Abstract

Globalisation creates a demand for international rule-making, and the WTO remains the forum for creating binding and enforceable international trade rules. The WTO therefore needs an effective decision-making system capable of resolving diverging interests. Although the WTO Agreement foresees votes where consensus cannot be reached, the practice of consensus dominates the process of decision-making. The GATT 1947 only spoke of voting and did not mention the word ‘consensus’. The WTO Agreement reflects the evolution from votes to consensus in the GATT 1947. Nevertheless, voting was not abandoned in the WTO Agreement. Although the consensus practice generally works well, it nevertheless implies the risk of deadlock and in particular the Membership’s inability to respond legislatively where it disagrees with a panel’s or the Appellate Body’s legal interpretation. Consensus also inherently favours the status quo and can make it extremely difficult to achieve change. Consensus does not provide for equality (in terms of decision and influence) because not every Member has the same ability to maintain vetoes. Consensus of course has many advantages, but it is questionable whether it is also more democratic than the majority rule.

The relinquishment of the political decision-making system strengthens

the (quasi-)judiciary beyond what was foreseen, thereby threatening its legitimacy. It would be fruitful to make operational the currently unused authoritative interpretation provided by the WTO Agreement and to support the quest for consensus in the WTO with a visible shadow of a vote. This would help encourage Members to block consensus only in defence of important interests. Increasing the implicit costs of the veto would be another instrument of this kind. Use should be made of Rule 33 of the WTO’s Rules of Procedure which requires subordinate Councils, Committees and other bodies of the WTO to refer a matter upwards whenever they are unable to reach consensus. For reducing the gap in effectiveness between the WTO’s political bodies and its dispute settlement, the most realistic options are those that involve small gradual steps. The intention of this paper is not to call for a hasty recourse to voting or for voting by small margins. The suggestion is that of reviving the possibility of voting as a potential solution of last resort.

I. Introduction

The World Trade Organization (WTO) currently has a membership of 148 sovereign States and independent customs territories. Its agreements cover some 95% of international trade and regulate the trade of goods and services as well as the protection of intellectual property rights. Its membership comes close to that of a universal organisation, even more so if one considers that a significant proportion of the remaining non-Members are currently negotiating their accession to the WTO. The reason why the WTO is important and unique, however, is also that it has been and continues to be the forum in which trade negotiations take place at the worldwide level in subsequent rounds. These negotiations result in binding international agreements that can be enforced in a highly effective, compulsory and exclusive quasi-judiciary.

Together with other factors, the strong increase of international trade (significantly faster than the growth of world GDP) and of other aspects of economic interaction (e.g. investment) has resulted in an increased international economic interdependence. In view of this interdependence, nation-state governments cannot regulate effectively any more in many areas,[1] which is why effective international governance is needed in order to manage globalisation. At a time when global governance is more necessary than ever before,[2]the WTO is a forum in which the international community can achieve many important things, given its rule-making vocation, its broad membership and its effective enforcement mechanism.

Yet, the WTO has not always presented itself as an organisation that pursues its agenda effectively and easily. Sometimes, or even regularly, it goes through periods of crisis and perceived or threatening paralysis. The ministerial conferences of Seattle and Cancún are two recent examples. Signs of inefficiency regularly also appear in Geneva, be it in the context of negotiating new agreements or in the context of the revision or even just application of the existing trade rules. A recent high profile example is the TRIPS and public health issue.[3] Arguably, the pressure imposed by the upcoming Cancún ministerial was essential in achieving a provisional solution in late August 2003, whereas the post-Cancún talks on a long-term solution have again been deadlocked. The big failures of course attract even more attention and also generate more reflection and criticism regarding the WTO’s effectiveness. The recent Cancún collapse certainly has aroused doubts about the organisation’s effectiveness[4] and has prompted people to see the WTO at a ‘cross roads’[5] or even in a ‘constitutional moment which parallels the creation of the GATT 1947 and the WTO.’[6]

Of course, the failure of Cancún, and previously that of Seattle, was rooted in the divide that exists between Members of the WTO on many issues of substance (agriculture, trade and investment, trade and competition, market access etc.).[7] In that sense, it has been said that the failures were programmed. It is obvious that if all Members were in agreement over the substance, the world trading system would work smoothly and there would be no crises. However, substantive divergences are to some extent a normal phenomenon (even though it is at times difficult to understand the existing divergence on particular topics). The fact that conflicting interests are a frequent reality in just about every polity or other organisation is precisely the reason why it is so important to have in place institutions that can balance these diverging interests. The situation of conflict is thus the situation in which an effective decision-making mechanism is most needed to resolve contentious issues. This is what has not worked well enough in the recent past.

In this sense, the European Commissioner for Trade, Pascal Lamy, has ascribed the collapse of Seattle and Cancún to the WTO’s ‘medieval’ decision-making process. After Cancún, discussions have taken place on the need and possibility of reforming the WTO. Given the prominence of the outstanding substantive issues, the post-Cancún discussions, however, soon returned to the specifics of the Doha mandate. Therefore, it seems worthwhile and important to continue to devote attention to the question of whether the WTO shows institutional deficiencies and how they could be addressed with a view to improvement. In terms of how to measure improvement, effectiveness and efficiency are of course not the sole considerations. The importance of transparency, participation and accountability, as well as other aspects of democratic legitimacy should in no way be discounted. Yet, it is submitted that promoting these higher values will be difficult or at least insufficient if the decision-making system is not effective.

A reform of the current system can mean two basically different things: on the one hand, one can think of changing the rules on decision-making in the WTO Agreement (along with changing the practice). On the other hand, reform can mean exploring the scope for improvement within the framework of the existing rules, i.e. changing the practice, but not the rules. We believe that, for both dogmatic and pragmatic reasons, the latter exercise should receive priority over the former. First, it would be extremely difficult to achieve a modification of the rules on decision making in the present context where the adoption of new multilateral trade rules is in general rather difficult.[8] Second, before resorting to proposing legislative change, one should explore the existing rules and the extent to which improvements are possible within their limits without formal change, as only such an exercise can reveal the need, if any, for legislative change.

With this in mind, we propose to take a closer look at the rules on decision- and rule-making in the WTO Agreement. Of course, there are other levels on which, more in the sense of fine-tuning, improvements can and should be explored, e.g. how ministerial conferences are organised, the role of the chairs etc. This paper, however, undertakes a more fundamental critique, also with the intention of keeping the debate on the WTO’s institutional reform alive, even when negotiations are currently gaining momentum again. There is also hope that the institutional debate will be revitalised when the Consultative Board set up by Director-General Supachai Panitchpakdi in June 2003 and chaired by former Director-General Peter Sutherland publishes its report.

II. Procedures for Making, Revising and Implementing Trade Rules in the WTO

One needs to distinguish between rule-making and decision-making, as these exercises are different in nature from a constitutional point of view. The formal rules of the WTO reflect this distinction, even though it largely disappears in the organisation’s practice.

• As is well known, the process of decision-making in the WTO is dominated by the practice of consensus. As is also well known, consensus means that ‘no Member, present at the meeting when the decision is taken, formally objects to the proposed decision’.[9] Often, at least one Member objects to a proposal, and in those circumstances, the next step is typically a protracted effort to reach consensus by overcoming the existing resistance, e.g. by finding a compromise. If this does not work, no decision is taken.

This contrasts with Article IX:1 of the WTO Agreement, which does not require consensus for all cases. While the first sentence states that ‘[t]he WTO shall continue the practice of decision-making by consensus followed under GATT 1947’, the second sentence allows votes: ‘except as otherwise provided, where a decision cannot be arrived at by consensus, the matter at issue shall be decided by voting’. Those decisions are reached with a (simple) majority of the votes cast. An exception is Article 2.4 of the DSU, according to which the DSB decides by consensus, with the notable exception of the reverse consensus mechanism for the key steps of a dispute settlement procedure.[10] Hence, except for the DSB, the bodies of the WTO would normally decide according to a two-step approach: consensus if possible, otherwise vote.

The Rules of Procedure contain quite detailed rules on how votes would take place. Rule 16 of the Rules of Procedure for Sessions of the Ministerial Conference and of the Rules of Procedure for Meetings of the General Council provides that a majority of Members must be present for votes to take place (quorum). Rules 29/34 specify that when decisions are required to be taken by vote, such votes be taken by ballot but that the representative of any Member may request, or the Chairperson suggest, that a vote be taken by raising cards or roll call. Where the WTO Agreement requires a vote by a qualified majority of all Members, the Ministerial Conference/General Council may decide that the vote be taken by airmail ballots or ballots transmitted by telegraph or telefacsimile. The respective Annex 1 of these Rules of Procedure contains further details for such airmail/telex/telefax ballots, inter alia a notice to be sent to each Member and a time-limit of a maximum of 30 days.[11]The Councils, Committees and other subordinate bodies of the WTO, however, are required by Rule 33 of their respective Rules of Procedure to refer a matter to the General Council whenever they are unable to reach a decision by consensus.[12]

• When rules are made from scratch, i.e. new international agreements adopted, it is no wonder that consensus generally governs the procedure. After all, the signatories of the agreement are to ratify the text (‘express their consent to be bound’), which is even more than consensus because a subject of international law becomes party to the agreement only by express (and typically written) consent.

Nevertheless, it is worth pointing out that Article 9(2) of the Vienna Convention on the Law of Treaties foresees that the ‘adoption of the text of a treaty at an international conference takes place by the vote of two-thirds of the States present and voting, unless by the same majority they shall decide to apply a different rule.’ Yet, this only goes for the adoption of the text which does not yet result in the States being bound. A majority vote in which up to a third of the negotiation participants are outvoted, therefore risks reducing the number of States that will later sign up (and ratify). Even if there are in many cases other reasons for non-signing or non-ratification, it is interesting to point out that the number of signatories of many UN-sponsored conventions is far below the number of conference participants, in fact this is the fate of the Vienna Convention itself.

In international trade, it is desirable that the number of countries that sign up to the agreements be as large as possible for economic[13]and legal reasons.[14] It is therefore productive if trade agreements are shaped in such a manner that, if possible, all become a party. This involves a search for compromises, persuasion and sometimes a certain degree of pressure on other States. Sometimes, an agreement with partial reach is better than no agreement, and in those cases plurilateral agreements are the best choice. However, Article X:9 of the WTO Agreement requires consensus of the Ministerial Conference for adding a plurilateral agreement to Annex 4.[15]

The required consent of every single State for that State to be bound by an international agreement constitutes an in-built preference for the status quo in international law (by default, this status quo amounts to a lack of legal disciplines, otherwise the status quo comprises those legal disciplines that have emerged so far). This contrasts with domestic democracies (representative or direct) where simple majority votes are formally neutral on making or not making, unmaking or changing rules. Obviously, in comparison, international rule-making is highly cumbersome and less effective, possibly more cumbersome and less efficient than it should be in the light of today’s demand for international governance in a world of increased international interdependence and eroding independence of single States as regulators.

a) Amendment

The default rule on amendments in public international law is Article 40 of the Vienna Convention, according to which an amendment does not require the consent of all parties, but obviously no party is bound by the amendment unless it gives its consent. With one exception, Article X of the WTO Agreement is stricter. It first provides that the Ministerial Conference must approve an amendment proposal with a two-thirds majority of the Members, if it cannot reach consensus within 90 days. Then, two thirds of the Members must accept (i.e. ratify) the amendment for it to become effective: for all Members, where the amendment does not alter substantive rights and obligations; for those who accept the amendment, where it does alter substantive rights and obligations.[16] The former procedure, however, requires a three-quarters majority decision by the Ministerial Conference. Amendments to the DSU require consensus, which is less burdensome than the normal amendment procedure in that no ratification is needed. Modifications of Articles IX and X of the WTO Agreement, of the most-favoured-nation treatment rules and of Article II of GATT 1994 (on bindings) require every Member’s consent.

b) Accession

A special form of amendment is the accession of a new Member to the WTO. Such accession is an amendment of the WTO Agreement because this Agreement is modified so as to cover an additional subject of international law. Legally, the standard WTO Accession Protocol amends the WTO Agreement by becoming an integral part of the WTO Agreement.[17] Nevertheless, the Accession Protocol is an agreement between the new Member and the WTO (Article XII:1 of the WTO Agreement), not an (amendment) agreement between the new and the old Members. In terms of decision-making, Article XII:2 stipulates that the Ministerial Conference approves the accession agreement by a two-thirds majority of the Members. Yet, when a new Member accedes, Article XIII permits Members to exclude the application of the WTO Agreement in relation to the new Member by so notifying the Ministerial Conference.

c) Renegotiation of Commitments

In the WTO Agreement, rights and obligations are also set out in each Member’s schedule of commitments. As is known, this part of the Agreement accounts for the majority of the famous 25,000 pages. If a Member intends to modify or withdraw a GATT concession (typically a tariff concession), Article XXVIII of the GATT 1994 provides for the possibility to do so according to a procedure that is considerably lighter than the amendment procedure under Article X of the WTO Agreement. Preferably, that Member should reach agreement with the other Members primarily concerned (principal supplier(s) and Members holding an initial negotiation right) and with those having a substantial interest, i.e. only a subset of WTO Members. If no agreement is reached, the Member in question can nevertheless proceed (unilaterally) with the modification or withdrawal of its concession and the other Members with rights under Article XXVIII may then withdraw substantially equivalent concessions initially negotiated with that Member.

Article XXI of the GATS provides for a similar, but slightly stricter procedure for a Member that wishes to modify a commitment it has made in its services schedule.

d) Waiver

In exceptional circumstances, the Ministerial Conference and the General Council may waive particular WTO obligations of any given WTO Member by a three-quarters vote.[18] Waivers are exemptions for certain Members from specific WTO obligations. They must be temporary (although they can be extended) and reviewed annually.

e) Authoritative Interpretation

A special instrument foreseen in the WTO Agreement that can be used to refine or revise multilateral trade rules is the interpretation provided for in Article IX:2. This instrument is an invention of the Uruguay Round, it did not exist under the GATT 1947. In Article 3.9 of the DSU, it is referred to as the ‘authoritative interpretation’, which allows it to be distinguished from the kind of interpretation performed by panels and the Appellate Body in clarifying the provisions of the WTO Agreement.[19] Article IX:2 attributes the responsibility for adopting such interpretations to the Ministerial Conference and the General Council and stipulates a decision by three-quarters majority of the Members and – for interpretations of the GATT, multilateral agreements on trade in goods, the GATS and the TRIPS Agreement – that there has been a recommendation by the respective Council (for Trade in Goods/Services/Intellectual Property). Article IX:2 also states that it must not be used in a manner that would undermine the amendment provisions in Article X.

Although the legal effect of an authoritative interpretation is not spelt out in Article IX:2 of the WTO Agreement, it is relatively clear that such an interpretation would bind all Members.[20] It has also been suggested that, unlike panel and Appellate Body reports and DSB rulings and recommendations,[21] an authoritative interpretation may add to or diminish rights and obligations of Members under the WTO Agreement.[22] This latter aspect is somewhat contested, also on the basis of the last sentence of Article IX:2, which prohibits undermining the amendment provisions. Yet, if an authoritative interpretation were not able to modify the law, it could only clarify existing obligations in accordance with the Vienna Convention interpretation rules. This does not make much sense for a decision emanating from a political organ and it would excessively narrow the purpose for which an authoritative interpretation could be used. Also, it would arguably mean that the Appellate Body could revise any such interpretation as regards whether it constitutes a ‘permissible’ interpretation of the relevant provisions under the customary rules of interpretation (in order to be valid). Otherwise, if an ultra-vires interpretation were nevertheless to be binding on (i.e. non-reviewable by) the Appellate Body, the whole question would be academic. Thus, the General Council arguably need not apply the rules of treaty interpretation in formulating an ‘authoritative interpretation’, in other words it may modify WTO law (arguably this would also be the case in the event of an authoritative interpretation intended to fill gaps or ambiguities in the texts of the agreements). The attribute ‘authoritative’ would seem to further support this thesis. The prohibition on undermining the amendment provisions in Article IX:2, last sentence, certainly imposes a limit on the extent to which the authoritative interpretation can serve for the purpose of revising trade rules. However, the verb ‘undermine’ is relatively strong, which permits reading that proviso somewhat restrictively, also to avoid making Article IX:2 redundant and void of effect. According to such reading, not every fine-tuning of a provision of the WTO Agreement would immediately ‘undermine’ the amendment provisions.

Theoretically speaking, the authoritative interpretation under Article IX:2 of the WTO Agreement is of high potential relevance. It gives the political bodies of the WTO an opportunity to refine existing trade rules. This can serve to determine the scope of rules in a prospective manner, but also to correct an interpretation given by a panel or the Appellate Body, whose rulings can no longer easily be blocked.[23] The quasi-automaticity of the adoption of dispute settlement reports makes the authoritative interpretation a necessary instrument of checks and balance vis-à-vis the WTO’s quasi-judiciary. If, unlike under the GATT 1947, individual WTO Members can no longer veto the adoption of a report, in fact even an overwhelming majority of WTO Members could not do so as long as one Member (presumably the winner, but in fact any Member) insists on adoption, corrections of jurisprudential developments should be possible to allow for a legislative response. The authoritative interpretation should perhaps not be viewed exclusively through the lens of dispute settlement, but for the reasons mentioned, such a strong correlation exists, as Article 3.9 of the DSU corroborates.

Whether an authoritative interpretation that is adopted during a pending dispute (i.e. after the panel has been established and before the Appellate Body issues its report) has legal effect on the outcome of this dispute depends on the relevant point in time for the legal evaluation of the matter in dispute. Although the WTO jurisprudence is somewhat unclear and arguably also inconsistent on this point, the tendency is to focus on the facts (the challenged measure) as they existed at the time of the establishment of the panel. This would theoretically preclude taking account of subsequent legal modifications, as far as substantive obligations are concerned. Nevertheless, at the stage of implementation, a losing Member would arguably have to (and be entitled to) be guided by the authoritative interpretation adopted in the intervening period. It might therefore be worthwhile to clarify this issue so as to prevent interference with pending disputes (and resistance from the party fearing a disadvantage for its litigation) by limiting any effect of an authoritative interpretation to other (future) cases.

B. How the WTO rules came about and how they were intended to operate

If one compares the WTO rules on decision-making, in particular Article IX, with the consensus-dominated practice of the WTO, one may wonder why such rules were incorporated that foresee votes when consensus cannot be achieved. When the WTO Agreement was drafted in the Uruguay Round, was there a belief that this would remain dead letter? In the search for a response, it is worth exploring the historical background. This is, on the one hand, the law and the practice under the GATT 1947. On the other hand, it is worthwhile to explore, to an unavoidably limited extent, the intentions and expectations of the negotiators during the Uruguay Round when they formulated the WTO Agreement, notably Article IX.

1.The GATT 1947 and evolving practice

When one reviews the institutional provisions of the GATT 1947, it becomes clear that Article XXX of the GATT 1947 inspired Article X of the WTO Agreement on amendments, and Article XXXIII of the GATT 1947 inspired Article XII of the WTO Agreement on accessions. Article XXV:4 of the GATT 1947 states that: ‘Except as otherwise provided for in this Agreement, decisions of the Contracting Parties shall be taken by a majority of the votes cast’.[24] Article XXV:3 gave each contracting party one vote. Special majorities were called for in Articles XXIV:10, XXV:5 and XXXIII. Article XXIV:10 provided for a two-thirds majority for approving a regional trade agreement that does not fully comply with the requirements of Article XXIV:5-9. Article XXV:5 provided for waivers of obligations but required that ‘any such decision shall be approved by a two-thirds majority of the votes cast and that such majority shall comprise more than half of the contracting parties.’

Voting did take place, but routinely only on decisions for waivers under Article XXV:5 and on accessions under Article XXXIII of the GATT 1947. In relation to other business, the Contracting Parties did not usually proceed to a formal vote in reaching decisions, but the Chairperson took the sense of the meeting.[25] Even on waivers, a consensus in the GATT Council very often preceded the votes. Notable exceptions prove this rule, and one such situation occurred in 1990 at the annual session of the Contracting Parties when the EEC requested a vote by roll call on a waiver for the German Democratic Republic’s trade preferences to former Soviet bloc countries. Despite the surprise and confusion this caused to many delegated who did not even have time to seek instructions, the unperturbed Chairman applied the existing procedures and immediately proceeded to the vote by roll call.

In the early days of the GATT, the Chairman of the Contracting Parties often resolved questions of interpretation through rulings that were tacitly or expressly accepted or put to a roll-call vote.[26] Over the years, decision-making by consensus became increasingly prevalent with the number of developing countries entering the international system in the wave of decolonisation and their accumulation of large voting majorities, although this is not a sufficient explanation if one looks at the early days of the GATT.[27] The GATT Analytical Index stated in 1995 that the most recent recorded decision of the Contracting Parties adopted by vote, other than decisions on waivers or accession, was in 1959.[28] However, the United States called for and obtained a vote in 1985 on whether to hold a special session of the Contracting Parties for the purpose of launching a new round of negotiations (the Uruguay Round).

The GATT 1947 is thus a partial answer to the question of where the rules on voting in the WTO Agreement come from. When the WTO Agreement was drafted, the evolution from votes to consensus (a term that did not even appear in the GATT 1947) was reflected in Article IX:1 of the WTO Agreement by making consensus the first choice. Article XVI:1 reinforced this by stipulating that the WTO be ‘guided by the decisions, procedures and customary practices followed by the Contracting Parties to GATT 1947’. Yet, it is important to note that voting was not abandoned in the text of the new Agreement and one can also say that the text gives it a more prominent role (vote when no consensus) than the GATT practice had (outside the area of waivers and accessions). This justifies posing the question about the intentions and expectations of Uruguay Round negotiators.

2.Uruguay Round negotiations on the establishment of a World Trade Organization

One must recall that the negotiations on the Multilateral Trade Organization (MTO, later in the negotiations to become the WTO) as an institution and international organisation had started between the EEC (later becoming the EC) and Canada, subsequently they also involved Mexico. This resulted in the ‘Dunkel Draft’ version of the Agreement Establishing the MTO. The United States joined in only at a later stage when it became interested in the MTO as a vehicle for the single undertaking. The United States disliked the draft text because it perceived the decision-making rules to be stronger than those in the existing GATT and considered these to be a threat to its sovereignty. It was already clear that there would be a strong majority of developing countries in the MTO/WTO. Some countries therefore intended to avoid the risk of frequent votes which, if taken along developed versus developing lines, could have resulted in majorities adverse to their interests. Thus, in the fall of 1993, some major players worked hard on previous drafts in order to constrain the decision-making process.[29]

One of the main objectives of the United States was to change the MTO text and to make it as difficult as possible to take decisions. The main concerns were (1) that developing countries would try to use the decision-making voting rules to get out of their obligations later on (note that Footnote 4 to Article IX:3 of the WTO Agreement exceptionally requires consensus for waivers of transition periods) and (2) that the United States’ sovereignty would be undermined by amendments forced through by quickly formed majorities. The latter was ultimately protected by a return to the GATT approach for amendments with an impact on rights and obligations. Also, the United States successfully fought for the combination of the three-quarters majority rule and the prohibition to undermine the amendment procedure in the context of authoritative interpretations (Article IX:2).

On other types of decisions, the ultimate compromise consisted in laying down a two-step approach: first consensus, and if necessary, as a second step, voting. In a way, this codified ‘consensus’ which had previously not been part of the GATT text, but had been the practice. At the same time, one must note that the two-step approach does not reflect a practice in which practically no votes took place other than on accessions and waivers (that are regulated elsewhere than in Article IX:2). Thus, the fact that voting was not eliminated from the MTO draft and that qualified majorities were introduced (for interpretations) and increased (for waivers) where qualified majorities had existed in the GATT, can be taken as a sign that the Members involved, notably the United States, at the time accepted the idea that votes on such matters would take place.

C. Comparison with other international organisations

It is interesting to compare the situation in the GATT/WTO with that of other international organisations. It would seem that only international organisations operating at the universal level would be relevant for this comparison. The fact that organisations of regional integration sometimes have more advanced decision-making mechanisms is due to their higher level of ambition in terms of integration. Therefore, and also because of the more limited diversity in their membership, these organisations do not lend themselves to a comparison with the WTO.

For reasons of space and brevity, this paper will not set out in detail the decision-making process of multilateral or (nearly) universal organisations other than the WTO. For present purposes, it should merely be pointed out that voting is an in-built mechanism in the United Nations both in the General Assembly and in the Security Council. Every State has one vote. In the General Assembly, however, a trend towards consensus rather than formal votes has emerged over the past decades. In the Security Council, the body that adopts decisions that are binding for all UN members including decisions on war and peace, votes are standard practice. In addition, only a fraction of the UN members, 15 States, are represented in the Security Council. This fact, however, may be precisely the reason why voting does not seem to be an issue. With its 15 members, the Security Council automatically is a representative body acting on behalf of the entire membership. This may make it more acceptable that decisions be taken by a majority, since the authority is a representative one anyway.

The Bretton Woods organisations have voting mechanisms but, unlike in most other organisations, not every member has the same number of votes. Instead, weighted voting applies. The principle is the same in the other development banks for Asia, Africa and America. Nevertheless, also in these financial institutions, many decisions are adopted by consensus, not through formal votes.

III. Subsequent Practice

A. Decision-Making Practice of the WTO (as Compared to the GATT)

The WTO did not only continue the practice of decision-making by consensus as it had emerged under the GATT 1947. Soon, the WTO even replaced votes with consensus where votes had existed in the GATT, such as in relation to accessions and waivers. Practice differed only in 1995, the first year of the WTO when the General Council submitted the draft decisions on the accession of Ecuador and on certain draft waivers to a vote by postal ballot (after reaching consensus on the contents of these decisions).[30] Thereafter, the General Council agreed on a statement by the Chair that waivers and accessions would be decided by consensus. Nevertheless, the statement makes clear that where consensus is not achieved, the matter shall be decided by voting and that the agreed procedure does not preclude a Member from requesting a vote.[31]

The systemically important tool of authoritative interpretations has remained completely unused. Only once did a Member, the European Communities, attempt to obtain an interpretation in order to resolve the so-called ‘sequencing’ issue regarding the relationship between Article 21.5 and 22.2 of the DSU. The European Communities specifically suggested that the decision-making procedure foreseen in Article IX:2 of the WTO Agreement could be used without delay if consensus could not be achieved.[32] It might also be noted that there has been only one proposal to use Article X to amend the WTO Agreement, on the same issue.[33]

Thus, no legislative response came from the Membership in the famous amicus curiae row, in which an overwhelming majority of the Members fiercely criticized the Appellate Body for transgressing its competences by stating that panels and the Appellate Body itself could accept unsolicited briefs.[34]

In the TRIPS and public health saga, the implementation of paragraph 6 of the Doha Declaration was not achieved within the December 2002 deadline. Up until August 2003,[35]consensus could not be reached because one single Member felt unable to abandon its resistance against the proposed draft waiver. The question at issue was presented by some to be one of life or death for thousands of people in Africa. Yet, no Member considered requesting a vote.

Recently, towards the beginning of the Cotton dispute, Brazil requested the Dispute Settlement Body (DSB) to appoint a facilitator pursuant to Annex V of the Agreement on Subsidies and Countervailing Measures, an appointment which the United States opposed. At some point, the question was put to the Chair whether consensus was truly necessary for that appointment. One would think that Article 2.4 of the DSU, which generally requires consensus (and rules out votes) for DSB decisions under the DSU does not apply to a decision under a different agreement, so that Article IX:1 of the WTO Agreement applies. The Secretariat (Legal Affairs Division) nevertheless took and maintained the position that the appointment was only possible through affirmative consensus.[36] In the end, the DSB did not appoint any facilitator and thus failed to fulfil its obligation under the Agreement on Subsidies and Countervailing Measures.

The daily practice of the WTO offers quite a few other examples where the consensus requirement has resulted in a deadlock. As examples one could adduce the rules on derestriction of documents, the observership of other international organizations, Iran’s accession request and the consistent inability of the Committee on Regional Trade Agreements to reach a conclusion on the free-trade or customs union agreements it reviews. The extent to which that deadlock is protracted and whether a breakthrough is possible at some point obviously depends on such factors as the political context of the question at issue and whether trade-offs are made.

B. Effects

The result is that it is in theory possible for any Member to block any decision. If consensus cannot be achieved, no vote takes place, contrary to what Article IX:1 of the WTO Agreement suggests. The flipside is that even an overwhelming majority of Members are not able to achieve what they want to decide if at least one Member maintains a veto.

Such a decision-making structure contains an in-built preference for the status quo. It is much easier to maintain the current legal situation than to achieve change.[37]

The practical impossibility of a vote means that the negotiations in search of a consensus do not even take place in the shadow of a threatening vote. The only shadow that exists is the shadow of public exposure for the Member(s) opposing the consensus and the shadow of a crisis for the organisation. In purely mathematic terms, one has to recognise that the likelihood of at least one Member opposing a decision increases with the number of Members. This creates a real danger of paralysis.[38]

Whom does this situation favour? Sometimes it is said that the consensus requirement favours the small Members, sometimes it is said that the developed countries benefit most, since they are a minority.[39] Yet, each of these propositions makes the assumption that the respective group would typically find itself in a minority in which it could be outvoted. Formally speaking, consensus protects every single Member, whoever may be in a minority.

Does consensus provide for equality? In theory yes, because any single Member can block any decision. Where all must (at least tacitly) agree, it does not even matter whether or not all Members have the same amount of votes, given that a single opposing Member is sufficient for blocking a decision. Thus, consensus also operates as a way to avoid dealing with the respective weights of different Members’ votes.[40] The proposition that consensus provides for equality among Members, however, is flawed in that it wrongly assumes that any Member is equally able to sustain a veto.[41] Where a Member is alone in opposing a decision, it can find itself in quite some isolation and exposed to quite some pressure which arguably only robust, big Members can sustain for an extended period of time.

Accordingly, it seems unavoidable that the proposed texts that emerge in a negotiation process reflect the views of different Members to very different degrees. These texts arguably give more weight to the positions of Members who are less likely to give up their veto than to the views of Members with weaker consensus resistance capacity. This capacity tends to be linked to their size and importance in international trade.[42] In a way, therefore, consensus is a partial substitute for weighted voting.[43]

It has been said that the negotiation process which is overshadowed by the danger of any Member’s veto tends to be less transparent because negotiations take place in informal mode and are often not recorded. Yet, it would seem that this is not inherent in this type of negotiations and can equally be the case where a formal majority vote marks the end of the procedure.

IV. Potential and desirability of improvements

A. Advantage of the current system

The advantages of consensus are obvious and should in no way be downplayed. Where a decision is taken on the basis of consensus, it will tend to enjoy broad support, at least no one expressly opposed it. The decision achieved through negotiations resulting in a mutually satisfactory compromise also means that no one loses face.[44] There is often no open battle, at least no open tensions emerge from the situation, and implementation is secured. One could say that the consensus requirement protects the delicate balance between international regulation and national sovereignty.[45]

Consensus is powerful and effective if the majority wishes to secure the cooperation of the minority in the implementation of the decision. In that sense, majority voting can be ineffective and damaging if it risks alienating powerful or disaffected minorities.[46] Consensus is built on a broader and often stronger basis. One must note, however, that this is likely to affect the substance because the search for consensus regularly involves the search for a compromise solution that is somehow acceptable to all. The outcome will in this way also reflect the stake that various Members have in what is at issue, and their influence.[47]

One can thus consider as an advantage the fact that no decisions are likely to be taken against the opposition of the big and mighty, who generally need to implement the decision for it to have practical value. One can further argue that (due to the need to actively object) the consensus system often makes reaching a decision easier than voting where a certain threshold of affirmative votes must be reached (passive abstention does not contribute to reaching the required number of votes). Put differently, consensus is possible even where several Members do not agree with every aspect of a decision, but they do not formally object. In addition, the WTO does not apply the quorum requirements of the Rules of Procedure when decision-making is by consensus, which is convenient where meetings have a level of attendance below the quorum (majority of the Membership, i.e. 75 Members at present). Accordingly, the supermajorities required for certain votes are even more difficult to reach in the meeting room, and recourse to postal ballot or telefax does not necessarily yield a high response rate within the deadline. Thus, an important reason for the replacement of votes by consensus on accessions and waivers after Ecuador’s accession was the fact that the postal ballot votes regarding that accession arrived in small numbers and late.

It has also been said that consensus is not necessarily popular, but that for both developed and developing WTO Members it is the least bad alternative. Developed countries fear being outvoted, while developing countries fear being presented with faits accomplis.[48]

Even if not all Members’ interests are protected under the consensus system, because not every Member can oppose any disliked decision, vital interests are. One would have to qualify this, however, by saying that this is true only if these interests are threatened by a modification of rules (a decision to be adopted), not if vital interests create the need for some decision. In the latter case, the consensus requirement makes it extremely difficult to pursue those vital interests. Consensus therefore creates a trade-off between the ability of easily objecting and the difficulty of achieving desired decisions.

In its public relations, the WTO also lauds consensus as being more democratic than majority rule.[49] We do not intend to enter into a political theory debate at this point. However, there are some doubts about this beautiful sounding democracy argument, if one thinks of the situation in which a decision supported by an overwhelming majority of Members is blocked by one or several governments, and possibly by governments that lack democratic legitimacy at the domestic level.

B. The problems with the current system

The current practice potentially threatens the effectiveness of the political decision-making process of the WTO. It seems that the GATT decision-making system worked well because there were far fewer countries and the issues were less complex.[50] Also, the membership was less diverse than it is nowadays.

It is not the intention of this paper to downplay the mentioned advantages of consensus or to give the impression that the WTO decision-making process does not work at all. In fact, one must acknowledge that the consensus system works remarkably well in many and even most instances, and not all of them are necessarily easy topics (an example would be the appointment of Appellate Body members). Yet, the problematic flipside of the mentioned advantages are the known and unknown decisions that are not adopted and those that would be substantially different. Also, even where the decision-making mechanism works, the outcomes are bound to be the lowest common denominator and the process can take an excessively long time. Although this ineffectiveness has its roots in a voluntary choice by the Members of the WTO, it can become a real problem in certain circumstances. First, the WTO does not deliver where there are real demands for rule-making in the face of today’s economic interdependence.[51] The dispute settlement system is neither able nor authorised to meet all these demands by way of dynamic interpretation of the provisions at issue. Second, there is an inherent danger of potential crisis and paralysis,[52]and of the WTO losing its importance when it does not deliver on topics that are difficult and important.

Thus, if Members’ efforts to find compromises do not take place in the shadow of a possible vote, the existing shadow of institutional crisis is not an appropriate substitute, as such a crisis is remote from single issues. This is even more the case for the shadow of a WTO that loses its importance,[53]a shadow that may have worked in Doha. Also, both these shadows of crisis and reduced relevance are too negative in nature for the day-to-day operation of an organization – perhaps similarly to the shadow of divorce not being a good tool for living a successful marriage day to day.

One can also say that under-use of the political/legislative decision-making systems, albeit voluntary, is not strengthening the legitimacy of the WTO.[54] Thus,

the inability of the political organs to reach difficult decisions is important already of itself. This inability is of course due to a lower level of convergence among Members on matters of trade policy than existed at earlier periods of time and also during the Uruguay Round. Nevertheless, it is submitted that these substantive differences could be less visible and less detrimental if the decision-making process were more effective.

In addition, the political paralysis becomes problematic when seen in the context of the active and effective dispute settlement system that has been created in the Uruguay Round. Indeed, the contrast between the very burdensome political decision-making process and the highly effective,(quasi‑)automatic dispute settlement system appears like an institutional paradox,[55]when previously, under the GATT, both areas were dominated by the consensus rule. Of course, there are inherent differences between these two kinds of processes, that make it impossible to simply transfer the mechanism used in dispute settlement (where a small and odd number of independent adjudicators must decide on the basis of the law) to the political/legislative area (where a large number of government officials bound by instructions are free to adopt decisions of open content or not to adopt them). Nevertheless, the imbalance is problematic and in the long run also dangerous for the WTO.[56]

An independent (quasi-)judicial system, in which norms are clarified and thereby developed, should not be left without a (democratically more directly legitimised) counterweight. If legislative response to judicial developments is not available or not working, the independent (quasi-)judiciary becomes an uncontrolled decision-maker and is weakened in its legitimacy. In domestic systems, such mechanisms of legislative response are usually available and important from a democratic point of view. Although legislative reversals of judgments remain the exception, they do occur once in a while and are important in terms of determining who has the final say on what the law is.[57] In the WTO, legislative response is theoretically available, mainly in the form of amendments and authoritative interpretations. Yet, the mechanism does not actually work.[58] A good example is the amicus curiae issue, where one might have expected a legislative response, given the vehemence of the reactions of an overwhelming majority of the Members.[59] This example also demonstrates the detrimental tensions that can arise between the different ‘branches’ of the WTO, if such disagreements are not resolved, but instead remain a contentious issue.

In a way, it is puzzling that legislative response does not seem to work in the WTO, where the dispute settlement system formally depends on the political institutions. The consensus rule is probably not a sufficient explanation for this phenomenon, especially if one explains the strong adherence to consensus with the concern to protect national sovereignty and to avoid supranational decision-making authority. This does not sufficiently explain the dysfunction of legislative response because the resulting loss of Member-control is even greater when the questions at issue are surrendered to dispute settlement panels and the Appellate Body,[60]which, if asked and unable to avoid the issue, can be forced even to address the most sensitive questions.[61]

In that sense, one can observe that Members voluntarily surrender their decision-making powers in the interest of avoiding divisive votes. It has also been suggested that Members may have a (possibly partly unconscious) preference for deferring decisions to the judiciary because: this allows Members to take less clear positions on issues that are contentious also at the national level; lowest-common-denominator compromises result in ambiguity and give the possibility to later blame the dispute settlement system; linkages can be avoided; Members prefer to focus on specific cases rather than on rule-making for the future; dispute settlement can be used for domestic political ends, both by the complainant and the respondent.[62] This explanation is of course not complete, but it does contain plausible elements. Also, it does not apply to all potential questions, because not all of them are, for legal reasons, candidates for adjudication in dispute settlement. Conversely, not all questions dealt with in dispute settlement would lend themselves to legislative rule-making instead of the dispute, since Members often bring disputes when there is simply a breach of obligations (e.g. national treatment), which as such are not problematic or contentious, and when a waiver is merely a theoretical option.

As the example of the sequencing issue demonstrates, it also seems that Members are sometimes reluctant to resolve a single issue (by way of amendment or interpretation), but prefer to include this issue within larger negotiations where a concession might be obtained from those who propose the modification.

A further, also incomplete explanation for the strong adherence to consensus can probably be found in the human tendency of inertia. By this we mean the widespread preference for continuing to do things the way they have always been done, rather than trying out something new and foreign, especially when it is unclear what might be the consequences. For this and other reasons, one is likely to encounter this degree of risk-aversion among many of the WTO Members’ delegates in Geneva, many of whom have become acquainted with the WTO by observing its practice and have a strong sense of upholding institutional traditions. If one does not read the WTO Agreement, the Rules of Procedure or the few existing official documents on decision-making, one cannot be certain to come across the fact that the WTO Agreement mentions votes.[63]

In reference to a longstanding mantra and the title of the part of the conference for which this paper was initially written,[64]one might provocatively state that the WTO can claim to be a ‘member-driven’ organisation only if the Members actually sit on the driver’s seat and actually drive (forward), not if they merely press down the brake. Otherwise ‘member-driven’ is reduced to an indirect claim regarding who should not be driving the organisation.

Be it as it may, the under-use of the political decision-making mechanisms results in a dispute settlement system that is even stronger than according to the WTO’s formal design.[65] In the long run, this imbalance is unhealthy for the WTO as a whole and uncomfortable even for the dispute settlement system, notably the Appellate Body (due to an awareness that mistakes or disapproved legal interpretations will not be corrected). In addition to the lack of legislative response, there is a danger that issues best left to rule-making are handed over to the dispute settlement process,[66]a problem known also from domestic settings. Such deferral involves an inherent legitimacy problem.

C. Possible ways of improvements

As already explained in the introduction, it appears that the exploration of possible improvements should first and foremost take place within the framework of existing rules on decision-making. Indeed, it is rather unlikely that WTO Members are willing to revisit the basic rules on decision-making,[67] also against the background of how these rules came about in the Uruguay Round.[68]

If one takes the existing rules as a given framework, improvements are nevertheless possible at various levels. In order to increase the effectiveness of the decision-making system, it is worth exploring possibilities to mitigate the potential pitfalls of the consensus system as we know it. It is submitted that this should include the question of the role that the availability of voting, not necessarily holding that many actual votes, could play. Despite all the advantages of the consensus mechanism, it has been seen that it also brings about many downsides and periodically also the danger of paralysis. In some of these circumstances, voting may even appear as the lesser evil,[69]given its ability to resolve a contentious issue.

However, given the one-Member-one-vote principle, decisions on substantive matters that are based on simple majorities would not be representative of the realities of international trade, measured in terms of actual participation and weight of different WTO Members in international trade.[70]^,[71] Also, the imbalance between the United States (one vote) and the European Communities (25 votes)[72] would make it difficult to move to the acceptance of simple majority votes as seemingly foreseen by Article IX:1, second sentence, of the WTO Agreement.[73] This however, essentially argues against narrow simple majorities to serve as the basis for decisions. The arguments are not equally valid if one thinks of qualified majorities as they are foreseen in the WTO Agreement for certain questions, or even overwhelming majorities as they are likely to often exist in practice.

Also, it may not be at all necessary to actually hold votes on important substantive issues in order to achieve an improvement. One would normally consider that the fact that voting is available under the existing rules, alone, should limit the risk of consensus leading to a paralysis of the WTO.[74] This however can only work if the possibility of a vote is a shadow under which the quest for consensus takes place. This ‘shadow of a vote’ must be visible and not absent as it presently is, which leaves only the ‘shadow of isolation’, if only one or extremely few Members prevent consensus, and the ‘shadow of crisis’ for the WTO as a whole.

In this sense, a possible proposition would not be that of abolishing the rule of consensus, but of abolishing the taboo of majority decision-making.[75] Also, it should be recalled that voting or consensus is not a binary choice and that there may be many variants between the two that are worth exploring.[76] A possibility could be the distinction between decisions that can only be adopted by consensus from decisions that may be adopted by majority voting, as a matter of course or under certain qualifying conditions.[77]A plausible suggestion of this kind would be the introduction, in the practice of decision-making, of a distinction between procedural aspects and real substance.[78] This could facilitate overcoming the currently existing problem that even procedural issues of minor importance can get stuck in a deadlock or become the object of protracted consultations until consensus is reached. As a remedy, it has been proposed to use a written procedure for such decisions or to resort to voting when consensus cannot be reached within a certain deadline.[79] The value of such a modest step in the context of procedural issues should not be underestimated in terms of gradually revitalising the underlying dynamics of decision-making.

In a different way, improvements could be explored by aiming at reducing the likelihood of individual Members interjecting their veto and thereby blocking consensus. For this, one would have to find tools that would encourage Members not to block consensus or to do so only in defence of vital interests. The possibility of a vote, albeit distant, could be such an instrument. Increasing the implicit costs of the veto would be another, e.g. by resorting to Rule 33 of the Rules of Procedure (see below) or by finding other means of exposing the blocking Member to internal and external criticism etc. The objective could be a system in which individual Members refrain from blocking consensus if an overwhelming majority supports a decision.[80]

As a more radical tool, actual votes can of course have the effect of discouraging those in the minority from upholding their resistance. Formalising the decision-making procedure may at times also make it more difficult for individual Members to hide behind a lacking consensus when they would not openly vote against a proposal in the decisive situation of a vote.[81] Indeed, the fact that blocking consensus at times merely results in delay and further consultations/negotiations where a negative vote would result in final failure of the proposal reduces the cost of blocking consensus. If it is felt that individual Members are reluctant to join a majority because they do not want to openly forsake a special national interest, one could think of using secret ballots,[82] although this would obviously involve a problematic tension with the objective of transparency. In any event, a system in which it is more difficult for individual members to block decisions would also make it easier to overcome the resistance of domestic special interest groups.

In any event, one should think of more extensive use of Rule 33 of the respective Rules of Procedure of the subordinate Councils, Committees and other bodies of the WTO. This rule prescribes that they refer a matter to the General Council whenever they are unable to reach a decision by consensus. At present, this referral often does not take place when consensus cannot be reached.[83] Such referral could expose the contentious issue to higher visibility and politicise the debate, rather than leaving it at the lower, more technical level where consensus cannot be reached. This would increase the costs of blocking the proposal for the opposing Members and thus in certain cases encourage them to abandon their opposition.

Because of the reasons already mentioned, it would seem particularly valuable if the decision-making mechanism could be reinforced in a way that would reduce the gap in effectiveness between the WTO’s political bodies and its dispute settlement system. For this purpose, the so far unused authoritative interpretation pursuant to Article IX:2 of the WTO Agreement would seem to be an ideal tool for giving Members normative guidance in the context of ambiguous rules, instead of resorting to dispute settlement. The three-quarters majority requirement foreseen in Article IX:2 is already quite demanding such that it would be neither helpful nor justified to assume that authoritative interpretations can only be developed on the basis of consensus.[84]

In any event, it is likely that the most realistic options for procedural improvement would be those that could be introduced through small, seamless steps that gradually have an impact on the practice of the WTO. If these steps are less noticeable, they are more likely to be accepted by those Members who would otherwise resist formal institutional reforms. For achieving such improvements, courageous and visionary Chairpersons of various WTO bodies could play a pivotal role, and so could the Secretariat, by fulfilling its duty of impartially advising the chairs on procedural matters. In this context, it is worth recalling the Cotton incident in the DSB where several Members seemed ready to consider the applicability of voting, but the Secretariat was not.[85]

Further, in terms of the making of new trade rules, where not all Members are ready to sign up for new rules, but a critical mass is, those Members could resort to techniques such as those used for the telecommunications reference paper, financial services, the Information Technology Agreement, or plurilateral agreements,[86]and the non-participating Members should allow this to happen.

Although this has not been the focus of this paper, it should finally be pointed out that at a less formal level, the decision-making processes at the WTO could be facilitated through the reintroduction of a high-level steering group of senior capital-based trade officials.[87] Such a group, the Consultative Group of 18 existed in the GATT between 1975 and 1988. It would of course not be easy to make such a body acceptable to all Members by finding a balance between inclusiveness, flexibility and efficiency. Nevertheless, it may well be worth another effort if a system can be found that ensures rough representation of the various regions and interests in a system of rotation that gives even small Members a chance to be part of that group at some point in time.[88]