gatt et conseil du commerce des marchandises
Le GATT de 1947 et la tâche épuisante de sa signature
Les négociations de l'Accord général sur les tarifs douaniers et le commerce (GATT) se sont achevées officiellement le 30 octobre 1947.
Ce texte est fondé sur des recherches effectuées par Roy Santana, Conseiller à l'OMC. Il ne s'agit pas d'une publication officielle de l'OMC.
To mark the occasion, UN press release No. 469 proudly announced that:
"there is no parallel to this achievement in any previous trade negotiations, all of which have been on a more limited scale. The completion of such a large number of simultaneous negotiations of such broad scope in a little over six months is in itself a remarkable feat. Since April 1947, when negotiations were opened, nearly 1,000 scheduled meetings between the representatives of the countries concerned have taken place in Geneva. In addition, there were continuous discussions of a less formal character between the delegations concerned. … During the time when both tariff negotiations and Charter discussions were proceeding, the delegations and their staff numbered approximately 760 persons."
How did they manage to pull it off?
While the bitter differences on how to handle tariff preferences were still being addressed by Will Clayton, the lead US negotiator, and Sir Richard Stafford Cripps, the lead UK negotiator, and before the breakthrough of the "Brown-Helmore proposals", GATT negotiators had to face a number of legal and practical issues in order to conclude the agreement. This was particularly challenging because they had to do so while simultaneously being involved in a larger negotiation of the "Preparatory Committee of the United Nations Conference on Trade and Employment" (Prepcom), which was at the time drafting the "Havana Charter for an International Trade Organization" (Havana Charter) that would create the International Trade Organization (ITO). In other words, the Havana Charter and the GATT negotiations were closely intertwined but separate negotiating processes.
In 1947, and a year after the International Trade Organization (ITO) negotiations had been launched, a group of negotiators was concerned by the amount of time that would be necessary to conclude the Havana Charter and implement the results of the tariff negotiations. To this end, they devised a relatively simple approach to avoid the undue delay of the tariff concessions:
1. Countries would launch a series of bilateral tariff negotiations in April 1947. The best treatment offered to any party in those negotiations would be extended to all the other parties through the tariff schedules. Unfortunately, having hundreds of pages of tariff schedules would not be enough: in the absence of other rules, they could be easily nullified or impaired through non-tariff measures. A basic framework of rules was definitively needed, but where could they come from?
2. Since most of the work of the Havana Charter had focused precisely on developing trade rules and tackling a number of non-tariff measures, negotiators opted for the easy way out and decided to copy and slightly adjust those draft provisions. This would have the benefit of avoiding lengthy discussions and allowing for the quick implementation of the tariff concessions. While a big part of the text could be borrowed from the Havana Charter, it would still be necessary to draft some provisions to conclude a self-contained General Agreement.
3. The GATT would be a temporary vehicle to quickly obtain results on the tariff front. The larger rule-making and institutional discussions of the Havana Charter would continue and these were meant to prevail. To this end, once the negotiations on the "Havana Charter" had been concluded in 1948, and its provisions had entered into force, that important "T" (i.e. the tariff concessions) would then be transferred from the General Agreement and absorbed by the institutional framework of the ITO.
The men for the job
Now, how best to wrap these ideas and develop a General Agreement on Tariffs and Trade? Eric Wyndham White, Director of the Division on International Trade and Balance of Payments of the United Nations, was in charge of overseeing the larger work of the Prepcom, as well as the more limited GATT negotiations. However, given his multiple responsibilities, the task fell on the shoulders of his 29-year-old deputy, Julio Lacarte Muró, who would be assisted by Alan Renouf, legal counsel at the UN Secretariat.
A legal conundrum
As early as May 1947, negotiators were concerned by the considerable delays that could arise from the domestic legal procedures necessary to ratify an international agreement, and were eager to find ways to expedite the implementation of the tariff concessions. Negotiators had the view that "it would not be practicable unduly to defer application of the tariff concessions. In addition, it is desirable that early proof should be given to the world of the benefits accruing from the present negotiations." (see p. 8 of E/PC/T/135) With this goal in mind, diplomats began exploring "the possibility of arranging for the Agreement to be applied provisionally, in anticipation of its definitive entry into force". They decided to survey whether there was a way of provisionally applying the tariff concessions without having to obtain parliamentary approval and having to deposit an instrument of acceptance.
With a view to studying this question, Lacarte and Renouf circulated a questionnaire on 18 June 1947 (E/PC/T/100) asking participants what internal procedures would be required to bring the GATT into force and, in particular, to implement the tariff concessions at the national level. Most countries replied that both elements would require ratification by the national parliament and preparation of an instrument of acceptance, so the procedures could take several months after concluding the deal (see E/PC/T/135, p. 8).
Only a handful of countries, including Chile and the Customs Union of Syria and Lebanon, were in a position to apply the tariff concessions without parliamentary approval. Internal records suggest that the Netherlands questioned the overall approach, as they were dubious about submitting the GATT for approval to parliament while, at the same time, the Prepcom kept discussing very similar provisions for the Havana Charter. Why would the same or similar provisions have to be submitted twice to parliament as separate agreements? However, the Dutch changed their mind and did not block the approach.
While the official records reflect the hard legal questions and choices that had to be made, the internal records provide a unique window into the practical challenges faced by Lacarte and Renouf. These included "teleprinter messages" exchanged between the UN offices at Geneva and Lake Success, NY, as well as writing internal memoranda and other informal records.
One of the very practical considerations was whether the "U.N. seal" could be used for the signature of the General Agreement, which ended up being considered inappropriate as "the UN as an organization is not a party to it" (teleprinter message No. 27 of 17 September 1947). Another teleprinter message from the same date warned that the Spanish translation of the draft Havana Charter would take at least 10 more weeks if given priority, which meant that only the English and French versions could be copied. Spanish would need to be dropped for the GATT to conclude.
Can I sign? Sign what?
A more complicated issue had to do with the "credentials" of those who would be signing the General Agreement, which involved the question of whether the representatives in Geneva had to demonstrate that they had sufficient authority to sign on behalf of their countries. After initial discussions that began in June 1947, and based on the work by Lacarte and Renouf, instructions concerning credentials were circulated on 9 October 2017 (E/PC/T/237) and a subcommittee on credentials created a few days later (E/PC/T/257). This committee would ascertain whether credentials were in the proper form on a case by case basis.
Lacarte and Renouf spent a considerable amount of time chasing delegations and gathering information. At one point, a troubled Renouf questioned the approach as he considered that many of the delegates had insufficient credentials to sign a General Agreement. However, one of the teleprinter messages clarified that this initial concern could be overcome if, instead, they signed a "final act", which is a formal summary of the proceedings of a diplomatic conference. (See here). In that scenario, the representatives would have sufficient credentials, as the assumption would be that the authority to sign derived from the general authority that had been given to delegations when they had been accredited to the Prepcom. On that basis, it appeared reasonable "that heads of delegations could delegate responsibility to members of his delegation in course of ordinary committee business without securing specific authorization from home" (teleprinter message of 18 September 1947).
The discussion on credentials also led to a peculiar problem arising from important geopolitical changes that were taking place in parallel: the partition of British India. The Dominion of India had thus far been represented by a single delegation, but there were plans to adopt an "Indian Independence Bill" on 15 August 1947, which would result in the succession of that government by the governments of the Dominion of India and the Dominion of Pakistan.
Ironically, Sir Stafford Cripps, the lead UK negotiator who was clashing with Will Clayton, had in 1946 been a major player of this political event through his participation in the "Cabinet Mission Plan" that eventually led to India's partition. By July 1947, Lacarte and Renouf were asked to clarify whether a representative of Pakistan would also need to sign the Final Act. In a letter dated 28 July 1947, Lacarte informed Mr Banerji from the Indian delegation that, in Renouf's opinion, it was possible for Pakistan to sign. However, Renouf felt that "it would be preferable, although not absolutely essential, that the Agreement be signed on behalf of India and Pakistan by different individuals each possessing and producing the necessary credentials from the two governments concerned."
By August 1947, trade diplomats adopted key decisions for the future of the General Agreement. Firstly, they agreed that it would not be necessary to submit the GATT for approval to the larger Economic and Social Council, effectively decoupling it from the Havana Charter negotiations.
Secondly, rather than signing the General Agreement itself, a "final act" would record in brief terms what had taken place, and would attach the text of the GATT and the tariff schedules. This would effectively prevent further changes to the texts and the tariff concessions that had been exchanged.
Thirdly, the elements concerning the provisional application would be removed from the GATT itself and moved into a specially designed Protocol of Signature (see E/PC/T/135, p. 9). Unfortunately, there were also issues on that front. Following a series of discussions, it became apparent that only eight "key" countries would be in a position to sign the Protocol and six other countries would only be able to do it later on (Brazil, Cuba, New Zealand, Norway, South Africa, and Southern Rhodesia).
23 signed but six signatures were of particular importance
The efforts by Lacarte and Renouf finally paid off on 30 October 1947, when the Final Act of the General Agreement on Tariffs and Trade was signed by its 23 contracting parties, with English and French as the only authentic languages. On the same day, eight of those 23 countries were meant to sign the so-called "Protocol of Provisional Application" (PPA), thereby committing to provisionally apply the GATT and its tariff concessions: Australia, Belgium, Canada, France, Luxembourg, the Netherlands, the United Kingdom and the United States.
However, Press release No. 479 shows that only six key countries were able to sign that day, as Australia and France had to be given until 15 November to do it. The six key signatures were those of P. A. Forthomme (Belgium), L. D. Wilgress (Canada), the Grand-Duchy of Luxembourg (J. Storm), the Netherlands (A. B. Speekenbrink), the United Kingdom (T. M. Snow), and the United States (Winthrop. G. Brown).
Through the PPA, these countries undertook "to apply provisionally on and after 1 January 1948: (a) Parts I and III of the General Agreement on Tariffs and Trade, and (b) Part II of that Agreement to the fullest extent not inconsistent with existing legislation." Paragraphs 3 and 4 of the PPA allowed the remaining contracting parties to also provisionally apply the GATT at a later date.
The PPA is also a good reflection of the overall plan explained above. Part I of the GATT encompassed Articles I (MFN) and II (Schedules of concessions) and, most importantly, the Schedules of concessions themselves. Part II reflected the draft provisions that had been copied from the Havana Charter discussions, i.e. Articles III to XXIII. Part III, comprised of Articles XXIV to XXXIV, dealt with the "treaty" aspects of the GATT. (Note: "GATT 1947" does not incorporate a large number of subsequent changes.)
A few months later, other countries signed the PPA including Brazil, Ceylon (Sri Lanka), Cuba, Czechoslovakia, India, Burma (Myanmar), New Zealand, Norway, Pakistan, South Africa and Southern Rhodesia (Zimbabwe). Although Lebanon and Syria also submitted an acceptance, they individually withdrew from the GATT at the beginning of the 1950s. Chile and Uruguay failed to submit an acceptance in 1948, but became contracting parties subsequently, through a "Protocol of the Accession of Signatories of the Final Act". Many other countries quickly joined the GATT ranks either through the acceptance of dedicated protocols or through the figure of succession. For a detailed description of these accessions, see "GATT: Status of Legal Instruments".
Against all odds
Things did not go as originally planned in 1947. Although the Final Act of the Havana Charter was successfully signed in March 1948, it was not ratified by the US Congress so the ITO never came into existence. What is perhaps less known is that the GATT, as an international treaty, never entered into force either.
Unknown to the 23 delegations celebrating 70 years ago, it was the signature of the PPA by those six "key" countries, and not the 23 signatures of the "Final Act" of the GATT that ultimately provided the legal hook that saved the multilateral trading system. The PPA allowed for the provisional application of the GATT for almost 50 years, and was finally terminated on 1 January 1996, one year after the entry into force of the Marrakech Agreement Establishing the WTO (see para. 3 of GATT document L/7583).
The events that took place in 1947 and 1948 forged the professional careers of those involved in the negotiations. Alan Renouf eventually became the Secretary of the Department of Foreign Affairs of Australia, and the Australian ambassador to Washington. Eric Wyndham White and Julio Lacarte remained deeply involved in the early work of the GATT and eventually became legends of the multilateral trading system. And, thanks to the 10 week delay in the translation, it took decades for Spanish to become an official language in the GATT.
Roy Santana, WTO Counsellor
Acknowledgements and disclaimer: I would like to thank Dolores Halloran and János Volkai for their insightful comments and suggestions on an earlier draft, as well as Ali Khilji for pointing out the role of Sir Stafford Cripps with British India's partition. All errors are the author's. For more on the issue, see Douglas A. Irwin, Petros C. Mavroidis, Alan O. Sykes, "The Genesis of the GATT", Cambridge University Press, 29 May 2009.