Decision-Making: Consensus, Voting, Interpretations and Waivers

Waivers, worked

Article IX:3 and IX:4 together set the conditions, and they are strict: exceptional circumstances, a three fourths majority, and a decision that must state the exceptional circumstances, the terms and conditions governing the application of the waiver, and the date on which it terminates. Any waiver for more than a year is reviewed annually, and on review the Conference examines whether the circumstances still exist and the conditions have been met, and may extend, modify or terminate it.

A worked example the examiner will recognise. The TRIPS and public health waiver decided on 30 August 2003 allowed members to export medicines made under compulsory licence to countries without manufacturing capacity, which Article 31(f) would otherwise have prevented by requiring production predominantly for the domestic market. It was granted as a waiver first, then made permanent by amendment as Article 31bis, in force 23 January 2017. That sequence is the textbook illustration of a waiver used as a bridge to an amendment.

A second example shows the other use. The 1955 waiver to the United States on agricultural import restrictions is the reason agriculture escaped GATT discipline for forty years. A waiver is not always a humanitarian instrument.

The paralysis, stated fairly

The negotiating function has produced two multilateral agreements in thirty years: trade facilitation, in force 22 February 2017, and fisheries subsidies, in force 15 September 2025. The Doha Development Agenda launched in 2001 has never been concluded.

Three responses have grown up outside the consensus rule, and a strong answer names them. Plurilateral agreements among willing members, such as the Information Technology Agreement and the 2021 Reference Paper on Services Domestic Regulation, which bind their participants and, where the benefits are extended on a most favoured nation basis, benefit everyone. The Multi-Party Interim Appeal Arbitration Arrangement, which reconstructs an appeal by Article 25 arbitration among its participants without amending the DSU. And regional trade agreements, which is where the liberalisation that cannot pass in Geneva has actually happened.

Each of those weakens the multilateral system in order to keep something moving, which is the honest way to put the trade-off.

A worked example

A proposal comes to the General Council to extend the moratorium on customs duties on electronic transmissions. Follow what the rule actually requires.

The Chair puts the proposal. Under footnote 1 to Article IX:1 the body is deemed to have decided by consensus if no member present at the meeting formally objects. So the test is not agreement, and it is not a count of supporters.

One hundred and sixty members say nothing. Silence is consent for this purpose, and it is why a delegation with no instructions is under pressure to be absent rather than present and silent.

One member says it cannot join a consensus. The proposal fails. It does not matter that the objector is small, that every other member supports the proposal, or that the moratorium has been renewed at every conference since 1998.

Three consequences follow, and each is examinable.

First, the objector need give no reason. Nothing in Article IX requires one.

Second, the objection is a bargaining instrument. A member with an unrelated grievance can withhold consensus here to obtain movement elsewhere, and this is ordinary practice rather than abuse.

Third, the fallback is unusable. Article IX:1 permits a vote where consensus cannot be reached, by simple majority unless otherwise provided, and Article IX:2 permits an authoritative interpretation by three fourths. Neither has ever been used, because a member outvoted on a matter it regards as vital would be pushed towards leaving, and the membership has judged that risk to be worse than paralysis.

That is how the moratorium lapsed on 31 March 2026: not by a decision to end it, but by the absence of a decision to continue it.

What this does NOT mean

Consensus is not a veto in the Security Council sense. There is no class of privileged members; every member has the same power to object, which is formally egalitarian and practically favours the status quo.

And a decision by consensus is not a decision everyone likes. It is a decision nobody present is willing to be recorded as blocking, which is a much lower bar and is why package deals and constructive ambiguity are the currency of trade negotiation.

Quick revision

• Article IX:1: consensus continued from GATT; failing that, one member one vote and a majority of votes cast.
• Consensus is defined negatively in a footnote: no member present at the meeting formally objects. Absence does not block; silence does not block.
• Article IX:2: interpretations, exclusive to the Ministerial Conference and General Council, three fourths, may not undermine Article X.
• Article IX:3 and IX:4: waivers, exceptional circumstances, three fourths, must state circumstances, terms and termination date, reviewed annually if over a year.
• Worked: the TRIPS public health waiver of 30 August 2003, made permanent as Article 31bis in force 23 January 2017; and the 1955 agriculture waiver that cost the system a whole sector.
• Consequence: two multilateral agreements in thirty years. The workarounds are plurilaterals, the MPIA and regional agreements.

Test yourself

1. How is consensus defined, and why does the definition matter? The footnote to Article IX:1 defines it negatively: a body is deemed to have decided by consensus on a matter submitted for its consideration if no member present at the meeting when the decision is taken formally objects to the proposed decision. The definition matters in three ways. It distinguishes consensus from unanimity, because a member that is absent from the meeting cannot block and a member that is present but says nothing does not block, so the rule rewards attendance and rewards being explicit. It requires the objection to be formal, so that a member which dislikes a proposal but is unwilling to be recorded as its sole opponent will let it pass. And it fixes the moment of decision as the moment of objection, which makes the timing and sequencing of proposals a negotiating instrument in itself. The practical result is that a WTO decision is not one that everybody supports, but one that nobody present is willing to be seen to stop.

2. Explain the waiver power and give an example of its use. Article IX:3 permits the Ministerial Conference, in exceptional circumstances, to waive an obligation imposed on a member by the Marrakesh Agreement or any Multilateral Trade Agreement, by a three fourths majority, after a request considered under the consensus practice for a period not exceeding ninety days, requests concerning an Annex 1A, 1B or 1C agreement going first to the relevant council. Article IX:4 imposes the discipline: the decision must state the exceptional circumstances justifying it, the terms and conditions governing its application and the date on which it terminates, and any waiver granted for more than a year is reviewed annually to see whether the circumstances persist and the conditions are met, the Conference being able to extend, modify or terminate it. The best example is the decision of 30 August 2003 on TRIPS and public health, which waived the Article 31(f) requirement that production under a compulsory licence be predominantly for the domestic market so that medicines could be exported to countries without manufacturing capacity; it was later made permanent by the insertion of Article 31bis, in force on 23 January 2017. The 1955 waiver permitting United States agricultural import restrictions is the counter-example, since it removed a whole sector from discipline for a generation.

3. What has consensus cost the organisation, and how have members worked around it? It has cost the negotiating function. Since 1995 the membership has concluded only two multilateral agreements, on trade facilitation, in force on 22 February 2017, and on fisheries subsidies, in force on 15 September 2025, and the Doha Development Agenda launched in 2001 has never been brought to a conclusion. Three workarounds have grown up. Groups of willing members conclude plurilateral instruments, such as the Information Technology Agreement and the Reference Paper on Services Domestic Regulation agreed by sixty seven participants in December 2021, whose benefits are often extended on a most favoured nation basis so that non-participants gain without contributing. Participants in the Multi-Party Interim Appeal Arbitration Arrangement reconstruct an appellate stage between themselves by arbitration under Article 25 of the DSU, without amending the DSU itself. And members negotiate regional and bilateral agreements, which is where most trade liberalisation of the last two decades has actually occurred. Each workaround keeps something moving at the cost of the universality that was the point of the multilateral system.