsettled question is whether it enters as a shareholder or as a supplier, and that question has a structure.
In a market, a relationship is priced. In a weaponised system, it has to be placed, on two axes at once. Every economic tie a country holds carries a reading on each: indispensability, the cost to the other party of doing without you, and dependence, the cost to you of losing them. A nation’s position is where all its ties, taken together, put it, and that, not the size of its economy, decides how a weapon of trade treats it.
The idea has a lineage. Hirschman showed in 1945 that a country’s trade dependence is a form of power that a larger partner can convert into pressure, Keohane and Nye recast asymmetric interdependence as leverage in 1977, and Farrell and Newman formalised the network chokepoint in 2019. What Diagelo adds is a map of movement: plot indispensability against dependence and four positions appear, none of them a permanent address, each a place a country can move through, and that direction is what industrial policy is for. Select a case to read its position.
The framework’s sharpest claim is about targeting. A weapon of trade goes where pressure is cheapest, which is against the exposed. It hits the Tributary, replaceable and dependent at once, because there pressure converts into compliance at no cost to the aggressor. Lesotho’s apparel, reliant on a single channel it did not control, was a Tributary, and the 50% rate found it.
If coercion targets the lower-right, escape means leaving it, and there are two exits. Move up, raising indispensability by capturing more of the value chain at home, the path Indonesia took with nickel. Move right to left, cutting dependence by spreading sales, suppliers and routes across many partners, the path a continental market opens.
A framework that produces no testable output is a diagram with ambitions, so both axes are scored, and one is far harder than the other. Dependence is the tractable one. Score it as a Herfindahl-Hirschman index of a country’s export destinations, weighted by how revocable access to the largest market is: a binding treaty such as USMCA scores low, most-favoured-nation access in the middle, a unilateral preference such as AGOA or the EU’s Everything But Arms high, since only the grantor can pull one. Two countries can ship the same share to one buyer yet sit far apart, one under a contract, the other under a favour, and that gap is most of what the map is built to see.
Indispensability resists measurement, and the tempting shortcut is the error the brief is written against. Share of world supply counts how much you sell, not how hard you are to replace. High cobalt share now coexists with real substitution toward lithium-iron-phosphate chemistry, and high platinum share with thrifting and recycling, so treating reserves as power, the reflex of every commodity-nationalism piece, mistakes a large number for a strong position. It is built here from four inputs, two observable and two of judgement, set out in the rubric below so a reader can dispute a single cell rather than the whole map.
| Axis | Input | Where it comes from | Type |
|---|---|---|---|
| Dependence | Herfindahl-Hirschman index of export destinations | UN Comtrade | observable |
| Dependence | Revocability weight on the top market (treaty / MFN / preference) | Trade-agreement status | observable |
| Indispensability | Elasticity of substitution in the primary end use | Sector literature | judgement |
| Indispensability | Buyer inventory cover, in months | Industry & IEA data | observable |
| Indispensability | Time and capital to qualify an alternative source | Engineering estimates | judgement |
| Indispensability | Number of viable alternative suppliers at scale | USGS & trade data | observable |
That reluctance is the distinction a reserves-equal-power reading refuses to make: any honest map carries more error on the vertical axis, which is why the DRC can hold the largest cobalt endowment on earth and still bargain from a Fulcrum, not an Anchor.
No African economy has to invent its response from scratch. Four emerging economies have already played this system, each showing one exit from the Tributary, and its price.
Indonesia sat on the world’s largest nickel reserves, shipped them raw for decades, then in 2020 banned raw exports, forcing refining onshore. Export value roughly quintupled, from about $6 billion in 2013 to nearly $30 billion by 2022. The same move carries the catch: the refineries are Chinese-built and Chinese-run, the buyer of last resort is still China, and the WTO ruled the ban illegal. Kinshasa, studying this playbook for its own cobalt, should read both halves.
Mexico’s exports are replaceable, yet it did what few captured economies manage: it wrote its dependence into rules. Under USMCA, goods that meet regional-content thresholds enter the United States effectively duty-free, and when blanket tariffs loomed in 2025, compliance jumped from 45 to 89% in a year. Proximity plus a treaty became a shield, but one lasting only as long as the treaty, with a 2026 review now over every plant.
Vietnam built the archetypal China-plus-one economy, taking work leaving China and reaching $165 billion in electronics exports. It cut everyone’s dependence by offering a second address. Then the weapon found the ceiling: the 2025 framework set a 20% tariff and a 40% duty on transshipped goods, aimed at Chinese parts passing through. A corridor that only relabels is itself replaceable.
India held to autonomy, refusing a bloc while courting factories fleeing China, and now assembles roughly a quarter of the world’s iPhones. Autonomy is not free. When it kept buying discounted Russian oil, Washington stacked a penalty that pushed its rate to 50%, before rolling it back once Indian buyers adjusted. Scale bought room to absorb and negotiate. For smaller states the lesson is unforgiving: non-alignment is leverage only if you are costly to punish.
Read together, the four say one thing: there is no clean escape from a weaponised system, only better and worse positions within it. Indonesia bought power at the cost of a new dependence, Mexico security at the cost of a review clause; Vietnam bought relevance until the rules caught its shortcut, India room by being too big to bully cheaply. For most African economies, too small to deter and too specialised to be indispensable, none of these hands is playable alone.
Africa’s two exits map onto two instruments it already holds: deepen the internal market to cut dependence, keep more of the mineral chain at home to raise indispensability. Both are underway, both too slowly.
The most striking fact about African trade is how little of it is African: 15 to 18% of the total, against nearly 60% in Asia and 70% in Europe. Every point of that gap is dependence on markets that can revoke access at will, as 2025 proved. The African Continental Free Trade Area is built to close it, a single market of 1.4 billion people, 48 ratifying states, schedules agreed on more than 90% of goods. Trade under its rules has begun, and the World Bank’s 2020 assessment projects the goods agreement could lift 30 million from extreme poverty and 68 million from moderate poverty by 2035; a 2022 study raises that to 50 million, but only with free movement and deeper integration.
A larger internal market does more than add sales; it changes a country’s position. An apparel maker selling only to the United States is a Tributary; the same maker selling across a working continental market becomes a Skiff, replaceable still but no longer captured. The obstacles are known: infrastructure gaps, a trade-finance shortfall near $100 billion, unresolved rules of origin on autos and textiles. But these are problems of implementation, not design, and implementation is a choice.
Mineral wealth alone will not make Africa indispensable, because indispensability lives in processing, not extraction. The continent holds over half the world’s cobalt reserves and captures almost none of the battery’s margin. Closing that gap is the second exit, reached for unevenly: the DRC has tested export quotas, others have studied Indonesia’s downstreaming. Indonesia supplies the caution: forcing refining onshore captures value but can relocate the dependence.
The great powers are building competing clubs: an American minerals bloc, Chinese processing dominance, a European push for its own raw-materials security. A continent that picks one loses the leverage of being wanted by all. That leverage is real, since Africa’s minerals are courted from every direction, and it demands turning courtship into competing bids: to run the DRC’s cobalt, Zambia’s copper and South Africa’s platinum as auctions among rival buyers rather than tribute to one patron. But an auction needs a second bidder willing to raise the price, and that is the easy assumption to get wrong. China, the obvious counterparty, already owns the refining step and has walked before: when cobalt hit a nine-year low in 2025, Chinese operators idled Congolese capacity rather than hold it. A buyer that controls the exit feels no pressure to bid up the entrance. The floor under an auction is not automatic; a seller manufactures it by holding an alternative the incumbent can lose to, a route, a refinery, a rival contract. Absent that, the call for bids meets silence, and silence names its own price.
The two letters close the argument where they opened it. Maseru’s and Kinshasa’s arrived in the same system in the same year, and everything separating their fortunes is captured on the map. Lesotho sold a replaceable good through a channel it did not own, and the weapon found it. The DRC held something the world could not do without, and the weapon negotiated with it. Neither position was destiny. Lesotho’s apparel could anchor a regional garment market in place of a single foreign one. The DRC’s cobalt could be refined and owned at home, not merely dug and shipped. The difference between the two futures is not luck or geology. It is the deliberate work of moving on the map, up toward indispensability, across toward diversity, and out of the one corner where a weapon of trade does its cleanest work.
Every institution with a model has projected the continent’s trajectory; almost none of those trajectories were chosen by Africans deciding, deliberately, which end of the weapon to hold. The tariff war of 2025 did not create Africa’s exposure. It made the exposure legible, and with it the choice. A weapon is only as dangerous as the grip that holds it. For the first time, which end of it Africa holds is a decision rather than an inheritance, and the map is where that choice gets settled.
Tariffs make news; the reordering underneath makes the decade. Global trade is re-pricing from efficiency toward security, and the change outlasts any one dispute.
The lesson outlasts the tariff that taught it: where a country can sell now weighs as heavily as what it can make, and the widest list of destinations has become the safest place to stand.
Sources & method
Figures and claims draw on the institutions cited in context throughout this piece; datasets are as published at the dateline. Where estimates differ across sources, the piece states the range rather than a single point.