The Rounds: Geneva 1947 to Tokyo 1979

And it coincided with the arrival of development as a subject. Part IV of GATT, Articles XXXVI to XXXVIII, was added in 1965 and entered into force in 1966. It commits developed contracting parties to accord high priority to reducing barriers on products of export interest to less developed parties, and it introduces non-reciprocity: developed parties do not expect reciprocity for concessions made to less developed parties in trade negotiations. Its language is hortatory, using best endeavours rather than obligation, and that weakness is the standing complaint about it.

The Tokyo Round, and the problem it created

Its subject was the non-tariff barrier, because by the 1970s tariffs on industrial goods were low enough that the real obstacles were standards, subsidies, customs procedures and government purchasing.

It produced nine separate agreements, on subsidies and countervailing measures, anti-dumping, customs valuation, import licensing, technical barriers to trade, government procurement, bovine meat, dairy products and civil aircraft.

And every one of them was optional. A contracting party accepted the codes it chose. Two consequences followed. The obligations owed by any pair of parties depended on which codes each had signed, which is legally untidy and practically unpredictable. And most developing countries signed few or none, so the disciplines that mattered most to their exports did not protect them.

That defect has a name in the literature, GATT a la carte, and a cure in 1994, the single undertaking. Annexes 1, 2 and 3 of the Marrakesh Agreement are integral parts of it and binding on all members, by Article II:2, and the Tokyo Round codes were either folded into those Annexes or terminated.

A worked example of what a round actually does

Take a single concession and follow it. In a linear cut round, State A agrees to reduce its bound tariff on woven cotton fabric from twenty per cent to twelve. That concession is recorded in A's own Schedule, annexed to the Agreement, and by Article II:1(b) A may not thereafter apply ordinary customs duties on that item in excess of twelve per cent.

A negotiated the reduction with State B, its principal supplier. But by Article I:1, the twelve per cent rate must be accorded immediately and unconditionally to the like product originating in every other contracting party. State C, which never sat at the table, gets the same rate.

And if A later wants the rate back, it cannot simply raise it. It must proceed under Article XXVIII, renegotiating with the parties having a principal supplying or substantial interest and offering compensatory adjustment on other products, failing which those parties may withdraw substantially equivalent concessions of their own. That is what makes a binding a legal commitment rather than a policy statement.

What this does NOT mean

It does not mean each round simply cut tariffs further. Annecy, Torquay and the 1956 round achieved comparatively little, and the fourth round coincided with the agricultural waiver that took a whole sector out of the discipline in practice.

It does not mean the codes were worthless. They are the first real disciplines on subsidies, dumping, standards and valuation, and their texts are the drafts from which the Uruguay Round agreements were made. Their defect was optionality, not content.

And it does not mean Part IV solved anything for developing countries. It gave them non-reciprocity and priority in principle, and left them without a remedy. The operative instrument for preferences is not Part IV but the Enabling Clause of 28 November 1979, adopted at the end of the Tokyo Round, which is where the Generalized System of Preferences finds its legal basis.

Quick revision

• Seven rounds: Geneva 1947, Annecy 1949, Torquay 1950 to 1951, Geneva 1955 to 1956, Dillon 1960 to 1962, Kennedy 1964 to 1967, Tokyo 1973 to 1979.
• Technique: request and offer, principal supplier to principal buyer in the early rounds; linear cuts from the Kennedy Round.
• Kennedy: about a third off industrial tariffs, the first Anti-Dumping Code, and Part IV added in 1965 with non-reciprocity in hortatory language.
• Tokyo: about a third again, and nine optional codes on non-tariff measures, the GATT a la carte problem cured in 1994 by the single undertaking in Article II:2 of the Marrakesh Agreement.
• The Enabling Clause, 28 November 1979, is the legal basis of the Generalized System of Preferences, not Part IV.
• A concession lives in the member's own Schedule, is generalised by Article I, and can be taken back only through Article XXVIII renegotiation with compensation.

Test yourself

1. Why did GATT liberalise trade through rounds rather than continuously? Because a round solves three problems at once. It packages concessions, so that a government conceding market access in one sector can point its domestic politics at the access it has gained in another, which is what makes liberalisation survivable at home. It exploits the most favoured nation obligation in Article I, since a bargain struck between two principal traders must be extended immediately and unconditionally to every other party, so bilateral effort produces multilateral results. And it imposes a deadline, without which trade negotiations do not conclude, since there is always an advantage in waiting for the other side to improve its offer. The method also has a distributional consequence that the developing contracting parties complained of throughout: request and offer bargaining directs the negotiating effort to the products that the largest traders care about, and the free extension of the result to everyone else is worth less than it appears when the products chosen are not the ones you export.

2. What did the Kennedy and Tokyo Rounds add that the earlier rounds did not? The Kennedy Round of 1964 to 1967 added a technique and a subject. The technique was the linear cut, an agreed across the board percentage reduction with a negotiated exceptions list, which replaced product by product bargaining and produced a reduction of about one third in industrial tariffs in three years. The subject was the non-tariff measure, in the shape of the first Anti-Dumping Code elaborating Article VI, and alongside it Part IV on trade and development was added to the Agreement in 1965. The Tokyo Round of 1973 to 1979 made the non-tariff measure the main business, producing nine agreements on subsidies and countervailing measures, anti-dumping, customs valuation, import licensing, technical barriers, government procurement, civil aircraft and two commodity arrangements. Those texts are the drafts from which the Uruguay Round agreements were made, and the Enabling Clause, which gave the Generalized System of Preferences its legal basis, was adopted at the same time.

3. What was the defect of the Tokyo Round codes and how was it cured? Their defect was that they were optional. Each code was a separate agreement which a contracting party could accept or decline, so the obligations owed between any two parties depended on which instruments each had signed, and a party could take the benefit of the tariff bindings while standing outside the disciplines on subsidies or standards. The arrangement was nicknamed GATT a la carte. Its practical effect fell hardest on developing countries, most of which signed few codes and therefore could not invoke them against those that had. The cure was the single undertaking of the Uruguay Round, expressed in Article II:2 of the Marrakesh Agreement: the agreements in Annexes 1, 2 and 3 are integral parts of the Agreement and binding on all members, so accession is to the whole package. Only the Annex 4 plurilaterals remain optional, and by Article II:3 they create neither rights nor obligations for members that have not accepted them.