Washington closes its G20 trade presidency with a 28-economy Milwaukee Framework on steel overcapacity, a blocked consensus statement and an open question about whether plurilateral coalitions can substitute for multilateral agreement
MILWAUKEE, Oct. 6, 2026
The United States Trade Representative closed the books yesterday on a G20 trade ministerial that produced one durable institutional artifact, one notable consensus statement, and a public demonstration that the world’s largest economies can no longer agree on the central problem in global trade.
In a wrap-up release issued Oct. 5, USTR summarized the Milwaukee ministerial as a showcase of American manufacturing, detailing factory tours at Rockwell Automation and Harley-Davidson, a chief executive roundtable at the Milwaukee Clock Tower attended by European Union and Japanese officials, and a welcome dinner at the Milwaukee Art Museum built around American agricultural products. The choreography was deliberate. So was the substance underneath it, and the substance was more contested than the itinerary suggested.
The headline deliverable is the Milwaukee Framework, a plurilateral arrangement on global steel excess capacity adopted by 28 economies of the Global Forum on Steel Excess Capacity, including the European Union, Australia, Japan and South Korea. Ambassador Jamieson Greer announced it with the declaration that the United States was launching the framework to take coordinated action in the face of the steel production overcapacity situation.
China, the largest single source of the capacity in question, is not a member of the Global Forum and is not a party to the framework.
What the framework commits signatories to do
The Milwaukee Framework is a commitment instrument rather than a treaty, and its obligations are correspondingly general. Signatories agreed to phase out government subsidies that distort markets, to enhance information sharing in order to identify abnormal trade patterns, and to consider trade barriers against producers responsible for excess capacity.
Each of those three pillars does real work if implemented seriously.
Subsidy phase-out is the hardest and the most consequential. Steel overcapacity is sustained by state support that allows producers to operate below cost indefinitely, which is the mechanism that converts a cyclical price slump into a structural one. A credible multilateral subsidy discipline would address the cause rather than the symptom. The difficulty is that the framework’s signatories are, with a few exceptions, not the principal subsidizers, which means the commitment is largely self-binding among parties with the least to give up.
Information sharing is more immediately operational. Identifying abnormal trade patterns is the technical predicate for transshipment and circumvention enforcement, and in a market where capacity sits outside the agreement, circumvention is the primary channel through which it reaches signatory markets. A functioning data exchange among 28 economies would materially shorten the lag between a surge and a response.
The third pillar, coordinated trade barriers, is where the framework intersects with existing U.S. policy most directly. The United States already maintains expanded Section 232 duties on steel, aluminum and copper, strengthened in April and modified again in June. What the framework contemplates is that signatories act in parallel rather than sequentially, which matters because uncoordinated action simply redirects displaced volume from one open market to the next.
The scale of the problem
The figures explain the urgency. Global steel excess capacity stood at approximately 601 million metric tons on a 2024 baseline and is projected to reach 745 million metric tons by 2028, according to analysis of the Milwaukee discussions. Between 2026 and 2028, capacity is expected to expand by 139 million tons while demand grows by only 34 million tons, a ratio of roughly four units of new capacity for every unit of new demand.
That is not a cyclical imbalance. It is a structural one, and it is widening rather than closing.
The economics follow directly. State-backed producers can sustain losses for strategic reasons for periods that no commercially financed producer can match, which depresses global prices beyond anything commercial logic would justify and progressively eliminates privately financed capacity in open markets. The result is a one-way ratchet: capacity that exits in market economies does not return when prices recover, because the conditions that drove it out have not changed.
It is also worth noting what the data does not cover. There is no comparable quantified baseline for excess capacity in aluminum, solar photovoltaics, electric vehicles or critical minerals processing, all sectors where similar dynamics are widely believed to be operating. The Milwaukee Framework addresses the one sector where the measurement infrastructure exists, which is a reasonable place to start and an incomplete answer to the problem ministers were actually discussing.
Where the ministerial failed
The framework was the achievable outcome. The G20 consensus was not.
According to USTR’s own account, a handful of members rejected creating a pathway toward cooperative action on structural excess capacity, blocking consensus language in the Chair’s Statement. Greer described the outcome as severely disappointing to the U.S. presidency’s efforts, and separately expressed disappointment that G20 trade ministers did not reach consensus on forced labor.
The forced labor numbers are stark. The declaration on eliminating forced labor from global supply chains attracted three signatures out of twenty G20 members: the United States, Mexico and Argentina. USTR’s framing commended the ministers who signed. The arithmetic says the rest did not.
Only one substantive item achieved full consensus. G20 trade ministers agreed to condemn the weaponization of food through coercive trade actions, with the statement declaring that trade in food, agricultural products and agricultural inputs has been weaponized to exert coercive pressure for economic or political reasons. For U.S. agricultural exporters, who have absorbed the brunt of retaliatory measures across successive trade disputes, that language carries real symbolic weight even without binding effect.
Ministers also discussed global diesel supplies and expressed interest in reconsidering the most favored nation principle, the foundational non-discrimination rule of the World Trade Organization system. That a G20 trade ministerial can treat MFN as an open question is itself a significant marker of where the multilateral system now sits.
The attendance list told its own story. Poland participated as a special guest despite having no permanent G20 membership, with Polish Finance Minister Andrzej Domanski raising concerns about restrictions affecting Ukrainian agricultural exports. South Africa was not invited. Several delegations, including India and Brazil, used the margins for bilateral negotiations rather than the plenary.
Reaction and interpretation
Interpretations of the Milwaukee outcome diverge sharply along a single axis: whether plurilateral coalitions are a workaround for multilateral paralysis or a symptom of it.
The administration’s position is the former. In this reading, a 28-economy framework with operational commitments is worth considerably more than a consensus communique that commits no one to anything, and the inability to secure G20 agreement simply confirms that the G20 is no longer the right venue for enforcement-oriented trade policy. The coalition of the willing is the point.
Critics make the opposite argument. A framework that excludes the largest source of the problem, and that secures commitments primarily from economies that are themselves victims of overcapacity rather than causes of it, risks becoming an elaborate exercise in mutual reassurance. The analytical case against is direct: the Milwaukee Framework is the only operational multilateral instrument on steel overcapacity, and China is not in it.
European and Japanese officials, who participated in both the framework and the CEO roundtable, occupy a middle position. Both jurisdictions have their own steel safeguard and capacity concerns and have been building their own defensive instruments. For them, the framework’s value lies less in constraining non-members than in coordinating members, reducing the risk that a measure taken in one market simply floods another.
Industry reaction in the United States has been broadly supportive on the steel side and more cautious among steel-consuming manufacturers, who have absorbed the cost of Section 232 duties for years and who read coordinated barriers as a signal that relief is not coming.
Economic impact analysis
For the steel market, the framework’s near-term effect is informational rather than price-setting. Twenty-eight economies sharing data on abnormal trade patterns will tighten circumvention enforcement, which raises the cost and reduces the reliability of transshipment routes. That should modestly improve price realization in signatory markets over several quarters. It will not address the underlying capacity growth, because the capacity is being built outside the agreement.
For the broader trade system, the more important signal is the MFN discussion. If major economies are prepared to reconsider non-discrimination as a default rule, the direction of travel is toward a system of overlapping preferential and punitive arrangements in which a product’s duty treatment depends increasingly on its origin’s political relationships rather than on a common schedule. For multinational firms, that converts tariff planning from a classification exercise into a geopolitical one.
For the United States specifically, the Milwaukee outcome fits a pattern visible across the year. After the Supreme Court invalidated the IEEPA tariff architecture in February, the administration rebuilt its program on Section 232 and Section 301, both of which are unilateral instruments. The Milwaukee Framework is an attempt to add a plurilateral layer on top of that unilateral base. Whether the layer holds will depend on whether signatories follow through when following through is costly.
The Global Forum problem
Any assessment of the Milwaukee Framework has to begin with the institution it sits inside. The Global Forum on Steel Excess Capacity was created in the aftermath of the last major steel glut, and its history is instructive. It produced useful data, built a shared analytical vocabulary, and then lost its most important participant. China’s departure from the forum left an arrangement in which the economies most committed to addressing overcapacity are also the economies least responsible for creating it.
The Milwaukee Framework inherits that structure. Twenty-eight economies committing to phase out distorting subsidies is a meaningful statement about their own conduct and a weak instrument against anyone else’s. Its practical force therefore depends almost entirely on the third pillar, coordinated trade barriers, which is the only mechanism that reaches non-members.
That creates a design tension that officials involved have not resolved publicly. A framework justified as cooperative subsidy discipline operates in practice as a coordinated defensive bloc. Those are different things with different legal implications under existing World Trade Organization commitments, and the difference will matter if and when the framework’s measures are challenged.
There is a counterargument worth stating fairly. Coordinated defensive action among major open markets has a deterrent logic that unilateral action lacks. A producer deciding whether to add capacity assesses where the output will go. If the principal open markets act together and act predictably, the expected return on marginal capacity falls, and the investment case weakens before the capacity is built. On that reading, the framework is a signaling instrument aimed at investment decisions rather than an enforcement instrument aimed at existing tonnage, and it should be judged on capacity announcements over several years rather than on trade flows over several quarters.
A ministerial shaped by an unusual year
The Milwaukee meeting took place against a backdrop that no recent G20 trade host has had to manage.
The United States spent the first half of 2026 rebuilding its entire tariff architecture after the Supreme Court held in February that the International Emergency Economic Powers Act does not authorize presidential tariffs. Customs and Border Protection has since certified roughly $122 billion in refunds and interest for payment, and opened a further refund phase this week covering approximately 3,700 court plaintiffs with claims reported at around $134.7 billion.
The replacement program leaned heavily on the two instruments most visible at Milwaukee. Section 232 was expanded across metals in April and June, extended to polysilicon and derivatives at a 15 percent rate with minimum import prices in August, applied at 25 percent to certain advanced semiconductor articles, imposed at rates reaching 100 percent on drones and drone components, and applied at 100 percent to certain patented pharmaceutical products from Sept. 29. Section 301 investigations covering some 60 trade partners produced final forced labor tariffs effective July 24, with structural excess capacity investigations covering 16 partners still running.
That last proceeding is the direct domestic analogue of the Milwaukee discussion, and its existence shaped the room. Ministers were negotiating cooperative language on excess capacity with a host that already has a unilateral excess capacity tariff track in motion. Delegations reading that correctly understood that consensus language would constrain the United States less than it would constrain them, which is a plausible part of the explanation for why a handful of members blocked it.
Bilateral activity in the margins reflected the same calculation. India and Brazil both pursued separate negotiations during the summit rather than investing in the plenary outcome, a pattern consistent with a system in which bilateral accommodation delivers more than multilateral agreement.
Implications for importers, exporters and U.S. businesses
Steel-consuming manufacturers should plan on sustained elevated input costs. Nothing in the Milwaukee outcome points toward relief, and the coordinated barriers pillar points the other way. Firms that have been deferring capital decisions pending clarity on metals duties now have their clarity, and it is not favorable.
Importers of steel-derivative products face a specific and growing compliance burden. Enhanced information sharing among 28 economies is, in operational terms, enhanced scrutiny of origin and metal content declarations. Derivative product coverage under Section 232 has expanded repeatedly, and the metal content calculation on a mixed-material article is already one of the most error-prone determinations in customs practice. Importers relying on supplier attestations without independent verification are carrying more risk than they did a year ago.
Exporters of U.S. agricultural goods have a modest win and should treat it as such. The food weaponization consensus does not restrain any specific measure, but it establishes language that U.S. negotiators will invoke in future disputes, and it was the one item on which the G20 could agree.
Companies with supply chains touching partners under Section 301 excess capacity investigation, a proceeding covering 16 trade partners, should treat the Milwaukee failure as information about timing. The inability to secure cooperative multilateral action strengthens the internal case for unilateral determinations, and firms should not assume that a negotiated outcome will arrive before a tariff does.
More broadly, the Milwaukee ministerial confirms that the operative level of trade governance has moved from the multilateral to the plurilateral and, frequently, to the unilateral. Planning assumptions built on the expectation that disputes resolve through consensus institutions are increasingly out of date. The practical substitute is scenario planning around coalitions: which blocs a company’s suppliers sit in, which instruments those blocs are building, and how quickly a product’s duty treatment can change when a coalition acts.
The G20 presidency passes on, and with it the host’s ability to set the agenda. What stays behind is a 28-member framework with real commitments and a missing member, which is a reasonable summary of where trade policy stands at the end of 2026.
What to watch next
Three markers will show whether the Milwaukee Framework becomes an operating instrument or a communique with a city name attached.
The first is whether the information sharing pillar produces an actual data exchange with a schedule, a format and a secretariat. Commitments to share information are common in trade diplomacy and are frequently never operationalized. If signatory customs administrations begin exchanging capacity and flow data on a fixed cadence within the next two quarters, the framework is real. If the pillar is still described in future tense a year from now, it is not.
The second is whether any signatory takes a trade measure and explicitly cites the framework as its basis. Coordinated action is only observable when someone acts. The European Union and Japan both have active steel defensive instruments and both have domestic pressure to use them. A measure framed as a Milwaukee Framework action, rather than as a purely national safeguard, would establish the precedent the arrangement needs.
The third is the Section 301 excess capacity proceeding covering 16 trade partners. If final determinations in that case arrive before the framework has produced anything operational, the practical message to partners will be that the plurilateral layer is decorative and the unilateral instrument is the real one. That would make the next host’s job harder and would reduce the incentive for any economy to invest political capital in the framework’s development.
For companies, none of these markers requires specialist monitoring. They will be visible in ordinary trade press and in the Federal Register. What they will determine is whether the duty environment for metals and metal-intensive goods in signatory markets converges, which would simplify multi-market sourcing decisions, or continues to diverge, which would keep the current patchwork in place.
The Milwaukee ministerial produced less than its host wanted and more than its critics expected. Twenty-eight economies now have a named instrument and a stated set of obligations on the largest structural distortion in global goods trade. Whether that proves to be a foundation or a monument is a question for 2027.