Skip to main content Accessibility help
Hostname: page-component-5c569c448b-nqqt6 Total loading time: 0.382 Render date: 2022-07-06T12:35:25.211Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true } hasContentIssue true


Published online by Cambridge University Press:  14 April 2008

World Bank and CEPR
University of Michigan
Rights & Permissions[Opens in a new window]


Review Article
Copyright © Bernard Hoekman and Robert Howse 2008


In the early 1970s, the EC was a net importer of sugar, buying some 2.5 million tonnes on the world market. As a result of high intervention prices for sugar – averaging more than double the world price since the mid 1980s, and rising to a multiple of four in 2000 – production expanded so much that the EC became a net exporter. In the late 1990s and early 2000s, the EC's net exports of sugar were in the range of four to five million tonnes. EC policies resulted in the EC becoming a major player on world markets: in 2005, it was the second largest exporter in the world (after Brazil), accounting for 12% of world exports. At the same time, the EC was also the world's fourth largest importer, an idiosyncrasy that reflected the preferential access granted to African, Caribbean, and Pacific (ACP) countries under the ACP Sugar Protocol.

In the Uruguay Round, the EC scheduled export subsidy commitments for sugar of €499 million and 1.2735 million tonnes. In ECSugar, Australia, Brazil, and Thailand argued that the EC provided export subsidies in excess of its WTO export-subsidy commitment levels for sugar, in part through de facto cross-subsidization of exports as a result of guaranteeing high annual intervention (support) prices for a given quantity of EC sugar, and in part as the result of re-exporting an amount of sugar equivalent to what the EC imported from ACP countries on a preferential basis. The EC argued that re-exports of ACP sugar were covered by its export subsidy commitments through a footnote in its schedule, and that the acceptance of the inclusion of this footnote by WTO Members implied that its schedule effectively altered its obligations under the Agreement on Agriculture (AoA). The EC also held that, because WTO Members had not previously raised objections to exports of so-called ‘C sugar’ as being subsidized, the complainants were not acting in good faith and were estopped from invoking dispute proceedings.Footnote 1

We address each of these dimensions of the case in this paper. We start in Section 1 with a brief description of EC policy and some of its likely effects. Section 2 provides an economic and legal analysis of the cross-subsidization argument brought by the complainants. Section 3 turns to the consistency of the EC measures with the EC's WTO obligations, as well as the estoppel defense argument of the EC. Section 4 discusses the finding of the Appellate Body (AB) that the Panel erred in its invocation of the principle of judicial economy to not address claims that aspects of the WTO Agreement on Subsidies and Countervailing Measures (SCM) had been violated. Section 5 presents concluding remarks.

1. EC sugar policy

Subsidies play a major role in the EC sugar industry.Footnote 2 In 1999–2001, the value of gross receipts of sugar producers in the EC was more than double the value of their output measured at world prices: the producer nominal assistance coefficient was 2.1. In this period, total OECD support for sugar was equivalent to about half of global exports ($6.35bn compared to $11.6bn), similar in value to the total exports of sugar of all developing countries ($6.5bn). The EC accounted for 43% of the $6.35bn in OECD support for sugar. Much of this support was provided through very high border protection – around 90% for the EC.

The EC policy regime for sugar complements high intervention or support prices for sugar with production quotas. These are of two types: so-called quota A and quota B. The sum of A and B quotas determines the maximum amount of sugar that may be sold in the EC market in a given year. All excess production must be exported – if not, producers are penalized.Footnote 3 Production of A and B sugar benefits from the high intervention price in the EC. Excess production by an EC producer over and above its A+B quotas does not benefit from the high intervention/support price. In EC-speak, such excess output is called C sugar. Note that there is no physical difference between these various categories of sugar; there is one world price for sugar, be it A, B, or C sugar.

The producer price for A sugar is greater than the producer price received for B sugar. Both are less than the basic intervention or support price as a result of a levy that is used to finance the export subsidies needed to sell excess production on the world market. Thus, producer prices of A and B sugar are endogenous; they are a function of how much of B is exported if the sum of A+B quotas exceeds total EC consumption at the (fixed) intervention price for sugar. The levy on A sugar is 2%, i.e. the producer price received is 2% below the ‘basic’ support price confronting buyers. In the case of B output, the levy varies, ranging up to a maximum of 37.5%.Footnote 4

In addition to (part of) the B quota, the EC also exports an amount of sugar that is equal to what it imports from ACP countries under its preferential access program (the Cotonou Convention). The total EC import commitment vis-à-vis the ACP is some 1.3 million tonnes, although in some years this has been expanded under a special (additional) preference program. As a result, the maximum amount imported from the ACP has ranged up to 1.6 million tonnes. Given that EC production exceeds consumption at the intervention price, in effect all the ACP sugar is ‘re-sold’ on the world market. Thus, the effects of the ACP sugar protocol imports on the EC market are ‘sterilized’ by exporting the amount imported.Footnote 5 As the ACP sugar is bought at the intervention price, the export sales incur a significant loss, which is absorbed by the EC budget (taxpayers). These costs are clearly export subsidies, and are recognized as such by the Panel and the EC.

The resulting EC-wide market for sugar is illustrated in Figure 1. Of some 18–20 million tonnes of sugar produced per year in the EC, some 15% is exported. In 2001, the EC exported 4.1 million tonnes: the quantity represented by line segment TW in Figure 1. However, the EC had only scheduled 1.2735 million metric tonnes of sugar (line segment TU in Figure 1) as being eligible for export subsidies.Footnote 6

Figure 1.

The excess exports of 2.8 million tonnes were comprised of 1.3 million metric tonnes of ACP sugar (line segment UV) and 1.5 million tonnes of C sugar (line segment VW). Thus, EC exports consist of B-quota sugar (benefiting from scheduled export subsidies financed by the levy on EC quota production) plus the (equivalent of) ACP sugar imports (also benefiting from a direct export subsidy, financed by the EC budget) plus C production. Only the latter does not benefit from direct export subsidies.

Two groups finance the sugar scheme: EC consumers of sugar, who pay two to four times the world price for sugar, and EC taxpayers (in order to re-export ACP sugar). The rest of the world, non-ACP producers that do not benefit from preferential access and confront lower world prices as a result of EC subsidization of exports of sugar, incur losses as a result of the policy. Two groups gain from the policy: EC and ACP producers. As illustrated in Figure 2, the rents EC producers obtain are determined by the high producer prices in the EC: P a for A quota (2% less than the support price), and P b for B-quota sugar (up to 37.5% less than the basic price, illustrated by area ‘b’ in Figure 2). The magnitude of the rents depends on the cost functions of producers: the more efficient the producer, the greater the rents. This also applies to ACP sugar protocol countries that have export quotas into the EC market. Low-cost producers – for simplicity, say those that can sell at the world price and cover their costs – will get a rent equal to the shaded area in Figure 2: areas A+B+C as a result of the A sugar quota for which they get P a; area D as a result of the B quota (for which they get P b). A high-cost producer – with marginal cost schedule MC h in Figure 2 – will only obtain a rent equal to area A.Footnote 7

Figure 2.

A key question in the dispute revolved around what producers do with the rents; in particular, whether they use them to cross-subsidize production and exports of C sugar. A second major issue in the case concerned the total volume of exports of sugar, which the complainants argued exceeded the EC's schedule of export subsidy commitments. The following two sections discuss these two dimensions of the dispute.

2. Cross-subsidization of C sugar – legal and economic analysis

Article 9.1(c) of the AoA defines export subsidies subject to reduction commitments as including:

[P]ayments on the export of an agricultural product that are financed by virtue of governmental action, whether or not a charge on the public account is involved, including payments that are financed from the proceeds of a levy imposed on the agricultural product concerned or on an agricultural product from which the exported product is derived[.]

The Panel found that the measures complained of fell within this provision, because: (1) Producers of C sugar were able to purchase C sugar beet at below cost; (2) the EC required C sugar to be exported; (3) the below-cost sales were financed through sales of A and B sugar at above-market prices, which were determined at least in part by EC regulations. It is clear that the legal theory here is one of cross-subsidization: above-normal profits from the sale of one product subsidize the export of another. The Panel argued that the sugar regime in the EC resulted in cross-subsidization of all EC sugar output, including C production that does not benefit from direct subsidies on export and is sold at world prices. The basic argument of the Panel is that rents on A and B quota sugar can be used to cover below-cost sales of C output. While in general such a cross-subsidy need not be an export subsidy, under EC policy, if C production is not exported (or used under the carry-forward provision), it is subject to high financial penalties – thus there is a clear incentive to export C sugar.

From an economic perspective, a key question is why farmers would produce C sugar given that it does not benefit from direct support (is sold at the world price). One possibility is fixed costs of production (scale economies): more output lowers average unit costs, perhaps by enough to make total production profitable if it goes beyond the level of A and B output levels. Another possibility is that some EC producers may be efficient and able to sell at the world price without incurring losses. A third possibility is uncertainty – farmers may plant more than is needed to generate the A+B level of output as a form of insurance: to make sure that they will be able to meet long-term contracts they have with refiners. As these are very profitable, overplanting ensures that if harvests are bad, farmers will still be able to fully satisfy their contracts, and thus capture all of the rents they can make on quota sugar. If harvests are good, however, there will be excess production – i.e., C sugar.

If the world price for C sugar is such that the resulting total average cost of production falls because of scale economies – i.e., there are fixed costs/fixed factors – the rents derived from quota (A and B) production can be used to cross-subsidize C production and enhance profits. Given a level of output B, defined as the maximum amount on which firms can get the high support price (more generally the share of output on which producers get a subsidy/rent), as long as the world price exceeds the marginal costs of producing this level of output, farmers will have an incentive to produce more. They will produce up to the point where marginal revenue (determined by the world price) equals marginal cost of production. This optimal level of output will exceed the quota level of sales. In general, De Gorter, Just, and Kropp (Reference De Gorter, David and Jaclyn2008) show that cross-subsidization is possible for a variety of permutations of production costs, world price levels and support prices implied by a quota level B. A number of possible situations where cross-subsidization can occur are illustrated in Figures 3–6. These include situations where firms make losses at the quota level of output and situations where the world price is less than the average variable costs of production.Footnote 8

Figure 3. Positive profits at B, cross-subsidization to Q* (P*>AVC*)

Figure 4. Positive profits at B, possible cross-subsidization to Q* (P*<AVC*)

Figure 5. Negative profits at B, possible cross-subsidization to Q* (P*>AVC*)

Figure 6. Negative profits at B, possible cross-subsidization to Q* (P*<AVC*)

De Gorter, Just, and Kropp (Reference De Gorter, David and Jaclyn2008) note that, even if marginal-output decisions are determined by the world price, not the quota (support) price, the existence of the rents on quota sugar may deter exit and thus result in output that is too high. There may also be firms that would be unprofitable producing at quota (support) level B but that become profitable by increasing their output at the margin because of the support that is offered for output up to level B (see also Figures 3–6). Thus, total output is likely to exceed what would be observed under free trade because there are producers in the market only because of the subsidy.

Only if there are producers that can cover their marginal costs at the world price without subsidies is there no cross-subsidization. In such cases, the EC policy will simply be a redistributive one: firms are allowed to capture rents, but they would be in the market anyway. A problem in assessing the validity of this possibility is that there is little information on costs of production, so it is not known to what extent there are such firms in the EC. However, the likeliest presumption is that there will only be few such producers in the EC.

Turning to the insurance motive possibility for C production and exports, if there is rational overproduction to deal with uncertainty ex ante, Gohin and Bureau (Reference Gohin and Jean-Christophe2006) show that the cross-subsidy conclusion still applies: the EC quota/support system acts as a cross-subsidy in such situations because the likelihood of ‘overproduction’ is a positive function of the difference between the EC intervention and the world price. However, they also point out that the carry-forward provisions of the sugar regime attenuate the incentive to overproduce, as they allow for quota rights to be transferred to the next year.

If a more long-run view is taken, the cross-subsidization argument based on fixed costs becomes less compelling. Producers should adjust production factors so that these are optimal for producing at level B. Gohin and Bureau (Reference Gohin and Jean-Christophe2006) argue that, as the EC regime has been in place for decades, such adjustments should have been made in the past. Moreover, they note that it is not clear that there actually are significant scale economies in beet production (fixed costs), although the possibility of such costs and long-lived assets is more likely for processors/refiners. These considerations make the De Gorter et al. finding that cross-subsidization is not limited to instances where world prices are less than average total costs of production more important in bolstering the economic argument for cross subsidization. Of particular importance here is the exit-deterring influence of policies that result in de facto cross-subsidization of exports.

Thus, it is indeed possible, based on the known facts, that cross-subsidization would be an effect of the EC regime. However, it is also conceptually possible that sugar producers could decide to do otherwise with the above-normal profits from A and B sales than subsidize sales of C beet.Footnote 9 As a matter of law then, is the requirement of governmental action satisfied if the governmental conduct in question will not necessarily induce the subsidization complained of? We are inclined to answer in the affirmative and thus to support the findings of the Panel and the Appellate Body. Art. 9.1(c) requires only that ‘payments on the export of an agricultural product’ be ‘financed by virtue of governmental action’ (emphasis added). In other words, all that is necessary is for the government to have done something ‘by virtue of which’ the payments are financially enabled; it is not required to show that the government has mandated or caused the payments to be actually made, much less that the governmental action itself is in the form of a ‘payment’, however so defined. Thus, whether or not such payments are made might still be at the discretion of nongovernmental actors; if they are made, and it is governmental action that results in the financial capacity to make them, they are covered by 9.1(c). It was not in dispute that the high prices for A and B sugar and sugar beet were a consequence, in part, if not large part, of the EC regulatory regime. The question then becomes, as a matter of proof, what the complainant must show in order to establish that below-cost sales of C beet were enabled by the profits generated through these high prices. Should there be a rebuttable presumption, based on standard economic axioms, that, absent above-normal profits on one product, it is simply not viable to sell another product at below-cost? This would leave it for the defendant to explain how, despite the standard economic axioms, in the particular case at hand, the firm would have been able to finance the below-cost sales in the absence of the above-normal profits due to government-influenced or government-determined above-market prices for other products. Certainly, the EC does not seem to have offered a very clear explanation of how the behavior in question with respect to C might have occurred anyhow in the absence of above-normal A and B profits.


A concern may be raised that, in opening the door to ‘cross-subsidization’ claims, the Appellate Body has considerably expanded the scope for Members to challenge subsidies in WTO law, as well, arguably, to legally countervail them. It is our view that in fact there will be limited scope for such claims outside the context of 9.1(c) of the AoA. The SCM Agreement requires that, for a measure to be considered a subsidy, there must be a ‘financial contribution by government’ (emphasis added), not merely that a payment (by anyone) be financially enabled ‘by virtue of governmental action’. The AoA is clearly addressed to dealing with the pervasive distortion of competition in agricultural markets through the combined and interactive effects of a wide variety of government policies: levies, export subsidies, quotas and price support mechanisms. It is thus understandable that ‘cross-subsidization’ would be disciplined by such a comprehensive agreement. The SCM Agreement is more limited, on the other hand: unlike the case with the EC's own internal law on state aids, the WTO rules on subsidies exist in the absence of agreement among WTO Members on antitrust or anticompetition law principles; indeed whether such principles should be part of the WTO system at all is intensely controversial. The definition of ‘financial contribution’ in the SCM Agreement makes clear that there is only a very narrow range of circumstances where a ‘financial contribution’ by government would exist by virtue of payments through a nongovernmental entity: SCM 1 provides that a financial contribution exists where ‘government makes payments to a funding mechanism, or entrusts or directs a private body to carry out one or more of the type of functions illustrated in (i) to (iii)’. These functions are: direct payments or transfers, forgoing revenue otherwise due, and provision of goods and services other than general infrastructure and purchase of goods.

In the case of provision of goods and services, a benefit is deemed to be conferred only if these are provided at below-market prices. Merely enabling a private entity to provide goods or services at below market prices would seem to fall short of ‘entrusting’ or ‘directing’ such an entity to do so. In the US–Export Restraint case, Canada challenged provisions of US countervailing duty law and regulation that included export restraints as among the kinds of countervailable subsidies. The context of this complaint was the longstanding dispute between Canada and the US over softwood lumber; US producers wanted export restraints on Canadian logs to be treated as subsidies to Canadian lumber producers, on the grounds that the effect of these restraints was to increase supply and reduce the price of such logs in the Canadian market, thereby reducing the input costs of Canadian lumber producers. Unlike in EC–Sugar, in this case the government was altering the conditions of the market such as effectively to compel, and not merely to enable, sales at lower prices. Nevertheless, the Panel held:

Government entrustment or direction is very different from the situation in which the government intervenes in the market in some way, which may or may not have a particular result simply based on the given factual circumstances and the exercise of free choice by the actors in that market. Indeed, governments intervene in markets in various ways, and with various policy or profit objectives, and these interventions might have various results, including results that are not intended by, or that are even undesirable for, the government.' (para. 8.31)

According to the Panel, there must be an explicit act of delegation or command to the private body to engage in the conduct in question for a situation of entrustment or direction to exist. In our view, this interpretation of ‘financial contribution’ provides very limited opportunities for making ‘cross-subsidization’ arguments under the SCM Agreement.

At the same time, cross-subsidization must not be confused with the issue of ‘pass-through’ subsidies, that is the question of whether by providing an actual financial contribution in respect of one product, a government confers a benefit and creates a subsidy on another. Thus, again in the softwood lumber dispute, most of the measures besides the export restraints that were the basis for US agency findings of subsidization, conferred a financial contribution with respect to the production of logs, not lumber (fees for the harvesting of timber at arguably below-market rates). The claim was that by subsidizing the production of logs, the government was reducing the price of the crucial input in the production of lumber and thereby conferring a ‘benefit’ or competitive advantage on Canadian lumber producers. The approach of the Appellate Body to situations where a financial contribution is conferred on one industry but the benefit and thus the subsidy is with respect to another is as follows:

The phrase ‘subsid[ies] bestowed … indirectly’, as used in Article VI:3, implies that financial contributions by the government to the production of inputs used in manufacturing products subject to an investigation are not, in principle, excluded from the amount of subsidies that may be through the imposition of countervailing duties on the processed product. Where the producer of the input is not the same entity as the producer of the processed product, it cannot be presumed, however, that the subsidy bestowed on the input passes through to the processed product. In such case, it is necessary to analyze to what extent subsidies on inputs may be included in the determination of the total amount of subsidies bestowed upon processed products. For it is only the subsidies determined to have been granted upon the processed products that may be offset by levying countervailing duties on those products. In our view, it would not be possible to determine whether countervailing duties levied on the processed product are in excess of the amount of the total subsidy accruing to that product, without establishing whether, and in what amount, subsidies bestowed on the producer of the input flowed through, downstream, to the producer of the product processed from that input. Because Article VI:3 permits offsetting, through countervailing duties, no more than the ‘subsidy determined to have been granted … directly or indirectly, on the manufacture [or] production … of such product’, it follows that Members must not impose duties to offset an amount of the input subsidy that has not passed through to the countervailed processed products. It is only the amount by which an indirect subsidy granted to producers of inputs flows through to the processed product, together with the amount of subsidy bestowed directly on producers of the processed product, that may be offset through the imposition of countervailing duties. (paras. 140–141)

3. Consistency of the measures complained of with the EC's schedule under the Agreement on Agriculture

Interpretation of Footnote 1 of the EC's schedule

The EC argued that, even if the measures complained of constituted export subsidies within the meaning of the Agreement on Agriculture, the payments in question fell within the ceilings to which the EC had bound itself in its schedule pursuant to the AoA. The AoA requires WTO Members to make commitments to reduce export subsidies on agricultural products (Article 9.1). Article 3.3 provides that, if an export subsidy is not listed in the Member's schedule, it may not be granted. Article 9.2(b)(iv) of the AoA stipulates that a Member's budgetary outlays for export subsidies and the quantities benefiting from such subsidies, at the conclusion of the implementation period shall be no greater than 64% and 79% of the 1986–1990 base period levels, respectively. For developing country Members these percentages shall be 76% and 86%, respectively. Thus, if a WTO Member wishes to continue to grant an export subsidy, (1) it must be scheduled; (2) it must be subject to reduction commitments; and (3) these commitments must be such that the percentage reductions in budgetary outlays and quantities specified in 9.2(b)(iv) of the AoA are achieved by the end of the implementation period.

The EC did in fact schedule reduction commitments with respect to subsidized exports on sugar. These are contained in Section II, Part IV of the EC Schedule and are summarized by the Appellate Body as follows:

(i) the ‘base quantity level’ (the average of the quantity of subsidized exports of sugar during the base period 1986–1990) was 1,612,000 tonnes, and this quantity level would be progressively reduced to 1,273,500 tonnes in the year 2000 as the ‘final quantity commitment level’ for sugar; and (ii) the ‘base outlay level’ (the average of the budgetary outlay on subsidized exports of sugar during the base period 1986–1990) was €779.9 million, and this budgetary outlay level would be progressively reduced to €499.1 million in the year 2000 as the ‘final [budgetary] outlay commitment level’ for sugar. (para. 159)

The EC did not contest that, if the payments complained were found to constitute export subsidies, by virtue of these payments the EC would have failed to meet the reduction commitments to which it was bound in the above-noted provisions of its Schedule. However, the EC claimed that, by virtue of Footnote 1 to its Schedule, the total ceiling it had bound itself to achieve, in quantity terms, was not 1,273,500 tonnes, but this amount plus a maximum of an additional 1.6 million tonnes. The EC essentially argued that the purpose of Footnote 1 was to allow it to meet its commitments in the body of its schedule while continuing to subsidize exports of sugar of ACP and Indian origin up to 1.6 million tonnes.

Footnote 1 reads as follows: ‘Does not include exports of sugar of ACP and Indian origin on which the Community is not making any reduction commitments. The average of export in the period 1986 to 1990 amounted to 1,6 mio t.’ The EC version of the meaning of this footnote depended on a certain interpretation of both the first and the second sentence. The AB considered each of these interpretations in turn. In the case of the first sentence, the EC claimed that its intent was not to exclude ‘exports of sugar of ACP and Indian origin’ as such from reduction commitments, but only the quantities of exports equivalent to such exports from its scheduled ceiling for reduction commitments, thereby in effect raising the total ceiling by that amount. The EC advanced this interpretation, one assumes, because it was aware that under the AoA there was no basis for excluding a specific subset of export subsidies from reduction commitments.

The basis the EC gave for its nonliteral reading of the first sentence was the negotiating history of the Uruguay Round. As summarized by the AB, the EC's claim was that ‘it was well known to all parties at the time of conclusion of the WTO Agreement that the EC did not grant export refunds only on the re-export of sugar originally of ACP/Indian origin, but granted export refunds for a quantity equivalent to such exports’(para. 176). The EC referred to two documents from the travaux to support its assertion. The first was a March 1992 letter to the negotiating parties in which the EC stated that it ‘ha[d] not included the volume of sugar corresponding to its imports of sugar from ACP countries’ (emphasis added). The second was an Australian memorandum dated 31 January 1994 that described Footnote 1 as covering ‘direct export restitutions (corresponding to [the European Communities’] imports of sugar from ACP countries and India)'.

The AB rejected the EC's interpretation, stating:

Like the Panel, we are not persuaded by these arguments, which rely on the presumed knowledge of other Members of the World Trade Organization (the ‘WTO’) on the export subsidy practices of the European Communities with respect to ACP/India sugar. We note that the Complaining Parties have rebutted the interpretations put forward by the European Communities on the terms, context, and negotiating history of Footnote 1. We also wish to note that the European Communities was unable to clarify at the oral hearing, why, having written the aforementioned letter in March 1992, it did not consider it necessary to use that same language in Footnote 1 in the December 1993 text, or to use language that plainly shows that the exports referred to were those equivalent in volume to the sugar that was imported from ACP countries and India. In any event, we are of the view that the European Communities' submissions do not alter the plain meaning of the first sentence of Footnote 1, so as to make it cover the exports of sugar equivalent in volume to the European Communities' imports of sugar from ACP countries and India. (para. 180)

With respect to the second sentence of the footnote, the EC claimed that the statement ‘The average of export in the period 1986 to 1990 amounted to 1,6 mio t’ amounted to a legal undertaking that the total amount of sugar exempted from the reduction commitments in the body of the EC's schedule would not exceed 1.6 million tonnes. In effect, the EC was saying that this was a cap or limitation that satisfied the requirement of the AoA that all export subsidies be included in overall reduction commitments in Members' schedules.

The AB rejected this interpretation, stating:

The fact that the second sentence makes a specific reference to the average of such exports in the base period does not necessarily lead to an inference that there is a commitment in the sentence to ‘limit’ the quantity of subsidization to that level, particularly when the first sentence says that the European Communities is not making ‘any reduction commitments’. Nor does the practice of the European Communities providing subsidies on exports equivalent to its actual imports from ACP countries and India lead to an inference that that practice flows from a commitment contained in the second sentence of Footnote 1.

Lastly, we see merit in the Panel's reference to the notification practice of the European Communities to the WTO Committee on Agriculture to conclude that this practice does not support the interpretation advanced by the European Communities. If Footnote 1 does contain a commitment on the part of the European Communities with respect to the export subsidies provided by it on exports of sugar equivalent in volume to its actual imports from ACP countries and India or 1.6 million tonnes (albeit such a commitment is only a ‘limitation’ commitment and not a ‘reduction’ commitment), we fail to see why the European Communities did not notify the WTO Committee on Agriculture of the status of its compliance with that commitment throughout the period of implementation of the Agreement on Agriculture. The fact that the European Communities did not do so undermines its interpretation of the second sentence of Footnote 1. (paras. 186–187)


In our view, the AB's rejection of the EC's ‘everybody knows’ reading of Footnote 1 is a correct approach to treaty interpretation under the Vienna Convention on the Law of Treaties (VCLT) and is also desirable from the perspective of fundamental systemic concerns in the WTO legal system, namely internal and external transparency. The VCLT Article 32 permits recourse to the negotiating history only ‘in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31: (a) leaves the meaning ambiguous or obscure; or (b) leads to a result which is manifestly absurd or unreasonable’. There is nothing ‘ambiguous or obscure’ in the language ‘does not include exports of ACP/India Sugar on which the community is not making any reduction commitments’. The words plainly say that the EC is not making any reduction commitments on these particular exports.

The question of whether the plain meaning leads to a result that is ‘absurd or unreasonable’ is somewhat more difficult. It is arguable that the plain meaning is ‘absurd or unreasonable’ as it would lead to the result that the EC would have made a scheduling that violates the unqualified obligation in AoA 9.1 to schedule reduction commitments on export subsidies and in AoA 3.3 not to grant any subsidy that has not been scheduled. The fact is, however, that even if one reads the second sentence as containing a legally binding commitment to limit the quantity subsidized to 1.6 million tonnes, the nonliteral meaning urged by the EC would not save it from the result of noncompliance with the AoA, since a limitation to this amount could not be construed, and indeed the EC was not urging it to be construed, as a commitment to reduction of the amount by the percentage and in the time frame required by the AoA.

In sum, if what is ‘absurd or unreasonable’ in the result of the literal interpretation of the first sentence of the footnote is that it would entail a scheduling by the EC contrary to the AoA, that result could not be avoided through recourse to the negotiating history in the manner suggested by the EC. Thus, under Article 32 of the VCLT, there was no basis for the EC to demand that the AB rely on the negotiating history.

As for the AB's rejection of the view that the second sentence of the Footnote contains a binding legal commitment to limit – albeit not reduce – subsidization to 1.6 million tonnes, the AB's view that a mere statement of the practice of a state party cannot without some further wording be understood as binding in international law is broadly consistent with general international law doctrine. To establish an obligation of customary international law, not only is evidence of state practice required but proof of intent to follow the practice as a matter of international legal obligation (opinio juris). Perhaps even more relevant is the law as it relates to binding unilateral declarations of states. As the adopted panel in the US – S. 301 case noted, a statement by a state should not be taken to give rise to a legal obligation unless strict conditions are met, including a manifest intent to be bound by the statement (US – S. 301 of the Tariff Act, para. 7.118 and accompanying footnote). In this respect, we note that the AB adverted to a significant piece of evidence indicating that the EC had not intended to be bound, namely that the EC had not notified the WTO Agriculture Committee of its compliance with this purported commitment, as it was required to do under the AoA.

More generally, the EC's contention that its Schedule should be interpreted based on insider knowledge of the trade negotiator community about the history of its export subsidization practices is at odds with policies of internal and external transparency that are important to the legitimacy of the WTO. Such knowledge is obviously available only to a limited group of persons and is not documented in any public transparent venue.Footnote 10 At most, those negotiating parties most intimately engaged in negotiations with the EC over export subsidies would have become aware of the practices in question (even assuming that they could infer from awareness of the practices the particular interpretation of the legal commitments in question urged by the EC). Scheduled commitments indicate the terms and conditions on which economic actors may compete in the global marketplace with the scheduling Member; it is important for advancing the goals of the trading system that such terms and conditions be accessible to economic actors and to all interested constituencies, for that matter, and that they be secure and legally predictable (see US – S. 301 of the Tariff Act on the importance of security and predictability to advancing the goals of the WTO system).

Consistency of the EC reduction commitments with the EC's obligations under the Agreement on Agriculture

Having determined that the ordinary meaning of Footnote 1 to the EC schedule was that the EC was purporting to exempt export subsidies on ACP/India sugar from reduction commitments altogether, it is hardly surprising that the AB would then go on to determine that this exemption was inconsistent with the AoA.

First of all, the AB held that reduction commitments must be scheduled both in terms of budgetary outlays for export subsidies and the quantities benefiting from such subsidies; thus, even if the AB had been able to agree with the meaning of Footnote 1 advanced by the EC, Footnote 1 would still have run afoul of the AoA by failing to mention any limitation or reduction of budgetary outlays, instead referring exclusively to quantities.

The EC argued that there was no specific provision of the AoA that required scheduling both in terms of budgetary outlays and quantities subsidized. However, the AB relied on the nature of the underlying obligation to make reduction commitments as set forth in AoA 3.3, which reads as follows:

Subject to the provisions of paragraphs 2(b) and 4 of Article 9, a Member shall not provide export subsidies listed in paragraph 1 of Article 9 in respect of the agricultural products or groups of products specified in Section II of Part IV of its Schedule in excess of the budgetary outlay and quantity commitment levels specified therein and shall not provide such subsidies in respect of any agricultural product not specified in that Section of its Schedule.

Concerning the implications of this obligation, the AB suggested:

In our view, the use of the conjunctive ‘and’, and the corresponding use of the word ‘levels’ in the plural, suggest that the drafters of the Agreement intended that both types of commitments must be specified in a Member's Schedule in respect of any export subsidy listed in Article 9.1. Had the drafters intended that a Member could specify one or the other of the two forms of commitments, they would have chosen the disjunctive ‘or’ and correspondingly used the word ‘level’ in the singular. Given the choice, Members would choose only one or the other type of commitment, but not both, so as to minimize their obligations. Therefore, it appears to us that the drafters intended to ensure that export subsidy commitments are specified in Members' Schedules in terms of both budgetary outlay and quantity commitments, by using the word ‘and’ as well as the word ‘levels’ in the text of Article 3.3. (para. 193)

The AB further noted:

We find contextual support for the above interpretation in Article 9.2(b)(iv) of the Agreement on Agriculture, which provides that Member's budgetary outlays for export subsidies and the quantities benefiting from such subsidies, at the conclusion of the implementation period, are no greater than 64 per cent and 79 per cent of the 1986–1990 base period levels, respectively. For developing country Members these percentages shall be 76 and 86 per cent, respectively. This provision prescribes the export subsidy commitment levels to be reached at the conclusion of the implementation period (and to be maintained thereafter), and those commitment levels are expressed in terms of both budgetary outlays and quantities. We do not see how a Member could comply with Article 9.2(b)(iv), or for that matter Article 9.2(a), without having specified its export subsidy commitments in terms of both budgetary outlays and quantities. We also consider it significant that both Article 9.2(b)(iii) and Article 9.2(b)(iv) use the expression ‘budgetary outlays for export subsidies and the quantities benefiting from such subsidies’. (emphasis added) This shows the drafters' recognition of the need to address the budgetary outlays and quantities together. (para. 194)

More fundamentally, however, the AB went on to affirm the obvious respect in which Footnote 1 was inconsistent with the substantive obligations of the Agreement on Agriculture: Art 9.1 simply does not provide any ‘carve out’ or safe haven for particular subsidies in respect of the obligation to make reduction commitments. The AB noted:

The chapeau of Article 9.1 says that the subsidies listed in that Article ‘are subject to reduction commitments under this Agreement’. The export subsidies given to ACP/India equivalent sugar, which admittedly fall within the ambit of Article 9.1(a), are therefore subject to reduction commitments. Furthermore, as noted by the Panel, the provisions of Article 9.2(b)(iv) apply to Members that take advantage of the flexibility provisions of Article 9.2(b). Article 9.2(b)(iv) specifies the reduction levels to be achieved at the conclusion of the implementation period with respect to both budgetary outlays and quantities. The provisions of Article 9.2(b)(iv) lend contextual support to the view that export subsidies listed in Article 9.1 are subject to reduction commitments. We further note that Article 9.2(a)(i) and (ii) also make it clear that both budgetary outlay and quantity commitments specified in a Member's Schedule for each year of the implementation period are ‘reduction’ commitments. It follows that the export subsidies provided to ACP/India equivalent sugar are subject to reduction commitments in terms of Article 9.1 of the Agreement on Agriculture. (para. 206)

Finally, the AB addressed an argument of the ACP countries, third parties in the litigation, that in certain circumstances the AoA permits limitations of export subsidies in lieu of reduction commitments (this argument of course assumes the validity of the EC reading of Footnote 1, namely that this footnote articulates a legally binding limitation on export subsidies, if not a literal reduction commitment; since the AB had already rejected such a reading of Footnote 1, its remarks on the ACP argument are, strictly speaking, obiter dicta). The AB opined:

The ACP CountriesFootnote 11 have argued that ‘limiting subsidization’, without reducing budgetary outlays or quantity, is permissible under Article 3.1 of the Agreement on Agriculture, which says that ‘[t]he domestic support and export subsidy commitments in Part IV of each Member's Schedule constitute commitments limiting subsidization’ (emphasis added). According to the ACP Countries, ‘the basic obligations of the Agreement with respect to a scheduled product such as sugar are found in Articles 3.3 and 8. These paramount provisions are framed around and are directly linked to what is specified by the respective member limiting subsidization in the sense of requirements in Article 3.1 of the Agreement.’ In response to questioning at the oral hearing, the ACP Countries also emphasized that Article 3.1 permits ‘limiting’ subsidization, and only what is specified in a Member's Schedule governs its obligation with respect to subsidization.

We are not persuaded by this argument. We do not see Article 3.1 as permitting a Member to limit subsidization to whatever commitment it chooses to specify in its Schedule without regard to Members' obligations under the Agreement on Agriculture. Rather, with respect to export subsidy commitments, we see Article 3.1 as requiring a Member to limit its subsidization to the budgetary outlay and quantity reduction commitments specified in its Schedule in accordance with the provisions of the Agreement on Agriculture. This is also clear from the provisions of Article 9.2(a) of the Agreement, which requires adherence by a Member in each year of the implementation period to the budgetary outlay and quantity ‘reduction commitments’, as specified in the Member's Schedule. (paras. 208–209)


The AoA clearly requires WTO Members to achieve the result of reduction commitments both in terms of budgetary outlay and quantity of exports subsidies: inferring from this obviously intended result the requirement that schedules be structured in those terms is, however, not a self-evident interpretive move. One reading of the Chile–Price Band and Argentina–Footwear AB Reports is that, absent an explicit textual requirement, a Member should be able to express and apply its liberalization commitments as it pleases, provided that it achieves the result of operating within the ceilings that it is legally bound to conform to. In these cases, the defending WTO Members had used formulas or methodologies for determining tariffs in customs administration that deviated from the ad valorem or specific methods used for scheduling; the approach of the AB in these cases was to determine whether, regardless of the formula or methodology used, the result was the possibility of protection being imposed in excess of the underlying legal obligation not to exceed the scheduled ad valorem MFN-bound tariff rate. From this perspective, arguably it would not be necessary to require that both budgetary outlay and quantity be scheduled, as long as within the specified time period EC policies were adjusted so as to ensure the required reductions in both actually occurred.

Did the AB therefore take an unjustified interpretative leap in reading in a requirement that schedules specify both budgetary outlay and quantity? We think not. In Chile–Price Band and Argentina–Footwear, there was an objective benchmark already in the defending Member's schedule against which it was possible to assess whether the obligation of result had been met, namely a tariff ceiling expressed in ad valorem terms. In the case of export subsidies under the AoA, determining whether the underlying obligation of result has been met, i.e. whether a certain percentage reduction in both budgetary outlay and quantity has been achieved, is impossible unless the Member in question specifies at a minimum the base outlays and quantities against which the percentage reductions can be measured over the specified time period. The fact is that transparency and specificity in scheduling and reporting are essential to legal security in a regime such as this, where Members are required to improve market access through the adjustment of labyrinthine and often opaque (in design and/or administration) domestic policy schemes.

In any event, we see the other interpretations of the AoA made by the AB in this part of its report as straightforward and logical, and not meriting specific commentary. Certainly, we agree with the AB that the use of the word ‘limiting’ does not necessarily imply that, in the absence of operative language, the AoA entertains the possibility of Members discharging their obligations under 9.1 through limits other than, or not including, the manifestly required reduction commitments.

Can a WTO Member limit or reduce its obligations under the AoA through scheduling?

The EC argued that, even if Footnote 1 was inconsistent with the AoA, by the acceptance of its schedule in the Uruguay Round negotiations, the EC had effectively altered its obligations under the AoA, or that the schedule somehow had an equal legal status to the AoA itself such that the obligations of the EC under the AoA should be read such that they were consistent with the rights reserved under the schedule. The AB rejected this argument. It observed:

Article 8 of the Agreement on Agriculture, entitled ‘Export Competition Commitments’, reads: Each Member undertakes not to provide export subsidies otherwise than in conformity with this Agreement and with the commitments as specified in that Member's Schedule. It is clear from the plain wording of Article 8 that Members are prohibited from providing export subsidies otherwise than in conformity with the Agreement on Agriculture and the commitments as specified in their Schedules. Thus, compliance with both is obligatory. As compliance with the provisions of the Agreement on Agriculture is obligatory, it is clear that the commitments specified in a Member's Schedule must be in conformity with the provisions of the Agreement. Only then would the export subsidies be in compliance with the requirements of Article 8. (paras. 215–216)

The EC had argued that the notion that scheduling cannot be used to limit or qualify obligations under a WTO treaty was derived in an unwarranted manner from the ruling in an old GATT case, coincidentally also concerning sugar. Thus, according to the EC, the panel in that case had depended on specific language in the GATT that addressed the possibility of qualifying obligations that were specific to the GATT regime, and the reasoning in question could not be transferred mutatis mutandis to the interpretation of the AoA.

The AB disagreed with the EC, finding that the panel had articulated a more general jurisprudential notion:

The GATT panel in US – Sugar did not rely solely on the language of Article II:1(b) of the GATT 1947 in making its ruling, as the European Communities suggests. Instead, the panel's reasoning was that, in the absence of a specific provision that would entitle Members to depart from their obligations under the GATT 1947, Members were not entitled to do so. Thus, the GATT panel in US – Sugar concluded: Article II:1(b) does not permit contracting parties to qualify their obligations under other provisions of the General Agreement and that the provisions in the United States GATT Schedule of Concessions can consequently not justify the maintenance of quantitative restrictions … inconsistent with the application of Article XI:1. Similarly, in this case, we find no provision under the Agreement on Agriculture that authorizes Members to depart, in their Schedules, from their obligations under that Agreement. Indeed, as we have noted, Article 8 requires that, in providing export subsidies, Members must comply with the provisions of both the Agreement on Agriculture and the export subsidy commitments specified in their Schedules. This is possible only if the commitments in the Schedules are in conformity with the provisions of the Agreement on Agriculture. Thus, we see no basis for the European Communities' assertion that it could depart from the obligations under the Agreement on Agriculture through the claimed commitment provided in Footnote 1. (paras. 219–220)

Finally, the AB rejected the EC argument that its schedule was equal in hierarchy of legal norms with the AoA itself and thus that the obligations of the AoA should be interpreted in a manner consistent with the schedule. The AB noted:

that Article 21 of the Agreement on Agriculture provides that: ‘[t]he provisions of [the] GATT 1994 and of other Multilateral Trade Agreements in Annex 1A to the WTO Agreement shall apply subject to the provisions of this Agreement.’ In other words, Members explicitly recognized that there may be conflicts between the Agreement on Agriculture and the GATT 1994, and explicitly provided, through Article 21, that the Agreement on Agriculture would prevail to the extent of such conflicts. Similarly, the General interpretative note to Annex 1A to the WTO Agreement states that, ‘[i]n the event of conflict between a provision of the [GATT 1994] and a provision of another agreement in Annex 1A …, the provision of the other agreement shall prevail to the extent of the conflict.’ The Agreement on Agriculture is contained in Annex 1A to the WTO Agreement.

The AB then observed:

As a separate matter, we note that the European Communities asserts that Footnote 1 was ‘negotiated’ with its partners in the Uruguay Round negotiations and that it has been ‘respected’. Accordingly, Footnote 1 forms part of the treaty ratified by the WTO Members. Similarly, the ACP Countries allege that Footnote 1 ‘was negotiated and agreed upon’ or acquiesced in by the Complaining Parties before the end of the Uruguay Round. The Panel found, however, that ‘[t]he evidence and submissions produced by all parties show that the Complainants did not agree to any European Communities’ deviations from the Agreement on Agriculture. The Panel concluded that ‘participants in the Uruguay Round and WTO Members did not agree to the European Communities’ inclusion of Footnote 1 as an agreed departure from the European Communities' basic obligations under the Agreement on Agriculture.' Accordingly, we see no basis in the Panel Reports for the contention of the European Communities and the ACP Countries that the Complaining Parties or the WTO Members negotiated or agreed to Footnote 1 as a departure from the European Communities' obligations under the Agreement on Agriculture


According to Article 19 of the VCLT, ‘a State may, when signing, ratifying, accepting, approving or acceding to a treaty, formulate a reservation unless: (a) the reservation is prohibited by the treaty; (b) the treaty provides that only specified reservations, which do not include the reservation in question, may be made; or (c) in cases not falling under subparagraphs (a) and (b), the reservation is incompatible with the object and purpose of the treaty’. The findings of the US–Sugar Panel did not apparently take into account this provision of the VCLT and indeed seem to be inconsistent with it, in as much as the Panel declared an opposite default rule, namely that no reservations are permitted from a treaty unless the text explicitly allows them. Nevertheless, the default rule of the US–Sugar Panel has now been incorporated into the WTO system as a default rule, by virtue of the WTO Agreement Article XVI:5, which provides that ‘Reservations in respect of any of the provisions of the Multilateral Trade Agreements may only be made to the extent provided for in those Agreements’. By virtue of VCLT 19(b), this lex specialis must prevail.

Moreover, the WTO Agreement elaborates a very specific detailed procedure that must be followed where a WTO obligation is to be waived for a particular WTO Member (Article IX:3–5). Even if it existed as the EC claimed, acquiescence or casual agreement among negotiators to Footnote 1 as a departure from the obligations of the AoA could hardly substitute for the waiver procedure outlined in Article IX of the WTO Agreement, and which is in any event only available in ‘exceptional circumstances’. Finally, there is an issue as to whether Article 41 of the VCLT, which deals with modifications of treaties that change obligations only between certain parties, would apply to a situation where all of the parties to a multilateral treaty modify the obligations of a single party towards all of them (the situation of waiver). However, in any case, Article 41 on its terms defers to any limitations or procedures in the treaty itself concerning such modifications, and this brings us back to the lex specialis of waivers in Article IX of the WTO Agreement.

The above analysis clearly supports the result of the AB ruling. We find it odd, however, in terms of the hierarchy of sources of law in the WTO system, that the AB would rely on a GATT panel ruling rather than on the provisions of the WTO Agreement, the foundational charter of the WTO legal system.

Estoppel and good faith

The EC claimed that the Complaining Parties acted inconsistently with Article 3.10 of the DSU and the principle of good faith by exercising their rights under the DSU in an ‘unreasonable’ and ‘abusive manner’ by ‘seeking to exploit what would be, at most, an excusable scheduling error in order to secure a manifestly unfair advantage’, and further that they were estopped from bringing DSU proceedings because they, along with all the other participants in the Uruguay Round negotiations, had given the EC to understand that they did not consider C sugar to benefit from export subsidies; here, the EC pointed to the nonreaction of Uruguay Round participants when the EC failed to include C sugar in its base levels. The Panel found that this mere nonreaction was insufficient to establish the kind of representation upon which an estoppel might be basedFootnote 12 and that there was no direct evidence of an actual shared understanding by Uruguay Round participants concerning C sugar. But the Panel also expressed serious doubts as to whether the principle of estoppel was applicable to WTO disputes.

In addressing the issue of estoppel, the Appellate Body echoed the uncertainty of the Panel as to the applicability of the principle of estoppel to WTO dispute-settlement proceedings. The AB opined:

The principle of estoppel has never been applied by the Appellate Body. Moreover, the notion of estoppel, as advanced by the European Communities, would appear to inhibit the ability of WTO Members to initiate a WTO dispute settlement proceeding. We see little in the DSU that explicitly limits the rights of WTO Members to bring an action; WTO Members must exercise their ‘judgement as to whether action under these procedures would be fruitful’, by virtue of Article 3.7 of the DSU, and they must engage in dispute settlement procedures in good faith, by virtue of Article 3.10 of the DSU. This latter obligation covers, in our view, the entire spectrum of dispute settlement, from the point of initiation of a case through implementation. Thus, even assuming arguendo that the principle of estoppel could apply in the WTO, its application would fall within these narrow parameters set out in the DSU. (para. 311)

The AB went on, however, to make no finding of law concerning the applicability of estoppel in WTO proceedings; it held that even assuming applicability along the lines that the EC was arguing, the Panel had determined that the EC had not proven the facts necessary to establish an estoppel.


The AB's approach here is incoherent. It is impossible to establish what facts would be needed to prove an estoppel without examining the nature and scope of the principle itself, as circumscribed or limited (if at all) by the WTO treaty texts. For example, if acquiescence or silence can be a basis for estoppel, it is far from clear that the findings of the Panel concerning the lack of proof of explicit statements or representations or the lack of proof of a ‘shared understanding’ would be enough to reject a claim of estoppel.

The legal authorities are divided as to whether estoppel includes acquiescence upon which a state may reasonably rely or whether estoppel and acquiescence are separate equitable grounds for refusing a legal claim.Footnote 13 In the Guatemala–Cement case, the panel contrasted estoppel with a situation of ‘acquiescence’, which meant ‘silence in the face of events that call for a reaction of some sort’(para. 8.23). In the Temple of Preah Vihear case, the International Court of Justice found that Thailand, in not objecting to a map filed as an annex to an agreement with Cambodia, could not claim that it had sovereignty over a temple that was drawn on the map as within Cambodian territory. The ICJ also noted that in the 50 years or more since the map was filed, Thailand's subsequent conduct was consistent with the view that the temple was within Cambodian jurisdiction.Footnote 14 In his separate opinion in the case, Judge Fitzmaurice appeared to take the view that acquiescence could be understood as one basis for estoppel rather than a separate legal principle:

The principle of preclusion is the nearest equivalent in the field of international law to the common-law rule of estoppel, though perhaps not applied under such strict limiting conditions (and it is certainly applied as a rule of substance and not merely as one of evidence or procedure). It is quite distinct theoretically from the notion of acquiescence. But acquiescence can operate as a preclusion or estoppel in certain cases, for instance where silence, on an occasion where there was a duty or need to speak or act, implies agreement, or a waiver of rights, and can be regarded as a representation to that effect.

Victor Rodríguez Cedeño, Special Rapporteur to the International Law Commission, noted in 2005 that in the Preah Vihear case, ‘in the reasoning of the Court the elements of estoppel and acquiescence are combined’.Footnote 15

The correct way for the AB to proceed in considering the EC's estoppel claim was to begin by considering first the nature and scope of estoppel as a general principle of international law, and, secondly, whether the provisions of the WTO treaties have altered or narrowed its applicability in the WTO. Only then could the AB properly conclude that the facts, as determined by the Panel, did not support a valid claim of estoppel.

In the passage cited above, without first considering the nature and scope of estoppel as a general principle, the AB nevertheless suggests, in dicta, that its application is somehow narrowed by the terms of the DSU. The notion seems to be that DSU 3.7 and 3.10 somehow replace and circumscribe the broader concept of equity in general international law.

It is true that 3.10 expresses one element of equity, namely ‘good faith’. It is unclear why this should lead to the conclusion that 3.10 thereby narrows or limits the application of other principles of international law based on equity, such as estoppel. In the Shrimp–Turtle case, the AB had held that ‘The chapeau of Article XX is, in fact, but one expression of the principle of good faith. This principle, at once a general principle of law and a general principle of international law, controls the exercise of rights by states’ (para. 158). Notably, the AB did not conclude that by expressing one aspect of the principle of good faith in the chapeau, the WTO Membership had limited or replaced good faith as a general principle of international law with a narrower treaty obligation; on the contrary, the AB stated that its responsibility was to ‘interpret the language of the chapeau, seeking additional interpretative guidance, as appropriate, from the general principles of international law’(para. 158; emphasis added). Even earlier, in the EC–Meat Hormones Report, para. 124, the AB had found, although reflected in a specific provision in the SPS Agreement (5.7, which allows preliminary measures in the absence of sufficient scientific evidence), was not exhausted by 5.7; it could thus affect the application of other provisions of the SPS.

It is difficult to reconcile the approach taken to estoppel in EC–Sugar with that taken to good faith in US–Shrimp. Perhaps it is significant that the EC was invoking estoppel as a bar to legal proceedings in EC–Sugar. We do agree with the AB to the extent that it is saying that the invocation of general principles of international law as a bar to adjudicative jurisdiction can be problematic in a legal system such as the WTO where the bases for jurisdiction of the dispute-settlement organs are set out explicitly by treaty, and provide virtually no discretion to panels to decline to adjudicate a dispute on broad equitable considerations. However, the AB never clearly distinguishes the issue of the general applicability of the principle of estoppel in the WTO with the issue of whether estoppel can function as a jurisdictional bar (which is what the EC appeared to be arguing). Indeed, quite apart from the WTO context, it is far from clear that estoppel as a general principle of international law can or should be understood as a bar to jurisdiction rather than to the advancement of a specific legal claim.

As for the AB's treatment of the EC claim concerning ‘good faith’, it rather summarily upheld the Panel's rejection of that claim. This is not surprising, for the same considerations that were rejected as a foundation for the estoppel claim were the basis, it seems, for the (partly separate or additional) claim of a violation of ‘good faith’.

4. Judicial economy

In WTO dispute settlement, a situation often arises where a complainant makes claims under a variety of provisions in a covered agreement or under several covered agreements. Often, if the adjudicator finds a violation of one or more provisions of an agreement, they will not go on to consider the other claims, including those under a different covered agreement. This practice is known as judicial economy. The rationale behind judicial economy is that a finding of violation under one provision or one agreement is adequate to ‘resolve’ the dispute, in that it is sufficient to give rise to the remedy that the complainant is seeking, usually withdrawal of the measures complained of, or their modification in such a way as to permit the market access or the competitive conditions that are the complainant's underlying objectives. The Panel in EC–Sugar, having found that the EC payments complained of were in violation of the AoA, did not go on to consider the complainants' claims that provisions of the SCM Agreement had also been violated. This use of judicial economy raised a serious legal issue on appeal, however, because the SCM Agreement contains a remedy for illegal export subsidization that would not be provided by the DSU in the case of violations of the AoA. This special remedy is provided for in SCM 4.7: ‘If the measure in question is found to be a prohibited subsidy, the panel shall recommend that the subsidizing Member withdraw the subsidy without delay. In this regard, the panel shall specify in its recommendation the time-period within which the measure must be withdrawn.’

In rejecting the Panel's use of judicial economy as a legal error the AB held:

In this case, the Panel's findings under Articles 3 and 8 of the Agreement on Agriculture were not sufficient to ‘fully resolve’ the dispute. This is because, in declining to rule on the Complaining Parties' claims under Article 3 of the SCM Agreement, the Panel precluded the possibility of a remedy being made available to the Complaining Parties, pursuant to Article 4.7 of the SCM Agreement, in the event of the Panel finding in favour of the Complaining Parties with respect to their claims under Article 3 of the SCM Agreement. Moreover, in declining to rule on the Complaining Parties' claims under Article 3 of the SCM Agreement, the Panel failed to discharge its obligation under Article 11 of the DSU by failing to make ‘such other findings as will assist the DSB in making the recommendations or in giving the rulings provided for in the covered agreements, namely, a recommendation or ruling by the DSB pursuant to Article 4.7. This constitutes false judicial economy and legal error. (para. 335)

Having found the Panel in error, the AB nevertheless did not go on to complete the analysis, i.e. to consider the claims that the Panel ought to have considered under the AoA.Footnote 16 The AB gave the following reasons for refusing to complete the analysis:

In several previous disputes, the Appellate Body examined an issue ‘not specifically addressed by the panel, in order to complete the legal analysis and resolve the dispute between the parties’. However, the Appellate Body has declined to complete the legal analysis where ‘the factual findings of the panel and the undisputed facts in the panel record’ did not provide a sufficient basis for the legal analysis by the Appellate Body. Moreover, as Article 17.6 of the DSU limits appeals to ‘issues of law covered in the panel report and legal interpretations developed by the panel’, the Appellate Body has also previously declined to complete the legal analysis of a panel in circumstances where that would involve addressing claims ‘which the panel had not examined at all’. In addition, the Appellate Body has indicated that it may complete the analysis only if the provision that a panel has not examined is ‘closely related’ to a provision that the panel has examined, and that the two are ‘part of a logical continuum’.

Turning to the specific case before us, we note that the Complaining Parties argue that their claims under the SCM Agreement are closely related to their claims under the Agreement on Agriculture. We are not persuaded, however, that Articles 3, 8, and 9.1 of the Agreement on Agriculture, on the one hand, and Articles 3.1(a), 3.2, and items (a) and (d) of the Illustrative List of the SCM Agreement, on the other hand, are ‘closely related’, because the issues presented under the two Agreements are different in several respects.

Furthermore, in the instant case, we note that the Panel made reference to the limited arguments made by the Complaining Parties under the SCM Agreement: [T]he Complainants [had] not set forth their claims under Article 3 of the SCM Agreement in quite as clear and unambiguous a manner as under the Agreement on Agriculture. Rather, [they] focused on their claims under the Agreement on Agriculture.

Although, on appeal, the Complaining Parties did argue their claims under the SCM Agreement to some extent, they did not address, in a sufficient manner, the question whether Article 3 of the SCM Agreement applies to export subsidies listed in Article 9.1 of the Agreement on Agriculture that are provided to scheduled agricultural products in excess of a responding Member's commitment levels. We believe that, in the light of Article 21 of the Agreement on Agriculture and the chapeau of Article 3 of the SCM Agreement, the question of the applicability of the SCM Agreement to the export subsidies in this dispute raises a number of complex issues. We also consider that, in the absence of a full exploration of these issues, completing the analysis might affect the due process rights of the participants.

Moreover, we do not have the requisite factual findings to complete the legal analysis. In particular, we do not have sufficient facts before us, as would be necessary to specify the period of time for withdrawal, as required by Article 4.7 of the SCM Agreement. We note in this respect that, when specifying what period would represent ‘without delay’, panels have taken into account, inter alia, ‘the nature of the measures and the difficulties likely to be faced in implementing the recommendation’. Based on our reading of the Panel Reports and the Panel record, we fail to see any evidence therein regarding the nature of the measures that would be required to ‘withdraw’ the subsidy, which would permit us to make a recommendation under Article 4.7. Hence, even if we were able to examine the Complaining Parties' claims under the SCM Agreement and, even if we were to conclude that the SCM Agreement applies in the circumstances of this dispute and that the European Communities acted inconsistently with its obligations under the SCM Agreement, we would not necessarily be in a position to make a recommendation under Article 4.7 as to the time period for withdrawal of the subsidy.


The existence of an explicit discrete remedy in the SCM Agreement not provided in the AoA made this a relatively easy case for rejecting the use of judicial economy. However, the underlying issue of principle, whether adjudicating a given claim is necessary to fully or adequately resolve a dispute, given that other claims have already been found to issue in violations, is a much trickier one. There are many instances where the violation of a particular provision of a covered agreement will not necessarily trigger a remedy that will achieve the complainant's objectives. Take, for instance, the SPS Agreement; a complaining Member may be challenging a measure as a violation of the national treatment and MFN provisions of SPS, as not based on an adequate risk assessment, as well as not least-trade-restrictive (5.6). If a panel were, for instance, to stop at a finding that the defending Member's risk assessment was inadequate under 5.1, a sufficient remedy would be for the defending Member to conduct a competent risk assessment (although there is always the possibility that such an assessment would not provide a basis for the measures and they would have to be withdrawn, this would not necessarily by any means be the case). However, if the panel went on to make findings on the other claims under SPS this might well result in the defending Member being required to alter substantive aspects of the measure itself, or withdraw it altogether. More generally, different aspects of a measure may raise different barriers to trade, each of which is of importance to the complaining Member. If these different aspects violate different legal provisions of the covered agreements, then judicial economy will undermine the legal rights of the complaining Member, since a finding limited to one or other of the aspects will only result normally in a requirement that the Member address those aspects, not withdraw the measure entirely.

In sum, to honor truly the underlying principle deployed by the AB in its ruling on judicial economy, panels will have to be much more careful in their decisions concerning judicial economy. A further systemic consideration here is that, where there is an action under DSU 21.5 concerning the adequacy of the defending Member's implementation of a panel ruling, the 21.5 panel will normally consider the consistency of the implementing measures, not only with those provisions of the covered agreements with which the original measures were found to be noncompliant in the first panel ruling, but with any other provisions that the complaining Member wishes to claim under in the 21.5 proceeding (provided, of course, they are within the terms of reference as established by the request for a 21.5 panel (Canada – Aircraft 21.5 Report of the Appellate Body; US – Shrimp 21.5, Report of the Appellate Body)). This extends even to claims never made in the original proceedings and must surely include those that were fully argued but over which the original panel decided to exercise judicial economy. From this point of view, judicial economy may not be particularly ‘economic’ and only have the odd (and arguably undesirable) result of delaying the adjudication of certain claims to the implementation stage of the dispute.

5. Concluding remarks

Our analysis suggests that the arguments of the Panel regarding cross-subsidization can be defended on both legal and economic grounds. Indeed, the economic reasoning applied by the Panel is conservative in that it covers only a subset of the possible situations that can give rise to cross-subsidization in situations where only part of production benefits from support policies. Although the reasoning of the Panel – that support policies that create rents on a portion of a firm's output may allow for the cross-subsidization of exports – is quite general, the potential for similar arguments to be applied to instances involving nonagricultural products where governments have not made explicit export subsidy commitments is likely to be very limited. Arguments that it does not matter whether the cross-subsidies are in fact contingent on exports, but that what matters is whether de facto a policy generates incentives to cross-subsidize exports, must be limited to agricultural products for which export subsidy commitments have been made. Thus, this case does not open up possibilities for countries to argue for imposition of countervailing measures on the basis of alleged cross-subsidies, or to policies that generate rents on other crops or products than those that are exported by a producer (e.g., to multi-product firms). Nor can high border protection that is not associated with a formal support price system of the EC sugar type and export subsidy reduction commitments give rise to claims of illegal export subsidization.

While not relevant from a WTO-law perspective, an important aspect of this case, from an economic-development perspective, concerns the impact of the EC having to cut back exports significantly. Not only will domestic production have to fall substantially, the Panel/AB findings imply that the EC cannot re-export the amount of imported ACP sugar back on to the world market. To maintain the past levels of ACP imports, given a total export limit of 1.27 million tonnes, would require a very large cut in EC output. The needed cut is even larger if account is taken of the prospective additional sugar exports by LDCs to the EC under EBA. The 2006 CAP sugar reforms lowered intervention prices and will reduce output, and substantially diminish the rents available for ACP sugar producers. A number of ACP producers will be forced to exit sugar production because they are too high cost. This will give rise to adjustment costs in these countries. Although the CAP reform embodies significant adjustment assistance for affected EC producers, the amount allocated to ACP producers for this purpose is limited.

The fact that the EC did not ‘safeguard’ the ACP sugar export volume in the reform of its sugar regime is an exogenous shock for the ACP that they could not have foreseen, even though it had been clear for some time that the EC was likely to change the program in the context of its decision to negotiate reciprocal trade agreements to replace the Cotonou convention, and that there would also be some ‘erosion’ of rents as a result of the EBA initiative to grant duty-free, quota-free access to LDC exports of sugar as of 2009. Given that the source of the shock was the scheduling error (lack of scheduling of the ACP sugar quota), there would appear to be a case for the EC to make the ACP whole again.Footnote 17

We are grateful to Paola Conconi and participants at the American Law Institute meeting for helpful comments on an earlier draft. The views expressed are personal and should not be attributed to the World Bank.


1 The principle of estoppel is defined in Section 2 below.

2 Figures in this paragraph are from Mitchell (Reference Mitchell, Aksoy and Beghin2005).

3 However, there is some allowance for carry forward – up to 20% of a firm's A+B quota per year.

4 The producer price of A sugar would be equal to that of B sugar if there were no exports of B sugar.

5 The extent to which this is actually a physical re-export of the cane sugar imported, or constitutes exports of beet sugar is unknown. Some 98.5% of total EC output is from sugar beet, whereas ACP sugar imports are derived from sugar cane. If sugar is sugar, i.e. is a homogenous product, it does not matter. In this paper, we assume that sugar is homogenous.

6 The value of these subsidies was capped in the EC's WTO commitments at €499 million.

7 This description abstracts from various factors that will determine the incidence of any rents. In practice, the rents will be shared with importers/distributors in the case of ACP producers, and with processors/refiners in the case of EC producers.

8 We are grateful to Harry de Gorter (Cornell University) for sharing these graphs.

9 We stress that while it is certainly conceptually possible for producers not to use A+B rents to subsidize C production, and that this justifies the legal analysis that follows, the practical relevance of the argument for the case at hand is limited as it is premised on alternative uses of these excess profits/rents not generating higher returns, be it through investment or consumption.

10 In US–Shrimp, the Appellate Body held that the ‘sprit’ of transparency, as reflected in the obligations in Article X of the GATT, should inform interpretation of other WTO provisions, especially where a WTO Member is claiming ‘exception.’ (paras. 182–183). While the ‘letter’ of Article X applies to domestic regulations and their administration, and the schedule is an international legal instrument, the schedule provides the basis on which economic actors from other WTO Members can form expectations concerning the terms and conditions of competition with the scheduling Member; therefore the ‘spirit’ of Article X is definitely engaged where a Member schedules so as to express in a nontransparent manner a limitation or exception to what would normally be expected to be its liberalization obligations.

11 The third-participant ACP countries in this appeal are: Barbados, Belize, Côte d'Ivoire, Fiji, Guyana, Jamaica, Kenya, Madagascar, Malawi, Mauritius, St. Kitts & Nevis, Swaziland, Tanzania, and Trinidad & Tobago.

12 Estoppel refers to the notion that some kinds of representations and conduct are such that the party responsible for them may be precluded from repudiating them, meaning often that the party in question may have legal responsibility for the reliance of other parties on the representations and conduct. Brownlie (Reference Brownlie2003), following Bowett, has suggested three conditions that must be fulfilled to establish an estoppel: ‘(1) a statement of fact that is clear and unambiguous; (2) this statement must be voluntary, unconditional, and authorized; and (3) there must be reliance in good faith upon the statement either to the detriment of the party so relying on the statement or to the advantage of the party making the statement.’ In the Argentina – Poultry case, Argentina unsuccessfully invoked this definition of estoppel to argue that by commencing proceedings in MERCOSUR in relation to the same dispute (on antidumping duties), Brazil was estopped from bringing a further action in the WTO; while finding that the conditions stated by Brownlie had not been met in that case, the Panel seemed open to the view that this definition could be an adequate statement of the definition of estoppel as a general principle of law.

13 See Sinclair (Reference Sinclair1995, p. 110).

14 Case concerning the Temple of Preah Vihear, Merits, Judgment of 15 June 1962: I.C.J. Reports 1962, p. 32.

15 International Law Commission (2005, para. 161).

16 In domestic legal systems, where an appellate tribunal has found errors of law in the judgment of a trial court, it may remand the matter to the trial level so that the trial court can apply the law as clarified by the appellate tribunal to the facts of the dispute. Under the DSU, the AB does not possess a comparable authority to remand a matter to the original panel. In these circumstances, the AB has engaged in a practice it calls ‘completing the analysis.’ This means that, having found errors of law in the panel ruling, the AB will not simply leave matters at reversing or modifying the panel's legal findings, but will proceed to apply the law as the AB has clarified it to the existing factual record of the case, as developed in the panel proceeding.

17 From a global-welfare perspective, the appropriate response is arguably to use non-trade instruments to provide financial assistance to the ACP, i.e. to offset the loss created by the Panel/AB decision.


Brownlie, Ian (2003), Principles of Public International Law, sixth edition, Oxford: Oxford University Press.Google Scholar
Fransden, Soren, Hans, Jensen, Wusheng, Yu, and Aage, Walter-Jorgensen (2003), ‘Reform of EU Sugar Policy: Price Cuts Versus Quota Reductions’, European Review of Agricultural Economics, 30(1): 126.Google Scholar
Gohin, Alexandre and Jean-Christophe, Bureau (2006), ‘Modeling the EU Sugar Supply to Assess Sectoral Policy Reforms’, European Review of Agricultural Economics, 33(2): 223247.CrossRefGoogle Scholar
De Gorter, Harry, David, Just, and Jaclyn, Kropp (2008), ‘Cross-subsidization due to Infra-Marginal Support in Agriculture: A General Theory and Empirical Evidence’, American Journal of Agricultural Economics, 90(1): 4254.CrossRefGoogle Scholar
International Law Commission (2005), ‘Eighth Report on unilateral acts of States’, 57th session, A/CN.4/557.Google Scholar
Mitchell, Donald (2005), ‘Sugar Policies: An Opportunity for Change’, in Aksoy, M. and Beghin, J. (eds.) Global Agricultural Trade and Developing Countries, Washington, DC: The World Bank.Google Scholar
Sinclair, Ian (1995), ‘Estoppel and Acquiescence’, Essays in Honour of Sir Robert Jennings, Cambridge: Cambridge University Press.Google Scholar
Figure 0

Figure 1.

Figure 1

Figure 2.

Figure 2

Figure 3. Positive profits at B, cross-subsidization to Q* (P*>AVC*)

Figure 3

Figure 4. Positive profits at B, possible cross-subsidization to Q* (P*<AVC*)

Figure 4

Figure 5. Negative profits at B, possible cross-subsidization to Q* (P*>AVC*)

Figure 5

Figure 6. Negative profits at B, possible cross-subsidization to Q* (P*<AVC*)

You have Access
Cited by

Save article to Kindle

To save this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the or variations. ‘’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats

Save article to Dropbox

To save this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about saving content to Dropbox.

Available formats

Save article to Google Drive

To save this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about saving content to Google Drive.

Available formats

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *